diff --git a/.gitignore b/.gitignore index d1c8be9..3c16adc 100644 --- a/.gitignore +++ b/.gitignore @@ -3,8 +3,10 @@ /passt.avx2 /pasta /pasta.avx2 +/passt-repair /qrap /pasta.1 /seccomp.h +/seccomp_repair.h /c*.json README.plain.md diff --git a/Makefile b/Makefile index 258d298..3328f83 100644 --- a/Makefile +++ b/Makefile @@ -16,9 +16,12 @@ VERSION ?= $(shell git describe --tags HEAD 2>/dev/null || echo "unknown\ versio DUAL_STACK_SOCKETS := 1 TARGET ?= $(shell $(CC) -dumpmachine) +$(if $(TARGET),,$(error Failed to get target architecture)) # Get 'uname -m'-like architecture description for target -TARGET_ARCH := $(shell echo $(TARGET) | cut -f1 -d- | tr [A-Z] [a-z]) -TARGET_ARCH := $(shell echo $(TARGET_ARCH) | sed 's/powerpc/ppc/') +TARGET_ARCH := $(firstword $(subst -, ,$(TARGET))) +TARGET_ARCH := $(patsubst [:upper:],[:lower:],$(TARGET_ARCH)) +TARGET_ARCH := $(patsubst arm%,arm,$(TARGET_ARCH)) +TARGET_ARCH := $(subst powerpc,ppc,$(TARGET_ARCH)) # On some systems enabling optimization also enables source fortification, # automagically. Do not override it. @@ -27,7 +30,7 @@ ifeq ($(shell $(CC) -O2 -dM -E - < /dev/null 2>&1 | grep ' _FORTIFY_SOURCE ' > / FORTIFY_FLAG := -D_FORTIFY_SOURCE=2 endif -FLAGS := -Wall -Wextra -Wno-format-zero-length +FLAGS := -Wall -Wextra -Wno-format-zero-length -Wformat-security FLAGS += -pedantic -std=c11 -D_XOPEN_SOURCE=700 -D_GNU_SOURCE FLAGS += $(FORTIFY_FLAG) -O2 -pie -fPIE FLAGS += -DPAGE_SIZE=$(shell getconf PAGE_SIZE) @@ -36,18 +39,21 @@ FLAGS += -DDUAL_STACK_SOCKETS=$(DUAL_STACK_SOCKETS) PASST_SRCS = arch.c arp.c checksum.c conf.c dhcp.c dhcpv6.c flow.c fwd.c \ icmp.c igmp.c inany.c iov.c ip.c isolation.c lineread.c log.c mld.c \ - ndp.c netlink.c packet.c passt.c pasta.c pcap.c pif.c tap.c tcp.c \ - tcp_buf.c tcp_splice.c udp.c udp_flow.c util.c + ndp.c netlink.c migrate.c packet.c passt.c pasta.c pcap.c pif.c \ + repair.c tap.c tcp.c tcp_buf.c tcp_splice.c tcp_vu.c udp.c udp_flow.c \ + udp_vu.c util.c vhost_user.c virtio.c vu_common.c QRAP_SRCS = qrap.c -SRCS = $(PASST_SRCS) $(QRAP_SRCS) +PASST_REPAIR_SRCS = passt-repair.c +SRCS = $(PASST_SRCS) $(QRAP_SRCS) $(PASST_REPAIR_SRCS) -MANPAGES = passt.1 pasta.1 qrap.1 +MANPAGES = passt.1 pasta.1 qrap.1 passt-repair.1 PASST_HEADERS = arch.h arp.h checksum.h conf.h dhcp.h dhcpv6.h flow.h fwd.h \ flow_table.h icmp.h icmp_flow.h inany.h iov.h ip.h isolation.h \ - lineread.h log.h ndp.h netlink.h packet.h passt.h pasta.h pcap.h pif.h \ - siphash.h tap.h tcp.h tcp_buf.h tcp_conn.h tcp_internal.h tcp_splice.h \ - udp.h udp_flow.h util.h + lineread.h log.h migrate.h ndp.h netlink.h packet.h passt.h pasta.h \ + pcap.h pif.h repair.h siphash.h tap.h tcp.h tcp_buf.h tcp_conn.h \ + tcp_internal.h tcp_splice.h tcp_vu.h udp.h udp_flow.h udp_internal.h \ + udp_vu.h util.h vhost_user.h virtio.h vu_common.h HEADERS = $(PASST_HEADERS) seccomp.h C := \#include <sys/random.h>\nint main(){int a=getrandom(0, 0, 0);} @@ -68,9 +74,9 @@ mandir ?= $(datarootdir)/man man1dir ?= $(mandir)/man1 ifeq ($(TARGET_ARCH),x86_64) -BIN := passt passt.avx2 pasta pasta.avx2 qrap +BIN := passt passt.avx2 pasta pasta.avx2 qrap passt-repair else -BIN := passt pasta qrap +BIN := passt pasta qrap passt-repair endif all: $(BIN) $(MANPAGES) docs @@ -79,7 +85,10 @@ static: FLAGS += -static -DGLIBC_NO_STATIC_NSS static: clean all seccomp.h: seccomp.sh $(PASST_SRCS) $(PASST_HEADERS) - @ EXTRA_SYSCALLS="$(EXTRA_SYSCALLS)" ARCH="$(TARGET_ARCH)" CC="$(CC)" ./seccomp.sh $(PASST_SRCS) $(PASST_HEADERS) + @ EXTRA_SYSCALLS="$(EXTRA_SYSCALLS)" ARCH="$(TARGET_ARCH)" CC="$(CC)" ./seccomp.sh seccomp.h $(PASST_SRCS) $(PASST_HEADERS) + +seccomp_repair.h: seccomp.sh $(PASST_REPAIR_SRCS) + @ ARCH="$(TARGET_ARCH)" CC="$(CC)" ./seccomp.sh seccomp_repair.h $(PASST_REPAIR_SRCS) passt: $(PASST_SRCS) $(HEADERS) $(CC) $(FLAGS) $(CFLAGS) $(CPPFLAGS) $(PASST_SRCS) -o passt $(LDFLAGS) @@ -97,15 +106,19 @@ pasta.avx2 pasta.1 pasta: pasta%: passt% qrap: $(QRAP_SRCS) passt.h $(CC) $(FLAGS) $(CFLAGS) $(CPPFLAGS) -DARCH=\"$(TARGET_ARCH)\" $(QRAP_SRCS) -o qrap $(LDFLAGS) +passt-repair: $(PASST_REPAIR_SRCS) seccomp_repair.h + $(CC) $(FLAGS) $(CFLAGS) $(CPPFLAGS) $(PASST_REPAIR_SRCS) -o passt-repair $(LDFLAGS) + valgrind: EXTRA_SYSCALLS += rt_sigprocmask rt_sigtimedwait rt_sigaction \ - rt_sigreturn getpid gettid kill clock_gettime mmap \ - mmap2 munmap open unlink gettimeofday futex + rt_sigreturn getpid gettid kill clock_gettime \ + mmap|mmap2 munmap open unlink gettimeofday futex \ + statx readlink valgrind: FLAGS += -g -DVALGRIND valgrind: all .PHONY: clean clean: - $(RM) $(BIN) *~ *.o seccomp.h pasta.1 \ + $(RM) $(BIN) *~ *.o seccomp.h seccomp_repair.h pasta.1 \ passt.tar passt.tar.gz *.deb *.rpm \ passt.pid README.plain.md diff --git a/README.md b/README.md index 752e59f..54fed07 100644 --- a/README.md +++ b/README.md @@ -321,7 +321,7 @@ speeding up local connections, and usually requiring NAT. _pasta_: protocol * ✅ 4 to 50 times IPv4 TCP throughput of existing, conceptually similar solutions depending on MTU (UDP and IPv6 hard to compare) -* 🛠 [_vhost-user_ support](https://bugs.passt.top/show_bug.cgi?id=25) for +* ✅ [_vhost-user_ support](https://bugs.passt.top/show_bug.cgi?id=25) for maximum one copy on every data path and lower request-response latency * ⌚ [multithreading](https://bugs.passt.top/show_bug.cgi?id=13) * ⌚ [raw IP socket support](https://bugs.passt.top/show_bug.cgi?id=14) if diff --git a/checksum.c b/checksum.c index c673993..0894eca 100644 --- a/checksum.c +++ b/checksum.c @@ -85,7 +85,7 @@ */ /* NOLINTNEXTLINE(clang-diagnostic-unknown-attributes) */ __attribute__((optimize("-fno-strict-aliasing"))) -uint32_t sum_16b(const void *buf, size_t len) +static uint32_t sum_16b(const void *buf, size_t len) { const uint16_t *p = buf; uint32_t sum = 0; @@ -107,7 +107,7 @@ uint32_t sum_16b(const void *buf, size_t len) * * Return: 16-bit folded sum */ -uint16_t csum_fold(uint32_t sum) +static uint16_t csum_fold(uint32_t sum) { while (sum >> 16) sum = (sum & 0xffff) + (sum >> 16); @@ -161,29 +161,42 @@ uint32_t proto_ipv4_header_psum(uint16_t l4len, uint8_t protocol, return psum; } +/** + * csum() - Compute TCP/IP-style checksum + * @buf: Input buffer + * @len: Input length + * @init: Initial 32-bit checksum, 0 for no pre-computed checksum + * + * Return: 16-bit folded, complemented checksum + */ +/* NOLINTNEXTLINE(clang-diagnostic-unknown-attributes) */ +__attribute__((optimize("-fno-strict-aliasing"))) /* See csum_16b() */ +static uint16_t csum(const void *buf, size_t len, uint32_t init) +{ + return (uint16_t)~csum_fold(csum_unfolded(buf, len, init)); +} + /** * csum_udp4() - Calculate and set checksum for a UDP over IPv4 packet * @udp4hr: UDP header, initialised apart from checksum * @saddr: IPv4 source address * @daddr: IPv4 destination address - * @iov: Pointer to the array of IO vectors - * @iov_cnt: Length of the array - * @offset: UDP payload offset in the iovec array + * @data: UDP payload (as IO vector tail) */ void csum_udp4(struct udphdr *udp4hr, struct in_addr saddr, struct in_addr daddr, - const struct iovec *iov, int iov_cnt, size_t offset) + struct iov_tail *data) { /* UDP checksums are optional, so don't bother */ udp4hr->check = 0; if (UDP4_REAL_CHECKSUMS) { - uint16_t l4len = iov_size(iov, iov_cnt) - offset + - sizeof(struct udphdr); + uint16_t l4len = iov_tail_size(data) + sizeof(struct udphdr); uint32_t psum = proto_ipv4_header_psum(l4len, IPPROTO_UDP, saddr, daddr); + psum = csum_unfolded(udp4hr, sizeof(struct udphdr), psum); - udp4hr->check = csum_iov(iov, iov_cnt, offset, psum); + udp4hr->check = csum_iov_tail(data, psum); } } @@ -231,22 +244,20 @@ uint32_t proto_ipv6_header_psum(uint16_t payload_len, uint8_t protocol, * @udp6hr: UDP header, initialised apart from checksum * @saddr: Source address * @daddr: Destination address - * @iov: Pointer to the array of IO vectors - * @iov_cnt: Length of the array - * @offset: UDP payload offset in the iovec array + * @data: UDP payload (as IO vector tail) */ void csum_udp6(struct udphdr *udp6hr, const struct in6_addr *saddr, const struct in6_addr *daddr, - const struct iovec *iov, int iov_cnt, size_t offset) + struct iov_tail *data) { - uint16_t l4len = iov_size(iov, iov_cnt) - offset + - sizeof(struct udphdr); + uint16_t l4len = iov_tail_size(data) + sizeof(struct udphdr); uint32_t psum = proto_ipv6_header_psum(l4len, IPPROTO_UDP, saddr, daddr); + udp6hr->check = 0; psum = csum_unfolded(udp6hr, sizeof(struct udphdr), psum); - udp6hr->check = csum_iov(iov, iov_cnt, offset, psum); + udp6hr->check = csum_iov_tail(data, psum); } /** @@ -456,7 +467,8 @@ uint32_t csum_unfolded(const void *buf, size_t len, uint32_t init) intptr_t align = ROUND_UP((intptr_t)buf, sizeof(__m256i)); unsigned int pad = align - (intptr_t)buf; - if (len < pad) + /* Don't mix sum_16b() and csum_avx2() with odd padding lengths */ + if (pad & 1 || len < pad) pad = len; if (pad) @@ -486,46 +498,23 @@ uint32_t csum_unfolded(const void *buf, size_t len, uint32_t init) #endif /* !__AVX2__ */ /** - * csum() - Compute TCP/IP-style checksum - * @buf: Input buffer - * @len: Input length - * @init: Initial 32-bit checksum, 0 for no pre-computed checksum - * - * Return: 16-bit folded, complemented checksum - */ -/* NOLINTNEXTLINE(clang-diagnostic-unknown-attributes) */ -__attribute__((optimize("-fno-strict-aliasing"))) /* See csum_16b() */ -uint16_t csum(const void *buf, size_t len, uint32_t init) -{ - return (uint16_t)~csum_fold(csum_unfolded(buf, len, init)); -} - -/** - * csum_iov() - Calculates the unfolded checksum over an array of IO vectors - * - * @iov Pointer to the array of IO vectors - * @n Length of the array - * @offset: Offset of the data to checksum within the full data length + * csum_iov_tail() - Calculate unfolded checksum for the tail of an IO vector + * @tail: IO vector tail to checksum * @init Initial 32-bit checksum, 0 for no pre-computed checksum * * Return: 16-bit folded, complemented checksum */ -uint16_t csum_iov(const struct iovec *iov, size_t n, size_t offset, - uint32_t init) +uint16_t csum_iov_tail(struct iov_tail *tail, uint32_t init) { - unsigned int i; - size_t first; - - i = iov_skip_bytes(iov, n, offset, &first); - if (i >= n) - return (uint16_t)~csum_fold(init); - - init = csum_unfolded((char *)iov[i].iov_base + first, - iov[i].iov_len - first, init); - i++; - - for (; i < n; i++) - init = csum_unfolded(iov[i].iov_base, iov[i].iov_len, init); + if (iov_tail_prune(tail)) { + size_t i; + init = csum_unfolded((char *)tail->iov[0].iov_base + tail->off, + tail->iov[0].iov_len - tail->off, init); + for (i = 1; i < tail->cnt; i++) { + const struct iovec *iov = &tail->iov[i]; + init = csum_unfolded(iov->iov_base, iov->iov_len, init); + } + } return (uint16_t)~csum_fold(init); } diff --git a/checksum.h b/checksum.h index 31ba322..683a09b 100644 --- a/checksum.h +++ b/checksum.h @@ -9,9 +9,8 @@ struct udphdr; struct icmphdr; struct icmp6hdr; +struct iov_tail; -uint32_t sum_16b(const void *buf, size_t len); -uint16_t csum_fold(uint32_t sum); uint16_t csum_unaligned(const void *buf, size_t len, uint32_t init); uint16_t csum_ip4_header(uint16_t l3len, uint8_t protocol, struct in_addr saddr, struct in_addr daddr); @@ -19,20 +18,18 @@ uint32_t proto_ipv4_header_psum(uint16_t l4len, uint8_t protocol, struct in_addr saddr, struct in_addr daddr); void csum_udp4(struct udphdr *udp4hr, struct in_addr saddr, struct in_addr daddr, - const struct iovec *iov, int iov_cnt, size_t offset); + struct iov_tail *data); void csum_icmp4(struct icmphdr *icmp4hr, const void *payload, size_t dlen); uint32_t proto_ipv6_header_psum(uint16_t payload_len, uint8_t protocol, const struct in6_addr *saddr, const struct in6_addr *daddr); void csum_udp6(struct udphdr *udp6hr, const struct in6_addr *saddr, const struct in6_addr *daddr, - const struct iovec *iov, int iov_cnt, size_t offset); + struct iov_tail *data); void csum_icmp6(struct icmp6hdr *icmp6hr, const struct in6_addr *saddr, const struct in6_addr *daddr, const void *payload, size_t dlen); uint32_t csum_unfolded(const void *buf, size_t len, uint32_t init); -uint16_t csum(const void *buf, size_t len, uint32_t init); -uint16_t csum_iov(const struct iovec *iov, size_t n, size_t offset, - uint32_t init); +uint16_t csum_iov_tail(struct iov_tail *tail, uint32_t init); #endif /* CHECKSUM_H */ diff --git a/conf.c b/conf.c index c6bffc4..a6d7e22 100644 --- a/conf.c +++ b/conf.c @@ -16,6 +16,7 @@ #include <errno.h> #include <fcntl.h> #include <getopt.h> +#include <libgen.h> #include <string.h> #include <sched.h> #include <sys/types.h> @@ -45,6 +46,7 @@ #include "lineread.h" #include "isolation.h" #include "log.h" +#include "vhost_user.h" #define NETNS_RUN_DIR "/run/netns" @@ -122,6 +124,75 @@ static int parse_port_range(const char *s, char **endptr, return 0; } +/** + * conf_ports_range_except() - Set up forwarding for a range of ports minus a + * bitmap of exclusions + * @c: Execution context + * @optname: Short option name, t, T, u, or U + * @optarg: Option argument (port specification) + * @fwd: Pointer to @fwd_ports to be updated + * @addr: Listening address + * @ifname: Listening interface + * @first: First port to forward + * @last: Last port to forward + * @exclude: Bitmap of ports to exclude + * @to: Port to translate @first to when forwarding + * @weak: Ignore errors, as long as at least one port is mapped + */ +static void conf_ports_range_except(const struct ctx *c, char optname, + const char *optarg, struct fwd_ports *fwd, + const union inany_addr *addr, + const char *ifname, + uint16_t first, uint16_t last, + const uint8_t *exclude, uint16_t to, + bool weak) +{ + bool bound_one = false; + unsigned i; + int ret; + + if (first == 0) { + die("Can't forward port 0 for option '-%c %s'", + optname, optarg); + } + + for (i = first; i <= last; i++) { + if (bitmap_isset(exclude, i)) + continue; + + if (bitmap_isset(fwd->map, i)) { + warn( +"Altering mapping of already mapped port number: %s", optarg); + } + + bitmap_set(fwd->map, i); + fwd->delta[i] = to - first; + + if (optname == 't') + ret = tcp_sock_init(c, addr, ifname, i); + else if (optname == 'u') + ret = udp_sock_init(c, 0, addr, ifname, i); + else + /* No way to check in advance for -T and -U */ + ret = 0; + + if (ret == -ENFILE || ret == -EMFILE) { + die("Can't open enough sockets for port specifier: %s", + optarg); + } + + if (!ret) { + bound_one = true; + } else if (!weak) { + die("Failed to bind port %u (%s) for option '-%c %s'", + i, strerror_(-ret), optname, optarg); + } + } + + if (!bound_one) + die("Failed to bind any port for '-%c %s'", optname, optarg); +} + /** * conf_ports() - Parse port configuration options, initialise UDP/TCP sockets * @c: Execution context @@ -134,10 +205,9 @@ static void conf_ports(const struct ctx *c, char optname, const char *optarg, { union inany_addr addr_buf = inany_any6, *addr = &addr_buf; char buf[BUFSIZ], *spec, *ifname = NULL, *p; - bool exclude_only = true, bound_one = false; uint8_t exclude[PORT_BITMAP_SIZE] = { 0 }; + bool exclude_only = true; unsigned i; - int ret; if (!strcmp(optarg, "none")) { if (fwd->mode) @@ -172,32 +242,15 @@ static void conf_ports(const struct ctx *c, char optname, const char *optarg, fwd->mode = FWD_ALL; - /* Skip port 0. It has special meaning for many socket APIs, so - * trying to bind it is not really safe. - */ - for (i = 1; i < NUM_PORTS; i++) { + /* Exclude ephemeral ports */ + for (i = 0; i < NUM_PORTS; i++) if (fwd_port_is_ephemeral(i)) - continue; - - bitmap_set(fwd->map, i); - if (optname == 't') { - ret = tcp_sock_init(c, NULL, NULL, i); - if (ret == -ENFILE || ret == -EMFILE) - goto enfile; - if (!ret) - bound_one = true; - } else if (optname == 'u') { - ret = udp_sock_init(c, 0, NULL, NULL, i); - if (ret == -ENFILE || ret == -EMFILE) - goto enfile; - if (!ret) - bound_one = true; - } - } - - if (!bound_one) - goto bind_all_fail; + bitmap_set(exclude, i); + conf_ports_range_except(c, optname, optarg, fwd, + NULL, NULL, + 1, NUM_PORTS - 1, exclude, + 1, true); return; } @@ -274,37 +327,15 @@ static void conf_ports(const struct ctx *c, char optname, const char *optarg, } while ((p = next_chunk(p, ','))); if (exclude_only) { - /* Skip port 0. It has special meaning for many socket APIs, so - * trying to bind it is not really safe. - */ - for (i = 1; i < NUM_PORTS; i++) { - if (fwd_port_is_ephemeral(i) || - bitmap_isset(exclude, i)) - continue; - - bitmap_set(fwd->map, i); - - if (optname == 't') { - ret = tcp_sock_init(c, addr, ifname, i); - if (ret == -ENFILE || ret == -EMFILE) - goto enfile; - if (!ret) - bound_one = true; - } else if (optname == 'u') { - ret = udp_sock_init(c, 0, addr, ifname, i); - if (ret == -ENFILE || ret == -EMFILE) - goto enfile; - if (!ret) - bound_one = true; - } else { - /* No way to check in advance for -T and -U */ - bound_one = true; - } - } - - if (!bound_one) - goto bind_all_fail; + /* Exclude ephemeral ports */ + for (i = 0; i < NUM_PORTS; i++) + if (fwd_port_is_ephemeral(i)) + bitmap_set(exclude, i); + conf_ports_range_except(c, optname, optarg, fwd, + addr, ifname, + 1, NUM_PORTS - 1, exclude, + 1, true); return; } @@ -333,40 +364,18 @@ static void conf_ports(const struct ctx *c, char optname, const char *optarg, if ((*p != '\0') && (*p != ',')) /* Garbage after the ranges */ goto bad; - for (i = orig_range.first; i <= orig_range.last; i++) { - if (bitmap_isset(fwd->map, i)) - warn( -"Altering mapping of already mapped port number: %s", optarg); - - if (bitmap_isset(exclude, i)) - continue; - - bitmap_set(fwd->map, i); - - fwd->delta[i] = mapped_range.first - orig_range.first; - - ret = 0; - if (optname == 't') - ret = tcp_sock_init(c, addr, ifname, i); - else if (optname == 'u') - ret = udp_sock_init(c, 0, addr, ifname, i); - if (ret) - goto bind_fail; - } + conf_ports_range_except(c, optname, optarg, fwd, + addr, ifname, + orig_range.first, orig_range.last, + exclude, + mapped_range.first, false); } while ((p = next_chunk(p, ','))); return; -enfile: - die("Can't open enough sockets for port specifier: %s", optarg); bad: die("Invalid port specifier %s", optarg); mode_conflict: die("Port forwarding mode '%s' conflicts with previous mode", optarg); -bind_fail: - die("Failed to bind port %u (%s) for option '-%c %s', exiting", - i, strerror(-ret), optname, optarg); -bind_all_fail: - die("Failed to bind any port for '-%c %s', exiting", optname, optarg); } /** @@ -405,6 +414,76 @@ static unsigned add_dns6(struct ctx *c, const struct in6_addr *addr, return 1; } +/** + * add_dns_resolv4() - Possibly add one IPv4 nameserver from host's resolv.conf + * @c: Execution context + * @ns: Nameserver address + * @idx: Pointer to index of current IPv4 resolver entry, set on return + */ +static void add_dns_resolv4(struct ctx *c, struct in_addr *ns, unsigned *idx) +{ + if (IN4_IS_ADDR_UNSPECIFIED(&c->ip4.dns_host)) + c->ip4.dns_host = *ns; + + /* Special handling if guest or container can only access local + * addresses via redirect, or if the host gateway is also a resolver and + * we shadow its address + */ + if (IN4_IS_ADDR_LOOPBACK(ns) || + IN4_ARE_ADDR_EQUAL(ns, &c->ip4.map_host_loopback)) { + if (IN4_IS_ADDR_UNSPECIFIED(&c->ip4.dns_match)) { + if (IN4_IS_ADDR_UNSPECIFIED(&c->ip4.map_host_loopback)) + return; /* Address unreachable */ + + *ns = c->ip4.map_host_loopback; + c->ip4.dns_match = c->ip4.map_host_loopback; + } else { + /* No general host mapping, but requested for DNS + * (--dns-forward and --no-map-gw): advertise resolver + * address from --dns-forward, and map that to loopback + */ + *ns = c->ip4.dns_match; + } + } + + *idx += add_dns4(c, ns, *idx); +} + +/** + * add_dns_resolv6() - Possibly add one IPv6 nameserver from host's resolv.conf + * @c: Execution context + * @ns: Nameserver address + * @idx: Pointer to index of current IPv6 resolver entry, set on return + */ +static void add_dns_resolv6(struct ctx *c, struct in6_addr *ns, unsigned *idx) +{ + if (IN6_IS_ADDR_UNSPECIFIED(&c->ip6.dns_host)) + c->ip6.dns_host = *ns; + + /* Special handling if guest or container can only access local + * addresses via redirect, or if the host gateway is also a resolver and + * we shadow its address + */ + if (IN6_IS_ADDR_LOOPBACK(ns) || + IN6_ARE_ADDR_EQUAL(ns, &c->ip6.map_host_loopback)) { + if (IN6_IS_ADDR_UNSPECIFIED(&c->ip6.dns_match)) { + if (IN6_IS_ADDR_UNSPECIFIED(&c->ip6.map_host_loopback)) + return; /* Address unreachable */ + + *ns = c->ip6.map_host_loopback; + c->ip6.dns_match = c->ip6.map_host_loopback; + } else { + /* No general host mapping, but requested for DNS + * (--dns-forward and --no-map-gw): advertise resolver + * address from --dns-forward, and map that to loopback + */ + *ns = c->ip6.dns_match; + } + } + + *idx += add_dns6(c, ns, *idx); +} + /** * add_dns_resolv() - Possibly add ns from host resolv.conf to configuration * @c: Execution context @@ -421,44 +500,11 @@ static void add_dns_resolv(struct ctx *c, const char *nameserver, struct in6_addr ns6; struct in_addr ns4; - if (idx4 && inet_pton(AF_INET, nameserver, &ns4)) { - if (IN4_IS_ADDR_UNSPECIFIED(&c->ip4.dns_host)) - c->ip4.dns_host = ns4; + if (idx4 && inet_pton(AF_INET, nameserver, &ns4)) + add_dns_resolv4(c, &ns4, idx4); - /* Guest or container can only access local addresses via - * redirect - */ - if (IN4_IS_ADDR_LOOPBACK(&ns4)) { - if (IN4_IS_ADDR_UNSPECIFIED(&c->ip4.map_host_loopback)) - return; - - ns4 = c->ip4.map_host_loopback; - if (IN4_IS_ADDR_UNSPECIFIED(&c->ip4.dns_match)) - c->ip4.dns_match = c->ip4.map_host_loopback; - } - - *idx4 += add_dns4(c, &ns4, *idx4); - } - - if (idx6 && inet_pton(AF_INET6, nameserver, &ns6)) { - if (IN6_IS_ADDR_UNSPECIFIED(&c->ip6.dns_host)) - c->ip6.dns_host = ns6; - - /* Guest or container can only access local addresses via - * redirect - */ - if (IN6_IS_ADDR_LOOPBACK(&ns6)) { - if (IN6_IS_ADDR_UNSPECIFIED(&c->ip6.map_host_loopback)) - return; - - ns6 = c->ip6.map_host_loopback; - - if (IN6_IS_ADDR_UNSPECIFIED(&c->ip6.dns_match)) - c->ip6.dns_match = c->ip6.map_host_loopback; - } - - *idx6 += add_dns6(c, &ns6, *idx6); - } + if (idx6 && inet_pton(AF_INET6, nameserver, &ns6)) + add_dns_resolv6(c, &ns6, idx6); } /** @@ -654,7 +700,7 @@ static unsigned int conf_ip4(unsigned int ifi, struct ip4_ctx *ip4) &ip4->guest_gw); if (rc < 0) { debug("Couldn't discover IPv4 gateway address: %s", - strerror(-rc)); + strerror_(-rc)); return 0; } } @@ -664,7 +710,7 @@ static unsigned int conf_ip4(unsigned int ifi, struct ip4_ctx *ip4) &ip4->addr, &ip4->prefix_len, NULL); if (rc < 0) { debug("Couldn't discover IPv4 address: %s", - strerror(-rc)); + strerror_(-rc)); return 0; } } @@ -728,7 +774,7 @@ static unsigned int conf_ip6(unsigned int ifi, struct ip6_ctx *ip6) rc = nl_route_get_def(nl_sock, ifi, AF_INET6, &ip6->guest_gw); if (rc < 0) { debug("Couldn't discover IPv6 gateway address: %s", - strerror(-rc)); + strerror_(-rc)); return 0; } } @@ -737,7 +783,7 @@ static unsigned int conf_ip6(unsigned int ifi, struct ip6_ctx *ip6) IN6_IS_ADDR_UNSPECIFIED(&ip6->addr) ? &ip6->addr : NULL, &prefix_len, &ip6->our_tap_ll); if (rc < 0) { - debug("Couldn't discover IPv6 address: %s", strerror(-rc)); + debug("Couldn't discover IPv6 address: %s", strerror_(-rc)); return 0; } @@ -768,7 +814,7 @@ static void conf_ip6_local(struct ip6_ctx *ip6) * usage() - Print usage, exit with given status code * @name: Executable name * @f: Stream to print usage info to - * @status: Status code for exit() + * @status: Status code for _exit() */ static void usage(const char *name, FILE *f, int status) { @@ -807,9 +853,17 @@ static void usage(const char *name, FILE *f, int status) " default: same interface name as external one\n"); } else { FPRINTF(f, - " -s, --socket PATH UNIX domain socket path\n" + " -s, --socket, --socket-path PATH UNIX domain socket path\n" " default: probe free path starting from " UNIX_SOCK_PATH "\n", 1); + FPRINTF(f, + " --vhost-user Enable vhost-user mode\n" + " UNIX domain socket is provided by -s option\n" + " --print-capabilities print back-end capabilities in JSON format,\n" + " only meaningful for vhost-user mode\n"); + FPRINTF(f, + " --repair-path PATH path for passt-repair(1)\n" + " default: append '.repair' to UNIX domain path\n"); } FPRINTF(f, @@ -848,7 +902,9 @@ static void usage(const char *name, FILE *f, int status) FPRINTF(f, " default: use addresses from /etc/resolv.conf\n"); FPRINTF(f, " -S, --search LIST Space-separated list, search domains\n" - " a single, empty option disables the DNS search list\n"); + " a single, empty option disables the DNS search list\n" + " -H, --hostname NAME Hostname to configure client with\n" + " --fqdn NAME FQDN to configure client with\n"); if (strstr(name, "pasta")) FPRINTF(f, " default: don't use any search list\n"); else @@ -919,7 +975,7 @@ static void usage(const char *name, FILE *f, int status) " SPEC is as described for TCP above\n" " default: none\n"); - exit(status); + _exit(status); pasta_opts: @@ -957,8 +1013,7 @@ pasta_opts: " -U, --udp-ns SPEC UDP port forwarding to init namespace\n" " SPEC is as described above\n" " default: auto\n" - " --host-lo-to-ns-lo DEPRECATED:\n" - " Translate host-loopback forwards to\n" + " --host-lo-to-ns-lo Translate host-loopback forwards to\n" " namespace loopback\n" " --userns NSPATH Target user namespace to join\n" " --netns PATH|NAME Target network namespace to join\n" @@ -971,9 +1026,49 @@ pasta_opts: " Don't copy all routes to namespace\n" " --no-copy-addrs DEPRECATED:\n" " Don't copy all addresses to namespace\n" - " --ns-mac-addr ADDR Set MAC address on tap interface\n"); + " --ns-mac-addr ADDR Set MAC address on tap interface\n" + " --no-splice Disable inbound socket splicing\n"); - exit(status); + _exit(status); +} + +/** + * conf_mode() - Determine passt/pasta's operating mode from command line + * @argc: Argument count + * @argv: Command line arguments + * + * Return: mode to operate in, PASTA or PASST + */ +enum passt_modes conf_mode(int argc, char *argv[]) +{ + int vhost_user = 0; + const struct option optvu[] = { + {"vhost-user", no_argument, &vhost_user, 1 }, + { 0 }, + }; + char argv0[PATH_MAX], *basearg0; + int name; + + optind = 0; + do { + name = getopt_long(argc, argv, "-:", optvu, NULL); + } while (name != -1); + + if (vhost_user) + return MODE_VU; + + if (argc < 1) + die("Cannot determine argv[0]"); + + strncpy(argv0, argv[0], PATH_MAX - 1); + basearg0 = basename(argv0); + if (strstr(basearg0, "pasta")) + return MODE_PASTA; + + if (strstr(basearg0, "passt")) + return MODE_PASST; + + die("Cannot determine mode, invoke as \"passt\" or \"pasta\""); } /** @@ -1210,6 +1305,8 @@ static void conf_nat(const char *arg, struct in_addr *addr4, *addr6 = in6addr_any; if (no_map_gw) *no_map_gw = 1; + + return; } if (inet_pton(AF_INET6, arg, addr6) && @@ -1233,8 +1330,25 @@ static void conf_nat(const char *arg, struct in_addr *addr4, */ static void conf_open_files(struct ctx *c) { - if (c->mode != MODE_PASTA && c->fd_tap == -1) - c->fd_tap_listen = tap_sock_unix_open(c->sock_path); + if (c->mode != MODE_PASTA && c->fd_tap == -1) { + c->fd_tap_listen = sock_unix(c->sock_path); + + if (c->mode == MODE_VU && strcmp(c->repair_path, "none")) { + if (!*c->repair_path && + snprintf_check(c->repair_path, + sizeof(c->repair_path), "%s.repair", + c->sock_path)) { + warn("passt-repair path %s not usable", + c->repair_path); + c->fd_repair_listen = -1; + } else { + c->fd_repair_listen = sock_unix(c->repair_path); + } + } else { + c->fd_repair_listen = -1; + } + c->fd_repair = -1; + } if (*c->pidfile) { c->pidfile_fd = output_file_open(c->pidfile, O_WRONLY); @@ -1306,6 +1420,7 @@ void conf(struct ctx *c, int argc, char **argv) {"outbound", required_argument, NULL, 'o' }, {"dns", required_argument, NULL, 'D' }, {"search", required_argument, NULL, 'S' }, + {"hostname", required_argument, NULL, 'H' }, {"no-tcp", no_argument, &c->no_tcp, 1 }, {"no-udp", no_argument, &c->no_udp, 1 }, {"no-icmp", no_argument, &c->no_icmp, 1 }, @@ -1313,6 +1428,7 @@ void conf(struct ctx *c, int argc, char **argv) {"no-dhcpv6", no_argument, &c->no_dhcpv6, 1 }, {"no-ndp", no_argument, &c->no_ndp, 1 }, {"no-ra", no_argument, &c->no_ra, 1 }, + {"no-splice", no_argument, &c->no_splice, 1 }, {"freebind", no_argument, &c->freebind, 1 }, {"no-map-gw", no_argument, &no_map_gw, 1 }, {"ipv4-only", no_argument, NULL, '4' }, @@ -1345,18 +1461,26 @@ void conf(struct ctx *c, int argc, char **argv) {"map-guest-addr", required_argument, NULL, 22 }, {"host-lo-to-ns-lo", no_argument, NULL, 23 }, {"dns-host", required_argument, NULL, 24 }, + {"vhost-user", no_argument, NULL, 25 }, + + /* vhost-user backend program convention */ + {"print-capabilities", no_argument, NULL, 26 }, + {"socket-path", required_argument, NULL, 's' }, + {"fqdn", required_argument, NULL, 27 }, + {"repair-path", required_argument, NULL, 28 }, { 0 }, }; + const char *optstring = "+dqfel:hs:F:I:p:P:m:a:n:M:g:i:o:D:S:H:461t:u:T:U:"; const char *logname = (c->mode == MODE_PASTA) ? "pasta" : "passt"; char userns[PATH_MAX] = { 0 }, netns[PATH_MAX] = { 0 }; bool copy_addrs_opt = false, copy_routes_opt = false; enum fwd_ports_mode fwd_default = FWD_NONE; bool v4_only = false, v6_only = false; unsigned dns4_idx = 0, dns6_idx = 0; + unsigned long max_mtu = IP_MAX_MTU; struct fqdn *dnss = c->dns_search; unsigned int ifi4 = 0, ifi6 = 0; const char *logfile = NULL; - const char *optstring; size_t logsize = 0; char *runas = NULL; long fd_tap_opt; @@ -1367,11 +1491,11 @@ void conf(struct ctx *c, int argc, char **argv) if (c->mode == MODE_PASTA) { c->no_dhcp_dns = c->no_dhcp_dns_search = 1; fwd_default = FWD_AUTO; - optstring = "+dqfel:hF:I:p:P:m:a:n:M:g:i:o:D:S:46t:u:T:U:"; - } else { - optstring = "+dqfel:hs:F:p:P:m:a:n:M:g:i:o:D:S:461t:u:"; } + if (tap_l2_max_len(c) - ETH_HLEN < max_mtu) + max_mtu = tap_l2_max_len(c) - ETH_HLEN; + c->mtu = ROUND_DOWN(max_mtu, sizeof(uint32_t)); c->tcp.fwd_in.mode = c->tcp.fwd_out.mode = FWD_UNSET; c->udp.fwd_in.mode = c->udp.fwd_out.mode = FWD_UNSET; memcpy(c->our_tap_mac, MAC_OUR_LAA, ETH_ALEN); @@ -1470,7 +1594,7 @@ void conf(struct ctx *c, int argc, char **argv) FPRINTF(stdout, c->mode == MODE_PASTA ? "pasta " : "passt "); FPRINTF(stdout, VERSION_BLOB); - exit(EXIT_SUCCESS); + _exit(EXIT_SUCCESS); case 15: ret = snprintf(c->ip4.ifname_out, sizeof(c->ip4.ifname_out), "%s", optarg); @@ -1538,6 +1662,27 @@ void conf(struct ctx *c, int argc, char **argv) break; die("Invalid host nameserver address: %s", optarg); + case 25: + /* Already handled in conf_mode() */ + ASSERT(c->mode == MODE_VU); + break; + case 26: + vu_print_capabilities(); + break; + case 27: + if (snprintf_check(c->fqdn, PASST_MAXDNAME, + "%s", optarg)) + die("Invalid FQDN: %s", optarg); + break; + case 28: + if (c->mode != MODE_VU && strcmp(optarg, "none")) + die("--repair-path is for vhost-user mode only"); + + if (snprintf_check(c->repair_path, + sizeof(c->repair_path), "%s", + optarg)) + die("Invalid passt-repair path: %s", optarg); + break; case 'd': c->debug = 1; @@ -1557,6 +1702,9 @@ void conf(struct ctx *c, int argc, char **argv) c->foreground = 1; break; case 's': + if (c->mode == MODE_PASTA) + die("-s is for passt / vhost-user mode only"); + ret = snprintf(c->sock_path, sizeof(c->sock_path), "%s", optarg); if (ret <= 0 || ret >= (int)sizeof(c->sock_path)) @@ -1569,7 +1717,8 @@ void conf(struct ctx *c, int argc, char **argv) fd_tap_opt = strtol(optarg, NULL, 0); if (errno || - fd_tap_opt <= STDERR_FILENO || fd_tap_opt > INT_MAX) + (fd_tap_opt != STDIN_FILENO && fd_tap_opt <= STDERR_FILENO) || + fd_tap_opt > INT_MAX) die("Invalid --fd: %s", optarg); c->fd_tap = fd_tap_opt; @@ -1577,6 +1726,9 @@ void conf(struct ctx *c, int argc, char **argv) *c->sock_path = 0; break; case 'I': + if (c->mode != MODE_PASTA) + die("-I is for pasta mode only"); + ret = snprintf(c->pasta_ifn, IFNAMSIZ, "%s", optarg); if (ret <= 0 || ret >= IFNAMSIZ) @@ -1596,20 +1748,24 @@ void conf(struct ctx *c, int argc, char **argv) die("Invalid PID file: %s", optarg); break; - case 'm': + case 'm': { + unsigned long mtu; + char *e; + errno = 0; - c->mtu = strtol(optarg, NULL, 0); + mtu = strtoul(optarg, &e, 0); - if (!c->mtu) { - c->mtu = -1; - break; - } - - if (c->mtu < ETH_MIN_MTU || c->mtu > (int)ETH_MAX_MTU || - errno) + if (errno || *e) die("Invalid MTU: %s", optarg); + if (mtu > max_mtu) { + die("MTU %lu too large (max %lu)", + mtu, max_mtu); + } + + c->mtu = mtu; break; + } case 'a': if (inet_pton(AF_INET6, optarg, &c->ip6.addr) && !IN6_IS_ADDR_UNSPECIFIED(&c->ip6.addr) && @@ -1708,6 +1864,11 @@ void conf(struct ctx *c, int argc, char **argv) die("Cannot use DNS search domain %s", optarg); break; + case 'H': + if (snprintf_check(c->hostname, PASST_MAXDNAME, + "%s", optarg)) + die("Invalid hostname: %s", optarg); + break; case '4': v4_only = true; v6_only = false; @@ -1724,11 +1885,16 @@ void conf(struct ctx *c, int argc, char **argv) break; case 't': case 'u': - case 'T': - case 'U': case 'D': /* Handle these later, once addresses are configured */ break; + case 'T': + case 'U': + if (c->mode != MODE_PASTA) + die("-%c is for pasta mode only", name); + + /* Handle properly later, once addresses are configured */ + break; case 'h': usage(argv[0], stdout, EXIT_SUCCESS); break; @@ -1739,6 +1905,9 @@ void conf(struct ctx *c, int argc, char **argv) } } while (name != -1); + if (c->mode != MODE_PASTA) + c->no_splice = 1; + if (c->mode == MODE_PASTA && !c->pasta_conf_ns) { if (copy_routes_opt) die("--no-copy-routes needs --config-net"); @@ -1773,9 +1942,21 @@ void conf(struct ctx *c, int argc, char **argv) c->ifi4 = conf_ip4(ifi4, &c->ip4); if (!v4_only) c->ifi6 = conf_ip6(ifi6, &c->ip6); + + if (c->ifi4 && c->mtu < IPV4_MIN_MTU) { + warn("MTU %"PRIu16" is too small for IPv4 (minimum %u)", + c->mtu, IPV4_MIN_MTU); + } + if (c->ifi6 && c->mtu < IPV6_MIN_MTU) { + warn("MTU %"PRIu16" is too small for IPv6 (minimum %u)", + c->mtu, IPV6_MIN_MTU); + } + if ((*c->ip4.ifname_out && !c->ifi4) || (*c->ip6.ifname_out && !c->ifi6)) die("External interface not usable"); + + if (!c->ifi4 && !c->ifi6) { info("No external interface as template, switch to local mode"); @@ -1802,8 +1983,8 @@ void conf(struct ctx *c, int argc, char **argv) if (c->ifi4 && IN4_IS_ADDR_UNSPECIFIED(&c->ip4.guest_gw)) c->no_dhcp = 1; - /* Inbound port options & DNS can be parsed now (after IPv4/IPv6 - * settings) + /* Inbound port options and DNS can be parsed now, after IPv4/IPv6 + * settings */ fwd_probe_ephemeral(); udp_portmap_clear(); @@ -1897,9 +2078,6 @@ void conf(struct ctx *c, int argc, char **argv) c->no_dhcpv6 = 1; } - if (!c->mtu) - c->mtu = ROUND_DOWN(ETH_MAX_MTU - ETH_HLEN, sizeof(uint32_t)); - get_dns(c); if (!*c->pasta_ifn) { diff --git a/conf.h b/conf.h index 9d2143d..b45ad74 100644 --- a/conf.h +++ b/conf.h @@ -6,6 +6,7 @@ #ifndef CONF_H #define CONF_H +enum passt_modes conf_mode(int argc, char *argv[]); void conf(struct ctx *c, int argc, char **argv); #endif /* CONF_H */ diff --git a/contrib/apparmor/usr.bin.passt b/contrib/apparmor/usr.bin.passt index 9568189..62a4514 100644 --- a/contrib/apparmor/usr.bin.passt +++ b/contrib/apparmor/usr.bin.passt @@ -27,4 +27,25 @@ profile passt /usr/bin/passt{,.avx2} { owner @{HOME}/** w, # pcap(), pidfile_open(), # pidfile_write() + + # Workaround: libvirt's profile comes with a passt subprofile which includes, + # in turn, <abstractions/passt>, and adds libvirt-specific rules on top, to + # allow passt (when started by libvirtd) to write socket and PID files in the + # location requested by libvirtd itself, and to execute passt itself. + # + # However, when libvirt runs as unprivileged user, the mechanism based on + # virt-aa-helper, designed to build per-VM profiles as guests are started, + # doesn't work. The helper needs to create and load profiles on the fly, which + # can't be done by unprivileged users, of course. + # + # As a result, libvirtd runs unconfined if guests are started by unprivileged + # users, starting passt unconfined as well, which means that passt runs under + # its own stand-alone profile (this one), which implies in turn that execve() + # of /usr/bin/passt is not allowed, and socket and PID files can't be written. + # + # Duplicate libvirt-specific rules here as long as this is not solved in + # libvirt's profile itself. + /usr/bin/passt r, + owner @{run}/user/[0-9]*/libvirt/qemu/run/passt/* rw, + owner @{run}/libvirt/qemu/passt/* rw, } diff --git a/contrib/apparmor/usr.bin.passt-repair b/contrib/apparmor/usr.bin.passt-repair new file mode 100644 index 0000000..901189d --- /dev/null +++ b/contrib/apparmor/usr.bin.passt-repair @@ -0,0 +1,29 @@ +# SPDX-License-Identifier: GPL-2.0-or-later +# +# PASST - Plug A Simple Socket Transport +# for qemu/UNIX domain socket mode +# +# PASTA - Pack A Subtle Tap Abstraction +# for network namespace/tap device mode +# +# contrib/apparmor/usr.bin.passt-repair - AppArmor profile for passt-repair(1) +# +# Copyright (c) 2025 Red Hat GmbH +# Author: Stefano Brivio <sbrivio@redhat.com> + +abi <abi/3.0>, + +#include <tunables/global> + +profile passt-repair /usr/bin/passt-repair { + #include <abstractions/base> + /** rw, # passt's ".repair" socket might be anywhere + unix (connect, receive, send) type=stream, + + capability dac_override, # connect to passt's socket as root + capability net_admin, # currently needed for TCP_REPAIR socket option + capability net_raw, # what TCP_REPAIR should require instead + + network unix stream, # connect and use UNIX domain socket + network inet stream, # use TCP sockets +} diff --git a/contrib/fedora/passt.spec b/contrib/fedora/passt.spec index 7950fb9..745cf01 100644 --- a/contrib/fedora/passt.spec +++ b/contrib/fedora/passt.spec @@ -44,7 +44,7 @@ Requires(preun): %{name} Requires(preun): policycoreutils %description selinux -This package adds SELinux enforcement to passt(1) and pasta(1). +This package adds SELinux enforcement to passt(1), pasta(1), passt-repair(1). %prep %setup -q -n passt-%{git_hash} @@ -82,6 +82,7 @@ make -f %{_datadir}/selinux/devel/Makefile install -p -m 644 -D passt.pp %{buildroot}%{_datadir}/selinux/packages/%{selinuxtype}/passt.pp install -p -m 644 -D passt.if %{buildroot}%{_datadir}/selinux/devel/include/distributed/passt.if install -p -m 644 -D pasta.pp %{buildroot}%{_datadir}/selinux/packages/%{selinuxtype}/pasta.pp +install -p -m 644 -D passt-repair.pp %{buildroot}%{_datadir}/selinux/packages/%{selinuxtype}/passt-repair.pp popd %pre selinux @@ -90,11 +91,13 @@ popd %post selinux %selinux_modules_install -s %{selinuxtype} %{_datadir}/selinux/packages/%{selinuxtype}/passt.pp %selinux_modules_install -s %{selinuxtype} %{_datadir}/selinux/packages/%{selinuxtype}/pasta.pp +%selinux_modules_install -s %{selinuxtype} %{_datadir}/selinux/packages/%{selinuxtype}/passt-repair.pp %postun selinux if [ $1 -eq 0 ]; then %selinux_modules_uninstall -s %{selinuxtype} passt %selinux_modules_uninstall -s %{selinuxtype} pasta + %selinux_modules_uninstall -s %{selinuxtype} passt-repair fi %posttrans selinux @@ -108,9 +111,11 @@ fi %{_bindir}/passt %{_bindir}/pasta %{_bindir}/qrap +%{_bindir}/passt-repair %{_mandir}/man1/passt.1* %{_mandir}/man1/pasta.1* %{_mandir}/man1/qrap.1* +%{_mandir}/man1/passt-repair.1* %ifarch x86_64 %{_bindir}/passt.avx2 %{_mandir}/man1/passt.avx2.1* @@ -122,6 +127,7 @@ fi %{_datadir}/selinux/packages/%{selinuxtype}/passt.pp %{_datadir}/selinux/devel/include/distributed/passt.if %{_datadir}/selinux/packages/%{selinuxtype}/pasta.pp +%{_datadir}/selinux/packages/%{selinuxtype}/passt-repair.pp %changelog {{{ passt_git_changelog }}} diff --git a/contrib/selinux/passt-repair.fc b/contrib/selinux/passt-repair.fc new file mode 100644 index 0000000..bcd526e --- /dev/null +++ b/contrib/selinux/passt-repair.fc @@ -0,0 +1,11 @@ +# SPDX-License-Identifier: GPL-2.0-or-later +# +# PASST - Plug A Simple Socket Transport +# for qemu/UNIX domain socket mode +# +# contrib/selinux/passt-repair.fc - SELinux: File Context for passt-repair +# +# Copyright (c) 2025 Red Hat GmbH +# Author: Stefano Brivio <sbrivio@redhat.com> + +/usr/bin/passt-repair system_u:object_r:passt_repair_exec_t:s0 diff --git a/contrib/selinux/passt-repair.te b/contrib/selinux/passt-repair.te new file mode 100644 index 0000000..7157dfb --- /dev/null +++ b/contrib/selinux/passt-repair.te @@ -0,0 +1,87 @@ +# SPDX-License-Identifier: GPL-2.0-or-later +# +# PASST - Plug A Simple Socket Transport +# for qemu/UNIX domain socket mode +# +# contrib/selinux/passt-repair.te - SELinux: Type Enforcement for passt-repair +# +# Copyright (c) 2025 Red Hat GmbH +# Author: Stefano Brivio <sbrivio@redhat.com> + +policy_module(passt-repair, 0.1) + +require { + type unconfined_t; + type passt_t; + role unconfined_r; + class process transition; + + class file { read execute execute_no_trans entrypoint open map }; + class capability { dac_override net_admin net_raw }; + class chr_file { append open getattr read write ioctl }; + + class unix_stream_socket { create connect sendto }; + class sock_file { read write }; + + class tcp_socket { read setopt write }; + + type console_device_t; + type user_devpts_t; + type user_tmp_t; + + # Workaround: passt-repair needs to needs to access socket files + # that passt, started by libvirt, might create under different + # labels, depending on whether passt is started as root or not. + # + # However, libvirt doesn't maintain its own policy, which makes + # updates particularly complicated. To avoid breakage in the short + # term, deal with that in passt's own policy. + type qemu_var_run_t; + type virt_var_run_t; +} + +type passt_repair_t; +domain_type(passt_repair_t); +type passt_repair_exec_t; +corecmd_executable_file(passt_repair_exec_t); + +role unconfined_r types passt_repair_t; + +allow passt_repair_t passt_repair_exec_t:file { read execute execute_no_trans entrypoint open map }; +type_transition unconfined_t passt_repair_exec_t:process passt_repair_t; +allow unconfined_t passt_repair_t:process transition; + +allow passt_repair_t self:capability { dac_override dac_read_search net_admin net_raw }; +allow passt_repair_t self:capability2 bpf; + +allow passt_repair_t console_device_t:chr_file { append open getattr read write ioctl }; +allow passt_repair_t user_devpts_t:chr_file { append open getattr read write ioctl }; + +allow passt_repair_t unconfined_t:unix_stream_socket { connectto read write }; +allow passt_repair_t passt_t:unix_stream_socket { connectto read write }; +allow passt_repair_t user_tmp_t:unix_stream_socket { connectto read write }; + +allow passt_repair_t user_tmp_t:dir { getattr read search watch }; + +allow passt_repair_t unconfined_t:sock_file { getattr read write }; +allow passt_repair_t passt_t:sock_file { getattr read write }; +allow passt_repair_t user_tmp_t:sock_file { getattr read write }; + +allow passt_repair_t unconfined_t:tcp_socket { read setopt write }; +allow passt_repair_t passt_t:tcp_socket { read setopt write }; + +# Workaround: passt-repair needs to needs to access socket files +# that passt, started by libvirt, might create under different +# labels, depending on whether passt is started as root or not. +# +# However, libvirt doesn't maintain its own policy, which makes +# updates particularly complicated. To avoid breakage in the short +# term, deal with that in passt's own policy. +allow passt_repair_t qemu_var_run_t:unix_stream_socket { connectto read write }; +allow passt_repair_t virt_var_run_t:unix_stream_socket { connectto read write }; + +allow passt_repair_t qemu_var_run_t:dir { getattr read search watch }; +allow passt_repair_t virt_var_run_t:dir { getattr read search watch }; + +allow passt_repair_t qemu_var_run_t:sock_file { getattr read write }; +allow passt_repair_t virt_var_run_t:sock_file { getattr read write }; diff --git a/contrib/selinux/passt.te b/contrib/selinux/passt.te index c6cea34..eb9ce72 100644 --- a/contrib/selinux/passt.te +++ b/contrib/selinux/passt.te @@ -20,9 +20,19 @@ require { type fs_t; type tmp_t; type user_tmp_t; + type user_home_t; type tmpfs_t; type root_t; + # Workaround: passt --vhost-user needs to map guest memory, but + # libvirt doesn't maintain its own policy, which makes updates + # particularly complicated. To avoid breakage in the short term, + # deal with it in passt's own policy. + type svirt_image_t; + type svirt_tmpfs_t; + type svirt_t; + type null_device_t; + class file { ioctl getattr setattr create read write unlink open relabelto execute execute_no_trans map }; class dir { search write add_name remove_name mounton }; class chr_file { append read write open getattr ioctl }; @@ -38,8 +48,8 @@ require { type net_conf_t; type proc_net_t; type node_t; - class tcp_socket { create accept listen name_bind name_connect }; - class udp_socket { create accept listen }; + class tcp_socket { create accept listen name_bind name_connect getattr ioctl }; + class udp_socket { create accept listen getattr }; class icmp_socket { bind create name_bind node_bind setopt read write }; class sock_file { create unlink write }; @@ -80,6 +90,9 @@ allow passt_t root_t:dir mounton; allow passt_t tmp_t:dir { add_name mounton remove_name write }; allow passt_t tmpfs_t:filesystem mount; allow passt_t fs_t:filesystem unmount; +allow passt_t user_home_t:dir search; +allow passt_t user_tmp_t:fifo_file append; +allow passt_t user_tmp_t:file map; manage_files_pattern(passt_t, user_tmp_t, user_tmp_t) files_pid_filetrans(passt_t, user_tmp_t, file) @@ -119,11 +132,19 @@ corenet_udp_sendrecv_all_ports(passt_t) allow passt_t node_t:icmp_socket { name_bind node_bind }; allow passt_t port_t:icmp_socket name_bind; -allow passt_t self:tcp_socket { create getopt setopt connect bind listen accept shutdown read write }; -allow passt_t self:udp_socket { create getopt setopt connect bind read write }; +allow passt_t self:tcp_socket { create getopt setopt connect bind listen accept shutdown read write getattr ioctl }; +allow passt_t self:udp_socket { create getopt setopt connect bind read write getattr }; allow passt_t self:icmp_socket { bind create setopt read write }; allow passt_t user_tmp_t:dir { add_name write }; allow passt_t user_tmp_t:file { create open }; allow passt_t user_tmp_t:sock_file { create read write unlink }; allow passt_t unconfined_t:unix_stream_socket { read write }; + +# Workaround: passt --vhost-user needs to map guest memory, but +# libvirt doesn't maintain its own policy, which makes updates +# particularly complicated. To avoid breakage in the short term, +# deal with it in passt's own policy. +allow passt_t svirt_image_t:file { read write map }; +allow passt_t svirt_tmpfs_t:file { read write map }; +allow passt_t null_device_t:chr_file map; diff --git a/contrib/selinux/pasta.te b/contrib/selinux/pasta.te index 69be081..89c8043 100644 --- a/contrib/selinux/pasta.te +++ b/contrib/selinux/pasta.te @@ -18,6 +18,7 @@ require { type bin_t; type user_home_t; type user_home_dir_t; + type user_tmp_t; type fs_t; type tmp_t; type tmpfs_t; @@ -56,8 +57,10 @@ require { attribute port_type; type port_t; type http_port_t; + type http_cache_port_t; type ssh_port_t; type reserved_port_t; + type unreserved_port_t; type dns_port_t; type dhcpc_port_t; type chronyd_port_t; @@ -122,8 +125,8 @@ domain_auto_trans(pasta_t, ping_exec_t, ping_t); allow pasta_t nsfs_t:file { open read }; -allow pasta_t user_home_t:dir getattr; -allow pasta_t user_home_t:file { open read getattr setattr }; +allow pasta_t user_home_t:dir { getattr search }; +allow pasta_t user_home_t:file { open read getattr setattr execute execute_no_trans map}; allow pasta_t user_home_dir_t:dir { search getattr open add_name read write }; allow pasta_t user_home_dir_t:file { create open read write }; allow pasta_t tmp_t:dir { add_name mounton remove_name write }; @@ -133,6 +136,11 @@ allow pasta_t root_t:dir mounton; manage_files_pattern(pasta_t, pasta_pid_t, pasta_pid_t) files_pid_filetrans(pasta_t, pasta_pid_t, file) +allow pasta_t user_tmp_t:dir { add_name remove_name search write }; +allow pasta_t user_tmp_t:fifo_file append; +allow pasta_t user_tmp_t:file { create open write }; +allow pasta_t user_tmp_t:sock_file { create unlink }; + allow pasta_t console_device_t:chr_file { open write getattr ioctl }; allow pasta_t user_devpts_t:chr_file { getattr read write ioctl }; logging_send_syslog_msg(pasta_t) @@ -160,6 +168,8 @@ allow pasta_t self:udp_socket create_stream_socket_perms; allow pasta_t reserved_port_t:udp_socket name_bind; allow pasta_t llmnr_port_t:tcp_socket name_bind; allow pasta_t llmnr_port_t:udp_socket name_bind; +allow pasta_t http_cache_port_t:tcp_socket { name_bind name_connect }; +allow pasta_t unreserved_port_t:udp_socket name_bind; corenet_udp_sendrecv_generic_node(pasta_t) corenet_udp_bind_generic_node(pasta_t) allow pasta_t node_t:icmp_socket { name_bind node_bind }; @@ -171,7 +181,7 @@ allow pasta_t init_t:lnk_file read; allow pasta_t init_t:unix_stream_socket connectto; allow pasta_t init_t:dbus send_msg; allow pasta_t init_t:system status; -allow pasta_t unconfined_t:dir search; +allow pasta_t unconfined_t:dir { read search }; allow pasta_t unconfined_t:file read; allow pasta_t unconfined_t:lnk_file read; allow pasta_t self:process { setpgid setcap }; @@ -192,8 +202,6 @@ allow pasta_t sysctl_net_t:dir search; allow pasta_t sysctl_net_t:file { open read write }; allow pasta_t kernel_t:system module_request; -allow pasta_t nsfs_t:file read; - allow pasta_t proc_t:dir mounton; allow pasta_t proc_t:filesystem mount; allow pasta_t net_conf_t:lnk_file read; diff --git a/dhcp.c b/dhcp.c index d8515aa..b0de04b 100644 --- a/dhcp.c +++ b/dhcp.c @@ -63,6 +63,11 @@ static struct opt opts[255]; #define OPT_MIN 60 /* RFC 951 */ +/* Total option size (excluding end option) is 576 (RFC 2131), minus + * offset of options (268), minus end option (1). + */ +#define OPT_MAX 307 + /** * dhcp_init() - Initialise DHCP options */ @@ -122,7 +127,7 @@ struct msg { uint8_t sname[64]; uint8_t file[128]; uint32_t magic; - uint8_t o[308]; + uint8_t o[OPT_MAX + 1 /* End option */ ]; } __attribute__((__packed__)); /** @@ -130,15 +135,28 @@ struct msg { * @m: Message to fill * @o: Option number * @offset: Current offset within options field, updated on insertion + * + * Return: false if m has space to write the option, true otherwise */ -static void fill_one(struct msg *m, int o, int *offset) +static bool fill_one(struct msg *m, int o, int *offset) { + size_t slen = opts[o].slen; + + /* If we don't have space to write the option, then just skip */ + if (*offset + 2 /* code and length of option */ + slen > OPT_MAX) + return true; + m->o[*offset] = o; - m->o[*offset + 1] = opts[o].slen; - memcpy(&m->o[*offset + 2], opts[o].s, opts[o].slen); + m->o[*offset + 1] = slen; + + /* Move to option */ + *offset += 2; + + memcpy(&m->o[*offset], opts[o].s, slen); opts[o].sent = 1; - *offset += 2 + opts[o].slen; + *offset += slen; + return false; } /** @@ -151,9 +169,6 @@ static int fill(struct msg *m) { int i, o, offset = 0; - m->op = BOOTREPLY; - m->secs = 0; - for (o = 0; o < 255; o++) opts[o].sent = 0; @@ -162,21 +177,23 @@ static int fill(struct msg *m) * Put it there explicitly, unless requested via option 55. */ if (opts[55].clen > 0 && !memchr(opts[55].c, 53, opts[55].clen)) - fill_one(m, 53, &offset); + if (fill_one(m, 53, &offset)) + debug("DHCP: skipping option 53"); for (i = 0; i < opts[55].clen; i++) { o = opts[55].c[i]; if (opts[o].slen != -1) - fill_one(m, o, &offset); + if (fill_one(m, o, &offset)) + debug("DHCP: skipping option %i", o); } for (o = 0; o < 255; o++) { if (opts[o].slen != -1 && !opts[o].sent) - fill_one(m, o, &offset); + if (fill_one(m, o, &offset)) + debug("DHCP: skipping option %i", o); } m->o[offset++] = 255; - m->o[offset++] = 0; if (offset < OPT_MIN) { memset(&m->o[offset], 0, OPT_MIN - offset); @@ -291,8 +308,9 @@ int dhcp(const struct ctx *c, const struct pool *p) const struct ethhdr *eh; const struct iphdr *iph; const struct udphdr *uh; + struct msg const *m; + struct msg reply; unsigned int i; - struct msg *m; eh = packet_get(p, 0, offset, sizeof(*eh), NULL); offset += sizeof(*eh); @@ -321,6 +339,22 @@ int dhcp(const struct ctx *c, const struct pool *p) m->op != BOOTREQUEST) return -1; + reply.op = BOOTREPLY; + reply.htype = m->htype; + reply.hlen = m->hlen; + reply.hops = 0; + reply.xid = m->xid; + reply.secs = 0; + reply.flags = m->flags; + reply.ciaddr = m->ciaddr; + reply.yiaddr = c->ip4.addr; + reply.siaddr = 0; + reply.giaddr = m->giaddr; + memcpy(&reply.chaddr, m->chaddr, sizeof(reply.chaddr)); + memset(&reply.sname, 0, sizeof(reply.sname)); + memset(&reply.file, 0, sizeof(reply.file)); + reply.magic = m->magic; + offset += offsetof(struct msg, o); for (i = 0; i < ARRAY_SIZE(opts); i++) @@ -364,7 +398,6 @@ int dhcp(const struct ctx *c, const struct pool *p) info(" from %s", eth_ntop(m->chaddr, macstr, sizeof(macstr))); - m->yiaddr = c->ip4.addr; mask.s_addr = htonl(0xffffffff << (32 - c->ip4.prefix_len)); memcpy(opts[1].s, &mask, sizeof(mask)); memcpy(opts[3].s, &c->ip4.guest_gw, sizeof(c->ip4.guest_gw)); @@ -384,7 +417,7 @@ int dhcp(const struct ctx *c, const struct pool *p) &c->ip4.guest_gw, sizeof(c->ip4.guest_gw)); } - if (c->mtu != -1) { + if (c->mtu) { opts[26].slen = 2; opts[26].s[0] = c->mtu / 256; opts[26].s[1] = c->mtu % 256; @@ -398,17 +431,41 @@ int dhcp(const struct ctx *c, const struct pool *p) if (!opts[6].slen) opts[6].slen = -1; + opt_len = strlen(c->hostname); + if (opt_len > 0) { + opts[12].slen = opt_len; + memcpy(opts[12].s, &c->hostname, opt_len); + } + + opt_len = strlen(c->fqdn); + if (opt_len > 0) { + opt_len += 3 /* flags */ + + 2; /* Length byte for first label, and terminator */ + + if (sizeof(opts[81].s) >= opt_len) { + opts[81].s[0] = 0x4; /* flags (E) */ + opts[81].s[1] = 0xff; /* RCODE1 */ + opts[81].s[2] = 0xff; /* RCODE2 */ + + encode_domain_name((char *)opts[81].s + 3, c->fqdn); + + opts[81].slen = opt_len; + } else { + debug("DHCP: client FQDN option doesn't fit, skipping"); + } + } + if (!c->no_dhcp_dns_search) opt_set_dns_search(c, sizeof(m->o)); - dlen = offsetof(struct msg, o) + fill(m); + dlen = offsetof(struct msg, o) + fill(&reply); if (m->flags & FLAG_BROADCAST) dst = in4addr_broadcast; else dst = c->ip4.addr; - tap_udp4_send(c, c->ip4.our_tap_addr, 67, dst, 68, m, dlen); + tap_udp4_send(c, c->ip4.our_tap_addr, 67, dst, 68, &reply, dlen); return 1; } diff --git a/dhcpv6.c b/dhcpv6.c index 0523bba..ba16c66 100644 --- a/dhcpv6.c +++ b/dhcpv6.c @@ -48,6 +48,7 @@ struct opt_hdr { # define STATUS_NOTONLINK htons_constant(4) # define OPT_DNS_SERVERS htons_constant(23) # define OPT_DNS_SEARCH htons_constant(24) +# define OPT_CLIENT_FQDN htons_constant(39) #define STR_NOTONLINK "Prefix not appropriate for link." uint16_t l; @@ -58,6 +59,9 @@ struct opt_hdr { sizeof(struct opt_hdr)) #define OPT_VSIZE(x) (sizeof(struct opt_##x) - \ sizeof(struct opt_hdr)) +#define OPT_MAX_SIZE IPV6_MIN_MTU - (sizeof(struct ipv6hdr) + \ + sizeof(struct udphdr) + \ + sizeof(struct msg_hdr)) /** * struct opt_client_id - DHCPv6 Client Identifier option @@ -140,7 +144,9 @@ struct opt_ia_addr { struct opt_status_code { struct opt_hdr hdr; uint16_t code; - char status_msg[sizeof(STR_NOTONLINK) - 1]; + /* "nonstring" is only supported since clang 23 */ + /* NOLINTNEXTLINE(clang-diagnostic-unknown-attributes) */ + __attribute__((nonstring)) char status_msg[sizeof(STR_NOTONLINK) - 1]; } __attribute__((packed)); /** @@ -163,6 +169,18 @@ struct opt_dns_search { char list[MAXDNSRCH * NS_MAXDNAME]; } __attribute__((packed)); +/** + * struct opt_client_fqdn - Client FQDN option (RFC 4704) + * @hdr: Option header + * @flags: Flags described by RFC 4704 + * @domain_name: Client FQDN + */ +struct opt_client_fqdn { + struct opt_hdr hdr; + uint8_t flags; + char domain_name[PASST_MAXDNAME]; +} __attribute__((packed)); + /** * struct msg_hdr - DHCPv6 client/server message header * @type: DHCP message type @@ -193,6 +211,7 @@ struct msg_hdr { * @client_id: Client Identifier, variable length * @dns_servers: DNS Recursive Name Server, here just for storage size * @dns_search: Domain Search List, here just for storage size + * @client_fqdn: Client FQDN, variable length */ static struct resp_t { struct msg_hdr hdr; @@ -203,6 +222,7 @@ static struct resp_t { struct opt_client_id client_id; struct opt_dns_servers dns_servers; struct opt_dns_search dns_search; + struct opt_client_fqdn client_fqdn; } __attribute__((__packed__)) resp = { { 0 }, SERVER_ID, @@ -228,6 +248,10 @@ static struct resp_t { { { OPT_DNS_SEARCH, 0, }, { 0 }, }, + + { { OPT_CLIENT_FQDN, 0, }, + 0, { 0 }, + }, }; static const struct opt_status_code sc_not_on_link = { @@ -346,7 +370,6 @@ static size_t dhcpv6_dns_fill(const struct ctx *c, char *buf, int offset) { struct opt_dns_servers *srv = NULL; struct opt_dns_search *srch = NULL; - char *p = NULL; int i; if (c->no_dhcp_dns) @@ -383,34 +406,81 @@ search: if (!name_len) continue; + name_len += 2; /* Length byte for first label, and terminator */ + if (name_len > + NS_MAXDNAME + 1 /* Length byte for first label */ || + name_len > 255) { + debug("DHCP: DNS search name '%s' too long, skipping", + c->dns_search[i].n); + continue; + } + if (!srch) { srch = (struct opt_dns_search *)(buf + offset); offset += sizeof(struct opt_hdr); srch->hdr.t = OPT_DNS_SEARCH; srch->hdr.l = 0; - p = srch->list; } - *p = '.'; - p = stpncpy(p + 1, c->dns_search[i].n, name_len); - p++; - srch->hdr.l += name_len + 2; - offset += name_len + 2; + encode_domain_name(buf + offset, c->dns_search[i].n); + + srch->hdr.l += name_len; + offset += name_len; + } - if (srch) { - for (i = 0; i < srch->hdr.l; i++) { - if (srch->list[i] == '.') { - srch->list[i] = strcspn(srch->list + i + 1, - "."); - } - } + if (srch) srch->hdr.l = htons(srch->hdr.l); - } return offset; } +/** + * dhcpv6_client_fqdn_fill() - Fill in client FQDN option + * @c: Execution context + * @buf: Response message buffer where options will be appended + * @offset: Offset in message buffer for new options + * + * Return: updated length of response message buffer. + */ +static size_t dhcpv6_client_fqdn_fill(const struct pool *p, const struct ctx *c, + char *buf, int offset) + +{ + struct opt_client_fqdn const *req_opt; + struct opt_client_fqdn *o; + size_t opt_len; + + opt_len = strlen(c->fqdn); + if (opt_len == 0) { + return offset; + } + + opt_len += 2; /* Length byte for first label, and terminator */ + if (opt_len > OPT_MAX_SIZE - (offset + + sizeof(struct opt_hdr) + + 1 /* flags */ )) { + debug("DHCPv6: client FQDN option doesn't fit, skipping"); + return offset; + } + + o = (struct opt_client_fqdn *)(buf + offset); + encode_domain_name(o->domain_name, c->fqdn); + req_opt = (struct opt_client_fqdn *)dhcpv6_opt(p, &(size_t){ 0 }, + OPT_CLIENT_FQDN); + if (req_opt && req_opt->flags & 0x01 /* S flag */) + o->flags = 0x02 /* O flag */; + else + o->flags = 0x00; + + opt_len++; + + o->hdr.t = OPT_CLIENT_FQDN; + o->hdr.l = htons(opt_len); + + return offset + sizeof(struct opt_hdr) + opt_len; +} + /** * dhcpv6() - Check if this is a DHCPv6 message, reply as needed * @c: Execution context @@ -544,6 +614,7 @@ int dhcpv6(struct ctx *c, const struct pool *p, n = offsetof(struct resp_t, client_id) + sizeof(struct opt_hdr) + ntohs(client_id->l); n = dhcpv6_dns_fill(c, (char *)&resp, n); + n = dhcpv6_client_fqdn_fill(p, c, (char *)&resp, n); resp.hdr.xid = mh->xid; diff --git a/doc/migration/.gitignore b/doc/migration/.gitignore new file mode 100644 index 0000000..59cb765 --- /dev/null +++ b/doc/migration/.gitignore @@ -0,0 +1,2 @@ +/source +/target diff --git a/doc/migration/Makefile b/doc/migration/Makefile new file mode 100644 index 0000000..04f6891 --- /dev/null +++ b/doc/migration/Makefile @@ -0,0 +1,20 @@ +# SPDX-License-Identifier: GPL-2.0-or-later +# +# PASST - Plug A Simple Socket Transport +# for qemu/UNIX domain socket mode +# +# PASTA - Pack A Subtle Tap Abstraction +# for network namespace/tap device mode +# +# Copyright (c) 2025 Red Hat GmbH +# Author: Stefano Brivio <sbrivio@redhat.com> + +TARGETS = source target +CFLAGS = -Wall -Wextra -pedantic + +all: $(TARGETS) + +$(TARGETS): %: %.c + +clean: + rm -f $(TARGETS) diff --git a/doc/migration/README b/doc/migration/README new file mode 100644 index 0000000..375603b --- /dev/null +++ b/doc/migration/README @@ -0,0 +1,51 @@ +<!--- +SPDX-License-Identifier: GPL-2.0-or-later +Copyright (c) 2025 Red Hat GmbH +Author: Stefano Brivio <sbrivio@redhat.com> +--> + +Migration +========= + +These test programs show a migration of a TCP connection from one process to +another using the TCP_REPAIR socket option. + +The two processes are a mock of the matching implementation in passt(1), and run +unprivileged, so they rely on the passt-repair helper to connect to them and set +or clear TCP_REPAIR on the connection socket, transferred to the helper using +SCM_RIGHTS. + +The passt-repair helper needs to have the CAP_NET_ADMIN capability, or run as +root. + +Example of usage +---------------- + +* Start the test server + + $ nc -l 9999 + +* Start the source side of the TCP client (mock of the source instance of passt) + + $ ./source 127.0.0.1 9999 9998 /tmp/repair.sock + +* The client sends a test string, and waits for a connection from passt-repair + + # passt-repair /tmp/repair.sock + +* The socket is now in repair mode, and `source` dumps sequences, then exits + + sending sequence: 3244673313 + receiving sequence: 2250449386 + +* Continue the connection on the target side, restarting from those sequences + + $ ./target 127.0.0.1 9999 9998 /tmp/repair.sock 3244673313 2250449386 + +* The target side now waits for a connection from passt-repair + + # passt-repair /tmp/repair.sock + +* The target side asks passt-repair to switch the socket to repair mode, sets up + the TCP sequences, then asks passt-repair to clear repair mode, and sends a + test string to the server diff --git a/doc/migration/source.c b/doc/migration/source.c new file mode 100644 index 0000000..d44ebf1 --- /dev/null +++ b/doc/migration/source.c @@ -0,0 +1,92 @@ +// SPDX-License-Identifier: GPL-2.0-or-later + +/* PASST - Plug A Simple Socket Transport + * for qemu/UNIX domain socket mode + * + * PASTA - Pack A Subtle Tap Abstraction + * for network namespace/tap device mode + * + * doc/migration/source.c - Mock of TCP migration source, use with passt-repair + * + * Copyright (c) 2025 Red Hat GmbH + * Author: Stefano Brivio <sbrivio@redhat.com> + */ + +#include <arpa/inet.h> +#include <sys/types.h> +#include <sys/socket.h> +#include <sys/un.h> +#include <stdint.h> +#include <stdio.h> +#include <stdlib.h> +#include <limits.h> +#include <unistd.h> +#include <netdb.h> +#include <netinet/tcp.h> + +int main(int argc, char **argv) +{ + struct sockaddr_in a = { AF_INET, htons(atoi(argv[3])), { 0 }, { 0 } }; + struct addrinfo hints = { 0, AF_UNSPEC, SOCK_STREAM, 0, 0, + NULL, NULL, NULL }; + struct sockaddr_un a_helper = { AF_UNIX, { 0 } }; + int seq, s, s_helper; + int8_t cmd; + struct iovec iov = { &cmd, sizeof(cmd) }; + char buf[CMSG_SPACE(sizeof(int))]; + struct msghdr msg = { NULL, 0, &iov, 1, buf, sizeof(buf), 0 }; + struct cmsghdr *cmsg = CMSG_FIRSTHDR(&msg); + socklen_t seqlen = sizeof(int); + struct addrinfo *r; + + (void)argc; + + if (argc != 5) { + fprintf(stderr, "%s DST_ADDR DST_PORT SRC_PORT HELPER_PATH\n", + argv[0]); + return -1; + } + + strcpy(a_helper.sun_path, argv[4]); + getaddrinfo(argv[1], argv[2], &hints, &r); + + /* Connect socket to server and send some data */ + s = socket(r->ai_family, SOCK_STREAM, IPPROTO_TCP); + setsockopt(s, SOL_SOCKET, SO_REUSEADDR, &((int){ 1 }), sizeof(int)); + bind(s, (struct sockaddr *)&a, sizeof(a)); + connect(s, r->ai_addr, r->ai_addrlen); + send(s, "before migration\n", sizeof("before migration\n"), 0); + + /* Wait for helper */ + s_helper = socket(AF_UNIX, SOCK_STREAM, 0); + unlink(a_helper.sun_path); + bind(s_helper, (struct sockaddr *)&a_helper, sizeof(a_helper)); + listen(s_helper, 1); + s_helper = accept(s_helper, NULL, NULL); + + /* Set up message for helper, with socket */ + cmsg->cmsg_level = SOL_SOCKET; + cmsg->cmsg_type = SCM_RIGHTS; + cmsg->cmsg_len = CMSG_LEN(sizeof(int)); + memcpy(CMSG_DATA(cmsg), &s, sizeof(s)); + + /* Send command to helper: turn repair mode on, wait for reply */ + cmd = TCP_REPAIR_ON; + sendmsg(s_helper, &msg, 0); + recv(s_helper, &((int8_t){ 0 }), 1, 0); + + /* Terminate helper */ + close(s_helper); + + /* Get sending sequence */ + seq = TCP_SEND_QUEUE; + setsockopt(s, SOL_TCP, TCP_REPAIR_QUEUE, &seq, sizeof(seq)); + getsockopt(s, SOL_TCP, TCP_QUEUE_SEQ, &seq, &seqlen); + fprintf(stdout, "%u ", seq); + + /* Get receiving sequence */ + seq = TCP_RECV_QUEUE; + setsockopt(s, SOL_TCP, TCP_REPAIR_QUEUE, &seq, sizeof(seq)); + getsockopt(s, SOL_TCP, TCP_QUEUE_SEQ, &seq, &seqlen); + fprintf(stdout, "%u\n", seq); +} diff --git a/doc/migration/target.c b/doc/migration/target.c new file mode 100644 index 0000000..f7d3108 --- /dev/null +++ b/doc/migration/target.c @@ -0,0 +1,102 @@ +// SPDX-License-Identifier: GPL-2.0-or-later + +/* PASST - Plug A Simple Socket Transport + * for qemu/UNIX domain socket mode + * + * PASTA - Pack A Subtle Tap Abstraction + * for network namespace/tap device mode + * + * doc/migration/target.c - Mock of TCP migration target, use with passt-repair + * + * Copyright (c) 2025 Red Hat GmbH + * Author: Stefano Brivio <sbrivio@redhat.com> + */ + +#include <arpa/inet.h> +#include <sys/types.h> +#include <sys/socket.h> +#include <sys/un.h> +#include <stdint.h> +#include <stdio.h> +#include <stdlib.h> +#include <limits.h> +#include <unistd.h> +#include <netdb.h> +#include <netinet/tcp.h> + +int main(int argc, char **argv) +{ + struct sockaddr_in a = { AF_INET, htons(atoi(argv[3])), { 0 }, { 0 } }; + struct addrinfo hints = { 0, AF_UNSPEC, SOCK_STREAM, 0, 0, + NULL, NULL, NULL }; + struct sockaddr_un a_helper = { AF_UNIX, { 0 } }; + int s, s_helper, seq; + int8_t cmd; + struct iovec iov = { &cmd, sizeof(cmd) }; + char buf[CMSG_SPACE(sizeof(int))]; + struct msghdr msg = { NULL, 0, &iov, 1, buf, sizeof(buf), 0 }; + struct cmsghdr *cmsg = CMSG_FIRSTHDR(&msg); + struct addrinfo *r; + + (void)argc; + + strcpy(a_helper.sun_path, argv[4]); + getaddrinfo(argv[1], argv[2], &hints, &r); + + if (argc != 7) { + fprintf(stderr, + "%s DST_ADDR DST_PORT SRC_PORT HELPER_PATH SSEQ RSEQ\n", + argv[0]); + return -1; + } + + /* Prepare socket, bind to source port */ + s = socket(r->ai_family, SOCK_STREAM, IPPROTO_TCP); + setsockopt(s, SOL_SOCKET, SO_REUSEADDR, &((int){ 1 }), sizeof(int)); + bind(s, (struct sockaddr *)&a, sizeof(a)); + + /* Wait for helper */ + s_helper = socket(AF_UNIX, SOCK_STREAM, 0); + unlink(a_helper.sun_path); + bind(s_helper, (struct sockaddr *)&a_helper, sizeof(a_helper)); + listen(s_helper, 1); + s_helper = accept(s_helper, NULL, NULL); + + /* Set up message for helper, with socket */ + cmsg->cmsg_level = SOL_SOCKET; + cmsg->cmsg_type = SCM_RIGHTS; + cmsg->cmsg_len = CMSG_LEN(sizeof(int)); + memcpy(CMSG_DATA(cmsg), &s, sizeof(s)); + + /* Send command to helper: turn repair mode on, wait for reply */ + cmd = TCP_REPAIR_ON; + sendmsg(s_helper, &msg, 0); + recv(s_helper, &((int){ 0 }), 1, 0); + + /* Set sending sequence */ + seq = TCP_SEND_QUEUE; + setsockopt(s, SOL_TCP, TCP_REPAIR_QUEUE, &seq, sizeof(seq)); + seq = atoi(argv[5]); + setsockopt(s, SOL_TCP, TCP_QUEUE_SEQ, &seq, sizeof(seq)); + + /* Set receiving sequence */ + seq = TCP_RECV_QUEUE; + setsockopt(s, SOL_TCP, TCP_REPAIR_QUEUE, &seq, sizeof(seq)); + seq = atoi(argv[6]); + setsockopt(s, SOL_TCP, TCP_QUEUE_SEQ, &seq, sizeof(seq)); + + /* Connect setting kernel state only, without actual SYN / handshake */ + connect(s, r->ai_addr, r->ai_addrlen); + + /* Send command to helper: turn repair mode off, wait for reply */ + cmd = TCP_REPAIR_OFF; + sendmsg(s_helper, &msg, 0); + + recv(s_helper, &((int8_t){ 0 }), 1, 0); + + /* Terminate helper */ + close(s_helper); + + /* Send some more data */ + send(s, "after migration\n", sizeof("after migration\n"), 0); +} diff --git a/doc/platform-requirements/.gitignore b/doc/platform-requirements/.gitignore index 3b5a10a..f6272cf 100644 --- a/doc/platform-requirements/.gitignore +++ b/doc/platform-requirements/.gitignore @@ -1,3 +1,4 @@ +/listen-vs-repair /reuseaddr-priority /recv-zero /udp-close-dup diff --git a/doc/platform-requirements/Makefile b/doc/platform-requirements/Makefile index 6a7d374..83930ef 100644 --- a/doc/platform-requirements/Makefile +++ b/doc/platform-requirements/Makefile @@ -3,8 +3,8 @@ # Copyright Red Hat # Author: David Gibson <david@gibson.dropbear.id.au> -TARGETS = reuseaddr-priority recv-zero udp-close-dup -SRCS = reuseaddr-priority.c recv-zero.c udp-close-dup.c +TARGETS = reuseaddr-priority recv-zero udp-close-dup listen-vs-repair +SRCS = reuseaddr-priority.c recv-zero.c udp-close-dup.c listen-vs-repair.c CFLAGS = -Wall all: cppcheck clang-tidy $(TARGETS:%=check-%) diff --git a/doc/platform-requirements/common.h b/doc/platform-requirements/common.h index 8844b1e..e85fc2b 100644 --- a/doc/platform-requirements/common.h +++ b/doc/platform-requirements/common.h @@ -15,6 +15,7 @@ #include <stdio.h> #include <stdlib.h> +__attribute__((format(printf, 1, 2), noreturn)) static inline void die(const char *fmt, ...) { va_list ap; diff --git a/doc/platform-requirements/listen-vs-repair.c b/doc/platform-requirements/listen-vs-repair.c new file mode 100644 index 0000000..d31fe3f --- /dev/null +++ b/doc/platform-requirements/listen-vs-repair.c @@ -0,0 +1,128 @@ +// SPDX-License-Identifier: GPL-2.0-or-later + +/* liste-vs-repair.c + * + * Do listening sockets have address conflicts with sockets under repair + * ==================================================================== + * + * When we accept() an incoming connection the accept()ed socket will have the + * same local address as the listening socket. This can be a complication on + * migration. On the migration target we've already set up listening sockets + * according to the command line. However to restore connections that we're + * migrating in we need to bind the new sockets to the same address, which would + * be an address conflict on the face of it. This test program verifies that + * enabling repair mode before bind() correctly suppresses that conflict. + * + * Copyright Red Hat + * Author: David Gibson <david@gibson.dropbear.id.au> + */ + +/* NOLINTNEXTLINE(bugprone-reserved-identifier,cert-dcl37-c,cert-dcl51-cpp) */ +#define _GNU_SOURCE + +#include <arpa/inet.h> +#include <errno.h> +#include <linux/netlink.h> +#include <linux/rtnetlink.h> +#include <net/if.h> +#include <netinet/in.h> +#include <netinet/tcp.h> +#include <sched.h> +#include <stdbool.h> +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <unistd.h> + +#include "common.h" + +#define PORT 13256U +#define CPORT 13257U + +/* 127.0.0.1:PORT */ +static const struct sockaddr_in addr = SOCKADDR_INIT(INADDR_LOOPBACK, PORT); + +/* 127.0.0.1:CPORT */ +static const struct sockaddr_in caddr = SOCKADDR_INIT(INADDR_LOOPBACK, CPORT); + +/* Put ourselves into a network sandbox */ +static void net_sandbox(void) +{ + /* NOLINTNEXTLINE(altera-struct-pack-align) */ + const struct req_t { + struct nlmsghdr nlh; + struct ifinfomsg ifm; + } __attribute__((packed)) req = { + .nlh.nlmsg_type = RTM_NEWLINK, + .nlh.nlmsg_flags = NLM_F_REQUEST, + .nlh.nlmsg_len = sizeof(req), + .nlh.nlmsg_seq = 1, + .ifm.ifi_family = AF_UNSPEC, + .ifm.ifi_index = 1, + .ifm.ifi_flags = IFF_UP, + .ifm.ifi_change = IFF_UP, + }; + int nl; + + if (unshare(CLONE_NEWUSER | CLONE_NEWNET)) + die("unshare(): %s\n", strerror(errno)); + + /* Bring up lo in the new netns */ + nl = socket(AF_NETLINK, SOCK_RAW | SOCK_CLOEXEC, NETLINK_ROUTE); + if (nl < 0) + die("Can't create netlink socket: %s\n", strerror(errno)); + + if (send(nl, &req, sizeof(req), 0) < 0) + die("Netlink send(): %s\n", strerror(errno)); + close(nl); +} + +static void check(void) +{ + int s1, s2, op; + + s1 = socket(AF_INET, SOCK_STREAM, IPPROTO_TCP); + if (s1 < 0) + die("socket() 1: %s\n", strerror(errno)); + + if (bind(s1, (struct sockaddr *)&addr, sizeof(addr))) + die("bind() 1: %s\n", strerror(errno)); + + if (listen(s1, 0)) + die("listen(): %s\n", strerror(errno)); + + s2 = socket(AF_INET, SOCK_STREAM, IPPROTO_TCP); + if (s2 < 0) + die("socket() 2: %s\n", strerror(errno)); + + op = TCP_REPAIR_ON; + if (setsockopt(s2, SOL_TCP, TCP_REPAIR, &op, sizeof(op))) + die("TCP_REPAIR: %s\n", strerror(errno)); + + if (bind(s2, (struct sockaddr *)&addr, sizeof(addr))) + die("bind() 2: %s\n", strerror(errno)); + + if (connect(s2, (struct sockaddr *)&caddr, sizeof(caddr))) + die("connect(): %s\n", strerror(errno)); + + op = TCP_REPAIR_OFF_NO_WP; + if (setsockopt(s2, SOL_TCP, TCP_REPAIR, &op, sizeof(op))) + die("TCP_REPAIR: %s\n", strerror(errno)); + + close(s1); + close(s2); +} + +int main(int argc, char *argv[]) +{ + (void)argc; + (void)argv; + + net_sandbox(); + + check(); + + printf("Repair mode appears to properly suppress conflicts with listening sockets\n"); + + exit(0); +} diff --git a/doc/platform-requirements/reuseaddr-priority.c b/doc/platform-requirements/reuseaddr-priority.c index 701b6ff..af39a39 100644 --- a/doc/platform-requirements/reuseaddr-priority.c +++ b/doc/platform-requirements/reuseaddr-priority.c @@ -46,13 +46,13 @@ /* Different cases for receiving socket configuration */ enum sock_type { /* Socket is bound to 0.0.0.0:DSTPORT and not connected */ - SOCK_BOUND_ANY = 0, + SOCK_BOUND_ANY, /* Socket is bound to 127.0.0.1:DSTPORT and not connected */ - SOCK_BOUND_LO = 1, + SOCK_BOUND_LO, /* Socket is bound to 0.0.0.0:DSTPORT and connected to 127.0.0.1:SRCPORT */ - SOCK_CONNECTED = 2, + SOCK_CONNECTED, NUM_SOCK_TYPES, }; diff --git a/epoll_type.h b/epoll_type.h index 0ad1efa..12ac59b 100644 --- a/epoll_type.h +++ b/epoll_type.h @@ -22,8 +22,8 @@ enum epoll_type { EPOLL_TYPE_TCP_TIMER, /* UDP "listening" sockets */ EPOLL_TYPE_UDP_LISTEN, - /* UDP socket for replies on a specific flow */ - EPOLL_TYPE_UDP_REPLY, + /* UDP socket for a specific flow */ + EPOLL_TYPE_UDP, /* ICMP/ICMPv6 ping sockets */ EPOLL_TYPE_PING, /* inotify fd watching for end of netns (pasta) */ @@ -36,6 +36,14 @@ enum epoll_type { EPOLL_TYPE_TAP_PASST, /* socket listening for qemu socket connections */ EPOLL_TYPE_TAP_LISTEN, + /* vhost-user command socket */ + EPOLL_TYPE_VHOST_CMD, + /* vhost-user kick event socket */ + EPOLL_TYPE_VHOST_KICK, + /* TCP_REPAIR helper listening socket */ + EPOLL_TYPE_REPAIR_LISTEN, + /* TCP_REPAIR helper socket */ + EPOLL_TYPE_REPAIR, EPOLL_NUM_TYPES, }; diff --git a/flow.c b/flow.c index 1ea112b..6a5c8aa 100644 --- a/flow.c +++ b/flow.c @@ -19,6 +19,7 @@ #include "inany.h" #include "flow.h" #include "flow_table.h" +#include "repair.h" const char *flow_state_str[] = { [FLOW_STATE_FREE] = "FREE", @@ -52,6 +53,13 @@ const uint8_t flow_proto[] = { static_assert(ARRAY_SIZE(flow_proto) == FLOW_NUM_TYPES, "flow_proto[] doesn't match enum flow_type"); +#define foreach_established_tcp_flow(flow) \ + flow_foreach_of_type((flow), FLOW_TCP) \ + if (!tcp_flow_is_established(&(flow)->tcp)) \ + /* NOLINTNEXTLINE(bugprone-branch-clone) */ \ + continue; \ + else + /* Global Flow Table */ /** @@ -73,7 +81,7 @@ static_assert(ARRAY_SIZE(flow_proto) == FLOW_NUM_TYPES, * * Free cluster list * flow_first_free gives the index of the first (lowest index) free cluster. - * Each free cluster has the index of the next free cluster, or MAX_FLOW if + * Each free cluster has the index of the next free cluster, or FLOW_MAX if * it is the last free cluster. Together these form a linked list of free * clusters, in strictly increasing order of index. * @@ -259,11 +267,13 @@ int flowside_connect(const struct ctx *c, int s, /** flow_log_ - Log flow-related message * @f: flow the message is related to + * @newline: Append newline at the end of the message, if missing * @pri: Log priority * @fmt: Format string * @...: printf-arguments */ -void flow_log_(const struct flow_common *f, int pri, const char *fmt, ...) +void flow_log_(const struct flow_common *f, bool newline, int pri, + const char *fmt, ...) { const char *type_or_state; char msg[BUFSIZ]; @@ -279,7 +289,7 @@ void flow_log_(const struct flow_common *f, int pri, const char *fmt, ...) else type_or_state = FLOW_TYPE(f); - logmsg(true, false, pri, + logmsg(newline, false, pri, "Flow %u (%s): %s", flow_idx(f), type_or_state, msg); } @@ -299,7 +309,7 @@ void flow_log_details_(const struct flow_common *f, int pri, const struct flowside *tgt = &f->side[TGTSIDE]; if (state >= FLOW_STATE_TGT) - flow_log_(f, pri, + flow_log_(f, true, pri, "%s [%s]:%hu -> [%s]:%hu => %s [%s]:%hu -> [%s]:%hu", pif_name(f->pif[INISIDE]), inany_ntop(&ini->eaddr, estr0, sizeof(estr0)), @@ -312,7 +322,7 @@ void flow_log_details_(const struct flow_common *f, int pri, inany_ntop(&tgt->eaddr, estr1, sizeof(estr1)), tgt->eport); else if (state >= FLOW_STATE_INI) - flow_log_(f, pri, "%s [%s]:%hu -> [%s]:%hu => ?", + flow_log_(f, true, pri, "%s [%s]:%hu -> [%s]:%hu => ?", pif_name(f->pif[INISIDE]), inany_ntop(&ini->eaddr, estr0, sizeof(estr0)), ini->eport, @@ -333,7 +343,7 @@ static void flow_set_state(struct flow_common *f, enum flow_state state) ASSERT(oldstate < FLOW_NUM_STATES); f->state = state; - flow_log_(f, LOG_DEBUG, "%s -> %s", flow_state_str[oldstate], + flow_log_(f, true, LOG_DEBUG, "%s -> %s", flow_state_str[oldstate], FLOW_STATE(f)); flow_log_details_(f, LOG_DEBUG, MAX(state, oldstate)); @@ -386,18 +396,27 @@ const struct flowside *flow_initiate_af(union flow *flow, uint8_t pif, * @flow: Flow to change state * @pif: pif of the initiating side * @ssa: Source socket address + * @daddr: Destination address (may be NULL) * @dport: Destination port * * Return: pointer to the initiating flowside information */ -const struct flowside *flow_initiate_sa(union flow *flow, uint8_t pif, - const union sockaddr_inany *ssa, - in_port_t dport) +struct flowside *flow_initiate_sa(union flow *flow, uint8_t pif, + const union sockaddr_inany *ssa, + const union inany_addr *daddr, + in_port_t dport) { struct flowside *ini = &flow->f.side[INISIDE]; - inany_from_sockaddr(&ini->eaddr, &ini->eport, ssa); - if (inany_v4(&ini->eaddr)) + if (inany_from_sockaddr(&ini->eaddr, &ini->eport, ssa) < 0) { + char str[SOCKADDR_STRLEN]; + + ASSERT_WITH_MSG(0, "Bad socket address %s", + sockaddr_ntop(ssa, str, sizeof(str))); + } + if (daddr) + ini->oaddr = *daddr; + else if (inany_v4(&ini->eaddr)) ini->oaddr = inany_any4; else ini->oaddr = inany_any6; @@ -414,8 +433,8 @@ const struct flowside *flow_initiate_sa(union flow *flow, uint8_t pif, * * Return: pointer to the target flowside information */ -const struct flowside *flow_target(const struct ctx *c, union flow *flow, - uint8_t proto) +struct flowside *flow_target(const struct ctx *c, union flow *flow, + uint8_t proto) { char estr[INANY_ADDRSTRLEN], fstr[INANY_ADDRSTRLEN]; struct flow_common *f = &flow->f; @@ -597,12 +616,7 @@ static uint64_t flow_sidx_hash(const struct ctx *c, flow_sidx_t sidx) const struct flowside *side = &f->side[sidx.sidei]; uint8_t pif = f->pif[sidx.sidei]; - /* For the hash table to work, entries must have complete endpoint - * information, and at least a forwarding port. - */ - ASSERT(pif != PIF_NONE && !inany_is_unspecified(&side->eaddr) && - side->eport != 0 && side->oport != 0); - + ASSERT(pif != PIF_NONE); return flow_hash(c, FLOW_PROTO(f), pif, side); } @@ -746,19 +760,30 @@ flow_sidx_t flow_lookup_af(const struct ctx *c, * @proto: Protocol of the flow (IP L4 protocol number) * @pif: Interface of the flow * @esa: Socket address of the endpoint + * @oaddr: Our address (may be NULL) * @oport: Our port number * * Return: sidx of the matching flow & side, FLOW_SIDX_NONE if not found */ flow_sidx_t flow_lookup_sa(const struct ctx *c, uint8_t proto, uint8_t pif, - const void *esa, in_port_t oport) + const void *esa, + const union inany_addr *oaddr, in_port_t oport) { struct flowside side = { .oport = oport, }; - inany_from_sockaddr(&side.eaddr, &side.eport, esa); - if (inany_v4(&side.eaddr)) + if (inany_from_sockaddr(&side.eaddr, &side.eport, esa) < 0) { + char str[SOCKADDR_STRLEN]; + + warn("Flow lookup on bad socket address %s", + sockaddr_ntop(esa, str, sizeof(str))); + return FLOW_SIDX_NONE; + } + + if (oaddr) + side.oaddr = *oaddr; + else if (inany_v4(&side.eaddr)) side.oaddr = inany_any4; else side.oaddr = inany_any6; @@ -775,8 +800,9 @@ void flow_defer_handler(const struct ctx *c, const struct timespec *now) { struct flow_free_cluster *free_head = NULL; unsigned *last_next = &flow_first_free; + bool to_free[FLOW_MAX] = { 0 }; bool timer = false; - unsigned idx; + union flow *flow; if (timespec_diff_ms(now, &flow_timer_run) >= FLOW_TIMER_INTERVAL) { timer = true; @@ -785,49 +811,12 @@ void flow_defer_handler(const struct ctx *c, const struct timespec *now) ASSERT(!flow_new_entry); /* Incomplete flow at end of cycle */ - for (idx = 0; idx < FLOW_MAX; idx++) { - union flow *flow = &flowtab[idx]; + /* Check which flows we might need to close first, but don't free them + * yet as it's not safe to do that in the middle of flow_foreach(). + */ + flow_foreach(flow) { bool closed = false; - switch (flow->f.state) { - case FLOW_STATE_FREE: { - unsigned skip = flow->free.n; - - /* First entry of a free cluster must have n >= 1 */ - ASSERT(skip); - - if (free_head) { - /* Merge into preceding free cluster */ - free_head->n += flow->free.n; - flow->free.n = flow->free.next = 0; - } else { - /* New free cluster, add to chain */ - free_head = &flow->free; - *last_next = idx; - last_next = &free_head->next; - } - - /* Skip remaining empty entries */ - idx += skip - 1; - continue; - } - - case FLOW_STATE_NEW: - case FLOW_STATE_INI: - case FLOW_STATE_TGT: - case FLOW_STATE_TYPED: - /* Incomplete flow at end of cycle */ - ASSERT(false); - break; - - case FLOW_STATE_ACTIVE: - /* Nothing to do */ - break; - - default: - ASSERT(false); - } - switch (flow->f.type) { case FLOW_TYPE_NONE: ASSERT(false); @@ -846,7 +835,7 @@ void flow_defer_handler(const struct ctx *c, const struct timespec *now) closed = icmp_ping_timer(c, &flow->ping, now); break; case FLOW_UDP: - closed = udp_flow_defer(&flow->udp); + closed = udp_flow_defer(c, &flow->udp, now); if (!closed && timer) closed = udp_flow_timer(c, &flow->udp, now); break; @@ -855,30 +844,321 @@ void flow_defer_handler(const struct ctx *c, const struct timespec *now) ; } - if (closed) { - flow_set_state(&flow->f, FLOW_STATE_FREE); - memset(flow, 0, sizeof(*flow)); + to_free[FLOW_IDX(flow)] = closed; + } + + /* Second step: actually free the flows */ + flow_foreach_slot(flow) { + switch (flow->f.state) { + case FLOW_STATE_FREE: { + unsigned skip = flow->free.n; + + /* First entry of a free cluster must have n >= 1 */ + ASSERT(skip); if (free_head) { - /* Add slot to current free cluster */ - ASSERT(idx == FLOW_IDX(free_head) + free_head->n); - free_head->n++; + /* Merge into preceding free cluster */ + free_head->n += flow->free.n; flow->free.n = flow->free.next = 0; } else { - /* Create new free cluster */ + /* New free cluster, add to chain */ free_head = &flow->free; - free_head->n = 1; - *last_next = idx; + *last_next = FLOW_IDX(flow); last_next = &free_head->next; } - } else { - free_head = NULL; + + /* Skip remaining empty entries */ + flow += skip - 1; + continue; + } + + case FLOW_STATE_NEW: + case FLOW_STATE_INI: + case FLOW_STATE_TGT: + case FLOW_STATE_TYPED: + /* Incomplete flow at end of cycle */ + ASSERT(false); + break; + + case FLOW_STATE_ACTIVE: + if (to_free[FLOW_IDX(flow)]) { + flow_set_state(&flow->f, FLOW_STATE_FREE); + memset(flow, 0, sizeof(*flow)); + + if (free_head) { + /* Add slot to current free cluster */ + ASSERT(FLOW_IDX(flow) == + FLOW_IDX(free_head) + free_head->n); + free_head->n++; + flow->free.n = flow->free.next = 0; + } else { + /* Create new free cluster */ + free_head = &flow->free; + free_head->n = 1; + *last_next = FLOW_IDX(flow); + last_next = &free_head->next; + } + } else { + free_head = NULL; + } + break; + + default: + ASSERT(false); } } *last_next = FLOW_MAX; } +/** + * flow_migrate_source_rollback() - Disable repair mode, return failure + * @c: Execution context + * @bound: No need to roll back flow indices >= @bound + * @ret: Negative error code + * + * Return: @ret + */ +static int flow_migrate_source_rollback(struct ctx *c, unsigned bound, int ret) +{ + union flow *flow; + + debug("...roll back migration"); + + foreach_established_tcp_flow(flow) { + if (FLOW_IDX(flow) >= bound) + break; + if (tcp_flow_repair_off(c, &flow->tcp)) + die("Failed to roll back TCP_REPAIR mode"); + } + + if (repair_flush(c)) + die("Failed to roll back TCP_REPAIR mode"); + + return ret; +} + +/** + * flow_migrate_need_repair() - Do we need to set repair mode for any flow? + * + * Return: true if repair mode is needed, false otherwise + */ +static bool flow_migrate_need_repair(void) +{ + union flow *flow; + + foreach_established_tcp_flow(flow) + return true; + + return false; +} + +/** + * flow_migrate_repair_all() - Turn repair mode on or off for all flows + * @c: Execution context + * @enable: Switch repair mode on if set, off otherwise + * + * Return: 0 on success, negative error code on failure + */ +static int flow_migrate_repair_all(struct ctx *c, bool enable) +{ + union flow *flow; + int rc; + + /* If we don't have a repair helper, there's nothing we can do */ + if (c->fd_repair < 0) + return 0; + + foreach_established_tcp_flow(flow) { + if (enable) + rc = tcp_flow_repair_on(c, &flow->tcp); + else + rc = tcp_flow_repair_off(c, &flow->tcp); + + if (rc) { + debug("Can't %s repair mode: %s", + enable ? "enable" : "disable", strerror_(-rc)); + return flow_migrate_source_rollback(c, FLOW_IDX(flow), + rc); + } + } + + if ((rc = repair_flush(c))) { + debug("Can't %s repair mode: %s", + enable ? "enable" : "disable", strerror_(-rc)); + return flow_migrate_source_rollback(c, FLOW_IDX(flow), rc); + } + + return 0; +} + +/** + * flow_migrate_source_pre() - Prepare flows for migration: enable repair mode + * @c: Execution context + * @stage: Migration stage information (unused) + * @fd: Migration file descriptor (unused) + * + * Return: 0 on success, positive error code on failure + */ +int flow_migrate_source_pre(struct ctx *c, const struct migrate_stage *stage, + int fd) +{ + int rc; + + (void)stage; + (void)fd; + + if (flow_migrate_need_repair()) + repair_wait(c); + + if ((rc = flow_migrate_repair_all(c, true))) + return -rc; + + return 0; +} + +/** + * flow_migrate_source() - Dump all the remaining information and send data + * @c: Execution context (unused) + * @stage: Migration stage information (unused) + * @fd: Migration file descriptor + * + * Return: 0 on success, positive error code on failure + */ +int flow_migrate_source(struct ctx *c, const struct migrate_stage *stage, + int fd) +{ + uint32_t count = 0; + bool first = true; + union flow *flow; + int rc; + + (void)c; + (void)stage; + + /* If we don't have a repair helper, we can't migrate TCP flows */ + if (c->fd_repair >= 0) { + foreach_established_tcp_flow(flow) + count++; + } + + count = htonl(count); + if (write_all_buf(fd, &count, sizeof(count))) { + rc = errno; + err_perror("Can't send flow count (%u)", ntohl(count)); + return flow_migrate_source_rollback(c, FLOW_MAX, rc); + } + + debug("Sending %u flows", ntohl(count)); + + if (!count) + return 0; + + /* Dump and send information that can be stored in the flow table. + * + * Limited rollback options here: if we fail to transfer any data (that + * is, on the first flow), undo everything and resume. Otherwise, the + * stream might now be inconsistent, and we might have closed listening + * TCP sockets, so just terminate. + */ + foreach_established_tcp_flow(flow) { + rc = tcp_flow_migrate_source(fd, &flow->tcp); + if (rc) { + flow_err(flow, "Can't send data: %s", + strerror_(-rc)); + if (!first) + die("Inconsistent migration state, exiting"); + + return flow_migrate_source_rollback(c, FLOW_MAX, -rc); + } + + first = false; + } + + /* And then "extended" data (including window data we saved previously): + * the target needs to set repair mode on sockets before it can set + * this stuff, but it needs sockets (and flows) for that. + * + * This also closes sockets so that the target can start connecting + * theirs: you can't sendmsg() to queues (using the socket) if the + * socket is not connected (EPIPE), not even in repair mode. And the + * target needs to restore queues now because we're sending the data. + * + * So, no rollback here, just try as hard as we can. Tolerate per-flow + * failures but not if the stream might be inconsistent (reported here + * as EIO). + */ + foreach_established_tcp_flow(flow) { + rc = tcp_flow_migrate_source_ext(fd, &flow->tcp); + if (rc) { + flow_err(flow, "Can't send extended data: %s", + strerror_(-rc)); + + if (rc == -EIO) + die("Inconsistent migration state, exiting"); + } + } + + return 0; +} + +/** + * flow_migrate_target() - Receive flows and insert in flow table + * @c: Execution context + * @stage: Migration stage information (unused) + * @fd: Migration file descriptor + * + * Return: 0 on success, positive error code on failure + */ +int flow_migrate_target(struct ctx *c, const struct migrate_stage *stage, + int fd) +{ + uint32_t count; + unsigned i; + int rc; + + (void)stage; + + if (read_all_buf(fd, &count, sizeof(count))) + return errno; + + count = ntohl(count); + debug("Receiving %u flows", count); + + if (!count) + return 0; + + repair_wait(c); + + if ((rc = flow_migrate_repair_all(c, true))) + return -rc; + + repair_flush(c); + + /* TODO: flow header with type, instead? */ + for (i = 0; i < count; i++) { + rc = tcp_flow_migrate_target(c, fd); + if (rc) { + flow_dbg(FLOW(i), "Migration data failure, abort: %s", + strerror_(-rc)); + return -rc; + } + } + + repair_flush(c); + + for (i = 0; i < count; i++) { + rc = tcp_flow_migrate_target_ext(c, &flowtab[i].tcp, fd); + if (rc) { + flow_dbg(FLOW(i), "Migration data failure, abort: %s", + strerror_(-rc)); + return -rc; + } + } + + return 0; +} + /** * flow_init() - Initialise flow related data structures */ diff --git a/flow.h b/flow.h index 24ba3ef..cac618a 100644 --- a/flow.h +++ b/flow.h @@ -243,18 +243,27 @@ flow_sidx_t flow_lookup_af(const struct ctx *c, const void *eaddr, const void *oaddr, in_port_t eport, in_port_t oport); flow_sidx_t flow_lookup_sa(const struct ctx *c, uint8_t proto, uint8_t pif, - const void *esa, in_port_t oport); + const void *esa, + const union inany_addr *oaddr, in_port_t oport); union flow; void flow_init(void); void flow_defer_handler(const struct ctx *c, const struct timespec *now); +int flow_migrate_source_early(struct ctx *c, const struct migrate_stage *stage, + int fd); +int flow_migrate_source_pre(struct ctx *c, const struct migrate_stage *stage, + int fd); +int flow_migrate_source(struct ctx *c, const struct migrate_stage *stage, + int fd); +int flow_migrate_target(struct ctx *c, const struct migrate_stage *stage, + int fd); -void flow_log_(const struct flow_common *f, int pri, const char *fmt, ...) - __attribute__((format(printf, 3, 4))); - -#define flow_log(f_, pri, ...) flow_log_(&(f_)->f, (pri), __VA_ARGS__) +void flow_log_(const struct flow_common *f, bool newline, int pri, + const char *fmt, ...) + __attribute__((format(printf, 4, 5))); +#define flow_log(f_, pri, ...) flow_log_(&(f_)->f, true, (pri), __VA_ARGS__) #define flow_dbg(f, ...) flow_log((f), LOG_DEBUG, __VA_ARGS__) #define flow_err(f, ...) flow_log((f), LOG_ERR, __VA_ARGS__) @@ -264,6 +273,16 @@ void flow_log_(const struct flow_common *f, int pri, const char *fmt, ...) flow_dbg((f), __VA_ARGS__); \ } while (0) +#define flow_log_perror_(f, pri, ...) \ + do { \ + int errno_ = errno; \ + flow_log_((f), false, (pri), __VA_ARGS__); \ + logmsg(true, true, (pri), ": %s", strerror_(errno_)); \ + } while (0) + +#define flow_dbg_perror(f_, ...) flow_log_perror_(&(f_)->f, LOG_DEBUG, __VA_ARGS__) +#define flow_perror(f_, ...) flow_log_perror_(&(f_)->f, LOG_ERR, __VA_ARGS__) + void flow_log_details_(const struct flow_common *f, int pri, enum flow_state state); #define flow_log_details(f_, pri) \ diff --git a/flow_table.h b/flow_table.h index f15db53..3f3f4b7 100644 --- a/flow_table.h +++ b/flow_table.h @@ -50,6 +50,42 @@ extern union flow flowtab[]; #define flow_foreach_sidei(sidei_) \ for ((sidei_) = INISIDE; (sidei_) < SIDES; (sidei_)++) + +/** + * flow_foreach_slot() - Step through each flow table entry + * @flow: Takes values of pointer to each flow table entry + * + * Includes FREE slots. + */ +#define flow_foreach_slot(flow) \ + for ((flow) = flowtab; FLOW_IDX(flow) < FLOW_MAX; (flow)++) + +/** + * flow_foreach() - Step through each active flow + * @flow: Takes values of pointer to each active flow + */ +#define flow_foreach(flow) \ + flow_foreach_slot((flow)) \ + if ((flow)->f.state == FLOW_STATE_FREE) \ + (flow) += (flow)->free.n - 1; \ + else if ((flow)->f.state != FLOW_STATE_ACTIVE) { \ + flow_err((flow), "Bad flow state during traversal"); \ + continue; \ + } else + +/** + * flow_foreach_of_type() - Step through each active flow of given type + * @flow: Takes values of pointer to each flow + * @type_: Type of flow to traverse + */ +#define flow_foreach_of_type(flow, type_) \ + flow_foreach((flow)) \ + if ((flow)->f.type != (type_)) \ + /* NOLINTNEXTLINE(bugprone-branch-clone) */ \ + continue; \ + else + + /** flow_idx() - Index of flow from common structure * @f: Common flow fields pointer * @@ -57,6 +93,7 @@ extern union flow flowtab[]; */ static inline unsigned flow_idx(const struct flow_common *f) { + /* NOLINTNEXTLINE(clang-analyzer-security.PointerSub) */ return (union flow *)f - flowtab; } @@ -161,15 +198,16 @@ const struct flowside *flow_initiate_af(union flow *flow, uint8_t pif, sa_family_t af, const void *saddr, in_port_t sport, const void *daddr, in_port_t dport); -const struct flowside *flow_initiate_sa(union flow *flow, uint8_t pif, - const union sockaddr_inany *ssa, - in_port_t dport); +struct flowside *flow_initiate_sa(union flow *flow, uint8_t pif, + const union sockaddr_inany *ssa, + const union inany_addr *daddr, + in_port_t dport); const struct flowside *flow_target_af(union flow *flow, uint8_t pif, sa_family_t af, const void *saddr, in_port_t sport, const void *daddr, in_port_t dport); -const struct flowside *flow_target(const struct ctx *c, union flow *flow, - uint8_t proto); +struct flowside *flow_target(const struct ctx *c, union flow *flow, + uint8_t proto); union flow *flow_set_type(union flow *flow, enum flow_type type); #define FLOW_SET_TYPE(flow_, t_, var_) (&flow_set_type((flow_), (t_))->var_) diff --git a/fwd.c b/fwd.c index 0b7f8b1..250cf56 100644 --- a/fwd.c +++ b/fwd.c @@ -323,6 +323,30 @@ static bool fwd_guest_accessible(const struct ctx *c, return fwd_guest_accessible6(c, &addr->a6); } +/** + * nat_outbound() - Apply address translation for outbound (TAP to HOST) + * @c: Execution context + * @addr: Input address (as seen on TAP interface) + * @translated: Output address (as seen on HOST interface) + * + * Only handles translations that depend *only* on the address. Anything + * related to specific ports or flows is handled elsewhere. + */ +static void nat_outbound(const struct ctx *c, const union inany_addr *addr, + union inany_addr *translated) +{ + if (inany_equals4(addr, &c->ip4.map_host_loopback)) + *translated = inany_loopback4; + else if (inany_equals6(addr, &c->ip6.map_host_loopback)) + *translated = inany_loopback6; + else if (inany_equals4(addr, &c->ip4.map_guest_addr)) + *translated = inany_from_v4(c->ip4.addr); + else if (inany_equals6(addr, &c->ip6.map_guest_addr)) + translated->a6 = c->ip6.addr; + else + *translated = *addr; +} + /** * fwd_nat_from_tap() - Determine to forward a flow from the tap interface * @c: Execution context @@ -342,16 +366,8 @@ uint8_t fwd_nat_from_tap(const struct ctx *c, uint8_t proto, else if (is_dns_flow(proto, ini) && inany_equals6(&ini->oaddr, &c->ip6.dns_match)) tgt->eaddr.a6 = c->ip6.dns_host; - else if (inany_equals4(&ini->oaddr, &c->ip4.map_host_loopback)) - tgt->eaddr = inany_loopback4; - else if (inany_equals6(&ini->oaddr, &c->ip6.map_host_loopback)) - tgt->eaddr = inany_loopback6; - else if (inany_equals4(&ini->oaddr, &c->ip4.map_guest_addr)) - tgt->eaddr = inany_from_v4(c->ip4.addr); - else if (inany_equals6(&ini->oaddr, &c->ip6.map_guest_addr)) - tgt->eaddr.a6 = c->ip6.addr; else - tgt->eaddr = ini->oaddr; + nat_outbound(c, &ini->oaddr, &tgt->eaddr); tgt->eport = ini->oport; @@ -402,7 +418,7 @@ uint8_t fwd_nat_from_splice(const struct ctx *c, uint8_t proto, else tgt->eaddr = inany_loopback6; - /* Preserve the specific loopback adddress used, but let the kernel pick + /* Preserve the specific loopback address used, but let the kernel pick * a source port on the target side */ tgt->oaddr = ini->eaddr; @@ -423,6 +439,42 @@ uint8_t fwd_nat_from_splice(const struct ctx *c, uint8_t proto, return PIF_HOST; } +/** + * nat_inbound() - Apply address translation for inbound (HOST to TAP) + * @c: Execution context + * @addr: Input address (as seen on HOST interface) + * @translated: Output address (as seen on TAP interface) + * + * Return: true on success, false if it couldn't translate the address + * + * Only handles translations that depend *only* on the address. Anything + * related to specific ports or flows is handled elsewhere. + */ +bool nat_inbound(const struct ctx *c, const union inany_addr *addr, + union inany_addr *translated) +{ + if (!IN4_IS_ADDR_UNSPECIFIED(&c->ip4.map_host_loopback) && + inany_equals4(addr, &in4addr_loopback)) { + /* Specifically 127.0.0.1, not 127.0.0.0/8 */ + *translated = inany_from_v4(c->ip4.map_host_loopback); + } else if (!IN6_IS_ADDR_UNSPECIFIED(&c->ip6.map_host_loopback) && + inany_equals6(addr, &in6addr_loopback)) { + translated->a6 = c->ip6.map_host_loopback; + } else if (!IN4_IS_ADDR_UNSPECIFIED(&c->ip4.map_guest_addr) && + inany_equals4(addr, &c->ip4.addr)) { + *translated = inany_from_v4(c->ip4.map_guest_addr); + } else if (!IN6_IS_ADDR_UNSPECIFIED(&c->ip6.map_guest_addr) && + inany_equals6(addr, &c->ip6.addr)) { + translated->a6 = c->ip6.map_guest_addr; + } else if (fwd_guest_accessible(c, addr)) { + *translated = *addr; + } else { + return false; + } + + return true; +} + /** * fwd_nat_from_host() - Determine to forward a flow from the host interface * @c: Execution context @@ -443,7 +495,7 @@ uint8_t fwd_nat_from_host(const struct ctx *c, uint8_t proto, else if (proto == IPPROTO_UDP) tgt->eport += c->udp.fwd_in.delta[tgt->eport]; - if (c->mode == MODE_PASTA && inany_is_loopback(&ini->eaddr) && + if (!c->no_splice && inany_is_loopback(&ini->eaddr) && (proto == IPPROTO_TCP || proto == IPPROTO_UDP)) { /* spliceable */ @@ -479,20 +531,7 @@ uint8_t fwd_nat_from_host(const struct ctx *c, uint8_t proto, return PIF_SPLICE; } - if (!IN4_IS_ADDR_UNSPECIFIED(&c->ip4.map_host_loopback) && - inany_equals4(&ini->eaddr, &in4addr_loopback)) { - /* Specifically 127.0.0.1, not 127.0.0.0/8 */ - tgt->oaddr = inany_from_v4(c->ip4.map_host_loopback); - } else if (!IN6_IS_ADDR_UNSPECIFIED(&c->ip6.map_host_loopback) && - inany_equals6(&ini->eaddr, &in6addr_loopback)) { - tgt->oaddr.a6 = c->ip6.map_host_loopback; - } else if (!IN4_IS_ADDR_UNSPECIFIED(&c->ip4.map_guest_addr) && - inany_equals4(&ini->eaddr, &c->ip4.addr)) { - tgt->oaddr = inany_from_v4(c->ip4.map_guest_addr); - } else if (!IN6_IS_ADDR_UNSPECIFIED(&c->ip6.map_guest_addr) && - inany_equals6(&ini->eaddr, &c->ip6.addr)) { - tgt->oaddr.a6 = c->ip6.map_guest_addr; - } else if (!fwd_guest_accessible(c, &ini->eaddr)) { + if (!nat_inbound(c, &ini->eaddr, &tgt->oaddr)) { if (inany_v4(&ini->eaddr)) { if (IN4_IS_ADDR_UNSPECIFIED(&c->ip4.our_tap_addr)) /* No source address we can use */ @@ -501,8 +540,6 @@ uint8_t fwd_nat_from_host(const struct ctx *c, uint8_t proto, } else { tgt->oaddr.a6 = c->ip6.our_tap_ll; } - } else { - tgt->oaddr = ini->eaddr; } tgt->oport = ini->eport; diff --git a/fwd.h b/fwd.h index 3562f3c..0458a3c 100644 --- a/fwd.h +++ b/fwd.h @@ -7,6 +7,7 @@ #ifndef FWD_H #define FWD_H +union inany_addr; struct flowside; /* Number of ports for both TCP and UDP */ @@ -47,6 +48,8 @@ void fwd_scan_ports_udp(struct fwd_ports *fwd, const struct fwd_ports *rev, const struct fwd_ports *tcp_rev); void fwd_scan_ports_init(struct ctx *c); +bool nat_inbound(const struct ctx *c, const union inany_addr *addr, + union inany_addr *translated); uint8_t fwd_nat_from_tap(const struct ctx *c, uint8_t proto, const struct flowside *ini, struct flowside *tgt); uint8_t fwd_nat_from_splice(const struct ctx *c, uint8_t proto, diff --git a/hooks/pre-push b/hooks/pre-push index 33a2052..8dbfa5f 100755 --- a/hooks/pre-push +++ b/hooks/pre-push @@ -56,6 +56,7 @@ cd .. make pkgs scp passt passt.avx2 passt.1 qrap qrap.1 "${USER_HOST}:${BIN}" scp pasta pasta.avx2 pasta.1 "${USER_HOST}:${BIN}" +scp passt-repair passt-repair.1 "${USER_HOST}:${BIN}" ssh "${USER_HOST}" "rm -f ${BIN}/*.deb" ssh "${USER_HOST}" "rm -f ${BIN}/*.rpm" diff --git a/icmp.c b/icmp.c index f514dbc..7e2b342 100644 --- a/icmp.c +++ b/icmp.c @@ -85,7 +85,7 @@ void icmp_sock_handler(const struct ctx *c, union epoll_ref ref) n = recvfrom(ref.fd, buf, sizeof(buf), 0, &sr.sa, &sl); if (n < 0) { - flow_err(pingf, "recvfrom() error: %s", strerror(errno)); + flow_perror(pingf, "recvfrom() error"); return; } @@ -150,7 +150,7 @@ unexpected: static void icmp_ping_close(const struct ctx *c, const struct icmp_ping_flow *pingf) { - epoll_ctl(c->epollfd, EPOLL_CTL_DEL, pingf->sock, NULL); + epoll_del(c, pingf->sock); close(pingf->sock); flow_hash_remove(c, FLOW_SIDX(pingf, INISIDE)); } @@ -300,8 +300,7 @@ int icmp_tap_handler(const struct ctx *c, uint8_t pif, sa_family_t af, pif_sockaddr(c, &sa, &sl, PIF_HOST, &tgt->eaddr, 0); if (sendto(pingf->sock, pkt, l4len, MSG_NOSIGNAL, &sa.sa, sl) < 0) { - flow_dbg(pingf, "failed to relay request to socket: %s", - strerror(errno)); + flow_dbg_perror(pingf, "failed to relay request to socket"); } else { flow_dbg(pingf, "echo request to socket, ID: %"PRIu16", seq: %"PRIu16, diff --git a/inany.h b/inany.h index 6a12c29..7ca5cbd 100644 --- a/inany.h +++ b/inany.h @@ -237,23 +237,30 @@ static inline void inany_from_af(union inany_addr *aa, } /** inany_from_sockaddr - Extract IPv[46] address and port number from sockaddr - * @aa: Pointer to store IPv[46] address + * @dst: Pointer to store IPv[46] address (output) * @port: Pointer to store port number, host order - * @addr: AF_INET or AF_INET6 socket address + * @addr: Socket address + * + * Return: 0 on success, -1 on error (bad address family) */ -static inline void inany_from_sockaddr(union inany_addr *aa, in_port_t *port, - const union sockaddr_inany *sa) +static inline int inany_from_sockaddr(union inany_addr *dst, in_port_t *port, + const void *addr) { + const union sockaddr_inany *sa = (const union sockaddr_inany *)addr; + if (sa->sa_family == AF_INET6) { - inany_from_af(aa, AF_INET6, &sa->sa6.sin6_addr); + inany_from_af(dst, AF_INET6, &sa->sa6.sin6_addr); *port = ntohs(sa->sa6.sin6_port); - } else if (sa->sa_family == AF_INET) { - inany_from_af(aa, AF_INET, &sa->sa4.sin_addr); - *port = ntohs(sa->sa4.sin_port); - } else { - /* Not valid to call with other address families */ - ASSERT(0); + return 0; } + + if (sa->sa_family == AF_INET) { + inany_from_af(dst, AF_INET, &sa->sa4.sin_addr); + *port = ntohs(sa->sa4.sin_port); + return 0; + } + + return -1; } /** inany_siphash_feed- Fold IPv[46] address into an in-progress siphash diff --git a/iov.c b/iov.c index 3f9e229..91e87a7 100644 --- a/iov.c +++ b/iov.c @@ -26,7 +26,8 @@ #include "iov.h" -/* iov_skip_bytes() - Skip leading bytes of an IO vector +/** + * iov_skip_bytes() - Skip leading bytes of an IO vector * @iov: IO vector * @n: Number of entries in @iov * @skip: Number of leading bytes of @iov to skip @@ -56,8 +57,8 @@ size_t iov_skip_bytes(const struct iovec *iov, size_t n, } /** - * iov_from_buf - Copy data from a buffer to an I/O vector (struct iovec) - * efficiently. + * iov_from_buf() - Copy data from a buffer to an I/O vector (struct iovec) + * efficiently. * * @iov: Pointer to the array of struct iovec describing the * scatter/gather I/O vector. @@ -68,7 +69,6 @@ size_t iov_skip_bytes(const struct iovec *iov, size_t n, * * Returns: The number of bytes successfully copied. */ -/* cppcheck-suppress unusedFunction */ size_t iov_from_buf(const struct iovec *iov, size_t iov_cnt, size_t offset, const void *buf, size_t bytes) { @@ -97,8 +97,8 @@ size_t iov_from_buf(const struct iovec *iov, size_t iov_cnt, } /** - * iov_to_buf - Copy data from a scatter/gather I/O vector (struct iovec) to - * a buffer efficiently. + * iov_to_buf() - Copy data from a scatter/gather I/O vector (struct iovec) to + * a buffer efficiently. * * @iov: Pointer to the array of struct iovec describing the scatter/gather * I/O vector. @@ -137,8 +137,8 @@ size_t iov_to_buf(const struct iovec *iov, size_t iov_cnt, } /** - * iov_size - Calculate the total size of a scatter/gather I/O vector - * (struct iovec). + * iov_size() - Calculate the total size of a scatter/gather I/O vector + * (struct iovec). * * @iov: Pointer to the array of struct iovec describing the * scatter/gather I/O vector. @@ -156,3 +156,95 @@ size_t iov_size(const struct iovec *iov, size_t iov_cnt) return len; } + +/** + * iov_tail_prune() - Remove any unneeded buffers from an IOV tail + * @tail: IO vector tail (modified) + * + * If an IOV tail's offset is large enough, it may not include any bytes from + * the first (or first several) buffers in the underlying IO vector. Modify the + * tail's representation so it contains the same logical bytes, but only + * includes buffers that are actually needed. This will avoid stepping through + * unnecessary elements of the underlying IO vector on future operations. + * + * Return: true if the tail still contains any bytes, otherwise false + */ +bool iov_tail_prune(struct iov_tail *tail) +{ + size_t i; + + i = iov_skip_bytes(tail->iov, tail->cnt, tail->off, &tail->off); + tail->iov += i; + tail->cnt -= i; + + return !!tail->cnt; +} + +/** + * iov_tail_size - Calculate the total size of an IO vector tail + * @tail: IO vector tail + * + * Returns: The total size in bytes. + */ +size_t iov_tail_size(struct iov_tail *tail) +{ + iov_tail_prune(tail); + return iov_size(tail->iov, tail->cnt) - tail->off; +} + +/** + * iov_peek_header_() - Get pointer to a header from an IOV tail + * @tail: IOV tail to get header from + * @len: Length of header to get, in bytes + * @align: Required alignment of header, in bytes + * + * @tail may be pruned, but will represent the same bytes as before. + * + * Returns: Pointer to the first @len logical bytes of the tail, NULL if that + * overruns the IO vector, is not contiguous or doesn't have the + * requested alignment. + */ +/* cppcheck-suppress [staticFunction,unmatchedSuppression] */ +void *iov_peek_header_(struct iov_tail *tail, size_t len, size_t align) +{ + char *p; + + if (!iov_tail_prune(tail)) + return NULL; /* Nothing left */ + + if (tail->off + len < tail->off) + return NULL; /* Overflow */ + + if (tail->off + len > tail->iov[0].iov_len) + return NULL; /* Not contiguous */ + + p = (char *)tail->iov[0].iov_base + tail->off; + if ((uintptr_t)p % align) + return NULL; /* not aligned */ + + return p; +} + +/** + * iov_remove_header_() - Remove a header from an IOV tail + * @tail: IOV tail to remove header from (modified) + * @len: Length of header to remove, in bytes + * @align: Required alignment of header, in bytes + * + * On success, @tail is updated so that it longer includes the bytes of the + * returned header. + * + * Returns: Pointer to the first @len logical bytes of the tail, NULL if that + * overruns the IO vector, is not contiguous or doesn't have the + * requested alignment. + */ +void *iov_remove_header_(struct iov_tail *tail, size_t len, size_t align) +{ + char *p = iov_peek_header_(tail, len, align); + + if (!p) + return NULL; + + tail->off = tail->off + len; + return p; +} diff --git a/iov.h b/iov.h index a9e1722..9855bf0 100644 --- a/iov.h +++ b/iov.h @@ -28,4 +28,80 @@ size_t iov_from_buf(const struct iovec *iov, size_t iov_cnt, size_t iov_to_buf(const struct iovec *iov, size_t iov_cnt, size_t offset, void *buf, size_t bytes); size_t iov_size(const struct iovec *iov, size_t iov_cnt); + +/* + * DOC: Theory of Operation, struct iov_tail + * + * Sometimes a single logical network frame is split across multiple buffers, + * represented by an IO vector (struct iovec[]). We often want to process this + * one header / network layer at a time. So, it's useful to maintain a "tail" + * of the vector representing the parts we haven't yet extracted. + * + * The headers we extract need not line up with buffer boundaries (though we do + * assume they're contiguous within a single buffer for now). So, we could + * represent that tail as another struct iovec[], but that would mean copying + * the whole array of struct iovecs, just so we can adjust the offset and length + * on the first one. + * + * So, instead represent the tail as pointer into an existing struct iovec[], + * with an explicit offset for where the "tail" starts within it. If we extract + * enough headers that some buffers of the original vector no longer contain + * part of the tail, we (lazily) advance our struct iovec * to the first buffer + * we still need, and adjust the vector length and offset to match. + */ + +/** + * struct iov_tail - An IO vector which may have some headers logically removed + * @iov: IO vector + * @cnt: Number of entries in @iov + * @off: Current offset in @iov + */ +struct iov_tail { + const struct iovec *iov; + size_t cnt, off; +}; + +/** + * IOV_TAIL() - Create a new IOV tail + * @iov_: IO vector to create tail from + * @cnt_: Length of the IO vector at @iov_ + * @off_: Byte offset in the IO vector where the tail begins + */ +#define IOV_TAIL(iov_, cnt_, off_) \ + (struct iov_tail){ .iov = (iov_), .cnt = (cnt_), .off = (off_) } + +bool iov_tail_prune(struct iov_tail *tail); +size_t iov_tail_size(struct iov_tail *tail); +void *iov_peek_header_(struct iov_tail *tail, size_t len, size_t align); +void *iov_remove_header_(struct iov_tail *tail, size_t len, size_t align); + +/** + * IOV_PEEK_HEADER() - Get typed pointer to a header from an IOV tail + * @tail_: IOV tail to get header from + * @type_: Data type of the header + * + * @tail_ may be pruned, but will represent the same bytes as before. + * + * Returns: Pointer of type (@type_ *) located at the start of @tail_, NULL if + * we can't get a contiguous and aligned pointer. + */ +#define IOV_PEEK_HEADER(tail_, type_) \ + ((type_ *)(iov_peek_header_((tail_), \ + sizeof(type_), __alignof__(type_)))) + +/** + * IOV_REMOVE_HEADER() - Remove and return typed header from an IOV tail + * @tail_: IOV tail to remove header from (modified) + * @type_: Data type of the header to remove + * + * On success, @tail_ is updated so that it longer includes the bytes of the + * returned header. + * + * Returns: Pointer of type (@type_ *) located at the old start of @tail_, NULL + * if we can't get a contiguous and aligned pointer. + */ +#define IOV_REMOVE_HEADER(tail_, type_) \ + ((type_ *)(iov_remove_header_((tail_), \ + sizeof(type_), __alignof__(type_)))) + #endif /* IOVEC_H */ diff --git a/ip.h b/ip.h index 1544dbf..24509d9 100644 --- a/ip.h +++ b/ip.h @@ -36,13 +36,14 @@ .tos = 0, \ .tot_len = 0, \ .id = 0, \ - .frag_off = 0, \ + .frag_off = htons(IP_DF), \ .ttl = 0xff, \ .protocol = (proto), \ .saddr = 0, \ .daddr = 0, \ } #define L2_BUF_IP4_PSUM(proto) ((uint32_t)htons_constant(0x4500) + \ + (uint32_t)htons_constant(IP_DF) + \ (uint32_t)htons(0xff00 | (proto))) @@ -90,10 +91,34 @@ struct ipv6_opt_hdr { */ } __attribute__((packed)); /* required for some archs */ +/** + * ip6_set_flow_lbl() - Set flow label in an IPv6 header + * @ip6h: Pointer to IPv6 header, updated + * @flow: Set @ip6h flow label to the low 20 bits of this integer + */ +static inline void ip6_set_flow_lbl(struct ipv6hdr *ip6h, uint32_t flow) +{ + ip6h->flow_lbl[0] = (flow >> 16) & 0xf; + ip6h->flow_lbl[1] = (flow >> 8) & 0xff; + ip6h->flow_lbl[2] = (flow >> 0) & 0xff; +} + +/** ip6_get_flow_lbl() - Get flow label from an IPv6 header + * @ip6h: Pointer to IPv6 header + * + * Return: flow label from @ip6h as an integer (<= 20 bits) + */ +static inline uint32_t ip6_get_flow_lbl(const struct ipv6hdr *ip6h) +{ + return (ip6h->flow_lbl[0] & 0xf) << 16 | + ip6h->flow_lbl[1] << 8 | + ip6h->flow_lbl[2]; +} + char *ipv6_l4hdr(const struct pool *p, int idx, size_t offset, uint8_t *proto, size_t *dlen); -/* IPv6 link-local all-nodes multicast adddress, ff02::1 */ +/* IPv6 link-local all-nodes multicast address, ff02::1 */ static const struct in6_addr in6addr_ll_all_nodes = { .s6_addr = { 0xff, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, @@ -104,4 +129,11 @@ static const struct in6_addr in6addr_ll_all_nodes = { /* IPv4 Limited Broadcast (RFC 919, Section 7), 255.255.255.255 */ static const struct in_addr in4addr_broadcast = { 0xffffffff }; +#ifndef IPV4_MIN_MTU +#define IPV4_MIN_MTU 68 +#endif +#ifndef IPV6_MIN_MTU +#define IPV6_MIN_MTU 1280 +#endif + #endif /* IP_H */ diff --git a/isolation.c b/isolation.c index 45fba1e..bbcd23b 100644 --- a/isolation.c +++ b/isolation.c @@ -129,7 +129,7 @@ static void drop_caps_ep_except(uint64_t keep) * additional layer of protection. Executing this requires * CAP_SETPCAP, which we will have within our userns. * - * Note that dropping capabilites from the bounding set limits + * Note that dropping capabilities from the bounding set limits * exec()ed processes, but does not remove them from the effective or * permitted sets, so it doesn't reduce our own capabilities. */ @@ -174,8 +174,8 @@ static void clamp_caps(void) * Should: * - drop unneeded capabilities * - close all open files except for standard streams and the one from --fd - * Musn't: - * - remove filesytem access (we need to access files during setup) + * Mustn't: + * - remove filesystem access (we need to access files during setup) */ void isolate_initial(int argc, char **argv) { @@ -194,7 +194,7 @@ void isolate_initial(int argc, char **argv) * * It's debatable whether it's useful to drop caps when we * retain SETUID and SYS_ADMIN, but we might as well. We drop - * further capabilites in isolate_user() and + * further capabilities in isolate_user() and * isolate_prefork(). */ keep = BIT(CAP_NET_BIND_SERVICE) | BIT(CAP_SETUID) | BIT(CAP_SETGID) | @@ -379,12 +379,21 @@ void isolate_postfork(const struct ctx *c) prctl(PR_SET_DUMPABLE, 0); - if (c->mode == MODE_PASTA) { - prog.len = (unsigned short)ARRAY_SIZE(filter_pasta); - prog.filter = filter_pasta; - } else { + switch (c->mode) { + case MODE_PASST: prog.len = (unsigned short)ARRAY_SIZE(filter_passt); prog.filter = filter_passt; + break; + case MODE_PASTA: + prog.len = (unsigned short)ARRAY_SIZE(filter_pasta); + prog.filter = filter_pasta; + break; + case MODE_VU: + prog.len = (unsigned short)ARRAY_SIZE(filter_vu); + prog.filter = filter_vu; + break; + default: + ASSERT(0); } if (prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0) || diff --git a/log.c b/log.c index 239c8ce..5d7d76f 100644 --- a/log.c +++ b/log.c @@ -56,7 +56,7 @@ bool log_stderr = true; /* Not daemonised, no shell spawned */ * * Return: pointer to @now, or NULL if there was an error retrieving the time */ -const struct timespec *logtime(struct timespec *ts) +static const struct timespec *logtime(struct timespec *ts) { if (clock_gettime(CLOCK_MONOTONIC, ts)) return NULL; @@ -249,6 +249,30 @@ static void logfile_write(bool newline, bool cont, int pri, log_written += n; } +/** + * passt_vsyslog() - vsyslog() implementation not using heap memory + * @newline: Append newline at the end of the message, if missing + * @pri: Facility and level map, same as priority for vsyslog() + * @format: Same as vsyslog() format + * @ap: Same as vsyslog() ap + */ +static void passt_vsyslog(bool newline, int pri, const char *format, va_list ap) +{ + char buf[BUFSIZ]; + int n; + + /* Send without timestamp, the system logger should add it */ + n = snprintf(buf, BUFSIZ, "<%i> %s: ", pri, log_ident); + + n += vsnprintf(buf + n, BUFSIZ - n, format, ap); + + if (newline && format[strlen(format)] != '\n') + n += snprintf(buf + n, BUFSIZ - n, "\n"); + + if (log_sock >= 0 && send(log_sock, buf, n, 0) != n && log_stderr) + FPRINTF(stderr, "Failed to send %i bytes to syslog\n", n); +} + /** * vlogmsg() - Print or send messages to log or output files as configured * @newline: Append newline at the end of the message, if missing @@ -257,6 +281,7 @@ static void logfile_write(bool newline, bool cont, int pri, * @format: Message * @ap: Variable argument list */ +/* cppcheck-suppress [staticFunction,unmatchedSuppression] */ void vlogmsg(bool newline, bool cont, int pri, const char *format, va_list ap) { bool debug_print = (log_mask & LOG_MASK(LOG_DEBUG)) && log_file == -1; @@ -322,7 +347,7 @@ void logmsg_perror(int pri, const char *format, ...) vlogmsg(false, false, pri, format, ap); va_end(ap); - logmsg(true, true, pri, ": %s", strerror(errno_copy)); + logmsg(true, true, pri, ": %s", strerror_(errno_copy)); } /** @@ -373,35 +398,11 @@ void __setlogmask(int mask) setlogmask(mask); } -/** - * passt_vsyslog() - vsyslog() implementation not using heap memory - * @newline: Append newline at the end of the message, if missing - * @pri: Facility and level map, same as priority for vsyslog() - * @format: Same as vsyslog() format - * @ap: Same as vsyslog() ap - */ -void passt_vsyslog(bool newline, int pri, const char *format, va_list ap) -{ - char buf[BUFSIZ]; - int n; - - /* Send without timestamp, the system logger should add it */ - n = snprintf(buf, BUFSIZ, "<%i> %s: ", pri, log_ident); - - n += vsnprintf(buf + n, BUFSIZ - n, format, ap); - - if (newline && format[strlen(format)] != '\n') - n += snprintf(buf + n, BUFSIZ - n, "\n"); - - if (log_sock >= 0 && send(log_sock, buf, n, 0) != n && log_stderr) - FPRINTF(stderr, "Failed to send %i bytes to syslog\n", n); -} - /** * logfile_init() - Open log file and write header with PID, version, path * @name: Identifier for header: passt or pasta * @path: Path to log file - * @size: Maximum size of log file: log_cut_size is calculatd here + * @size: Maximum size of log file: log_cut_size is calculated here */ void logfile_init(const char *name, const char *path, size_t size) { diff --git a/log.h b/log.h index a30b091..08aa88c 100644 --- a/log.h +++ b/log.h @@ -32,13 +32,13 @@ void logmsg_perror(int pri, const char *format, ...) #define die(...) \ do { \ err(__VA_ARGS__); \ - exit(EXIT_FAILURE); \ + _exit(EXIT_FAILURE); \ } while (0) #define die_perror(...) \ do { \ err_perror(__VA_ARGS__); \ - exit(EXIT_FAILURE); \ + _exit(EXIT_FAILURE); \ } while (0) extern int log_trace; @@ -55,7 +55,6 @@ void trace_init(int enable); void __openlog(const char *ident, int option, int facility); void logfile_init(const char *name, const char *path, size_t size); -void passt_vsyslog(bool newline, int pri, const char *format, va_list ap); void __setlogmask(int mask); #endif /* LOG_H */ diff --git a/migrate.c b/migrate.c new file mode 100644 index 0000000..48d63a0 --- /dev/null +++ b/migrate.c @@ -0,0 +1,304 @@ +// SPDX-License-Identifier: GPL-2.0-or-later + +/* PASST - Plug A Simple Socket Transport + * for qemu/UNIX domain socket mode + * + * PASTA - Pack A Subtle Tap Abstraction + * for network namespace/tap device mode + * + * migrate.c - Migration sections, layout, and routines + * + * Copyright (c) 2025 Red Hat GmbH + * Author: Stefano Brivio <sbrivio@redhat.com> + */ + +#include <errno.h> +#include <sys/uio.h> + +#include "util.h" +#include "ip.h" +#include "passt.h" +#include "inany.h" +#include "flow.h" +#include "flow_table.h" + +#include "migrate.h" +#include "repair.h" + +/* Magic identifier for migration data */ +#define MIGRATE_MAGIC 0xB1BB1D1B0BB1D1B0 + +/** + * struct migrate_seen_addrs_v1 - Migratable guest addresses for v1 state stream + * @addr6: Observed guest IPv6 address + * @addr6_ll: Observed guest IPv6 link-local address + * @addr4: Observed guest IPv4 address + * @mac: Observed guest MAC address + */ +struct migrate_seen_addrs_v1 { + struct in6_addr addr6; + struct in6_addr addr6_ll; + struct in_addr addr4; + unsigned char mac[ETH_ALEN]; +} __attribute__((packed)); + +/** + * seen_addrs_source_v1() - Copy and send guest observed addresses from source + * @c: Execution context + * @stage: Migration stage, unused + * @fd: File descriptor for state transfer + * + * Return: 0 on success, positive error code on failure + */ +/* cppcheck-suppress [constParameterCallback, unmatchedSuppression] */ +static int seen_addrs_source_v1(struct ctx *c, + const struct migrate_stage *stage, int fd) +{ + struct migrate_seen_addrs_v1 addrs = { + .addr6 = c->ip6.addr_seen, + .addr6_ll = c->ip6.addr_ll_seen, + .addr4 = c->ip4.addr_seen, + }; + + (void)stage; + + memcpy(addrs.mac, c->guest_mac, sizeof(addrs.mac)); + + if (write_all_buf(fd, &addrs, sizeof(addrs))) + return errno; + + return 0; +} + +/** + * seen_addrs_target_v1() - Receive and use guest observed addresses on target + * @c: Execution context + * @stage: Migration stage, unused + * @fd: File descriptor for state transfer + * + * Return: 0 on success, positive error code on failure + */ +static int seen_addrs_target_v1(struct ctx *c, + const struct migrate_stage *stage, int fd) +{ + struct migrate_seen_addrs_v1 addrs; + + (void)stage; + + if (read_all_buf(fd, &addrs, sizeof(addrs))) + return errno; + + c->ip6.addr_seen = addrs.addr6; + c->ip6.addr_ll_seen = addrs.addr6_ll; + c->ip4.addr_seen = addrs.addr4; + memcpy(c->guest_mac, addrs.mac, sizeof(c->guest_mac)); + + return 0; +} + +/* Stages for version 2 */ +static const struct migrate_stage stages_v2[] = { + { + .name = "observed addresses", + .source = seen_addrs_source_v1, + .target = seen_addrs_target_v1, + }, + { + .name = "prepare flows", + .source = flow_migrate_source_pre, + .target = NULL, + }, + { + .name = "transfer flows", + .source = flow_migrate_source, + .target = flow_migrate_target, + }, + { 0 }, +}; + +/* Supported encoding versions, from latest (most preferred) to oldest */ +static const struct migrate_version versions[] = { + { 2, stages_v2, }, + /* v1 was released, but not widely used. It had bad endianness for the + * MSS and omitted timestamps, which meant it usually wouldn't work. + * Therefore we don't attempt to support compatibility with it. + */ + { 0 }, +}; + +/* Current encoding version */ +#define CURRENT_VERSION (&versions[0]) + +/** + * migrate_source() - Migration as source, send state to hypervisor + * @c: Execution context + * @fd: File descriptor for state transfer + * + * Return: 0 on success, positive error code on failure + */ +static int migrate_source(struct ctx *c, int fd) +{ + const struct migrate_version *v = CURRENT_VERSION; + const struct migrate_header header = { + .magic = htonll_constant(MIGRATE_MAGIC), + .version = htonl(v->id), + .compat_version = htonl(v->id), + }; + const struct migrate_stage *s; + int ret; + + if (write_all_buf(fd, &header, sizeof(header))) { + ret = errno; + err("Can't send migration header: %s, abort", strerror_(ret)); + return ret; + } + + for (s = v->s; s->name; s++) { + if (!s->source) + continue; + + debug("Source side migration stage: %s", s->name); + + if ((ret = s->source(c, s, fd))) { + err("Source migration stage: %s: %s, abort", s->name, + strerror_(ret)); + return ret; + } + } + + return 0; +} + +/** + * migrate_target_read_header() - Read header in target + * @fd: Descriptor for state transfer + * + * Return: version structure on success, NULL on failure with errno set + */ +static const struct migrate_version *migrate_target_read_header(int fd) +{ + const struct migrate_version *v; + struct migrate_header h; + uint32_t id, compat_id; + + if (read_all_buf(fd, &h, sizeof(h))) + return NULL; + + id = ntohl(h.version); + compat_id = ntohl(h.compat_version); + + debug("Source magic: 0x%016" PRIx64 ", version: %u, compat: %u", + ntohll(h.magic), id, compat_id); + + if (ntohll(h.magic) != MIGRATE_MAGIC || !id || !compat_id) { + err("Invalid incoming device state"); + errno = EINVAL; + return NULL; + } + + for (v = versions; v->id; v++) + if (v->id <= id && v->id >= compat_id) + return v; + + errno = ENOTSUP; + err("Unsupported device state version: %u", id); + return NULL; +} + +/** + * migrate_target() - Migration as target, receive state from hypervisor + * @c: Execution context + * @fd: File descriptor for state transfer + * + * Return: 0 on success, positive error code on failure + */ +static int migrate_target(struct ctx *c, int fd) +{ + const struct migrate_version *v; + const struct migrate_stage *s; + int ret; + + if (!(v = migrate_target_read_header(fd))) + return errno; + + for (s = v->s; s->name; s++) { + if (!s->target) + continue; + + debug("Target side migration stage: %s", s->name); + + if ((ret = s->target(c, s, fd))) { + err("Target migration stage: %s: %s, abort", s->name, + strerror_(ret)); + return ret; + } + } + + return 0; +} + +/** + * migrate_init() - Set up things necessary for migration + * @c: Execution context + */ +void migrate_init(struct ctx *c) +{ + c->device_state_result = -1; +} + +/** + * migrate_close() - Close migration channel and connection to passt-repair + * @c: Execution context + */ +void migrate_close(struct ctx *c) +{ + if (c->device_state_fd != -1) { + debug("Closing migration channel, fd: %d", c->device_state_fd); + close(c->device_state_fd); + c->device_state_fd = -1; + c->device_state_result = -1; + } + + repair_close(c); +} + +/** + * migrate_request() - Request a migration of device state + * @c: Execution context + * @fd: fd to transfer state + * @target: Are we the target of the migration? + */ +void migrate_request(struct ctx *c, int fd, bool target) +{ + debug("Migration requested, fd: %d (was %d)", fd, c->device_state_fd); + + if (c->device_state_fd != -1) + migrate_close(c); + + c->device_state_fd = fd; + c->migrate_target = target; +} + +/** + * migrate_handler() - Send/receive passt internal state to/from hypervisor + * @c: Execution context + */ +void migrate_handler(struct ctx *c) +{ + int rc; + + if (c->device_state_fd < 0) + return; + + debug("Handling migration request from fd: %d, target: %d", + c->device_state_fd, c->migrate_target); + + if (c->migrate_target) + rc = migrate_target(c, c->device_state_fd); + else + rc = migrate_source(c, c->device_state_fd); + + migrate_close(c); + + c->device_state_result = rc; +} diff --git a/migrate.h b/migrate.h new file mode 100644 index 0000000..2c51cd9 --- /dev/null +++ b/migrate.h @@ -0,0 +1,51 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later + * Copyright (c) 2025 Red Hat GmbH + * Author: Stefano Brivio <sbrivio@redhat.com> + */ + +#ifndef MIGRATE_H +#define MIGRATE_H + +/** + * struct migrate_header - Migration header from source + * @magic: 0xB1BB1D1B0BB1D1B0, network order + * @version: Highest known, target aborts if too old, network order + * @compat_version: Lowest version compatible with @version, target aborts + * if too new, network order + */ +struct migrate_header { + uint64_t magic; + uint32_t version; + uint32_t compat_version; +} __attribute__((packed)); + +/** + * struct migrate_stage - Callbacks and parameters for one stage of migration + * @name: Stage name (for debugging) + * @source: Callback to implement this stage on the source + * @target: Callback to implement this stage on the target + */ +struct migrate_stage { + const char *name; + int (*source)(struct ctx *c, const struct migrate_stage *stage, int fd); + int (*target)(struct ctx *c, const struct migrate_stage *stage, int fd); + + /* Add here separate rollback callbacks if needed */ +}; + +/** + * struct migrate_version - Stages for a particular protocol version + * @id: Version number, host order + * @s: Ordered array of stages, NULL-terminated + */ +struct migrate_version { + uint32_t id; + const struct migrate_stage *s; +}; + +void migrate_init(struct ctx *c); +void migrate_close(struct ctx *c); +void migrate_request(struct ctx *c, int fd, bool target); +void migrate_handler(struct ctx *c); + +#endif /* MIGRATE_H */ diff --git a/ndp.c b/ndp.c index 37bf7a3..b664034 100644 --- a/ndp.c +++ b/ndp.c @@ -256,7 +256,7 @@ static void ndp_ra(const struct ctx *c, const struct in6_addr *dst) ptr = &ra.var[0]; - if (c->mtu != -1) { + if (c->mtu) { struct opt_mtu *mtu = (struct opt_mtu *)ptr; *mtu = (struct opt_mtu) { .header = { @@ -328,6 +328,7 @@ static void ndp_ra(const struct ctx *c, const struct in6_addr *dst) memcpy(&ra.source_ll.mac, c->our_tap_mac, ETH_ALEN); + /* NOLINTNEXTLINE(clang-analyzer-security.PointerSub) */ ndp_send(c, dst, &ra, ptr - (unsigned char *)&ra); } diff --git a/netlink.c b/netlink.c index 4aba2a3..a052504 100644 --- a/netlink.c +++ b/netlink.c @@ -297,6 +297,10 @@ unsigned int nl_get_ext_if(int s, sa_family_t af) if (!thisifi) continue; /* No interface for this route */ + /* Skip 'lo': we should test IFF_LOOPBACK, but keep it simple */ + if (thisifi == 1) + continue; + /* Skip routes to link-local addresses */ if (af == AF_INET && dst && IN4_IS_PREFIX_LINKLOCAL(dst, rtm->rtm_dst_len)) @@ -320,7 +324,7 @@ unsigned int nl_get_ext_if(int s, sa_family_t af) } if (status < 0) - warn("netlink: RTM_GETROUTE failed: %s", strerror(-status)); + warn("netlink: RTM_GETROUTE failed: %s", strerror_(-status)); if (defifi) { if (ndef > 1) { @@ -351,7 +355,7 @@ unsigned int nl_get_ext_if(int s, sa_family_t af) * * Return: true if a gateway was found, false otherwise */ -bool nl_route_get_def_multipath(struct rtattr *rta, void *gw) +static bool nl_route_get_def_multipath(struct rtattr *rta, void *gw) { int nh_len = RTA_PAYLOAD(rta); struct rtnexthop *rtnh; diff --git a/packet.c b/packet.c index ccfc846..72c6158 100644 --- a/packet.c +++ b/packet.c @@ -22,12 +22,74 @@ #include "util.h" #include "log.h" +/** + * packet_check_range() - Check if a memory range is valid for a pool + * @p: Packet pool + * @ptr: Start of desired data range + * @len: Length of desired data range + * @func: For tracing: name of calling function + * @line: For tracing: caller line of function call + * + * Return: 0 if the range is valid, -1 otherwise + */ +static int packet_check_range(const struct pool *p, const char *ptr, size_t len, + const char *func, int line) +{ + if (len > PACKET_MAX_LEN) { + debug("packet range length %zu (max %zu), %s:%i", + len, PACKET_MAX_LEN, func, line); + return -1; + } + + if (p->buf_size == 0) { + int ret; + + ret = vu_packet_check_range((void *)p->buf, ptr, len); + + if (ret == -1) + debug("cannot find region, %s:%i", func, line); + + return ret; + } + + if (ptr < p->buf) { + debug("packet range start %p before buffer start %p, %s:%i", + (void *)ptr, (void *)p->buf, func, line); + return -1; + } + + if (len > p->buf_size) { + debug("packet range length %zu larger than buffer %zu, %s:%i", + len, p->buf_size, func, line); + return -1; + } + + if ((size_t)(ptr - p->buf) > p->buf_size - len) { + debug("packet range %p, len %zu after buffer end %p, %s:%i", + (void *)ptr, len, (void *)(p->buf + p->buf_size), + func, line); + return -1; + } + + return 0; +} +/** + * pool_full() - Is a packet pool full? + * @p: Pointer to packet pool + * + * Return: true if the pool is full, false if more packets can be added + */ +bool pool_full(const struct pool *p) +{ + return p->count >= p->size; +} + /** * packet_add_do() - Add data as packet descriptor to given pool * @p: Existing pool * @len: Length of new descriptor * @start: Start of data - * @func: For tracing: name of calling function, NULL means no trace() + * @func: For tracing: name of calling function * @line: For tracing: caller line of function call */ void packet_add_do(struct pool *p, size_t len, const char *start, @@ -35,44 +97,63 @@ void packet_add_do(struct pool *p, size_t len, const char *start, { size_t idx = p->count; - if (idx >= p->size) { - trace("add packet index %zu to pool with size %zu, %s:%i", + if (pool_full(p)) { + debug("add packet index %zu to pool with size %zu, %s:%i", idx, p->size, func, line); return; } - if (start < p->buf) { - trace("add packet start %p before buffer start %p, %s:%i", - (void *)start, (void *)p->buf, func, line); + if (packet_check_range(p, start, len, func, line)) return; - } - if (start + len > p->buf + p->buf_size) { - trace("add packet start %p, length: %zu, buffer end %p, %s:%i", - (void *)start, len, (void *)(p->buf + p->buf_size), - func, line); - return; - } - - if (len > UINT16_MAX) { - trace("add packet length %zu, %s:%i", len, func, line); - return; - } - -#if UINTPTR_MAX == UINT64_MAX - if ((uintptr_t)start - (uintptr_t)p->buf > UINT32_MAX) { - trace("add packet start %p, buffer start %p, %s:%i", - (void *)start, (void *)p->buf, func, line); - return; - } -#endif - - p->pkt[idx].offset = start - p->buf; - p->pkt[idx].len = len; + p->pkt[idx].iov_base = (void *)start; + p->pkt[idx].iov_len = len; p->count++; } +/** + * packet_get_try_do() - Get data range from packet descriptor from given pool + * @p: Packet pool + * @idx: Index of packet descriptor in pool + * @offset: Offset of data range in packet descriptor + * @len: Length of desired data range + * @left: Length of available data after range, set on return, can be NULL + * @func: For tracing: name of calling function + * @line: For tracing: caller line of function call + * + * Return: pointer to start of data range, NULL on invalid range or descriptor + */ +void *packet_get_try_do(const struct pool *p, size_t idx, size_t offset, + size_t len, size_t *left, const char *func, int line) +{ + char *ptr; + + ASSERT_WITH_MSG(p->count <= p->size, + "Corrupt pool count: %zu, size: %zu, %s:%i", + p->count, p->size, func, line); + + if (idx >= p->count) { + debug("packet %zu from pool count: %zu, %s:%i", + idx, p->count, func, line); + return NULL; + } + + if (offset > p->pkt[idx].iov_len || + len > (p->pkt[idx].iov_len - offset)) + return NULL; + + ptr = (char *)p->pkt[idx].iov_base + offset; + + ASSERT_WITH_MSG(!packet_check_range(p, ptr, len, func, line), + "Corrupt packet pool, %s:%i", func, line); + + if (left) + *left = p->pkt[idx].iov_len - offset - len; + + return ptr; +} + /** * packet_get_do() - Get data range from packet descriptor from given pool * @p: Packet pool @@ -80,52 +161,24 @@ void packet_add_do(struct pool *p, size_t len, const char *start, * @offset: Offset of data range in packet descriptor * @len: Length of desired data range * @left: Length of available data after range, set on return, can be NULL - * @func: For tracing: name of calling function, NULL means no trace() + * @func: For tracing: name of calling function * @line: For tracing: caller line of function call * - * Return: pointer to start of data range, NULL on invalid range or descriptor + * Return: as packet_get_try_do() but log a trace message when returning NULL */ -void *packet_get_do(const struct pool *p, size_t idx, size_t offset, - size_t len, size_t *left, const char *func, int line) +void *packet_get_do(const struct pool *p, const size_t idx, + size_t offset, size_t len, size_t *left, + const char *func, int line) { - if (idx >= p->size || idx >= p->count) { - if (func) { - trace("packet %zu from pool size: %zu, count: %zu, " - "%s:%i", idx, p->size, p->count, func, line); - } - return NULL; + void *r = packet_get_try_do(p, idx, offset, len, left, func, line); + + if (!r) { + trace("missing packet data length %zu, offset %zu from " + "length %zu, %s:%i", + len, offset, p->pkt[idx].iov_len, func, line); } - if (len > UINT16_MAX || len + offset > UINT32_MAX) { - if (func) { - trace("packet data length %zu, offset %zu, %s:%i", - len, offset, func, line); - } - return NULL; - } - - if (p->pkt[idx].offset + len + offset > p->buf_size) { - if (func) { - trace("packet offset plus length %zu from size %zu, " - "%s:%i", p->pkt[idx].offset + len + offset, - p->buf_size, func, line); - } - return NULL; - } - - if (len + offset > p->pkt[idx].len) { - if (func) { - trace("data length %zu, offset %zu from length %u, " - "%s:%i", len, offset, p->pkt[idx].len, - func, line); - } - return NULL; - } - - if (left) - *left = p->pkt[idx].len - offset - len; - - return p->buf + p->pkt[idx].offset + offset; + return r; } /** diff --git a/packet.h b/packet.h index a784b07..c94780a 100644 --- a/packet.h +++ b/packet.h @@ -6,20 +6,17 @@ #ifndef PACKET_H #define PACKET_H -/** - * struct desc - Generic offset-based descriptor within buffer - * @offset: Offset of descriptor relative to buffer start, 32-bit limit - * @len: Length of descriptor, host order, 16-bit limit - */ -struct desc { - uint32_t offset; - uint16_t len; -}; +#include <stdbool.h> + +/* Maximum size of a single packet stored in pool, including headers */ +#define PACKET_MAX_LEN ((size_t)UINT16_MAX) /** * struct pool - Generic pool of packets stored in a buffer - * @buf: Buffer storing packet descriptors - * @buf_size: Total size of buffer + * @buf: Buffer storing packet descriptors, + * a struct vu_dev_region array for passt vhost-user mode + * @buf_size: Total size of buffer, + * 0 for passt vhost-user mode * @size: Number of usable descriptors for the pool * @count: Number of used descriptors for the pool * @pkt: Descriptors: see macros below @@ -29,32 +26,36 @@ struct pool { size_t buf_size; size_t size; size_t count; - struct desc pkt[1]; + struct iovec pkt[]; }; +int vu_packet_check_range(void *buf, const char *ptr, size_t len); void packet_add_do(struct pool *p, size_t len, const char *start, const char *func, int line); +void *packet_get_try_do(const struct pool *p, const size_t idx, + size_t offset, size_t len, size_t *left, + const char *func, int line); void *packet_get_do(const struct pool *p, const size_t idx, size_t offset, size_t len, size_t *left, const char *func, int line); +bool pool_full(const struct pool *p); void pool_flush(struct pool *p); #define packet_add(p, len, start) \ packet_add_do(p, len, start, __func__, __LINE__) +#define packet_get_try(p, idx, offset, len, left) \ + packet_get_try_do(p, idx, offset, len, left, __func__, __LINE__) #define packet_get(p, idx, offset, len, left) \ packet_get_do(p, idx, offset, len, left, __func__, __LINE__) -#define packet_get_try(p, idx, offset, len, left) \ - packet_get_do(p, idx, offset, len, left, NULL, 0) - #define PACKET_POOL_DECL(_name, _size, _buf) \ struct _name ## _t { \ char *buf; \ size_t buf_size; \ size_t size; \ size_t count; \ - struct desc pkt[_size]; \ + struct iovec pkt[_size]; \ } #define PACKET_POOL_INIT_NOCAST(_size, _buf, _buf_size) \ diff --git a/passt-repair.1 b/passt-repair.1 new file mode 100644 index 0000000..e65aadd --- /dev/null +++ b/passt-repair.1 @@ -0,0 +1,74 @@ +.\" SPDX-License-Identifier: GPL-2.0-or-later +.\" Copyright (c) 2025 Red Hat GmbH +.\" Author: Stefano Brivio <sbrivio@redhat.com> +.TH passt-repair 1 + +.SH NAME +.B passt-repair +\- Helper setting TCP_REPAIR socket options for \fBpasst\fR(1) + +.SH SYNOPSIS +.B passt-repair +\fIPATH\fR + +.SH DESCRIPTION + +.B passt-repair +is a privileged helper setting and clearing repair mode on TCP sockets on behalf +of \fBpasst\fR(1), as instructed via single-byte commands over a UNIX domain +socket. + +It can be used to migrate TCP connections between guests without granting +additional capabilities to \fBpasst\fR(1) itself: to migrate TCP connections, +\fBpasst\fR(1) leverages repair mode, which needs the \fBCAP_NET_ADMIN\fR +capability (see \fBcapabilities\fR(7)) to be set or cleared. + +If \fIPATH\fR represents a UNIX domain socket, \fBpasst-repair\fR(1) attempts to +connect to it. If it is a directory, \fBpasst-repair\fR(1) waits until a file +ending with \fI.repair\fR appears in it, and then attempts to connect to it. + +.SH PROTOCOL + +\fBpasst-repair\fR(1) connects to \fBpasst\fR(1) using the socket specified via +\fI--repair-path\fR option in \fBpasst\fR(1) itself. By default, the name is the +same as the UNIX domain socket used for guest communication, suffixed by +\fI.repair\fR. + +The messages consist of one 8-bit signed integer that can be \fITCP_REPAIR_ON\fR +(1), \fITCP_REPAIR_OFF\fR (0), or \fITCP_REPAIR_OFF_NO_WP\fR (-1), as defined by +the Linux kernel user API, and one to SCM_MAX_FD (253) sockets as SCM_RIGHTS +(see \fBunix\fR(7)) ancillary message, sent by the server, \fBpasst\fR(1). + +The client, \fBpasst-repair\fR(1), replies with the same byte (and no ancillary +message) to indicate success, and closes the connection on failure. + +The server closes the connection on error or completion. + +.SH NOTES + +\fBpasst-repair\fR(1) can be granted the \fBCAP_NET_ADMIN\fR capability +(preferred, as it limits privileges to the strictly necessary ones), or it can +be run as root. + +.SH AUTHOR + +Stefano Brivio <sbrivio@redhat.com>. + +.SH REPORTING BUGS + +Please report issues on the bug tracker at https://bugs.passt.top/, or +send a message to the passt-user@passt.top mailing list, see +https://lists.passt.top/. + +.SH COPYRIGHT + +Copyright (c) 2025 Red Hat GmbH. + +\fBpasst-repair\fR is free software: you can redistribute them and/or modify +them under the terms of the GNU General Public License as published by the Free +Software Foundation, either version 2 of the License, or (at your option) any +later version. + +.SH SEE ALSO + +\fBpasst\fR(1), \fBqemu\fR(1), \fBcapabilities\fR(7), \fBunix\fR(7). diff --git a/passt-repair.c b/passt-repair.c new file mode 100644 index 0000000..ff1c44f --- /dev/null +++ b/passt-repair.c @@ -0,0 +1,266 @@ +// SPDX-License-Identifier: GPL-2.0-or-later + +/* PASST - Plug A Simple Socket Transport + * for qemu/UNIX domain socket mode + * + * PASTA - Pack A Subtle Tap Abstraction + * for network namespace/tap device mode + * + * passt-repair.c - Privileged helper to set/clear TCP_REPAIR on sockets + * + * Copyright (c) 2025 Red Hat GmbH + * Author: Stefano Brivio <sbrivio@redhat.com> + * + * Connect to passt via UNIX domain socket, receive sockets via SCM_RIGHTS along + * with byte commands mapping to TCP_REPAIR values, and switch repair mode on or + * off. Reply by echoing the command. Exit on EOF. + */ + +#include <sys/inotify.h> +#include <sys/prctl.h> +#include <sys/types.h> +#include <sys/socket.h> +#include <sys/stat.h> +#include <sys/un.h> +#include <errno.h> +#include <stdbool.h> +#include <stddef.h> +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <limits.h> +#include <unistd.h> +#include <netdb.h> + +#include <netinet/tcp.h> + +#include <linux/audit.h> +#include <linux/capability.h> +#include <linux/filter.h> +#include <linux/seccomp.h> + +#include "seccomp_repair.h" + +#define SCM_MAX_FD 253 /* From Linux kernel (include/net/scm.h), not in UAPI */ +#define REPAIR_EXT ".repair" +#define REPAIR_EXT_LEN strlen(REPAIR_EXT) + +/** + * main() - Entry point and whole program with loop + * @argc: Argument count, must be 2 + * @argv: Argument: path of UNIX domain socket to connect to + * + * Return: 0 on success (EOF), 1 on error, 2 on usage error + * + * #syscalls:repair connect setsockopt write close exit_group + * #syscalls:repair socket s390x:socketcall i686:socketcall + * #syscalls:repair recvfrom recvmsg arm:recv ppc64le:recv + * #syscalls:repair sendto sendmsg arm:send ppc64le:send + * #syscalls:repair stat|statx stat64|statx statx + * #syscalls:repair fstat|fstat64 newfstatat|fstatat64 + * #syscalls:repair inotify_init1 inotify_add_watch + */ +int main(int argc, char **argv) +{ + char buf[CMSG_SPACE(sizeof(int) * SCM_MAX_FD)] + __attribute__ ((aligned(__alignof__(struct cmsghdr)))); + struct sockaddr_un a = { AF_UNIX, "" }; + int fds[SCM_MAX_FD], s, ret, i, n = 0; + bool inotify_dir = false; + struct sock_fprog prog; + int8_t cmd = INT8_MAX; + struct cmsghdr *cmsg; + struct msghdr msg; + struct iovec iov; + size_t cmsg_len; + struct stat sb; + int op; + + prctl(PR_SET_DUMPABLE, 0); + + prog.len = (unsigned short)sizeof(filter_repair) / + sizeof(filter_repair[0]); + prog.filter = filter_repair; + if (prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0) || + prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, &prog)) { + fprintf(stderr, "Failed to apply seccomp filter"); + _exit(1); + } + + iov = (struct iovec){ &cmd, sizeof(cmd) }; + msg = (struct msghdr){ .msg_name = NULL, .msg_namelen = 0, + .msg_iov = &iov, .msg_iovlen = 1, + .msg_control = buf, + .msg_controllen = sizeof(buf), + .msg_flags = 0 }; + cmsg = CMSG_FIRSTHDR(&msg); + + if (argc != 2) { + fprintf(stderr, "Usage: %s PATH\n", argv[0]); + _exit(2); + } + + if ((s = socket(AF_UNIX, SOCK_STREAM, 0)) < 0) { + fprintf(stderr, "Failed to create AF_UNIX socket: %i\n", errno); + _exit(1); + } + + if ((stat(argv[1], &sb))) { + fprintf(stderr, "Can't stat() %s: %i\n", argv[1], errno); + _exit(1); + } + + if ((sb.st_mode & S_IFMT) == S_IFDIR) { + char buf[sizeof(struct inotify_event) + NAME_MAX + 1] + __attribute__ ((aligned(__alignof__(struct inotify_event)))); + const struct inotify_event *ev = NULL; + char path[PATH_MAX + 1]; + bool found = false; + ssize_t n; + int fd; + + if ((fd = inotify_init1(IN_CLOEXEC)) < 0) { + fprintf(stderr, "inotify_init1: %i\n", errno); + _exit(1); + } + + if (inotify_add_watch(fd, argv[1], IN_CREATE) < 0) { + fprintf(stderr, "inotify_add_watch: %i\n", errno); + _exit(1); + } + + do { + char *p; + + n = read(fd, buf, sizeof(buf)); + if (n < 0) { + fprintf(stderr, "inotify read: %i", errno); + _exit(1); + } + buf[n - 1] = '\0'; + + if (n < (ssize_t)sizeof(*ev)) { + fprintf(stderr, "Short inotify read: %zi", n); + continue; + } + + for (p = buf; p < buf + n; p += sizeof(*ev) + ev->len) { + ev = (const struct inotify_event *)p; + + if (ev->len >= REPAIR_EXT_LEN && + !memcmp(ev->name + + strnlen(ev->name, ev->len) - + REPAIR_EXT_LEN, + REPAIR_EXT, REPAIR_EXT_LEN)) { + found = true; + break; + } + } + } while (!found); + + if (ev->len > NAME_MAX + 1 || ev->name[ev->len - 1] != '\0') { + fprintf(stderr, "Invalid filename from inotify\n"); + _exit(1); + } + + snprintf(path, sizeof(path), "%s/%s", argv[1], ev->name); + if ((stat(path, &sb))) { + fprintf(stderr, "Can't stat() %s: %i\n", path, errno); + _exit(1); + } + + ret = snprintf(a.sun_path, sizeof(a.sun_path), "%s", path); + inotify_dir = true; + } else { + ret = snprintf(a.sun_path, sizeof(a.sun_path), "%s", argv[1]); + } + + if (ret <= 0 || ret >= (int)sizeof(a.sun_path)) { + fprintf(stderr, "Invalid socket path"); + _exit(2); + } + + if ((sb.st_mode & S_IFMT) != S_IFSOCK) { + fprintf(stderr, "%s is not a socket\n", a.sun_path); + _exit(2); + } + + while (connect(s, (struct sockaddr *)&a, sizeof(a))) { + if (inotify_dir && errno == ECONNREFUSED) + continue; + + fprintf(stderr, "Failed to connect to %s: %s\n", a.sun_path, + strerror(errno)); + _exit(1); + } + +loop: + ret = recvmsg(s, &msg, 0); + if (ret < 0) { + if (errno == ECONNRESET) { + ret = 0; + } else { + fprintf(stderr, "Failed to read message: %i\n", errno); + _exit(1); + } + } + + if (!ret) /* Done */ + _exit(0); + + if (!cmsg || + cmsg->cmsg_len < CMSG_LEN(sizeof(int)) || + cmsg->cmsg_len > CMSG_LEN(sizeof(int) * SCM_MAX_FD) || + cmsg->cmsg_type != SCM_RIGHTS) { + fprintf(stderr, "No/bad ancillary data from peer\n"); + _exit(1); + } + + /* No inverse formula for CMSG_LEN(x), and building one with CMSG_LEN(0) + * works but there's no guarantee it does. Search the whole domain. + */ + for (i = 1; i <= SCM_MAX_FD; i++) { + if (CMSG_LEN(sizeof(int) * i) == cmsg->cmsg_len) { + n = i; + break; + } + } + if (!n) { + cmsg_len = cmsg->cmsg_len; /* socklen_t is 'unsigned' on musl */ + fprintf(stderr, "Invalid ancillary data length %zu from peer\n", + cmsg_len); + _exit(1); + } + + memcpy(fds, CMSG_DATA(cmsg), sizeof(int) * n); + + if (cmd != TCP_REPAIR_ON && cmd != TCP_REPAIR_OFF && + cmd != TCP_REPAIR_OFF_NO_WP) { + fprintf(stderr, "Unsupported command 0x%04x\n", cmd); + _exit(1); + } + + op = cmd; + + for (i = 0; i < n; i++) { + if (setsockopt(fds[i], SOL_TCP, TCP_REPAIR, &op, sizeof(op))) { + fprintf(stderr, + "Setting TCP_REPAIR to %i on socket %i: %s", op, + fds[i], strerror(errno)); + _exit(1); + } + + /* Close _our_ copy */ + close(fds[i]); + } + + /* Confirm setting by echoing the command back */ + if (send(s, &cmd, sizeof(cmd), 0) < 0) { + fprintf(stderr, "Reply to %i: %s\n", op, strerror(errno)); + _exit(1); + } + + goto loop; + + return 0; +} diff --git a/passt.1 b/passt.1 index 15c8338..60066c2 100644 --- a/passt.1 +++ b/passt.1 @@ -401,15 +401,44 @@ Enable IPv6-only operation. IPv4 traffic will be ignored. By default, IPv4 operation is enabled as long as at least an IPv4 route and an interface address are configured on a given host interface. +.TP +.BR \-H ", " \-\-hostname " " \fIname +Hostname to configure the client with. +Send \fIname\fR as DHCP option 12 (hostname). + +.TP +.BR \-\-fqdn " " \fIname +FQDN to configure the client with. +Send \fIname\fR as Client FQDN: DHCP option 81 and DHCPv6 option 39. + .SS \fBpasst\fR-only options .TP -.BR \-s ", " \-\-socket " " \fIpath +.BR \-s ", " \-\-socket-path ", " \-\-socket " " \fIpath Path for UNIX domain socket used by \fBqemu\fR(1) or \fBqrap\fR(1) to connect to \fBpasst\fR. Default is to probe a free socket, not accepting connections, starting from \fI/tmp/passt_1.socket\fR to \fI/tmp/passt_64.socket\fR. +.TP +.BR \-\-vhost-user +Enable vhost-user. The vhost-user command socket is provided by \fB--socket\fR. + +.TP +.BR \-\-print-capabilities +Print back-end capabilities in JSON format, only meaningful for vhost-user mode. + +.TP +.BR \-\-repair-path " " \fIpath +Path for UNIX domain socket used by the \fBpasst-repair\fR(1) helper to connect +to \fBpasst\fR in order to set or clear the TCP_REPAIR option on sockets, during +migration. \fB--repair-path none\fR disables this interface (if you need to +specify a socket path called "none" you can prefix the path by \fI./\fR). + +Default, for \-\-vhost-user mode only, is to append \fI.repair\fR to the path +chosen for the hypervisor UNIX domain socket. No socket is created if not in +\-\-vhost-user mode. + .TP .BR \-F ", " \-\-fd " " \fIFD Pass a pre-opened, connected socket to \fBpasst\fR. Usually the socket is opened @@ -614,7 +643,7 @@ Configure UDP port forwarding from target namespace to init namespace. Default is \fBauto\fR. .TP -.BR \-\-host-lo-to-ns-lo " " (DEPRECATED) +.BR \-\-host-lo-to-ns-lo If specified, connections forwarded with \fB\-t\fR and \fB\-u\fR from the host's loopback address will appear on the loopback address in the guest as well. Without this option such forwarded packets will appear @@ -687,6 +716,11 @@ Configure MAC address \fIaddr\fR on the tap interface in the namespace. Default is to let the tap driver build a pseudorandom hardware address. +.TP +.BR \-\-no-splice +Disable the bypass path for inbound, local traffic. See the section \fBHandling +of local traffic in pasta\fR in the \fBNOTES\fR for more details. + .SH EXAMPLES .SS \fBpasta @@ -928,10 +962,16 @@ with destination 127.0.0.10, and the default IPv4 gateway is 192.0.2.1, while the last observed source address from guest or namespace is 192.0.2.2, this will be translated to a connection from 192.0.2.1 to 192.0.2.2. -Similarly, for traffic coming from guest or namespace, packets with -destination address corresponding to the \fB\-\-map-host-loopback\fR -address will have their destination address translated to a loopback -address. +Similarly, for traffic coming from guest or namespace, packets with destination +address corresponding to the \fB\-\-map-host-loopback\fR address will have their +destination address translated to a loopback address. + +As an exception, traffic identified as DNS, originally directed to the +\fB\-\-map-host-loopback\fR address, if this address matches a resolver address +on the host, is \fBnot\fR translated to loopback, but rather handled in the same +way as if specified as \-\-dns-forward address, if no such option was given. +In the common case where the host gateway also acts a resolver, this avoids that +the host mapping shadows the gateway/resolver itself. .SS Handling of local traffic in pasta diff --git a/passt.c b/passt.c index 06e0a33..388d10f 100644 --- a/passt.c +++ b/passt.c @@ -50,6 +50,9 @@ #include "log.h" #include "tcp_splice.h" #include "ndp.h" +#include "vu_common.h" +#include "migrate.h" +#include "repair.h" #define EPOLL_EVENTS 8 @@ -65,13 +68,17 @@ char *epoll_type_str[] = { [EPOLL_TYPE_TCP_LISTEN] = "listening TCP socket", [EPOLL_TYPE_TCP_TIMER] = "TCP timer", [EPOLL_TYPE_UDP_LISTEN] = "listening UDP socket", - [EPOLL_TYPE_UDP_REPLY] = "UDP reply socket", + [EPOLL_TYPE_UDP] = "UDP flow socket", [EPOLL_TYPE_PING] = "ICMP/ICMPv6 ping socket", [EPOLL_TYPE_NSQUIT_INOTIFY] = "namespace inotify watch", [EPOLL_TYPE_NSQUIT_TIMER] = "namespace timer watch", [EPOLL_TYPE_TAP_PASTA] = "/dev/net/tun device", [EPOLL_TYPE_TAP_PASST] = "connected qemu socket", [EPOLL_TYPE_TAP_LISTEN] = "listening qemu socket", + [EPOLL_TYPE_VHOST_CMD] = "vhost-user command socket", + [EPOLL_TYPE_VHOST_KICK] = "vhost-user kick socket", + [EPOLL_TYPE_REPAIR_LISTEN] = "TCP_REPAIR helper listening socket", + [EPOLL_TYPE_REPAIR] = "TCP_REPAIR helper socket", }; static_assert(ARRAY_SIZE(epoll_type_str) == EPOLL_NUM_TYPES, "epoll_type_str[] doesn't match enum epoll_type"); @@ -159,11 +166,11 @@ void proto_update_l2_buf(const unsigned char *eth_d, const unsigned char *eth_s) * * #syscalls exit_group */ -void exit_handler(int signal) +static void exit_handler(int signal) { (void)signal; - exit(EXIT_SUCCESS); + _exit(EXIT_SUCCESS); } /** @@ -177,14 +184,13 @@ void exit_handler(int signal) * #syscalls socket getsockopt setsockopt s390x:socketcall i686:socketcall close * #syscalls bind connect recvfrom sendto shutdown * #syscalls arm:recv ppc64le:recv arm:send ppc64le:send - * #syscalls accept4|accept listen epoll_ctl epoll_wait|epoll_pwait epoll_pwait + * #syscalls accept4 accept listen epoll_ctl epoll_wait|epoll_pwait epoll_pwait * #syscalls clock_gettime arm:clock_gettime64 i686:clock_gettime64 */ int main(int argc, char **argv) { struct epoll_event events[EPOLL_EVENTS]; int nfds, i, devnull_fd = -1; - char argv0[PATH_MAX], *name; struct ctx c = { 0 }; struct rlimit limit; struct timespec now; @@ -198,6 +204,7 @@ int main(int argc, char **argv) isolate_initial(argc, argv); c.pasta_netns_fd = c.fd_tap = c.pidfile_fd = -1; + c.device_state_fd = -1; sigemptyset(&sa.sa_mask); sa.sa_flags = 0; @@ -205,27 +212,18 @@ int main(int argc, char **argv) sigaction(SIGTERM, &sa, NULL); sigaction(SIGQUIT, &sa, NULL); - if (argc < 1) - exit(EXIT_FAILURE); + c.mode = conf_mode(argc, argv); - strncpy(argv0, argv[0], PATH_MAX - 1); - name = basename(argv0); - if (strstr(name, "pasta")) { + if (c.mode == MODE_PASTA) { sa.sa_handler = pasta_child_handler; if (sigaction(SIGCHLD, &sa, NULL)) die_perror("Couldn't install signal handlers"); - - if (signal(SIGPIPE, SIG_IGN) == SIG_ERR) - die_perror("Couldn't set disposition for SIGPIPE"); - - c.mode = MODE_PASTA; - } else if (strstr(name, "passt")) { - c.mode = MODE_PASST; - } else { - exit(EXIT_FAILURE); } - madvise(pkt_buf, TAP_BUF_BYTES, MADV_HUGEPAGE); + if (signal(SIGPIPE, SIG_IGN) == SIG_ERR) + die_perror("Couldn't set disposition for SIGPIPE"); + + madvise(pkt_buf, sizeof(pkt_buf), MADV_HUGEPAGE); c.epollfd = epoll_create1(EPOLL_CLOEXEC); if (c.epollfd == -1) @@ -245,7 +243,7 @@ int main(int argc, char **argv) pasta_netns_quit_init(&c); - tap_sock_init(&c); + tap_backend_init(&c); random_init(&c); @@ -255,7 +253,7 @@ int main(int argc, char **argv) flow_init(); if ((!c.no_udp && udp_init(&c)) || (!c.no_tcp && tcp_init(&c))) - exit(EXIT_FAILURE); + _exit(EXIT_FAILURE); proto_update_l2_buf(c.guest_mac, c.our_tap_mac); @@ -341,12 +339,24 @@ loop: case EPOLL_TYPE_UDP_LISTEN: udp_listen_sock_handler(&c, ref, eventmask, &now); break; - case EPOLL_TYPE_UDP_REPLY: - udp_reply_sock_handler(&c, ref, eventmask, &now); + case EPOLL_TYPE_UDP: + udp_sock_handler(&c, ref, eventmask, &now); break; case EPOLL_TYPE_PING: icmp_sock_handler(&c, ref); break; + case EPOLL_TYPE_VHOST_CMD: + vu_control_handler(c.vdev, c.fd_tap, eventmask); + break; + case EPOLL_TYPE_VHOST_KICK: + vu_kick_cb(c.vdev, ref, &now); + break; + case EPOLL_TYPE_REPAIR_LISTEN: + repair_listen_handler(&c, eventmask); + break; + case EPOLL_TYPE_REPAIR: + repair_handler(&c, eventmask); + break; default: /* Can't happen */ ASSERT(0); @@ -355,5 +365,7 @@ loop: post_handler(&c, &now); + migrate_handler(&c); + goto loop; } diff --git a/passt.h b/passt.h index 799ee50..8693794 100644 --- a/passt.h +++ b/passt.h @@ -20,11 +20,13 @@ union epoll_ref; #include "siphash.h" #include "ip.h" #include "inany.h" +#include "migrate.h" #include "flow.h" #include "icmp.h" #include "fwd.h" #include "tcp.h" #include "udp.h" +#include "vhost_user.h" /* Default address for our end on the tap interface. Bit 0 of byte 0 must be 0 * (unicast) and bit 1 of byte 1 must be 1 (locally administered). Otherwise @@ -43,6 +45,7 @@ union epoll_ref; * @icmp: ICMP-specific reference part * @data: Data handled by protocol handlers * @nsdir_fd: netns dirfd for fallback timer checking if namespace is gone + * @queue: vhost-user queue index for this fd * @u64: Opaque reference for epoll_ctl() and epoll_wait() */ union epoll_ref { @@ -58,6 +61,7 @@ union epoll_ref { union udp_listen_epoll_ref udp; uint32_t data; int nsdir_fd; + int queue; }; }; uint64_t u64; @@ -65,12 +69,9 @@ union epoll_ref { static_assert(sizeof(union epoll_ref) <= sizeof(union epoll_data), "epoll_ref must have same size as epoll_data"); -#define TAP_BUF_BYTES \ - ROUND_DOWN(((ETH_MAX_MTU + sizeof(uint32_t)) * 128), PAGE_SIZE) -#define TAP_MSGS \ - DIV_ROUND_UP(TAP_BUF_BYTES, ETH_ZLEN - 2 * ETH_ALEN + sizeof(uint32_t)) +/* Large enough for ~128 maximum size frames */ +#define PKT_BUF_BYTES (8UL << 20) -#define PKT_BUF_BYTES MAX(TAP_BUF_BYTES, 0) extern char pkt_buf [PKT_BUF_BYTES]; extern char *epoll_type_str[]; @@ -94,6 +95,7 @@ struct fqdn { enum passt_modes { MODE_PASST, MODE_PASTA, + MODE_VU, }; /** @@ -189,6 +191,7 @@ struct ip6_ctx { * @foreground: Run in foreground, don't log to stderr by default * @nofile: Maximum number of open files (ulimit -n) * @sock_path: Path for UNIX domain socket + * @repair_path: TCP_REPAIR helper path, can be "none", empty for default * @pcap: Path for packet capture file * @pidfile: Path to PID file, empty string if not configured * @pidfile_fd: File descriptor for PID file, -1 if none @@ -199,12 +202,16 @@ struct ip6_ctx { * @epollfd: File descriptor for epoll instance * @fd_tap_listen: File descriptor for listening AF_UNIX socket, if any * @fd_tap: AF_UNIX socket, tuntap device, or pre-opened socket + * @fd_repair_listen: File descriptor for listening TCP_REPAIR socket, if any + * @fd_repair: Connected AF_UNIX socket for TCP_REPAIR helper * @our_tap_mac: Pasta/passt's MAC on the tap link * @guest_mac: MAC address of guest or namespace, seen or configured * @hash_secret: 128-bit secret for siphash functions * @ifi4: Template interface for IPv4, -1: none, 0: IPv4 disabled * @ip: IPv4 configuration * @dns_search: DNS search list + * @hostname: Guest hostname + * @fqdn: Guest FQDN * @ifi6: Template interface for IPv6, -1: none, 0: IPv6 disabled * @ip6: IPv6 configuration * @pasta_ifn: Name of namespace interface for pasta @@ -225,10 +232,15 @@ struct ip6_ctx { * @no_dhcpv6: Disable DHCPv6 server * @no_ndp: Disable NDP handler altogether * @no_ra: Disable router advertisements + * @no_splice: Disable socket splicing for inbound traffic * @host_lo_to_ns_lo: Map host loopback addresses to ns loopback addresses * @freebind: Allow binding of non-local addresses for forwarding * @low_wmem: Low probed net.core.wmem_max * @low_rmem: Low probed net.core.rmem_max + * @vdev: vhost-user device + * @device_state_fd: Device state migration channel + * @device_state_result: Device state migration result + * @migrate_target: Are we the target, on the next migration request? */ struct ctx { enum passt_modes mode; @@ -238,6 +250,7 @@ struct ctx { int foreground; int nofile; char sock_path[UNIX_PATH_MAX]; + char repair_path[UNIX_PATH_MAX]; char pcap[PATH_MAX]; char pidfile[PATH_MAX]; @@ -254,8 +267,12 @@ struct ctx { int epollfd; int fd_tap_listen; int fd_tap; + int fd_repair_listen; + int fd_repair; unsigned char our_tap_mac[ETH_ALEN]; unsigned char guest_mac[ETH_ALEN]; + uint16_t mtu; + uint64_t hash_secret[2]; int ifi4; @@ -263,6 +280,9 @@ struct ctx { struct fqdn dns_search[MAXDNSRCH]; + char hostname[PASST_MAXDNAME]; + char fqdn[PASST_MAXDNAME]; + int ifi6; struct ip6_ctx ip6; @@ -277,7 +297,6 @@ struct ctx { int no_icmp; struct icmp_ctx icmp; - int mtu; int no_dns; int no_dns_search; int no_dhcp_dns; @@ -286,11 +305,19 @@ struct ctx { int no_dhcpv6; int no_ndp; int no_ra; + int no_splice; int host_lo_to_ns_lo; int freebind; int low_wmem; int low_rmem; + + struct vu_dev *vdev; + + /* Migration */ + int device_state_fd; + int device_state_result; + bool migrate_target; }; void proto_update_l2_buf(const unsigned char *eth_d, diff --git a/pasta.c b/pasta.c index 96dacc3..017fa32 100644 --- a/pasta.c +++ b/pasta.c @@ -73,12 +73,12 @@ void pasta_child_handler(int signal) !waitid(P_PID, pasta_child_pid, &infop, WEXITED | WNOHANG)) { if (infop.si_pid == pasta_child_pid) { if (infop.si_code == CLD_EXITED) - exit(infop.si_status); + _exit(infop.si_status); /* If killed by a signal, si_status is the number. * Follow common shell convention of returning it + 128. */ - exit(infop.si_status + 128); + _exit(infop.si_status + 128); /* Nothing to do, detached PID namespace going away */ } @@ -169,10 +169,12 @@ void pasta_open_ns(struct ctx *c, const char *netns) * struct pasta_spawn_cmd_arg - Argument for pasta_spawn_cmd() * @exe: Executable to run * @argv: Command and arguments to run + * @ctx: Context to read config from */ struct pasta_spawn_cmd_arg { const char *exe; char *const *argv; + struct ctx *c; }; /** @@ -186,6 +188,7 @@ static int pasta_spawn_cmd(void *arg) { char hostname[HOST_NAME_MAX + 1] = HOSTNAME_PREFIX; const struct pasta_spawn_cmd_arg *a; + size_t conf_hostname_len; sigset_t set; /* We run in a detached PID and mount namespace: mount /proc over */ @@ -195,9 +198,15 @@ static int pasta_spawn_cmd(void *arg) if (write_file("/proc/sys/net/ipv4/ping_group_range", "0 0")) warn("Cannot set ping_group_range, ICMP requests might fail"); - if (!gethostname(hostname + sizeof(HOSTNAME_PREFIX) - 1, - HOST_NAME_MAX + 1 - sizeof(HOSTNAME_PREFIX)) || - errno == ENAMETOOLONG) { + a = (const struct pasta_spawn_cmd_arg *)arg; + + conf_hostname_len = strlen(a->c->hostname); + if (conf_hostname_len > 0) { + if (sethostname(a->c->hostname, conf_hostname_len)) + warn("Unable to set configured hostname"); + } else if (!gethostname(hostname + sizeof(HOSTNAME_PREFIX) - 1, + HOST_NAME_MAX + 1 - sizeof(HOSTNAME_PREFIX)) || + errno == ENAMETOOLONG) { hostname[HOST_NAME_MAX] = '\0'; if (sethostname(hostname, strlen(hostname))) warn("Unable to set pasta-prefixed hostname"); @@ -208,7 +217,6 @@ static int pasta_spawn_cmd(void *arg) sigaddset(&set, SIGUSR1); sigwaitinfo(&set, NULL); - a = (const struct pasta_spawn_cmd_arg *)arg; execvp(a->exe, a->argv); die_perror("Failed to start command or shell"); @@ -230,6 +238,7 @@ void pasta_start_ns(struct ctx *c, uid_t uid, gid_t gid, struct pasta_spawn_cmd_arg arg = { .exe = argv[0], .argv = argv, + .c = c, }; char uidmap[BUFSIZ], gidmap[BUFSIZ]; char *sh_argv[] = { NULL, NULL }; @@ -296,7 +305,7 @@ void pasta_ns_conf(struct ctx *c) rc = nl_link_set_flags(nl_sock_ns, 1 /* lo */, IFF_UP, IFF_UP); if (rc < 0) die("Couldn't bring up loopback interface in namespace: %s", - strerror(-rc)); + strerror_(-rc)); /* Get or set MAC in target namespace */ if (MAC_IS_ZERO(c->guest_mac)) @@ -305,12 +314,12 @@ void pasta_ns_conf(struct ctx *c) rc = nl_link_set_mac(nl_sock_ns, c->pasta_ifi, c->guest_mac); if (rc < 0) die("Couldn't set MAC address in namespace: %s", - strerror(-rc)); + strerror_(-rc)); if (c->pasta_conf_ns) { unsigned int flags = IFF_UP; - if (c->mtu != -1) + if (c->mtu) nl_link_set_mtu(nl_sock_ns, c->pasta_ifi, c->mtu); if (c->ifi6) /* Avoid duplicate address detection on link up */ @@ -332,7 +341,7 @@ void pasta_ns_conf(struct ctx *c) if (rc < 0) { die("Couldn't set IPv4 address(es) in namespace: %s", - strerror(-rc)); + strerror_(-rc)); } if (c->ip4.no_copy_routes) { @@ -346,7 +355,7 @@ void pasta_ns_conf(struct ctx *c) if (rc < 0) { die("Couldn't set IPv4 route(s) in guest: %s", - strerror(-rc)); + strerror_(-rc)); } } @@ -355,13 +364,13 @@ void pasta_ns_conf(struct ctx *c) &c->ip6.addr_ll_seen); if (rc < 0) { warn("Can't get LL address from namespace: %s", - strerror(-rc)); + strerror_(-rc)); } rc = nl_addr_set_ll_nodad(nl_sock_ns, c->pasta_ifi); if (rc < 0) { warn("Can't set nodad for LL in namespace: %s", - strerror(-rc)); + strerror_(-rc)); } /* We dodged DAD: re-enable neighbour solicitations */ @@ -382,7 +391,7 @@ void pasta_ns_conf(struct ctx *c) if (rc < 0) { die("Couldn't set IPv6 address(es) in namespace: %s", - strerror(-rc)); + strerror_(-rc)); } if (c->ip6.no_copy_routes) { @@ -397,7 +406,7 @@ void pasta_ns_conf(struct ctx *c) if (rc < 0) { die("Couldn't set IPv6 route(s) in guest: %s", - strerror(-rc)); + strerror_(-rc)); } } } @@ -446,18 +455,18 @@ void pasta_netns_quit_init(const struct ctx *c) return; if ((dir_fd = open(c->netns_dir, O_CLOEXEC | O_RDONLY)) < 0) - die("netns dir open: %s, exiting", strerror(errno)); + die("netns dir open: %s, exiting", strerror_(errno)); if (fstatfs(dir_fd, &s) || s.f_type == DEVPTS_SUPER_MAGIC || s.f_type == PROC_SUPER_MAGIC || s.f_type == SYSFS_MAGIC) try_inotify = false; if (try_inotify && (fd = inotify_init1(flags)) < 0) - warn("inotify_init1(): %s, use a timer", strerror(errno)); + warn("inotify_init1(): %s, use a timer", strerror_(errno)); if (fd >= 0 && inotify_add_watch(fd, c->netns_dir, IN_DELETE) < 0) { warn("inotify_add_watch(): %s, use a timer", - strerror(errno)); + strerror_(errno)); close(fd); fd = -1; } @@ -489,17 +498,23 @@ void pasta_netns_quit_init(const struct ctx *c) */ void pasta_netns_quit_inotify_handler(struct ctx *c, int inotify_fd) { - char buf[sizeof(struct inotify_event) + NAME_MAX + 1]; - const struct inotify_event *in_ev = (struct inotify_event *)buf; + char buf[sizeof(struct inotify_event) + NAME_MAX + 1] + __attribute__ ((aligned(__alignof__(struct inotify_event)))); + const struct inotify_event *ev; + ssize_t n; + char *p; - if (read(inotify_fd, buf, sizeof(buf)) < (ssize_t)sizeof(*in_ev)) + if ((n = read(inotify_fd, buf, sizeof(buf))) < (ssize_t)sizeof(*ev)) return; - if (strncmp(in_ev->name, c->netns_base, sizeof(c->netns_base))) - return; + for (p = buf; p < buf + n; p += sizeof(*ev) + ev->len) { + ev = (const struct inotify_event *)p; - info("Namespace %s is gone, exiting", c->netns_base); - exit(EXIT_SUCCESS); + if (!strncmp(ev->name, c->netns_base, sizeof(c->netns_base))) { + info("Namespace %s is gone, exiting", c->netns_base); + _exit(EXIT_SUCCESS); + } + } } /** @@ -525,7 +540,7 @@ void pasta_netns_quit_timer_handler(struct ctx *c, union epoll_ref ref) return; info("Namespace %s is gone, exiting", c->netns_base); - exit(EXIT_SUCCESS); + _exit(EXIT_SUCCESS); } close(fd); diff --git a/pcap.c b/pcap.c index 23205dd..e95aa6f 100644 --- a/pcap.c +++ b/pcap.c @@ -33,33 +33,12 @@ #include "log.h" #include "pcap.h" #include "iov.h" +#include "tap.h" #define PCAP_VERSION_MINOR 4 static int pcap_fd = -1; -/* See pcap.h from libpcap, or pcap-savefile(5) */ -static const struct { - uint32_t magic; -#define PCAP_MAGIC 0xa1b2c3d4 - - uint16_t major; -#define PCAP_VERSION_MAJOR 2 - - uint16_t minor; -#define PCAP_VERSION_MINOR 4 - - int32_t thiszone; - uint32_t sigfigs; - uint32_t snaplen; - - uint32_t linktype; -#define PCAP_LINKTYPE_ETHERNET 1 -} pcap_hdr = { - PCAP_MAGIC, PCAP_VERSION_MAJOR, PCAP_VERSION_MINOR, 0, 0, ETH_MAX_MTU, - PCAP_LINKTYPE_ETHERNET -}; - struct pcap_pkthdr { uint32_t tv_sec; uint32_t tv_usec; @@ -143,7 +122,6 @@ void pcap_multiple(const struct iovec *iov, size_t frame_parts, unsigned int n, * @iovcnt: Number of buffers (@iov entries) * @offset: Offset of the L2 frame within the full data length */ -/* cppcheck-suppress unusedFunction */ void pcap_iov(const struct iovec *iov, size_t iovcnt, size_t offset) { struct timespec now = { 0 }; @@ -163,6 +141,29 @@ void pcap_iov(const struct iovec *iov, size_t iovcnt, size_t offset) */ void pcap_init(struct ctx *c) { + /* See pcap.h from libpcap, or pcap-savefile(5) */ +#define PCAP_MAGIC 0xa1b2c3d4 +#define PCAP_VERSION_MAJOR 2 +#define PCAP_VERSION_MINOR 4 +#define PCAP_LINKTYPE_ETHERNET 1 + const struct { + uint32_t magic; + uint16_t major; + uint16_t minor; + + int32_t thiszone; + uint32_t sigfigs; + uint32_t snaplen; + + uint32_t linktype; + } pcap_hdr = { + .magic = PCAP_MAGIC, + .major = PCAP_VERSION_MAJOR, + .minor = PCAP_VERSION_MINOR, + .snaplen = tap_l2_max_len(c), + .linktype = PCAP_LINKTYPE_ETHERNET + }; + if (pcap_fd != -1) return; diff --git a/repair.c b/repair.c new file mode 100644 index 0000000..149fe51 --- /dev/null +++ b/repair.c @@ -0,0 +1,255 @@ +// SPDX-License-Identifier: GPL-2.0-or-later + +/* PASST - Plug A Simple Socket Transport + * for qemu/UNIX domain socket mode + * + * PASTA - Pack A Subtle Tap Abstraction + * for network namespace/tap device mode + * + * repair.c - Interface (server) for passt-repair, set/clear TCP_REPAIR + * + * Copyright (c) 2025 Red Hat GmbH + * Author: Stefano Brivio <sbrivio@redhat.com> + */ + +#include <errno.h> +#include <sys/socket.h> +#include <sys/uio.h> + +#include "util.h" +#include "ip.h" +#include "passt.h" +#include "inany.h" +#include "flow.h" +#include "flow_table.h" + +#include "repair.h" + +#define SCM_MAX_FD 253 /* From Linux kernel (include/net/scm.h), not in UAPI */ + +/* Wait for a while for TCP_REPAIR helper to connect if it's not there yet */ +#define REPAIR_ACCEPT_TIMEOUT_MS 10 +#define REPAIR_ACCEPT_TIMEOUT_US (REPAIR_ACCEPT_TIMEOUT_MS * 1000) + +/* Pending file descriptors for next repair_flush() call, or command change */ +static int repair_fds[SCM_MAX_FD]; + +/* Pending command: flush pending file descriptors if it changes */ +static int8_t repair_cmd; + +/* Number of pending file descriptors set in @repair_fds */ +static int repair_nfds; + +/** + * repair_sock_init() - Start listening for connections on helper socket + * @c: Execution context + */ +void repair_sock_init(const struct ctx *c) +{ + union epoll_ref ref = { .type = EPOLL_TYPE_REPAIR_LISTEN }; + struct epoll_event ev = { 0 }; + + if (c->fd_repair_listen == -1) + return; + + if (listen(c->fd_repair_listen, 0)) { + err_perror("listen() on repair helper socket, won't migrate"); + return; + } + + ref.fd = c->fd_repair_listen; + ev.events = EPOLLIN | EPOLLHUP | EPOLLET; + ev.data.u64 = ref.u64; + if (epoll_ctl(c->epollfd, EPOLL_CTL_ADD, c->fd_repair_listen, &ev)) + err_perror("repair helper socket epoll_ctl(), won't migrate"); +} + +/** + * repair_listen_handler() - Handle events on TCP_REPAIR helper listening socket + * @c: Execution context + * @events: epoll events + */ +void repair_listen_handler(struct ctx *c, uint32_t events) +{ + union epoll_ref ref = { .type = EPOLL_TYPE_REPAIR }; + struct epoll_event ev = { 0 }; + struct ucred ucred; + socklen_t len; + + if (events != EPOLLIN) { + debug("Spurious event 0x%04x on TCP_REPAIR helper socket", + events); + return; + } + + len = sizeof(ucred); + + /* Another client is already connected: accept and close right away. */ + if (c->fd_repair != -1) { + int discard = accept4(c->fd_repair_listen, NULL, NULL, + SOCK_NONBLOCK); + + if (discard == -1) + return; + + if (!getsockopt(discard, SOL_SOCKET, SO_PEERCRED, &ucred, &len)) + info("Discarding TCP_REPAIR helper, PID %i", ucred.pid); + + close(discard); + return; + } + + if ((c->fd_repair = accept4(c->fd_repair_listen, NULL, NULL, 0)) < 0) { + debug_perror("accept4() on TCP_REPAIR helper listening socket"); + return; + } + + if (!getsockopt(c->fd_repair, SOL_SOCKET, SO_PEERCRED, &ucred, &len)) + info("Accepted TCP_REPAIR helper, PID %i", ucred.pid); + + ref.fd = c->fd_repair; + ev.events = EPOLLHUP | EPOLLET; + ev.data.u64 = ref.u64; + if (epoll_ctl(c->epollfd, EPOLL_CTL_ADD, c->fd_repair, &ev)) { + debug_perror("epoll_ctl() on TCP_REPAIR helper socket"); + close(c->fd_repair); + c->fd_repair = -1; + } +} + +/** + * repair_close() - Close connection to TCP_REPAIR helper + * @c: Execution context + */ +void repair_close(struct ctx *c) +{ + debug("Closing TCP_REPAIR helper socket"); + + epoll_ctl(c->epollfd, EPOLL_CTL_DEL, c->fd_repair, NULL); + close(c->fd_repair); + c->fd_repair = -1; +} + +/** + * repair_handler() - Handle EPOLLHUP and EPOLLERR on TCP_REPAIR helper socket + * @c: Execution context + * @events: epoll events + */ +void repair_handler(struct ctx *c, uint32_t events) +{ + (void)events; + + repair_close(c); +} + +/** + * repair_wait() - Wait (with timeout) for TCP_REPAIR helper to connect + * @c: Execution context + */ +void repair_wait(struct ctx *c) +{ + struct timeval tv = { .tv_sec = 0, + .tv_usec = (long)(REPAIR_ACCEPT_TIMEOUT_US) }; + static_assert(REPAIR_ACCEPT_TIMEOUT_US < 1000 * 1000, + ".tv_usec is greater than 1000 * 1000"); + + if (c->fd_repair >= 0 || c->fd_repair_listen == -1) + return; + + if (setsockopt(c->fd_repair_listen, SOL_SOCKET, SO_RCVTIMEO, + &tv, sizeof(tv))) { + err_perror("Set timeout on TCP_REPAIR listening socket"); + return; + } + + repair_listen_handler(c, EPOLLIN); + + tv.tv_usec = 0; + if (setsockopt(c->fd_repair_listen, SOL_SOCKET, SO_RCVTIMEO, + &tv, sizeof(tv))) + err_perror("Clear timeout on TCP_REPAIR listening socket"); +} + +/** + * repair_flush() - Flush current set of sockets to helper, with current command + * @c: Execution context + * + * Return: 0 on success, negative error code on failure + */ +int repair_flush(struct ctx *c) +{ + char buf[CMSG_SPACE(sizeof(int) * SCM_MAX_FD)] + __attribute__ ((aligned(__alignof__(struct cmsghdr)))) = { 0 }; + struct iovec iov = { &repair_cmd, sizeof(repair_cmd) }; + struct cmsghdr *cmsg; + struct msghdr msg; + int8_t reply; + + if (!repair_nfds) + return 0; + + msg = (struct msghdr){ .msg_name = NULL, .msg_namelen = 0, + .msg_iov = &iov, .msg_iovlen = 1, + .msg_control = buf, + .msg_controllen = CMSG_SPACE(sizeof(int) * + repair_nfds), + .msg_flags = 0 }; + cmsg = CMSG_FIRSTHDR(&msg); + + cmsg->cmsg_level = SOL_SOCKET; + cmsg->cmsg_type = SCM_RIGHTS; + cmsg->cmsg_len = CMSG_LEN(sizeof(int) * repair_nfds); + memcpy(CMSG_DATA(cmsg), repair_fds, sizeof(int) * repair_nfds); + + repair_nfds = 0; + + if (sendmsg(c->fd_repair, &msg, 0) < 0) { + int ret = -errno; + err_perror("Failed to send sockets to TCP_REPAIR helper"); + repair_close(c); + return ret; + } + + if (recv(c->fd_repair, &reply, sizeof(reply), 0) < 0) { + int ret = -errno; + err_perror("Failed to receive reply from TCP_REPAIR helper"); + repair_close(c); + return ret; + } + + if (reply != repair_cmd) { + err("Unexpected reply from TCP_REPAIR helper: %d", reply); + repair_close(c); + return -ENXIO; + } + + return 0; +} + +/** + * repair_set() - Add socket to TCP_REPAIR set with given command + * @c: Execution context + * @s: Socket to add + * @cmd: TCP_REPAIR_ON, TCP_REPAIR_OFF, or TCP_REPAIR_OFF_NO_WP + * + * Return: 0 on success, negative error code on failure + */ +int repair_set(struct ctx *c, int s, int cmd) +{ + int rc; + + if (repair_nfds && repair_cmd != cmd) { + if ((rc = repair_flush(c))) + return rc; + } + + repair_cmd = cmd; + repair_fds[repair_nfds++] = s; + + if (repair_nfds >= SCM_MAX_FD) { + if ((rc = repair_flush(c))) + return rc; + } + + return 0; +} diff --git a/repair.h b/repair.h new file mode 100644 index 0000000..1d37922 --- /dev/null +++ b/repair.h @@ -0,0 +1,17 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later + * Copyright (c) 2025 Red Hat GmbH + * Author: Stefano Brivio <sbrivio@redhat.com> + */ + +#ifndef REPAIR_H +#define REPAIR_H + +void repair_sock_init(const struct ctx *c); +void repair_listen_handler(struct ctx *c, uint32_t events); +void repair_handler(struct ctx *c, uint32_t events); +void repair_close(struct ctx *c); +void repair_wait(struct ctx *c); +int repair_flush(struct ctx *c); +int repair_set(struct ctx *c, int s, int cmd); + +#endif /* REPAIR_H */ diff --git a/seccomp.sh b/seccomp.sh index 6499c58..a7bc417 100755 --- a/seccomp.sh +++ b/seccomp.sh @@ -14,8 +14,10 @@ # Author: Stefano Brivio <sbrivio@redhat.com> TMP="$(mktemp)" -IN="$@" OUT="$(mktemp)" +OUT_FINAL="${1}" +shift +IN="$@" [ -z "${ARCH}" ] && ARCH="$(uname -m)" [ -z "${CC}" ] && CC="cc" @@ -253,7 +255,7 @@ for __p in ${__profiles}; do __calls="${__calls} ${EXTRA_SYSCALLS:-}" __calls="$(filter ${__calls})" - cols="$(stty -a | sed -n 's/.*columns \([0-9]*\).*/\1/p' || :)" 2>/dev/null + cols="$(stty -a 2>/dev/null | sed -n 's/.*columns \([0-9]*\).*/\1/p' || :)" 2>/dev/null case $cols in [0-9]*) col_args="-w ${cols}";; *) col_args="";; esac echo "seccomp profile ${__p} allows: ${__calls}" | tr '\n' ' ' | fmt -t ${col_args} @@ -268,4 +270,4 @@ for __p in ${__profiles}; do gen_profile "${__p}" ${__calls} done -mv "${OUT}" seccomp.h +mv "${OUT}" "${OUT_FINAL}" diff --git a/tap.c b/tap.c index 5347df4..d630f6d 100644 --- a/tap.c +++ b/tap.c @@ -56,16 +56,70 @@ #include "netlink.h" #include "pasta.h" #include "packet.h" +#include "repair.h" #include "tap.h" #include "log.h" +#include "vhost_user.h" +#include "vu_common.h" + +/* Maximum allowed frame lengths (including L2 header) */ + +/* Verify that an L2 frame length limit is large enough to contain the header, + * but small enough to fit in the packet pool + */ +#define CHECK_FRAME_LEN(len) \ + static_assert((len) >= ETH_HLEN && (len) <= PACKET_MAX_LEN, \ + #len " has bad value") + +CHECK_FRAME_LEN(L2_MAX_LEN_PASTA); +CHECK_FRAME_LEN(L2_MAX_LEN_PASST); +CHECK_FRAME_LEN(L2_MAX_LEN_VU); + +/* We try size the packet pools so that we can use a single batch for the entire + * packet buffer. This might be exceeded for vhost-user, though, which uses its + * own buffers rather than pkt_buf. + * + * This is just a tuning parameter, the code will work with slightly more + * overhead if it's incorrect. So, we estimate based on the minimum practical + * frame size - an empty UDP datagram - rather than the minimum theoretical + * frame size. + * + * FIXME: Profile to work out how big this actually needs to be to amortise + * per-batch syscall overheads + */ +#define TAP_MSGS_IP4 \ + DIV_ROUND_UP(sizeof(pkt_buf), \ + ETH_HLEN + sizeof(struct iphdr) + sizeof(struct udphdr)) +#define TAP_MSGS_IP6 \ + DIV_ROUND_UP(sizeof(pkt_buf), \ + ETH_HLEN + sizeof(struct ipv6hdr) + sizeof(struct udphdr)) /* IPv4 (plus ARP) and IPv6 message batches from tap/guest to IP handlers */ -static PACKET_POOL_NOINIT(pool_tap4, TAP_MSGS, pkt_buf); -static PACKET_POOL_NOINIT(pool_tap6, TAP_MSGS, pkt_buf); +static PACKET_POOL_NOINIT(pool_tap4, TAP_MSGS_IP4, pkt_buf); +static PACKET_POOL_NOINIT(pool_tap6, TAP_MSGS_IP6, pkt_buf); #define TAP_SEQS 128 /* Different L4 tuples in one batch */ #define FRAGMENT_MSG_RATE 10 /* # seconds between fragment warnings */ +/** + * tap_l2_max_len() - Maximum frame size (including L2 header) for current mode + * @c: Execution context + */ +unsigned long tap_l2_max_len(const struct ctx *c) +{ + /* NOLINTBEGIN(bugprone-branch-clone): values can be the same */ + switch (c->mode) { + case MODE_PASST: + return L2_MAX_LEN_PASST; + case MODE_PASTA: + return L2_MAX_LEN_PASTA; + case MODE_VU: + return L2_MAX_LEN_VU; + } + /* NOLINTEND(bugprone-branch-clone) */ + ASSERT(0); +} + /** * tap_send_single() - Send a single frame * @c: Execution context @@ -78,16 +132,22 @@ void tap_send_single(const struct ctx *c, const void *data, size_t l2len) struct iovec iov[2]; size_t iovcnt = 0; - if (c->mode == MODE_PASST) { + switch (c->mode) { + case MODE_PASST: iov[iovcnt] = IOV_OF_LVALUE(vnet_len); iovcnt++; + /* fall through */ + case MODE_PASTA: + iov[iovcnt].iov_base = (void *)data; + iov[iovcnt].iov_len = l2len; + iovcnt++; + + tap_send_frames(c, iov, iovcnt, 1); + break; + case MODE_VU: + vu_send_single(c, data, l2len); + break; } - - iov[iovcnt].iov_base = (void *)data; - iov[iovcnt].iov_len = l2len; - iovcnt++; - - tap_send_frames(c, iov, iovcnt, 1); } /** @@ -113,7 +173,7 @@ const struct in6_addr *tap_ip6_daddr(const struct ctx *c, * * Return: pointer at which to write the packet's payload */ -static void *tap_push_l2h(const struct ctx *c, void *buf, uint16_t proto) +void *tap_push_l2h(const struct ctx *c, void *buf, uint16_t proto) { struct ethhdr *eh = (struct ethhdr *)buf; @@ -134,8 +194,8 @@ static void *tap_push_l2h(const struct ctx *c, void *buf, uint16_t proto) * * Return: pointer at which to write the packet's payload */ -static void *tap_push_ip4h(struct iphdr *ip4h, struct in_addr src, - struct in_addr dst, size_t l4len, uint8_t proto) +void *tap_push_ip4h(struct iphdr *ip4h, struct in_addr src, + struct in_addr dst, size_t l4len, uint8_t proto) { uint16_t l3len = l4len + sizeof(*ip4h); @@ -144,13 +204,43 @@ static void *tap_push_ip4h(struct iphdr *ip4h, struct in_addr src, ip4h->tos = 0; ip4h->tot_len = htons(l3len); ip4h->id = 0; - ip4h->frag_off = 0; + ip4h->frag_off = htons(IP_DF); ip4h->ttl = 255; ip4h->protocol = proto; ip4h->saddr = src.s_addr; ip4h->daddr = dst.s_addr; ip4h->check = csum_ip4_header(l3len, proto, src, dst); - return ip4h + 1; + return (char *)ip4h + sizeof(*ip4h); +} + +/** + * tap_push_uh4() - Build UDPv4 header with checksum + * @c: Execution context + * @src: IPv4 source address + * @sport: UDP source port + * @dst: IPv4 destination address + * @dport: UDP destination port + * @in: UDP payload contents (not including UDP header) + * @dlen: UDP payload length (not including UDP header) + * + * Return: pointer at which to write the packet's payload + */ +void *tap_push_uh4(struct udphdr *uh, struct in_addr src, in_port_t sport, + struct in_addr dst, in_port_t dport, + const void *in, size_t dlen) +{ + size_t l4len = dlen + sizeof(struct udphdr); + const struct iovec iov = { + .iov_base = (void *)in, + .iov_len = dlen + }; + struct iov_tail payload = IOV_TAIL(&iov, 1, 0); + + uh->source = htons(sport); + uh->dest = htons(dport); + uh->len = htons(l4len); + csum_udp4(uh, src, dst, &payload); + return (char *)uh + sizeof(*uh); } /** @@ -160,7 +250,7 @@ static void *tap_push_ip4h(struct iphdr *ip4h, struct in_addr src, * @sport: UDP source port * @dst: IPv4 destination address * @dport: UDP destination port - * @in: UDP payload contents (not including UDP header) + * @in: UDP payload contents (not including UDP header) * @dlen: UDP payload length (not including UDP header) */ void tap_udp4_send(const struct ctx *c, struct in_addr src, in_port_t sport, @@ -171,18 +261,9 @@ void tap_udp4_send(const struct ctx *c, struct in_addr src, in_port_t sport, char buf[USHRT_MAX]; struct iphdr *ip4h = tap_push_l2h(c, buf, ETH_P_IP); struct udphdr *uh = tap_push_ip4h(ip4h, src, dst, l4len, IPPROTO_UDP); - char *data = (char *)(uh + 1); - const struct iovec iov = { - .iov_base = (void *)in, - .iov_len = dlen - }; + char *data = tap_push_uh4(uh, src, sport, dst, dport, in, dlen); - uh->source = htons(sport); - uh->dest = htons(dport); - uh->len = htons(l4len); - csum_udp4(uh, src, dst, &iov, 1, 0); memcpy(data, in, dlen); - tap_send_single(c, buf, dlen + (data - buf)); } @@ -219,10 +300,9 @@ void tap_icmp4_send(const struct ctx *c, struct in_addr src, struct in_addr dst, * * Return: pointer at which to write the packet's payload */ -static void *tap_push_ip6h(struct ipv6hdr *ip6h, - const struct in6_addr *src, - const struct in6_addr *dst, - size_t l4len, uint8_t proto, uint32_t flow) +void *tap_push_ip6h(struct ipv6hdr *ip6h, + const struct in6_addr *src, const struct in6_addr *dst, + size_t l4len, uint8_t proto, uint32_t flow) { ip6h->payload_len = htons(l4len); ip6h->priority = 0; @@ -231,10 +311,40 @@ static void *tap_push_ip6h(struct ipv6hdr *ip6h, ip6h->hop_limit = 255; ip6h->saddr = *src; ip6h->daddr = *dst; - ip6h->flow_lbl[0] = (flow >> 16) & 0xf; - ip6h->flow_lbl[1] = (flow >> 8) & 0xff; - ip6h->flow_lbl[2] = (flow >> 0) & 0xff; - return ip6h + 1; + ip6_set_flow_lbl(ip6h, flow); + return (char *)ip6h + sizeof(*ip6h); +} + +/** + * tap_push_uh6() - Build UDPv6 header with checksum + * @c: Execution context + * @src: IPv6 source address + * @sport: UDP source port + * @dst: IPv6 destination address + * @dport: UDP destination port + * @flow: Flow label + * @in: UDP payload contents (not including UDP header) + * @dlen: UDP payload length (not including UDP header) + * + * Return: pointer at which to write the packet's payload + */ +void *tap_push_uh6(struct udphdr *uh, + const struct in6_addr *src, in_port_t sport, + const struct in6_addr *dst, in_port_t dport, + void *in, size_t dlen) +{ + size_t l4len = dlen + sizeof(struct udphdr); + const struct iovec iov = { + .iov_base = in, + .iov_len = dlen + }; + struct iov_tail payload = IOV_TAIL(&iov, 1, 0); + + uh->source = htons(sport); + uh->dest = htons(dport); + uh->len = htons(l4len); + csum_udp6(uh, src, dst, &payload); + return (char *)uh + sizeof(*uh); } /** @@ -245,7 +355,7 @@ static void *tap_push_ip6h(struct ipv6hdr *ip6h, * @dst: IPv6 destination address * @dport: UDP destination port * @flow: Flow label - * @in: UDP payload contents (not including UDP header) + * @in: UDP payload contents (not including UDP header) * @dlen: UDP payload length (not including UDP header) */ void tap_udp6_send(const struct ctx *c, @@ -258,18 +368,9 @@ void tap_udp6_send(const struct ctx *c, struct ipv6hdr *ip6h = tap_push_l2h(c, buf, ETH_P_IPV6); struct udphdr *uh = tap_push_ip6h(ip6h, src, dst, l4len, IPPROTO_UDP, flow); - char *data = (char *)(uh + 1); - const struct iovec iov = { - .iov_base = in, - .iov_len = dlen - }; + char *data = tap_push_uh6(uh, src, sport, dst, dport, in, dlen); - uh->source = htons(sport); - uh->dest = htons(dport); - uh->len = htons(l4len); - csum_udp6(uh, src, dst, &iov, 1, 0); memcpy(data, in, dlen); - tap_send_single(c, buf, dlen + (data - buf)); } @@ -414,10 +515,18 @@ size_t tap_send_frames(const struct ctx *c, const struct iovec *iov, if (!nframes) return 0; - if (c->mode == MODE_PASTA) + switch (c->mode) { + case MODE_PASTA: m = tap_send_frames_pasta(c, iov, bufs_per_frame, nframes); - else + break; + case MODE_PASST: m = tap_send_frames_passt(c, iov, bufs_per_frame, nframes); + break; + case MODE_VU: + /* fall through */ + default: + ASSERT(0); + } if (m < nframes) debug("tap: failed to send %zu frames of %zu", @@ -450,6 +559,7 @@ PACKET_POOL_DECL(pool_l4, UIO_MAXIOV, pkt_buf); * struct l4_seq4_t - Message sequence for one protocol handler call, IPv4 * @msgs: Count of messages in sequence * @protocol: Protocol number + * @ttl: Time to live * @source: Source port * @dest: Destination port * @saddr: Source address @@ -458,6 +568,7 @@ PACKET_POOL_DECL(pool_l4, UIO_MAXIOV, pkt_buf); */ static struct tap4_l4_t { uint8_t protocol; + uint8_t ttl; uint16_t source; uint16_t dest; @@ -472,14 +583,17 @@ static struct tap4_l4_t { * struct l4_seq6_t - Message sequence for one protocol handler call, IPv6 * @msgs: Count of messages in sequence * @protocol: Protocol number + * @flow_lbl: IPv6 flow label * @source: Source port * @dest: Destination port * @saddr: Source address * @daddr: Destination address + * @hop_limit: Hop limit * @msg: Array of messages that can be handled in a single call */ static struct tap6_l4_t { uint8_t protocol; + uint32_t flow_lbl :20; uint16_t source; uint16_t dest; @@ -487,6 +601,8 @@ static struct tap6_l4_t { struct in6_addr saddr; struct in6_addr daddr; + uint8_t hop_limit; + struct pool_l4_t p; } tap6_l4[TAP_SEQS /* Arbitrary: TAP_MSGS in theory, so limit in users */]; @@ -675,7 +791,8 @@ resume: #define L4_MATCH(iph, uh, seq) \ ((seq)->protocol == (iph)->protocol && \ (seq)->source == (uh)->source && (seq)->dest == (uh)->dest && \ - (seq)->saddr.s_addr == (iph)->saddr && (seq)->daddr.s_addr == (iph)->daddr) + (seq)->saddr.s_addr == (iph)->saddr && \ + (seq)->daddr.s_addr == (iph)->daddr && (seq)->ttl == (iph)->ttl) #define L4_SET(iph, uh, seq) \ do { \ @@ -684,6 +801,7 @@ resume: (seq)->dest = (uh)->dest; \ (seq)->saddr.s_addr = (iph)->saddr; \ (seq)->daddr.s_addr = (iph)->daddr; \ + (seq)->ttl = (iph)->ttl; \ } while (0) if (seq && L4_MATCH(iph, uh, seq) && seq->p.count < UIO_MAXIOV) @@ -725,14 +843,14 @@ append: for (k = 0; k < p->count; ) k += tcp_tap_handler(c, PIF_TAP, AF_INET, &seq->saddr, &seq->daddr, - p, k, now); + 0, p, k, now); } else if (seq->protocol == IPPROTO_UDP) { if (c->no_udp) continue; for (k = 0; k < p->count; ) k += udp_tap_handler(c, PIF_TAP, AF_INET, &seq->saddr, &seq->daddr, - p, k, now); + seq->ttl, p, k, now); } } @@ -853,16 +971,20 @@ resume: ((seq)->protocol == (proto) && \ (seq)->source == (uh)->source && \ (seq)->dest == (uh)->dest && \ + (seq)->flow_lbl == ip6_get_flow_lbl(ip6h) && \ IN6_ARE_ADDR_EQUAL(&(seq)->saddr, saddr) && \ - IN6_ARE_ADDR_EQUAL(&(seq)->daddr, daddr)) + IN6_ARE_ADDR_EQUAL(&(seq)->daddr, daddr) && \ + (seq)->hop_limit == (ip6h)->hop_limit) #define L4_SET(ip6h, proto, uh, seq) \ do { \ (seq)->protocol = (proto); \ (seq)->source = (uh)->source; \ (seq)->dest = (uh)->dest; \ + (seq)->flow_lbl = ip6_get_flow_lbl(ip6h); \ (seq)->saddr = *saddr; \ (seq)->daddr = *daddr; \ + (seq)->hop_limit = (ip6h)->hop_limit; \ } while (0) if (seq && L4_MATCH(ip6h, proto, uh, seq) && @@ -906,14 +1028,14 @@ append: for (k = 0; k < p->count; ) k += tcp_tap_handler(c, PIF_TAP, AF_INET6, &seq->saddr, &seq->daddr, - p, k, now); + seq->flow_lbl, p, k, now); } else if (seq->protocol == IPPROTO_UDP) { if (c->no_udp) continue; for (k = 0; k < p->count; ) k += udp_tap_handler(c, PIF_TAP, AF_INET6, &seq->saddr, &seq->daddr, - p, k, now); + seq->hop_limit, p, k, now); } } @@ -948,8 +1070,10 @@ void tap_handler(struct ctx *c, const struct timespec *now) * @c: Execution context * @l2len: Total L2 packet length * @p: Packet buffer + * @now: Current timestamp */ -void tap_add_packet(struct ctx *c, ssize_t l2len, char *p) +void tap_add_packet(struct ctx *c, ssize_t l2len, char *p, + const struct timespec *now) { const struct ethhdr *eh; @@ -965,9 +1089,17 @@ void tap_add_packet(struct ctx *c, ssize_t l2len, char *p) switch (ntohs(eh->h_proto)) { case ETH_P_ARP: case ETH_P_IP: + if (pool_full(pool_tap4)) { + tap4_handler(c, pool_tap4, now); + pool_flush(pool_tap4); + } packet_add(pool_tap4, l2len, p); break; case ETH_P_IPV6: + if (pool_full(pool_tap6)) { + tap6_handler(c, pool_tap6, now); + pool_flush(pool_tap6); + } packet_add(pool_tap6, l2len, p); break; default: @@ -979,17 +1111,19 @@ void tap_add_packet(struct ctx *c, ssize_t l2len, char *p) * tap_sock_reset() - Handle closing or failure of connect AF_UNIX socket * @c: Execution context */ -static void tap_sock_reset(struct ctx *c) +void tap_sock_reset(struct ctx *c) { info("Client connection closed%s", c->one_off ? ", exiting" : ""); if (c->one_off) - exit(EXIT_SUCCESS); + _exit(EXIT_SUCCESS); /* Close the connected socket, wait for a new connection */ - epoll_ctl(c->epollfd, EPOLL_CTL_DEL, c->fd_tap, NULL); + epoll_del(c, c->fd_tap); close(c->fd_tap); c->fd_tap = -1; + if (c->mode == MODE_VU) + vu_cleanup(c->vdev); } /** @@ -1016,7 +1150,7 @@ static void tap_passt_input(struct ctx *c, const struct timespec *now) do { n = recv(c->fd_tap, pkt_buf + partial_len, - TAP_BUF_BYTES - partial_len, MSG_DONTWAIT); + sizeof(pkt_buf) - partial_len, MSG_DONTWAIT); } while ((n < 0) && errno == EINTR); if (n < 0) { @@ -1033,7 +1167,7 @@ static void tap_passt_input(struct ctx *c, const struct timespec *now) while (n >= (ssize_t)sizeof(uint32_t)) { uint32_t l2len = ntohl_unaligned(p); - if (l2len < sizeof(struct ethhdr) || l2len > ETH_MAX_MTU) { + if (l2len < sizeof(struct ethhdr) || l2len > L2_MAX_LEN_PASST) { err("Bad frame size from guest, resetting connection"); tap_sock_reset(c); return; @@ -1046,7 +1180,7 @@ static void tap_passt_input(struct ctx *c, const struct timespec *now) p += sizeof(uint32_t); n -= sizeof(uint32_t); - tap_add_packet(c, l2len, p); + tap_add_packet(c, l2len, p, now); p += l2len; n -= l2len; @@ -1087,8 +1221,10 @@ static void tap_pasta_input(struct ctx *c, const struct timespec *now) tap_flush_pools(); - for (n = 0; n <= (ssize_t)(TAP_BUF_BYTES - ETH_MAX_MTU); n += len) { - len = read(c->fd_tap, pkt_buf + n, ETH_MAX_MTU); + for (n = 0; + n <= (ssize_t)(sizeof(pkt_buf) - L2_MAX_LEN_PASTA); + n += len) { + len = read(c->fd_tap, pkt_buf + n, L2_MAX_LEN_PASTA); if (len == 0) { die("EOF on tap device, exiting"); @@ -1106,10 +1242,10 @@ static void tap_pasta_input(struct ctx *c, const struct timespec *now) /* Ignore frames of bad length */ if (len < (ssize_t)sizeof(struct ethhdr) || - len > (ssize_t)ETH_MAX_MTU) + len > (ssize_t)L2_MAX_LEN_PASTA) continue; - tap_add_packet(c, len, pkt_buf + n); + tap_add_packet(c, len, pkt_buf + n, now); } tap_handler(c, now); @@ -1132,72 +1268,35 @@ void tap_handler_pasta(struct ctx *c, uint32_t events, } /** - * tap_sock_unix_open() - Create and bind AF_UNIX socket - * @sock_path: Socket path. If empty, set on return (UNIX_SOCK_PATH as prefix) - * - * Return: socket descriptor on success, won't return on failure + * tap_backend_show_hints() - Give help information to start QEMU + * @c: Execution context */ -int tap_sock_unix_open(char *sock_path) +static void tap_backend_show_hints(struct ctx *c) { - int fd = socket(AF_UNIX, SOCK_STREAM | SOCK_CLOEXEC, 0); - struct sockaddr_un addr = { - .sun_family = AF_UNIX, - }; - int i; - - if (fd < 0) - die_perror("Failed to open UNIX domain socket"); - - for (i = 1; i < UNIX_SOCK_MAX; i++) { - char *path = addr.sun_path; - int ex, ret; - - if (*sock_path) - memcpy(path, sock_path, UNIX_PATH_MAX); - else if (snprintf_check(path, UNIX_PATH_MAX - 1, - UNIX_SOCK_PATH, i)) - die_perror("Can't build UNIX domain socket path"); - - ex = socket(AF_UNIX, SOCK_STREAM | SOCK_NONBLOCK | SOCK_CLOEXEC, - 0); - if (ex < 0) - die_perror("Failed to check for UNIX domain conflicts"); - - ret = connect(ex, (const struct sockaddr *)&addr, sizeof(addr)); - if (!ret || (errno != ENOENT && errno != ECONNREFUSED && - errno != EACCES)) { - if (*sock_path) - die("Socket path %s already in use", path); - - close(ex); - continue; - } - close(ex); - - unlink(path); - ret = bind(fd, (const struct sockaddr *)&addr, sizeof(addr)); - if (*sock_path && ret) - die_perror("Failed to bind UNIX domain socket"); - - if (!ret) - break; + switch (c->mode) { + case MODE_PASTA: + /* No hints */ + break; + case MODE_PASST: + info("\nYou can now start qemu (>= 7.2, with commit 13c6be96618c):"); + info(" kvm ... -device virtio-net-pci,netdev=s -netdev stream,id=s,server=off,addr.type=unix,addr.path=%s", + c->sock_path); + info("or qrap, for earlier qemu versions:"); + info(" ./qrap 5 kvm ... -net socket,fd=5 -net nic,model=virtio"); + break; + case MODE_VU: + info("You can start qemu with:"); + info(" kvm ... -chardev socket,id=chr0,path=%s -netdev vhost-user,id=netdev0,chardev=chr0 -device virtio-net,netdev=netdev0 -object memory-backend-memfd,id=memfd0,share=on,size=$RAMSIZE -numa node,memdev=memfd0\n", + c->sock_path); + break; } - - if (i == UNIX_SOCK_MAX) - die_perror("Failed to bind UNIX domain socket"); - - info("UNIX domain socket bound at %s", addr.sun_path); - if (!*sock_path) - memcpy(sock_path, addr.sun_path, UNIX_PATH_MAX); - - return fd; } /** * tap_sock_unix_init() - Start listening for connections on AF_UNIX socket * @c: Execution context */ -static void tap_sock_unix_init(struct ctx *c) +static void tap_sock_unix_init(const struct ctx *c) { union epoll_ref ref = { .type = EPOLL_TYPE_TAP_LISTEN }; struct epoll_event ev = { 0 }; @@ -1208,12 +1307,33 @@ static void tap_sock_unix_init(struct ctx *c) ev.events = EPOLLIN | EPOLLET; ev.data.u64 = ref.u64; epoll_ctl(c->epollfd, EPOLL_CTL_ADD, c->fd_tap_listen, &ev); +} - info("\nYou can now start qemu (>= 7.2, with commit 13c6be96618c):"); - info(" kvm ... -device virtio-net-pci,netdev=s -netdev stream,id=s,server=off,addr.type=unix,addr.path=%s", - c->sock_path); - info("or qrap, for earlier qemu versions:"); - info(" ./qrap 5 kvm ... -net socket,fd=5 -net nic,model=virtio"); +/** + * tap_start_connection() - start a new connection + * @c: Execution context + */ +static void tap_start_connection(const struct ctx *c) +{ + struct epoll_event ev = { 0 }; + union epoll_ref ref = { 0 }; + + ref.fd = c->fd_tap; + switch (c->mode) { + case MODE_PASST: + ref.type = EPOLL_TYPE_TAP_PASST; + break; + case MODE_PASTA: + ref.type = EPOLL_TYPE_TAP_PASTA; + break; + case MODE_VU: + ref.type = EPOLL_TYPE_VHOST_CMD; + break; + } + + ev.events = EPOLLIN | EPOLLRDHUP; + ev.data.u64 = ref.u64; + epoll_ctl(c->epollfd, EPOLL_CTL_ADD, c->fd_tap, &ev); } /** @@ -1223,8 +1343,6 @@ static void tap_sock_unix_init(struct ctx *c) */ void tap_listen_handler(struct ctx *c, uint32_t events) { - union epoll_ref ref = { .type = EPOLL_TYPE_TAP_PASST }; - struct epoll_event ev = { 0 }; int v = INT_MAX / 2; struct ucred ucred; socklen_t len; @@ -1263,10 +1381,7 @@ void tap_listen_handler(struct ctx *c, uint32_t events) setsockopt(c->fd_tap, SOL_SOCKET, SO_SNDBUF, &v, sizeof(v))) trace("tap: failed to set SO_SNDBUF to %i", v); - ref.fd = c->fd_tap; - ev.events = EPOLLIN | EPOLLRDHUP; - ev.data.u64 = ref.u64; - epoll_ctl(c->epollfd, EPOLL_CTL_ADD, c->fd_tap, &ev); + tap_start_connection(c); } /** @@ -1310,58 +1425,61 @@ static int tap_ns_tun(void *arg) */ static void tap_sock_tun_init(struct ctx *c) { - union epoll_ref ref = { .type = EPOLL_TYPE_TAP_PASTA }; - struct epoll_event ev = { 0 }; - NS_CALL(tap_ns_tun, c); if (c->fd_tap == -1) die("Failed to set up tap device in namespace"); pasta_ns_conf(c); - ref.fd = c->fd_tap; - ev.events = EPOLLIN | EPOLLRDHUP; - ev.data.u64 = ref.u64; - epoll_ctl(c->epollfd, EPOLL_CTL_ADD, c->fd_tap, &ev); + tap_start_connection(c); } /** - * tap_sock_init() - Create and set up AF_UNIX socket or tuntap file descriptor - * @c: Execution context + * tap_sock_update_pool() - Set the buffer base and size for the pool of packets + * @base: Buffer base + * @size Buffer size */ -void tap_sock_init(struct ctx *c) +void tap_sock_update_pool(void *base, size_t size) { - size_t sz = sizeof(pkt_buf); int i; - pool_tap4_storage = PACKET_INIT(pool_tap4, TAP_MSGS, pkt_buf, sz); - pool_tap6_storage = PACKET_INIT(pool_tap6, TAP_MSGS, pkt_buf, sz); + pool_tap4_storage = PACKET_INIT(pool_tap4, TAP_MSGS_IP4, base, size); + pool_tap6_storage = PACKET_INIT(pool_tap6, TAP_MSGS_IP6, base, size); for (i = 0; i < TAP_SEQS; i++) { - tap4_l4[i].p = PACKET_INIT(pool_l4, UIO_MAXIOV, pkt_buf, sz); - tap6_l4[i].p = PACKET_INIT(pool_l4, UIO_MAXIOV, pkt_buf, sz); + tap4_l4[i].p = PACKET_INIT(pool_l4, UIO_MAXIOV, base, size); + tap6_l4[i].p = PACKET_INIT(pool_l4, UIO_MAXIOV, base, size); + } +} + +/** + * tap_backend_init() - Create and set up AF_UNIX socket or + * tuntap file descriptor + * @c: Execution context + */ +void tap_backend_init(struct ctx *c) +{ + if (c->mode == MODE_VU) { + tap_sock_update_pool(NULL, 0); + vu_init(c); + } else { + tap_sock_update_pool(pkt_buf, sizeof(pkt_buf)); } if (c->fd_tap != -1) { /* Passed as --fd */ - struct epoll_event ev = { 0 }; - union epoll_ref ref; - ASSERT(c->one_off); - ref.fd = c->fd_tap; - if (c->mode == MODE_PASST) - ref.type = EPOLL_TYPE_TAP_PASST; - else - ref.type = EPOLL_TYPE_TAP_PASTA; - - ev.events = EPOLLIN | EPOLLRDHUP; - ev.data.u64 = ref.u64; - epoll_ctl(c->epollfd, EPOLL_CTL_ADD, c->fd_tap, &ev); + tap_start_connection(c); return; } - if (c->mode == MODE_PASTA) { + switch (c->mode) { + case MODE_PASTA: tap_sock_tun_init(c); - } else { + break; + case MODE_VU: + repair_sock_init(c); + /* fall through */ + case MODE_PASST: tap_sock_unix_init(c); /* In passt mode, we don't know the guest's MAC address until it @@ -1369,5 +1487,8 @@ void tap_sock_init(struct ctx *c) * first packets will reach it. */ memset(&c->guest_mac, 0xff, sizeof(c->guest_mac)); + break; } + + tap_backend_show_hints(c); } diff --git a/tap.h b/tap.h index 85f1e84..6fe3d15 100644 --- a/tap.h +++ b/tap.h @@ -6,7 +6,32 @@ #ifndef TAP_H #define TAP_H -#define ETH_HDR_INIT(proto) { .h_proto = htons_constant(proto) } +/** L2_MAX_LEN_PASTA - Maximum frame length for pasta mode (with L2 header) + * + * The kernel tuntap device imposes a maximum frame size of 65535 including + * 'hard_header_len' (14 bytes for L2 Ethernet in the case of "tap" mode). + */ +#define L2_MAX_LEN_PASTA USHRT_MAX + +/** L2_MAX_LEN_PASST - Maximum frame length for passt mode (with L2 header) + * + * The only structural limit the QEMU socket protocol imposes on frames is + * (2^32-1) bytes, but that would be ludicrously long in practice. For now, + * limit it somewhat arbitrarily to 65535 bytes. FIXME: Work out an appropriate + * limit with more precision. + */ +#define L2_MAX_LEN_PASST USHRT_MAX + +/** L2_MAX_LEN_VU - Maximum frame length for vhost-user mode (with L2 header) + * + * vhost-user allows multiple buffers per frame, each of which can be quite + * large, so the inherent frame size limit is rather large. Much larger than is + * actually useful for IP. For now limit arbitrarily to 65535 bytes. FIXME: + * Work out an appropriate limit with more precision. + */ +#define L2_MAX_LEN_VU USHRT_MAX + +struct udphdr; /** * struct tap_hdr - tap backend specific headers @@ -40,9 +65,27 @@ static inline struct iovec tap_hdr_iov(const struct ctx *c, */ static inline void tap_hdr_update(struct tap_hdr *thdr, size_t l2len) { - thdr->vnet_len = htonl(l2len); + if (thdr) + thdr->vnet_len = htonl(l2len); } +unsigned long tap_l2_max_len(const struct ctx *c); +void *tap_push_l2h(const struct ctx *c, void *buf, uint16_t proto); +void *tap_push_ip4h(struct iphdr *ip4h, struct in_addr src, + struct in_addr dst, size_t l4len, uint8_t proto); +void *tap_push_uh4(struct udphdr *uh, struct in_addr src, in_port_t sport, + struct in_addr dst, in_port_t dport, + const void *in, size_t dlen); +void *tap_push_uh6(struct udphdr *uh, + const struct in6_addr *src, in_port_t sport, + const struct in6_addr *dst, in_port_t dport, + void *in, size_t dlen); +void *tap_push_ip4h(struct iphdr *ip4h, struct in_addr src, + struct in_addr dst, size_t l4len, uint8_t proto); +void *tap_push_ip6h(struct ipv6hdr *ip6h, + const struct in6_addr *src, + const struct in6_addr *dst, + size_t l4len, uint8_t proto, uint32_t flow); void tap_udp4_send(const struct ctx *c, struct in_addr src, in_port_t sport, struct in_addr dst, in_port_t dport, const void *in, size_t dlen); @@ -50,6 +93,9 @@ void tap_icmp4_send(const struct ctx *c, struct in_addr src, struct in_addr dst, const void *in, size_t l4len); const struct in6_addr *tap_ip6_daddr(const struct ctx *c, const struct in6_addr *src); +void *tap_push_ip6h(struct ipv6hdr *ip6h, + const struct in6_addr *src, const struct in6_addr *dst, + size_t l4len, uint8_t proto, uint32_t flow); void tap_udp6_send(const struct ctx *c, const struct in6_addr *src, in_port_t sport, const struct in6_addr *dst, in_port_t dport, @@ -68,9 +114,12 @@ void tap_handler_pasta(struct ctx *c, uint32_t events, void tap_handler_passt(struct ctx *c, uint32_t events, const struct timespec *now); int tap_sock_unix_open(char *sock_path); -void tap_sock_init(struct ctx *c); +void tap_sock_reset(struct ctx *c); +void tap_sock_update_pool(void *base, size_t size); +void tap_backend_init(struct ctx *c); void tap_flush_pools(void); void tap_handler(struct ctx *c, const struct timespec *now); -void tap_add_packet(struct ctx *c, ssize_t l2len, char *p); +void tap_add_packet(struct ctx *c, ssize_t l2len, char *p, + const struct timespec *now); #endif /* TAP_H */ diff --git a/tcp.c b/tcp.c index 1eb85bb..0ac298a 100644 --- a/tcp.c +++ b/tcp.c @@ -280,6 +280,7 @@ #include <stddef.h> #include <string.h> #include <sys/epoll.h> +#include <sys/ioctl.h> #include <sys/socket.h> #include <sys/timerfd.h> #include <sys/types.h> @@ -287,6 +288,8 @@ #include <time.h> #include <arpa/inet.h> +#include <linux/sockios.h> + #include "checksum.h" #include "util.h" #include "iov.h" @@ -299,11 +302,28 @@ #include "log.h" #include "inany.h" #include "flow.h" +#include "repair.h" #include "linux_dep.h" #include "flow_table.h" #include "tcp_internal.h" #include "tcp_buf.h" +#include "tcp_vu.h" + +#ifndef __USE_MISC +/* From Linux UAPI, missing in netinet/tcp.h provided by musl */ +struct tcp_repair_opt { + __u32 opt_code; + __u32 opt_val; +}; + +enum { + TCP_NO_QUEUE, + TCP_RECV_QUEUE, + TCP_SEND_QUEUE, + TCP_QUEUES_NR, +}; +#endif /* MSS rounding: see SET_MSS() */ #define MSS_DEFAULT 536 @@ -325,6 +345,19 @@ ((conn)->events & (SOCK_FIN_RCVD | TAP_FIN_RCVD))) #define CONN_HAS(conn, set) (((conn)->events & (set)) == (set)) +/* Buffers to migrate pending data from send and receive queues. No, they don't + * use memory if we don't use them. And we're going away after this, so splurge. + */ +#define TCP_MIGRATE_SND_QUEUE_MAX (64 << 20) +#define TCP_MIGRATE_RCV_QUEUE_MAX (64 << 20) +uint8_t tcp_migrate_snd_queue [TCP_MIGRATE_SND_QUEUE_MAX]; +uint8_t tcp_migrate_rcv_queue [TCP_MIGRATE_RCV_QUEUE_MAX]; + +#define TCP_MIGRATE_RESTORE_CHUNK_MIN 1024 /* Try smaller when above this */ + +/* "Extended" data (not stored in the flow table) for TCP flow migration */ +static struct tcp_tap_transfer_ext migrate_ext[FLOW_MAX]; + static const char *tcp_event_str[] __attribute((__unused__)) = { "SOCK_ACCEPTED", "TAP_SYN_RCVD", "ESTABLISHED", "TAP_SYN_ACK_SENT", @@ -337,14 +370,14 @@ static const char *tcp_state_str[] __attribute((__unused__)) = { "SYN_RCVD", /* approximately maps to TAP_SYN_ACK_SENT */ /* Passive close: */ - "CLOSE_WAIT", "CLOSE_WAIT", "LAST_ACK", "LAST_ACK", "LAST_ACK", + "CLOSE_WAIT", "CLOSE_WAIT", "CLOSE_WAIT", "LAST_ACK", "LAST_ACK", /* Active close (+5): */ "CLOSING", "FIN_WAIT_1", "FIN_WAIT_1", "FIN_WAIT_2", "TIME_WAIT", }; static const char *tcp_flag_str[] __attribute((__unused__)) = { "STALLED", "LOCAL", "ACTIVE_CLOSE", "ACK_TO_TAP_DUE", - "ACK_FROM_TAP_DUE", + "ACK_FROM_TAP_DUE", "ACK_FROM_TAP_BLOCKS", }; /* Listening sockets, used for automatic port forwarding in pasta mode only */ @@ -401,19 +434,20 @@ static struct tcp_tap_conn *conn_at_sidx(flow_sidx_t sidx) } /** - * tcp_set_peek_offset() - Set SO_PEEK_OFF offset on a socket if supported - * @s: Socket to update + * tcp_set_peek_offset() - Set SO_PEEK_OFF offset on connection if supported + * @conn: Pointer to the TCP connection structure * @offset: Offset in bytes * * Return: -1 when it fails, 0 otherwise. */ -int tcp_set_peek_offset(int s, int offset) +int tcp_set_peek_offset(const struct tcp_tap_conn *conn, int offset) { if (!peek_offset_cap) return 0; - if (setsockopt(s, SOL_SOCKET, SO_PEEK_OFF, &offset, sizeof(offset))) { - err("Failed to set SO_PEEK_OFF to %i in socket %i", offset, s); + if (setsockopt(conn->sock, SOL_SOCKET, SO_PEEK_OFF, + &offset, sizeof(offset))) { + flow_perror(conn, "Failed to set SO_PEEK_OFF to %i", offset); return -1; } return 0; @@ -435,8 +469,12 @@ static uint32_t tcp_conn_epoll_events(uint8_t events, uint8_t conn_flags) if (events & TAP_FIN_SENT) return EPOLLET; - if (conn_flags & STALLED) - return EPOLLIN | EPOLLOUT | EPOLLRDHUP | EPOLLET; + if (conn_flags & STALLED) { + if (conn_flags & ACK_FROM_TAP_BLOCKS) + return EPOLLRDHUP | EPOLLET; + + return EPOLLIN | EPOLLRDHUP | EPOLLET; + } return EPOLLIN | EPOLLRDHUP; } @@ -463,9 +501,9 @@ static int tcp_epoll_ctl(const struct ctx *c, struct tcp_tap_conn *conn) if (conn->events == CLOSED) { if (conn->in_epoll) - epoll_ctl(c->epollfd, EPOLL_CTL_DEL, conn->sock, &ev); + epoll_del(c, conn->sock); if (conn->timer != -1) - epoll_ctl(c->epollfd, EPOLL_CTL_DEL, conn->timer, &ev); + epoll_del(c, conn->timer); return 0; } @@ -514,8 +552,7 @@ static void tcp_timer_ctl(const struct ctx *c, struct tcp_tap_conn *conn) fd = timerfd_create(CLOCK_MONOTONIC, 0); if (fd == -1 || fd > FD_REF_MAX) { - flow_dbg(conn, "failed to get timer: %s", - strerror(errno)); + flow_dbg_perror(conn, "failed to get timer"); if (fd > -1) close(fd); conn->timer = -1; @@ -524,8 +561,7 @@ static void tcp_timer_ctl(const struct ctx *c, struct tcp_tap_conn *conn) conn->timer = fd; if (epoll_ctl(c->epollfd, EPOLL_CTL_ADD, conn->timer, &ev)) { - flow_dbg(conn, "failed to add timer: %s", - strerror(errno)); + flow_dbg_perror(conn, "failed to add timer"); close(conn->timer); conn->timer = -1; return; @@ -550,7 +586,7 @@ static void tcp_timer_ctl(const struct ctx *c, struct tcp_tap_conn *conn) (unsigned long long)it.it_value.tv_nsec / 1000 / 1000); if (timerfd_settime(conn->timer, 0, &it, NULL)) - flow_err(conn, "failed to set timer: %s", strerror(errno)); + flow_perror(conn, "failed to set timer"); } /** @@ -734,124 +770,30 @@ static void tcp_get_sndbuf(struct tcp_tap_conn *conn) } /** - * tcp_sock_set_bufsize() - Set SO_RCVBUF and SO_SNDBUF to maximum values + * tcp_sock_set_nodelay() - Set TCP_NODELAY option (disable Nagle's algorithm) * @s: Socket, can be -1 to avoid check in the caller */ -static void tcp_sock_set_bufsize(const struct ctx *c, int s) +static void tcp_sock_set_nodelay(int s) { - int v = INT_MAX / 2; /* Kernel clamps and rounds, no need to check */ - if (s == -1) return; - if (!c->low_rmem && setsockopt(s, SOL_SOCKET, SO_RCVBUF, &v, sizeof(v))) - trace("TCP: failed to set SO_RCVBUF to %i", v); - - if (!c->low_wmem && setsockopt(s, SOL_SOCKET, SO_SNDBUF, &v, sizeof(v))) - trace("TCP: failed to set SO_SNDBUF to %i", v); + if (setsockopt(s, SOL_TCP, TCP_NODELAY, &((int){ 1 }), sizeof(int))) + debug("TCP: failed to set TCP_NODELAY on socket %i", s); } /** - * tcp_update_check_tcp4() - Calculate TCP checksum for IPv4 - * @iph: IPv4 header - * @iov: Pointer to the array of IO vectors - * @iov_cnt: Length of the array - * @l4offset: IPv4 payload offset in the iovec array + * tcp_update_csum() - Calculate TCP checksum + * @psum: Unfolded partial checksum of the IPv4 or IPv6 pseudo-header + * @th: TCP header (updated) + * @payload: TCP payload */ -static void tcp_update_check_tcp4(const struct iphdr *iph, - const struct iovec *iov, int iov_cnt, - size_t l4offset) +static void tcp_update_csum(uint32_t psum, struct tcphdr *th, + struct iov_tail *payload) { - uint16_t l4len = ntohs(iph->tot_len) - sizeof(struct iphdr); - struct in_addr saddr = { .s_addr = iph->saddr }; - struct in_addr daddr = { .s_addr = iph->daddr }; - size_t check_ofs; - uint16_t *check; - int check_idx; - uint32_t sum; - char *ptr; - - sum = proto_ipv4_header_psum(l4len, IPPROTO_TCP, saddr, daddr); - - check_idx = iov_skip_bytes(iov, iov_cnt, - l4offset + offsetof(struct tcphdr, check), - &check_ofs); - - if (check_idx >= iov_cnt) { - err("TCP4 buffer is too small, iov size %zd, check offset %zd", - iov_size(iov, iov_cnt), - l4offset + offsetof(struct tcphdr, check)); - return; - } - - if (check_ofs + sizeof(*check) > iov[check_idx].iov_len) { - err("TCP4 checksum field memory is not contiguous " - "check_ofs %zd check_idx %d iov_len %zd", - check_ofs, check_idx, iov[check_idx].iov_len); - return; - } - - ptr = (char *)iov[check_idx].iov_base + check_ofs; - if ((uintptr_t)ptr & (__alignof__(*check) - 1)) { - err("TCP4 checksum field is not correctly aligned in memory"); - return; - } - - check = (uint16_t *)ptr; - - *check = 0; - *check = csum_iov(iov, iov_cnt, l4offset, sum); -} - -/** - * tcp_update_check_tcp6() - Calculate TCP checksum for IPv6 - * @ip6h: IPv6 header - * @iov: Pointer to the array of IO vectors - * @iov_cnt: Length of the array - * @l4offset: IPv6 payload offset in the iovec array - */ -static void tcp_update_check_tcp6(const struct ipv6hdr *ip6h, - const struct iovec *iov, int iov_cnt, - size_t l4offset) -{ - uint16_t l4len = ntohs(ip6h->payload_len); - size_t check_ofs; - uint16_t *check; - int check_idx; - uint32_t sum; - char *ptr; - - sum = proto_ipv6_header_psum(l4len, IPPROTO_TCP, &ip6h->saddr, - &ip6h->daddr); - - check_idx = iov_skip_bytes(iov, iov_cnt, - l4offset + offsetof(struct tcphdr, check), - &check_ofs); - - if (check_idx >= iov_cnt) { - err("TCP6 buffer is too small, iov size %zd, check offset %zd", - iov_size(iov, iov_cnt), - l4offset + offsetof(struct tcphdr, check)); - return; - } - - if (check_ofs + sizeof(*check) > iov[check_idx].iov_len) { - err("TCP6 checksum field memory is not contiguous " - "check_ofs %zd check_idx %d iov_len %zd", - check_ofs, check_idx, iov[check_idx].iov_len); - return; - } - - ptr = (char *)iov[check_idx].iov_base + check_ofs; - if ((uintptr_t)ptr & (__alignof__(*check) - 1)) { - err("TCP6 checksum field is not correctly aligned in memory"); - return; - } - - check = (uint16_t *)ptr; - - *check = 0; - *check = csum_iov(iov, iov_cnt, l4offset, sum); + th->check = 0; + psum = csum_unfolded(th, sizeof(*th), psum); + th->check = csum_iov_tail(payload, psum); } /** @@ -966,138 +908,80 @@ static void tcp_fill_header(struct tcphdr *th, } /** - * tcp_fill_headers4() - Fill 802.3, IPv4, TCP headers in pre-cooked buffers + * tcp_fill_headers() - Fill 802.3, IP, TCP headers * @conn: Connection pointer * @taph: tap backend specific header - * @iph: Pointer to IPv4 header - * @bp: Pointer to TCP header followed by TCP payload - * @dlen: TCP payload length - * @check: Checksum, if already known + * @ip4h: Pointer to IPv4 header, or NULL + * @ip6h: Pointer to IPv6 header, or NULL + * @th: Pointer to TCP header + * @payload: TCP payload + * @ip4_check: IPv4 checksum, if already known * @seq: Sequence number for this segment * @no_tcp_csum: Do not set TCP checksum - * - * Return: The IPv4 payload length, host order */ -static size_t tcp_fill_headers4(const struct tcp_tap_conn *conn, - struct tap_hdr *taph, - struct iphdr *iph, struct tcp_payload_t *bp, - size_t dlen, const uint16_t *check, - uint32_t seq, bool no_tcp_csum) +void tcp_fill_headers(const struct tcp_tap_conn *conn, + struct tap_hdr *taph, + struct iphdr *ip4h, struct ipv6hdr *ip6h, + struct tcphdr *th, struct iov_tail *payload, + const uint16_t *ip4_check, uint32_t seq, bool no_tcp_csum) { const struct flowside *tapside = TAPFLOW(conn); - const struct in_addr *src4 = inany_v4(&tapside->oaddr); - const struct in_addr *dst4 = inany_v4(&tapside->eaddr); - size_t l4len = dlen + sizeof(bp->th); - size_t l3len = l4len + sizeof(*iph); + size_t l4len = iov_tail_size(payload) + sizeof(*th); + size_t l3len = l4len; + uint32_t psum = 0; - ASSERT(src4 && dst4); + if (ip4h) { + const struct in_addr *src4 = inany_v4(&tapside->oaddr); + const struct in_addr *dst4 = inany_v4(&tapside->eaddr); - iph->tot_len = htons(l3len); - iph->saddr = src4->s_addr; - iph->daddr = dst4->s_addr; + ASSERT(src4 && dst4); - iph->check = check ? *check : - csum_ip4_header(l3len, IPPROTO_TCP, *src4, *dst4); + l3len += + sizeof(*ip4h); - tcp_fill_header(&bp->th, conn, seq); + ip4h->tot_len = htons(l3len); + ip4h->saddr = src4->s_addr; + ip4h->daddr = dst4->s_addr; - if (no_tcp_csum) { - bp->th.check = 0; - } else { - const struct iovec iov = { - .iov_base = bp, - .iov_len = ntohs(iph->tot_len) - sizeof(struct iphdr), - }; + if (ip4_check) + ip4h->check = *ip4_check; + else + ip4h->check = csum_ip4_header(l3len, IPPROTO_TCP, + *src4, *dst4); - tcp_update_check_tcp4(iph, &iov, 1, 0); + if (!no_tcp_csum) { + psum = proto_ipv4_header_psum(l4len, IPPROTO_TCP, + *src4, *dst4); + } } + if (ip6h) { + l3len += sizeof(*ip6h); + + ip6h->payload_len = htons(l4len); + ip6h->saddr = tapside->oaddr.a6; + ip6h->daddr = tapside->eaddr.a6; + + ip6h->hop_limit = 255; + ip6h->version = 6; + ip6h->nexthdr = IPPROTO_TCP; + + ip6_set_flow_lbl(ip6h, conn->sock); + + if (!no_tcp_csum) { + psum = proto_ipv6_header_psum(l4len, IPPROTO_TCP, + &ip6h->saddr, + &ip6h->daddr); + } + } + + tcp_fill_header(th, conn, seq); + + if (no_tcp_csum) + th->check = 0; + else + tcp_update_csum(psum, th, payload); + tap_hdr_update(taph, l3len + sizeof(struct ethhdr)); - - return l4len; -} - -/** - * tcp_fill_headers6() - Fill 802.3, IPv6, TCP headers in pre-cooked buffers - * @conn: Connection pointer - * @taph: tap backend specific header - * @ip6h: Pointer to IPv6 header - * @bp: Pointer to TCP header followed by TCP payload - * @dlen: TCP payload length - * @check: Checksum, if already known - * @seq: Sequence number for this segment - * @no_tcp_csum: Do not set TCP checksum - * - * Return: The IPv6 payload length, host order - */ -static size_t tcp_fill_headers6(const struct tcp_tap_conn *conn, - struct tap_hdr *taph, - struct ipv6hdr *ip6h, struct tcp_payload_t *bp, - size_t dlen, uint32_t seq, bool no_tcp_csum) -{ - const struct flowside *tapside = TAPFLOW(conn); - size_t l4len = dlen + sizeof(bp->th); - - ip6h->payload_len = htons(l4len); - ip6h->saddr = tapside->oaddr.a6; - ip6h->daddr = tapside->eaddr.a6; - - ip6h->hop_limit = 255; - ip6h->version = 6; - ip6h->nexthdr = IPPROTO_TCP; - - ip6h->flow_lbl[0] = (conn->sock >> 16) & 0xf; - ip6h->flow_lbl[1] = (conn->sock >> 8) & 0xff; - ip6h->flow_lbl[2] = (conn->sock >> 0) & 0xff; - - tcp_fill_header(&bp->th, conn, seq); - - if (no_tcp_csum) { - bp->th.check = 0; - } else { - const struct iovec iov = { - .iov_base = bp, - .iov_len = ntohs(ip6h->payload_len) - }; - - tcp_update_check_tcp6(ip6h, &iov, 1, 0); - } - - tap_hdr_update(taph, l4len + sizeof(*ip6h) + sizeof(struct ethhdr)); - - return l4len; -} - -/** - * tcp_l2_buf_fill_headers() - Fill 802.3, IP, TCP headers in pre-cooked buffers - * @conn: Connection pointer - * @iov: Pointer to an array of iovec of TCP pre-cooked buffers - * @dlen: TCP payload length - * @check: Checksum, if already known - * @seq: Sequence number for this segment - * @no_tcp_csum: Do not set TCP checksum - * - * Return: IP payload length, host order - */ -size_t tcp_l2_buf_fill_headers(const struct tcp_tap_conn *conn, - struct iovec *iov, size_t dlen, - const uint16_t *check, uint32_t seq, - bool no_tcp_csum) -{ - const struct flowside *tapside = TAPFLOW(conn); - const struct in_addr *a4 = inany_v4(&tapside->oaddr); - - if (a4) { - return tcp_fill_headers4(conn, iov[TCP_IOV_TAP].iov_base, - iov[TCP_IOV_IP].iov_base, - iov[TCP_IOV_PAYLOAD].iov_base, dlen, - check, seq, no_tcp_csum); - } - - return tcp_fill_headers6(conn, iov[TCP_IOV_TAP].iov_base, - iov[TCP_IOV_IP].iov_base, - iov[TCP_IOV_PAYLOAD].iov_base, dlen, - seq, no_tcp_csum); } /** @@ -1255,7 +1139,7 @@ int tcp_prepare_flags(const struct ctx *c, struct tcp_tap_conn *conn, if (flags & SYN) { int mss; - if (c->mtu == -1) { + if (!c->mtu) { mss = tinfo.tcpi_snd_mss; } else { mss = c->mtu - sizeof(struct tcphdr); @@ -1275,7 +1159,7 @@ int tcp_prepare_flags(const struct ctx *c, struct tcp_tap_conn *conn, *opts = TCP_SYN_OPTS(mss, conn->ws_to_tap); *optlen = sizeof(*opts); - } else if (!(flags & RST)) { + } else { flags |= ACK; } @@ -1314,6 +1198,9 @@ int tcp_prepare_flags(const struct ctx *c, struct tcp_tap_conn *conn, static int tcp_send_flag(const struct ctx *c, struct tcp_tap_conn *conn, int flags) { + if (c->mode == MODE_VU) + return tcp_vu_send_flag(c, conn, flags); + return tcp_buf_send_flag(c, conn, flags); } @@ -1327,8 +1214,8 @@ void tcp_rst_do(const struct ctx *c, struct tcp_tap_conn *conn) if (conn->events == CLOSED) return; - if (!tcp_send_flag(c, conn, RST)) - conn_event(c, conn, CLOSED); + tcp_send_flag(c, conn, RST); + conn_event(c, conn, CLOSED); } /** @@ -1403,12 +1290,11 @@ int tcp_conn_pool_sock(int pool[]) /** * tcp_conn_new_sock() - Open and prepare new socket for connection - * @c: Execution context * @af: Address family * * Return: socket number on success, negative code if socket creation failed */ -static int tcp_conn_new_sock(const struct ctx *c, sa_family_t af) +static int tcp_conn_new_sock(sa_family_t af) { int s; @@ -1422,19 +1308,18 @@ static int tcp_conn_new_sock(const struct ctx *c, sa_family_t af) if (s < 0) return -errno; - tcp_sock_set_bufsize(c, s); + tcp_sock_set_nodelay(s); return s; } /** * tcp_conn_sock() - Obtain a connectable socket in the host/init namespace - * @c: Execution context * @af: Address family (AF_INET or AF_INET6) * * Return: Socket fd on success, -errno on failure */ -int tcp_conn_sock(const struct ctx *c, sa_family_t af) +int tcp_conn_sock(sa_family_t af) { int *pool = af == AF_INET6 ? init_sock_pool6 : init_sock_pool4; int s; @@ -1445,11 +1330,11 @@ int tcp_conn_sock(const struct ctx *c, sa_family_t af) /* If the pool is empty we just open a new one without refilling the * pool to keep latency down. */ - if ((s = tcp_conn_new_sock(c, af)) >= 0) + if ((s = tcp_conn_new_sock(af)) >= 0) return s; err("TCP: Unable to open socket for new connection: %s", - strerror(-s)); + strerror_(-s)); return -1; } @@ -1499,10 +1384,10 @@ static void tcp_bind_outbound(const struct ctx *c, if (bind(s, &bind_sa.sa, sl)) { char sstr[INANY_ADDRSTRLEN]; - flow_dbg(conn, - "Can't bind TCP outbound socket to %s:%hu: %s", - inany_ntop(&tgt->oaddr, sstr, sizeof(sstr)), - tgt->oport, strerror(errno)); + flow_dbg_perror(conn, + "Can't bind TCP outbound socket to %s:%hu", + inany_ntop(&tgt->oaddr, sstr, sizeof(sstr)), + tgt->oport); } } @@ -1511,9 +1396,9 @@ static void tcp_bind_outbound(const struct ctx *c, if (setsockopt(s, SOL_SOCKET, SO_BINDTODEVICE, c->ip4.ifname_out, strlen(c->ip4.ifname_out))) { - flow_dbg(conn, "Can't bind IPv4 TCP socket to" - " interface %s: %s", c->ip4.ifname_out, - strerror(errno)); + flow_dbg_perror(conn, + "Can't bind IPv4 TCP socket to interface %s", + c->ip4.ifname_out); } } } else if (bind_sa.sa_family == AF_INET6) { @@ -1521,9 +1406,9 @@ static void tcp_bind_outbound(const struct ctx *c, if (setsockopt(s, SOL_SOCKET, SO_BINDTODEVICE, c->ip6.ifname_out, strlen(c->ip6.ifname_out))) { - flow_dbg(conn, "Can't bind IPv6 TCP socket to" - " interface %s: %s", c->ip6.ifname_out, - strerror(errno)); + flow_dbg_perror(conn, + "Can't bind IPv6 TCP socket to interface %s", + c->ip6.ifname_out); } } } @@ -1539,6 +1424,8 @@ static void tcp_bind_outbound(const struct ctx *c, * @opts: Pointer to start of options * @optlen: Bytes in options: caller MUST ensure available length * @now: Current timestamp + * + * #syscalls:vu getsockname */ static void tcp_conn_from_tap(const struct ctx *c, sa_family_t af, const void *saddr, const void *daddr, @@ -1547,9 +1434,10 @@ static void tcp_conn_from_tap(const struct ctx *c, sa_family_t af, { in_port_t srcport = ntohs(th->source); in_port_t dstport = ntohs(th->dest); - const struct flowside *ini, *tgt; + const struct flowside *ini; struct tcp_tap_conn *conn; union sockaddr_inany sa; + struct flowside *tgt; union flow *flow; int s = -1, mss; uint64_t hash; @@ -1583,7 +1471,7 @@ static void tcp_conn_from_tap(const struct ctx *c, sa_family_t af, goto cancel; } - if ((s = tcp_conn_sock(c, af)) < 0) + if ((s = tcp_conn_sock(af)) < 0) goto cancel; pif_sockaddr(c, &sa, &sl, PIF_HOST, &tgt->eaddr, tgt->eport); @@ -1604,12 +1492,13 @@ static void tcp_conn_from_tap(const struct ctx *c, sa_family_t af, } else { /* Not a local, bound destination, inconclusive test */ close(s); - if ((s = tcp_conn_sock(c, af)) < 0) + if ((s = tcp_conn_sock(af)) < 0) goto cancel; } conn->sock = s; conn->timer = -1; + conn->listening_sock = -1; conn_event(c, conn, TAP_SYN_RCVD); conn->wnd_to_tap = WINDOW_DEFAULT; @@ -1654,6 +1543,14 @@ static void tcp_conn_from_tap(const struct ctx *c, sa_family_t af, } tcp_epoll_ctl(c, conn); + + if (c->mode == MODE_VU) { /* To rebind to same oport after migration */ + sl = sizeof(sa); + if (getsockname(s, &sa.sa, &sl) || + inany_from_sockaddr(&tgt->oaddr, &tgt->oport, &sa) < 0) + err_perror("Can't get local address for socket %i", s); + } + FLOW_ACTIVATE(conn); return; @@ -1707,6 +1604,9 @@ static int tcp_sock_consume(const struct tcp_tap_conn *conn, uint32_t ack_seq) */ static int tcp_data_from_sock(const struct ctx *c, struct tcp_tap_conn *conn) { + if (c->mode == MODE_VU) + return tcp_vu_data_from_sock(c, conn); + return tcp_buf_data_from_sock(c, conn); } @@ -1771,8 +1671,10 @@ static int tcp_data_from_tap(const struct ctx *c, struct tcp_tap_conn *conn, tcp_send_flag(c, conn, ACK); tcp_timer_ctl(c, conn); - if (p->count == 1) + if (p->count == 1) { + tcp_tap_window_update(conn, ntohs(th->window)); return 1; + } continue; } @@ -1855,7 +1757,7 @@ static int tcp_data_from_tap(const struct ctx *c, struct tcp_tap_conn *conn, "fast re-transmit, ACK: %u, previous sequence: %u", max_ack_seq, conn->seq_to_tap); conn->seq_to_tap = max_ack_seq; - if (tcp_set_peek_offset(conn->sock, 0)) { + if (tcp_set_peek_offset(conn, 0)) { tcp_rst(c, conn); return -1; } @@ -1952,7 +1854,7 @@ static void tcp_conn_from_sock_finish(const struct ctx *c, conn->seq_ack_to_tap = conn->seq_from_tap; conn_event(c, conn, ESTABLISHED); - if (tcp_set_peek_offset(conn->sock, 0)) { + if (tcp_set_peek_offset(conn, 0)) { tcp_rst(c, conn); return; } @@ -1965,6 +1867,75 @@ static void tcp_conn_from_sock_finish(const struct ctx *c, tcp_data_from_sock(c, conn); } +/** + * tcp_rst_no_conn() - Send RST in response to a packet with no connection + * @c: Execution context + * @af: Address family, AF_INET or AF_INET6 + * @saddr: Source address of the packet we're responding to + * @daddr: Destination address of the packet we're responding to + * @flow_lbl: IPv6 flow label (ignored for IPv4) + * @th: TCP header of the packet we're responding to + * @l4len: Packet length, including TCP header + */ +static void tcp_rst_no_conn(const struct ctx *c, int af, + const void *saddr, const void *daddr, + uint32_t flow_lbl, + const struct tcphdr *th, size_t l4len) +{ + struct iov_tail payload = IOV_TAIL(NULL, 0, 0); + struct tcphdr *rsth; + char buf[USHRT_MAX]; + uint32_t psum = 0; + size_t rst_l2len; + + /* Don't respond to RSTs without a connection */ + if (th->rst) + return; + + if (af == AF_INET) { + struct iphdr *ip4h = tap_push_l2h(c, buf, ETH_P_IP); + const struct in_addr *rst_src = daddr; + const struct in_addr *rst_dst = saddr; + + rsth = tap_push_ip4h(ip4h, *rst_src, *rst_dst, + sizeof(*rsth), IPPROTO_TCP); + psum = proto_ipv4_header_psum(sizeof(*rsth), IPPROTO_TCP, + *rst_src, *rst_dst); + + } else { + struct ipv6hdr *ip6h = tap_push_l2h(c, buf, ETH_P_IPV6); + const struct in6_addr *rst_src = daddr; + const struct in6_addr *rst_dst = saddr; + + rsth = tap_push_ip6h(ip6h, rst_src, rst_dst, + sizeof(*rsth), IPPROTO_TCP, flow_lbl); + psum = proto_ipv6_header_psum(sizeof(*rsth), IPPROTO_TCP, + rst_src, rst_dst); + } + + memset(rsth, 0, sizeof(*rsth)); + + rsth->source = th->dest; + rsth->dest = th->source; + rsth->rst = 1; + rsth->doff = sizeof(*rsth) / 4UL; + + /* Sequence matching logic from RFC 9293 section 3.10.7.1 */ + if (th->ack) { + rsth->seq = th->ack_seq; + } else { + size_t dlen = l4len - th->doff * 4UL; + uint32_t ack = ntohl(th->seq) + dlen; + + rsth->ack_seq = htonl(ack); + rsth->ack = 1; + } + + tcp_update_csum(psum, rsth, &payload); + rst_l2len = ((char *)rsth - buf) + sizeof(*rsth); + tap_send_single(c, buf, rst_l2len); +} + /** * tcp_tap_handler() - Handle packets from tap and state transitions * @c: Execution context @@ -1972,6 +1943,7 @@ static void tcp_conn_from_sock_finish(const struct ctx *c, * @af: Address family, AF_INET or AF_INET6 * @saddr: Source address * @daddr: Destination address + * @flow_lbl: IPv6 flow label (ignored for IPv4) * @p: Pool of TCP packets, with TCP headers * @idx: Index of first packet in pool to process * @now: Current timestamp @@ -1979,7 +1951,7 @@ static void tcp_conn_from_sock_finish(const struct ctx *c, * Return: count of consumed packets */ int tcp_tap_handler(const struct ctx *c, uint8_t pif, sa_family_t af, - const void *saddr, const void *daddr, + const void *saddr, const void *daddr, uint32_t flow_lbl, const struct pool *p, int idx, const struct timespec *now) { struct tcp_tap_conn *conn; @@ -2012,6 +1984,8 @@ int tcp_tap_handler(const struct ctx *c, uint8_t pif, sa_family_t af, if (opts && th->syn && !th->ack) tcp_conn_from_tap(c, af, saddr, daddr, th, opts, optlen, now); + else + tcp_rst_no_conn(c, af, saddr, daddr, flow_lbl, th, len); return 1; } @@ -2041,11 +2015,14 @@ int tcp_tap_handler(const struct ctx *c, uint8_t pif, sa_family_t af, /* Establishing connection from tap */ if (conn->events & TAP_SYN_RCVD) { + if (th->syn && !th->ack && !th->fin) + return 1; /* SYN retry: ignore and keep waiting */ + if (!(conn->events & TAP_SYN_ACK_SENT)) goto reset; conn_event(c, conn, ESTABLISHED); - if (tcp_set_peek_offset(conn->sock, 0)) + if (tcp_set_peek_offset(conn, 0)) goto reset; if (th->fin) { @@ -2071,7 +2048,10 @@ int tcp_tap_handler(const struct ctx *c, uint8_t pif, sa_family_t af, /* Established connections not accepting data from tap */ if (conn->events & TAP_FIN_RCVD) { + tcp_sock_consume(conn, ntohl(th->ack_seq)); tcp_update_seqack_from_tap(c, conn, ntohl(th->ack_seq)); + tcp_tap_window_update(conn, ntohs(th->window)); + tcp_data_from_sock(c, conn); if (conn->events & SOCK_FIN_RCVD && conn->seq_ack_from_tap == conn->seq_to_tap) @@ -2091,10 +2071,27 @@ int tcp_tap_handler(const struct ctx *c, uint8_t pif, sa_family_t af, ack_due = 1; if ((conn->events & TAP_FIN_RCVD) && !(conn->events & SOCK_FIN_SENT)) { + socklen_t sl; + struct tcp_info tinfo; + shutdown(conn->sock, SHUT_WR); conn_event(c, conn, SOCK_FIN_SENT); tcp_send_flag(c, conn, ACK); ack_due = 0; + + /* If we received a FIN, but the socket is in TCP_ESTABLISHED + * state, it must be a migrated socket. The kernel saw the FIN + * on the source socket, but not on the target socket. + * + * Approximate the effect of that FIN: as we're sending a FIN + * out ourselves, the socket is now in a state equivalent to + * LAST_ACK. Now that we sent the FIN out, close it with a RST. + */ + sl = sizeof(tinfo); + getsockopt(conn->sock, SOL_TCP, TCP_INFO, &tinfo, &sl); + if (tinfo.tcpi_state == TCP_ESTABLISHED && + conn->events & SOCK_FIN_RCVD) + goto reset; } if (ack_due) @@ -2177,9 +2174,10 @@ static void tcp_tap_conn_from_sock(const struct ctx *c, union flow *flow, void tcp_listen_handler(const struct ctx *c, union epoll_ref ref, const struct timespec *now) { - const struct flowside *ini; + struct tcp_tap_conn *conn; union sockaddr_inany sa; socklen_t sl = sizeof(sa); + struct flowside *ini; union flow *flow; int s; @@ -2192,11 +2190,23 @@ void tcp_listen_handler(const struct ctx *c, union epoll_ref ref, if (s < 0) goto cancel; - /* FIXME: When listening port has a specific bound address, record that - * as our address + conn = (struct tcp_tap_conn *)flow; + conn->listening_sock = ref.fd; + + tcp_sock_set_nodelay(s); + + /* FIXME: If useful: when the listening port has a specific bound + * address, record that as our address, as implemented for vhost-user + * mode only, below. */ ini = flow_initiate_sa(flow, ref.tcp_listen.pif, &sa, - ref.tcp_listen.port); + NULL, ref.tcp_listen.port); + + if (c->mode == MODE_VU) { /* Rebind to same address after migration */ + if (getsockname(s, &sa.sa, &sl) || + inany_from_sockaddr(&ini->oaddr, &ini->oport, &sa) < 0) + err_perror("Can't get local address for socket %i", s); + } if (!inany_is_unicast(&ini->eaddr) || ini->eport == 0) { char sastr[SOCKADDR_STRLEN]; @@ -2252,7 +2262,7 @@ void tcp_timer_handler(const struct ctx *c, union epoll_ref ref) * and we just set the timer to a new point in the future: discard it. */ if (timerfd_gettime(conn->timer, &check_armed)) - flow_err(conn, "failed to read timer: %s", strerror(errno)); + flow_perror(conn, "failed to read timer"); if (check_armed.it_value.tv_sec || check_armed.it_value.tv_nsec) return; @@ -2274,7 +2284,9 @@ void tcp_timer_handler(const struct ctx *c, union epoll_ref ref) flow_dbg(conn, "ACK timeout, retry"); conn->retrans++; conn->seq_to_tap = conn->seq_ack_from_tap; - if (tcp_set_peek_offset(conn->sock, 0)) { + if (!conn->wnd_from_tap) + conn->wnd_from_tap = 1; /* Zero-window probe */ + if (tcp_set_peek_offset(conn, 0)) { tcp_rst(c, conn); } else { tcp_data_from_sock(c, conn); @@ -2292,8 +2304,7 @@ void tcp_timer_handler(const struct ctx *c, union epoll_ref ref) * ~ACK_TO_TAP_DUE or ~ACK_FROM_TAP_DUE. */ if (timerfd_settime(conn->timer, 0, &new, &old)) - flow_err(conn, "failed to set timer: %s", - strerror(errno)); + flow_perror(conn, "failed to set timer"); if (old.it_value.tv_sec == ACT_TIMEOUT) { flow_dbg(conn, "activity timeout"); @@ -2339,8 +2350,10 @@ void tcp_sock_handler(const struct ctx *c, union epoll_ref ref, if (events & EPOLLIN) tcp_data_from_sock(c, conn); - if (events & EPOLLOUT) - tcp_update_seqack_wnd(c, conn, false, NULL); + if (events & EPOLLOUT) { + if (tcp_update_seqack_wnd(c, conn, false, NULL)) + tcp_send_flag(c, conn, ACK); + } return; } @@ -2393,7 +2406,6 @@ static int tcp_sock_init_one(const struct ctx *c, const union inany_addr *addr, if (s < 0) return s; - tcp_sock_set_bufsize(c, s); return s; } @@ -2450,9 +2462,7 @@ static void tcp_ns_sock_init4(const struct ctx *c, in_port_t port) s = pif_sock_l4(c, EPOLL_TYPE_TCP_LISTEN, PIF_SPLICE, &inany_loopback4, NULL, port, tref.u32); - if (s >= 0) - tcp_sock_set_bufsize(c, s); - else + if (s < 0) s = -1; if (c->tcp.fwd_out.mode == FWD_AUTO) @@ -2476,9 +2486,7 @@ static void tcp_ns_sock_init6(const struct ctx *c, in_port_t port) s = pif_sock_l4(c, EPOLL_TYPE_TCP_LISTEN, PIF_SPLICE, &inany_loopback6, NULL, port, tref.u32); - if (s >= 0) - tcp_sock_set_bufsize(c, s); - else + if (s < 0) s = -1; if (c->tcp.fwd_out.mode == FWD_AUTO) @@ -2490,7 +2498,7 @@ static void tcp_ns_sock_init6(const struct ctx *c, in_port_t port) * @c: Execution context * @port: Port, host order */ -void tcp_ns_sock_init(const struct ctx *c, in_port_t port) +static void tcp_ns_sock_init(const struct ctx *c, in_port_t port) { ASSERT(!c->no_tcp); @@ -2526,13 +2534,12 @@ static int tcp_ns_socks_init(void *arg) /** * tcp_sock_refill_pool() - Refill one pool of pre-opened sockets - * @c: Execution context * @pool: Pool of sockets to refill * @af: Address family to use * * Return: 0 on success, negative error code if there was at least one error */ -int tcp_sock_refill_pool(const struct ctx *c, int pool[], sa_family_t af) +int tcp_sock_refill_pool(int pool[], sa_family_t af) { int i; @@ -2542,7 +2549,7 @@ int tcp_sock_refill_pool(const struct ctx *c, int pool[], sa_family_t af) if (pool[i] >= 0) continue; - if ((fd = tcp_conn_new_sock(c, af)) < 0) + if ((fd = tcp_conn_new_sock(af)) < 0) return fd; pool[i] = fd; @@ -2558,16 +2565,16 @@ int tcp_sock_refill_pool(const struct ctx *c, int pool[], sa_family_t af) static void tcp_sock_refill_init(const struct ctx *c) { if (c->ifi4) { - int rc = tcp_sock_refill_pool(c, init_sock_pool4, AF_INET); + int rc = tcp_sock_refill_pool(init_sock_pool4, AF_INET); if (rc < 0) warn("TCP: Error refilling IPv4 host socket pool: %s", - strerror(-rc)); + strerror_(-rc)); } if (c->ifi6) { - int rc = tcp_sock_refill_pool(c, init_sock_pool6, AF_INET6); + int rc = tcp_sock_refill_pool(init_sock_pool6, AF_INET6); if (rc < 0) warn("TCP: Error refilling IPv6 host socket pool: %s", - strerror(-rc)); + strerror_(-rc)); } } @@ -2749,3 +2756,980 @@ void tcp_timer(struct ctx *c, const struct timespec *now) if (c->mode == MODE_PASTA) tcp_splice_refill(c); } + +/** + * tcp_flow_is_established() - Was the connection established? Includes closing + * @conn: Pointer to the TCP connection structure + * + * Return: true if the connection was established, false otherwise + */ +bool tcp_flow_is_established(const struct tcp_tap_conn *conn) +{ + return conn->events & ESTABLISHED; +} + +/** + * tcp_flow_repair_on() - Enable repair mode for a single TCP flow + * @c: Execution context + * @conn: Pointer to the TCP connection structure + * + * Return: 0 on success, negative error code on failure + */ +int tcp_flow_repair_on(struct ctx *c, const struct tcp_tap_conn *conn) +{ + int rc = 0; + + if (conn->sock < 0) + return 0; + + if ((rc = repair_set(c, conn->sock, TCP_REPAIR_ON))) + err("Failed to set TCP_REPAIR"); + + return rc; +} + +/** + * tcp_flow_repair_off() - Clear repair mode for a single TCP flow + * @c: Execution context + * @conn: Pointer to the TCP connection structure + * + * Return: 0 on success, negative error code on failure + */ +int tcp_flow_repair_off(struct ctx *c, const struct tcp_tap_conn *conn) +{ + int rc = 0; + + if (conn->sock < 0) + return 0; + + if ((rc = repair_set(c, conn->sock, TCP_REPAIR_OFF))) + err("Failed to clear TCP_REPAIR"); + + return rc; +} + +/** + * tcp_flow_dump_tinfo() - Dump window scale, tcpi_state, tcpi_options + * @conn: Pointer to the TCP connection structure + * @t: Extended migration data + * + * Return: 0 on success, negative error code on failure + */ +static int tcp_flow_dump_tinfo(const struct tcp_tap_conn *conn, + struct tcp_tap_transfer_ext *t) +{ + struct tcp_info tinfo; + socklen_t sl; + + sl = sizeof(tinfo); + if (getsockopt(conn->sock, SOL_TCP, TCP_INFO, &tinfo, &sl)) { + int rc = -errno; + flow_perror(conn, "Querying TCP_INFO"); + return rc; + } + + t->snd_ws = tinfo.tcpi_snd_wscale; + t->rcv_ws = tinfo.tcpi_rcv_wscale; + t->tcpi_state = tinfo.tcpi_state; + t->tcpi_options = tinfo.tcpi_options; + + return 0; +} + +/** + * tcp_flow_dump_mss() - Dump MSS clamp (not current MSS) via TCP_MAXSEG + * @conn: Pointer to the TCP connection structure + * @t: Extended migration data + * + * Return: 0 on success, negative error code on failure + */ +static int tcp_flow_dump_mss(const struct tcp_tap_conn *conn, + struct tcp_tap_transfer_ext *t) +{ + socklen_t sl = sizeof(t->mss); + int val; + + if (getsockopt(conn->sock, SOL_TCP, TCP_MAXSEG, &val, &sl)) { + int rc = -errno; + flow_perror(conn, "Getting MSS"); + return rc; + } + + t->mss = (uint32_t)val; + + return 0; +} + + +/** + * tcp_flow_dump_timestamp() - Dump RFC 7323 timestamp via TCP_TIMESTAMP + * @conn: Pointer to the TCP connection structure + * @t: Extended migration data (tcpi_options must be populated) + * + * Return: 0 on success, negative error code on failure + */ +static int tcp_flow_dump_timestamp(const struct tcp_tap_conn *conn, + struct tcp_tap_transfer_ext *t) +{ + int val = 0; + + if (t->tcpi_options & TCPI_OPT_TIMESTAMPS) { + socklen_t sl = sizeof(val); + + if (getsockopt(conn->sock, SOL_TCP, TCP_TIMESTAMP, &val, &sl)) { + int rc = -errno; + flow_perror(conn, "Getting RFC 7323 timestamp"); + return rc; + } + } + + t->timestamp = (uint32_t)val; + return 0; +} + +/** + * tcp_flow_repair_timestamp() - Restore RFC 7323 timestamp via TCP_TIMESTAMP + * @conn: Pointer to the TCP connection structure + * @t: Extended migration data + * + * Return: 0 on success, negative error code on failure + */ +static int tcp_flow_repair_timestamp(const struct tcp_tap_conn *conn, + const struct tcp_tap_transfer_ext *t) +{ + int val = (int)t->timestamp; + + if (t->tcpi_options & TCPI_OPT_TIMESTAMPS) { + if (setsockopt(conn->sock, SOL_TCP, TCP_TIMESTAMP, + &val, sizeof(val))) { + int rc = -errno; + flow_perror(conn, "Setting RFC 7323 timestamp"); + return rc; + } + } + + return 0; +} + +/** + * tcp_flow_dump_wnd() - Dump current tcp_repair_window parameters + * @conn: Pointer to the TCP connection structure + * @t: Extended migration data + * + * Return: 0 on success, negative error code on failure + */ +static int tcp_flow_dump_wnd(const struct tcp_tap_conn *conn, + struct tcp_tap_transfer_ext *t) +{ + struct tcp_repair_window wnd; + socklen_t sl = sizeof(wnd); + + if (getsockopt(conn->sock, IPPROTO_TCP, TCP_REPAIR_WINDOW, &wnd, &sl)) { + int rc = -errno; + flow_perror(conn, "Getting window repair data"); + return rc; + } + + t->snd_wl1 = wnd.snd_wl1; + t->snd_wnd = wnd.snd_wnd; + t->max_window = wnd.max_window; + t->rcv_wnd = wnd.rcv_wnd; + t->rcv_wup = wnd.rcv_wup; + + /* If we received a FIN, we also need to adjust window parameters. + * + * This must be called after tcp_flow_dump_tinfo(), for t->tcpi_state. + */ + if (t->tcpi_state == TCP_CLOSE_WAIT || t->tcpi_state == TCP_LAST_ACK) { + t->rcv_wup--; + t->rcv_wnd++; + } + + return 0; +} + +/** + * tcp_flow_repair_wnd() - Restore window parameters from extended data + * @conn: Pointer to the TCP connection structure + * @t: Extended migration data + * + * Return: 0 on success, negative error code on failure + */ +static int tcp_flow_repair_wnd(const struct tcp_tap_conn *conn, + const struct tcp_tap_transfer_ext *t) +{ + struct tcp_repair_window wnd; + + wnd.snd_wl1 = t->snd_wl1; + wnd.snd_wnd = t->snd_wnd; + wnd.max_window = t->max_window; + wnd.rcv_wnd = t->rcv_wnd; + wnd.rcv_wup = t->rcv_wup; + + if (setsockopt(conn->sock, IPPROTO_TCP, TCP_REPAIR_WINDOW, + &wnd, sizeof(wnd))) { + int rc = -errno; + flow_perror(conn, "Setting window data"); + return rc; + } + + return 0; +} + +/** + * tcp_flow_select_queue() - Select queue (receive or send) for next operation + * @conn: Connection to select queue for + * @queue: TCP_RECV_QUEUE or TCP_SEND_QUEUE + * + * Return: 0 on success, negative error code on failure + */ +static int tcp_flow_select_queue(const struct tcp_tap_conn *conn, int queue) +{ + if (setsockopt(conn->sock, SOL_TCP, TCP_REPAIR_QUEUE, + &queue, sizeof(queue))) { + int rc = -errno; + flow_perror(conn, "Selecting TCP_SEND_QUEUE"); + return rc; + } + + return 0; +} + +/** + * tcp_flow_dump_sndqueue() - Dump send queue, length of sent and not sent data + * @conn: Connection to dump queue for + * @t: Extended migration data + * + * Return: 0 on success, negative error code on failure + * + * #syscalls:vu ioctl + */ +static int tcp_flow_dump_sndqueue(const struct tcp_tap_conn *conn, + struct tcp_tap_transfer_ext *t) +{ + int s = conn->sock; + ssize_t rc; + + if (ioctl(s, SIOCOUTQ, &t->sndq) < 0) { + rc = -errno; + flow_perror(conn, "Getting send queue size"); + return rc; + } + + if (ioctl(s, SIOCOUTQNSD, &t->notsent) < 0) { + rc = -errno; + flow_perror(conn, "Getting not sent count"); + return rc; + } + + /* If we sent a FIN, SIOCOUTQ and SIOCOUTQNSD are one greater than the + * actual pending queue length, because they are based on the sequence + * numbers, not directly on the buffer contents. + * + * This must be called after tcp_flow_dump_tinfo(), for t->tcpi_state. + */ + if (t->tcpi_state == TCP_FIN_WAIT1 || t->tcpi_state == TCP_FIN_WAIT2 || + t->tcpi_state == TCP_LAST_ACK || t->tcpi_state == TCP_CLOSING) { + if (t->sndq) + t->sndq--; + if (t->notsent) + t->notsent--; + } + + if (t->notsent > t->sndq) { + flow_err(conn, + "Invalid notsent count socket %i, send: %u, not sent: %u", + s, t->sndq, t->notsent); + return -EINVAL; + } + + if (t->sndq > TCP_MIGRATE_SND_QUEUE_MAX) { + flow_err(conn, + "Send queue too large to migrate socket %i: %u bytes", + s, t->sndq); + return -ENOBUFS; + } + + rc = recv(s, tcp_migrate_snd_queue, + MIN(t->sndq, TCP_MIGRATE_SND_QUEUE_MAX), MSG_PEEK); + if (rc < 0) { + if (errno == EAGAIN) { /* EAGAIN means empty */ + rc = 0; + } else { + rc = -errno; + flow_perror(conn, "Can't read send queue"); + return rc; + } + } + + if ((uint32_t)rc < t->sndq) { + flow_err(conn, "Short read migrating send queue"); + return -ENXIO; + } + + t->notsent = MIN(t->notsent, t->sndq); + + return 0; +} + +/** + * tcp_flow_repair_queue() - Restore contents of a given (pre-selected) queue + * @conn: Connection to repair queue for + * @len: Length of data to be restored + * @buf: Buffer with content of pending data queue + * + * Return: 0 on success, negative error code on failure + */ +static int tcp_flow_repair_queue(const struct tcp_tap_conn *conn, + size_t len, uint8_t *buf) +{ + size_t chunk = len; + uint8_t *p = buf; + + while (len > 0) { + ssize_t rc = send(conn->sock, p, MIN(len, chunk), 0); + + if (rc < 0) { + if ((errno == ENOBUFS || errno == ENOMEM) && + chunk >= TCP_MIGRATE_RESTORE_CHUNK_MIN) { + chunk /= 2; + continue; + } + + rc = -errno; + flow_perror(conn, "Can't write queue"); + return rc; + } + + len -= rc; + p += rc; + } + + return 0; +} + +/** + * tcp_flow_dump_seq() - Dump current sequence of pre-selected queue + * @conn: Pointer to the TCP connection structure + * @v: Sequence value, set on return + * + * Return: 0 on success, negative error code on failure + */ +static int tcp_flow_dump_seq(const struct tcp_tap_conn *conn, uint32_t *v) +{ + socklen_t sl = sizeof(*v); + + if (getsockopt(conn->sock, SOL_TCP, TCP_QUEUE_SEQ, v, &sl)) { + int rc = -errno; + flow_perror(conn, "Dumping sequence"); + return rc; + } + + return 0; +} + +/** + * tcp_flow_repair_seq() - Restore sequence for pre-selected queue + * @conn: Connection to repair sequences for + * @v: Sequence value to be set + * + * Return: 0 on success, negative error code on failure + */ +static int tcp_flow_repair_seq(const struct tcp_tap_conn *conn, + const uint32_t *v) +{ + if (setsockopt(conn->sock, SOL_TCP, TCP_QUEUE_SEQ, v, sizeof(*v))) { + int rc = -errno; + flow_perror(conn, "Setting sequence"); + return rc; + } + + return 0; +} + +/** + * tcp_flow_dump_rcvqueue() - Dump receive queue and its length, seal/block it + * @conn: Pointer to the TCP connection structure + * @t: Extended migration data + * + * Return: 0 on success, negative error code on failure + * + * #syscalls:vu ioctl + */ +static int tcp_flow_dump_rcvqueue(const struct tcp_tap_conn *conn, + struct tcp_tap_transfer_ext *t) +{ + int s = conn->sock; + ssize_t rc; + + if (ioctl(s, SIOCINQ, &t->rcvq) < 0) { + rc = -errno; + err_perror("Get receive queue size, socket %i", s); + return rc; + } + + /* If we received a FIN, SIOCINQ is one greater than the actual number + * of bytes on the queue, because it's based on the sequence number + * rather than directly on the buffer contents. + * + * This must be called after tcp_flow_dump_tinfo(), for t->tcpi_state. + */ + if (t->rcvq && + (t->tcpi_state == TCP_CLOSE_WAIT || t->tcpi_state == TCP_LAST_ACK)) + t->rcvq--; + + if (t->rcvq > TCP_MIGRATE_RCV_QUEUE_MAX) { + flow_err(conn, + "Receive queue too large to migrate socket: %u bytes", + t->rcvq); + return -ENOBUFS; + } + + rc = recv(s, tcp_migrate_rcv_queue, t->rcvq, MSG_PEEK); + if (rc < 0) { + if (errno == EAGAIN) { /* EAGAIN means empty */ + rc = 0; + } else { + rc = -errno; + flow_perror(conn, "Can't read receive queue"); + return rc; + } + } + + if ((uint32_t)rc < t->rcvq) { + flow_err(conn, "Short read migrating receive queue"); + return -ENXIO; + } + + return 0; +} + +/** + * tcp_flow_repair_opt() - Set repair "options" (MSS, scale, SACK, timestamps) + * @conn: Pointer to the TCP connection structure + * @t: Extended migration data + * + * Return: 0 on success, negative error code on failure + */ +static int tcp_flow_repair_opt(const struct tcp_tap_conn *conn, + const struct tcp_tap_transfer_ext *t) +{ + const struct tcp_repair_opt opts[] = { + { TCPOPT_WINDOW, t->snd_ws + (t->rcv_ws << 16) }, + { TCPOPT_MAXSEG, t->mss }, + { TCPOPT_SACK_PERMITTED, 0 }, + { TCPOPT_TIMESTAMP, 0 }, + }; + socklen_t sl; + + sl = sizeof(opts[0]) * (2 + + !!(t->tcpi_options & TCPI_OPT_SACK) + + !!(t->tcpi_options & TCPI_OPT_TIMESTAMPS)); + + if (setsockopt(conn->sock, SOL_TCP, TCP_REPAIR_OPTIONS, opts, sl)) { + int rc = -errno; + flow_perror(conn, "Setting repair options"); + return rc; + } + + return 0; +} + +/** + * tcp_flow_migrate_source() - Send data (flow table) for flow, close listening + * @fd: Descriptor for state migration + * @conn: Pointer to the TCP connection structure + * + * Return: 0 on success, negative error code on failure + */ +int tcp_flow_migrate_source(int fd, struct tcp_tap_conn *conn) +{ + struct tcp_tap_transfer t = { + .retrans = conn->retrans, + .ws_from_tap = conn->ws_from_tap, + .ws_to_tap = conn->ws_to_tap, + .events = conn->events, + + .tap_mss = htonl(MSS_GET(conn)), + + .sndbuf = htonl(conn->sndbuf), + + .flags = conn->flags, + .seq_dup_ack_approx = conn->seq_dup_ack_approx, + + .wnd_from_tap = htons(conn->wnd_from_tap), + .wnd_to_tap = htons(conn->wnd_to_tap), + + .seq_to_tap = htonl(conn->seq_to_tap), + .seq_ack_from_tap = htonl(conn->seq_ack_from_tap), + .seq_from_tap = htonl(conn->seq_from_tap), + .seq_ack_to_tap = htonl(conn->seq_ack_to_tap), + .seq_init_from_tap = htonl(conn->seq_init_from_tap), + }; + + memcpy(&t.pif, conn->f.pif, sizeof(t.pif)); + memcpy(&t.side, conn->f.side, sizeof(t.side)); + + if (write_all_buf(fd, &t, sizeof(t))) { + int rc = -errno; + err_perror("Can't write migration data, socket %i", conn->sock); + return rc; + } + + if (conn->listening_sock != -1 && !fcntl(conn->listening_sock, F_GETFD)) + close(conn->listening_sock); + + return 0; +} + +/** + * tcp_flow_migrate_source_ext() - Dump queues, close sockets, send final data + * @fd: Descriptor for state migration + * @conn: Pointer to the TCP connection structure + * + * Return: 0 on success, negative (not -EIO) on failure, -EIO on sending failure + */ +int tcp_flow_migrate_source_ext(int fd, const struct tcp_tap_conn *conn) +{ + uint32_t peek_offset = conn->seq_to_tap - conn->seq_ack_from_tap; + struct tcp_tap_transfer_ext *t = &migrate_ext[FLOW_IDX(conn)]; + int s = conn->sock; + int rc; + + /* Disable SO_PEEK_OFF, it will make accessing the queues in repair mode + * weird. + */ + if (tcp_set_peek_offset(conn, -1)) { + rc = -errno; + goto fail; + } + + if ((rc = tcp_flow_dump_tinfo(conn, t))) + goto fail; + + if ((rc = tcp_flow_dump_mss(conn, t))) + goto fail; + + if ((rc = tcp_flow_dump_timestamp(conn, t))) + goto fail; + + if ((rc = tcp_flow_dump_wnd(conn, t))) + goto fail; + + if ((rc = tcp_flow_select_queue(conn, TCP_SEND_QUEUE))) + goto fail; + + if ((rc = tcp_flow_dump_sndqueue(conn, t))) + goto fail; + + if ((rc = tcp_flow_dump_seq(conn, &t->seq_snd))) + goto fail; + + if ((rc = tcp_flow_select_queue(conn, TCP_RECV_QUEUE))) + goto fail; + + if ((rc = tcp_flow_dump_rcvqueue(conn, t))) + goto fail; + + if ((rc = tcp_flow_dump_seq(conn, &t->seq_rcv))) + goto fail; + + close(s); + + /* Adjustments unrelated to FIN segments: sequence numbers we dumped are + * based on the end of the queues. + */ + t->seq_rcv -= t->rcvq; + t->seq_snd -= t->sndq; + + flow_dbg(conn, "Extended migration data, socket %i sequences send %u receive %u", + s, t->seq_snd, t->seq_rcv); + flow_dbg(conn, " pending queues: send %u not sent %u receive %u", + t->sndq, t->notsent, t->rcvq); + flow_dbg(conn, " window: snd_wl1 %u snd_wnd %u max %u rcv_wnd %u rcv_wup %u", + t->snd_wl1, t->snd_wnd, t->max_window, t->rcv_wnd, t->rcv_wup); + flow_dbg(conn, " SO_PEEK_OFF %s offset=%"PRIu32, + peek_offset_cap ? "enabled" : "disabled", peek_offset); + + /* Endianness fix-ups */ + t->seq_snd = htonl(t->seq_snd); + t->seq_rcv = htonl(t->seq_rcv); + t->sndq = htonl(t->sndq); + t->notsent = htonl(t->notsent); + t->rcvq = htonl(t->rcvq); + t->mss = htonl(t->mss); + t->timestamp = htonl(t->timestamp); + + t->snd_wl1 = htonl(t->snd_wl1); + t->snd_wnd = htonl(t->snd_wnd); + t->max_window = htonl(t->max_window); + t->rcv_wnd = htonl(t->rcv_wnd); + t->rcv_wup = htonl(t->rcv_wup); + + if (write_all_buf(fd, t, sizeof(*t))) { + flow_perror(conn, "Failed to write extended data"); + return -EIO; + } + + if (write_all_buf(fd, tcp_migrate_snd_queue, ntohl(t->sndq))) { + flow_perror(conn, "Failed to write send queue data"); + return -EIO; + } + + if (write_all_buf(fd, tcp_migrate_rcv_queue, ntohl(t->rcvq))) { + flow_perror(conn, "Failed to write receive queue data"); + return -EIO; + } + + return 0; + +fail: + /* For any type of failure dumping data, write an invalid extended data + * descriptor that allows us to keep the stream in sync, but tells the + * target to skip the flow. If we fail to transfer data, that's fatal: + * return -EIO in that case (and only in that case). + */ + t->tcpi_state = 0; /* Not defined: tell the target to skip this flow */ + + if (write_all_buf(fd, t, sizeof(*t))) { + flow_perror(conn, "Failed to write extended data"); + return -EIO; + } + + if (rc == -EIO) /* but not a migration data transfer failure */ + return -ENODATA; + + return rc; +} + +/** + * tcp_flow_repair_socket() - Open and bind socket, request repair mode + * @c: Execution context + * @conn: Pointer to the TCP connection structure + * + * Return: 0 on success, negative error code on failure + */ +static int tcp_flow_repair_socket(struct ctx *c, struct tcp_tap_conn *conn) +{ + sa_family_t af = CONN_V4(conn) ? AF_INET : AF_INET6; + int s, rc; + + if ((conn->sock = socket(af, SOCK_STREAM | SOCK_NONBLOCK | SOCK_CLOEXEC, + IPPROTO_TCP)) < 0) { + rc = -errno; + flow_perror(conn, "Failed to create socket for migrated flow"); + return rc; + } + s = conn->sock; + + if (setsockopt(s, SOL_SOCKET, SO_REUSEADDR, &(int){ 1 }, sizeof(int))) + flow_dbg_perror(conn, "Failed to set SO_REUSEADDR on socket %i", + s); + + tcp_sock_set_nodelay(s); + + if ((rc = tcp_flow_repair_on(c, conn))) + goto err; + + return 0; + +err: + close(s); + conn->sock = -1; + return rc; +} + +/** + * tcp_flow_repair_bind() - Bind socket in repair mode + * @c: Execution context + * @conn: Pointer to the TCP connection structure + * + * Return: 0 on success, negative error code on failure + */ +static int tcp_flow_repair_bind(const struct ctx *c, struct tcp_tap_conn *conn) +{ + const struct flowside *sockside = HOSTFLOW(conn); + union sockaddr_inany a; + socklen_t sl; + + pif_sockaddr(c, &a, &sl, PIF_HOST, &sockside->oaddr, sockside->oport); + + if (bind(conn->sock, &a.sa, sizeof(a))) { + int rc = -errno; + flow_perror(conn, "Failed to bind socket for migrated flow"); + return rc; + } + + return 0; +} + +/** + * tcp_flow_repair_connect() - Connect socket in repair mode, then turn it off + * @c: Execution context + * @conn: Pointer to the TCP connection structure + * + * Return: 0 on success, negative error code on failure + */ +static int tcp_flow_repair_connect(const struct ctx *c, + struct tcp_tap_conn *conn) +{ + const struct flowside *tgt = HOSTFLOW(conn); + int rc; + + rc = flowside_connect(c, conn->sock, PIF_HOST, tgt); + if (rc) { + rc = -errno; + flow_perror(conn, "Failed to connect migrated socket"); + return rc; + } + + conn->in_epoll = 0; + conn->timer = -1; + conn->listening_sock = -1; + + return 0; +} + +/** + * tcp_flow_migrate_target() - Receive data (flow table part) for flow, insert + * @c: Execution context + * @fd: Descriptor for state migration + * + * Return: 0 on success, negative on fatal failure, but 0 on single flow failure + */ +int tcp_flow_migrate_target(struct ctx *c, int fd) +{ + struct tcp_tap_transfer t; + struct tcp_tap_conn *conn; + union flow *flow; + int rc; + + if (!(flow = flow_alloc())) { + err("Flow table full on migration target"); + return 0; + } + + if (read_all_buf(fd, &t, sizeof(t))) { + flow_perror(flow, "Failed to receive migration data"); + flow_alloc_cancel(flow); + return -errno; + } + + flow->f.state = FLOW_STATE_TGT; + memcpy(&flow->f.pif, &t.pif, sizeof(flow->f.pif)); + memcpy(&flow->f.side, &t.side, sizeof(flow->f.side)); + conn = FLOW_SET_TYPE(flow, FLOW_TCP, tcp); + + conn->retrans = t.retrans; + conn->ws_from_tap = t.ws_from_tap; + conn->ws_to_tap = t.ws_to_tap; + conn->events = t.events; + + conn->sndbuf = htonl(t.sndbuf); + + conn->flags = t.flags; + conn->seq_dup_ack_approx = t.seq_dup_ack_approx; + + MSS_SET(conn, ntohl(t.tap_mss)); + + conn->wnd_from_tap = ntohs(t.wnd_from_tap); + conn->wnd_to_tap = ntohs(t.wnd_to_tap); + + conn->seq_to_tap = ntohl(t.seq_to_tap); + conn->seq_ack_from_tap = ntohl(t.seq_ack_from_tap); + conn->seq_from_tap = ntohl(t.seq_from_tap); + conn->seq_ack_to_tap = ntohl(t.seq_ack_to_tap); + conn->seq_init_from_tap = ntohl(t.seq_init_from_tap); + + if ((rc = tcp_flow_repair_socket(c, conn))) { + flow_err(flow, "Can't set up socket: %s, drop", strerror_(-rc)); + /* Can't leave the flow in an incomplete state */ + FLOW_ACTIVATE(conn); + return 0; + } + + flow_hash_insert(c, TAP_SIDX(conn)); + FLOW_ACTIVATE(conn); + + return 0; +} + +/** + * tcp_flow_migrate_target_ext() - Receive extended data for flow, set, connect + * @c: Execution context + * @conn: Connection entry to complete with extra data + * @fd: Descriptor for state migration + * + * Return: 0 on success, negative on fatal failure, but 0 on single flow failure + */ +int tcp_flow_migrate_target_ext(struct ctx *c, struct tcp_tap_conn *conn, int fd) +{ + uint32_t peek_offset = conn->seq_to_tap - conn->seq_ack_from_tap; + struct tcp_tap_transfer_ext t; + int s = conn->sock, rc; + + if (read_all_buf(fd, &t, sizeof(t))) { + rc = -errno; + flow_perror(conn, "Failed to read extended data"); + return rc; + } + + if (!t.tcpi_state) { /* Source wants us to skip this flow */ + flow_err(conn, "Dropping as requested by source"); + goto fail; + } + + /* Endianness fix-ups */ + t.seq_snd = ntohl(t.seq_snd); + t.seq_rcv = ntohl(t.seq_rcv); + t.sndq = ntohl(t.sndq); + t.notsent = ntohl(t.notsent); + t.rcvq = ntohl(t.rcvq); + t.mss = ntohl(t.mss); + t.timestamp = ntohl(t.timestamp); + + t.snd_wl1 = ntohl(t.snd_wl1); + t.snd_wnd = ntohl(t.snd_wnd); + t.max_window = ntohl(t.max_window); + t.rcv_wnd = ntohl(t.rcv_wnd); + t.rcv_wup = ntohl(t.rcv_wup); + + flow_dbg(conn, + "Extended migration data, socket %i sequences send %u receive %u", + s, t.seq_snd, t.seq_rcv); + flow_dbg(conn, " pending queues: send %u not sent %u receive %u", + t.sndq, t.notsent, t.rcvq); + flow_dbg(conn, + " window: snd_wl1 %u snd_wnd %u max %u rcv_wnd %u rcv_wup %u", + t.snd_wl1, t.snd_wnd, t.max_window, t.rcv_wnd, t.rcv_wup); + flow_dbg(conn, " SO_PEEK_OFF %s offset=%"PRIu32, + peek_offset_cap ? "enabled" : "disabled", peek_offset); + + if (t.sndq > TCP_MIGRATE_SND_QUEUE_MAX || t.notsent > t.sndq || + t.rcvq > TCP_MIGRATE_RCV_QUEUE_MAX) { + flow_err(conn, + "Bad queues socket %i, send: %u, not sent: %u, receive: %u", + s, t.sndq, t.notsent, t.rcvq); + return -EINVAL; + } + + if (read_all_buf(fd, tcp_migrate_snd_queue, t.sndq)) { + rc = -errno; + flow_perror(conn, "Failed to read send queue data"); + return rc; + } + + if (read_all_buf(fd, tcp_migrate_rcv_queue, t.rcvq)) { + rc = -errno; + flow_perror(conn, "Failed to read receive queue data"); + return rc; + } + + if (conn->sock < 0) + /* We weren't able to create the socket, discard flow */ + goto fail; + + if (tcp_flow_repair_bind(c, conn)) + goto fail; + + if (tcp_flow_repair_timestamp(conn, &t)) + goto fail; + + if (tcp_flow_select_queue(conn, TCP_SEND_QUEUE)) + goto fail; + + if (tcp_flow_repair_seq(conn, &t.seq_snd)) + goto fail; + + if (tcp_flow_select_queue(conn, TCP_RECV_QUEUE)) + goto fail; + + if (tcp_flow_repair_seq(conn, &t.seq_rcv)) + goto fail; + + if (tcp_flow_repair_connect(c, conn)) + goto fail; + + if (tcp_flow_repair_queue(conn, t.rcvq, tcp_migrate_rcv_queue)) + goto fail; + + if (tcp_flow_select_queue(conn, TCP_SEND_QUEUE)) + goto fail; + + if (tcp_flow_repair_queue(conn, t.sndq - t.notsent, + tcp_migrate_snd_queue)) + goto fail; + + if (tcp_flow_repair_opt(conn, &t)) + goto fail; + + /* If we sent a FIN sent and it was acknowledged (TCP_FIN_WAIT2), don't + * send it out, because we already sent it for sure. + * + * Call shutdown(x, SHUT_WR) in repair mode, so that we move to + * FIN_WAIT_1 (tcp_shutdown()) without sending anything + * (goto in tcp_write_xmit()). + */ + if (t.tcpi_state == TCP_FIN_WAIT2) { + int v; + + v = TCP_SEND_QUEUE; + if (setsockopt(s, SOL_TCP, TCP_REPAIR_QUEUE, &v, sizeof(v))) + flow_perror(conn, "Selecting repair queue"); + else + shutdown(s, SHUT_WR); + } + + if (tcp_flow_repair_wnd(conn, &t)) + goto fail; + + tcp_flow_repair_off(c, conn); + repair_flush(c); + + if (t.notsent) { + if (tcp_flow_repair_queue(conn, t.notsent, + tcp_migrate_snd_queue + + (t.sndq - t.notsent))) { + /* This sometimes seems to fail for unclear reasons. + * Don't fail the whole migration, just reset the flow + * and carry on to the next one. + */ + goto fail; + } + } + + /* If we sent a FIN but it wasn't acknowledged yet (TCP_FIN_WAIT1), send + * it out, because we don't know if we already sent it. + * + * Call shutdown(x, SHUT_WR) *not* in repair mode, which moves us to + * TCP_FIN_WAIT1. + */ + if (t.tcpi_state == TCP_FIN_WAIT1) + shutdown(s, SHUT_WR); + + if (tcp_set_peek_offset(conn, peek_offset)) + goto fail; + + tcp_send_flag(c, conn, ACK); + tcp_data_from_sock(c, conn); + + if ((rc = tcp_epoll_ctl(c, conn))) { + flow_dbg(conn, + "Failed to subscribe to epoll for migrated socket: %s", + strerror_(-rc)); + goto fail; + } + + return 0; + +fail: + if (conn->sock >= 0) { + tcp_flow_repair_off(c, conn); + repair_flush(c); + } + + conn->flags = 0; /* Not waiting for ACK, don't schedule timer */ + tcp_rst(c, conn); + + return 0; +} diff --git a/tcp.h b/tcp.h index cf30744..234a803 100644 --- a/tcp.h +++ b/tcp.h @@ -16,7 +16,7 @@ void tcp_listen_handler(const struct ctx *c, union epoll_ref ref, void tcp_sock_handler(const struct ctx *c, union epoll_ref ref, uint32_t events); int tcp_tap_handler(const struct ctx *c, uint8_t pif, sa_family_t af, - const void *saddr, const void *daddr, + const void *saddr, const void *daddr, uint32_t flow_lbl, const struct pool *p, int idx, const struct timespec *now); int tcp_sock_init(const struct ctx *c, const union inany_addr *addr, const char *ifname, in_port_t port); @@ -25,7 +25,6 @@ void tcp_timer(struct ctx *c, const struct timespec *now); void tcp_defer_handler(struct ctx *c); void tcp_update_l2_buf(const unsigned char *eth_d, const unsigned char *eth_s); -int tcp_set_peek_offset(int s, int offset); extern bool peek_offset_cap; diff --git a/tcp_buf.c b/tcp_buf.c index d29c1a9..0530563 100644 --- a/tcp_buf.c +++ b/tcp_buf.c @@ -125,7 +125,7 @@ static void tcp_revert_seq(const struct ctx *c, struct tcp_tap_conn **conns, conn->seq_to_tap = seq; peek_offset = conn->seq_to_tap - conn->seq_ack_from_tap; - if (tcp_set_peek_offset(conn->sock, peek_offset)) + if (tcp_set_peek_offset(conn, peek_offset)) tcp_rst(c, conn); } } @@ -147,6 +147,35 @@ void tcp_payload_flush(const struct ctx *c) tcp_payload_used = 0; } +/** + * tcp_buf_fill_headers() - Fill 802.3, IP, TCP headers in pre-cooked buffers + * @conn: Connection pointer + * @iov: Pointer to an array of iovec of TCP pre-cooked buffers + * @check: Checksum, if already known + * @seq: Sequence number for this segment + * @no_tcp_csum: Do not set TCP checksum + */ +static void tcp_l2_buf_fill_headers(const struct tcp_tap_conn *conn, + struct iovec *iov, const uint16_t *check, + uint32_t seq, bool no_tcp_csum) +{ + struct iov_tail tail = IOV_TAIL(&iov[TCP_IOV_PAYLOAD], 1, 0); + struct tcphdr *th = IOV_REMOVE_HEADER(&tail, struct tcphdr); + struct tap_hdr *taph = iov[TCP_IOV_TAP].iov_base; + const struct flowside *tapside = TAPFLOW(conn); + const struct in_addr *a4 = inany_v4(&tapside->oaddr); + struct ipv6hdr *ip6h = NULL; + struct iphdr *ip4h = NULL; + + if (a4) + ip4h = iov[TCP_IOV_IP].iov_base; + else + ip6h = iov[TCP_IOV_IP].iov_base; + + tcp_fill_headers(conn, taph, ip4h, ip6h, th, &tail, + check, seq, no_tcp_csum); +} + /** * tcp_buf_send_flag() - Send segment with flags to tap (no payload) * @c: Execution context @@ -181,8 +210,10 @@ int tcp_buf_send_flag(const struct ctx *c, struct tcp_tap_conn *conn, int flags) return ret; tcp_payload_used++; - l4len = tcp_l2_buf_fill_headers(conn, iov, optlen, NULL, seq, false); + l4len = optlen + sizeof(struct tcphdr); iov[TCP_IOV_PAYLOAD].iov_len = l4len; + tcp_l2_buf_fill_headers(conn, iov, NULL, seq, false); + if (flags & DUP_ACK) { struct iovec *dup_iov = tcp_l2_iov[tcp_payload_used++]; @@ -208,14 +239,14 @@ int tcp_buf_send_flag(const struct ctx *c, struct tcp_tap_conn *conn, int flags) * @dlen: TCP payload length * @no_csum: Don't compute IPv4 checksum, use the one from previous buffer * @seq: Sequence number to be sent + * @push: Set PSH flag, last segment in a batch */ static void tcp_data_to_tap(const struct ctx *c, struct tcp_tap_conn *conn, - ssize_t dlen, int no_csum, uint32_t seq) + ssize_t dlen, int no_csum, uint32_t seq, bool push) { struct tcp_payload_t *payload; const uint16_t *check = NULL; struct iovec *iov; - size_t l4len; conn->seq_to_tap = seq + dlen; tcp_frame_conns[tcp_payload_used] = conn; @@ -238,8 +269,9 @@ static void tcp_data_to_tap(const struct ctx *c, struct tcp_tap_conn *conn, payload->th.th_x2 = 0; payload->th.th_flags = 0; payload->th.ack = 1; - l4len = tcp_l2_buf_fill_headers(conn, iov, dlen, check, seq, false); - iov[TCP_IOV_PAYLOAD].iov_len = l4len; + payload->th.psh = push; + iov[TCP_IOV_PAYLOAD].iov_len = dlen + sizeof(struct tcphdr); + tcp_l2_buf_fill_headers(conn, iov, check, seq, false); if (++tcp_payload_used > TCP_FRAMES_MEM - 1) tcp_payload_flush(c); } @@ -272,13 +304,14 @@ int tcp_buf_data_from_sock(const struct ctx *c, struct tcp_tap_conn *conn) conn->seq_ack_from_tap, conn->seq_to_tap); conn->seq_to_tap = conn->seq_ack_from_tap; already_sent = 0; - if (tcp_set_peek_offset(s, 0)) { + if (tcp_set_peek_offset(conn, 0)) { tcp_rst(c, conn); return -1; } } if (!wnd_scaled || already_sent >= wnd_scaled) { + conn_flag(c, conn, ACK_FROM_TAP_BLOCKS); conn_flag(c, conn, STALLED); conn_flag(c, conn, ACK_FROM_TAP_DUE); return 0; @@ -329,6 +362,9 @@ int tcp_buf_data_from_sock(const struct ctx *c, struct tcp_tap_conn *conn) return -errno; } + if (already_sent) /* No new data and EAGAIN: set EPOLLET */ + conn_flag(c, conn, STALLED); + return 0; } @@ -354,6 +390,7 @@ int tcp_buf_data_from_sock(const struct ctx *c, struct tcp_tap_conn *conn) return 0; } + conn_flag(c, conn, ~ACK_FROM_TAP_BLOCKS); conn_flag(c, conn, ~STALLED); send_bufs = DIV_ROUND_UP(len, mss); @@ -367,11 +404,14 @@ int tcp_buf_data_from_sock(const struct ctx *c, struct tcp_tap_conn *conn) seq = conn->seq_to_tap; for (i = 0; i < send_bufs; i++) { int no_csum = i && i != send_bufs - 1 && tcp_payload_used; + bool push = false; - if (i == send_bufs - 1) + if (i == send_bufs - 1) { dlen = last_len; + push = true; + } - tcp_data_to_tap(c, conn, dlen, no_csum, seq); + tcp_data_to_tap(c, conn, dlen, no_csum, seq, push); seq += dlen; } diff --git a/tcp_conn.h b/tcp_conn.h index 6ae0511..35d813d 100644 --- a/tcp_conn.h +++ b/tcp_conn.h @@ -19,6 +19,7 @@ * @tap_mss: MSS advertised by tap/guest, rounded to 2 ^ TCP_MSS_BITS * @sock: Socket descriptor number * @events: Connection events, implying connection states + * @listening_sock: Listening socket this socket was accept()ed from, or -1 * @timer: timerfd descriptor for timeout events * @flags: Connection flags representing internal attributes * @sndbuf: Sending buffer in kernel, rounded to 2 ^ SNDBUF_BITS @@ -68,6 +69,7 @@ struct tcp_tap_conn { #define CONN_STATE_BITS /* Setting these clears other flags */ \ (SOCK_ACCEPTED | TAP_SYN_RCVD | ESTABLISHED) + int listening_sock; int timer :FD_REF_BITS; @@ -77,6 +79,7 @@ struct tcp_tap_conn { #define ACTIVE_CLOSE BIT(2) #define ACK_TO_TAP_DUE BIT(3) #define ACK_FROM_TAP_DUE BIT(4) +#define ACK_FROM_TAP_BLOCKS BIT(5) #define SNDBUF_BITS 24 unsigned int sndbuf :SNDBUF_BITS; @@ -95,6 +98,95 @@ struct tcp_tap_conn { uint32_t seq_init_from_tap; }; +/** + * struct tcp_tap_transfer - Migrated TCP data, flow table part, network order + * @pif: Interfaces for each side of the flow + * @side: Addresses and ports for each side of the flow + * @retrans: Number of retransmissions occurred due to ACK_TIMEOUT + * @ws_from_tap: Window scaling factor advertised from tap/guest + * @ws_to_tap: Window scaling factor advertised to tap/guest + * @events: Connection events, implying connection states + * @tap_mss: MSS advertised by tap/guest, rounded to 2 ^ TCP_MSS_BITS + * @sndbuf: Sending buffer in kernel, rounded to 2 ^ SNDBUF_BITS + * @flags: Connection flags representing internal attributes + * @seq_dup_ack_approx: Last duplicate ACK number sent to tap + * @wnd_from_tap: Last window size from tap, unscaled (as received) + * @wnd_to_tap: Sending window advertised to tap, unscaled (as sent) + * @seq_to_tap: Next sequence for packets to tap + * @seq_ack_from_tap: Last ACK number received from tap + * @seq_from_tap: Next sequence for packets from tap (not actually sent) + * @seq_ack_to_tap: Last ACK number sent to tap + * @seq_init_from_tap: Initial sequence number from tap +*/ +struct tcp_tap_transfer { + uint8_t pif[SIDES]; + struct flowside side[SIDES]; + + uint8_t retrans; + uint8_t ws_from_tap; + uint8_t ws_to_tap; + uint8_t events; + + uint32_t tap_mss; + + uint32_t sndbuf; + + uint8_t flags; + uint8_t seq_dup_ack_approx; + + uint16_t wnd_from_tap; + uint16_t wnd_to_tap; + + uint32_t seq_to_tap; + uint32_t seq_ack_from_tap; + uint32_t seq_from_tap; + uint32_t seq_ack_to_tap; + uint32_t seq_init_from_tap; +} __attribute__((packed, aligned(__alignof__(uint32_t)))); + +/** + * struct tcp_tap_transfer_ext - Migrated TCP data, outside flow, network order + * @seq_snd: Socket-side send sequence + * @seq_rcv: Socket-side receive sequence + * @sndq: Length of pending send queue (unacknowledged / not sent) + * @notsent: Part of pending send queue that wasn't sent out yet + * @rcvq: Length of pending receive queue + * @mss: Socket-side MSS clamp + * @timestamp: RFC 7323 timestamp + * @snd_wl1: Next sequence used in window probe (next sequence - 1) + * @snd_wnd: Socket-side sending window + * @max_window: Window clamp + * @rcv_wnd: Socket-side receive window + * @rcv_wup: rcv_nxt on last window update sent + * @snd_ws: Window scaling factor, send + * @rcv_ws: Window scaling factor, receive + * @tcpi_state: Connection state in TCP_INFO style (enum, tcp_states.h) + * @tcpi_options: TCPI_OPT_* constants (timestamps, selective ACK) + */ +struct tcp_tap_transfer_ext { + uint32_t seq_snd; + uint32_t seq_rcv; + + uint32_t sndq; + uint32_t notsent; + uint32_t rcvq; + + uint32_t mss; + uint32_t timestamp; + + /* We can't just use struct tcp_repair_window: we need network order */ + uint32_t snd_wl1; + uint32_t snd_wnd; + uint32_t max_window; + uint32_t rcv_wnd; + uint32_t rcv_wup; + + uint8_t snd_ws; + uint8_t rcv_ws; + uint8_t tcpi_state; + uint8_t tcpi_options; +} __attribute__((packed, aligned(__alignof__(uint32_t)))); + /** * struct tcp_splice_conn - Descriptor for a spliced TCP connection * @f: Generic flow information @@ -139,11 +231,23 @@ extern int init_sock_pool4 [TCP_SOCK_POOL_SIZE]; extern int init_sock_pool6 [TCP_SOCK_POOL_SIZE]; bool tcp_flow_defer(const struct tcp_tap_conn *conn); + +int tcp_flow_repair_on(struct ctx *c, const struct tcp_tap_conn *conn); +int tcp_flow_repair_off(struct ctx *c, const struct tcp_tap_conn *conn); + +int tcp_flow_migrate_source(int fd, struct tcp_tap_conn *conn); +int tcp_flow_migrate_source_ext(int fd, const struct tcp_tap_conn *conn); + +int tcp_flow_migrate_target(struct ctx *c, int fd); +int tcp_flow_migrate_target_ext(struct ctx *c, struct tcp_tap_conn *conn, int fd); + +bool tcp_flow_is_established(const struct tcp_tap_conn *conn); + bool tcp_splice_flow_defer(struct tcp_splice_conn *conn); void tcp_splice_timer(const struct ctx *c, struct tcp_splice_conn *conn); int tcp_conn_pool_sock(int pool[]); -int tcp_conn_sock(const struct ctx *c, sa_family_t af); -int tcp_sock_refill_pool(const struct ctx *c, int pool[], sa_family_t af); +int tcp_conn_sock(sa_family_t af); +int tcp_sock_refill_pool(int pool[], sa_family_t af); void tcp_splice_refill(const struct ctx *c); #endif /* TCP_CONN_H */ diff --git a/tcp_internal.h b/tcp_internal.h index c846f60..36c6533 100644 --- a/tcp_internal.h +++ b/tcp_internal.h @@ -38,9 +38,13 @@ #define OPT_SACK 5 #define OPT_TS 8 -#define TAPSIDE(conn_) ((conn_)->f.pif[1] == PIF_TAP) -#define TAPFLOW(conn_) (&((conn_)->f.side[TAPSIDE(conn_)])) -#define TAP_SIDX(conn_) (FLOW_SIDX((conn_), TAPSIDE(conn_))) +#define TAPSIDE(conn_) ((conn_)->f.pif[1] == PIF_TAP) +#define TAPFLOW(conn_) (&((conn_)->f.side[TAPSIDE(conn_)])) +#define TAP_SIDX(conn_) (FLOW_SIDX((conn_), TAPSIDE(conn_))) + +#define HOSTSIDE(conn_) ((conn_)->f.pif[1] == PIF_HOST) +#define HOSTFLOW(conn_) (&((conn_)->f.side[HOSTSIDE(conn_)])) +#define HOST_SIDX(conn_) (FLOW_SIDX((conn_), TAPSIDE(conn_))) #define CONN_V4(conn) (!!inany_v4(&TAPFLOW(conn)->oaddr)) #define CONN_V6(conn) (!CONN_V4(conn)) @@ -162,14 +166,17 @@ void tcp_rst_do(const struct ctx *c, struct tcp_tap_conn *conn); struct tcp_info_linux; -size_t tcp_l2_buf_fill_headers(const struct tcp_tap_conn *conn, - struct iovec *iov, size_t dlen, - const uint16_t *check, uint32_t seq, - bool no_tcp_csum); +void tcp_fill_headers(const struct tcp_tap_conn *conn, + struct tap_hdr *taph, + struct iphdr *ip4h, struct ipv6hdr *ip6h, + struct tcphdr *th, struct iov_tail *payload, + const uint16_t *ip4_check, uint32_t seq, bool no_tcp_csum); + int tcp_update_seqack_wnd(const struct ctx *c, struct tcp_tap_conn *conn, bool force_seq, struct tcp_info_linux *tinfo); int tcp_prepare_flags(const struct ctx *c, struct tcp_tap_conn *conn, int flags, struct tcphdr *th, struct tcp_syn_opts *opts, size_t *optlen); +int tcp_set_peek_offset(const struct tcp_tap_conn *conn, int offset); #endif /* TCP_INTERNAL_H */ diff --git a/tcp_splice.c b/tcp_splice.c index 93f8bce..60455d6 100644 --- a/tcp_splice.c +++ b/tcp_splice.c @@ -28,7 +28,7 @@ * - FIN_SENT_0: FIN (write shutdown) sent to accepted socket * - FIN_SENT_1: FIN (write shutdown) sent to target socket * - * #syscalls:pasta pipe2|pipe fcntl arm:fcntl64 ppc64:fcntl64 i686:fcntl64 + * #syscalls:pasta pipe2|pipe fcntl arm:fcntl64 ppc64:fcntl64|fcntl i686:fcntl64 */ #include <sched.h> @@ -131,8 +131,12 @@ static void tcp_splice_conn_epoll_events(uint16_t events, ev[1].events = EPOLLOUT; } - flow_foreach_sidei(sidei) - ev[sidei].events |= (events & OUT_WAIT(sidei)) ? EPOLLOUT : 0; + flow_foreach_sidei(sidei) { + if (events & OUT_WAIT(sidei)) { + ev[sidei].events |= EPOLLOUT; + ev[!sidei].events &= ~EPOLLIN; + } + } } /** @@ -160,7 +164,7 @@ static int tcp_splice_epoll_ctl(const struct ctx *c, if (epoll_ctl(c->epollfd, m, conn->s[0], &ev[0]) || epoll_ctl(c->epollfd, m, conn->s[1], &ev[1])) { int ret = -errno; - flow_err(conn, "ERROR on epoll_ctl(): %s", strerror(errno)); + flow_perror(conn, "ERROR on epoll_ctl()"); return ret; } @@ -200,8 +204,8 @@ static void conn_flag_do(const struct ctx *c, struct tcp_splice_conn *conn, } if (flag == CLOSING) { - epoll_ctl(c->epollfd, EPOLL_CTL_DEL, conn->s[0], NULL); - epoll_ctl(c->epollfd, EPOLL_CTL_DEL, conn->s[1], NULL); + epoll_del(c, conn->s[0]); + epoll_del(c, conn->s[1]); } } @@ -313,8 +317,8 @@ static int tcp_splice_connect_finish(const struct ctx *c, if (conn->pipe[sidei][0] < 0) { if (pipe2(conn->pipe[sidei], O_NONBLOCK | O_CLOEXEC)) { - flow_err(conn, "cannot create %d->%d pipe: %s", - sidei, !sidei, strerror(errno)); + flow_perror(conn, "cannot create %d->%d pipe", + sidei, !sidei); conn_flag(c, conn, CLOSING); return -EIO; } @@ -348,9 +352,10 @@ static int tcp_splice_connect(const struct ctx *c, struct tcp_splice_conn *conn) uint8_t tgtpif = conn->f.pif[TGTSIDE]; union sockaddr_inany sa; socklen_t sl; + int one = 1; if (tgtpif == PIF_HOST) - conn->s[1] = tcp_conn_sock(c, af); + conn->s[1] = tcp_conn_sock(af); else if (tgtpif == PIF_SPLICE) conn->s[1] = tcp_conn_sock_ns(c, af); else @@ -359,18 +364,27 @@ static int tcp_splice_connect(const struct ctx *c, struct tcp_splice_conn *conn) if (conn->s[1] < 0) return -1; - if (setsockopt(conn->s[1], SOL_TCP, TCP_QUICKACK, - &((int){ 1 }), sizeof(int))) { + if (setsockopt(conn->s[1], SOL_TCP, TCP_QUICKACK, &one, sizeof(one))) { flow_trace(conn, "failed to set TCP_QUICKACK on socket %i", conn->s[1]); } + if (setsockopt(conn->s[0], SOL_TCP, TCP_NODELAY, &one, sizeof(one))) { + flow_trace(conn, "failed to set TCP_NODELAY on socket %i", + conn->s[0]); + } + + if (setsockopt(conn->s[1], SOL_TCP, TCP_NODELAY, &one, sizeof(one))) { + flow_trace(conn, "failed to set TCP_NODELAY on socket %i", + conn->s[1]); + } + pif_sockaddr(c, &sa, &sl, tgtpif, &tgt->eaddr, tgt->eport); if (connect(conn->s[1], &sa.sa, sl)) { if (errno != EINPROGRESS) { flow_trace(conn, "Couldn't connect socket for splice: %s", - strerror(errno)); + strerror_(errno)); return -errno; } @@ -468,11 +482,10 @@ void tcp_splice_sock_handler(struct ctx *c, union epoll_ref ref, rc = getsockopt(ref.fd, SOL_SOCKET, SO_ERROR, &err, &sl); if (rc) - flow_err(conn, "Error retrieving SO_ERROR: %s", - strerror(errno)); + flow_perror(conn, "Error retrieving SO_ERROR"); else flow_trace(conn, "Error event on socket: %s", - strerror(err)); + strerror_(err)); goto close; } @@ -507,20 +520,21 @@ swap: int more = 0; retry: - readlen = splice(conn->s[fromsidei], NULL, - conn->pipe[fromsidei][1], NULL, - c->tcp.pipe_size, - SPLICE_F_MOVE | SPLICE_F_NONBLOCK); - flow_trace(conn, "%zi from read-side call", readlen); - if (readlen < 0) { - if (errno == EINTR) - goto retry; + do + readlen = splice(conn->s[fromsidei], NULL, + conn->pipe[fromsidei][1], NULL, + c->tcp.pipe_size, + SPLICE_F_MOVE | SPLICE_F_NONBLOCK); + while (readlen < 0 && errno == EINTR); - if (errno != EAGAIN) - goto close; - } else if (!readlen) { + if (readlen < 0 && errno != EAGAIN) + goto close; + + flow_trace(conn, "%zi from read-side call", readlen); + + if (!readlen) { eof = 1; - } else { + } else if (readlen > 0) { never_read = 0; if (readlen >= (long)c->tcp.pipe_size * 90 / 100) @@ -530,10 +544,16 @@ retry: conn_flag(c, conn, lowat_act_flag); } -eintr: - written = splice(conn->pipe[fromsidei][0], NULL, - conn->s[!fromsidei], NULL, c->tcp.pipe_size, - SPLICE_F_MOVE | more | SPLICE_F_NONBLOCK); + do + written = splice(conn->pipe[fromsidei][0], NULL, + conn->s[!fromsidei], NULL, + c->tcp.pipe_size, + SPLICE_F_MOVE | more | SPLICE_F_NONBLOCK); + while (written < 0 && errno == EINTR); + + if (written < 0 && errno != EAGAIN) + goto close; + flow_trace(conn, "%zi from write-side call (passed %zi)", written, c->tcp.pipe_size); @@ -542,7 +562,7 @@ eintr: if (readlen >= (long)c->tcp.pipe_size * 10 / 100) continue; - if (conn->flags & lowat_set_flag && + if (!(conn->flags & lowat_set_flag) && readlen > (long)c->tcp.pipe_size / 10) { int lowat = c->tcp.pipe_size / 4; @@ -551,7 +571,7 @@ eintr: &lowat, sizeof(lowat))) { flow_trace(conn, "Setting SO_RCVLOWAT %i: %s", - lowat, strerror(errno)); + lowat, strerror_(errno)); } else { conn_flag(c, conn, lowat_set_flag); conn_flag(c, conn, lowat_act_flag); @@ -565,12 +585,6 @@ eintr: conn->written[fromsidei] += written > 0 ? written : 0; if (written < 0) { - if (errno == EINTR) - goto eintr; - - if (errno != EAGAIN) - goto close; - if (conn->read[fromsidei] == conn->written[fromsidei]) break; @@ -693,16 +707,16 @@ static int tcp_sock_refill_ns(void *arg) ns_enter(c); if (c->ifi4) { - int rc = tcp_sock_refill_pool(c, ns_sock_pool4, AF_INET); + int rc = tcp_sock_refill_pool(ns_sock_pool4, AF_INET); if (rc < 0) warn("TCP: Error refilling IPv4 ns socket pool: %s", - strerror(-rc)); + strerror_(-rc)); } if (c->ifi6) { - int rc = tcp_sock_refill_pool(c, ns_sock_pool6, AF_INET6); + int rc = tcp_sock_refill_pool(ns_sock_pool6, AF_INET6); if (rc < 0) warn("TCP: Error refilling IPv6 ns socket pool: %s", - strerror(-rc)); + strerror_(-rc)); } return 0; diff --git a/tcp_vu.c b/tcp_vu.c new file mode 100644 index 0000000..57587cc --- /dev/null +++ b/tcp_vu.c @@ -0,0 +1,474 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* tcp_vu.c - TCP L2 vhost-user management functions + * + * Copyright Red Hat + * Author: Laurent Vivier <lvivier@redhat.com> + */ + +#include <errno.h> +#include <stddef.h> +#include <stdint.h> + +#include <netinet/ip.h> +#include <netinet/tcp.h> + +#include <sys/socket.h> + +#include <netinet/if_ether.h> +#include <linux/virtio_net.h> + +#include "util.h" +#include "ip.h" +#include "passt.h" +#include "siphash.h" +#include "inany.h" +#include "vhost_user.h" +#include "tcp.h" +#include "pcap.h" +#include "flow.h" +#include "tcp_conn.h" +#include "flow_table.h" +#include "tcp_vu.h" +#include "tap.h" +#include "tcp_internal.h" +#include "checksum.h" +#include "vu_common.h" +#include <time.h> + +static struct iovec iov_vu[VIRTQUEUE_MAX_SIZE + 1]; +static struct vu_virtq_element elem[VIRTQUEUE_MAX_SIZE]; +static int head[VIRTQUEUE_MAX_SIZE + 1]; + +/** + * tcp_vu_hdrlen() - return the size of the header in level 2 frame (TCP) + * @v6: Set for IPv6 packet + * + * Return: Return the size of the header + */ +static size_t tcp_vu_hdrlen(bool v6) +{ + size_t hdrlen; + + hdrlen = sizeof(struct virtio_net_hdr_mrg_rxbuf) + + sizeof(struct ethhdr) + sizeof(struct tcphdr); + + if (v6) + hdrlen += sizeof(struct ipv6hdr); + else + hdrlen += sizeof(struct iphdr); + + return hdrlen; +} + +/** + * tcp_vu_send_flag() - Send segment with flags to vhost-user (no payload) + * @c: Execution context + * @conn: Connection pointer + * @flags: TCP flags: if not set, send segment only if ACK is due + * + * Return: negative error code on connection reset, 0 otherwise + */ +int tcp_vu_send_flag(const struct ctx *c, struct tcp_tap_conn *conn, int flags) +{ + struct vu_dev *vdev = c->vdev; + struct vu_virtq *vq = &vdev->vq[VHOST_USER_RX_QUEUE]; + size_t optlen, hdrlen; + struct vu_virtq_element flags_elem[2]; + struct ipv6hdr *ip6h = NULL; + struct iphdr *ip4h = NULL; + struct iovec flags_iov[2]; + struct tcp_syn_opts *opts; + struct iov_tail payload; + struct tcphdr *th; + struct ethhdr *eh; + uint32_t seq; + int elem_cnt; + int nb_ack; + int ret; + + hdrlen = tcp_vu_hdrlen(CONN_V6(conn)); + + vu_set_element(&flags_elem[0], NULL, &flags_iov[0]); + + elem_cnt = vu_collect(vdev, vq, &flags_elem[0], 1, + hdrlen + sizeof(struct tcp_syn_opts), NULL); + if (elem_cnt != 1) + return -1; + + ASSERT(flags_elem[0].in_sg[0].iov_len >= + hdrlen + sizeof(struct tcp_syn_opts)); + + vu_set_vnethdr(vdev, flags_elem[0].in_sg[0].iov_base, 1); + + eh = vu_eth(flags_elem[0].in_sg[0].iov_base); + + memcpy(eh->h_dest, c->guest_mac, sizeof(eh->h_dest)); + memcpy(eh->h_source, c->our_tap_mac, sizeof(eh->h_source)); + + if (CONN_V4(conn)) { + eh->h_proto = htons(ETH_P_IP); + + ip4h = vu_ip(flags_elem[0].in_sg[0].iov_base); + *ip4h = (struct iphdr)L2_BUF_IP4_INIT(IPPROTO_TCP); + + th = vu_payloadv4(flags_elem[0].in_sg[0].iov_base); + } else { + eh->h_proto = htons(ETH_P_IPV6); + + ip6h = vu_ip(flags_elem[0].in_sg[0].iov_base); + *ip6h = (struct ipv6hdr)L2_BUF_IP6_INIT(IPPROTO_TCP); + th = vu_payloadv6(flags_elem[0].in_sg[0].iov_base); + } + + memset(th, 0, sizeof(*th)); + th->doff = sizeof(*th) / 4; + th->ack = 1; + + seq = conn->seq_to_tap; + opts = (struct tcp_syn_opts *)(th + 1); + ret = tcp_prepare_flags(c, conn, flags, th, opts, &optlen); + if (ret <= 0) { + vu_queue_rewind(vq, 1); + return ret; + } + + flags_elem[0].in_sg[0].iov_len = hdrlen + optlen; + payload = IOV_TAIL(flags_elem[0].in_sg, 1, hdrlen); + + tcp_fill_headers(conn, NULL, ip4h, ip6h, th, &payload, + NULL, seq, !*c->pcap); + + if (*c->pcap) { + pcap_iov(&flags_elem[0].in_sg[0], 1, + sizeof(struct virtio_net_hdr_mrg_rxbuf)); + } + nb_ack = 1; + + if (flags & DUP_ACK) { + vu_set_element(&flags_elem[1], NULL, &flags_iov[1]); + + elem_cnt = vu_collect(vdev, vq, &flags_elem[1], 1, + flags_elem[0].in_sg[0].iov_len, NULL); + if (elem_cnt == 1 && + flags_elem[1].in_sg[0].iov_len >= + flags_elem[0].in_sg[0].iov_len) { + memcpy(flags_elem[1].in_sg[0].iov_base, + flags_elem[0].in_sg[0].iov_base, + flags_elem[0].in_sg[0].iov_len); + nb_ack++; + + if (*c->pcap) { + pcap_iov(&flags_elem[1].in_sg[0], 1, + sizeof(struct virtio_net_hdr_mrg_rxbuf)); + } + } + } + + vu_flush(vdev, vq, flags_elem, nb_ack); + + return 0; +} + +/** tcp_vu_sock_recv() - Receive datastream from socket into vhost-user buffers + * @c: Execution context + * @conn: Connection pointer + * @v6: Set for IPv6 connections + * @already_sent: Number of bytes already sent + * @fillsize: Maximum bytes to fill in guest-side receiving window + * @iov_cnt: number of iov (output) + * + * Return: Number of iov entries used to store the data or negative error code + */ +static ssize_t tcp_vu_sock_recv(const struct ctx *c, + const struct tcp_tap_conn *conn, bool v6, + uint32_t already_sent, size_t fillsize, + int *iov_cnt, int *head_cnt) +{ + struct vu_dev *vdev = c->vdev; + struct vu_virtq *vq = &vdev->vq[VHOST_USER_RX_QUEUE]; + struct msghdr mh_sock = { 0 }; + uint16_t mss = MSS_GET(conn); + int s = conn->sock; + ssize_t ret, len; + size_t hdrlen; + int elem_cnt; + int i; + + *iov_cnt = 0; + + hdrlen = tcp_vu_hdrlen(v6); + + vu_init_elem(elem, &iov_vu[1], VIRTQUEUE_MAX_SIZE); + + elem_cnt = 0; + *head_cnt = 0; + while (fillsize > 0 && elem_cnt < VIRTQUEUE_MAX_SIZE) { + struct iovec *iov; + size_t frame_size, dlen; + int cnt; + + cnt = vu_collect(vdev, vq, &elem[elem_cnt], + VIRTQUEUE_MAX_SIZE - elem_cnt, + MIN(mss, fillsize) + hdrlen, &frame_size); + if (cnt == 0) + break; + + dlen = frame_size - hdrlen; + + /* reserve space for headers in iov */ + iov = &elem[elem_cnt].in_sg[0]; + ASSERT(iov->iov_len >= hdrlen); + iov->iov_base = (char *)iov->iov_base + hdrlen; + iov->iov_len -= hdrlen; + head[(*head_cnt)++] = elem_cnt; + + fillsize -= dlen; + elem_cnt += cnt; + } + + if (peek_offset_cap) { + mh_sock.msg_iov = iov_vu + 1; + mh_sock.msg_iovlen = elem_cnt; + } else { + iov_vu[0].iov_base = tcp_buf_discard; + iov_vu[0].iov_len = already_sent; + + mh_sock.msg_iov = iov_vu; + mh_sock.msg_iovlen = elem_cnt + 1; + } + + do + ret = recvmsg(s, &mh_sock, MSG_PEEK); + while (ret < 0 && errno == EINTR); + + if (ret < 0) { + vu_queue_rewind(vq, elem_cnt); + return -errno; + } + + if (!peek_offset_cap) + ret -= already_sent; + + /* adjust iov number and length of the last iov */ + len = ret; + for (i = 0; len && i < elem_cnt; i++) { + struct iovec *iov = &elem[i].in_sg[0]; + + if (iov->iov_len > (size_t)len) + iov->iov_len = len; + + len -= iov->iov_len; + } + /* adjust head count */ + while (*head_cnt > 0 && head[*head_cnt - 1] >= i) + (*head_cnt)--; + + /* mark end of array */ + head[*head_cnt] = i; + *iov_cnt = i; + + /* release unused buffers */ + vu_queue_rewind(vq, elem_cnt - i); + + /* restore space for headers in iov */ + for (i = 0; i < *head_cnt; i++) { + struct iovec *iov = &elem[head[i]].in_sg[0]; + + iov->iov_base = (char *)iov->iov_base - hdrlen; + iov->iov_len += hdrlen; + } + + return ret; +} + +/** + * tcp_vu_prepare() - Prepare the frame header + * @c: Execution context + * @conn: Connection pointer + * @iov: Pointer to the array of IO vectors + * @iov_cnt: Number of entries in @iov + * @check: Checksum, if already known + * @no_tcp_csum: Do not set TCP checksum + * @push: Set PSH flag, last segment in a batch + */ +static void tcp_vu_prepare(const struct ctx *c, struct tcp_tap_conn *conn, + struct iovec *iov, size_t iov_cnt, + const uint16_t **check, bool no_tcp_csum, bool push) +{ + const struct flowside *toside = TAPFLOW(conn); + bool v6 = !(inany_v4(&toside->eaddr) && inany_v4(&toside->oaddr)); + size_t hdrlen = tcp_vu_hdrlen(v6); + struct iov_tail payload = IOV_TAIL(iov, iov_cnt, hdrlen); + char *base = iov[0].iov_base; + struct ipv6hdr *ip6h = NULL; + struct iphdr *ip4h = NULL; + struct tcphdr *th; + struct ethhdr *eh; + + /* we guess the first iovec provided by the guest can embed + * all the headers needed by L2 frame + */ + ASSERT(iov[0].iov_len >= hdrlen); + + eh = vu_eth(base); + + memcpy(eh->h_dest, c->guest_mac, sizeof(eh->h_dest)); + memcpy(eh->h_source, c->our_tap_mac, sizeof(eh->h_source)); + + /* initialize header */ + + if (!v6) { + eh->h_proto = htons(ETH_P_IP); + + ip4h = vu_ip(base); + *ip4h = (struct iphdr)L2_BUF_IP4_INIT(IPPROTO_TCP); + th = vu_payloadv4(base); + } else { + eh->h_proto = htons(ETH_P_IPV6); + + ip6h = vu_ip(base); + *ip6h = (struct ipv6hdr)L2_BUF_IP6_INIT(IPPROTO_TCP); + + th = vu_payloadv6(base); + } + + memset(th, 0, sizeof(*th)); + th->doff = sizeof(*th) / 4; + th->ack = 1; + th->psh = push; + + tcp_fill_headers(conn, NULL, ip4h, ip6h, th, &payload, + *check, conn->seq_to_tap, no_tcp_csum); + if (ip4h) + *check = &ip4h->check; +} + +/** + * tcp_vu_data_from_sock() - Handle new data from socket, queue to vhost-user, + * in window + * @c: Execution context + * @conn: Connection pointer + * + * Return: Negative on connection reset, 0 otherwise + */ +int tcp_vu_data_from_sock(const struct ctx *c, struct tcp_tap_conn *conn) +{ + uint32_t wnd_scaled = conn->wnd_from_tap << conn->ws_from_tap; + struct vu_dev *vdev = c->vdev; + struct vu_virtq *vq = &vdev->vq[VHOST_USER_RX_QUEUE]; + ssize_t len, previous_dlen; + int i, iov_cnt, head_cnt; + size_t hdrlen, fillsize; + int v6 = CONN_V6(conn); + uint32_t already_sent; + const uint16_t *check; + + if (!vu_queue_enabled(vq) || !vu_queue_started(vq)) { + debug("Got packet, but RX virtqueue not usable yet"); + return 0; + } + + already_sent = conn->seq_to_tap - conn->seq_ack_from_tap; + + if (SEQ_LT(already_sent, 0)) { + /* RFC 761, section 2.1. */ + flow_trace(conn, "ACK sequence gap: ACK for %u, sent: %u", + conn->seq_ack_from_tap, conn->seq_to_tap); + conn->seq_to_tap = conn->seq_ack_from_tap; + already_sent = 0; + if (tcp_set_peek_offset(conn, 0)) { + tcp_rst(c, conn); + return -1; + } + } + + if (!wnd_scaled || already_sent >= wnd_scaled) { + conn_flag(c, conn, ACK_FROM_TAP_BLOCKS); + conn_flag(c, conn, STALLED); + conn_flag(c, conn, ACK_FROM_TAP_DUE); + return 0; + } + + /* Set up buffer descriptors we'll fill completely and partially. */ + + fillsize = wnd_scaled - already_sent; + + /* collect the buffers from vhost-user and fill them with the + * data from the socket + */ + len = tcp_vu_sock_recv(c, conn, v6, already_sent, fillsize, + &iov_cnt, &head_cnt); + if (len < 0) { + if (len != -EAGAIN && len != -EWOULDBLOCK) { + tcp_rst(c, conn); + return len; + } + + if (already_sent) /* No new data and EAGAIN: set EPOLLET */ + conn_flag(c, conn, STALLED); + + return 0; + } + + if (!len) { + if (already_sent) { + conn_flag(c, conn, STALLED); + } else if ((conn->events & (SOCK_FIN_RCVD | TAP_FIN_SENT)) == + SOCK_FIN_RCVD) { + int ret = tcp_vu_send_flag(c, conn, FIN | ACK); + if (ret) { + tcp_rst(c, conn); + return ret; + } + + conn_event(c, conn, TAP_FIN_SENT); + } + + return 0; + } + + conn_flag(c, conn, ~ACK_FROM_TAP_BLOCKS); + conn_flag(c, conn, ~STALLED); + + /* Likely, some new data was acked too. */ + tcp_update_seqack_wnd(c, conn, false, NULL); + + /* initialize headers */ + /* iov_vu is an array of buffers and the buffer size can be + * smaller than the frame size we want to use but with + * num_buffer we can merge several virtio iov buffers in one packet + * we need only to set the packet headers in the first iov and + * num_buffer to the number of iov entries + */ + + hdrlen = tcp_vu_hdrlen(v6); + for (i = 0, previous_dlen = -1, check = NULL; i < head_cnt; i++) { + struct iovec *iov = &elem[head[i]].in_sg[0]; + int buf_cnt = head[i + 1] - head[i]; + ssize_t dlen = iov_size(iov, buf_cnt) - hdrlen; + bool push = i == head_cnt - 1; + + vu_set_vnethdr(vdev, iov->iov_base, buf_cnt); + + /* The IPv4 header checksum varies only with dlen */ + if (previous_dlen != dlen) + check = NULL; + previous_dlen = dlen; + + tcp_vu_prepare(c, conn, iov, buf_cnt, &check, !*c->pcap, push); + + if (*c->pcap) { + pcap_iov(iov, buf_cnt, + sizeof(struct virtio_net_hdr_mrg_rxbuf)); + } + + conn->seq_to_tap += dlen; + } + + /* send packets */ + vu_flush(vdev, vq, elem, iov_cnt); + + conn_flag(c, conn, ACK_FROM_TAP_DUE); + + return 0; +} diff --git a/tcp_vu.h b/tcp_vu.h new file mode 100644 index 0000000..6ab6057 --- /dev/null +++ b/tcp_vu.h @@ -0,0 +1,12 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* Copyright Red Hat + * Author: Laurent Vivier <lvivier@redhat.com> + */ + +#ifndef TCP_VU_H +#define TCP_VU_H + +int tcp_vu_send_flag(const struct ctx *c, struct tcp_tap_conn *conn, int flags); +int tcp_vu_data_from_sock(const struct ctx *c, struct tcp_tap_conn *conn); + +#endif /*TCP_VU_H */ diff --git a/test/.gitignore b/test/.gitignore index 6dd4790..3573444 100644 --- a/test/.gitignore +++ b/test/.gitignore @@ -8,5 +8,6 @@ QEMU_EFI.fd *.raw.xz *.bin nstool +rampstream guest-key guest-key.pub diff --git a/test/Makefile b/test/Makefile index 5e49047..bf63db8 100644 --- a/test/Makefile +++ b/test/Makefile @@ -52,7 +52,8 @@ UBUNTU_IMGS = $(UBUNTU_OLD_IMGS) $(UBUNTU_NEW_IMGS) DOWNLOAD_ASSETS = mbuto podman \ $(DEBIAN_IMGS) $(FEDORA_IMGS) $(OPENSUSE_IMGS) $(UBUNTU_IMGS) -TESTDATA_ASSETS = small.bin big.bin medium.bin +TESTDATA_ASSETS = small.bin big.bin medium.bin \ + rampstream LOCAL_ASSETS = mbuto.img mbuto.mem.img podman/bin/podman QEMU_EFI.fd \ $(DEBIAN_IMGS:%=prepared-%) $(FEDORA_IMGS:%=prepared-%) \ $(UBUNTU_NEW_IMGS:%=prepared-%) \ @@ -85,7 +86,7 @@ podman/bin/podman: pull-podman guest-key guest-key.pub: ssh-keygen -f guest-key -N '' -mbuto.img: passt.mbuto mbuto/mbuto guest-key.pub $(TESTDATA_ASSETS) +mbuto.img: passt.mbuto mbuto/mbuto guest-key.pub rampstream-check.sh $(TESTDATA_ASSETS) ./mbuto/mbuto -p ./$< -c lz4 -f $@ mbuto.mem.img: passt.mem.mbuto mbuto ../passt.avx2 diff --git a/test/lib/layout b/test/lib/layout index 4d03572..fddcdc4 100644 --- a/test/lib/layout +++ b/test/lib/layout @@ -135,7 +135,7 @@ layout_two_guests() { get_info_cols pane_watch_contexts ${PANE_GUEST_1} "guest #1 in namespace #1" qemu_1 guest_1 - pane_watch_contexts ${PANE_GUEST_2} "guest #2 in namespace #2" qemu_2 guest_2 + pane_watch_contexts ${PANE_GUEST_2} "guest #2 in namespace #1" qemu_2 guest_2 tmux send-keys -l -t ${PANE_INFO} 'while cat '"$STATEBASE/log_pipe"'; do :; done' tmux send-keys -t ${PANE_INFO} -N 100 C-m @@ -143,13 +143,66 @@ layout_two_guests() { pane_watch_contexts ${PANE_HOST} host host pane_watch_contexts ${PANE_PASST_1} "passt #1 in namespace #1" pasta_1 passt_1 - pane_watch_contexts ${PANE_PASST_2} "passt #2 in namespace #2" pasta_2 passt_2 + pane_watch_contexts ${PANE_PASST_2} "passt #2 in namespace #1" pasta_1 passt_2 info_layout "two guests, two passt instances, in namespaces" sleep 1 } +# layout_migrate() - Two guest panes, two passt panes, two passt-repair panes, +# plus host and log +layout_migrate() { + sleep 1 + + tmux kill-pane -a -t 0 + cmd_write 0 clear + + tmux split-window -v -t passt_test + tmux split-window -h -l '33%' + tmux split-window -h -t passt_test:1.1 + + tmux split-window -h -l '35%' -t passt_test:1.0 + tmux split-window -v -t passt_test:1.0 + + tmux split-window -v -t passt_test:1.4 + tmux split-window -v -t passt_test:1.6 + + tmux split-window -v -t passt_test:1.3 + + PANE_GUEST_1=0 + PANE_GUEST_2=1 + PANE_INFO=2 + PANE_MON=3 + PANE_HOST=4 + PANE_PASST_REPAIR_1=5 + PANE_PASST_1=6 + PANE_PASST_REPAIR_2=7 + PANE_PASST_2=8 + + get_info_cols + + pane_watch_contexts ${PANE_GUEST_1} "guest #1 in namespace #1" qemu_1 guest_1 + pane_watch_contexts ${PANE_GUEST_2} "guest #2 in namespace #2" qemu_2 guest_2 + + tmux send-keys -l -t ${PANE_INFO} 'while cat '"$STATEBASE/log_pipe"'; do :; done' + tmux send-keys -t ${PANE_INFO} -N 100 C-m + tmux select-pane -t ${PANE_INFO} -T "test log" + + pane_watch_contexts ${PANE_MON} "QEMU monitor" mon mon + + pane_watch_contexts ${PANE_HOST} host host + pane_watch_contexts ${PANE_PASST_REPAIR_1} "passt-repair #1 in namespace #1" repair_1 passt_repair_1 + pane_watch_contexts ${PANE_PASST_1} "passt #1 in namespace #1" pasta_1 passt_1 + + pane_watch_contexts ${PANE_PASST_REPAIR_2} "passt-repair #2 in namespace #2" repair_2 passt_repair_2 + pane_watch_contexts ${PANE_PASST_2} "passt #2 in namespace #2" pasta_2 passt_2 + + info_layout "two guests, two passt + passt-repair instances, in namespaces" + + sleep 1 +} + # layout_demo_pasta() - Four panes for pasta demo layout_demo_pasta() { sleep 1 diff --git a/test/lib/perf_report b/test/lib/perf_report index d1ef50b..c4ec817 100755 --- a/test/lib/perf_report +++ b/test/lib/perf_report @@ -49,6 +49,21 @@ td:empty { visibility: hidden; } __passt_tcp_LINE__ __passt_udp_LINE__ </table> +</li><li><p>passt with vhost-user support</p> +<table class="passt" width="70%"> + <tr> + <th/> + <th id="perf_passt_vu_tcp" colspan="__passt_vu_tcp_cols__">TCP, __passt_vu_tcp_threads__ at __passt_vu_tcp_freq__ GHz</th> + <th id="perf_passt_vu_udp" colspan="__passt_vu_udp_cols__">UDP, __passt_vu_udp_threads__ at __passt_vu_udp_freq__ GHz</th> + </tr> + <tr> + <td align="right">MTU:</td> + __passt_vu_tcp_header__ + __passt_vu_udp_header__ + </tr> + __passt_vu_tcp_LINE__ __passt_vu_udp_LINE__ +</table> + <style type="text/CSS"> table.pasta_local td { border: 0px solid; padding: 6px; line-height: 1; } table.pasta_local td { text-align: right; } diff --git a/test/lib/setup b/test/lib/setup index 5338393..575bc21 100755 --- a/test/lib/setup +++ b/test/lib/setup @@ -15,8 +15,7 @@ INITRAMFS="${BASEPATH}/mbuto.img" VCPUS="$( [ $(nproc) -ge 8 ] && echo 6 || echo $(( $(nproc) / 2 + 1 )) )" -__mem_kib="$(sed -n 's/MemTotal:[ ]*\([0-9]*\) kB/\1/p' /proc/meminfo)" -VMEM="$((${__mem_kib} / 1024 / 4))" +MEM_KIB="$(sed -n 's/MemTotal:[ ]*\([0-9]*\) kB/\1/p' /proc/meminfo)" QEMU_ARCH="$(uname -m)" [ "${QEMU_ARCH}" = "i686" ] && QEMU_ARCH=i386 @@ -46,24 +45,38 @@ setup_passt() { [ ${PCAP} -eq 1 ] && __opts="${__opts} -p ${LOGDIR}/passt.pcap" [ ${DEBUG} -eq 1 ] && __opts="${__opts} -d" [ ${TRACE} -eq 1 ] && __opts="${__opts} --trace" + [ ${VHOST_USER} -eq 1 ] && __opts="${__opts} --vhost-user" context_run passt "make clean" context_run passt "make valgrind" - context_run_bg passt "valgrind --max-stackframe=$((4 * 1024 * 1024)) --trace-children=yes --vgdb=no --error-exitcode=1 --suppressions=test/valgrind.supp ./passt ${__opts} -s ${STATESETUP}/passt.socket -f -t 10001 -u 10001 -P ${STATESETUP}/passt.pid" + context_run_bg passt "valgrind --max-stackframe=$((4 * 1024 * 1024)) --trace-children=yes --vgdb=no --error-exitcode=1 --suppressions=test/valgrind.supp ./passt ${__opts} -s ${STATESETUP}/passt.socket -f -t 10001 -u 10001 -H hostname1 --fqdn fqdn1.passt.test -P ${STATESETUP}/passt.pid" # pidfile isn't created until passt is listening wait_for [ -f "${STATESETUP}/passt.pid" ] + __vmem="$((${MEM_KIB} / 1024 / 4))" + if [ ${VHOST_USER} -eq 1 ]; then + __vmem="$(((${__vmem} + 500) / 1000))G" + __qemu_netdev=" \ + -chardev socket,id=c,path=${STATESETUP}/passt.socket \ + -netdev vhost-user,id=v,chardev=c \ + -device virtio-net,netdev=v \ + -object memory-backend-memfd,id=m,share=on,size=${__vmem} \ + -numa node,memdev=m" + else + __qemu_netdev="-device virtio-net-pci,netdev=s \ + -netdev stream,id=s,server=off,addr.type=unix,addr.path=${STATESETUP}/passt.socket" + fi + GUEST_CID=94557 context_run_bg qemu 'qemu-system-'"${QEMU_ARCH}" \ ' -machine accel=kvm' \ - ' -m '${VMEM}' -cpu host -smp '${VCPUS} \ + ' -m '${__vmem}' -cpu host -smp '${VCPUS} \ ' -kernel '"${KERNEL}" \ ' -initrd '${INITRAMFS}' -nographic -serial stdio' \ ' -nodefaults' \ ' -append "console=ttyS0 mitigations=off apparmor=0" ' \ - ' -device virtio-net-pci,netdev=s0 ' \ - " -netdev stream,id=s0,server=off,addr.type=unix,addr.path=${STATESETUP}/passt.socket " \ + " ${__qemu_netdev}" \ " -pidfile ${STATESETUP}/qemu.pid" \ " -device vhost-vsock-pci,guest-cid=$GUEST_CID" @@ -142,29 +155,43 @@ setup_passt_in_ns() { [ ${PCAP} -eq 1 ] && __opts="${__opts} -p ${LOGDIR}/passt_in_pasta.pcap" [ ${DEBUG} -eq 1 ] && __opts="${__opts} -d" [ ${TRACE} -eq 1 ] && __opts="${__opts} --trace" + [ ${VHOST_USER} -eq 1 ] && __opts="${__opts} --vhost-user" if [ ${VALGRIND} -eq 1 ]; then context_run passt "make clean" context_run passt "make valgrind" - context_run_bg passt "valgrind --max-stackframe=$((4 * 1024 * 1024)) --trace-children=yes --vgdb=no --error-exitcode=1 --suppressions=test/valgrind.supp ./passt -f ${__opts} -s ${STATESETUP}/passt.socket -t 10001,10011,10021,10031 -u 10001,10011,10021,10031 -P ${STATESETUP}/passt.pid --map-host-loopback ${__map_ns4} --map-host-loopback ${__map_ns6}" + context_run_bg passt "valgrind --max-stackframe=$((4 * 1024 * 1024)) --trace-children=yes --vgdb=no --error-exitcode=1 --suppressions=test/valgrind.supp ./passt -f ${__opts} -s ${STATESETUP}/passt.socket -H hostname1 --fqdn fqdn1.passt.test -t 10001,10011,10021,10031 -u 10001,10011,10021,10031 -P ${STATESETUP}/passt.pid --map-host-loopback ${__map_ns4} --map-host-loopback ${__map_ns6}" else context_run passt "make clean" context_run passt "make" - context_run_bg passt "./passt -f ${__opts} -s ${STATESETUP}/passt.socket -t 10001,10011,10021,10031 -u 10001,10011,10021,10031 -P ${STATESETUP}/passt.pid --map-host-loopback ${__map_ns4} --map-host-loopback ${__map_ns6}" + context_run_bg passt "./passt -f ${__opts} -s ${STATESETUP}/passt.socket -H hostname1 --fqdn fqdn1.passt.test -t 10001,10011,10021,10031 -u 10001,10011,10021,10031 -P ${STATESETUP}/passt.pid --map-host-loopback ${__map_ns4} --map-host-loopback ${__map_ns6}" fi wait_for [ -f "${STATESETUP}/passt.pid" ] + __vmem="$((${MEM_KIB} / 1024 / 4))" + if [ ${VHOST_USER} -eq 1 ]; then + __vmem="$(((${__vmem} + 500) / 1000))G" + __qemu_netdev=" \ + -chardev socket,id=c,path=${STATESETUP}/passt.socket \ + -netdev vhost-user,id=v,chardev=c \ + -device virtio-net,netdev=v \ + -object memory-backend-memfd,id=m,share=on,size=${__vmem} \ + -numa node,memdev=m" + else + __qemu_netdev="-device virtio-net-pci,netdev=s \ + -netdev stream,id=s,server=off,addr.type=unix,addr.path=${STATESETUP}/passt.socket" + fi + GUEST_CID=94557 context_run_bg qemu 'qemu-system-'"${QEMU_ARCH}" \ ' -machine accel=kvm' \ ' -M accel=kvm:tcg' \ - ' -m '${VMEM}' -cpu host -smp '${VCPUS} \ + ' -m '${__vmem}' -cpu host -smp '${VCPUS} \ ' -kernel '"${KERNEL}" \ ' -initrd '${INITRAMFS}' -nographic -serial stdio' \ ' -nodefaults' \ ' -append "console=ttyS0 mitigations=off apparmor=0" ' \ - ' -device virtio-net-pci,netdev=s0 ' \ - " -netdev stream,id=s0,server=off,addr.type=unix,addr.path=${STATESETUP}/passt.socket " \ + " ${__qemu_netdev}" \ " -pidfile ${STATESETUP}/qemu.pid" \ " -device vhost-vsock-pci,guest-cid=$GUEST_CID" @@ -214,11 +241,126 @@ setup_two_guests() { [ ${PCAP} -eq 1 ] && __opts="${__opts} -p ${LOGDIR}/passt_1.pcap" [ ${DEBUG} -eq 1 ] && __opts="${__opts} -d" [ ${TRACE} -eq 1 ] && __opts="${__opts} --trace" + [ ${VHOST_USER} -eq 1 ] && __opts="${__opts} --vhost-user" + + context_run_bg passt_1 "./passt -s ${STATESETUP}/passt_1.socket -P ${STATESETUP}/passt_1.pid -f ${__opts} --fqdn fqdn1.passt.test -H hostname1 -t 10001 -u 10001" + wait_for [ -f "${STATESETUP}/passt_1.pid" ] + + __opts= + [ ${PCAP} -eq 1 ] && __opts="${__opts} -p ${LOGDIR}/passt_2.pcap" + [ ${DEBUG} -eq 1 ] && __opts="${__opts} -d" + [ ${TRACE} -eq 1 ] && __opts="${__opts} --trace" + [ ${VHOST_USER} -eq 1 ] && __opts="${__opts} --vhost-user" + + context_run_bg passt_2 "./passt -s ${STATESETUP}/passt_2.socket -P ${STATESETUP}/passt_2.pid -f ${__opts} --hostname hostname2 --fqdn fqdn2 -t 10004 -u 10004" + wait_for [ -f "${STATESETUP}/passt_2.pid" ] + + __vmem="$((${MEM_KIB} / 1024 / 4))" + if [ ${VHOST_USER} -eq 1 ]; then + __vmem="$(((${__vmem} + 500) / 1000))G" + __qemu_netdev1=" \ + -chardev socket,id=c,path=${STATESETUP}/passt_1.socket \ + -netdev vhost-user,id=v,chardev=c \ + -device virtio-net,netdev=v \ + -object memory-backend-memfd,id=m,share=on,size=${__vmem} \ + -numa node,memdev=m" + __qemu_netdev2=" \ + -chardev socket,id=c,path=${STATESETUP}/passt_2.socket \ + -netdev vhost-user,id=v,chardev=c \ + -device virtio-net,netdev=v \ + -object memory-backend-memfd,id=m,share=on,size=${__vmem} \ + -numa node,memdev=m" + else + __qemu_netdev1="-device virtio-net-pci,netdev=s \ + -netdev stream,id=s,server=off,addr.type=unix,addr.path=${STATESETUP}/passt_1.socket" + __qemu_netdev2="-device virtio-net-pci,netdev=s \ + -netdev stream,id=s,server=off,addr.type=unix,addr.path=${STATESETUP}/passt_2.socket" + fi + + GUEST_1_CID=94557 + context_run_bg qemu_1 'qemu-system-'"${QEMU_ARCH}" \ + ' -M accel=kvm:tcg' \ + ' -m '${__vmem}' -cpu host -smp '${VCPUS} \ + ' -kernel '"${KERNEL}" \ + ' -initrd '${INITRAMFS}' -nographic -serial stdio' \ + ' -nodefaults' \ + ' -append "console=ttyS0 mitigations=off apparmor=0" ' \ + " ${__qemu_netdev1}" \ + " -pidfile ${STATESETUP}/qemu_1.pid" \ + " -device vhost-vsock-pci,guest-cid=$GUEST_1_CID" + + GUEST_2_CID=94558 + context_run_bg qemu_2 'qemu-system-'"${QEMU_ARCH}" \ + ' -M accel=kvm:tcg' \ + ' -m '${__vmem}' -cpu host -smp '${VCPUS} \ + ' -kernel '"${KERNEL}" \ + ' -initrd '${INITRAMFS}' -nographic -serial stdio' \ + ' -nodefaults' \ + ' -append "console=ttyS0 mitigations=off apparmor=0" ' \ + " ${__qemu_netdev2}" \ + " -pidfile ${STATESETUP}/qemu_2.pid" \ + " -device vhost-vsock-pci,guest-cid=$GUEST_2_CID" + + context_setup_guest guest_1 ${GUEST_1_CID} + context_setup_guest guest_2 ${GUEST_2_CID} +} + +# setup_migrate() - Set up two namespace, run qemu, passt/passt-repair in both +setup_migrate() { + context_setup_host host + context_setup_host mon + context_setup_host pasta_1 + context_setup_host pasta_2 + + layout_migrate + + # Ports: + # + # guest #1 | guest #2 | ns #1 | host + # --------- |-----------|-----------|------------ + # 10001 as server | | to guest | to ns #1 + # 10002 | | as server | to ns #1 + # 10003 | | to init | as server + # 10004 | as server | to guest | to ns #1 + + __opts= + [ ${PCAP} -eq 1 ] && __opts="${__opts} -p ${LOGDIR}/pasta_1.pcap" + [ ${DEBUG} -eq 1 ] && __opts="${__opts} -d" + [ ${TRACE} -eq 1 ] && __opts="${__opts} --trace" + + __map_host4=192.0.2.1 + __map_host6=2001:db8:9a55::1 + __map_ns4=192.0.2.2 + __map_ns6=2001:db8:9a55::2 + + # Option 1: send stuff via spliced path in pasta + # context_run_bg pasta_1 "./pasta ${__opts} -P ${STATESETUP}/pasta_1.pid -t 10001,10002 -T 10003 -u 10001,10002 -U 10003 --config-net ${NSTOOL} hold ${STATESETUP}/ns1.hold" + # Option 2: send stuff via tap (--map-guest-addr) instead (useful to see capture of full migration) + context_run_bg pasta_1 "./pasta ${__opts} -P ${STATESETUP}/pasta_1.pid -t 10001,10002,10004 -T 10003 -u 10001,10002,10004 -U 10003 --map-guest-addr ${__map_host4} --map-guest-addr ${__map_host6} --config-net ${NSTOOL} hold ${STATESETUP}/ns1.hold" + context_setup_nstool passt_1 ${STATESETUP}/ns1.hold + context_setup_nstool passt_repair_1 ${STATESETUP}/ns1.hold + + context_setup_nstool passt_2 ${STATESETUP}/ns1.hold + context_setup_nstool passt_repair_2 ${STATESETUP}/ns1.hold + + context_setup_nstool qemu_1 ${STATESETUP}/ns1.hold + context_setup_nstool qemu_2 ${STATESETUP}/ns1.hold + + __ifname="$(context_run qemu_1 "ip -j link show | jq -rM '.[] | select(.link_type == \"ether\").ifname'")" + + sleep 1 + + __opts="--vhost-user" + [ ${PCAP} -eq 1 ] && __opts="${__opts} -p ${LOGDIR}/passt_1.pcap" + [ ${DEBUG} -eq 1 ] && __opts="${__opts} -d" + [ ${TRACE} -eq 1 ] && __opts="${__opts} --trace" context_run_bg passt_1 "./passt -s ${STATESETUP}/passt_1.socket -P ${STATESETUP}/passt_1.pid -f ${__opts} -t 10001 -u 10001" wait_for [ -f "${STATESETUP}/passt_1.pid" ] - __opts= + context_run_bg passt_repair_1 "./passt-repair ${STATESETUP}/passt_1.socket.repair" + + __opts="--vhost-user" [ ${PCAP} -eq 1 ] && __opts="${__opts} -p ${LOGDIR}/passt_2.pcap" [ ${DEBUG} -eq 1 ] && __opts="${__opts} -d" [ ${TRACE} -eq 1 ] && __opts="${__opts} --trace" @@ -226,34 +368,52 @@ setup_two_guests() { context_run_bg passt_2 "./passt -s ${STATESETUP}/passt_2.socket -P ${STATESETUP}/passt_2.pid -f ${__opts} -t 10004 -u 10004" wait_for [ -f "${STATESETUP}/passt_2.pid" ] + context_run_bg passt_repair_2 "./passt-repair ${STATESETUP}/passt_2.socket.repair" + + __vmem="512M" # Keep migration fast + __qemu_netdev1=" \ + -chardev socket,id=c,path=${STATESETUP}/passt_1.socket \ + -netdev vhost-user,id=v,chardev=c \ + -device virtio-net,netdev=v \ + -object memory-backend-memfd,id=m,share=on,size=${__vmem} \ + -numa node,memdev=m" + __qemu_netdev2=" \ + -chardev socket,id=c,path=${STATESETUP}/passt_2.socket \ + -netdev vhost-user,id=v,chardev=c \ + -device virtio-net,netdev=v \ + -object memory-backend-memfd,id=m,share=on,size=${__vmem} \ + -numa node,memdev=m" + GUEST_1_CID=94557 context_run_bg qemu_1 'qemu-system-'"${QEMU_ARCH}" \ ' -M accel=kvm:tcg' \ - ' -m '${VMEM}' -cpu host -smp '${VCPUS} \ + ' -m '${__vmem}' -cpu host -smp '${VCPUS} \ ' -kernel '"${KERNEL}" \ ' -initrd '${INITRAMFS}' -nographic -serial stdio' \ ' -nodefaults' \ ' -append "console=ttyS0 mitigations=off apparmor=0" ' \ - ' -device virtio-net-pci,netdev=s0 ' \ - " -netdev stream,id=s0,server=off,addr.type=unix,addr.path=${STATESETUP}/passt_1.socket " \ + " ${__qemu_netdev1}" \ " -pidfile ${STATESETUP}/qemu_1.pid" \ - " -device vhost-vsock-pci,guest-cid=$GUEST_1_CID" + " -device vhost-vsock-pci,guest-cid=$GUEST_1_CID" \ + " -monitor unix:${STATESETUP}/qemu_1_mon.sock,server,nowait" GUEST_2_CID=94558 context_run_bg qemu_2 'qemu-system-'"${QEMU_ARCH}" \ ' -M accel=kvm:tcg' \ - ' -m '${VMEM}' -cpu host -smp '${VCPUS} \ + ' -m '${__vmem}' -cpu host -smp '${VCPUS} \ ' -kernel '"${KERNEL}" \ ' -initrd '${INITRAMFS}' -nographic -serial stdio' \ ' -nodefaults' \ ' -append "console=ttyS0 mitigations=off apparmor=0" ' \ - ' -device virtio-net-pci,netdev=s0 ' \ - " -netdev stream,id=s0,server=off,addr.type=unix,addr.path=${STATESETUP}/passt_2.socket " \ + " ${__qemu_netdev2}" \ " -pidfile ${STATESETUP}/qemu_2.pid" \ - " -device vhost-vsock-pci,guest-cid=$GUEST_2_CID" + " -device vhost-vsock-pci,guest-cid=$GUEST_2_CID" \ + " -monitor unix:${STATESETUP}/qemu_2_mon.sock,server,nowait" \ + " -incoming tcp:0:20005" context_setup_guest guest_1 ${GUEST_1_CID} - context_setup_guest guest_2 ${GUEST_2_CID} + # Only available after migration: + ( context_setup_guest guest_2 ${GUEST_2_CID} & ) } # teardown_context_watch() - Remove contexts and stop panes watching them @@ -326,7 +486,8 @@ teardown_two_guests() { context_wait pasta_1 context_wait pasta_2 - rm -f "${STATESETUP}/passt__[12].pid" "${STATESETUP}/pasta_[12].pid" + rm "${STATESETUP}/passt_1.pid" "${STATESETUP}/passt_2.pid" + rm "${STATESETUP}/pasta_1.pid" "${STATESETUP}/pasta_2.pid" teardown_context_watch ${PANE_HOST} host teardown_context_watch ${PANE_GUEST_1} qemu_1 guest_1 @@ -335,6 +496,30 @@ teardown_two_guests() { teardown_context_watch ${PANE_PASST_2} pasta_2 passt_2 } +# teardown_migrate() - Exit namespaces, kill qemu processes, passt and pasta +teardown_migrate() { + ${NSTOOL} exec ${STATESETUP}/ns1.hold -- kill $(cat "${STATESETUP}/qemu_1.pid") + ${NSTOOL} exec ${STATESETUP}/ns1.hold -- kill $(cat "${STATESETUP}/qemu_2.pid") + context_wait qemu_1 + context_wait qemu_2 + + ${NSTOOL} exec ${STATESETUP}/ns1.hold -- kill $(cat "${STATESETUP}/passt_2.pid") + context_wait passt_1 + context_wait passt_2 + ${NSTOOL} stop "${STATESETUP}/ns1.hold" + context_wait pasta_1 + + rm -f "${STATESETUP}/passt_1.pid" "${STATESETUP}/passt_2.pid" + rm -f "${STATESETUP}/pasta_1.pid" "${STATESETUP}/pasta_2.pid" + + teardown_context_watch ${PANE_HOST} host + + teardown_context_watch ${PANE_GUEST_1} qemu_1 guest_1 + teardown_context_watch ${PANE_GUEST_2} qemu_2 guest_2 + teardown_context_watch ${PANE_PASST_1} pasta_1 passt_1 + teardown_context_watch ${PANE_PASST_2} pasta_1 passt_2 +} + # teardown_demo_passt() - Exit namespace, kill qemu, passt and pasta teardown_demo_passt() { tmux send-keys -t ${PANE_GUEST} "C-c" diff --git a/test/lib/setup_ugly b/test/lib/setup_ugly index 4b2a077..2802cc3 100755 --- a/test/lib/setup_ugly +++ b/test/lib/setup_ugly @@ -33,7 +33,7 @@ setup_memory() { pane_or_context_run guest 'qemu-system-$(uname -m)' \ ' -machine accel=kvm' \ - ' -m '${VMEM}' -cpu host -smp '${VCPUS} \ + ' -m '$((${MEM_KIB} / 1024 / 4))' -cpu host -smp '${VCPUS} \ ' -kernel ' "/boot/vmlinuz-$(uname -r)" \ ' -initrd '${INITRAMFS_MEM}' -nographic -serial stdio' \ ' -nodefaults' \ diff --git a/test/lib/term b/test/lib/term index ed690de..089364c 100755 --- a/test/lib/term +++ b/test/lib/term @@ -19,6 +19,7 @@ STATUS_FILE_INDEX=0 STATUS_COLS= STATUS_PASS=0 STATUS_FAIL=0 +STATUS_SKIPPED=0 PR_RED='\033[1;31m' PR_GREEN='\033[1;32m' @@ -439,19 +440,21 @@ info_layout() { # status_test_ok() - Update counter of passed tests, log and display message status_test_ok() { STATUS_PASS=$((STATUS_PASS + 1)) - tmux set status-right "PASS: ${STATUS_PASS} | FAIL: ${STATUS_FAIL} | #(TZ="UTC" date -Iseconds)" + tmux set status-right "PASS: ${STATUS_PASS} | FAIL: ${STATUS_FAIL} | SKIPPED: ${STATUS_SKIPPED} | #(TZ="UTC" date -Iseconds)" info_passed } # status_test_fail() - Update counter of failed tests, log and display message status_test_fail() { STATUS_FAIL=$((STATUS_FAIL + 1)) - tmux set status-right "PASS: ${STATUS_PASS} | FAIL: ${STATUS_FAIL} | #(TZ="UTC" date -Iseconds)" + tmux set status-right "PASS: ${STATUS_PASS} | FAIL: ${STATUS_FAIL} | SKIPPED: ${STATUS_SKIPPED} | #(TZ="UTC" date -Iseconds)" info_failed } # status_test_fail() - Update counter of failed tests, log and display message status_test_skip() { + STATUS_SKIPPED=$((STATUS_SKIPPED + 1)) + tmux set status-right "PASS: ${STATUS_PASS} | FAIL: ${STATUS_FAIL} | SKIPPED: ${STATUS_SKIPPED} | #(TZ="UTC" date -Iseconds)" info_skipped } diff --git a/test/lib/test b/test/lib/test index e6726be..7349674 100755 --- a/test/lib/test +++ b/test/lib/test @@ -20,10 +20,7 @@ test_iperf3s() { __sctx="${1}" __port="${2}" - pane_or_context_run_bg "${__sctx}" \ - 'iperf3 -s -p'${__port}' & echo $! > s.pid' \ - - sleep 1 # Wait for server to be ready + pane_or_context_run "${__sctx}" 'iperf3 -s -p'${__port}' -D -I s.pid' } # test_iperf3k() - Kill iperf3 server @@ -31,7 +28,7 @@ test_iperf3s() { test_iperf3k() { __sctx="${1}" - pane_or_context_run "${__sctx}" 'kill -INT $(cat s.pid); rm s.pid' + pane_or_context_run "${__sctx}" 'kill -INT $(cat s.pid)' sleep 1 # Wait for kernel to free up ports } @@ -68,6 +65,45 @@ test_iperf3() { TEST_ONE_subs="$(list_add_pair "${TEST_ONE_subs}" "__${__var}__" "${__bw}" )" } +# test_iperf3m() - Ugly helper for iperf3 directive, guest migration variant +# $1: Variable name: to put the measure bandwidth into +# $2: Initial source/client context +# $3: Second source/client context the guest is moving to +# $4: Destination name or address for client +# $5: Port number, ${i} is translated to process index +# $6: Run time, in seconds +# $7: Client options +test_iperf3m() { + __var="${1}"; shift + __cctx="${1}"; shift + __cctx2="${1}"; shift + __dest="${1}"; shift + __port="${1}"; shift + __time="${1}"; shift + + pane_or_context_run "${__cctx}" 'rm -f c.json' + + # A 1s wait for connection on what's basically a local link + # indicates something is pretty wrong + __timeout=1000 + pane_or_context_run_bg "${__cctx}" \ + 'iperf3 -J -c '${__dest}' -p '${__port} \ + ' --connect-timeout '${__timeout} \ + ' -t'${__time}' -i0 '"${@}"' > c.json' \ + + __jval=".end.sum_received.bits_per_second" + + sleep $((${__time} + 3)) + + pane_or_context_output "${__cctx2}" \ + 'cat c.json' + + __bw=$(pane_or_context_output "${__cctx2}" \ + 'cat c.json | jq -rMs "map('${__jval}') | add"') + + TEST_ONE_subs="$(list_add_pair "${TEST_ONE_subs}" "__${__var}__" "${__bw}" )" +} + test_one_line() { __line="${1}" @@ -177,6 +213,12 @@ test_one_line() { "guest2w") pane_or_context_wait guest_2 || TEST_ONE_nok=1 ;; + "mon") + pane_or_context_run mon "${__arg}" || TEST_ONE_nok=1 + ;; + "monb") + pane_or_context_run_bg mon "${__arg}" + ;; "ns") pane_or_context_run ns "${__arg}" || TEST_ONE_nok=1 ;; @@ -292,6 +334,9 @@ test_one_line() { "iperf3") test_iperf3 ${__arg} ;; + "iperf3m") + test_iperf3m ${__arg} + ;; "set") TEST_ONE_subs="$(list_add_pair "${TEST_ONE_subs}" "__${__arg%% *}__" "${__arg#* }")" ;; diff --git a/test/migrate/basic b/test/migrate/basic new file mode 100644 index 0000000..3f11f7d --- /dev/null +++ b/test/migrate/basic @@ -0,0 +1,59 @@ +# SPDX-License-Identifier: GPL-2.0-or-later +# +# PASST - Plug A Simple Socket Transport +# for qemu/UNIX domain socket mode +# +# PASTA - Pack A Subtle Tap Abstraction +# for network namespace/tap device mode +# +# test/migrate/basic - Check basic migration functionality +# +# Copyright (c) 2025 Red Hat GmbH +# Author: Stefano Brivio <sbrivio@redhat.com> + +g1tools ip jq dhclient socat cat +htools ip jq + +set MAP_HOST4 192.0.2.1 +set MAP_HOST6 2001:db8:9a55::1 +set MAP_NS4 192.0.2.2 +set MAP_NS6 2001:db8:9a55::2 + +test Interface name +g1out IFNAME1 ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname' +hout HOST_IFNAME ip -j -4 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +hout HOST_IFNAME6 ip -j -6 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +check [ -n "__IFNAME1__" ] + +test DHCP: address +guest1 ip link set dev __IFNAME1__ up +guest1 /sbin/dhclient -4 __IFNAME1__ +g1out ADDR1 ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__IFNAME1__").addr_info[0].local' +hout HOST_ADDR ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__HOST_IFNAME__").addr_info[0].local' +check [ "__ADDR1__" = "__HOST_ADDR__" ] + +test DHCPv6: address +# Link is up now, wait for DAD to complete +guest1 while ip -j -6 addr show tentative | jq -e '.[].addr_info'; do sleep 0.1; done +guest1 /sbin/dhclient -6 __IFNAME1__ +# Wait for DAD to complete on the DHCP address +guest1 while ip -j -6 addr show tentative | jq -e '.[].addr_info'; do sleep 0.1; done +g1out ADDR1_6 ip -j -6 addr show|jq -rM '[.[] | select(.ifname == "__IFNAME1__").addr_info[] | select(.prefixlen == 128).local] | .[0]' +hout HOST_ADDR6 ip -j -6 addr show|jq -rM '[.[] | select(.ifname == "__HOST_IFNAME6__").addr_info[] | select(.scope == "global" and .deprecated != true).local] | .[0]' +check [ "__ADDR1_6__" = "__HOST_ADDR6__" ] + +test TCP/IPv4: guest1/guest2 > host +g1out GW1 ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway' +hostb socat -u TCP4-LISTEN:10006 OPEN:__STATESETUP__/msg,create,trunc +sleep 1 +# Option 1: via spliced path in pasta, namespace to host +# guest1b { printf "Hello from guest 1"; sleep 10; printf " and from guest 2\n"; } | socat -u STDIN TCP4:__GW1__:10003 +# Option 2: via --map-guest-addr (tap) in pasta, namespace to host +guest1b { printf "Hello from guest 1"; sleep 3; printf " and from guest 2\n"; } | socat -u STDIN TCP4:__MAP_HOST4__:10006 +sleep 1 + +mon echo "migrate tcp:0:20005" | socat -u STDIN UNIX:__STATESETUP__/qemu_1_mon.sock + +hostw +hout MSG cat __STATESETUP__/msg +check [ "__MSG__" = "Hello from guest 1 and from guest 2" ] diff --git a/test/migrate/basic_fin b/test/migrate/basic_fin new file mode 100644 index 0000000..aa61ec5 --- /dev/null +++ b/test/migrate/basic_fin @@ -0,0 +1,62 @@ +# SPDX-License-Identifier: GPL-2.0-or-later +# +# PASST - Plug A Simple Socket Transport +# for qemu/UNIX domain socket mode +# +# PASTA - Pack A Subtle Tap Abstraction +# for network namespace/tap device mode +# +# test/migrate/basic_fin - Outbound traffic across migration, half-closed socket +# +# Copyright (c) 2025 Red Hat GmbH +# Author: Stefano Brivio <sbrivio@redhat.com> + +g1tools ip jq dhclient socat cat +htools ip jq + +set MAP_HOST4 192.0.2.1 +set MAP_HOST6 2001:db8:9a55::1 +set MAP_NS4 192.0.2.2 +set MAP_NS6 2001:db8:9a55::2 + +test Interface name +g1out IFNAME1 ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname' +hout HOST_IFNAME ip -j -4 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +hout HOST_IFNAME6 ip -j -6 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +check [ -n "__IFNAME1__" ] + +test DHCP: address +guest1 ip link set dev __IFNAME1__ up +guest1 /sbin/dhclient -4 __IFNAME1__ +g1out ADDR1 ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__IFNAME1__").addr_info[0].local' +hout HOST_ADDR ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__HOST_IFNAME__").addr_info[0].local' +check [ "__ADDR1__" = "__HOST_ADDR__" ] + +test DHCPv6: address +# Link is up now, wait for DAD to complete +guest1 while ip -j -6 addr show tentative | jq -e '.[].addr_info'; do sleep 0.1; done +guest1 /sbin/dhclient -6 __IFNAME1__ +# Wait for DAD to complete on the DHCP address +guest1 while ip -j -6 addr show tentative | jq -e '.[].addr_info'; do sleep 0.1; done +g1out ADDR1_6 ip -j -6 addr show|jq -rM '[.[] | select(.ifname == "__IFNAME1__").addr_info[] | select(.prefixlen == 128).local] | .[0]' +hout HOST_ADDR6 ip -j -6 addr show|jq -rM '[.[] | select(.ifname == "__HOST_IFNAME6__").addr_info[] | select(.scope == "global" and .deprecated != true).local] | .[0]' +check [ "__ADDR1_6__" = "__HOST_ADDR6__" ] + +test TCP/IPv4: guest1, half-close, guest2 > host +g1out GW1 ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway' + +hostb echo FIN | socat TCP4-LISTEN:10006,shut-down STDIO,ignoreeof > __STATESETUP__/msg +#hostb socat -u TCP4-LISTEN:10006 OPEN:__STATESETUP__/msg,create,trunc + +#sleep 20 +# Option 1: via spliced path in pasta, namespace to host +# guest1b { printf "Hello from guest 1"; sleep 10; printf " and from guest 2\n"; } | socat -u STDIN TCP4:__GW1__:10003 +# Option 2: via --map-guest-addr (tap) in pasta, namespace to host +guest1b { printf "Hello from guest 1"; sleep 3; printf " and from guest 2\n"; } | socat -u STDIN TCP4:__MAP_HOST4__:10006 +sleep 1 + +mon echo "migrate tcp:0:20005" | socat -u STDIN UNIX:__STATESETUP__/qemu_1_mon.sock + +hostw +hout MSG cat __STATESETUP__/msg +check [ "__MSG__" = "Hello from guest 1 and from guest 2" ] diff --git a/test/migrate/bidirectional b/test/migrate/bidirectional new file mode 100644 index 0000000..4c04081 --- /dev/null +++ b/test/migrate/bidirectional @@ -0,0 +1,64 @@ +# SPDX-License-Identifier: GPL-2.0-or-later +# +# PASST - Plug A Simple Socket Transport +# for qemu/UNIX domain socket mode +# +# PASTA - Pack A Subtle Tap Abstraction +# for network namespace/tap device mode +# +# test/migrate/bidirectional - Check migration with messages in both directions +# +# Copyright (c) 2025 Red Hat GmbH +# Author: Stefano Brivio <sbrivio@redhat.com> + +g1tools ip jq dhclient socat cat +htools ip jq + +set MAP_HOST4 192.0.2.1 +set MAP_HOST6 2001:db8:9a55::1 +set MAP_NS4 192.0.2.2 +set MAP_NS6 2001:db8:9a55::2 + +test Interface name +g1out IFNAME1 ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname' +hout HOST_IFNAME ip -j -4 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +hout HOST_IFNAME6 ip -j -6 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +check [ -n "__IFNAME1__" ] + +test DHCP: address +guest1 ip link set dev __IFNAME1__ up +guest1 /sbin/dhclient -4 __IFNAME1__ +g1out ADDR1 ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__IFNAME1__").addr_info[0].local' +hout HOST_ADDR ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__HOST_IFNAME__").addr_info[0].local' +check [ "__ADDR1__" = "__HOST_ADDR__" ] + +test TCP/IPv4: guest1/guest2 > host, host > guest1/guest2 +g1out GW1 ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway' + +hostb socat -u TCP4-LISTEN:10006 OPEN:__STATESETUP__/msg,create,trunc +guest1b socat -u TCP4-LISTEN:10001 OPEN:msg,create,trunc +sleep 1 + +guest1b socat -u UNIX-RECV:proxy.sock,null-eof TCP4:__MAP_HOST4__:10006 +hostb socat -u UNIX-RECV:__STATESETUP__/proxy.sock,null-eof TCP4:__ADDR1__:10001 +sleep 1 +guest1 printf "Hello from guest 1" | socat -u STDIN UNIX:proxy.sock +host printf "Dear guest 1," | socat -u STDIN UNIX:__STATESETUP__/proxy.sock +sleep 1 + +mon echo "migrate tcp:0:20005" | socat -u STDIN UNIX:__STATESETUP__/qemu_1_mon.sock + +sleep 2 +guest2 printf " and from guest 2" | socat -u STDIN UNIX:proxy.sock,shut-null +host printf " you are now guest 2" | socat -u STDIN UNIX:__STATESETUP__/proxy.sock,shut-null + +hostw +# FIXME: guest2w doesn't work here because shell jobs are (also) from guest #1, +# use sleep 1 for the moment +sleep 1 + +hout MSG cat __STATESETUP__/msg +check [ "__MSG__" = "Hello from guest 1 and from guest 2" ] + +g2out MSG cat msg +check [ "__MSG__" = "Dear guest 1, you are now guest 2" ] diff --git a/test/migrate/bidirectional_fin b/test/migrate/bidirectional_fin new file mode 100644 index 0000000..1c13527 --- /dev/null +++ b/test/migrate/bidirectional_fin @@ -0,0 +1,64 @@ +# SPDX-License-Identifier: GPL-2.0-or-later +# +# PASST - Plug A Simple Socket Transport +# for qemu/UNIX domain socket mode +# +# PASTA - Pack A Subtle Tap Abstraction +# for network namespace/tap device mode +# +# test/migrate/bidirectional_fin - Both directions, half-closed sockets +# +# Copyright (c) 2025 Red Hat GmbH +# Author: Stefano Brivio <sbrivio@redhat.com> + +g1tools ip jq dhclient socat cat +htools ip jq + +set MAP_HOST4 192.0.2.1 +set MAP_HOST6 2001:db8:9a55::1 +set MAP_NS4 192.0.2.2 +set MAP_NS6 2001:db8:9a55::2 + +test Interface name +g1out IFNAME1 ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname' +hout HOST_IFNAME ip -j -4 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +hout HOST_IFNAME6 ip -j -6 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +check [ -n "__IFNAME1__" ] + +test DHCP: address +guest1 ip link set dev __IFNAME1__ up +guest1 /sbin/dhclient -4 __IFNAME1__ +g1out ADDR1 ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__IFNAME1__").addr_info[0].local' +hout HOST_ADDR ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__HOST_IFNAME__").addr_info[0].local' +check [ "__ADDR1__" = "__HOST_ADDR__" ] + +test TCP/IPv4: guest1/guest2 <- (half closed) -> host +g1out GW1 ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway' + +hostb echo FIN | socat TCP4-LISTEN:10006,shut-down STDIO,ignoreeof > __STATESETUP__/msg +guest1b echo FIN | socat TCP4-LISTEN:10001,shut-down STDIO,ignoreeof > msg +sleep 1 + +guest1b socat -u UNIX-RECV:proxy.sock,null-eof TCP4:__MAP_HOST4__:10006 +hostb socat -u UNIX-RECV:__STATESETUP__/proxy.sock,null-eof TCP4:__ADDR1__:10001 +sleep 1 +guest1 printf "Hello from guest 1" | socat -u STDIN UNIX:proxy.sock +host printf "Dear guest 1," | socat -u STDIN UNIX:__STATESETUP__/proxy.sock +sleep 1 + +mon echo "migrate tcp:0:20005" | socat -u STDIN UNIX:__STATESETUP__/qemu_1_mon.sock + +sleep 2 +guest2 printf " and from guest 2" | socat -u STDIN UNIX:proxy.sock,shut-null +host printf " you are now guest 2" | socat -u STDIN UNIX:__STATESETUP__/proxy.sock,shut-null + +hostw +# FIXME: guest2w doesn't work here because shell jobs are (also) from guest #1, +# use sleep 1 for the moment +sleep 1 + +hout MSG cat __STATESETUP__/msg +check [ "__MSG__" = "Hello from guest 1 and from guest 2" ] + +g2out MSG cat msg +check [ "__MSG__" = "Dear guest 1, you are now guest 2" ] diff --git a/test/migrate/iperf3_bidir6 b/test/migrate/iperf3_bidir6 new file mode 100644 index 0000000..4bfefb5 --- /dev/null +++ b/test/migrate/iperf3_bidir6 @@ -0,0 +1,58 @@ +# SPDX-License-Identifier: GPL-2.0-or-later +# +# PASST - Plug A Simple Socket Transport +# for qemu/UNIX domain socket mode +# +# PASTA - Pack A Subtle Tap Abstraction +# for network namespace/tap device mode +# +# test/migrate/iperf3_bidir6 - Migration behaviour with many bidirectional flows +# +# Copyright (c) 2025 Red Hat GmbH +# Author: Stefano Brivio <sbrivio@redhat.com> + +g1tools ip jq dhclient socat cat +htools ip jq + +set MAP_HOST4 192.0.2.1 +set MAP_HOST6 2001:db8:9a55::1 +set MAP_NS4 192.0.2.2 +set MAP_NS6 2001:db8:9a55::2 + +set THREADS 128 +set TIME 3 +set OMIT 0.1 +set OPTS -Z -P __THREADS__ -O__OMIT__ -N --bidir + +test Interface name +g1out IFNAME1 ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname' +hout HOST_IFNAME ip -j -4 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +hout HOST_IFNAME6 ip -j -6 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +check [ -n "__IFNAME1__" ] + +test DHCP: address +guest1 ip link set dev __IFNAME1__ up +guest1 /sbin/dhclient -4 __IFNAME1__ +g1out ADDR1 ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__IFNAME1__").addr_info[0].local' +hout HOST_ADDR ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__HOST_IFNAME__").addr_info[0].local' +check [ "__ADDR1__" = "__HOST_ADDR__" ] + +test DHCPv6: address +# Link is up now, wait for DAD to complete +guest1 while ip -j -6 addr show tentative | jq -e '.[].addr_info'; do sleep 0.1; done +guest1 /sbin/dhclient -6 __IFNAME1__ +# Wait for DAD to complete on the DHCP address +guest1 while ip -j -6 addr show tentative | jq -e '.[].addr_info'; do sleep 0.1; done +g1out ADDR1_6 ip -j -6 addr show|jq -rM '[.[] | select(.ifname == "__IFNAME1__").addr_info[] | select(.prefixlen == 128).local] | .[0]' +hout HOST_ADDR6 ip -j -6 addr show|jq -rM '[.[] | select(.ifname == "__HOST_IFNAME6__").addr_info[] | select(.scope == "global" and .deprecated != true).local] | .[0]' +check [ "__ADDR1_6__" = "__HOST_ADDR6__" ] + +test TCP/IPv6 host <-> guest flood, many flows, during migration + +monb sleep 1; echo "migrate tcp:0:20005" | socat -u STDIN UNIX:__STATESETUP__/qemu_1_mon.sock + +iperf3s host 10006 +iperf3m BW guest_1 guest_2 __MAP_HOST6__ 10006 __TIME__ __OPTS__ +bw __BW__ 1 2 + +iperf3k host diff --git a/test/migrate/iperf3_in4 b/test/migrate/iperf3_in4 new file mode 100644 index 0000000..c5f3916 --- /dev/null +++ b/test/migrate/iperf3_in4 @@ -0,0 +1,50 @@ +# SPDX-License-Identifier: GPL-2.0-or-later +# +# PASST - Plug A Simple Socket Transport +# for qemu/UNIX domain socket mode +# +# PASTA - Pack A Subtle Tap Abstraction +# for network namespace/tap device mode +# +# test/migrate/iperf3_in4 - Migration behaviour under inbound IPv4 flood +# +# Copyright (c) 2025 Red Hat GmbH +# Author: Stefano Brivio <sbrivio@redhat.com> + +g1tools ip jq dhclient socat cat +htools ip jq + +set MAP_HOST4 192.0.2.1 +set MAP_HOST6 2001:db8:9a55::1 +set MAP_NS4 192.0.2.2 +set MAP_NS6 2001:db8:9a55::2 + +guest1 /sbin/sysctl -w net.core.rmem_max=33554432 +guest1 /sbin/sysctl -w net.core.wmem_max=33554432 + +set THREADS 1 +set TIME 4 +set OMIT 0.1 +set OPTS -Z -P __THREADS__ -O__OMIT__ -N -R + +test Interface name +g1out IFNAME1 ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname' +hout HOST_IFNAME ip -j -4 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +check [ -n "__IFNAME1__" ] + +test DHCP: address +guest1 ip link set dev __IFNAME1__ up +guest1 /sbin/dhclient -4 __IFNAME1__ +g1out ADDR1 ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__IFNAME1__").addr_info[0].local' +hout HOST_ADDR ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__HOST_IFNAME__").addr_info[0].local' +check [ "__ADDR1__" = "__HOST_ADDR__" ] + +test TCP/IPv4 host to guest throughput during migration + +monb sleep 1; echo "migrate tcp:0:20005" | socat -u STDIN UNIX:__STATESETUP__/qemu_1_mon.sock + +iperf3s host 10006 +iperf3m BW guest_1 guest_2 __MAP_HOST4__ 10006 __TIME__ __OPTS__ +bw __BW__ 1 2 + +iperf3k host diff --git a/test/migrate/iperf3_in6 b/test/migrate/iperf3_in6 new file mode 100644 index 0000000..16cf504 --- /dev/null +++ b/test/migrate/iperf3_in6 @@ -0,0 +1,58 @@ +# SPDX-License-Identifier: GPL-2.0-or-later +# +# PASST - Plug A Simple Socket Transport +# for qemu/UNIX domain socket mode +# +# PASTA - Pack A Subtle Tap Abstraction +# for network namespace/tap device mode +# +# test/migrate/iperf3_in6 - Migration behaviour under inbound IPv6 flood +# +# Copyright (c) 2025 Red Hat GmbH +# Author: Stefano Brivio <sbrivio@redhat.com> + +g1tools ip jq dhclient socat cat +htools ip jq + +set MAP_HOST4 192.0.2.1 +set MAP_HOST6 2001:db8:9a55::1 +set MAP_NS4 192.0.2.2 +set MAP_NS6 2001:db8:9a55::2 + +set THREADS 4 +set TIME 3 +set OMIT 0.1 +set OPTS -Z -P __THREADS__ -O__OMIT__ -N -R + +test Interface name +g1out IFNAME1 ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname' +hout HOST_IFNAME ip -j -4 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +hout HOST_IFNAME6 ip -j -6 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +check [ -n "__IFNAME1__" ] + +test DHCP: address +guest1 ip link set dev __IFNAME1__ up +guest1 /sbin/dhclient -4 __IFNAME1__ +g1out ADDR1 ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__IFNAME1__").addr_info[0].local' +hout HOST_ADDR ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__HOST_IFNAME__").addr_info[0].local' +check [ "__ADDR1__" = "__HOST_ADDR__" ] + +test DHCPv6: address +# Link is up now, wait for DAD to complete +guest1 while ip -j -6 addr show tentative | jq -e '.[].addr_info'; do sleep 0.1; done +guest1 /sbin/dhclient -6 __IFNAME1__ +# Wait for DAD to complete on the DHCP address +guest1 while ip -j -6 addr show tentative | jq -e '.[].addr_info'; do sleep 0.1; done +g1out ADDR1_6 ip -j -6 addr show|jq -rM '[.[] | select(.ifname == "__IFNAME1__").addr_info[] | select(.prefixlen == 128).local] | .[0]' +hout HOST_ADDR6 ip -j -6 addr show|jq -rM '[.[] | select(.ifname == "__HOST_IFNAME6__").addr_info[] | select(.scope == "global" and .deprecated != true).local] | .[0]' +check [ "__ADDR1_6__" = "__HOST_ADDR6__" ] + +test TCP/IPv6 host to guest throughput during migration + +monb sleep 1; echo "migrate tcp:0:20005" | socat -u STDIN UNIX:__STATESETUP__/qemu_1_mon.sock + +iperf3s host 10006 +iperf3m BW guest_1 guest_2 __MAP_HOST6__ 10006 __TIME__ __OPTS__ +bw __BW__ 1 2 + +iperf3k host diff --git a/test/migrate/iperf3_many_out6 b/test/migrate/iperf3_many_out6 new file mode 100644 index 0000000..88133f2 --- /dev/null +++ b/test/migrate/iperf3_many_out6 @@ -0,0 +1,60 @@ +# SPDX-License-Identifier: GPL-2.0-or-later +# +# PASST - Plug A Simple Socket Transport +# for qemu/UNIX domain socket mode +# +# PASTA - Pack A Subtle Tap Abstraction +# for network namespace/tap device mode +# +# test/migrate/iperf3_many_out6 - Migration behaviour with many outbound flows +# +# Copyright (c) 2025 Red Hat GmbH +# Author: Stefano Brivio <sbrivio@redhat.com> + +g1tools ip jq dhclient socat cat +htools ip jq + +set MAP_HOST4 192.0.2.1 +set MAP_HOST6 2001:db8:9a55::1 +set MAP_NS4 192.0.2.2 +set MAP_NS6 2001:db8:9a55::2 + +set THREADS 16 +set TIME 3 +set OMIT 0.1 +set OPTS -Z -P __THREADS__ -O__OMIT__ -N -l 1M + +test Interface name +g1out IFNAME1 ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname' +hout HOST_IFNAME ip -j -4 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +hout HOST_IFNAME6 ip -j -6 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +check [ -n "__IFNAME1__" ] + +test DHCP: address +guest1 ip link set dev __IFNAME1__ up +guest1 /sbin/dhclient -4 __IFNAME1__ +g1out ADDR1 ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__IFNAME1__").addr_info[0].local' +hout HOST_ADDR ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__HOST_IFNAME__").addr_info[0].local' +check [ "__ADDR1__" = "__HOST_ADDR__" ] + +test DHCPv6: address +# Link is up now, wait for DAD to complete +guest1 while ip -j -6 addr show tentative | jq -e '.[].addr_info'; do sleep 0.1; done +guest1 /sbin/dhclient -6 __IFNAME1__ +# Wait for DAD to complete on the DHCP address +guest1 while ip -j -6 addr show tentative | jq -e '.[].addr_info'; do sleep 0.1; done +g1out ADDR1_6 ip -j -6 addr show|jq -rM '[.[] | select(.ifname == "__IFNAME1__").addr_info[] | select(.prefixlen == 128).local] | .[0]' +hout HOST_ADDR6 ip -j -6 addr show|jq -rM '[.[] | select(.ifname == "__HOST_IFNAME6__").addr_info[] | select(.scope == "global" and .deprecated != true).local] | .[0]' +check [ "__ADDR1_6__" = "__HOST_ADDR6__" ] + +test TCP/IPv6 guest to host flood, many flows, during migration + +test TCP/IPv6 host to guest throughput during migration + +monb sleep 1; echo "migrate tcp:0:20005" | socat -u STDIN UNIX:__STATESETUP__/qemu_1_mon.sock + +iperf3s host 10006 +iperf3m BW guest_1 guest_2 __MAP_HOST6__ 10006 __TIME__ __OPTS__ +bw __BW__ 1 2 + +iperf3k host diff --git a/test/migrate/iperf3_out4 b/test/migrate/iperf3_out4 new file mode 100644 index 0000000..968057b --- /dev/null +++ b/test/migrate/iperf3_out4 @@ -0,0 +1,47 @@ +# SPDX-License-Identifier: GPL-2.0-or-later +# +# PASST - Plug A Simple Socket Transport +# for qemu/UNIX domain socket mode +# +# PASTA - Pack A Subtle Tap Abstraction +# for network namespace/tap device mode +# +# test/migrate/iperf3_out4 - Migration behaviour under outbound IPv4 flood +# +# Copyright (c) 2025 Red Hat GmbH +# Author: Stefano Brivio <sbrivio@redhat.com> + +g1tools ip jq dhclient socat cat +htools ip jq + +set MAP_HOST4 192.0.2.1 +set MAP_HOST6 2001:db8:9a55::1 +set MAP_NS4 192.0.2.2 +set MAP_NS6 2001:db8:9a55::2 + +set THREADS 6 +set TIME 2 +set OMIT 0.1 +set OPTS -P __THREADS__ -O__OMIT__ -Z -N -l 1M + +test Interface name +g1out IFNAME1 ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname' +hout HOST_IFNAME ip -j -4 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +check [ -n "__IFNAME1__" ] + +test DHCP: address +guest1 ip link set dev __IFNAME1__ up +guest1 /sbin/dhclient -4 __IFNAME1__ +g1out ADDR1 ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__IFNAME1__").addr_info[0].local' +hout HOST_ADDR ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__HOST_IFNAME__").addr_info[0].local' +check [ "__ADDR1__" = "__HOST_ADDR__" ] + +test TCP/IPv4 guest to host throughput during migration + +monb sleep 1; echo "migrate tcp:0:20005" | socat -u STDIN UNIX:__STATESETUP__/qemu_1_mon.sock + +iperf3s host 10006 +iperf3m BW guest_1 guest_2 __MAP_HOST4__ 10006 __TIME__ __OPTS__ +bw __BW__ 1 2 + +iperf3k host diff --git a/test/migrate/iperf3_out6 b/test/migrate/iperf3_out6 new file mode 100644 index 0000000..21fbfcd --- /dev/null +++ b/test/migrate/iperf3_out6 @@ -0,0 +1,58 @@ +# SPDX-License-Identifier: GPL-2.0-or-later +# +# PASST - Plug A Simple Socket Transport +# for qemu/UNIX domain socket mode +# +# PASTA - Pack A Subtle Tap Abstraction +# for network namespace/tap device mode +# +# test/migrate/iperf3_out6 - Migration behaviour under outbound IPv6 flood +# +# Copyright (c) 2025 Red Hat GmbH +# Author: Stefano Brivio <sbrivio@redhat.com> + +g1tools ip jq dhclient socat cat +htools ip jq + +set MAP_HOST4 192.0.2.1 +set MAP_HOST6 2001:db8:9a55::1 +set MAP_NS4 192.0.2.2 +set MAP_NS6 2001:db8:9a55::2 + +set THREADS 6 +set TIME 2 +set OMIT 0.1 +set OPTS -P __THREADS__ -O__OMIT__ -Z -N -l 1M + +test Interface name +g1out IFNAME1 ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname' +hout HOST_IFNAME ip -j -4 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +hout HOST_IFNAME6 ip -j -6 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +check [ -n "__IFNAME1__" ] + +test DHCP: address +guest1 ip link set dev __IFNAME1__ up +guest1 /sbin/dhclient -4 __IFNAME1__ +g1out ADDR1 ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__IFNAME1__").addr_info[0].local' +hout HOST_ADDR ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__HOST_IFNAME__").addr_info[0].local' +check [ "__ADDR1__" = "__HOST_ADDR__" ] + +test DHCPv6: address +# Link is up now, wait for DAD to complete +guest1 while ip -j -6 addr show tentative | jq -e '.[].addr_info'; do sleep 0.1; done +guest1 /sbin/dhclient -6 __IFNAME1__ +# Wait for DAD to complete on the DHCP address +guest1 while ip -j -6 addr show tentative | jq -e '.[].addr_info'; do sleep 0.1; done +g1out ADDR1_6 ip -j -6 addr show|jq -rM '[.[] | select(.ifname == "__IFNAME1__").addr_info[] | select(.prefixlen == 128).local] | .[0]' +hout HOST_ADDR6 ip -j -6 addr show|jq -rM '[.[] | select(.ifname == "__HOST_IFNAME6__").addr_info[] | select(.scope == "global" and .deprecated != true).local] | .[0]' +check [ "__ADDR1_6__" = "__HOST_ADDR6__" ] + +test TCP/IPv6 guest to host throughput during migration + +monb sleep 1; echo "migrate tcp:0:20005" | socat -u STDIN UNIX:__STATESETUP__/qemu_1_mon.sock + +iperf3s host 10006 +iperf3m BW guest_1 guest_2 __MAP_HOST6__ 10006 __TIME__ __OPTS__ +bw __BW__ 1 2 + +iperf3k host diff --git a/test/migrate/rampstream_in b/test/migrate/rampstream_in new file mode 100644 index 0000000..df333ba --- /dev/null +++ b/test/migrate/rampstream_in @@ -0,0 +1,59 @@ +# SPDX-License-Identifier: GPL-2.0-or-later +# +# PASST - Plug A Simple Socket Transport +# for qemu/UNIX domain socket mode +# +# PASTA - Pack A Subtle Tap Abstraction +# for network namespace/tap device mode +# +# test/migrate/rampstream_in - Check sequence correctness with inbound ramp +# +# Copyright (c) 2025 Red Hat +# Author: David Gibson <david@gibson.dropbear.id.au> + +g1tools ip jq dhclient socat cat +htools ip jq + +set MAP_HOST4 192.0.2.1 +set MAP_HOST6 2001:db8:9a55::1 +set MAP_NS4 192.0.2.2 +set MAP_NS6 2001:db8:9a55::2 +set RAMPS 6000000 + +test Interface name +g1out IFNAME1 ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname' +hout HOST_IFNAME ip -j -4 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +hout HOST_IFNAME6 ip -j -6 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +check [ -n "__IFNAME1__" ] + +test DHCP: address +guest1 ip link set dev __IFNAME1__ up +guest1 /sbin/dhclient -4 __IFNAME1__ +g1out ADDR1 ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__IFNAME1__").addr_info[0].local' +hout HOST_ADDR ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__HOST_IFNAME__").addr_info[0].local' +check [ "__ADDR1__" = "__HOST_ADDR__" ] + +test DHCPv6: address +# Link is up now, wait for DAD to complete +guest1 while ip -j -6 addr show tentative | jq -e '.[].addr_info'; do sleep 0.1; done +guest1 /sbin/dhclient -6 __IFNAME1__ +# Wait for DAD to complete on the DHCP address +guest1 while ip -j -6 addr show tentative | jq -e '.[].addr_info'; do sleep 0.1; done +g1out ADDR1_6 ip -j -6 addr show|jq -rM '[.[] | select(.ifname == "__IFNAME1__").addr_info[] | select(.prefixlen == 128).local] | .[0]' +hout HOST_ADDR6 ip -j -6 addr show|jq -rM '[.[] | select(.ifname == "__HOST_IFNAME6__").addr_info[] | select(.scope == "global" and .deprecated != true).local] | .[0]' +check [ "__ADDR1_6__" = "__HOST_ADDR6__" ] + +test TCP/IPv4: sequence check, ramps, inbound +g1out GW1 ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway' +guest1b socat -u TCP4-LISTEN:10001 EXEC:"rampstream-check.sh __RAMPS__" +sleep 1 +hostb socat -u EXEC:"test/rampstream send __RAMPS__" TCP4:__ADDR1__:10001 + +sleep 1 + +monb echo "migrate tcp:0:20005" | socat -u STDIN UNIX:__STATESETUP__/qemu_1_mon.sock + +hostw + +guest2 cat rampstream.err +guest2 [ $(cat rampstream.status) -eq 0 ] diff --git a/test/migrate/rampstream_out b/test/migrate/rampstream_out new file mode 100644 index 0000000..8ed3229 --- /dev/null +++ b/test/migrate/rampstream_out @@ -0,0 +1,55 @@ +# SPDX-License-Identifier: GPL-2.0-or-later +# +# PASST - Plug A Simple Socket Transport +# for qemu/UNIX domain socket mode +# +# PASTA - Pack A Subtle Tap Abstraction +# for network namespace/tap device mode +# +# test/migrate/rampstream_out - Check sequence correctness with outbound ramp +# +# Copyright (c) 2025 Red Hat +# Author: David Gibson <david@gibson.dropbear.id.au> + +g1tools ip jq dhclient socat cat +htools ip jq + +set MAP_HOST4 192.0.2.1 +set MAP_HOST6 2001:db8:9a55::1 +set MAP_NS4 192.0.2.2 +set MAP_NS6 2001:db8:9a55::2 +set RAMPS 6000000 + +test Interface name +g1out IFNAME1 ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname' +hout HOST_IFNAME ip -j -4 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +hout HOST_IFNAME6 ip -j -6 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]' +check [ -n "__IFNAME1__" ] + +test DHCP: address +guest1 ip link set dev __IFNAME1__ up +guest1 /sbin/dhclient -4 __IFNAME1__ +g1out ADDR1 ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__IFNAME1__").addr_info[0].local' +hout HOST_ADDR ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__HOST_IFNAME__").addr_info[0].local' +check [ "__ADDR1__" = "__HOST_ADDR__" ] + +test DHCPv6: address +# Link is up now, wait for DAD to complete +guest1 while ip -j -6 addr show tentative | jq -e '.[].addr_info'; do sleep 0.1; done +guest1 /sbin/dhclient -6 __IFNAME1__ +# Wait for DAD to complete on the DHCP address +guest1 while ip -j -6 addr show tentative | jq -e '.[].addr_info'; do sleep 0.1; done +g1out ADDR1_6 ip -j -6 addr show|jq -rM '[.[] | select(.ifname == "__IFNAME1__").addr_info[] | select(.prefixlen == 128).local] | .[0]' +hout HOST_ADDR6 ip -j -6 addr show|jq -rM '[.[] | select(.ifname == "__HOST_IFNAME6__").addr_info[] | select(.scope == "global" and .deprecated != true).local] | .[0]' +check [ "__ADDR1_6__" = "__HOST_ADDR6__" ] + +test TCP/IPv4: sequence check, ramps, outbound +g1out GW1 ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway' +hostb socat -u TCP4-LISTEN:10006 EXEC:"test/rampstream check __RAMPS__" +sleep 1 +guest1b socat -u EXEC:"rampstream send __RAMPS__" TCP4:__MAP_HOST4__:10006 +sleep 1 + +mon echo "migrate tcp:0:20005" | socat -u STDIN UNIX:__STATESETUP__/qemu_1_mon.sock + +hostw diff --git a/test/passt.mbuto b/test/passt.mbuto index 138d365..5e00132 100755 --- a/test/passt.mbuto +++ b/test/passt.mbuto @@ -13,7 +13,8 @@ PROGS="${PROGS:-ash,dash,bash ip mount ls insmod mkdir ln cat chmod lsmod modprobe find grep mknod mv rm umount jq iperf3 dhclient hostname sed tr chown sipcalc cut socat dd strace ping tail killall sleep sysctl - nproc tcp_rr tcp_crr udp_rr which tee seq bc sshd ssh-keygen cmp}" + nproc tcp_rr tcp_crr udp_rr which tee seq bc sshd ssh-keygen cmp tcpdump + env}" # OpenSSH 9.8 introduced split binaries, with sshd being the daemon, and # sshd-session the per-session program. We need the latter as well, and the path @@ -31,7 +32,7 @@ LINKS="${LINKS:- DIRS="${DIRS} /tmp /usr/sbin /usr/share /var/log /var/lib /etc/ssh /run/sshd /root/.ssh" -COPIES="${COPIES} small.bin,/root/small.bin medium.bin,/root/medium.bin big.bin,/root/big.bin" +COPIES="${COPIES} small.bin,/root/small.bin medium.bin,/root/medium.bin big.bin,/root/big.bin rampstream,/bin/rampstream rampstream-check.sh,/bin/rampstream-check.sh" FIXUP="${FIXUP}"' mv /sbin/* /usr/sbin || : @@ -41,6 +42,7 @@ FIXUP="${FIXUP}"' #!/bin/sh LOG=/var/log/dhclient-script.log echo \${reason} \${interface} >> \$LOG +env >> \$LOG set >> \$LOG [ -n "\${new_interface_mtu}" ] && ip link set dev \${interface} mtu \${new_interface_mtu} @@ -54,7 +56,8 @@ set >> \$LOG [ -n "\${new_ip6_address}" ] && ip addr add \${new_ip6_address}/\${new_ip6_prefixlen} dev \${interface} [ -n "\${new_dhcp6_name_servers}" ] && for d in \${new_dhcp6_name_servers}; do echo "nameserver \${d}%\${interface}" >> /etc/resolv.conf; done [ -n "\${new_dhcp6_domain_search}" ] && (printf "search"; for d in \${new_dhcp6_domain_search}; do printf " %s" "\${d}"; done; printf "\n") >> /etc/resolv.conf -[ -n "\${new_host_name}" ] && hostname "\${new_host_name}" +[ -n "\${new_host_name}" ] && echo "\${new_host_name}" > /tmp/new_host_name +[ -n "\${new_fqdn_fqdn}" ] && echo "\${new_fqdn_fqdn}" > /tmp/new_fqdn_fqdn exit 0 EOF chmod 755 /sbin/dhclient-script @@ -65,6 +68,7 @@ EOF # sshd via vsock cat > /etc/passwd << EOF root:x:0:0:root:/root:/bin/sh +tcpdump:x:72:72:tcpdump:/:/sbin/nologin sshd:x:100:100:Privilege-separated SSH:/var/empty/sshd:/sbin/nologin EOF cat > /etc/shadow << EOF diff --git a/test/passt/dhcp b/test/passt/dhcp index 9925ab9..145f1ba 100644 --- a/test/passt/dhcp +++ b/test/passt/dhcp @@ -11,7 +11,7 @@ # Copyright (c) 2021 Red Hat GmbH # Author: Stefano Brivio <sbrivio@redhat.com> -gtools ip jq dhclient sed tr +gtools ip jq dhclient sed tr hostname htools ip jq sed tr head test Interface name @@ -47,7 +47,16 @@ gout SEARCH sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^searc hout HOST_SEARCH sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/,$//;s/$/\n/' check [ "__SEARCH__" = "__HOST_SEARCH__" ] +test DHCP: Hostname +gout NEW_HOST_NAME cat /tmp/new_host_name +check [ "__NEW_HOST_NAME__" = "hostname1" ] + +test DHCP: Client FQDN +gout NEW_FQDN_FQDN cat /tmp/new_fqdn_fqdn +check [ "__NEW_FQDN_FQDN__" = "fqdn1.passt.test" ] + test DHCPv6: address +guest rm /tmp/new_fqdn_fqdn guest /sbin/dhclient -6 __IFNAME__ # Wait for DAD to complete guest while ip -j -6 addr show tentative | jq -e '.[].addr_info'; do sleep 0.1; done @@ -70,3 +79,7 @@ test DHCPv6: search list gout SEARCH6 sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/,$//;s/$/\n/' hout HOST_SEARCH6 sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/,$//;s/$/\n/' check [ "__SEARCH6__" = "__HOST_SEARCH6__" ] + +test DHCPv6: Hostname +gout NEW_FQDN_FQDN cat /tmp/new_fqdn_fqdn +check [ "__NEW_FQDN_FQDN__" = "fqdn1.passt.test" ] diff --git a/test/passt_vu b/test/passt_vu new file mode 120000 index 0000000..22f1840 --- /dev/null +++ b/test/passt_vu @@ -0,0 +1 @@ +passt \ No newline at end of file diff --git a/test/passt_vu_in_ns b/test/passt_vu_in_ns new file mode 120000 index 0000000..3ff479e --- /dev/null +++ b/test/passt_vu_in_ns @@ -0,0 +1 @@ +passt_in_ns \ No newline at end of file diff --git a/test/pasta_podman/bats b/test/pasta_podman/bats index 6b1c575..2f07be8 100644 --- a/test/pasta_podman/bats +++ b/test/pasta_podman/bats @@ -23,4 +23,4 @@ check [ "__PASTA_BIN__" = "__WD__/pasta" ] test Podman system test with bats -host PODMAN="__PODMAN__" CONTAINERS_HELPER_BINARY_DIR="__WD__" bats test/podman/test/system/505-networking-pasta.bats +host PODMAN="__PODMAN__" CONTAINERS_HELPER_BINARY_DIR="__WD__" taskset -c 1 bats test/podman/test/system/505-networking-pasta.bats diff --git a/test/perf/passt_vu_tcp b/test/perf/passt_vu_tcp new file mode 100644 index 0000000..c4409b9 --- /dev/null +++ b/test/perf/passt_vu_tcp @@ -0,0 +1,211 @@ +# SPDX-License-Identifier: GPL-2.0-or-later +# +# PASST - Plug A Simple Socket Transport +# for qemu/UNIX domain socket mode +# +# PASTA - Pack A Subtle Tap Abstraction +# for network namespace/tap device mode +# +# test/perf/passt_vu_tcp - Check TCP performance in passt vhost-user mode +# +# Copyright (c) 2021 Red Hat GmbH +# Author: Stefano Brivio <sbrivio@redhat.com> + +gtools /sbin/sysctl ip jq nproc seq sleep iperf3 tcp_rr tcp_crr # From neper +nstools /sbin/sysctl ip jq nproc seq sleep iperf3 tcp_rr tcp_crr +htools bc head sed seq + +set MAP_NS4 192.0.2.2 +set MAP_NS6 2001:db8:9a55::2 + +test passt: throughput and latency + +guest /sbin/sysctl -w net.core.rmem_max=536870912 +guest /sbin/sysctl -w net.core.wmem_max=536870912 +guest /sbin/sysctl -w net.core.rmem_default=33554432 +guest /sbin/sysctl -w net.core.wmem_default=33554432 +guest /sbin/sysctl -w net.ipv4.tcp_rmem="4096 131072 268435456" +guest /sbin/sysctl -w net.ipv4.tcp_wmem="4096 131072 268435456" +guest /sbin/sysctl -w net.ipv4.tcp_timestamps=0 + +ns /sbin/sysctl -w net.ipv4.tcp_rmem="4096 524288 134217728" +ns /sbin/sysctl -w net.ipv4.tcp_wmem="4096 524288 134217728" +ns /sbin/sysctl -w net.ipv4.tcp_timestamps=0 + +gout IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname' + +hout FREQ_PROCFS (echo "scale=1"; sed -n 's/cpu MHz.*: \([0-9]*\)\..*$/(\1+10^2\/2)\/10^3/p' /proc/cpuinfo) | bc -l | head -n1 +hout FREQ_CPUFREQ (echo "scale=1"; printf '( %i + 10^5 / 2 ) / 10^6\n' $(cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq) ) | bc -l +hout FREQ [ -n "__FREQ_CPUFREQ__" ] && echo __FREQ_CPUFREQ__ || echo __FREQ_PROCFS__ + +set THREADS 6 +set TIME 2 +set OMIT 0.1 +set OPTS -Z -P __THREADS__ -O__OMIT__ -N + +info Throughput in Gbps, latency in µs, __THREADS__ threads at __FREQ__ GHz +report passt_vu tcp __THREADS__ __FREQ__ + +th MTU 256B 576B 1280B 1500B 9000B 65520B + + +tr TCP throughput over IPv6: guest to host +iperf3s ns 10002 + +bw - +bw - +guest ip link set dev __IFNAME__ mtu 1280 +iperf3 BW guest __MAP_NS6__ 10002 __TIME__ __OPTS__ -w 16M -l 1M +bw __BW__ 1.2 1.5 +guest ip link set dev __IFNAME__ mtu 1500 +iperf3 BW guest __MAP_NS6__ 10002 __TIME__ __OPTS__ -w 32M -l 1M +bw __BW__ 1.6 1.8 +guest ip link set dev __IFNAME__ mtu 9000 +iperf3 BW guest __MAP_NS6__ 10002 __TIME__ __OPTS__ -w 64M -l 1M +bw __BW__ 4.0 5.0 +guest ip link set dev __IFNAME__ mtu 65520 +iperf3 BW guest __MAP_NS6__ 10002 __TIME__ __OPTS__ -w 64M -l 1M +bw __BW__ 7.0 8.0 + +iperf3k ns + +tl TCP RR latency over IPv6: guest to host +lat - +lat - +lat - +lat - +lat - +nsb tcp_rr --nolog -6 +gout LAT tcp_rr --nolog -l1 -6 -c -H __MAP_NS6__ | sed -n 's/^throughput=\(.*\)/\1/p' +lat __LAT__ 200 150 + +tl TCP CRR latency over IPv6: guest to host +lat - +lat - +lat - +lat - +lat - +nsb tcp_crr --nolog -6 +gout LAT tcp_crr --nolog -l1 -6 -c -H __MAP_NS6__ | sed -n 's/^throughput=\(.*\)/\1/p' +lat __LAT__ 500 400 + +tr TCP throughput over IPv4: guest to host +iperf3s ns 10002 + +guest ip link set dev __IFNAME__ mtu 256 +iperf3 BW guest __MAP_NS4__ 10002 __TIME__ __OPTS__ -w 2M -l 1M +bw __BW__ 0.2 0.3 +guest ip link set dev __IFNAME__ mtu 576 +iperf3 BW guest __MAP_NS4__ 10002 __TIME__ __OPTS__ -w 4M -l 1M +bw __BW__ 0.5 0.8 +guest ip link set dev __IFNAME__ mtu 1280 +iperf3 BW guest __MAP_NS4__ 10002 __TIME__ __OPTS__ -w 8M -l 1M +bw __BW__ 1.2 1.5 +guest ip link set dev __IFNAME__ mtu 1500 +iperf3 BW guest __MAP_NS4__ 10002 __TIME__ __OPTS__ -w 16M -l 1M +bw __BW__ 1.6 1.8 +guest ip link set dev __IFNAME__ mtu 9000 +iperf3 BW guest __MAP_NS4__ 10002 __TIME__ __OPTS__ -w 64M -l 1M +bw __BW__ 4.0 5.0 +guest ip link set dev __IFNAME__ mtu 65520 +iperf3 BW guest __MAP_NS4__ 10002 __TIME__ __OPTS__ -w 64M -l 1M +bw __BW__ 7.0 8.0 + +iperf3k ns + +# Reducing MTU below 1280 deconfigures IPv6, get our address back +guest dhclient -6 -x +guest dhclient -6 __IFNAME__ + +tl TCP RR latency over IPv4: guest to host +lat - +lat - +lat - +lat - +lat - +nsb tcp_rr --nolog -4 +gout LAT tcp_rr --nolog -l1 -4 -c -H __MAP_NS4__ | sed -n 's/^throughput=\(.*\)/\1/p' +lat __LAT__ 200 150 + +tl TCP CRR latency over IPv4: guest to host +lat - +lat - +lat - +lat - +lat - +nsb tcp_crr --nolog -4 +gout LAT tcp_crr --nolog -l1 -4 -c -H __MAP_NS4__ | sed -n 's/^throughput=\(.*\)/\1/p' +lat __LAT__ 500 400 + +tr TCP throughput over IPv6: host to guest +iperf3s guest 10001 + +bw - +bw - +bw - +bw - +bw - +iperf3 BW ns ::1 10001 __TIME__ __OPTS__ -w 256M -l 16k +bw __BW__ 6.0 6.8 + +iperf3k guest + +tl TCP RR latency over IPv6: host to guest +lat - +lat - +lat - +lat - +lat - +guestb tcp_rr --nolog -P 10001 -C 10011 -6 +sleep 1 +nsout LAT tcp_rr --nolog -l1 -P 10001 -C 10011 -6 -c -H ::1 | sed -n 's/^throughput=\(.*\)/\1/p' +lat __LAT__ 200 150 + +tl TCP CRR latency over IPv6: host to guest +lat - +lat - +lat - +lat - +lat - +guestb tcp_crr --nolog -P 10001 -C 10011 -6 +sleep 1 +nsout LAT tcp_crr --nolog -l1 -P 10001 -C 10011 -6 -c -H ::1 | sed -n 's/^throughput=\(.*\)/\1/p' +lat __LAT__ 500 350 + + +tr TCP throughput over IPv4: host to guest +iperf3s guest 10001 + +bw - +bw - +bw - +bw - +bw - +iperf3 BW ns 127.0.0.1 10001 __TIME__ __OPTS__ -w 256M -l 16k +bw __BW__ 6.0 6.8 + +iperf3k guest + +tl TCP RR latency over IPv4: host to guest +lat - +lat - +lat - +lat - +lat - +guestb tcp_rr --nolog -P 10001 -C 10011 -4 +sleep 1 +nsout LAT tcp_rr --nolog -l1 -P 10001 -C 10011 -4 -c -H 127.0.0.1 | sed -n 's/^throughput=\(.*\)/\1/p' +lat __LAT__ 200 150 + +tl TCP CRR latency over IPv6: host to guest +lat - +lat - +lat - +lat - +lat - +guestb tcp_crr --nolog -P 10001 -C 10011 -4 +sleep 1 +nsout LAT tcp_crr --nolog -l1 -P 10001 -C 10011 -4 -c -H 127.0.0.1 | sed -n 's/^throughput=\(.*\)/\1/p' +lat __LAT__ 500 300 + +te diff --git a/test/perf/passt_vu_udp b/test/perf/passt_vu_udp new file mode 100644 index 0000000..943ac11 --- /dev/null +++ b/test/perf/passt_vu_udp @@ -0,0 +1,159 @@ +# SPDX-License-Identifier: GPL-2.0-or-later +# +# PASST - Plug A Simple Socket Transport +# for qemu/UNIX domain socket mode +# +# PASTA - Pack A Subtle Tap Abstraction +# for network namespace/tap device mode +# +# test/perf/passt_vu_udp - Check UDP performance in passt vhost-user mode +# +# Copyright (c) 2021 Red Hat GmbH +# Author: Stefano Brivio <sbrivio@redhat.com> + +gtools /sbin/sysctl ip jq nproc sleep iperf3 udp_rr # From neper +nstools ip jq sleep iperf3 udp_rr +htools bc head sed + +set MAP_NS4 192.0.2.2 +set MAP_NS6 2001:db8:9a55::2 + +test passt: throughput and latency + +guest /sbin/sysctl -w net.core.rmem_max=16777216 +guest /sbin/sysctl -w net.core.wmem_max=16777216 +guest /sbin/sysctl -w net.core.rmem_default=16777216 +guest /sbin/sysctl -w net.core.wmem_default=16777216 + +hout FREQ_PROCFS (echo "scale=1"; sed -n 's/cpu MHz.*: \([0-9]*\)\..*$/(\1+10^2\/2)\/10^3/p' /proc/cpuinfo) | bc -l | head -n1 +hout FREQ_CPUFREQ (echo "scale=1"; printf '( %i + 10^5 / 2 ) / 10^6\n' $(cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq) ) | bc -l +hout FREQ [ -n "__FREQ_CPUFREQ__" ] && echo __FREQ_CPUFREQ__ || echo __FREQ_PROCFS__ + +set THREADS 2 +set TIME 1 +set OPTS -u -P __THREADS__ --pacing-timer 1000 + +info Throughput in Gbps, latency in µs, __THREADS__ threads at __FREQ__ GHz + +report passt_vu udp __THREADS__ __FREQ__ + +th pktlen 256B 576B 1280B 1500B 9000B 65520B + +tr UDP throughput over IPv6: guest to host +iperf3s ns 10002 +# (datagram size) = (packet size) - 48: 40 bytes of IPv6 header, 8 of UDP header + +bw - +bw - +iperf3 BW guest __MAP_NS6__ 10002 __TIME__ __OPTS__ -b 3G -l 1232 +bw __BW__ 0.8 1.2 +iperf3 BW guest __MAP_NS6__ 10002 __TIME__ __OPTS__ -b 4G -l 1452 +bw __BW__ 1.0 1.5 +iperf3 BW guest __MAP_NS6__ 10002 __TIME__ __OPTS__ -b 10G -l 8952 +bw __BW__ 4.0 5.0 +iperf3 BW guest __MAP_NS6__ 10002 __TIME__ __OPTS__ -b 20G -l 64372 +bw __BW__ 4.0 5.0 + +iperf3k ns + +tl UDP RR latency over IPv6: guest to host +lat - +lat - +lat - +lat - +lat - +nsb udp_rr --nolog -6 +gout LAT udp_rr --nolog -6 -c -H __MAP_NS6__ | sed -n 's/^throughput=\(.*\)/\1/p' +lat __LAT__ 200 150 + + +tr UDP throughput over IPv4: guest to host +iperf3s ns 10002 +# (datagram size) = (packet size) - 28: 20 bytes of IPv4 header, 8 of UDP header + +iperf3 BW guest __MAP_NS4__ 10002 __TIME__ __OPTS__ -b 1G -l 228 +bw __BW__ 0.0 0.0 +iperf3 BW guest __MAP_NS4__ 10002 __TIME__ __OPTS__ -b 2G -l 548 +bw __BW__ 0.4 0.6 +iperf3 BW guest __MAP_NS4__ 10002 __TIME__ __OPTS__ -b 3G -l 1252 +bw __BW__ 0.8 1.2 +iperf3 BW guest __MAP_NS4__ 10002 __TIME__ __OPTS__ -b 4G -l 1472 +bw __BW__ 1.0 1.5 +iperf3 BW guest __MAP_NS4__ 10002 __TIME__ __OPTS__ -b 10G -l 8972 +bw __BW__ 4.0 5.0 +iperf3 BW guest __MAP_NS4__ 10002 __TIME__ __OPTS__ -b 20G -l 65492 +bw __BW__ 4.0 5.0 + +iperf3k ns + +tl UDP RR latency over IPv4: guest to host +lat - +lat - +lat - +lat - +lat - +nsb udp_rr --nolog -4 +gout LAT udp_rr --nolog -4 -c -H __MAP_NS4__ | sed -n 's/^throughput=\(.*\)/\1/p' +lat __LAT__ 200 150 + + +tr UDP throughput over IPv6: host to guest +iperf3s guest 10001 +# (datagram size) = (packet size) - 48: 40 bytes of IPv6 header, 8 of UDP header + +bw - +bw - +iperf3 BW ns ::1 10001 __TIME__ __OPTS__ -b 3G -l 1232 +bw __BW__ 0.8 1.2 +iperf3 BW ns ::1 10001 __TIME__ __OPTS__ -b 4G -l 1452 +bw __BW__ 1.0 1.5 +iperf3 BW ns ::1 10001 __TIME__ __OPTS__ -b 10G -l 8952 +bw __BW__ 3.0 4.0 +iperf3 BW ns ::1 10001 __TIME__ __OPTS__ -b 20G -l 64372 +bw __BW__ 3.0 4.0 + +iperf3k guest + +tl UDP RR latency over IPv6: host to guest +lat - +lat - +lat - +lat - +lat - +guestb udp_rr --nolog -P 10001 -C 10011 -6 +sleep 1 +nsout LAT udp_rr --nolog -P 10001 -C 10011 -6 -c -H ::1 | sed -n 's/^throughput=\(.*\)/\1/p' +lat __LAT__ 200 150 + + +tr UDP throughput over IPv4: host to guest +iperf3s guest 10001 +# (datagram size) = (packet size) - 28: 20 bytes of IPv4 header, 8 of UDP header + +iperf3 BW ns 127.0.0.1 10001 __TIME__ __OPTS__ -b 1G -l 228 +bw __BW__ 0.0 0.0 +iperf3 BW ns 127.0.0.1 10001 __TIME__ __OPTS__ -b 2G -l 548 +bw __BW__ 0.4 0.6 +iperf3 BW ns 127.0.0.1 10001 __TIME__ __OPTS__ -b 3G -l 1252 +bw __BW__ 0.8 1.2 +iperf3 BW ns 127.0.0.1 10001 __TIME__ __OPTS__ -b 4G -l 1472 +bw __BW__ 1.0 1.5 +iperf3 BW ns 127.0.0.1 10001 __TIME__ __OPTS__ -b 10G -l 8972 +bw __BW__ 3.0 4.0 +iperf3 BW ns 127.0.0.1 10001 __TIME__ __OPTS__ -b 20G -l 65492 +bw __BW__ 3.0 4.0 + +iperf3k guest + +tl UDP RR latency over IPv4: host to guest +lat - +lat - +lat - +lat - +lat - +guestb udp_rr --nolog -P 10001 -C 10011 -4 +sleep 1 +nsout LAT udp_rr --nolog -P 10001 -C 10011 -4 -c -H 127.0.0.1 | sed -n 's/^throughput=\(.*\)/\1/p' +lat __LAT__ 200 150 + +te diff --git a/test/rampstream-check.sh b/test/rampstream-check.sh new file mode 100755 index 0000000..c27acdb --- /dev/null +++ b/test/rampstream-check.sh @@ -0,0 +1,3 @@ +#! /bin/sh + +(rampstream check "$@" 2>&1; echo $? > rampstream.status) | tee rampstream.err diff --git a/test/rampstream.c b/test/rampstream.c new file mode 100644 index 0000000..8d81296 --- /dev/null +++ b/test/rampstream.c @@ -0,0 +1,143 @@ +// SPDX-License-Identifier: GPL-2.0-or-later + +/* rampstream - Generate a check and stream of bytes in a ramp pattern + * + * Copyright Red Hat + * Author: David Gibson <david@gibson.dropbear.id.au> + */ + +#include <stdio.h> +#include <stdlib.h> +#include <stdint.h> +#include <sys/types.h> +#include <unistd.h> +#include <errno.h> +#include <string.h> + +/* Length of the repeating ramp. This is a deliberately not a "round" number so + * that we're very likely to misalign with likely block or chunk sizes of the + * transport. That means we'll detect gaps in the stream, even if they occur + * neatly on block boundaries. Specifically this is the largest 8-bit prime. */ +#define RAMPLEN 251 + +#define INTERVAL 10000 + +#define ARRAY_SIZE(a) ((int)(sizeof(a) / sizeof((a)[0]))) + +#define die(...) \ + do { \ + fprintf(stderr, "rampstream: " __VA_ARGS__); \ + exit(1); \ + } while (0) + +static void usage(void) +{ + die("Usage:\n" + " rampstream send <number>\n" + " Generate a ramp pattern of bytes on stdout, repeated <number>\n" + " times\n" + " rampstream check <number>\n" + " Check a ramp pattern of bytes on stdin, repeater <number>\n" + " times\n"); +} + +static void ramp_send(unsigned long long num, const uint8_t *ramp) +{ + unsigned long long i; + + for (i = 0; i < num; i++) { + int off = 0; + ssize_t rc; + + if (i % INTERVAL == 0) + fprintf(stderr, "%llu...\r", i); + + while (off < RAMPLEN) { + rc = write(1, ramp + off, RAMPLEN - off); + if (rc < 0) { + if (errno == EINTR || + errno == EAGAIN || + errno == EWOULDBLOCK) + continue; + die("Error writing ramp: %s\n", + strerror(errno)); + } + if (rc == 0) + die("Zero length write\n"); + off += rc; + } + } +} + +static void ramp_check(unsigned long long num, const uint8_t *ramp) +{ + unsigned long long i; + + for (i = 0; i < num; i++) { + uint8_t buf[RAMPLEN]; + int off = 0; + ssize_t rc; + + if (i % INTERVAL == 0) + fprintf(stderr, "%llu...\r", i); + + while (off < RAMPLEN) { + rc = read(0, buf + off, RAMPLEN - off); + if (rc < 0) { + if (errno == EINTR || + errno == EAGAIN || + errno == EWOULDBLOCK) + continue; + die("Error reading ramp: %s\n", + strerror(errno)); + } + if (rc == 0) + die("Unexpected EOF, ramp %llu, byte %d\n", + i, off); + off += rc; + } + + if (memcmp(buf, ramp, sizeof(buf)) != 0) { + int j, k; + + for (j = 0; j < RAMPLEN; j++) + if (buf[j] != ramp[j]) + break; + for (k = j; k < RAMPLEN && k < j + 16; k++) + fprintf(stderr, + "Byte %d: expected 0x%02x, got 0x%02x\n", + k, ramp[k], buf[k]); + die("Data mismatch, ramp %llu, byte %d\n", i, j); + } + } +} + +int main(int argc, char *argv[]) +{ + const char *subcmd = argv[1]; + unsigned long long num; + uint8_t ramp[RAMPLEN]; + char *e; + int i; + + if (argc < 2) + usage(); + + errno = 0; + num = strtoull(argv[2], &e, 0); + if (*e || errno) + usage(); + + /* Initialize the ramp block */ + for (i = 0; i < RAMPLEN; i++) + ramp[i] = i; + + if (strcmp(subcmd, "send") == 0) + ramp_send(num, ramp); + else if (strcmp(subcmd, "check") == 0) + ramp_check(num, ramp); + else + usage(); + + exit(0); +} diff --git a/test/run b/test/run index 547a729..f73c311 100755 --- a/test/run +++ b/test/run @@ -93,6 +93,7 @@ run() { test memory/passt teardown memory + VHOST_USER=0 setup passt test passt/ndp test passt/dhcp @@ -115,7 +116,59 @@ run() { test two_guests/basic teardown two_guests + VHOST_USER=1 + setup passt_in_ns + test passt_vu/ndp + test passt_vu_in_ns/dhcp + test passt_vu_in_ns/icmp + test passt_vu_in_ns/tcp + test passt_vu_in_ns/udp + test passt_vu_in_ns/shutdown + teardown passt_in_ns + + setup two_guests + test two_guests_vu/basic + teardown two_guests + + setup migrate + test migrate/basic + teardown migrate + setup migrate + test migrate/basic_fin + teardown migrate + setup migrate + test migrate/bidirectional + teardown migrate + setup migrate + test migrate/bidirectional_fin + teardown migrate + setup migrate + test migrate/iperf3_out4 + teardown migrate + setup migrate + test migrate/iperf3_out6 + teardown migrate + setup migrate + test migrate/iperf3_in4 + teardown migrate + setup migrate + test migrate/iperf3_in6 + teardown migrate + setup migrate + test migrate/iperf3_bidir6 + teardown migrate + setup migrate + test migrate/iperf3_many_out6 + teardown migrate + setup migrate + test migrate/rampstream_in + teardown migrate + setup migrate + test migrate/rampstream_out + teardown migrate + VALGRIND=0 + VHOST_USER=0 setup passt_in_ns test passt/ndp test passt_in_ns/dhcp @@ -126,6 +179,15 @@ run() { test passt_in_ns/shutdown teardown passt_in_ns + VHOST_USER=1 + setup passt_in_ns + test passt_vu/ndp + test passt_vu_in_ns/dhcp + test perf/passt_vu_tcp + test perf/passt_vu_udp + test passt_vu_in_ns/shutdown + teardown passt_in_ns + # TODO: Make those faster by at least pre-installing gcc and make on # non-x86 images, then re-enable. skip_distro() { @@ -140,7 +202,7 @@ skip_distro() { perf_finish [ ${CI} -eq 1 ] && video_stop - log "PASS: ${STATUS_PASS}, FAIL: ${STATUS_FAIL}" + log "PASS: ${STATUS_PASS}, FAIL: ${STATUS_FAIL}, SKIPPED: ${STATUS_SKIPPED}" pause_continue \ "Press any key to keep test session open" \ @@ -161,7 +223,10 @@ run_selected() { __setup= for __test; do - if [ "${__test%%/*}" != "${__setup}" ]; then + # HACK: the migrate tests need the setup repeated for + # each test + if [ "${__test%%/*}" != "${__setup}" -o \ + "${__test%%/*}" = "migrate" ]; then [ -n "${__setup}" ] && teardown "${__setup}" __setup="${__test%%/*}" setup "${__setup}" @@ -171,7 +236,7 @@ run_selected() { done teardown "${__setup}" - log "PASS: ${STATUS_PASS}, FAIL: ${STATUS_FAIL}" + log "PASS: ${STATUS_PASS}, FAIL: ${STATUS_FAIL}, SKIPPED: ${STATUS_SKIPPED}" pause_continue \ "Press any key to keep test session open" \ @@ -242,4 +307,4 @@ fi tail -n1 ${LOGFILE} echo "Log at ${LOGFILE}" -exit $(tail -n1 ${LOGFILE} | sed -n 's/.*FAIL: \(.*\)$/\1/p') +exit $(tail -n1 ${LOGFILE} | sed -n 's/.*FAIL: \(.*\),.*$/\1/p') diff --git a/test/two_guests_vu b/test/two_guests_vu new file mode 120000 index 0000000..a8648fc --- /dev/null +++ b/test/two_guests_vu @@ -0,0 +1 @@ +two_guests \ No newline at end of file diff --git a/udp.c b/udp.c index 4be165f..ca28b37 100644 --- a/udp.c +++ b/udp.c @@ -39,27 +39,30 @@ * could receive packets from multiple flows, so we use a hash table match to * find the specific flow for a datagram. * - * When a UDP flow is initiated from a listening socket we take a duplicate of - * the socket and store it in uflow->s[INISIDE]. This will last for the + * Flow sockets + * ============ + * + * When a UDP flow targets a socket, we create a "flow" socket in + * uflow->s[TGTSIDE] both to deliver datagrams to the target side and receive + * replies on the target side. This socket is both bound and connected and has + * EPOLL_TYPE_UDP. The connect() means it will only receive datagrams + * associated with this flow, so the epoll reference directly points to the flow + * and we don't need a hash lookup. + * + * When a flow is initiated from a listening socket, we create a "flow" socket + * with the same bound address as the listening socket, but also connect()ed to + * the flow's peer. This is stored in uflow->s[INISIDE] and will last for the * lifetime of the flow, even if the original listening socket is closed due to * port auto-probing. The duplicate is used to deliver replies back to the * originating side. * - * Reply sockets - * ============= - * - * When a UDP flow targets a socket, we create a "reply" socket in - * uflow->s[TGTSIDE] both to deliver datagrams to the target side and receive - * replies on the target side. This socket is both bound and connected and has - * EPOLL_TYPE_UDP_REPLY. The connect() means it will only receive datagrams - * associated with this flow, so the epoll reference directly points to the flow - * and we don't need a hash lookup. - * - * NOTE: it's possible that the reply socket could have a bound address - * overlapping with an unrelated listening socket. We assume datagrams for the - * flow will come to the reply socket in preference to a listening socket. The - * sample program doc/platform-requirements/reuseaddr-priority.c documents and - * tests that assumption. + * NOTE: A flow socket can have a bound address overlapping with a listening + * socket. That will happen naturally for flows initiated from a socket, but is + * also possible (though unlikely) for tap initiated flows, depending on the + * source port. We assume datagrams for the flow will come to a connect()ed + * socket in preference to a listening socket. The sample program + * doc/platform-requirements/reuseaddr-priority.c documents and tests that + * assumption. * * "Spliced" flows * =============== @@ -71,8 +74,7 @@ * actually used; it doesn't make sense for datagrams and instead a pair of * recvmmsg() and sendmmsg() is used to forward the datagrams. * - * Note that a spliced flow will have *both* a duplicated listening socket and a - * reply socket (see above). + * Note that a spliced flow will have two flow sockets (see above). */ #include <sched.h> @@ -87,6 +89,8 @@ #include <netinet/in.h> #include <netinet/ip.h> #include <netinet/udp.h> +#include <netinet/ip_icmp.h> +#include <netinet/icmp6.h> #include <stdint.h> #include <stddef.h> #include <string.h> @@ -109,29 +113,25 @@ #include "pcap.h" #include "log.h" #include "flow_table.h" +#include "udp_internal.h" +#include "udp_vu.h" #define UDP_MAX_FRAMES 32 /* max # of frames to receive at once */ +/* Maximum UDP data to be returned in ICMP messages */ +#define ICMP4_MAX_DLEN 8 +#define ICMP6_MAX_DLEN (IPV6_MIN_MTU \ + - sizeof(struct udphdr) \ + - sizeof(struct ipv6hdr)) + /* "Spliced" sockets indexed by bound port (host order) */ static int udp_splice_ns [IP_VERSIONS][NUM_PORTS]; static int udp_splice_init[IP_VERSIONS][NUM_PORTS]; /* Static buffers */ -/** - * struct udp_payload_t - UDP header and data for inbound messages - * @uh: UDP header - * @data: UDP data - */ -static struct udp_payload_t { - struct udphdr uh; - char data[USHRT_MAX - sizeof(struct udphdr)]; -#ifdef __AVX2__ -} __attribute__ ((packed, aligned(32))) -#else -} __attribute__ ((packed, aligned(__alignof__(unsigned int)))) -#endif -udp_payload[UDP_MAX_FRAMES]; +/* UDP header and data for inbound messages */ +static struct udp_payload_t udp_payload[UDP_MAX_FRAMES]; /* Ethernet header for IPv4 frames */ static struct ethhdr udp4_eth_hdr; @@ -140,26 +140,31 @@ static struct ethhdr udp4_eth_hdr; static struct ethhdr udp6_eth_hdr; /** - * struct udp_meta_t - Pre-cooked headers and metadata for UDP packets + * struct udp_meta_t - Pre-cooked headers for UDP packets * @ip6h: Pre-filled IPv6 header (except for payload_len and addresses) * @ip4h: Pre-filled IPv4 header (except for tot_len and saddr) * @taph: Tap backend specific header - * @s_in: Source socket address, filled in by recvmmsg() - * @tosidx: sidx for the destination side of this datagram's flow */ static struct udp_meta_t { struct ipv6hdr ip6h; struct iphdr ip4h; struct tap_hdr taph; - - union sockaddr_inany s_in; - flow_sidx_t tosidx; } #ifdef __AVX2__ __attribute__ ((aligned(32))) #endif udp_meta[UDP_MAX_FRAMES]; +#define PKTINFO_SPACE \ + MAX(CMSG_SPACE(sizeof(struct in_pktinfo)), \ + CMSG_SPACE(sizeof(struct in6_pktinfo))) + +#define RECVERR_SPACE \ + MAX(CMSG_SPACE(sizeof(struct sock_extended_err) + \ + sizeof(struct sockaddr_in)), \ + CMSG_SPACE(sizeof(struct sock_extended_err) + \ + sizeof(struct sockaddr_in6))) + /** * enum udp_iov_idx - Indices for the buffers making up a single UDP frame * @UDP_IOV_TAP tap specific header @@ -236,8 +241,6 @@ static void udp_iov_init_one(const struct ctx *c, size_t i) tiov[UDP_IOV_TAP] = tap_hdr_iov(c, &meta->taph); tiov[UDP_IOV_PAYLOAD].iov_base = payload; - mh->msg_name = &meta->s_in; - mh->msg_namelen = sizeof(meta->s_in); mh->msg_iov = siov; mh->msg_iovlen = 1; } @@ -257,41 +260,6 @@ static void udp_iov_init(const struct ctx *c) udp_iov_init_one(c, i); } -/** - * udp_splice_prepare() - Prepare one datagram for splicing - * @mmh: Receiving mmsghdr array - * @idx: Index of the datagram to prepare - */ -static void udp_splice_prepare(struct mmsghdr *mmh, unsigned idx) -{ - udp_mh_splice[idx].msg_hdr.msg_iov->iov_len = mmh[idx].msg_len; -} - -/** - * udp_splice_send() - Send a batch of datagrams from socket to socket - * @c: Execution context - * @start: Index of batch's first datagram in udp[46]_l2_buf - * @n: Number of datagrams in batch - * @src: Source port for datagram (target side) - * @dst: Destination port for datagrams (target side) - * @ref: epoll reference for origin socket - * @now: Timestamp - */ -static void udp_splice_send(const struct ctx *c, size_t start, size_t n, - flow_sidx_t tosidx) -{ - const struct flowside *toside = flowside_at_sidx(tosidx); - const struct udp_flow *uflow = udp_at_sidx(tosidx); - uint8_t topif = pif_at_sidx(tosidx); - int s = uflow->s[tosidx.sidei]; - socklen_t sl; - - pif_sockaddr(c, &udp_splice_to, &sl, topif, - &toside->eaddr, toside->eport); - - sendmmsg(s, udp_mh_splice + start, n, MSG_NOSIGNAL); -} - /** * udp_update_hdr4() - Update headers for one IPv4 datagram * @ip4h: Pre-filled IPv4 header (except for tot_len and saddr) @@ -302,9 +270,9 @@ static void udp_splice_send(const struct ctx *c, size_t start, size_t n, * * Return: size of IPv4 payload (UDP header + data) */ -static size_t udp_update_hdr4(struct iphdr *ip4h, struct udp_payload_t *bp, - const struct flowside *toside, size_t dlen, - bool no_udp_csum) +size_t udp_update_hdr4(struct iphdr *ip4h, struct udp_payload_t *bp, + const struct flowside *toside, size_t dlen, + bool no_udp_csum) { const struct in_addr *src = inany_v4(&toside->oaddr); const struct in_addr *dst = inany_v4(&toside->eaddr); @@ -328,7 +296,8 @@ static size_t udp_update_hdr4(struct iphdr *ip4h, struct udp_payload_t *bp, .iov_base = bp->data, .iov_len = dlen }; - csum_udp4(&bp->uh, *src, *dst, &iov, 1, 0); + struct iov_tail data = IOV_TAIL(&iov, 1, 0); + csum_udp4(&bp->uh, *src, *dst, &data); } return l4len; @@ -345,9 +314,9 @@ static size_t udp_update_hdr4(struct iphdr *ip4h, struct udp_payload_t *bp, * * Return: size of IPv6 payload (UDP header + data) */ -static size_t udp_update_hdr6(struct ipv6hdr *ip6h, struct udp_payload_t *bp, - const struct flowside *toside, size_t dlen, - bool no_udp_csum) +size_t udp_update_hdr6(struct ipv6hdr *ip6h, struct udp_payload_t *bp, + const struct flowside *toside, size_t dlen, + bool no_udp_csum) { uint16_t l4len = dlen + sizeof(bp->uh); @@ -372,8 +341,8 @@ static size_t udp_update_hdr6(struct ipv6hdr *ip6h, struct udp_payload_t *bp, .iov_base = bp->data, .iov_len = dlen }; - csum_udp6(&bp->uh, &toside->oaddr.a6, &toside->eaddr.a6, - &iov, 1, 0); + struct iov_tail data = IOV_TAIL(&iov, 1, 0); + csum_udp6(&bp->uh, &toside->oaddr.a6, &toside->eaddr.a6, &data); } return l4len; @@ -413,28 +382,172 @@ static void udp_tap_prepare(const struct mmsghdr *mmh, (*tap_iov)[UDP_IOV_PAYLOAD].iov_len = l4len; } +/** + * udp_send_tap_icmp4() - Construct and send ICMPv4 to local peer + * @c: Execution context + * @ee: Extended error descriptor + * @toside: Destination side of flow + * @saddr: Address of ICMP generating node + * @in: First bytes (max 8) of original UDP message body + * @dlen: Length of the read part of original UDP message body + */ +static void udp_send_tap_icmp4(const struct ctx *c, + const struct sock_extended_err *ee, + const struct flowside *toside, + struct in_addr saddr, + const void *in, size_t dlen) +{ + struct in_addr oaddr = toside->oaddr.v4mapped.a4; + struct in_addr eaddr = toside->eaddr.v4mapped.a4; + in_port_t eport = toside->eport; + in_port_t oport = toside->oport; + struct { + struct icmphdr icmp4h; + struct iphdr ip4h; + struct udphdr uh; + char data[ICMP4_MAX_DLEN]; + } __attribute__((packed, aligned(__alignof__(max_align_t)))) msg; + size_t msglen = sizeof(msg) - sizeof(msg.data) + dlen; + size_t l4len = dlen + sizeof(struct udphdr); + + ASSERT(dlen <= ICMP4_MAX_DLEN); + memset(&msg, 0, sizeof(msg)); + msg.icmp4h.type = ee->ee_type; + msg.icmp4h.code = ee->ee_code; + if (ee->ee_type == ICMP_DEST_UNREACH && ee->ee_code == ICMP_FRAG_NEEDED) + msg.icmp4h.un.frag.mtu = htons((uint16_t) ee->ee_info); + + /* Reconstruct the original headers as returned in the ICMP message */ + tap_push_ip4h(&msg.ip4h, eaddr, oaddr, l4len, IPPROTO_UDP); + tap_push_uh4(&msg.uh, eaddr, eport, oaddr, oport, in, dlen); + memcpy(&msg.data, in, dlen); + + tap_icmp4_send(c, saddr, eaddr, &msg, msglen); +} + + +/** + * udp_send_tap_icmp6() - Construct and send ICMPv6 to local peer + * @c: Execution context + * @ee: Extended error descriptor + * @toside: Destination side of flow + * @saddr: Address of ICMP generating node + * @in: First bytes (max 1232) of original UDP message body + * @dlen: Length of the read part of original UDP message body + * @flow: IPv6 flow identifier + */ +static void udp_send_tap_icmp6(const struct ctx *c, + const struct sock_extended_err *ee, + const struct flowside *toside, + const struct in6_addr *saddr, + void *in, size_t dlen, uint32_t flow) +{ + const struct in6_addr *oaddr = &toside->oaddr.a6; + const struct in6_addr *eaddr = &toside->eaddr.a6; + in_port_t eport = toside->eport; + in_port_t oport = toside->oport; + struct { + struct icmp6_hdr icmp6h; + struct ipv6hdr ip6h; + struct udphdr uh; + char data[ICMP6_MAX_DLEN]; + } __attribute__((packed, aligned(__alignof__(max_align_t)))) msg; + size_t msglen = sizeof(msg) - sizeof(msg.data) + dlen; + size_t l4len = dlen + sizeof(struct udphdr); + + ASSERT(dlen <= ICMP6_MAX_DLEN); + memset(&msg, 0, sizeof(msg)); + msg.icmp6h.icmp6_type = ee->ee_type; + msg.icmp6h.icmp6_code = ee->ee_code; + if (ee->ee_type == ICMP6_PACKET_TOO_BIG) + msg.icmp6h.icmp6_dataun.icmp6_un_data32[0] = htonl(ee->ee_info); + + /* Reconstruct the original headers as returned in the ICMP message */ + tap_push_ip6h(&msg.ip6h, eaddr, oaddr, l4len, IPPROTO_UDP, flow); + tap_push_uh6(&msg.uh, eaddr, eport, oaddr, oport, in, dlen); + memcpy(&msg.data, in, dlen); + + tap_icmp6_send(c, saddr, eaddr, &msg, msglen); +} + +/** + * udp_pktinfo() - Retrieve packet destination address from cmsg + * @msg: msghdr into which message has been received + * @dst: (Local) destination address of message in @mh (output) + * + * Return: 0 on success, -1 if the information was missing (@dst is set to + * inany_any6). + */ +static int udp_pktinfo(struct msghdr *msg, union inany_addr *dst) +{ + struct cmsghdr *hdr; + + for (hdr = CMSG_FIRSTHDR(msg); hdr; hdr = CMSG_NXTHDR(msg, hdr)) { + if (hdr->cmsg_level == IPPROTO_IP && + hdr->cmsg_type == IP_PKTINFO) { + const struct in_pktinfo *i4 = (void *)CMSG_DATA(hdr); + + *dst = inany_from_v4(i4->ipi_addr); + return 0; + } + + if (hdr->cmsg_level == IPPROTO_IPV6 && + hdr->cmsg_type == IPV6_PKTINFO) { + const struct in6_pktinfo *i6 = (void *)CMSG_DATA(hdr); + + dst->a6 = i6->ipi6_addr; + return 0; + } + } + + debug("Missing PKTINFO cmsg on datagram"); + *dst = inany_any6; + return -1; +} + /** * udp_sock_recverr() - Receive and clear an error from a socket - * @s: Socket to receive from + * @c: Execution context + * @s: Socket to receive errors from + * @sidx: Flow and side of @s, or FLOW_SIDX_NONE if unknown + * @pif: Interface on which the error occurred + * (only used if @sidx == FLOW_SIDX_NONE) + * @port: Local port number of @s (only used if @sidx == FLOW_SIDX_NONE) * * Return: 1 if error received and processed, 0 if no more errors in queue, < 0 * if there was an error reading the queue * * #syscalls recvmsg */ -static int udp_sock_recverr(int s) +static int udp_sock_recverr(const struct ctx *c, int s, flow_sidx_t sidx, + uint8_t pif, in_port_t port) { + char buf[PKTINFO_SPACE + RECVERR_SPACE]; const struct sock_extended_err *ee; - const struct cmsghdr *hdr; - char buf[CMSG_SPACE(sizeof(*ee))]; + char data[ICMP6_MAX_DLEN]; + struct cmsghdr *hdr; + struct iovec iov = { + .iov_base = data, + .iov_len = sizeof(data) + }; + union sockaddr_inany src; struct msghdr mh = { - .msg_name = NULL, - .msg_namelen = 0, - .msg_iov = NULL, - .msg_iovlen = 0, + .msg_name = &src, + .msg_namelen = sizeof(src), + .msg_iov = &iov, + .msg_iovlen = 1, .msg_control = buf, .msg_controllen = sizeof(buf), }; + const struct flowside *fromside, *toside; + union inany_addr offender, otap; + char astr[INANY_ADDRSTRLEN]; + char sastr[SOCKADDR_STRLEN]; + const struct in_addr *o4; + in_port_t offender_port; + struct udp_flow *uflow; + uint8_t topif; + size_t dlen; ssize_t rc; rc = recvmsg(s, &mh, MSG_ERRQUEUE); @@ -451,33 +564,102 @@ static int udp_sock_recverr(int s) return -1; } - hdr = CMSG_FIRSTHDR(&mh); - if (!((hdr->cmsg_level == IPPROTO_IP && - hdr->cmsg_type == IP_RECVERR) || - (hdr->cmsg_level == IPPROTO_IPV6 && - hdr->cmsg_type == IPV6_RECVERR))) { - err("Unexpected cmsg reading error queue"); + for (hdr = CMSG_FIRSTHDR(&mh); hdr; hdr = CMSG_NXTHDR(&mh, hdr)) { + if ((hdr->cmsg_level == IPPROTO_IP && + hdr->cmsg_type == IP_RECVERR) || + (hdr->cmsg_level == IPPROTO_IPV6 && + hdr->cmsg_type == IPV6_RECVERR)) + break; + } + + if (!hdr) { + err("Missing RECVERR cmsg in error queue"); return -1; } ee = (const struct sock_extended_err *)CMSG_DATA(hdr); - /* TODO: When possible propagate and otherwise handle errors */ debug("%s error on UDP socket %i: %s", - str_ee_origin(ee), s, strerror(ee->ee_errno)); + str_ee_origin(ee), s, strerror_(ee->ee_errno)); + if (!flow_sidx_valid(sidx)) { + /* No hint from the socket, determine flow from addresses */ + union inany_addr dst; + + if (udp_pktinfo(&mh, &dst) < 0) { + debug("Missing PKTINFO on UDP error"); + return 1; + } + + sidx = flow_lookup_sa(c, IPPROTO_UDP, pif, &src, &dst, port); + if (!flow_sidx_valid(sidx)) { + debug("Ignoring UDP error without flow"); + return 1; + } + } else { + pif = pif_at_sidx(sidx); + } + + uflow = udp_at_sidx(sidx); + ASSERT(uflow); + fromside = &uflow->f.side[sidx.sidei]; + toside = &uflow->f.side[!sidx.sidei]; + topif = uflow->f.pif[!sidx.sidei]; + dlen = rc; + + if (inany_from_sockaddr(&offender, &offender_port, + SO_EE_OFFENDER(ee)) < 0) + goto fail; + + if (pif != PIF_HOST || topif != PIF_TAP) + /* XXX Can we support any other cases? */ + goto fail; + + /* If the offender *is* the endpoint, make sure our translation is + * consistent with the flow's translation. This matters if the flow + * endpoint has a port specific translation (like --dns-match). + */ + if (inany_equals(&offender, &fromside->eaddr)) + otap = toside->oaddr; + else if (!nat_inbound(c, &offender, &otap)) + goto fail; + + if (hdr->cmsg_level == IPPROTO_IP && + (o4 = inany_v4(&otap)) && inany_v4(&toside->eaddr)) { + dlen = MIN(dlen, ICMP4_MAX_DLEN); + udp_send_tap_icmp4(c, ee, toside, *o4, data, dlen); + return 1; + } + + if (hdr->cmsg_level == IPPROTO_IPV6 && !inany_v4(&toside->eaddr)) { + udp_send_tap_icmp6(c, ee, toside, &otap.a6, data, dlen, + FLOW_IDX(uflow)); + return 1; + } + +fail: + flow_dbg(uflow, "Can't propagate %s error from %s %s to %s %s", + str_ee_origin(ee), + pif_name(pif), + sockaddr_ntop(SO_EE_OFFENDER(ee), sastr, sizeof(sastr)), + pif_name(topif), + inany_ntop(&toside->eaddr, astr, sizeof(astr))); return 1; } /** * udp_sock_errs() - Process errors on a socket * @c: Execution context - * @s: Socket to receive from - * @events: epoll events bitmap + * @s: Socket to receive errors from + * @sidx: Flow and side of @s, or FLOW_SIDX_NONE if unknown + * @pif: Interface on which the error occurred + * (only used if @sidx == FLOW_SIDX_NONE) + * @port: Local port number of @s (only used if @sidx == FLOW_SIDX_NONE) * * Return: Number of errors handled, or < 0 if we have an unrecoverable error */ -static int udp_sock_errs(const struct ctx *c, int s, uint32_t events) +static int udp_sock_errs(const struct ctx *c, int s, flow_sidx_t sidx, + uint8_t pif, in_port_t port) { unsigned n_err = 0; socklen_t errlen; @@ -485,11 +667,8 @@ static int udp_sock_errs(const struct ctx *c, int s, uint32_t events) ASSERT(!c->no_udp); - if (!(events & EPOLLERR)) - return 0; /* Nothing to do */ - /* Empty the error queue */ - while ((rc = udp_sock_recverr(s)) > 0) + while ((rc = udp_sock_recverr(c, s, sidx, pif, port)) > 0) n_err += rc; if (rc < 0) @@ -503,7 +682,7 @@ static int udp_sock_errs(const struct ctx *c, int s, uint32_t events) } if (err) { - debug("Unqueued error on UDP socket %i: %s", s, strerror(err)); + debug("Unqueued error on UDP socket %i: %s", s, strerror_(err)); n_err++; } @@ -516,173 +695,262 @@ static int udp_sock_errs(const struct ctx *c, int s, uint32_t events) return n_err; } +/** + * udp_peek_addr() - Get source address for next packet + * @s: Socket to get information from + * @src: Socket address (output) + * @dst: (Local) destination address (output) + * + * Return: 0 if no more packets, 1 on success, -ve error code on error + */ +static int udp_peek_addr(int s, union sockaddr_inany *src, + union inany_addr *dst) +{ + char sastr[SOCKADDR_STRLEN], dstr[INANY_ADDRSTRLEN]; + char cmsg[PKTINFO_SPACE]; + struct msghdr msg = { + .msg_name = src, + .msg_namelen = sizeof(*src), + .msg_control = cmsg, + .msg_controllen = sizeof(cmsg), + }; + int rc; + + rc = recvmsg(s, &msg, MSG_PEEK | MSG_DONTWAIT); + if (rc < 0) { + if (errno == EAGAIN || errno == EWOULDBLOCK) + return 0; + return -errno; + } + + udp_pktinfo(&msg, dst); + + trace("Peeked UDP datagram: %s -> %s", + sockaddr_ntop(src, sastr, sizeof(sastr)), + inany_ntop(dst, dstr, sizeof(dstr))); + + return 1; +} + /** * udp_sock_recv() - Receive datagrams from a socket * @c: Execution context * @s: Socket to receive from - * @events: epoll events bitmap * @mmh mmsghdr array to receive into + * @n: Maximum number of datagrams to receive * * Return: Number of datagrams received * * #syscalls recvmmsg arm:recvmmsg_time64 i686:recvmmsg_time64 */ -static int udp_sock_recv(const struct ctx *c, int s, uint32_t events, - struct mmsghdr *mmh) +static int udp_sock_recv(const struct ctx *c, int s, struct mmsghdr *mmh, int n) { - /* For not entirely clear reasons (data locality?) pasta gets better - * throughput if we receive tap datagrams one at a atime. For small - * splice datagrams throughput is slightly better if we do batch, but - * it's slightly worse for large splice datagrams. Since we don't know - * before we receive whether we'll use tap or splice, always go one at a - * time for pasta mode. - */ - int n = (c->mode == MODE_PASTA ? 1 : UDP_MAX_FRAMES); - ASSERT(!c->no_udp); - if (!(events & EPOLLIN)) - return 0; - n = recvmmsg(s, mmh, n, 0, NULL); if (n < 0) { - err_perror("Error receiving datagrams"); + trace("Error receiving datagrams: %s", strerror_(errno)); + /* Bail out and let the EPOLLERR handler deal with it */ return 0; } return n; } +/** + * udp_sock_to_sock() - Forward datagrams from socket to socket + * @c: Execution context + * @from_s: Socket to receive datagrams from + * @n: Maximum number of datagrams to forward + * @tosidx: Flow & side to forward datagrams to + * + * #syscalls sendmmsg + */ +static void udp_sock_to_sock(const struct ctx *c, int from_s, int n, + flow_sidx_t tosidx) +{ + const struct flowside *toside = flowside_at_sidx(tosidx); + const struct udp_flow *uflow = udp_at_sidx(tosidx); + uint8_t topif = pif_at_sidx(tosidx); + int to_s = uflow->s[tosidx.sidei]; + socklen_t sl; + int i; + + if ((n = udp_sock_recv(c, from_s, udp_mh_recv, n)) <= 0) + return; + + for (i = 0; i < n; i++) { + udp_mh_splice[i].msg_hdr.msg_iov->iov_len + = udp_mh_recv[i].msg_len; + } + + pif_sockaddr(c, &udp_splice_to, &sl, topif, + &toside->eaddr, toside->eport); + + sendmmsg(to_s, udp_mh_splice, n, MSG_NOSIGNAL); +} + +/** + * udp_buf_sock_to_tap() - Forward datagrams from socket to tap + * @c: Execution context + * @s: Socket to read data from + * @n: Maximum number of datagrams to forward + * @tosidx: Flow & side to forward data from @s to + */ +static void udp_buf_sock_to_tap(const struct ctx *c, int s, int n, + flow_sidx_t tosidx) +{ + const struct flowside *toside = flowside_at_sidx(tosidx); + int i; + + if ((n = udp_sock_recv(c, s, udp_mh_recv, n)) <= 0) + return; + + for (i = 0; i < n; i++) + udp_tap_prepare(udp_mh_recv, i, toside, false); + + tap_send_frames(c, &udp_l2_iov[0][0], UDP_NUM_IOVS, n); +} + +/** + * udp_sock_fwd() - Forward datagrams from a possibly unconnected socket + * @c: Execution context + * @s: Socket to forward from + * @frompif: Interface to which @s belongs + * @port: Our (local) port number of @s + * @now: Current timestamp + */ +void udp_sock_fwd(const struct ctx *c, int s, uint8_t frompif, + in_port_t port, const struct timespec *now) +{ + union sockaddr_inany src; + union inany_addr dst; + int rc; + + while ((rc = udp_peek_addr(s, &src, &dst)) != 0) { + bool discard = false; + flow_sidx_t tosidx; + uint8_t topif; + + if (rc < 0) { + trace("Error peeking at socket address: %s", + strerror_(-rc)); + /* Clear errors & carry on */ + if (udp_sock_errs(c, s, FLOW_SIDX_NONE, + frompif, port) < 0) { + err( +"UDP: Unrecoverable error on listening socket: (%s port %hu)", + pif_name(frompif), port); + /* FIXME: what now? close/re-open socket? */ + } + continue; + } + + tosidx = udp_flow_from_sock(c, frompif, &dst, port, &src, now); + topif = pif_at_sidx(tosidx); + + if (pif_is_socket(topif)) { + udp_sock_to_sock(c, s, 1, tosidx); + } else if (topif == PIF_TAP) { + if (c->mode == MODE_VU) + udp_vu_sock_to_tap(c, s, 1, tosidx); + else + udp_buf_sock_to_tap(c, s, 1, tosidx); + } else if (flow_sidx_valid(tosidx)) { + struct udp_flow *uflow = udp_at_sidx(tosidx); + + flow_err(uflow, + "No support for forwarding UDP from %s to %s", + pif_name(frompif), pif_name(topif)); + discard = true; + } else { + debug("Discarding datagram without flow"); + discard = true; + } + + if (discard) { + struct msghdr msg = { 0 }; + + if (recvmsg(s, &msg, MSG_DONTWAIT) < 0) + debug_perror("Failed to discard datagram"); + } + } +} + /** * udp_listen_sock_handler() - Handle new data from socket * @c: Execution context * @ref: epoll reference * @events: epoll events bitmap * @now: Current timestamp - * - * #syscalls recvmmsg */ -void udp_listen_sock_handler(const struct ctx *c, union epoll_ref ref, - uint32_t events, const struct timespec *now) +void udp_listen_sock_handler(const struct ctx *c, + union epoll_ref ref, uint32_t events, + const struct timespec *now) { - const socklen_t sasize = sizeof(udp_meta[0].s_in); - int n, i; - - if (udp_sock_errs(c, ref.fd, events) < 0) { - err("UDP: Unrecoverable error on listening socket:" - " (%s port %hu)", pif_name(ref.udp.pif), ref.udp.port); - /* FIXME: what now? close/re-open socket? */ - return; - } - - if ((n = udp_sock_recv(c, ref.fd, events, udp_mh_recv)) <= 0) - return; - - /* We divide datagrams into batches based on how we need to send them, - * determined by udp_meta[i].tosidx. To avoid either two passes through - * the array, or recalculating tosidx for a single entry, we have to - * populate it one entry *ahead* of the loop counter. - */ - udp_meta[0].tosidx = udp_flow_from_sock(c, ref, &udp_meta[0].s_in, now); - udp_mh_recv[0].msg_hdr.msg_namelen = sasize; - for (i = 0; i < n; ) { - flow_sidx_t batchsidx = udp_meta[i].tosidx; - uint8_t batchpif = pif_at_sidx(batchsidx); - int batchstart = i; - - do { - if (pif_is_socket(batchpif)) { - udp_splice_prepare(udp_mh_recv, i); - } else if (batchpif == PIF_TAP) { - udp_tap_prepare(udp_mh_recv, i, - flowside_at_sidx(batchsidx), - false); - } - - if (++i >= n) - break; - - udp_meta[i].tosidx = udp_flow_from_sock(c, ref, - &udp_meta[i].s_in, - now); - udp_mh_recv[i].msg_hdr.msg_namelen = sasize; - } while (flow_sidx_eq(udp_meta[i].tosidx, batchsidx)); - - if (pif_is_socket(batchpif)) { - udp_splice_send(c, batchstart, i - batchstart, - batchsidx); - } else if (batchpif == PIF_TAP) { - tap_send_frames(c, &udp_l2_iov[batchstart][0], - UDP_NUM_IOVS, i - batchstart); - } else if (flow_sidx_valid(batchsidx)) { - flow_sidx_t fromsidx = flow_sidx_opposite(batchsidx); - struct udp_flow *uflow = udp_at_sidx(batchsidx); - - flow_err(uflow, - "No support for forwarding UDP from %s to %s", - pif_name(pif_at_sidx(fromsidx)), - pif_name(batchpif)); - } else { - debug("Discarding %d datagrams without flow", - i - batchstart); - } - } + if (events & (EPOLLERR | EPOLLIN)) + udp_sock_fwd(c, ref.fd, ref.udp.pif, ref.udp.port, now); } /** - * udp_reply_sock_handler() - Handle new data from flow specific socket + * udp_sock_handler() - Handle new data from flow specific socket * @c: Execution context * @ref: epoll reference * @events: epoll events bitmap * @now: Current timestamp - * - * #syscalls recvmmsg */ -void udp_reply_sock_handler(const struct ctx *c, union epoll_ref ref, - uint32_t events, const struct timespec *now) +void udp_sock_handler(const struct ctx *c, union epoll_ref ref, + uint32_t events, const struct timespec *now) { - flow_sidx_t tosidx = flow_sidx_opposite(ref.flowside); - const struct flowside *toside = flowside_at_sidx(tosidx); struct udp_flow *uflow = udp_at_sidx(ref.flowside); - uint8_t topif = pif_at_sidx(tosidx); - int n, i, from_s; ASSERT(!c->no_udp && uflow); - from_s = uflow->s[ref.flowside.sidei]; - - if (udp_sock_errs(c, from_s, events) < 0) { - flow_err(uflow, "Unrecoverable error on reply socket"); - flow_err_details(uflow); - udp_flow_close(c, uflow); - return; + if (events & EPOLLERR) { + if (udp_sock_errs(c, ref.fd, ref.flowside, PIF_NONE, 0) < 0) { + flow_err(uflow, "Unrecoverable error on flow socket"); + goto fail; + } } - if ((n = udp_sock_recv(c, from_s, events, udp_mh_recv)) <= 0) - return; + if (events & EPOLLIN) { + /* For not entirely clear reasons (data locality?) pasta gets + * better throughput if we receive tap datagrams one at a + * time. For small splice datagrams throughput is slightly + * better if we do batch, but it's slightly worse for large + * splice datagrams. Since we don't know the size before we + * receive, always go one at a time for pasta mode. + */ + size_t n = (c->mode == MODE_PASTA ? 1 : UDP_MAX_FRAMES); + flow_sidx_t tosidx = flow_sidx_opposite(ref.flowside); + uint8_t topif = pif_at_sidx(tosidx); + int s = ref.fd; - flow_trace(uflow, "Received %d datagrams on reply socket", n); - uflow->ts = now->tv_sec; + flow_trace(uflow, "Received data on reply socket"); + uflow->ts = now->tv_sec; - for (i = 0; i < n; i++) { - if (pif_is_socket(topif)) - udp_splice_prepare(udp_mh_recv, i); - else if (topif == PIF_TAP) - udp_tap_prepare(udp_mh_recv, i, toside, false); - /* Restore sockaddr length clobbered by recvmsg() */ - udp_mh_recv[i].msg_hdr.msg_namelen = sizeof(udp_meta[i].s_in); + if (pif_is_socket(topif)) { + udp_sock_to_sock(c, ref.fd, n, tosidx); + } else if (topif == PIF_TAP) { + if (c->mode == MODE_VU) { + udp_vu_sock_to_tap(c, s, UDP_MAX_FRAMES, + tosidx); + } else { + udp_buf_sock_to_tap(c, s, n, tosidx); + } + } else { + flow_err(uflow, + "No support for forwarding UDP from %s to %s", + pif_name(pif_at_sidx(ref.flowside)), + pif_name(topif)); + goto fail; + } } + return; - if (pif_is_socket(topif)) { - udp_splice_send(c, 0, n, tosidx); - } else if (topif == PIF_TAP) { - tap_send_frames(c, &udp_l2_iov[0][0], UDP_NUM_IOVS, n); - } else { - uint8_t frompif = pif_at_sidx(ref.flowside); - - flow_err(uflow, "No support for forwarding UDP from %s to %s", - pif_name(frompif), pif_name(topif)); - } +fail: + flow_err_details(uflow); + udp_flow_close(c, uflow); } /** @@ -692,6 +960,7 @@ void udp_reply_sock_handler(const struct ctx *c, union epoll_ref ref, * @af: Address family, AF_INET or AF_INET6 * @saddr: Source address * @daddr: Destination address + * @ttl: TTL or hop limit for packets to be sent in this call * @p: Pool of UDP packets, with UDP headers * @idx: Index of first packet to process * @now: Current timestamp @@ -702,7 +971,8 @@ void udp_reply_sock_handler(const struct ctx *c, union epoll_ref ref, */ int udp_tap_handler(const struct ctx *c, uint8_t pif, sa_family_t af, const void *saddr, const void *daddr, - const struct pool *p, int idx, const struct timespec *now) + uint8_t ttl, const struct pool *p, int idx, + const struct timespec *now) { const struct flowside *toside; struct mmsghdr mm[UIO_MAXIOV]; @@ -750,7 +1020,7 @@ int udp_tap_handler(const struct ctx *c, uint8_t pif, } toside = flowside_at_sidx(tosidx); - s = udp_at_sidx(tosidx)->s[tosidx.sidei]; + s = uflow->s[tosidx.sidei]; ASSERT(s >= 0); pif_sockaddr(c, &to_sa, &sl, topif, &toside->eaddr, toside->eport); @@ -781,6 +1051,24 @@ int udp_tap_handler(const struct ctx *c, uint8_t pif, mm[i].msg_hdr.msg_controllen = 0; mm[i].msg_hdr.msg_flags = 0; + if (ttl != uflow->ttl[tosidx.sidei]) { + uflow->ttl[tosidx.sidei] = ttl; + if (af == AF_INET) { + if (setsockopt(s, IPPROTO_IP, IP_TTL, + &ttl, sizeof(ttl)) < 0) + flow_perror(uflow, + "setsockopt IP_TTL"); + } else { + /* IPv6 hop_limit cannot be only 1 byte */ + int hop_limit = ttl; + + if (setsockopt(s, SOL_IPV6, IPV6_UNICAST_HOPS, + &hop_limit, sizeof(hop_limit)) < 0) + flow_perror(uflow, + "setsockopt IPV6_UNICAST_HOPS"); + } + } + count++; } diff --git a/udp.h b/udp.h index de2df6d..8f8531a 100644 --- a/udp.h +++ b/udp.h @@ -11,11 +11,12 @@ void udp_portmap_clear(void); void udp_listen_sock_handler(const struct ctx *c, union epoll_ref ref, uint32_t events, const struct timespec *now); -void udp_reply_sock_handler(const struct ctx *c, union epoll_ref ref, - uint32_t events, const struct timespec *now); +void udp_sock_handler(const struct ctx *c, union epoll_ref ref, + uint32_t events, const struct timespec *now); int udp_tap_handler(const struct ctx *c, uint8_t pif, sa_family_t af, const void *saddr, const void *daddr, - const struct pool *p, int idx, const struct timespec *now); + uint8_t ttl, const struct pool *p, int idx, + const struct timespec *now); int udp_sock_init(const struct ctx *c, int ns, const union inany_addr *addr, const char *ifname, in_port_t port); int udp_init(struct ctx *c); diff --git a/udp_flow.c b/udp_flow.c index b81be2c..4c6b3c2 100644 --- a/udp_flow.c +++ b/udp_flow.c @@ -9,10 +9,12 @@ #include <fcntl.h> #include <sys/uio.h> #include <unistd.h> +#include <netinet/udp.h> #include "util.h" #include "passt.h" #include "flow_table.h" +#include "udp_internal.h" #define UDP_CONN_TIMEOUT 180 /* s, timeout for ephemeral or local bind */ @@ -41,127 +43,145 @@ struct udp_flow *udp_at_sidx(flow_sidx_t sidx) */ void udp_flow_close(const struct ctx *c, struct udp_flow *uflow) { + unsigned sidei; + if (uflow->closed) return; /* Nothing to do */ - if (uflow->s[INISIDE] >= 0) { - /* The listening socket needs to stay in epoll */ - close(uflow->s[INISIDE]); - uflow->s[INISIDE] = -1; + flow_foreach_sidei(sidei) { + flow_hash_remove(c, FLOW_SIDX(uflow, sidei)); + if (uflow->s[sidei] >= 0) { + epoll_del(c, uflow->s[sidei]); + close(uflow->s[sidei]); + uflow->s[sidei] = -1; + } } - if (uflow->s[TGTSIDE] >= 0) { - /* But the flow specific one needs to be removed */ - epoll_ctl(c->epollfd, EPOLL_CTL_DEL, uflow->s[TGTSIDE], NULL); - close(uflow->s[TGTSIDE]); - uflow->s[TGTSIDE] = -1; - } - flow_hash_remove(c, FLOW_SIDX(uflow, INISIDE)); - if (!pif_is_socket(uflow->f.pif[TGTSIDE])) - flow_hash_remove(c, FLOW_SIDX(uflow, TGTSIDE)); - uflow->closed = true; } +/** + * udp_flow_sock() - Create, bind and connect a flow specific UDP socket + * @c: Execution context + * @uflow: UDP flow to open socket for + * @sidei: Side of @uflow to open socket for + * + * Return: fd of new socket on success, -ve error code on failure + */ +static int udp_flow_sock(const struct ctx *c, + struct udp_flow *uflow, unsigned sidei) +{ + const struct flowside *side = &uflow->f.side[sidei]; + uint8_t pif = uflow->f.pif[sidei]; + union { + flow_sidx_t sidx; + uint32_t data; + } fref = { .sidx = FLOW_SIDX(uflow, sidei) }; + int s; + + s = flowside_sock_l4(c, EPOLL_TYPE_UDP, pif, side, fref.data); + if (s < 0) { + flow_dbg_perror(uflow, "Couldn't open flow specific socket"); + return s; + } + + if (flowside_connect(c, s, pif, side) < 0) { + int rc = -errno; + + epoll_del(c, s); + close(s); + + flow_dbg_perror(uflow, "Couldn't connect flow socket"); + return rc; + } + + /* It's possible, if unlikely, that we could receive some packets in + * between the bind() and connect() which may or may not be for this + * flow. Being UDP we could just discard them, but it's not ideal. + * + * There's also a tricky case if a bunch of datagrams for a new flow + * arrive in rapid succession, the first going to the original listening + * socket and later ones going to this new socket. If we forwarded the + * datagrams from the new socket immediately here they would go before + * the datagram which established the flow. Again, not strictly wrong + * for UDP, but not ideal. + * + * So, we flag that the new socket is in a transient state where it + * might have datagrams for a different flow queued. Before the next + * epoll cycle, udp_flow_defer() will flush out any such datagrams, and + * thereafter everything on the new socket should be strictly for this + * flow. + */ + if (sidei) + uflow->flush1 = true; + else + uflow->flush0 = true; + + return s; +} + /** * udp_flow_new() - Common setup for a new UDP flow * @c: Execution context * @flow: Initiated flow - * @s_ini: Initiating socket (or -1) * @now: Timestamp * * Return: UDP specific flow, if successful, NULL on failure + * + * #syscalls getsockname */ static flow_sidx_t udp_flow_new(const struct ctx *c, union flow *flow, - int s_ini, const struct timespec *now) + const struct timespec *now) { - const struct flowside *ini = &flow->f.side[INISIDE]; struct udp_flow *uflow = NULL; const struct flowside *tgt; - uint8_t tgtpif; - - if (!inany_is_unicast(&ini->eaddr) || ini->eport == 0) { - flow_trace(flow, "Invalid endpoint to initiate UDP flow"); - goto cancel; - } + unsigned sidei; if (!(tgt = flow_target(c, flow, IPPROTO_UDP))) goto cancel; - tgtpif = flow->f.pif[TGTSIDE]; uflow = FLOW_SET_TYPE(flow, FLOW_UDP, udp); uflow->ts = now->tv_sec; uflow->s[INISIDE] = uflow->s[TGTSIDE] = -1; + uflow->ttl[INISIDE] = uflow->ttl[TGTSIDE] = 0; - if (s_ini >= 0) { - /* When using auto port-scanning the listening port could go - * away, so we need to duplicate the socket - */ - uflow->s[INISIDE] = fcntl(s_ini, F_DUPFD_CLOEXEC, 0); - if (uflow->s[INISIDE] < 0) { - flow_err(uflow, - "Couldn't duplicate listening socket: %s", - strerror(errno)); + flow_foreach_sidei(sidei) { + if (pif_is_socket(uflow->f.pif[sidei])) + if ((uflow->s[sidei] = udp_flow_sock(c, uflow, sidei)) < 0) + goto cancel; + } + + if (uflow->s[TGTSIDE] >= 0 && inany_is_unspecified(&tgt->oaddr)) { + /* When we target a socket, we connect() it, but might not + * always bind(), leaving the kernel to pick our address. In + * that case connect() will implicitly bind() the socket, but we + * need to determine its local address so that we can match + * reply packets back to the correct flow. Update the flow with + * the information from getsockname() */ + union sockaddr_inany sa; + socklen_t sl = sizeof(sa); + in_port_t port; + + if (getsockname(uflow->s[TGTSIDE], &sa.sa, &sl) < 0 || + inany_from_sockaddr(&uflow->f.side[TGTSIDE].oaddr, + &port, &sa) < 0) { + flow_perror(uflow, "Unable to determine local address"); + goto cancel; + } + if (port != tgt->oport) { + flow_err(uflow, "Unexpected local port"); goto cancel; } } - if (pif_is_socket(tgtpif)) { - struct mmsghdr discard[UIO_MAXIOV] = { 0 }; - union { - flow_sidx_t sidx; - uint32_t data; - } fref = { - .sidx = FLOW_SIDX(flow, TGTSIDE), - }; - int rc; - - uflow->s[TGTSIDE] = flowside_sock_l4(c, EPOLL_TYPE_UDP_REPLY, - tgtpif, tgt, fref.data); - if (uflow->s[TGTSIDE] < 0) { - flow_dbg(uflow, - "Couldn't open socket for spliced flow: %s", - strerror(errno)); - goto cancel; - } - - if (flowside_connect(c, uflow->s[TGTSIDE], tgtpif, tgt) < 0) { - flow_dbg(uflow, - "Couldn't connect flow socket: %s", - strerror(errno)); - goto cancel; - } - - /* It's possible, if unlikely, that we could receive some - * unrelated packets in between the bind() and connect() of this - * socket. For now we just discard these. We could consider - * trying to redirect these to an appropriate handler, if we - * need to. - */ - rc = recvmmsg(uflow->s[TGTSIDE], discard, ARRAY_SIZE(discard), - MSG_DONTWAIT, NULL); - if (rc >= ARRAY_SIZE(discard)) { - flow_dbg(uflow, - "Too many (%d) spurious reply datagrams", rc); - goto cancel; - } else if (rc > 0) { - flow_trace(uflow, - "Discarded %d spurious reply datagrams", rc); - } else if (errno != EAGAIN) { - flow_err(uflow, - "Unexpected error discarding datagrams: %s", - strerror(errno)); - } - } - - flow_hash_insert(c, FLOW_SIDX(uflow, INISIDE)); - - /* If the target side is a socket, it will be a reply socket that knows - * its own flowside. But if it's tap, then we need to look it up by - * hash. + /* Tap sides always need to be looked up by hash. Socket sides don't + * always, but sometimes do (receiving packets on a socket not specific + * to one flow). Unconditionally hash both sides so all our bases are + * covered */ - if (!pif_is_socket(tgtpif)) - flow_hash_insert(c, FLOW_SIDX(uflow, TGTSIDE)); + flow_foreach_sidei(sidei) + flow_hash_insert(c, FLOW_SIDX(uflow, sidei)); + FLOW_ACTIVATE(uflow); return FLOW_SIDX(uflow, TGTSIDE); @@ -174,28 +194,30 @@ cancel: } /** - * udp_flow_from_sock() - Find or create UDP flow for "listening" socket + * udp_flow_from_sock() - Find or create UDP flow for incoming datagram * @c: Execution context - * @ref: epoll reference of the receiving socket + * @pif: Interface the datagram is arriving from + * @dst: Our (local) address to which the datagram is arriving + * @port: Our (local) port number to which the datagram is arriving * @s_in: Source socket address, filled in by recvmmsg() * @now: Timestamp * - * #syscalls fcntl arm:fcntl64 ppc64:fcntl64 i686:fcntl64 + * #syscalls fcntl arm:fcntl64 ppc64:fcntl64|fcntl i686:fcntl64 * * Return: sidx for the destination side of the flow for this packet, or * FLOW_SIDX_NONE if we couldn't find or create a flow. */ -flow_sidx_t udp_flow_from_sock(const struct ctx *c, union epoll_ref ref, +flow_sidx_t udp_flow_from_sock(const struct ctx *c, uint8_t pif, + const union inany_addr *dst, in_port_t port, const union sockaddr_inany *s_in, const struct timespec *now) { + const struct flowside *ini; struct udp_flow *uflow; union flow *flow; flow_sidx_t sidx; - ASSERT(ref.type == EPOLL_TYPE_UDP_LISTEN); - - sidx = flow_lookup_sa(c, IPPROTO_UDP, ref.udp.pif, s_in, ref.udp.port); + sidx = flow_lookup_sa(c, IPPROTO_UDP, pif, s_in, dst, port); if ((uflow = udp_at_sidx(sidx))) { uflow->ts = now->tv_sec; return flow_sidx_opposite(sidx); @@ -205,13 +227,24 @@ flow_sidx_t udp_flow_from_sock(const struct ctx *c, union epoll_ref ref, char sastr[SOCKADDR_STRLEN]; debug("Couldn't allocate flow for UDP datagram from %s %s", - pif_name(ref.udp.pif), - sockaddr_ntop(s_in, sastr, sizeof(sastr))); + pif_name(pif), sockaddr_ntop(s_in, sastr, sizeof(sastr))); return FLOW_SIDX_NONE; } - flow_initiate_sa(flow, ref.udp.pif, s_in, ref.udp.port); - return udp_flow_new(c, flow, ref.fd, now); + ini = flow_initiate_sa(flow, pif, s_in, dst, port); + + if (!inany_is_unicast(&ini->eaddr) || + ini->eport == 0 || ini->oport == 0) { + /* In principle ini->oddr also must be specified, but when we've + * been initiated from a socket bound to 0.0.0.0 or ::, we don't + * know our address, so we have to leave it unpopulated. + */ + flow_err(flow, "Invalid endpoint on UDP recvfrom()"); + flow_alloc_cancel(flow); + return FLOW_SIDX_NONE; + } + + return udp_flow_new(c, flow, now); } /** @@ -233,6 +266,7 @@ flow_sidx_t udp_flow_from_tap(const struct ctx *c, in_port_t srcport, in_port_t dstport, const struct timespec *now) { + const struct flowside *ini; struct udp_flow *uflow; union flow *flow; flow_sidx_t sidx; @@ -256,19 +290,55 @@ flow_sidx_t udp_flow_from_tap(const struct ctx *c, return FLOW_SIDX_NONE; } - flow_initiate_af(flow, PIF_TAP, af, saddr, srcport, daddr, dstport); + ini = flow_initiate_af(flow, PIF_TAP, af, saddr, srcport, + daddr, dstport); - return udp_flow_new(c, flow, -1, now); + if (inany_is_unspecified(&ini->eaddr) || ini->eport == 0 || + inany_is_unspecified(&ini->oaddr) || ini->oport == 0) { + flow_dbg(flow, "Invalid endpoint on UDP packet"); + flow_alloc_cancel(flow); + return FLOW_SIDX_NONE; + } + + return udp_flow_new(c, flow, now); +} + +/** + * udp_flush_flow() - Flush datagrams that might not be for this flow + * @c: Execution context + * @uflow: Flow to handle + * @sidei: Side of the flow to flush + * @now: Current timestamp + */ +static void udp_flush_flow(const struct ctx *c, + const struct udp_flow *uflow, unsigned sidei, + const struct timespec *now) +{ + /* We don't know exactly where the datagrams will come from, but we know + * they'll have an interface and oport matching this flow */ + udp_sock_fwd(c, uflow->s[sidei], uflow->f.pif[sidei], + uflow->f.side[sidei].oport, now); } /** * udp_flow_defer() - Deferred per-flow handling (clean up aborted flows) + * @c: Execution context * @uflow: Flow to handle + * @now: Current timestamp * * Return: true if the connection is ready to free, false otherwise */ -bool udp_flow_defer(const struct udp_flow *uflow) +bool udp_flow_defer(const struct ctx *c, struct udp_flow *uflow, + const struct timespec *now) { + if (uflow->flush0) { + udp_flush_flow(c, uflow, INISIDE, now); + uflow->flush0 = false; + } + if (uflow->flush1) { + udp_flush_flow(c, uflow, TGTSIDE, now); + uflow->flush1 = false; + } return uflow->closed; } diff --git a/udp_flow.h b/udp_flow.h index 9a1b059..4c528e9 100644 --- a/udp_flow.h +++ b/udp_flow.h @@ -8,9 +8,12 @@ #define UDP_FLOW_H /** - * struct udp - Descriptor for a flow of UDP packets + * struct udp_flow - Descriptor for a flow of UDP packets * @f: Generic flow information + * @ttl: TTL or hop_limit for both sides * @closed: Flow is already closed + * @flush0: @s[0] may have datagrams queued for other flows + * @flush1: @s[1] may have datagrams queued for other flows * @ts: Activity timestamp * @s: Socket fd (or -1) for each side of the flow */ @@ -18,13 +21,19 @@ struct udp_flow { /* Must be first element */ struct flow_common f; - bool closed :1; + uint8_t ttl[SIDES]; + + bool closed :1, + flush0 :1, + flush1 :1; + time_t ts; int s[SIDES]; }; struct udp_flow *udp_at_sidx(flow_sidx_t sidx); -flow_sidx_t udp_flow_from_sock(const struct ctx *c, union epoll_ref ref, +flow_sidx_t udp_flow_from_sock(const struct ctx *c, uint8_t pif, + const union inany_addr *dst, in_port_t port, const union sockaddr_inany *s_in, const struct timespec *now); flow_sidx_t udp_flow_from_tap(const struct ctx *c, @@ -33,7 +42,8 @@ flow_sidx_t udp_flow_from_tap(const struct ctx *c, in_port_t srcport, in_port_t dstport, const struct timespec *now); void udp_flow_close(const struct ctx *c, struct udp_flow *uflow); -bool udp_flow_defer(const struct udp_flow *uflow); +bool udp_flow_defer(const struct ctx *c, struct udp_flow *uflow, + const struct timespec *now); bool udp_flow_timer(const struct ctx *c, struct udp_flow *uflow, const struct timespec *now); diff --git a/udp_internal.h b/udp_internal.h new file mode 100644 index 0000000..96d11cf --- /dev/null +++ b/udp_internal.h @@ -0,0 +1,34 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later + * Copyright (c) 2021 Red Hat GmbH + * Author: Stefano Brivio <sbrivio@redhat.com> + */ + +#ifndef UDP_INTERNAL_H +#define UDP_INTERNAL_H + +#include "tap.h" /* needed by udp_meta_t */ + +/** + * struct udp_payload_t - UDP header and data for inbound messages + * @uh: UDP header + * @data: UDP data + */ +struct udp_payload_t { + struct udphdr uh; + char data[USHRT_MAX - sizeof(struct udphdr)]; +#ifdef __AVX2__ +} __attribute__ ((packed, aligned(32))); +#else +} __attribute__ ((packed, aligned(__alignof__(unsigned int)))); +#endif + +size_t udp_update_hdr4(struct iphdr *ip4h, struct udp_payload_t *bp, + const struct flowside *toside, size_t dlen, + bool no_udp_csum); +size_t udp_update_hdr6(struct ipv6hdr *ip6h, struct udp_payload_t *bp, + const struct flowside *toside, size_t dlen, + bool no_udp_csum); +void udp_sock_fwd(const struct ctx *c, int s, uint8_t frompif, + in_port_t port, const struct timespec *now); + +#endif /* UDP_INTERNAL_H */ diff --git a/udp_vu.c b/udp_vu.c new file mode 100644 index 0000000..1f89509 --- /dev/null +++ b/udp_vu.c @@ -0,0 +1,225 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* udp_vu.c - UDP L2 vhost-user management functions + * + * Copyright Red Hat + * Author: Laurent Vivier <lvivier@redhat.com> + */ + +#include <unistd.h> +#include <assert.h> +#include <net/ethernet.h> +#include <net/if.h> +#include <netinet/in.h> +#include <netinet/ip.h> +#include <netinet/udp.h> +#include <stdint.h> +#include <stddef.h> +#include <sys/uio.h> +#include <linux/virtio_net.h> + +#include "checksum.h" +#include "util.h" +#include "ip.h" +#include "siphash.h" +#include "inany.h" +#include "passt.h" +#include "pcap.h" +#include "log.h" +#include "vhost_user.h" +#include "udp_internal.h" +#include "flow.h" +#include "flow_table.h" +#include "udp_flow.h" +#include "udp_vu.h" +#include "vu_common.h" + +static struct iovec iov_vu [VIRTQUEUE_MAX_SIZE]; +static struct vu_virtq_element elem [VIRTQUEUE_MAX_SIZE]; + +/** + * udp_vu_hdrlen() - return the size of the header in level 2 frame (UDP) + * @v6: Set for IPv6 packet + * + * Return: Return the size of the header + */ +static size_t udp_vu_hdrlen(bool v6) +{ + size_t hdrlen; + + hdrlen = sizeof(struct virtio_net_hdr_mrg_rxbuf) + + sizeof(struct ethhdr) + sizeof(struct udphdr); + + if (v6) + hdrlen += sizeof(struct ipv6hdr); + else + hdrlen += sizeof(struct iphdr); + + return hdrlen; +} + +/** + * udp_vu_sock_recv() - Receive datagrams from socket into vhost-user buffers + * @c: Execution context + * @s: Socket to receive from + * @v6: Set for IPv6 connections + * @dlen: Size of received data (output) + * + * Return: Number of iov entries used to store the datagram + */ +static int udp_vu_sock_recv(const struct ctx *c, int s, bool v6, ssize_t *dlen) +{ + struct vu_dev *vdev = c->vdev; + struct vu_virtq *vq = &vdev->vq[VHOST_USER_RX_QUEUE]; + int iov_cnt, idx, iov_used; + struct msghdr msg = { 0 }; + size_t off, hdrlen; + + ASSERT(!c->no_udp); + + /* compute L2 header length */ + hdrlen = udp_vu_hdrlen(v6); + + vu_init_elem(elem, iov_vu, VIRTQUEUE_MAX_SIZE); + + iov_cnt = vu_collect(vdev, vq, elem, VIRTQUEUE_MAX_SIZE, + IP_MAX_MTU + ETH_HLEN + + sizeof(struct virtio_net_hdr_mrg_rxbuf), + NULL); + if (iov_cnt == 0) + return 0; + + /* reserve space for the headers */ + ASSERT(iov_vu[0].iov_len >= hdrlen); + iov_vu[0].iov_base = (char *)iov_vu[0].iov_base + hdrlen; + iov_vu[0].iov_len -= hdrlen; + + /* read data from the socket */ + msg.msg_iov = iov_vu; + msg.msg_iovlen = iov_cnt; + + *dlen = recvmsg(s, &msg, 0); + if (*dlen < 0) { + vu_queue_rewind(vq, iov_cnt); + return 0; + } + + /* restore the pointer to the headers address */ + iov_vu[0].iov_base = (char *)iov_vu[0].iov_base - hdrlen; + iov_vu[0].iov_len += hdrlen; + + /* count the numbers of buffer filled by recvmsg() */ + idx = iov_skip_bytes(iov_vu, iov_cnt, *dlen + hdrlen, &off); + + /* adjust last iov length */ + if (idx < iov_cnt) + iov_vu[idx].iov_len = off; + iov_used = idx + !!off; + + vu_set_vnethdr(vdev, iov_vu[0].iov_base, iov_used); + + /* release unused buffers */ + vu_queue_rewind(vq, iov_cnt - iov_used); + + return iov_used; +} + +/** + * udp_vu_prepare() - Prepare the packet header + * @c: Execution context + * @toside: Address information for one side of the flow + * @dlen: Packet data length + * + * Return: Layer-4 length + */ +static size_t udp_vu_prepare(const struct ctx *c, + const struct flowside *toside, ssize_t dlen) +{ + struct ethhdr *eh; + size_t l4len; + + /* ethernet header */ + eh = vu_eth(iov_vu[0].iov_base); + + memcpy(eh->h_dest, c->guest_mac, sizeof(eh->h_dest)); + memcpy(eh->h_source, c->our_tap_mac, sizeof(eh->h_source)); + + /* initialize header */ + if (inany_v4(&toside->eaddr) && inany_v4(&toside->oaddr)) { + struct iphdr *iph = vu_ip(iov_vu[0].iov_base); + struct udp_payload_t *bp = vu_payloadv4(iov_vu[0].iov_base); + + eh->h_proto = htons(ETH_P_IP); + + *iph = (struct iphdr)L2_BUF_IP4_INIT(IPPROTO_UDP); + + l4len = udp_update_hdr4(iph, bp, toside, dlen, true); + } else { + struct ipv6hdr *ip6h = vu_ip(iov_vu[0].iov_base); + struct udp_payload_t *bp = vu_payloadv6(iov_vu[0].iov_base); + + eh->h_proto = htons(ETH_P_IPV6); + + *ip6h = (struct ipv6hdr)L2_BUF_IP6_INIT(IPPROTO_UDP); + + l4len = udp_update_hdr6(ip6h, bp, toside, dlen, true); + } + + return l4len; +} + +/** + * udp_vu_csum() - Calculate and set checksum for a UDP packet + * @toside: Address information for one side of the flow + * @iov_used: Number of used iov_vu items + */ +static void udp_vu_csum(const struct flowside *toside, int iov_used) +{ + const struct in_addr *src4 = inany_v4(&toside->oaddr); + const struct in_addr *dst4 = inany_v4(&toside->eaddr); + char *base = iov_vu[0].iov_base; + struct udp_payload_t *bp; + struct iov_tail data; + + if (src4 && dst4) { + bp = vu_payloadv4(base); + data = IOV_TAIL(iov_vu, iov_used, (char *)&bp->data - base); + csum_udp4(&bp->uh, *src4, *dst4, &data); + } else { + bp = vu_payloadv6(base); + data = IOV_TAIL(iov_vu, iov_used, (char *)&bp->data - base); + csum_udp6(&bp->uh, &toside->oaddr.a6, &toside->eaddr.a6, &data); + } +} + +/** + * udp_vu_sock_to_tap() - Forward datagrams from socket to tap + * @c: Execution context + * @s: Socket to read data from + * @n: Maximum number of datagrams to forward + * @tosidx: Flow & side to forward data from @s to + */ +void udp_vu_sock_to_tap(const struct ctx *c, int s, int n, flow_sidx_t tosidx) +{ + const struct flowside *toside = flowside_at_sidx(tosidx); + bool v6 = !(inany_v4(&toside->eaddr) && inany_v4(&toside->oaddr)); + struct vu_dev *vdev = c->vdev; + struct vu_virtq *vq = &vdev->vq[VHOST_USER_RX_QUEUE]; + int i; + + for (i = 0; i < n; i++) { + ssize_t dlen; + int iov_used; + + iov_used = udp_vu_sock_recv(c, s, v6, &dlen); + if (iov_used <= 0) + break; + + udp_vu_prepare(c, toside, dlen); + if (*c->pcap) { + udp_vu_csum(toside, iov_used); + pcap_iov(iov_vu, iov_used, + sizeof(struct virtio_net_hdr_mrg_rxbuf)); + } + vu_flush(vdev, vq, elem, iov_used); + } +} diff --git a/udp_vu.h b/udp_vu.h new file mode 100644 index 0000000..576b0e7 --- /dev/null +++ b/udp_vu.h @@ -0,0 +1,13 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* Copyright Red Hat + * Author: Laurent Vivier <lvivier@redhat.com> + */ + +#ifndef UDP_VU_H +#define UDP_VU_H + +void udp_vu_listen_sock_data(const struct ctx *c, union epoll_ref ref, + const struct timespec *now); +void udp_vu_sock_to_tap(const struct ctx *c, int s, int n, flow_sidx_t tosidx); + +#endif /* UDP_VU_H */ diff --git a/util.c b/util.c index 55cae3f..f5497d4 100644 --- a/util.c +++ b/util.c @@ -71,7 +71,7 @@ int sock_l4_sa(const struct ctx *c, enum epoll_type type, case EPOLL_TYPE_UDP_LISTEN: freebind = c->freebind; /* fallthrough */ - case EPOLL_TYPE_UDP_REPLY: + case EPOLL_TYPE_UDP: proto = IPPROTO_UDP; socktype = SOCK_DGRAM | SOCK_NONBLOCK; break; @@ -90,7 +90,7 @@ int sock_l4_sa(const struct ctx *c, enum epoll_type type, ret = -errno; if (fd < 0) { - warn("L4 socket: %s", strerror(-ret)); + warn("L4 socket: %s", strerror_(-ret)); return ret; } @@ -109,11 +109,15 @@ int sock_l4_sa(const struct ctx *c, enum epoll_type type, debug("Failed to set SO_REUSEADDR on socket %i", fd); if (proto == IPPROTO_UDP) { + int pktinfo = af == AF_INET ? IP_PKTINFO : IPV6_RECVPKTINFO; + int recverr = af == AF_INET ? IP_RECVERR : IPV6_RECVERR; int level = af == AF_INET ? IPPROTO_IP : IPPROTO_IPV6; - int opt = af == AF_INET ? IP_RECVERR : IPV6_RECVERR; - if (setsockopt(fd, level, opt, &y, sizeof(y))) + if (setsockopt(fd, level, recverr, &y, sizeof(y))) die_perror("Failed to set RECVERR on socket %i", fd); + + if (setsockopt(fd, level, pktinfo, &y, sizeof(y))) + die_perror("Failed to set PKTINFO on socket %i", fd); } if (ifname && *ifname) { @@ -162,7 +166,7 @@ int sock_l4_sa(const struct ctx *c, enum epoll_type type, if (type == EPOLL_TYPE_TCP_LISTEN && listen(fd, 128) < 0) { ret = -errno; - warn("TCP socket listen: %s", strerror(-ret)); + warn("TCP socket listen: %s", strerror_(-ret)); close(fd); return ret; } @@ -171,13 +175,75 @@ int sock_l4_sa(const struct ctx *c, enum epoll_type type, ev.data.u64 = ref.u64; if (epoll_ctl(c->epollfd, EPOLL_CTL_ADD, fd, &ev) == -1) { ret = -errno; - warn("L4 epoll_ctl: %s", strerror(-ret)); + warn("L4 epoll_ctl: %s", strerror_(-ret)); return ret; } return fd; } +/** + * sock_unix() - Create and bind AF_UNIX socket + * @sock_path: Socket path. If empty, set on return (UNIX_SOCK_PATH as prefix) + * + * Return: socket descriptor on success, won't return on failure + */ +int sock_unix(char *sock_path) +{ + int fd = socket(AF_UNIX, SOCK_STREAM | SOCK_CLOEXEC, 0); + struct sockaddr_un addr = { + .sun_family = AF_UNIX, + }; + int i; + + if (fd < 0) + die_perror("Failed to open UNIX domain socket"); + + for (i = 1; i < UNIX_SOCK_MAX; i++) { + char *path = addr.sun_path; + int ex, ret; + + if (*sock_path) + memcpy(path, sock_path, UNIX_PATH_MAX); + else if (snprintf_check(path, UNIX_PATH_MAX - 1, + UNIX_SOCK_PATH, i)) + die_perror("Can't build UNIX domain socket path"); + + ex = socket(AF_UNIX, SOCK_STREAM | SOCK_NONBLOCK | SOCK_CLOEXEC, + 0); + if (ex < 0) + die_perror("Failed to check for UNIX domain conflicts"); + + ret = connect(ex, (const struct sockaddr *)&addr, sizeof(addr)); + if (!ret || (errno != ENOENT && errno != ECONNREFUSED && + errno != EACCES)) { + if (*sock_path) + die("Socket path %s already in use", path); + + close(ex); + continue; + } + close(ex); + + unlink(path); + ret = bind(fd, (const struct sockaddr *)&addr, sizeof(addr)); + if (*sock_path && ret) + die_perror("Failed to bind UNIX domain socket"); + + if (!ret) + break; + } + + if (i == UNIX_SOCK_MAX) + die_perror("Failed to bind UNIX domain socket"); + + info("UNIX domain socket bound at %s", addr.sun_path); + if (!*sock_path) + memcpy(sock_path, addr.sun_path, UNIX_PATH_MAX); + + return fd; +} + /** * sock_probe_mem() - Check if setting high SO_SNDBUF and SO_RCVBUF is allowed * @c: Execution context @@ -405,7 +471,7 @@ void pidfile_write(int fd, pid_t pid) if (write(fd, pid_buf, n) < 0) { perror("PID file write"); - exit(EXIT_FAILURE); + _exit(EXIT_FAILURE); } close(fd); @@ -441,12 +507,12 @@ int __daemon(int pidfile_fd, int devnull_fd) if (pid == -1) { perror("fork"); - exit(EXIT_FAILURE); + _exit(EXIT_FAILURE); } if (pid) { pidfile_write(pidfile_fd, pid); - exit(EXIT_SUCCESS); + _exit(EXIT_SUCCESS); } if (setsid() < 0 || @@ -454,7 +520,7 @@ int __daemon(int pidfile_fd, int devnull_fd) dup2(devnull_fd, STDOUT_FILENO) < 0 || dup2(devnull_fd, STDERR_FILENO) < 0 || close(devnull_fd)) - exit(EXIT_FAILURE); + _exit(EXIT_FAILURE); return 0; } @@ -606,6 +672,90 @@ int write_remainder(int fd, const struct iovec *iov, size_t iovcnt, size_t skip) return 0; } +/** + * read_all_buf() - Fill a whole buffer from a file descriptor + * @fd: File descriptor + * @buf: Pointer to base of buffer + * @len: Length of buffer + * + * Return: 0 on success, -1 on error (with errno set) + * + * #syscalls read + */ +int read_all_buf(int fd, void *buf, size_t len) +{ + size_t left = len; + char *p = buf; + + while (left) { + ssize_t rc; + + ASSERT(left <= len); + + do + rc = read(fd, p, left); + while ((rc < 0) && errno == EINTR); + + if (rc < 0) + return -1; + + if (rc == 0) { + errno = ENODATA; + return -1; + } + + p += rc; + left -= rc; + } + return 0; +} + +/** + * read_remainder() - Read the tail of an IO vector from a file descriptor + * @fd: File descriptor + * @iov: IO vector + * @cnt: Number of entries in @iov + * @skip: Number of bytes of the vector to skip reading + * + * Return: 0 on success, -1 on error (with errno set) + * + * Note: mode-specific seccomp profiles need to enable readv() to use this. + */ +/* cppcheck-suppress unusedFunction */ +int read_remainder(int fd, const struct iovec *iov, size_t cnt, size_t skip) +{ + size_t i = 0, offset; + + while ((i += iov_skip_bytes(iov + i, cnt - i, skip, &offset)) < cnt) { + ssize_t rc; + + if (offset) { + ASSERT(offset < iov[i].iov_len); + /* Read the remainder of the partially read buffer */ + if (read_all_buf(fd, (char *)iov[i].iov_base + offset, + iov[i].iov_len - offset) < 0) + return -1; + i++; + } + + if (cnt == i) + break; + + /* Fill as many of the remaining buffers as we can */ + rc = readv(fd, &iov[i], cnt - i); + if (rc < 0) + return -1; + + if (rc == 0) { + errno = ENODATA; + return -1; + } + + skip = rc; + } + return 0; +} + /** sockaddr_ntop() - Convert a socket address to text format * @sa: Socket address * @dst: output buffer, minimum SOCKADDR_STRLEN bytes @@ -725,7 +875,9 @@ void close_open_files(int argc, char **argv) errno = 0; fd = strtol(optarg, NULL, 0); - if (errno || fd <= STDERR_FILENO || fd > INT_MAX) + if (errno || + (fd != STDIN_FILENO && fd <= STDERR_FILENO) || + fd > INT_MAX) die("Invalid --fd: %s", optarg); } } while (name != -1); @@ -837,3 +989,56 @@ void raw_random(void *buf, size_t buflen) if (random_read < buflen) die("Unexpected EOF on random data source"); } + +/** + * epoll_del() - Remove a file descriptor from our passt epoll + * @c: Execution context + * @fd: File descriptor to remove + */ +void epoll_del(const struct ctx *c, int fd) +{ + epoll_ctl(c->epollfd, EPOLL_CTL_DEL, fd, NULL); + +} + +/** + * encode_domain_name() - Encode domain name according to RFC 1035, section 3.1 + * @buf: Buffer to fill in with encoded domain name + * @domain_name: Input domain name string with terminator + * + * The buffer's 'buf' size has to be >= strlen(domain_name) + 2 + */ +void encode_domain_name(char *buf, const char *domain_name) +{ + size_t i; + char *p; + + buf[0] = strcspn(domain_name, "."); + p = buf + 1; + for (i = 0; domain_name[i]; i++) { + if (domain_name[i] == '.') + p[i] = strcspn(domain_name + i + 1, "."); + else + p[i] = domain_name[i]; + } + p[i] = 0L; +} + +/** + * abort_with_msg() - Print error message and abort + * @fmt: Format string + * @...: Format parameters + */ +void abort_with_msg(const char *fmt, ...) +{ + va_list ap; + + va_start(ap, fmt); + vlogmsg(true, false, LOG_CRIT, fmt, ap); + va_end(ap); + + /* This may actually cause a SIGSYS instead of SIGABRT, due to seccomp, + * but that will still get the job done. + */ + abort(); +} diff --git a/util.h b/util.h index 90428c4..5947337 100644 --- a/util.h +++ b/util.h @@ -31,12 +31,6 @@ #ifndef SECCOMP_RET_KILL_PROCESS #define SECCOMP_RET_KILL_PROCESS SECCOMP_RET_KILL #endif -#ifndef ETH_MAX_MTU -#define ETH_MAX_MTU USHRT_MAX -#endif -#ifndef ETH_MIN_MTU -#define ETH_MIN_MTU 68 -#endif #ifndef IP_MAX_MTU #define IP_MAX_MTU USHRT_MAX #endif @@ -67,27 +61,22 @@ #define STRINGIFY(x) #x #define STR(x) STRINGIFY(x) -#ifdef CPPCHECK_6936 +void abort_with_msg(const char *fmt, ...) + __attribute__((format(printf, 1, 2), noreturn)); + /* Some cppcheck versions get confused by aborts inside a loop, causing * it to give false positive uninitialised variable warnings later in * the function, because it doesn't realise the non-initialising path * already exited. See https://trac.cppcheck.net/ticket/13227 + * + * Therefore, avoid using the usual do while wrapper we use to force the macro + * to act like a single statement requiring a ';'. */ -#define ASSERT(expr) \ - ((expr) ? (void)0 : abort()) -#else +#define ASSERT_WITH_MSG(expr, ...) \ + ((expr) ? (void)0 : abort_with_msg(__VA_ARGS__)) #define ASSERT(expr) \ - do { \ - if (!(expr)) { \ - err("ASSERTION FAILED in %s (%s:%d): %s", \ - __func__, __FILE__, __LINE__, STRINGIFY(expr)); \ - /* This may actually SIGSYS, due to seccomp, \ - * but that will still get the job done \ - */ \ - abort(); \ - } \ - } while (0) -#endif + ASSERT_WITH_MSG((expr), "ASSERTION FAILED in %s (%s:%d): %s", \ + __func__, __FILE__, __LINE__, STRINGIFY(expr)) #ifdef P_tmpdir #define TMPDIR P_tmpdir @@ -122,14 +111,43 @@ (((x) & 0x0000ff00) << 8) | (((x) & 0x000000ff) << 24)) #endif +#ifndef __bswap_constant_32 +#define __bswap_constant_32(x) \ + ((((x) & 0xff000000) >> 24) | (((x) & 0x00ff0000) >> 8) | \ + (((x) & 0x0000ff00) << 8) | (((x) & 0x000000ff) << 24)) +#endif + +#ifndef __bswap_constant_64 +#define __bswap_constant_64(x) \ + ((((x) & 0xff00000000000000ULL) >> 56) | \ + (((x) & 0x00ff000000000000ULL) >> 40) | \ + (((x) & 0x0000ff0000000000ULL) >> 24) | \ + (((x) & 0x000000ff00000000ULL) >> 8) | \ + (((x) & 0x00000000ff000000ULL) << 8) | \ + (((x) & 0x0000000000ff0000ULL) << 24) | \ + (((x) & 0x000000000000ff00ULL) << 40) | \ + (((x) & 0x00000000000000ffULL) << 56)) +#endif + #if __BYTE_ORDER == __BIG_ENDIAN #define htons_constant(x) (x) #define htonl_constant(x) (x) +#define htonll_constant(x) (x) +#define ntohs_constant(x) (x) +#define ntohl_constant(x) (x) +#define ntohll_constant(x) (x) #else #define htons_constant(x) (__bswap_constant_16(x)) #define htonl_constant(x) (__bswap_constant_32(x)) +#define htonll_constant(x) (__bswap_constant_64(x)) +#define ntohs_constant(x) (__bswap_constant_16(x)) +#define ntohl_constant(x) (__bswap_constant_32(x)) +#define ntohll_constant(x) (__bswap_constant_64(x)) #endif +#define ntohll(x) (be64toh((x))) +#define htonll(x) (htobe64((x))) + /** * ntohl_unaligned() - Read 32-bit BE value from a possibly unaligned address * @p: Pointer to the BE value in memory @@ -144,7 +162,19 @@ static inline uint32_t ntohl_unaligned(const void *p) return ntohl(val); } +static inline void barrier(void) { __asm__ __volatile__("" ::: "memory"); } +#define smp_mb() do { barrier(); __atomic_thread_fence(__ATOMIC_SEQ_CST); } while (0) +#define smp_mb_release() do { barrier(); __atomic_thread_fence(__ATOMIC_RELEASE); } while (0) +#define smp_mb_acquire() do { barrier(); __atomic_thread_fence(__ATOMIC_ACQUIRE); } while (0) + +#define smp_wmb() smp_mb_release() +#define smp_rmb() smp_mb_acquire() + +#define qatomic_or(ptr, n) \ + ((void) __atomic_fetch_or(ptr, n, __ATOMIC_SEQ_CST)) + #define NS_FN_STACK_SIZE (1024 * 1024) /* 1MiB */ + int do_clone(int (*fn)(void *), char *stack_area, size_t stack_size, int flags, void *arg); #define NS_CALL(fn, arg) \ @@ -173,6 +203,7 @@ struct ctx; int sock_l4_sa(const struct ctx *c, enum epoll_type type, const void *sa, socklen_t sl, const char *ifname, bool v6only, uint32_t data); +int sock_unix(char *sock_path); void sock_probe_mem(struct ctx *c); long timespec_diff_ms(const struct timespec *a, const struct timespec *b); int64_t timespec_diff_us(const struct timespec *a, const struct timespec *b); @@ -191,6 +222,8 @@ int fls(unsigned long x); int write_file(const char *path, const char *buf); int write_all_buf(int fd, const void *buf, size_t len); int write_remainder(int fd, const struct iovec *iov, size_t iovcnt, size_t skip); +int read_all_buf(int fd, void *buf, size_t len); +int read_remainder(int fd, const struct iovec *iov, size_t cnt, size_t skip); void close_open_files(int argc, char **argv); bool snprintf_check(char *str, size_t size, const char *format, ...); @@ -264,6 +297,39 @@ static inline bool mod_between(unsigned x, unsigned i, unsigned j, unsigned m) #define FPRINTF(f, ...) (void)fprintf(f, __VA_ARGS__) void raw_random(void *buf, size_t buflen); +void epoll_del(const struct ctx *c, int fd); + +/* + * Starting from glibc 2.40.9000 and commit 25a5eb4010df ("string: strerror, + * strsignal cannot use buffer after dlmopen (bug 32026)"), strerror() needs + * getrandom(2) and brk(2) as it allocates memory for the locale-translated + * error description, but our seccomp profiles forbid both. + * + * Use the strerror_() wrapper instead, calling into strerrordesc_np() to get + * a static untranslated string. It's a GNU implementation, but also defined by + * bionic. + * + * If strerrordesc_np() is not defined (e.g. musl), call strerror(). C libraries + * not defining strerrordesc_np() are expected to provide strerror() + * implementations that are simple enough for us to call. + */ +__attribute__ ((weak)) const char *strerrordesc_np(int errnum); + +/** + * strerror_() - strerror() wrapper calling strerrordesc_np() if available + * @errnum: Error code + * + * Return: error description string + */ +static inline const char *strerror_(int errnum) +{ + if (strerrordesc_np) + return strerrordesc_np(errnum); + + return strerror(errnum); +} + +#define strerror(x) @ "Don't call strerror() directly, use strerror_() instead" /* * Workarounds for https://github.com/llvm/llvm-project/issues/58992 @@ -305,4 +371,17 @@ static inline int wrap_accept4(int sockfd, struct sockaddr *addr, #define accept4(s, addr, addrlen, flags) \ wrap_accept4((s), (addr), (addrlen), (flags)) +static inline int wrap_getsockname(int sockfd, struct sockaddr *addr, +/* cppcheck-suppress constParameterPointer */ + socklen_t *addrlen) +{ + sa_init(addr, addrlen); + return getsockname(sockfd, addr, addrlen); +} +#define getsockname(s, addr, addrlen) \ + wrap_getsockname((s), (addr), (addrlen)) + +#define PASST_MAXDNAME 254 /* 253 (RFC 1035) + 1 (the terminator) */ +void encode_domain_name(char *buf, const char *domain_name); + #endif /* UTIL_H */ diff --git a/vhost_user.c b/vhost_user.c new file mode 100644 index 0000000..e8377bb --- /dev/null +++ b/vhost_user.c @@ -0,0 +1,1213 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * vhost-user API, command management and virtio interface + * + * Copyright Red Hat + * Author: Laurent Vivier <lvivier@redhat.com> + * + * Some parts from QEMU subprojects/libvhost-user/libvhost-user.c + * licensed under the following terms: + * + * Copyright IBM, Corp. 2007 + * Copyright (c) 2016 Red Hat, Inc. + * + * Authors: + * Anthony Liguori <aliguori@us.ibm.com> + * Marc-André Lureau <mlureau@redhat.com> + * Victor Kaplansky <victork@redhat.com> + * + * This work is licensed under the terms of the GNU GPL, version 2 or + * later. See the COPYING file in the top-level directory. + */ + +#include <errno.h> +#include <fcntl.h> +#include <stdlib.h> +#include <stdio.h> +#include <stdint.h> +#include <stddef.h> +#include <string.h> +#include <assert.h> +#include <stdbool.h> +#include <inttypes.h> +#include <time.h> +#include <net/ethernet.h> +#include <netinet/in.h> +#include <sys/epoll.h> +#include <sys/eventfd.h> +#include <sys/mman.h> +#include <linux/vhost_types.h> +#include <linux/virtio_net.h> + +#include "util.h" +#include "passt.h" +#include "tap.h" +#include "vhost_user.h" +#include "pcap.h" +#include "migrate.h" + +/* vhost-user version we are compatible with */ +#define VHOST_USER_VERSION 1 + +static struct vu_dev vdev_storage; + +/** + * vu_print_capabilities() - print vhost-user capabilities + * this is part of the vhost-user backend + * convention. + */ +void vu_print_capabilities(void) +{ + info("{"); + info(" \"type\": \"net\""); + info("}"); + _exit(EXIT_SUCCESS); +} + +/** + * vu_request_to_string() - convert a vhost-user request number to its name + * @req: request number + * + * Return: the name of request number + */ +static const char *vu_request_to_string(unsigned int req) +{ + if (req < VHOST_USER_MAX) { +#define REQ(req) [req] = #req + static const char * const vu_request_str[VHOST_USER_MAX] = { + REQ(VHOST_USER_NONE), + REQ(VHOST_USER_GET_FEATURES), + REQ(VHOST_USER_SET_FEATURES), + REQ(VHOST_USER_SET_OWNER), + REQ(VHOST_USER_RESET_OWNER), + REQ(VHOST_USER_SET_MEM_TABLE), + REQ(VHOST_USER_SET_LOG_BASE), + REQ(VHOST_USER_SET_LOG_FD), + REQ(VHOST_USER_SET_VRING_NUM), + REQ(VHOST_USER_SET_VRING_ADDR), + REQ(VHOST_USER_SET_VRING_BASE), + REQ(VHOST_USER_GET_VRING_BASE), + REQ(VHOST_USER_SET_VRING_KICK), + REQ(VHOST_USER_SET_VRING_CALL), + REQ(VHOST_USER_SET_VRING_ERR), + REQ(VHOST_USER_GET_PROTOCOL_FEATURES), + REQ(VHOST_USER_SET_PROTOCOL_FEATURES), + REQ(VHOST_USER_GET_QUEUE_NUM), + REQ(VHOST_USER_SET_VRING_ENABLE), + REQ(VHOST_USER_SEND_RARP), + REQ(VHOST_USER_NET_SET_MTU), + REQ(VHOST_USER_SET_BACKEND_REQ_FD), + REQ(VHOST_USER_IOTLB_MSG), + REQ(VHOST_USER_SET_VRING_ENDIAN), + REQ(VHOST_USER_GET_CONFIG), + REQ(VHOST_USER_SET_CONFIG), + REQ(VHOST_USER_POSTCOPY_ADVISE), + REQ(VHOST_USER_POSTCOPY_LISTEN), + REQ(VHOST_USER_POSTCOPY_END), + REQ(VHOST_USER_GET_INFLIGHT_FD), + REQ(VHOST_USER_SET_INFLIGHT_FD), + REQ(VHOST_USER_GPU_SET_SOCKET), + REQ(VHOST_USER_VRING_KICK), + REQ(VHOST_USER_GET_MAX_MEM_SLOTS), + REQ(VHOST_USER_ADD_MEM_REG), + REQ(VHOST_USER_REM_MEM_REG), + REQ(VHOST_USER_SET_STATUS), + REQ(VHOST_USER_GET_STATUS), + REQ(VHOST_USER_GET_SHARED_OBJECT), + REQ(VHOST_USER_SET_DEVICE_STATE_FD), + REQ(VHOST_USER_CHECK_DEVICE_STATE), + }; +#undef REQ + return vu_request_str[req]; + } + + return "unknown"; +} + +/** + * qva_to_va() - Translate front-end (QEMU) virtual address to our virtual + * address + * @dev: vhost-user device + * @qemu_addr: front-end userspace address + * + * Return: the memory address in our process virtual address space. + */ +static void *qva_to_va(struct vu_dev *dev, uint64_t qemu_addr) +{ + unsigned int i; + + /* Find matching memory region. */ + for (i = 0; i < dev->nregions; i++) { + const struct vu_dev_region *r = &dev->regions[i]; + + if ((qemu_addr >= r->qva) && (qemu_addr < (r->qva + r->size))) { + /* NOLINTNEXTLINE(performance-no-int-to-ptr) */ + return (void *)(uintptr_t)(qemu_addr - r->qva + + r->mmap_addr + + r->mmap_offset); + } + } + + return NULL; +} + +/** + * vmsg_close_fds() - Close all file descriptors of a given message + * @vmsg: vhost-user message with the list of the file descriptors + */ +static void vmsg_close_fds(const struct vhost_user_msg *vmsg) +{ + int i; + + for (i = 0; i < vmsg->fd_num; i++) + close(vmsg->fds[i]); +} + +/** + * vmsg_set_reply_u64() - Set reply payload.u64 and clear request flags + * and fd_num + * @vmsg: vhost-user message + * @val: 64-bit value to reply + */ +static void vmsg_set_reply_u64(struct vhost_user_msg *vmsg, uint64_t val) +{ + vmsg->hdr.flags = 0; /* defaults will be set by vu_send_reply() */ + vmsg->hdr.size = sizeof(vmsg->payload.u64); + vmsg->payload.u64 = val; + vmsg->fd_num = 0; +} + +/** + * vu_message_read_default() - Read incoming vhost-user message from the + * front-end + * @conn_fd: vhost-user command socket + * @vmsg: vhost-user message + * + * Return: 0 if recvmsg() has been interrupted or if there's no data to read, + * 1 if a message has been received + */ +static int vu_message_read_default(int conn_fd, struct vhost_user_msg *vmsg) +{ + char control[CMSG_SPACE(VHOST_MEMORY_BASELINE_NREGIONS * + sizeof(int))] = { 0 }; + struct iovec iov = { + .iov_base = (char *)vmsg, + .iov_len = VHOST_USER_HDR_SIZE, + }; + struct msghdr msg = { + .msg_iov = &iov, + .msg_iovlen = 1, + .msg_control = control, + .msg_controllen = sizeof(control), + }; + ssize_t ret, sz_payload; + struct cmsghdr *cmsg; + + ret = recvmsg(conn_fd, &msg, MSG_DONTWAIT); + if (ret < 0) { + if (errno == EINTR || errno == EAGAIN || errno == EWOULDBLOCK) + return 0; + die_perror("vhost-user message receive (recvmsg)"); + } + + vmsg->fd_num = 0; + for (cmsg = CMSG_FIRSTHDR(&msg); cmsg != NULL; + cmsg = CMSG_NXTHDR(&msg, cmsg)) { + if (cmsg->cmsg_level == SOL_SOCKET && + cmsg->cmsg_type == SCM_RIGHTS) { + size_t fd_size; + + ASSERT(cmsg->cmsg_len >= CMSG_LEN(0)); + fd_size = cmsg->cmsg_len - CMSG_LEN(0); + ASSERT(fd_size <= sizeof(vmsg->fds)); + vmsg->fd_num = fd_size / sizeof(int); + memcpy(vmsg->fds, CMSG_DATA(cmsg), fd_size); + break; + } + } + + sz_payload = vmsg->hdr.size; + if ((size_t)sz_payload > sizeof(vmsg->payload)) { + die("vhost-user message request too big: %d," + " size: vmsg->size: %zd, " + "while sizeof(vmsg->payload) = %zu", + vmsg->hdr.request, sz_payload, sizeof(vmsg->payload)); + } + + if (sz_payload) { + do + ret = recv(conn_fd, &vmsg->payload, sz_payload, 0); + while (ret < 0 && errno == EINTR); + + if (ret < 0) + die_perror("vhost-user message receive"); + + if (ret == 0) + die("EOF on vhost-user message receive"); + + if (ret < sz_payload) + die("Short-read on vhost-user message receive"); + } + + return 1; +} + +/** + * vu_message_write() - Send a message to the front-end + * @conn_fd: vhost-user command socket + * @vmsg: vhost-user message + * + * #syscalls:vu sendmsg + */ +static void vu_message_write(int conn_fd, struct vhost_user_msg *vmsg) +{ + char control[CMSG_SPACE(VHOST_MEMORY_BASELINE_NREGIONS * sizeof(int))] = { 0 }; + struct iovec iov = { + .iov_base = (char *)vmsg, + .iov_len = VHOST_USER_HDR_SIZE + vmsg->hdr.size, + }; + struct msghdr msg = { + .msg_iov = &iov, + .msg_iovlen = 1, + .msg_control = control, + }; + int rc; + + ASSERT(vmsg->fd_num <= VHOST_MEMORY_BASELINE_NREGIONS); + if (vmsg->fd_num > 0) { + size_t fdsize = vmsg->fd_num * sizeof(int); + struct cmsghdr *cmsg; + + msg.msg_controllen = CMSG_SPACE(fdsize); + cmsg = CMSG_FIRSTHDR(&msg); + cmsg->cmsg_len = CMSG_LEN(fdsize); + cmsg->cmsg_level = SOL_SOCKET; + cmsg->cmsg_type = SCM_RIGHTS; + memcpy(CMSG_DATA(cmsg), vmsg->fds, fdsize); + } + + do + rc = sendmsg(conn_fd, &msg, 0); + while (rc < 0 && errno == EINTR); + + if (rc < 0) + die_perror("vhost-user message send"); + + if ((uint32_t)rc < VHOST_USER_HDR_SIZE + vmsg->hdr.size) + die("EOF on vhost-user message send"); +} + +/** + * vu_send_reply() - Update message flags and send it to front-end + * @conn_fd: vhost-user command socket + * @vmsg: vhost-user message + */ +static void vu_send_reply(int conn_fd, struct vhost_user_msg *vmsg) +{ + vmsg->hdr.flags &= ~VHOST_USER_VERSION_MASK; + vmsg->hdr.flags |= VHOST_USER_VERSION; + vmsg->hdr.flags |= VHOST_USER_REPLY_MASK; + + vu_message_write(conn_fd, vmsg); +} + +/** + * vu_get_features_exec() - Provide back-end features bitmask to front-end + * @vdev: vhost-user device + * @vmsg: vhost-user message + * + * Return: True as a reply is requested + */ +static bool vu_get_features_exec(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) +{ + uint64_t features = + 1ULL << VIRTIO_F_VERSION_1 | + 1ULL << VIRTIO_NET_F_MRG_RXBUF | + 1ULL << VHOST_F_LOG_ALL | + 1ULL << VHOST_USER_F_PROTOCOL_FEATURES; + + (void)vdev; + + vmsg_set_reply_u64(vmsg, features); + + debug("Sending back to guest u64: 0x%016"PRIx64, vmsg->payload.u64); + + return true; +} + +/** + * vu_set_enable_all_rings() - Enable/disable all the virtqueues + * @vdev: vhost-user device + * @enable: New virtqueues state + */ +static void vu_set_enable_all_rings(struct vu_dev *vdev, bool enable) +{ + uint16_t i; + + for (i = 0; i < VHOST_USER_MAX_QUEUES; i++) + vdev->vq[i].enable = enable; +} + +/** + * vu_set_features_exec() - Enable features of the back-end + * @vdev: vhost-user device + * @vmsg: vhost-user message + * + * Return: False as no reply is requested + */ +static bool vu_set_features_exec(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) +{ + debug("u64: 0x%016"PRIx64, vmsg->payload.u64); + + vdev->features = vmsg->payload.u64; + /* We only support devices conforming to VIRTIO 1.0 or + * later + */ + if (!vu_has_feature(vdev, VIRTIO_F_VERSION_1)) + die("virtio legacy devices aren't supported by passt"); + + if (!vu_has_feature(vdev, VHOST_USER_F_PROTOCOL_FEATURES)) + vu_set_enable_all_rings(vdev, true); + + return false; +} + +/** + * vu_set_owner_exec() - Session start flag, do nothing in our case + * @vdev: vhost-user device + * @vmsg: vhost-user message + * + * Return: False as no reply is requested + */ +static bool vu_set_owner_exec(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) +{ + (void)vdev; + (void)vmsg; + + return false; +} + +/** + * map_ring() - Convert ring front-end (QEMU) addresses to our process + * virtual address space. + * @vdev: vhost-user device + * @vq: Virtqueue + * + * Return: True if ring cannot be mapped to our address space + */ +static bool map_ring(struct vu_dev *vdev, struct vu_virtq *vq) +{ + vq->vring.desc = qva_to_va(vdev, vq->vra.desc_user_addr); + vq->vring.used = qva_to_va(vdev, vq->vra.used_user_addr); + vq->vring.avail = qva_to_va(vdev, vq->vra.avail_user_addr); + + debug("Setting virtq addresses:"); + debug(" vring_desc at %p", (void *)vq->vring.desc); + debug(" vring_used at %p", (void *)vq->vring.used); + debug(" vring_avail at %p", (void *)vq->vring.avail); + + return !(vq->vring.desc && vq->vring.used && vq->vring.avail); +} + +/** + * vu_set_mem_table_exec() - Sets the memory map regions to be able to + * translate the vring addresses. + * @vdev: vhost-user device + * @vmsg: vhost-user message + * + * Return: False as no reply is requested + * + * #syscalls:vu mmap|mmap2 munmap + */ +static bool vu_set_mem_table_exec(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) +{ + struct vhost_user_memory m = vmsg->payload.memory, *memory = &m; + unsigned int i; + + for (i = 0; i < vdev->nregions; i++) { + const struct vu_dev_region *r = &vdev->regions[i]; + + if (r->mmap_addr) { + /* NOLINTNEXTLINE(performance-no-int-to-ptr) */ + munmap((void *)(uintptr_t)r->mmap_addr, + r->size + r->mmap_offset); + } + } + vdev->nregions = memory->nregions; + + debug("vhost-user nregions: %u", memory->nregions); + for (i = 0; i < vdev->nregions; i++) { + struct vhost_user_memory_region *msg_region = &memory->regions[i]; + struct vu_dev_region *dev_region = &vdev->regions[i]; + void *mmap_addr; + + debug("vhost-user region %d", i); + debug(" guest_phys_addr: 0x%016"PRIx64, + msg_region->guest_phys_addr); + debug(" memory_size: 0x%016"PRIx64, + msg_region->memory_size); + debug(" userspace_addr 0x%016"PRIx64, + msg_region->userspace_addr); + debug(" mmap_offset 0x%016"PRIx64, + msg_region->mmap_offset); + + dev_region->gpa = msg_region->guest_phys_addr; + dev_region->size = msg_region->memory_size; + dev_region->qva = msg_region->userspace_addr; + dev_region->mmap_offset = msg_region->mmap_offset; + + /* We don't use offset argument of mmap() since the + * mapped address has to be page aligned. + */ + mmap_addr = mmap(0, dev_region->size + dev_region->mmap_offset, + PROT_READ | PROT_WRITE, MAP_SHARED | + MAP_NORESERVE, vmsg->fds[i], 0); + + if (mmap_addr == MAP_FAILED) + die_perror("vhost-user region mmap error"); + + dev_region->mmap_addr = (uint64_t)(uintptr_t)mmap_addr; + debug(" mmap_addr: 0x%016"PRIx64, + dev_region->mmap_addr); + + close(vmsg->fds[i]); + } + + for (i = 0; i < VHOST_USER_MAX_QUEUES; i++) { + if (vdev->vq[i].vring.desc) { + if (map_ring(vdev, &vdev->vq[i])) + die("remapping queue %d during setmemtable", i); + } + } + + /* As vu_packet_check_range() has no access to the number of + * memory regions, mark the end of the array with mmap_addr = 0 + */ + ASSERT(vdev->nregions < VHOST_USER_MAX_RAM_SLOTS - 1); + vdev->regions[vdev->nregions].mmap_addr = 0; + + tap_sock_update_pool(vdev->regions, 0); + + return false; +} + +/** + * vu_close_log() - Close the logging file descriptor + * @vdev: vhost-user device + */ +static void vu_close_log(struct vu_dev *vdev) +{ + if (vdev->log_table) { + if (munmap(vdev->log_table, vdev->log_size) != 0) + die_perror("close log munmap() error"); + vdev->log_table = NULL; + } + + if (vdev->log_call_fd != -1) { + close(vdev->log_call_fd); + vdev->log_call_fd = -1; + } +} + +/** + * vu_log_kick() - Inform the front-end that the log has been modified + * @vdev: vhost-user device + */ +static void vu_log_kick(const struct vu_dev *vdev) +{ + if (vdev->log_call_fd != -1) { + int rc; + + rc = eventfd_write(vdev->log_call_fd, 1); + if (rc == -1) + die_perror("vhost-user kick eventfd_write()"); + } +} + +/** + * vu_log_page() - Update logging table + * @log_table: Base address of the logging table + * @page: Page number that has been updated + */ +/* NOLINTNEXTLINE(readability-non-const-parameter) */ +static void vu_log_page(uint8_t *log_table, uint64_t page) +{ + qatomic_or(&log_table[page / 8], 1 << (page % 8)); +} + +/** + * vu_log_write() - Log memory write + * @vdev: vhost-user device + * @address: Memory address + * @length: Memory size + */ +void vu_log_write(const struct vu_dev *vdev, uint64_t address, uint64_t length) +{ + uint64_t page; + + if (!vdev->log_table || !length || + !vu_has_feature(vdev, VHOST_F_LOG_ALL)) + return; + + page = address / VHOST_LOG_PAGE; + while (page * VHOST_LOG_PAGE < address + length) { + vu_log_page(vdev->log_table, page); + page++; + } + vu_log_kick(vdev); +} + +/** + * vu_set_log_base_exec() - Set the memory log base + * @vdev: vhost-user device + * @vmsg: vhost-user message + * + * Return: True as a reply is requested + * + * #syscalls:vu mmap|mmap2 munmap + */ +static bool vu_set_log_base_exec(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) +{ + uint64_t log_mmap_size, log_mmap_offset; + void *base; + int fd; + + if (vmsg->fd_num != 1 || vmsg->hdr.size != sizeof(vmsg->payload.log)) + die("vhost-user: Invalid log_base message"); + + fd = vmsg->fds[0]; + log_mmap_offset = vmsg->payload.log.mmap_offset; + log_mmap_size = vmsg->payload.log.mmap_size; + + debug("vhost-user log mmap_offset: %"PRId64, log_mmap_offset); + debug("vhost-user log mmap_size: %"PRId64, log_mmap_size); + + base = mmap(0, log_mmap_size, PROT_READ | PROT_WRITE, MAP_SHARED, fd, + log_mmap_offset); + close(fd); + if (base == MAP_FAILED) + die("vhost-user log mmap error"); + + if (vdev->log_table) + munmap(vdev->log_table, vdev->log_size); + + vdev->log_table = base; + vdev->log_size = log_mmap_size; + + vmsg->hdr.size = sizeof(vmsg->payload.u64); + vmsg->fd_num = 0; + + return true; +} + +/** + * vu_set_log_fd_exec() - Set the eventfd used to report logging update + * @vdev: vhost-user device + * @vmsg: vhost-user message + * + * Return: False as no reply is requested + */ +static bool vu_set_log_fd_exec(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) +{ + if (vmsg->fd_num != 1) + die("Invalid log_fd message"); + + if (vdev->log_call_fd != -1) + close(vdev->log_call_fd); + + vdev->log_call_fd = vmsg->fds[0]; + + debug("Got log_call_fd: %d", vdev->log_call_fd); + + return false; +} + +/** + * vu_set_vring_num_exec() - Set the size of the queue (vring size) + * @vdev: vhost-user device + * @vmsg: vhost-user message + * + * Return: False as no reply is requested + */ +static bool vu_set_vring_num_exec(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) +{ + unsigned int idx = vmsg->payload.state.index; + unsigned int num = vmsg->payload.state.num; + + trace("State.index: %u", idx); + trace("State.num: %u", num); + vdev->vq[idx].vring.num = num; + + return false; +} + +/** + * vu_set_vring_addr_exec() - Set the addresses of the vring + * @vdev: vhost-user device + * @vmsg: vhost-user message + * + * Return: False as no reply is requested + */ +static bool vu_set_vring_addr_exec(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) +{ + /* We need to copy the payload to vhost_vring_addr structure + * to access index because address of vmsg->payload.addr + * can be unaligned as it is packed. + */ + struct vhost_vring_addr addr = vmsg->payload.addr; + struct vu_virtq *vq = &vdev->vq[addr.index]; + + debug("vhost_vring_addr:"); + debug(" index: %d", addr.index); + debug(" flags: %d", addr.flags); + debug(" desc_user_addr: 0x%016" PRIx64, + (uint64_t)addr.desc_user_addr); + debug(" used_user_addr: 0x%016" PRIx64, + (uint64_t)addr.used_user_addr); + debug(" avail_user_addr: 0x%016" PRIx64, + (uint64_t)addr.avail_user_addr); + debug(" log_guest_addr: 0x%016" PRIx64, + (uint64_t)addr.log_guest_addr); + + vq->vra = vmsg->payload.addr; + vq->vring.flags = addr.flags; + vq->vring.log_guest_addr = addr.log_guest_addr; + + if (map_ring(vdev, vq)) + die("Invalid vring_addr message"); + + vq->used_idx = le16toh(vq->vring.used->idx); + + if (vq->last_avail_idx != vq->used_idx) { + debug("Last avail index != used index: %u != %u", + vq->last_avail_idx, vq->used_idx); + } + + return false; +} +/** + * vu_set_vring_base_exec() - Sets the next index to use for descriptors + * in this vring + * @vdev: vhost-user device + * @vmsg: vhost-user message + * + * Return: False as no reply is requested + */ +static bool vu_set_vring_base_exec(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) +{ + unsigned int idx = vmsg->payload.state.index; + unsigned int num = vmsg->payload.state.num; + + debug("State.index: %u", idx); + debug("State.num: %u", num); + vdev->vq[idx].shadow_avail_idx = vdev->vq[idx].last_avail_idx = num; + + return false; +} + +/** + * vu_get_vring_base_exec() - Stops the vring and returns the current + * descriptor index or indices + * @vdev: vhost-user device + * @vmsg: vhost-user message + * + * Return: True as a reply is requested + */ +static bool vu_get_vring_base_exec(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) +{ + unsigned int idx = vmsg->payload.state.index; + + debug("State.index: %u", idx); + vmsg->payload.state.num = vdev->vq[idx].last_avail_idx; + vmsg->hdr.size = sizeof(vmsg->payload.state); + + vdev->vq[idx].started = false; + vdev->vq[idx].vring.avail = 0; + + if (vdev->vq[idx].call_fd != -1) { + close(vdev->vq[idx].call_fd); + vdev->vq[idx].call_fd = -1; + } + if (vdev->vq[idx].kick_fd != -1) { + epoll_del(vdev->context, vdev->vq[idx].kick_fd); + close(vdev->vq[idx].kick_fd); + vdev->vq[idx].kick_fd = -1; + } + + return true; +} + +/** + * vu_set_watch() - Add a file descriptor to the passt epoll file descriptor + * @vdev: vhost-user device + * @idx: queue index of the file descriptor to add + */ +static void vu_set_watch(const struct vu_dev *vdev, int idx) +{ + union epoll_ref ref = { + .type = EPOLL_TYPE_VHOST_KICK, + .fd = vdev->vq[idx].kick_fd, + .queue = idx + }; + struct epoll_event ev = { 0 }; + + ev.data.u64 = ref.u64; + ev.events = EPOLLIN; + epoll_ctl(vdev->context->epollfd, EPOLL_CTL_ADD, ref.fd, &ev); +} + +/** + * vu_check_queue_msg_file() - Check if a message is valid, + * close fds if NOFD bit is set + * @vmsg: vhost-user message + */ +static void vu_check_queue_msg_file(struct vhost_user_msg *vmsg) +{ + bool nofd = vmsg->payload.u64 & VHOST_USER_VRING_NOFD_MASK; + int idx = vmsg->payload.u64 & VHOST_USER_VRING_IDX_MASK; + + if (idx >= VHOST_USER_MAX_QUEUES) + die("Invalid vhost-user queue index: %u", idx); + + if (nofd) { + vmsg_close_fds(vmsg); + return; + } + + if (vmsg->fd_num != 1) + die("Invalid fds in vhost-user request: %d", vmsg->hdr.request); +} + +/** + * vu_set_vring_kick_exec() - Set the event file descriptor for adding buffers + * to the vring + * @vdev: vhost-user device + * @vmsg: vhost-user message + * + * Return: False as no reply is requested + */ +static bool vu_set_vring_kick_exec(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) +{ + bool nofd = vmsg->payload.u64 & VHOST_USER_VRING_NOFD_MASK; + int idx = vmsg->payload.u64 & VHOST_USER_VRING_IDX_MASK; + + debug("u64: 0x%016"PRIx64, vmsg->payload.u64); + + vu_check_queue_msg_file(vmsg); + + if (vdev->vq[idx].kick_fd != -1) { + epoll_del(vdev->context, vdev->vq[idx].kick_fd); + close(vdev->vq[idx].kick_fd); + vdev->vq[idx].kick_fd = -1; + } + + if (!nofd) + vdev->vq[idx].kick_fd = vmsg->fds[0]; + + debug("Got kick_fd: %d for vq: %d", vdev->vq[idx].kick_fd, idx); + + vdev->vq[idx].started = true; + + if (vdev->vq[idx].kick_fd != -1 && VHOST_USER_IS_QUEUE_TX(idx)) { + vu_set_watch(vdev, idx); + debug("Waiting for kicks on fd: %d for vq: %d", + vdev->vq[idx].kick_fd, idx); + } + + return false; +} + +/** + * vu_set_vring_call_exec() - Set the event file descriptor to signal when + * buffers are used + * @vdev: vhost-user device + * @vmsg: vhost-user message + * + * Return: False as no reply is requested + */ +static bool vu_set_vring_call_exec(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) +{ + bool nofd = vmsg->payload.u64 & VHOST_USER_VRING_NOFD_MASK; + int idx = vmsg->payload.u64 & VHOST_USER_VRING_IDX_MASK; + + debug("u64: 0x%016"PRIx64, vmsg->payload.u64); + + vu_check_queue_msg_file(vmsg); + + if (vdev->vq[idx].call_fd != -1) { + close(vdev->vq[idx].call_fd); + vdev->vq[idx].call_fd = -1; + } + + if (!nofd) + vdev->vq[idx].call_fd = vmsg->fds[0]; + + /* in case of I/O hang after reconnecting */ + if (vdev->vq[idx].call_fd != -1) + eventfd_write(vmsg->fds[0], 1); + + debug("Got call_fd: %d for vq: %d", vdev->vq[idx].call_fd, idx); + + return false; +} + +/** + * vu_set_vring_err_exec() - Set the event file descriptor to signal when + * error occurs + * @vdev: vhost-user device + * @vmsg: vhost-user message + * + * Return: False as no reply is requested + */ +static bool vu_set_vring_err_exec(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) +{ + bool nofd = vmsg->payload.u64 & VHOST_USER_VRING_NOFD_MASK; + int idx = vmsg->payload.u64 & VHOST_USER_VRING_IDX_MASK; + + debug("u64: 0x%016"PRIx64, vmsg->payload.u64); + + vu_check_queue_msg_file(vmsg); + + if (vdev->vq[idx].err_fd != -1) { + close(vdev->vq[idx].err_fd); + vdev->vq[idx].err_fd = -1; + } + + if (!nofd) + vdev->vq[idx].err_fd = vmsg->fds[0]; + + return false; +} + +/** + * vu_get_protocol_features_exec() - Provide the protocol (vhost-user) features + * to the front-end + * @vdev: vhost-user device + * @vmsg: vhost-user message + * + * Return: True as a reply is requested + */ +static bool vu_get_protocol_features_exec(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) +{ + uint64_t features = 1ULL << VHOST_USER_PROTOCOL_F_REPLY_ACK | + 1ULL << VHOST_USER_PROTOCOL_F_LOG_SHMFD | + 1ULL << VHOST_USER_PROTOCOL_F_DEVICE_STATE | + 1ULL << VHOST_USER_PROTOCOL_F_RARP; + + (void)vdev; + vmsg_set_reply_u64(vmsg, features); + + return true; +} + +/** + * vu_set_protocol_features_exec() - Enable protocol (vhost-user) features + * @vdev: vhost-user device + * @vmsg: vhost-user message + * + * Return: False as no reply is requested + */ +static bool vu_set_protocol_features_exec(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) +{ + uint64_t features = vmsg->payload.u64; + + debug("u64: 0x%016"PRIx64, features); + + vdev->protocol_features = vmsg->payload.u64; + + return false; +} + +/** + * vu_get_queue_num_exec() - Tell how many queues we support + * @vdev: vhost-user device + * @vmsg: vhost-user message + * + * Return: True as a reply is requested + */ +static bool vu_get_queue_num_exec(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) +{ + (void)vdev; + + vmsg_set_reply_u64(vmsg, VHOST_USER_MAX_QUEUES); + + return true; +} + +/** + * vu_set_vring_enable_exec() - Enable or disable corresponding vring + * @vdev: vhost-user device + * @vmsg: vhost-user message + * + * Return: False as no reply is requested + */ +static bool vu_set_vring_enable_exec(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) +{ + unsigned int enable = vmsg->payload.state.num; + unsigned int idx = vmsg->payload.state.index; + + debug("State.index: %u", idx); + debug("State.enable: %u", enable); + + if (idx >= VHOST_USER_MAX_QUEUES) + die("Invalid vring_enable index: %u", idx); + + vdev->vq[idx].enable = enable; + return false; +} + +/** + * vu_send_rarp_exec() - vhost-user specification says: "Broadcast a fake + * RARP to notify the migration is terminated", + * but passt doesn't need to update any ARP table, + * so do nothing to silence QEMU bogus error message + * @vdev: vhost-user device + * @vmsg: vhost-user message + * + * Return: False as no reply is requested + */ +static bool vu_send_rarp_exec(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) +{ + char macstr[ETH_ADDRSTRLEN]; + + (void)vdev; + + /* ignore the command */ + + debug("Ignore command VHOST_USER_SEND_RARP for %s", + eth_ntop((unsigned char *)&vmsg->payload.u64, macstr, + sizeof(macstr))); + + return false; +} + +/** + * vu_set_device_state_fd_exec() - Set the device state migration channel + * @vdev: vhost-user device + * @vmsg: vhost-user message + * + * Return: True as the reply contains 0 to indicate success + * and set bit 8 as we don't provide our own fd. + */ +static bool vu_set_device_state_fd_exec(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) +{ + unsigned int direction = vmsg->payload.transfer_state.direction; + unsigned int phase = vmsg->payload.transfer_state.phase; + + if (vmsg->fd_num != 1) + die("Invalid device_state_fd message"); + + if (phase != VHOST_USER_TRANSFER_STATE_PHASE_STOPPED) + die("Invalid device_state_fd phase: %d", phase); + + if (direction != VHOST_USER_TRANSFER_STATE_DIRECTION_SAVE && + direction != VHOST_USER_TRANSFER_STATE_DIRECTION_LOAD) + die("Invalid device_state_fd direction: %d", direction); + + migrate_request(vdev->context, vmsg->fds[0], + direction == VHOST_USER_TRANSFER_STATE_DIRECTION_LOAD); + + /* We don't provide a new fd for the data transfer */ + vmsg_set_reply_u64(vmsg, VHOST_USER_VRING_NOFD_MASK); + + return true; +} + +/** + * vu_check_device_state_exec() - Return device state migration result + * @vdev: vhost-user device + * @vmsg: vhost-user message + * + * Return: True as the reply contains the migration result + */ +/* cppcheck-suppress constParameterCallback */ +static bool vu_check_device_state_exec(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) +{ + vmsg_set_reply_u64(vmsg, vdev->context->device_state_result); + + return true; +} + +/** + * vu_init() - Initialize vhost-user device structure + * @c: execution context + */ +void vu_init(struct ctx *c) +{ + int i; + + c->vdev = &vdev_storage; + c->vdev->context = c; + for (i = 0; i < VHOST_USER_MAX_QUEUES; i++) { + c->vdev->vq[i] = (struct vu_virtq){ + .call_fd = -1, + .kick_fd = -1, + .err_fd = -1, + .notification = true, + }; + } + c->vdev->log_table = NULL; + c->vdev->log_call_fd = -1; + + migrate_init(c); +} + + +/** + * vu_cleanup() - Reset vhost-user device + * @vdev: vhost-user device + */ +void vu_cleanup(struct vu_dev *vdev) +{ + unsigned int i; + + for (i = 0; i < VHOST_USER_MAX_QUEUES; i++) { + struct vu_virtq *vq = &vdev->vq[i]; + + vq->started = false; + vq->notification = true; + + if (vq->call_fd != -1) { + close(vq->call_fd); + vq->call_fd = -1; + } + if (vq->err_fd != -1) { + close(vq->err_fd); + vq->err_fd = -1; + } + if (vq->kick_fd != -1) { + epoll_del(vdev->context, vq->kick_fd); + close(vq->kick_fd); + vq->kick_fd = -1; + } + + vq->vring.desc = 0; + vq->vring.used = 0; + vq->vring.avail = 0; + } + + for (i = 0; i < vdev->nregions; i++) { + const struct vu_dev_region *r = &vdev->regions[i]; + + if (r->mmap_addr) { + /* NOLINTNEXTLINE(performance-no-int-to-ptr) */ + munmap((void *)(uintptr_t)r->mmap_addr, + r->size + r->mmap_offset); + } + } + vdev->nregions = 0; + + vu_close_log(vdev); + + /* If we lose the VU dev, we also lose our migration channel */ + migrate_close(vdev->context); +} + +/** + * vu_sock_reset() - Reset connection socket + * @vdev: vhost-user device + */ +static void vu_sock_reset(struct vu_dev *vdev) +{ + tap_sock_reset(vdev->context); +} + +static bool (*vu_handle[VHOST_USER_MAX])(struct vu_dev *vdev, + struct vhost_user_msg *vmsg) = { + [VHOST_USER_GET_FEATURES] = vu_get_features_exec, + [VHOST_USER_SET_FEATURES] = vu_set_features_exec, + [VHOST_USER_GET_PROTOCOL_FEATURES] = vu_get_protocol_features_exec, + [VHOST_USER_SET_PROTOCOL_FEATURES] = vu_set_protocol_features_exec, + [VHOST_USER_GET_QUEUE_NUM] = vu_get_queue_num_exec, + [VHOST_USER_SET_OWNER] = vu_set_owner_exec, + [VHOST_USER_SET_MEM_TABLE] = vu_set_mem_table_exec, + [VHOST_USER_SET_LOG_BASE] = vu_set_log_base_exec, + [VHOST_USER_SET_LOG_FD] = vu_set_log_fd_exec, + [VHOST_USER_SET_VRING_NUM] = vu_set_vring_num_exec, + [VHOST_USER_SET_VRING_ADDR] = vu_set_vring_addr_exec, + [VHOST_USER_SET_VRING_BASE] = vu_set_vring_base_exec, + [VHOST_USER_GET_VRING_BASE] = vu_get_vring_base_exec, + [VHOST_USER_SET_VRING_KICK] = vu_set_vring_kick_exec, + [VHOST_USER_SET_VRING_CALL] = vu_set_vring_call_exec, + [VHOST_USER_SET_VRING_ERR] = vu_set_vring_err_exec, + [VHOST_USER_SET_VRING_ENABLE] = vu_set_vring_enable_exec, + [VHOST_USER_SEND_RARP] = vu_send_rarp_exec, + [VHOST_USER_SET_DEVICE_STATE_FD] = vu_set_device_state_fd_exec, + [VHOST_USER_CHECK_DEVICE_STATE] = vu_check_device_state_exec, +}; + +/** + * vu_control_handler() - Handle control commands for vhost-user + * @vdev: vhost-user device + * @fd: vhost-user message socket + * @events: epoll events + */ +void vu_control_handler(struct vu_dev *vdev, int fd, uint32_t events) +{ + struct vhost_user_msg vmsg = { 0 }; + bool need_reply, reply_requested; + int ret; + + if (events & (EPOLLRDHUP | EPOLLHUP | EPOLLERR)) { + vu_sock_reset(vdev); + return; + } + + ret = vu_message_read_default(fd, &vmsg); + if (ret == 0) { + vu_sock_reset(vdev); + return; + } + debug("================ Vhost user message ================"); + debug("Request: %s (%d)", vu_request_to_string(vmsg.hdr.request), + vmsg.hdr.request); + debug("Flags: 0x%x", vmsg.hdr.flags); + debug("Size: %u", vmsg.hdr.size); + + need_reply = vmsg.hdr.flags & VHOST_USER_NEED_REPLY_MASK; + + if (vmsg.hdr.request >= 0 && vmsg.hdr.request < VHOST_USER_MAX && + vu_handle[vmsg.hdr.request]) + reply_requested = vu_handle[vmsg.hdr.request](vdev, &vmsg); + else + die("Unhandled request: %d", vmsg.hdr.request); + + /* cppcheck-suppress legacyUninitvar */ + if (!reply_requested && need_reply) { + vmsg.payload.u64 = 0; + vmsg.hdr.flags = 0; + vmsg.hdr.size = sizeof(vmsg.payload.u64); + vmsg.fd_num = 0; + reply_requested = true; + } + + if (reply_requested) + vu_send_reply(fd, &vmsg); + + if (vmsg.hdr.request == VHOST_USER_CHECK_DEVICE_STATE && + vdev->context->device_state_result == 0 && + !vdev->context->migrate_target) { + info("Migration complete, exiting"); + _exit(EXIT_SUCCESS); + } +} diff --git a/vhost_user.h b/vhost_user.h new file mode 100644 index 0000000..f2ae2da --- /dev/null +++ b/vhost_user.h @@ -0,0 +1,247 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * vhost-user API, command management and virtio interface + * + * Copyright Red Hat + * Author: Laurent Vivier <lvivier@redhat.com> + */ + +/* some parts from subprojects/libvhost-user/libvhost-user.h */ + +#ifndef VHOST_USER_H +#define VHOST_USER_H + +#include "virtio.h" +#include "iov.h" + +#define VHOST_USER_F_PROTOCOL_FEATURES 30 +#define VHOST_LOG_PAGE 4096 + +#define VHOST_MEMORY_BASELINE_NREGIONS 8 + +/** + * enum vhost_user_protocol_feature - List of available vhost-user features + */ +enum vhost_user_protocol_feature { + VHOST_USER_PROTOCOL_F_MQ = 0, + VHOST_USER_PROTOCOL_F_LOG_SHMFD = 1, + VHOST_USER_PROTOCOL_F_RARP = 2, + VHOST_USER_PROTOCOL_F_REPLY_ACK = 3, + VHOST_USER_PROTOCOL_F_NET_MTU = 4, + VHOST_USER_PROTOCOL_F_BACKEND_REQ = 5, + VHOST_USER_PROTOCOL_F_CROSS_ENDIAN = 6, + VHOST_USER_PROTOCOL_F_CRYPTO_SESSION = 7, + VHOST_USER_PROTOCOL_F_PAGEFAULT = 8, + VHOST_USER_PROTOCOL_F_CONFIG = 9, + VHOST_USER_PROTOCOL_F_SLAVE_SEND_FD = 10, + VHOST_USER_PROTOCOL_F_HOST_NOTIFIER = 11, + VHOST_USER_PROTOCOL_F_INFLIGHT_SHMFD = 12, + VHOST_USER_PROTOCOL_F_INBAND_NOTIFICATIONS = 14, + VHOST_USER_PROTOCOL_F_CONFIGURE_MEM_SLOTS = 15, + VHOST_USER_PROTOCOL_F_STATUS = 16, + /* Feature 17 reserved for VHOST_USER_PROTOCOL_F_XEN_MMAP. */ + VHOST_USER_PROTOCOL_F_SHARED_OBJECT = 18, + VHOST_USER_PROTOCOL_F_DEVICE_STATE = 19, + + VHOST_USER_PROTOCOL_F_MAX +}; + +/** + * enum vhost_user_request - List of available vhost-user requests + */ +enum vhost_user_request { + VHOST_USER_NONE = 0, + VHOST_USER_GET_FEATURES = 1, + VHOST_USER_SET_FEATURES = 2, + VHOST_USER_SET_OWNER = 3, + VHOST_USER_RESET_OWNER = 4, + VHOST_USER_SET_MEM_TABLE = 5, + VHOST_USER_SET_LOG_BASE = 6, + VHOST_USER_SET_LOG_FD = 7, + VHOST_USER_SET_VRING_NUM = 8, + VHOST_USER_SET_VRING_ADDR = 9, + VHOST_USER_SET_VRING_BASE = 10, + VHOST_USER_GET_VRING_BASE = 11, + VHOST_USER_SET_VRING_KICK = 12, + VHOST_USER_SET_VRING_CALL = 13, + VHOST_USER_SET_VRING_ERR = 14, + VHOST_USER_GET_PROTOCOL_FEATURES = 15, + VHOST_USER_SET_PROTOCOL_FEATURES = 16, + VHOST_USER_GET_QUEUE_NUM = 17, + VHOST_USER_SET_VRING_ENABLE = 18, + VHOST_USER_SEND_RARP = 19, + VHOST_USER_NET_SET_MTU = 20, + VHOST_USER_SET_BACKEND_REQ_FD = 21, + VHOST_USER_IOTLB_MSG = 22, + VHOST_USER_SET_VRING_ENDIAN = 23, + VHOST_USER_GET_CONFIG = 24, + VHOST_USER_SET_CONFIG = 25, + VHOST_USER_CREATE_CRYPTO_SESSION = 26, + VHOST_USER_CLOSE_CRYPTO_SESSION = 27, + VHOST_USER_POSTCOPY_ADVISE = 28, + VHOST_USER_POSTCOPY_LISTEN = 29, + VHOST_USER_POSTCOPY_END = 30, + VHOST_USER_GET_INFLIGHT_FD = 31, + VHOST_USER_SET_INFLIGHT_FD = 32, + VHOST_USER_GPU_SET_SOCKET = 33, + VHOST_USER_VRING_KICK = 35, + VHOST_USER_GET_MAX_MEM_SLOTS = 36, + VHOST_USER_ADD_MEM_REG = 37, + VHOST_USER_REM_MEM_REG = 38, + VHOST_USER_SET_STATUS = 39, + VHOST_USER_GET_STATUS = 40, + VHOST_USER_GET_SHARED_OBJECT = 41, + VHOST_USER_SET_DEVICE_STATE_FD = 42, + VHOST_USER_CHECK_DEVICE_STATE = 43, + VHOST_USER_MAX +}; + +/** + * struct vhost_user_header - vhost-user message header + * @request: Request type of the message + * @flags: Request flags + * @size: The following payload size + */ +struct vhost_user_header { + enum vhost_user_request request; + +#define VHOST_USER_VERSION_MASK 0x3 +#define VHOST_USER_REPLY_MASK (0x1 << 2) +#define VHOST_USER_NEED_REPLY_MASK (0x1 << 3) + uint32_t flags; + uint32_t size; +} __attribute__ ((__packed__)); + +/** + * struct vhost_user_memory_region - Front-end shared memory region information + * @guest_phys_addr: Guest physical address of the region + * @memory_size: Memory size + * @userspace_addr: front-end (QEMU) userspace address + * @mmap_offset: region offset in the shared memory area + */ +struct vhost_user_memory_region { + uint64_t guest_phys_addr; + uint64_t memory_size; + uint64_t userspace_addr; + uint64_t mmap_offset; +}; + +/** + * struct vhost_user_memory - List of all the shared memory regions + * @nregions: Number of memory regions + * @padding: Padding + * @regions: Memory regions list + */ +struct vhost_user_memory { + uint32_t nregions; + uint32_t padding; + struct vhost_user_memory_region regions[VHOST_MEMORY_BASELINE_NREGIONS]; +}; + +/** + * struct vhost_user_log - Address and size of the shared memory region used + * to log page update + * @mmap_size: Size of the shared memory region + * @mmap_offset: Offset of the shared memory region + */ +struct vhost_user_log { + uint64_t mmap_size; + uint64_t mmap_offset; +}; + +/** + * struct vhost_user_transfer_device_state - Set the direction and phase + * of the backend device state fd + * @direction: Device state transfer direction (save or load) + * @phase: Migration phase (only stopped is supported) + */ +struct vhost_user_transfer_device_state { + uint32_t direction; +#define VHOST_USER_TRANSFER_STATE_DIRECTION_SAVE 0 +#define VHOST_USER_TRANSFER_STATE_DIRECTION_LOAD 1 + uint32_t phase; +#define VHOST_USER_TRANSFER_STATE_PHASE_STOPPED 0 +}; + +/** + * union vhost_user_payload - vhost-user message payload + * @u64: 64-bit payload + * @state: vring state payload + * @addr: vring addresses payload + * @memory: Memory regions information payload + * @log: Memory logging payload + * @transfer_state: Device state payload + */ +union vhost_user_payload { +#define VHOST_USER_VRING_IDX_MASK 0xff +#define VHOST_USER_VRING_NOFD_MASK (0x1 << 8) + uint64_t u64; + struct vhost_vring_state state; + struct vhost_vring_addr addr; + struct vhost_user_memory memory; + struct vhost_user_log log; + struct vhost_user_transfer_device_state transfer_state; +}; + +/** + * struct vhost_user_msg - vhost-user message + * @hdr: Message header + * @payload: Message payload + * @fds: File descriptors associated with the message + * in the ancillary data. + * (shared memory or event file descriptors) + * @fd_num: Number of file descriptors + */ +struct vhost_user_msg { + struct vhost_user_header hdr; + union vhost_user_payload payload; + + int fds[VHOST_MEMORY_BASELINE_NREGIONS]; + int fd_num; +} __attribute__ ((__packed__)); +#define VHOST_USER_HDR_SIZE sizeof(struct vhost_user_header) + +/* index of the RX virtqueue */ +#define VHOST_USER_RX_QUEUE 0 +/* index of the TX virtqueue */ +#define VHOST_USER_TX_QUEUE 1 + +/* in case of multiqueue, the RX and TX queues are interleaved */ +#define VHOST_USER_IS_QUEUE_TX(n) (n % 2) +#define VHOST_USER_IS_QUEUE_RX(n) (!(n % 2)) + +/* Default virtio-net header for passt */ +#define VU_HEADER ((struct virtio_net_hdr){ \ + .flags = VIRTIO_NET_HDR_F_DATA_VALID, \ + .gso_type = VIRTIO_NET_HDR_GSO_NONE, \ +}) + +/** + * vu_queue_enabled - Return state of a virtqueue + * @vq: virtqueue to check + * + * Return: true if the virqueue is enabled, false otherwise + */ +static inline bool vu_queue_enabled(const struct vu_virtq *vq) +{ + return vq->enable; +} + +/** + * vu_queue_started - Return state of a virtqueue + * @vq: virtqueue to check + * + * Return: true if the virqueue is started, false otherwise + */ +static inline bool vu_queue_started(const struct vu_virtq *vq) +{ + return vq->started; +} + +void vu_print_capabilities(void); +void vu_init(struct ctx *c); +void vu_cleanup(struct vu_dev *vdev); +void vu_log_write(const struct vu_dev *vdev, uint64_t address, + uint64_t length); +void vu_control_handler(struct vu_dev *vdev, int fd, uint32_t events); +#endif /* VHOST_USER_H */ diff --git a/virtio.c b/virtio.c new file mode 100644 index 0000000..83906aa --- /dev/null +++ b/virtio.c @@ -0,0 +1,752 @@ +// SPDX-License-Identifier: GPL-2.0-or-later AND BSD-3-Clause +/* + * virtio API, vring and virtqueue functions definition + * + * Copyright Red Hat + * Author: Laurent Vivier <lvivier@redhat.com> + */ + +/* Some parts copied from QEMU subprojects/libvhost-user/libvhost-user.c + * originally licensed under the following terms: + * + * -- + * + * Copyright IBM, Corp. 2007 + * Copyright (c) 2016 Red Hat, Inc. + * + * Authors: + * Anthony Liguori <aliguori@us.ibm.com> + * Marc-André Lureau <mlureau@redhat.com> + * Victor Kaplansky <victork@redhat.com> + * + * This work is licensed under the terms of the GNU GPL, version 2 or + * later. See the COPYING file in the top-level directory. + * + * Some parts copied from QEMU hw/virtio/virtio.c + * licensed under the following terms: + * + * Copyright IBM, Corp. 2007 + * + * Authors: + * Anthony Liguori <aliguori@us.ibm.com> + * + * This work is licensed under the terms of the GNU GPL, version 2. See + * the COPYING file in the top-level directory. + * + * -- + * + * virtq_used_event() and virtq_avail_event() from + * https://docs.oasis-open.org/virtio/virtio/v1.2/csd01/virtio-v1.2-csd01.html#x1-712000A + * licensed under the following terms: + * + * -- + * + * This header is BSD licensed so anyone can use the definitions + * to implement compatible drivers/servers. + * + * Copyright 2007, 2009, IBM Corporation + * Copyright 2011, Red Hat, Inc + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * 3. Neither the name of IBM nor the names of its contributors + * may be used to endorse or promote products derived from this software + * without specific prior written permission. + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ‘‘AS IS’’ AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL IBM OR CONTRIBUTORS BE LIABLE + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ + +#include <stddef.h> +#include <endian.h> +#include <string.h> +#include <errno.h> +#include <sys/eventfd.h> +#include <sys/socket.h> + +#include "util.h" +#include "virtio.h" +#include "vhost_user.h" + +#define VIRTQUEUE_MAX_SIZE 1024 + +/** + * vu_gpa_to_va() - Translate guest physical address to our virtual address. + * @dev: Vhost-user device + * @plen: Physical length to map (input), capped to region (output) + * @guest_addr: Guest physical address + * + * Return: virtual address in our address space of the guest physical address + */ +static void *vu_gpa_to_va(const struct vu_dev *dev, uint64_t *plen, + uint64_t guest_addr) +{ + unsigned int i; + + if (*plen == 0) + return NULL; + + /* Find matching memory region. */ + for (i = 0; i < dev->nregions; i++) { + const struct vu_dev_region *r = &dev->regions[i]; + + if ((guest_addr >= r->gpa) && + (guest_addr < (r->gpa + r->size))) { + if ((guest_addr + *plen) > (r->gpa + r->size)) + *plen = r->gpa + r->size - guest_addr; + /* NOLINTNEXTLINE(performance-no-int-to-ptr) */ + return (void *)(uintptr_t)(guest_addr - r->gpa + + r->mmap_addr + + r->mmap_offset); + } + } + + return NULL; +} + +/** + * vring_avail_flags() - Read the available ring flags + * @vq: Virtqueue + * + * Return: the available ring descriptor flags of the given virtqueue + */ +static inline uint16_t vring_avail_flags(const struct vu_virtq *vq) +{ + return le16toh(vq->vring.avail->flags); +} + +/** + * vring_avail_idx() - Read the available ring index + * @vq: Virtqueue + * + * Return: the available ring index of the given virtqueue + */ +static inline uint16_t vring_avail_idx(struct vu_virtq *vq) +{ + vq->shadow_avail_idx = le16toh(vq->vring.avail->idx); + + return vq->shadow_avail_idx; +} + +/** + * vring_avail_ring() - Read an available ring entry + * @vq: Virtqueue + * @i: Index of the entry to read + * + * Return: the ring entry content (head of the descriptor chain) + */ +static inline uint16_t vring_avail_ring(const struct vu_virtq *vq, int i) +{ + return le16toh(vq->vring.avail->ring[i]); +} + +/** + * virtq_used_event() - Get location of used event indices + * (only with VIRTIO_F_EVENT_IDX) + * @vq: Virtqueue + * + * Return: return the location of the used event index + */ +static inline uint16_t *virtq_used_event(const struct vu_virtq *vq) +{ + /* For backwards compat, used event index is at *end* of avail ring. */ + return &vq->vring.avail->ring[vq->vring.num]; +} + +/** + * vring_get_used_event() - Get the used event from the available ring + * @vq: Virtqueue + * + * Return: the used event (available only if VIRTIO_RING_F_EVENT_IDX is set) + * used_event is a performant alternative where the driver + * specifies how far the device can progress before a notification + * is required. + */ +static inline uint16_t vring_get_used_event(const struct vu_virtq *vq) +{ + return le16toh(*virtq_used_event(vq)); +} + +/** + * virtqueue_get_head() - Get the head of the descriptor chain for a given + * index + * @vq: Virtqueue + * @idx: Available ring entry index + * @head: Head of the descriptor chain + */ +static void virtqueue_get_head(const struct vu_virtq *vq, + unsigned int idx, unsigned int *head) +{ + /* Grab the next descriptor number they're advertising, and increment + * the index we've seen. + */ + *head = vring_avail_ring(vq, idx % vq->vring.num); + + /* If their number is silly, that's a fatal mistake. */ + if (*head >= vq->vring.num) + die("vhost-user: Guest says index %u is available", *head); +} + +/** + * virtqueue_read_indirect_desc() - Copy virtio ring descriptors from guest + * memory + * @dev: Vhost-user device + * @desc: Destination address to copy the descriptors to + * @addr: Guest memory address to copy from + * @len: Length of memory to copy + * + * Return: -1 if there is an error, 0 otherwise + */ +static int virtqueue_read_indirect_desc(const struct vu_dev *dev, + struct vring_desc *desc, + uint64_t addr, size_t len) +{ + uint64_t read_len; + + if (len > (VIRTQUEUE_MAX_SIZE * sizeof(struct vring_desc))) + return -1; + + if (len == 0) + return -1; + + while (len) { + const struct vring_desc *orig_desc; + + read_len = len; + orig_desc = vu_gpa_to_va(dev, &read_len, addr); + if (!orig_desc) + return -1; + + memcpy(desc, orig_desc, read_len); + len -= read_len; + addr += read_len; + /* NOLINTNEXTLINE(bugprone-sizeof-expression,cert-arr39-c) */ + desc += read_len / sizeof(struct vring_desc); + } + + return 0; +} + +/** + * enum virtqueue_read_desc_state - State in the descriptor chain + * @VIRTQUEUE_READ_DESC_ERROR: Found an invalid descriptor + * @VIRTQUEUE_READ_DESC_DONE: No more descriptors in the chain + * @VIRTQUEUE_READ_DESC_MORE: there are more descriptors in the chain + */ +enum virtqueue_read_desc_state { + VIRTQUEUE_READ_DESC_ERROR = -1, + VIRTQUEUE_READ_DESC_DONE = 0, /* end of chain */ + VIRTQUEUE_READ_DESC_MORE = 1, /* more buffers in chain */ +}; + +/** + * virtqueue_read_next_desc() - Read the the next descriptor in the chain + * @desc: Virtio ring descriptors + * @i: Index of the current descriptor + * @max: Maximum value of the descriptor index + * @next: Index of the next descriptor in the chain (output value) + * + * Return: current chain descriptor state (error, next, done) + */ +static int virtqueue_read_next_desc(const struct vring_desc *desc, + int i, unsigned int max, unsigned int *next) +{ + /* If this descriptor says it doesn't chain, we're done. */ + if (!(le16toh(desc[i].flags) & VRING_DESC_F_NEXT)) + return VIRTQUEUE_READ_DESC_DONE; + + /* Check they're not leading us off end of descriptors. */ + *next = le16toh(desc[i].next); + /* Make sure compiler knows to grab that: we don't want it changing! */ + smp_wmb(); + + if (*next >= max) + return VIRTQUEUE_READ_DESC_ERROR; + + return VIRTQUEUE_READ_DESC_MORE; +} + +/** + * vu_queue_empty() - Check if virtqueue is empty + * @vq: Virtqueue + * + * Return: true if the virtqueue is empty, false otherwise + */ +static bool vu_queue_empty(struct vu_virtq *vq) +{ + if (!vq->vring.avail) + return true; + + if (vq->shadow_avail_idx != vq->last_avail_idx) + return false; + + return vring_avail_idx(vq) == vq->last_avail_idx; +} + +/** + * vring_can_notify() - Check if a notification can be sent + * @dev: Vhost-user device + * @vq: Virtqueue + * + * Return: true if notification can be sent + */ +static bool vring_can_notify(const struct vu_dev *dev, struct vu_virtq *vq) +{ + uint16_t old, new; + bool v; + + /* We need to expose used array entries before checking used event. */ + smp_mb(); + + /* Always notify when queue is empty (when feature acknowledge) */ + if (vu_has_feature(dev, VIRTIO_F_NOTIFY_ON_EMPTY) && + !vq->inuse && vu_queue_empty(vq)) + return true; + + if (!vu_has_feature(dev, VIRTIO_RING_F_EVENT_IDX)) + return !(vring_avail_flags(vq) & VRING_AVAIL_F_NO_INTERRUPT); + + v = vq->signalled_used_valid; + vq->signalled_used_valid = true; + old = vq->signalled_used; + new = vq->signalled_used = vq->used_idx; + return !v || vring_need_event(vring_get_used_event(vq), new, old); +} + +/** + * vu_queue_notify() - Send a notification to the given virtqueue + * @dev: Vhost-user device + * @vq: Virtqueue + */ +void vu_queue_notify(const struct vu_dev *dev, struct vu_virtq *vq) +{ + if (!vq->vring.avail) + return; + + if (!vring_can_notify(dev, vq)) { + debug("vhost-user: virtqueue can skip notify..."); + return; + } + + if (eventfd_write(vq->call_fd, 1) < 0) + die_perror("Error writing vhost-user queue eventfd"); +} + +/** + * virtq_avail_event() - Get location of available event indices + * (only with VIRTIO_F_EVENT_IDX) + * @vq: Virtqueue + * + * Return: return the location of the available event index + */ +static inline uint16_t *virtq_avail_event(const struct vu_virtq *vq) +{ + /* For backwards compat, avail event index is at *end* of used ring. */ + return (uint16_t *)&vq->vring.used->ring[vq->vring.num]; +} + +/** + * vring_set_avail_event() - Set avail_event + * @vq: Virtqueue + * @val: Value to set to avail_event + * avail_event is used in the same way the used_event is in the + * avail_ring. + * avail_event is used to advise the driver that notifications + * are unnecessary until the driver writes entry with an index + * specified by avail_event into the available ring. + */ +static inline void vring_set_avail_event(const struct vu_virtq *vq, + uint16_t val) +{ + uint16_t val_le = htole16(val); + + if (!vq->notification) + return; + + memcpy(virtq_avail_event(vq), &val_le, sizeof(val_le)); +} + +/** + * virtqueue_map_desc() - Translate descriptor ring physical address into our + * virtual address space + * @dev: Vhost-user device + * @p_num_sg: First iov entry to use (input), + * first iov entry not used (output) + * @iov: Iov array to use to store buffer virtual addresses + * @max_num_sg: Maximum number of iov entries + * @pa: Guest physical address of the buffer to map into our virtual + * address + * @sz: Size of the buffer + * + * Return: false on error, true otherwise + */ +static bool virtqueue_map_desc(const struct vu_dev *dev, + unsigned int *p_num_sg, struct iovec *iov, + unsigned int max_num_sg, + uint64_t pa, size_t sz) +{ + unsigned int num_sg = *p_num_sg; + + ASSERT(num_sg < max_num_sg); + ASSERT(sz); + + while (sz) { + uint64_t len = sz; + + iov[num_sg].iov_base = vu_gpa_to_va(dev, &len, pa); + if (iov[num_sg].iov_base == NULL) + die("vhost-user: invalid address for buffers"); + iov[num_sg].iov_len = len; + num_sg++; + sz -= len; + pa += len; + } + + *p_num_sg = num_sg; + return true; +} + +/** + * vu_queue_map_desc() - Map the virtqueue descriptor ring into our virtual + * address space + * @dev: Vhost-user device + * @vq: Virtqueue + * @idx: First descriptor ring entry to map + * @elem: Virtqueue element to store descriptor ring iov + * + * Return: -1 if there is an error, 0 otherwise + */ +static int vu_queue_map_desc(const struct vu_dev *dev, + struct vu_virtq *vq, unsigned int idx, + struct vu_virtq_element *elem) +{ + const struct vring_desc *desc = vq->vring.desc; + struct vring_desc desc_buf[VIRTQUEUE_MAX_SIZE]; + unsigned int out_num = 0, in_num = 0; + unsigned int max = vq->vring.num; + unsigned int i = idx; + uint64_t read_len; + int rc; + + if (le16toh(desc[i].flags) & VRING_DESC_F_INDIRECT) { + unsigned int desc_len; + uint64_t desc_addr; + + if (le32toh(desc[i].len) % sizeof(struct vring_desc)) + die("vhost-user: Invalid size for indirect buffer table"); + + /* loop over the indirect descriptor table */ + desc_addr = le64toh(desc[i].addr); + desc_len = le32toh(desc[i].len); + max = desc_len / sizeof(struct vring_desc); + read_len = desc_len; + desc = vu_gpa_to_va(dev, &read_len, desc_addr); + if (desc && read_len != desc_len) { + /* Failed to use zero copy */ + desc = NULL; + if (!virtqueue_read_indirect_desc(dev, desc_buf, desc_addr, desc_len)) + desc = desc_buf; + } + if (!desc) + die("vhost-user: Invalid indirect buffer table"); + i = 0; + } + + /* Collect all the descriptors */ + do { + if (le16toh(desc[i].flags) & VRING_DESC_F_WRITE) { + if (!virtqueue_map_desc(dev, &in_num, elem->in_sg, + elem->in_num, + le64toh(desc[i].addr), + le32toh(desc[i].len))) + return -1; + } else { + if (in_num) + die("Incorrect order for descriptors"); + if (!virtqueue_map_desc(dev, &out_num, elem->out_sg, + elem->out_num, + le64toh(desc[i].addr), + le32toh(desc[i].len))) { + return -1; + } + } + + /* If we've got too many, that implies a descriptor loop. */ + if ((in_num + out_num) > max) + die("vhost-user: Loop in queue descriptor list"); + rc = virtqueue_read_next_desc(desc, i, max, &i); + } while (rc == VIRTQUEUE_READ_DESC_MORE); + + if (rc == VIRTQUEUE_READ_DESC_ERROR) + die("vhost-user: Failed to read descriptor list"); + + elem->index = idx; + elem->in_num = in_num; + elem->out_num = out_num; + + return 0; +} + +/** + * vu_queue_pop() - Pop an entry from the virtqueue + * @dev: Vhost-user device + * @vq: Virtqueue + * @elem: Virtqueue element to fill with the entry information + * + * Return: -1 if there is an error, 0 otherwise + */ +int vu_queue_pop(const struct vu_dev *dev, struct vu_virtq *vq, + struct vu_virtq_element *elem) +{ + unsigned int head; + int ret; + + if (!vq->vring.avail) + return -1; + + if (vu_queue_empty(vq)) + return -1; + + /* Needed after vu_queue_empty(), see comment in + * virtqueue_num_heads(). + */ + smp_rmb(); + + if (vq->inuse >= vq->vring.num) + die("vhost-user queue size exceeded"); + + virtqueue_get_head(vq, vq->last_avail_idx++, &head); + + if (vu_has_feature(dev, VIRTIO_RING_F_EVENT_IDX)) + vring_set_avail_event(vq, vq->last_avail_idx); + + ret = vu_queue_map_desc(dev, vq, head, elem); + + if (ret < 0) + return ret; + + vq->inuse++; + + return 0; +} + +/** + * vu_queue_detach_element() - Detach an element from the virtqueue + * @vq: Virtqueue + */ +void vu_queue_detach_element(struct vu_virtq *vq) +{ + vq->inuse--; + /* unmap, when DMA support is added */ +} + +/** + * vu_queue_unpop() - Push back the previously popped element from the virtqueue + * @vq: Virtqueue + */ +/* cppcheck-suppress unusedFunction */ +void vu_queue_unpop(struct vu_virtq *vq) +{ + vq->last_avail_idx--; + vu_queue_detach_element(vq); +} + +/** + * vu_queue_rewind() - Push back a given number of popped elements + * @vq: Virtqueue + * @num: Number of element to unpop + * + * Return: True on success, false if not + */ +bool vu_queue_rewind(struct vu_virtq *vq, unsigned int num) +{ + if (num > vq->inuse) + return false; + + vq->last_avail_idx -= num; + vq->inuse -= num; + return true; +} + +/** + * vring_used_write() - Write an entry in the used ring + * @dev: Vhost-user device + * @vq: Virtqueue + * @uelem: Entry to write + * @i: Index of the entry in the used ring + */ +static inline void vring_used_write(const struct vu_dev *vdev, + struct vu_virtq *vq, + const struct vring_used_elem *uelem, int i) +{ + struct vring_used *used = vq->vring.used; + + used->ring[i] = *uelem; + vu_log_write(vdev, vq->vring.log_guest_addr + + offsetof(struct vring_used, ring[i]), + sizeof(used->ring[i])); +} + +/** + * vu_log_queue_fill() - Log virtqueue memory update + * @dev: vhost-user device + * @vq: Virtqueue + * @index: Descriptor ring index + * @len: Size of the element + */ +static void vu_log_queue_fill(const struct vu_dev *vdev, struct vu_virtq *vq, + unsigned int index, unsigned int len) +{ + struct vring_desc desc_buf[VIRTQUEUE_MAX_SIZE]; + struct vring_desc *desc = vq->vring.desc; + unsigned int max, min; + unsigned num_bufs = 0; + uint64_t read_len; + + if (!vdev->log_table || !len || !vu_has_feature(vdev, VHOST_F_LOG_ALL)) + return; + + max = vq->vring.num; + + if (le16toh(desc[index].flags) & VRING_DESC_F_INDIRECT) { + unsigned int desc_len; + uint64_t desc_addr; + + if (le32toh(desc[index].len) % sizeof(struct vring_desc)) + die("Invalid size for indirect buffer table"); + + /* loop over the indirect descriptor table */ + desc_addr = le64toh(desc[index].addr); + desc_len = le32toh(desc[index].len); + max = desc_len / sizeof(struct vring_desc); + read_len = desc_len; + desc = vu_gpa_to_va(vdev, &read_len, desc_addr); + if (desc && read_len != desc_len) { + /* Failed to use zero copy */ + desc = NULL; + if (!virtqueue_read_indirect_desc(vdev, desc_buf, + desc_addr, + desc_len)) + desc = desc_buf; + } + + if (!desc) + die("Invalid indirect buffer table"); + + index = 0; + } + + do { + if (++num_bufs > max) + die("Looped descriptor"); + + if (le16toh(desc[index].flags) & VRING_DESC_F_WRITE) { + min = MIN(le32toh(desc[index].len), len); + vu_log_write(vdev, le64toh(desc[index].addr), min); + len -= min; + } + } while (len > 0 && + (virtqueue_read_next_desc(desc, index, max, &index) == + VIRTQUEUE_READ_DESC_MORE)); +} + + +/** + * vu_queue_fill_by_index() - Update information of a descriptor ring entry + * in the used ring + * @dev: Vhost-user device + * @vq: Virtqueue + * @index: Descriptor ring index + * @len: Size of the element + * @idx: Used ring entry index + */ +static void vu_queue_fill_by_index(const struct vu_dev *vdev, + struct vu_virtq *vq, + unsigned int index, unsigned int len, + unsigned int idx) +{ + struct vring_used_elem uelem; + + if (!vq->vring.avail) + return; + + vu_log_queue_fill(vdev, vq, index, len); + + idx = (idx + vq->used_idx) % vq->vring.num; + + uelem.id = htole32(index); + uelem.len = htole32(len); + vring_used_write(vdev, vq, &uelem, idx); +} + +/** + * vu_queue_fill() - Update information of a given element in the used ring + * @dev: Vhost-user device + * @vq: Virtqueue + * @elem: Element information to fill + * @len: Size of the element + * @idx: Used ring entry index + */ +void vu_queue_fill(const struct vu_dev *vdev, struct vu_virtq *vq, + const struct vu_virtq_element *elem, unsigned int len, + unsigned int idx) +{ + vu_queue_fill_by_index(vdev, vq, elem->index, len, idx); +} + +/** + * vring_used_idx_set() - Set the descriptor ring current index + * @dev: Vhost-user device + * @vq: Virtqueue + * @val: Value to set in the index + */ +static inline void vring_used_idx_set(const struct vu_dev *vdev, + struct vu_virtq *vq, uint16_t val) +{ + vq->vring.used->idx = htole16(val); + vu_log_write(vdev, vq->vring.log_guest_addr + + offsetof(struct vring_used, idx), + sizeof(vq->vring.used->idx)); + + vq->used_idx = val; +} + +/** + * vu_queue_flush() - Flush the virtqueue + * @dev: Vhost-user device + * @vq: Virtqueue + * @count: Number of entry to flush + */ +void vu_queue_flush(const struct vu_dev *vdev, struct vu_virtq *vq, + unsigned int count) +{ + uint16_t old, new; + + if (!vq->vring.avail) + return; + + /* Make sure buffer is written before we update index. */ + smp_wmb(); + + old = vq->used_idx; + new = old + count; + vring_used_idx_set(vdev, vq, new); + vq->inuse -= count; + if ((uint16_t)(new - vq->signalled_used) < (uint16_t)(new - old)) + vq->signalled_used_valid = false; +} diff --git a/virtio.h b/virtio.h new file mode 100644 index 0000000..d8beb88 --- /dev/null +++ b/virtio.h @@ -0,0 +1,188 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * virtio API, vring and virtqueue functions definition + * + * Copyright Red Hat + * Author: Laurent Vivier <lvivier@redhat.com> + */ + +#ifndef VIRTIO_H +#define VIRTIO_H + +#include <stdbool.h> +#include <linux/vhost_types.h> + +/* Maximum size of a virtqueue */ +#define VIRTQUEUE_MAX_SIZE 1024 + +/** + * struct vu_ring - Virtqueue rings + * @num: Size of the queue + * @desc: Descriptor ring + * @avail: Available ring + * @used: Used ring + * @log_guest_addr: Guest address for logging + * @flags: Vring flags + * VHOST_VRING_F_LOG is set if log address is valid + */ +struct vu_ring { + unsigned int num; + struct vring_desc *desc; + struct vring_avail *avail; + struct vring_used *used; + uint64_t log_guest_addr; + uint32_t flags; +}; + +/** + * struct vu_virtq - Virtqueue definition + * @vring: Virtqueue rings + * @last_avail_idx: Next head to pop + * @shadow_avail_idx: Last avail_idx read from VQ. + * @used_idx: Descriptor ring current index + * @signalled_used: Last used index value we have signalled on + * @signalled_used_valid: True if signalled_used if valid + * @notification: True if the queues notify (via event + * index or interrupt) + * @inuse: Number of entries in use + * @call_fd: The event file descriptor to signal when + * buffers are used. + * @kick_fd: The event file descriptor for adding + * buffers to the vring + * @err_fd: The event file descriptor to signal when + * error occurs + * @enable: True if the virtqueue is enabled + * @started: True if the virtqueue is started + * @vra: QEMU address of our rings + */ +struct vu_virtq { + struct vu_ring vring; + uint16_t last_avail_idx; + uint16_t shadow_avail_idx; + uint16_t used_idx; + uint16_t signalled_used; + bool signalled_used_valid; + bool notification; + unsigned int inuse; + int call_fd; + int kick_fd; + int err_fd; + unsigned int enable; + bool started; + struct vhost_vring_addr vra; +}; + +/** + * struct vu_dev_region - guest shared memory region + * @gpa: Guest physical address of the region + * @size: Memory size in bytes + * @qva: QEMU virtual address + * @mmap_offset: Offset where the region starts in the mapped memory + * @mmap_addr: Address of the mapped memory + */ +struct vu_dev_region { + uint64_t gpa; + uint64_t size; + uint64_t qva; + uint64_t mmap_offset; + uint64_t mmap_addr; +}; + +#define VHOST_USER_MAX_QUEUES 2 + +/* + * Set a reasonable maximum number of ram slots, which will be supported by + * any architecture. + */ +#define VHOST_USER_MAX_RAM_SLOTS 32 + +/** + * struct vu_dev - vhost-user device information + * @context: Execution context + * @nregions: Number of shared memory regions + * @regions: Guest shared memory regions + * @features: Vhost-user features + * @protocol_features: Vhost-user protocol features + * @log_call_fd: Eventfd to report logging update + * @log_size: Size of the logging memory region + * @log_table: Base of the logging memory region + */ +struct vu_dev { + struct ctx *context; + uint32_t nregions; + struct vu_dev_region regions[VHOST_USER_MAX_RAM_SLOTS]; + struct vu_virtq vq[VHOST_USER_MAX_QUEUES]; + uint64_t features; + uint64_t protocol_features; + int log_call_fd; + uint64_t log_size; + uint8_t *log_table; +}; + +/** + * struct vu_virtq_element - virtqueue element + * @index: Descriptor ring index + * @out_num: Number of outgoing iovec buffers + * @in_num: Number of incoming iovec buffers + * @in_sg: Incoming iovec buffers + * @out_sg: Outgoing iovec buffers + */ +struct vu_virtq_element { + unsigned int index; + unsigned int out_num; + unsigned int in_num; + struct iovec *in_sg; + struct iovec *out_sg; +}; + +/** + * has_feature() - Check a feature bit in a features set + * @features: Features set + * @fb: Feature bit to check + * + * Return: True if the feature bit is set + */ +static inline bool has_feature(uint64_t features, unsigned int fbit) +{ + return !!(features & (1ULL << fbit)); +} + +/** + * vu_has_feature() - Check if a virtio-net feature is available + * @vdev: Vhost-user device + * @fbit: Feature to check + * + * Return: True if the feature is available + */ +static inline bool vu_has_feature(const struct vu_dev *vdev, + unsigned int fbit) +{ + return has_feature(vdev->features, fbit); +} + +/** + * vu_has_protocol_feature() - Check if a vhost-user feature is available + * @vdev: Vhost-user device + * @fbit: Feature to check + * + * Return: True if the feature is available + */ +/* cppcheck-suppress unusedFunction */ +static inline bool vu_has_protocol_feature(const struct vu_dev *vdev, + unsigned int fbit) +{ + return has_feature(vdev->protocol_features, fbit); +} + +void vu_queue_notify(const struct vu_dev *dev, struct vu_virtq *vq); +int vu_queue_pop(const struct vu_dev *dev, struct vu_virtq *vq, + struct vu_virtq_element *elem); +void vu_queue_detach_element(struct vu_virtq *vq); +void vu_queue_unpop(struct vu_virtq *vq); +bool vu_queue_rewind(struct vu_virtq *vq, unsigned int num); +void vu_queue_fill(const struct vu_dev *vdev, struct vu_virtq *vq, + const struct vu_virtq_element *elem, unsigned int len, + unsigned int idx); +void vu_queue_flush(const struct vu_dev *vdev, struct vu_virtq *vq, + unsigned int count); +#endif /* VIRTIO_H */ diff --git a/vu_common.c b/vu_common.c new file mode 100644 index 0000000..5e6fd4a --- /dev/null +++ b/vu_common.c @@ -0,0 +1,309 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* Copyright Red Hat + * Author: Laurent Vivier <lvivier@redhat.com> + * + * common_vu.c - vhost-user common UDP and TCP functions + */ + +#include <errno.h> +#include <unistd.h> +#include <sys/uio.h> +#include <sys/eventfd.h> +#include <netinet/if_ether.h> +#include <linux/virtio_net.h> + +#include "util.h" +#include "passt.h" +#include "tap.h" +#include "vhost_user.h" +#include "pcap.h" +#include "vu_common.h" +#include "migrate.h" + +#define VU_MAX_TX_BUFFER_NB 2 + +/** + * vu_packet_check_range() - Check if a given memory zone is contained in + * a mapped guest memory region + * @buf: Array of the available memory regions + * @ptr: Start of desired data range + * @size: Length of desired data range + * + * Return: 0 if the zone is in a mapped memory region, -1 otherwise + */ +int vu_packet_check_range(void *buf, const char *ptr, size_t len) +{ + struct vu_dev_region *dev_region; + + for (dev_region = buf; dev_region->mmap_addr; dev_region++) { + uintptr_t base_addr = dev_region->mmap_addr + + dev_region->mmap_offset; + /* NOLINTNEXTLINE(performance-no-int-to-ptr) */ + const char *base = (const char *)base_addr; + + ASSERT(base_addr >= dev_region->mmap_addr); + + if (len <= dev_region->size && base <= ptr && + (size_t)(ptr - base) <= dev_region->size - len) + return 0; + } + + return -1; +} + +/** + * vu_init_elem() - initialize an array of virtqueue elements with 1 iov in each + * @elem: Array of virtqueue elements to initialize + * @iov: Array of iovec to assign to virtqueue element + * @elem_cnt: Number of virtqueue element + */ +void vu_init_elem(struct vu_virtq_element *elem, struct iovec *iov, int elem_cnt) +{ + int i; + + for (i = 0; i < elem_cnt; i++) + vu_set_element(&elem[i], NULL, &iov[i]); +} + +/** + * vu_collect() - collect virtio buffers from a given virtqueue + * @vdev: vhost-user device + * @vq: virtqueue to collect from + * @elem: Array of virtqueue element + * each element must be initialized with one iovec entry + * in the in_sg array. + * @max_elem: Number of virtqueue elements in the array + * @size: Maximum size of the data in the frame + * @frame_size: The total size of the buffers (output) + * + * Return: number of elements used to contain the frame + */ +int vu_collect(const struct vu_dev *vdev, struct vu_virtq *vq, + struct vu_virtq_element *elem, int max_elem, + size_t size, size_t *frame_size) +{ + size_t current_size = 0; + int elem_cnt = 0; + + while (current_size < size && elem_cnt < max_elem) { + struct iovec *iov; + int ret; + + ret = vu_queue_pop(vdev, vq, &elem[elem_cnt]); + if (ret < 0) + break; + + if (elem[elem_cnt].in_num < 1) { + warn("virtio-net receive queue contains no in buffers"); + vu_queue_detach_element(vq); + break; + } + + iov = &elem[elem_cnt].in_sg[0]; + + if (iov->iov_len > size - current_size) + iov->iov_len = size - current_size; + + current_size += iov->iov_len; + elem_cnt++; + + if (!vu_has_feature(vdev, VIRTIO_NET_F_MRG_RXBUF)) + break; + } + + if (frame_size) + *frame_size = current_size; + + return elem_cnt; +} + +/** + * vu_set_vnethdr() - set virtio-net headers + * @vdev: vhost-user device + * @vnethdr: Address of the header to set + * @num_buffers: Number of guest buffers of the frame + */ +void vu_set_vnethdr(const struct vu_dev *vdev, + struct virtio_net_hdr_mrg_rxbuf *vnethdr, + int num_buffers) +{ + vnethdr->hdr = VU_HEADER; + if (vu_has_feature(vdev, VIRTIO_NET_F_MRG_RXBUF)) + vnethdr->num_buffers = htole16(num_buffers); +} + +/** + * vu_flush() - flush all the collected buffers to the vhost-user interface + * @vdev: vhost-user device + * @vq: vhost-user virtqueue + * @elem: virtqueue elements array to send back to the virtqueue + * @elem_cnt: Length of the array + */ +void vu_flush(const struct vu_dev *vdev, struct vu_virtq *vq, + struct vu_virtq_element *elem, int elem_cnt) +{ + int i; + + for (i = 0; i < elem_cnt; i++) + vu_queue_fill(vdev, vq, &elem[i], elem[i].in_sg[0].iov_len, i); + + vu_queue_flush(vdev, vq, elem_cnt); + vu_queue_notify(vdev, vq); +} + +/** + * vu_handle_tx() - Receive data from the TX virtqueue + * @vdev: vhost-user device + * @index: index of the virtqueue + * @now: Current timestamp + */ +static void vu_handle_tx(struct vu_dev *vdev, int index, + const struct timespec *now) +{ + struct vu_virtq_element elem[VIRTQUEUE_MAX_SIZE]; + struct iovec out_sg[VIRTQUEUE_MAX_SIZE]; + struct vu_virtq *vq = &vdev->vq[index]; + int hdrlen = sizeof(struct virtio_net_hdr_mrg_rxbuf); + int out_sg_count; + int count; + + ASSERT(VHOST_USER_IS_QUEUE_TX(index)); + + tap_flush_pools(); + + count = 0; + out_sg_count = 0; + while (count < VIRTQUEUE_MAX_SIZE && + out_sg_count + VU_MAX_TX_BUFFER_NB <= VIRTQUEUE_MAX_SIZE) { + int ret; + + elem[count].out_num = VU_MAX_TX_BUFFER_NB; + elem[count].out_sg = &out_sg[out_sg_count]; + elem[count].in_num = 0; + elem[count].in_sg = NULL; + + ret = vu_queue_pop(vdev, vq, &elem[count]); + if (ret < 0) + break; + out_sg_count += elem[count].out_num; + + if (elem[count].out_num < 1) { + warn("virtio-net transmit queue contains no out buffers"); + break; + } + if (elem[count].out_num == 1) { + tap_add_packet(vdev->context, + elem[count].out_sg[0].iov_len - hdrlen, + (char *)elem[count].out_sg[0].iov_base + + hdrlen, now); + } else { + /* vnet header can be in a separate iovec */ + if (elem[count].out_num != 2) { + debug("virtio-net transmit queue contains more than one buffer ([%d]: %u)", + count, elem[count].out_num); + } else if (elem[count].out_sg[0].iov_len != (size_t)hdrlen) { + debug("virtio-net transmit queue entry not aligned on hdrlen ([%d]: %d != %zu)", + count, hdrlen, elem[count].out_sg[0].iov_len); + } else { + tap_add_packet(vdev->context, + elem[count].out_sg[1].iov_len, + (char *)elem[count].out_sg[1].iov_base, + now); + } + } + + count++; + } + tap_handler(vdev->context, now); + + if (count) { + int i; + + for (i = 0; i < count; i++) + vu_queue_fill(vdev, vq, &elem[i], 0, i); + vu_queue_flush(vdev, vq, count); + vu_queue_notify(vdev, vq); + } +} + +/** + * vu_kick_cb() - Called on a kick event to start to receive data + * @vdev: vhost-user device + * @ref: epoll reference information + * @now: Current timestamp + */ +void vu_kick_cb(struct vu_dev *vdev, union epoll_ref ref, + const struct timespec *now) +{ + eventfd_t kick_data; + ssize_t rc; + + rc = eventfd_read(ref.fd, &kick_data); + if (rc == -1) + die_perror("vhost-user kick eventfd_read()"); + + trace("vhost-user: got kick_data: %016"PRIx64" idx: %d", + kick_data, ref.queue); + if (VHOST_USER_IS_QUEUE_TX(ref.queue)) + vu_handle_tx(vdev, ref.queue, now); +} + +/** + * vu_send_single() - Send a buffer to the front-end using the RX virtqueue + * @c: execution context + * @buf: address of the buffer + * @size: size of the buffer + * + * Return: number of bytes sent, -1 if there is an error + */ +int vu_send_single(const struct ctx *c, const void *buf, size_t size) +{ + struct vu_dev *vdev = c->vdev; + struct vu_virtq *vq = &vdev->vq[VHOST_USER_RX_QUEUE]; + struct vu_virtq_element elem[VIRTQUEUE_MAX_SIZE]; + struct iovec in_sg[VIRTQUEUE_MAX_SIZE]; + size_t total; + int elem_cnt; + int i; + + trace("vu_send_single size %zu", size); + + if (!vu_queue_enabled(vq) || !vu_queue_started(vq)) { + debug("Got packet, but RX virtqueue not usable yet"); + return -1; + } + + vu_init_elem(elem, in_sg, VIRTQUEUE_MAX_SIZE); + + size += sizeof(struct virtio_net_hdr_mrg_rxbuf); + elem_cnt = vu_collect(vdev, vq, elem, VIRTQUEUE_MAX_SIZE, size, &total); + if (total < size) { + debug("vu_send_single: no space to send the data " + "elem_cnt %d size %zd", elem_cnt, total); + goto err; + } + + vu_set_vnethdr(vdev, in_sg[0].iov_base, elem_cnt); + + total -= sizeof(struct virtio_net_hdr_mrg_rxbuf); + + /* copy data from the buffer to the iovec */ + iov_from_buf(in_sg, elem_cnt, sizeof(struct virtio_net_hdr_mrg_rxbuf), + buf, total); + + if (*c->pcap) { + pcap_iov(in_sg, elem_cnt, + sizeof(struct virtio_net_hdr_mrg_rxbuf)); + } + + vu_flush(vdev, vq, elem, elem_cnt); + + trace("vhost-user sent %zu", total); + + return total; +err: + for (i = 0; i < elem_cnt; i++) + vu_queue_detach_element(vq); + + return -1; +} diff --git a/vu_common.h b/vu_common.h new file mode 100644 index 0000000..f538f23 --- /dev/null +++ b/vu_common.h @@ -0,0 +1,61 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later + * Copyright Red Hat + * Author: Laurent Vivier <lvivier@redhat.com> + * + * vhost-user common UDP and TCP functions + */ + +#ifndef VU_COMMON_H +#define VU_COMMON_H +#include <linux/virtio_net.h> + +static inline void *vu_eth(void *base) +{ + return ((char *)base + sizeof(struct virtio_net_hdr_mrg_rxbuf)); +} + +static inline void *vu_ip(void *base) +{ + return (struct ethhdr *)vu_eth(base) + 1; +} + +static inline void *vu_payloadv4(void *base) +{ + return (struct iphdr *)vu_ip(base) + 1; +} + +static inline void *vu_payloadv6(void *base) +{ + return (struct ipv6hdr *)vu_ip(base) + 1; +} + +/** + * vu_set_element() - Initialize a vu_virtq_element + * @elem: Element to initialize + * @out_sg: One out iovec entry to set in elem + * @in_sg: One in iovec entry to set in elem + */ +static inline void vu_set_element(struct vu_virtq_element *elem, + struct iovec *out_sg, struct iovec *in_sg) +{ + elem->out_num = !!out_sg; + elem->out_sg = out_sg; + elem->in_num = !!in_sg; + elem->in_sg = in_sg; +} + +void vu_init_elem(struct vu_virtq_element *elem, struct iovec *iov, + int elem_cnt); +int vu_collect(const struct vu_dev *vdev, struct vu_virtq *vq, + struct vu_virtq_element *elem, int max_elem, size_t size, + size_t *frame_size); +void vu_set_vnethdr(const struct vu_dev *vdev, + struct virtio_net_hdr_mrg_rxbuf *vnethdr, + int num_buffers); +void vu_flush(const struct vu_dev *vdev, struct vu_virtq *vq, + struct vu_virtq_element *elem, int elem_cnt); +void vu_kick_cb(struct vu_dev *vdev, union epoll_ref ref, + const struct timespec *now); +int vu_send_single(const struct ctx *c, const void *buf, size_t size); + +#endif /* VU_COMMON_H */