tap: Revert recently added checks in tap_handler_passt()

This reverts commit 198f87835d ("tap: Return -EIO from
tap_handler_passt() on inconsistent packet stream") and commit
510dace86c ("tap: Keep stream consistent if qemu length descriptor
spans two recv() calls").

I can hit occasional failures in perf/passt_tcp tests where we seem
to be getting excess data at the end of a recv(), and for some reason
I couldn't figure out yet, if we just ignore it, subsequent recv()
calls from qemu return correct data. If we close the connection, qemu
can't talk to us anymore, of course.

Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
This commit is contained in:
Stefano Brivio 2022-11-16 13:58:13 +01:00
parent b27d6d121c
commit 25dab96205

20
tap.c
View file

@ -747,31 +747,19 @@ redo:
return -ECONNRESET; return -ECONNRESET;
} }
while (n > 0) { while (n > (ssize_t)sizeof(uint32_t)) {
ssize_t len; ssize_t len = ntohl(*(uint32_t *)p);
/* Force receiving at least a complete length descriptor to
* avoid an inconsistent stream.
*/
if (n < (ssize_t)sizeof(uint32_t)) {
rem = recv(c->fd_tap, p + n,
(ssize_t)sizeof(uint32_t) - n, 0);
if ((n += rem) != (ssize_t)sizeof(uint32_t))
return -EIO;
}
len = ntohl(*(uint32_t *)p);
p += sizeof(uint32_t); p += sizeof(uint32_t);
n -= sizeof(uint32_t); n -= sizeof(uint32_t);
/* At most one packet might not fit in a single read, and this /* At most one packet might not fit in a single read, and this
* also needs to be blocking. * needs to be blocking.
*/ */
if (len > n) { if (len > n) {
rem = recv(c->fd_tap, p + n, len - n, 0); rem = recv(c->fd_tap, p + n, len - n, 0);
if ((n += rem) != len) if ((n += rem) != len)
return -EIO; return 0;
} }
/* Complete the partial read above before discarding a malformed /* Complete the partial read above before discarding a malformed