On Wed, Aug 25, 2021 at 2:07 PM Yucong Sun <fallentree@xxxxxx> wrote:
>
> This patch adds similar retry logic to more places where read() is used, to
> reduce flakyness in slow CI environment.
>
> Signed-off-by: Yucong Sun <fallentree@xxxxxx>
> ---
> .../selftests/bpf/prog_tests/sockmap_listen.c | 19 ++++++++++++++++---
> 1 file changed, 16 insertions(+), 3 deletions(-)
>
> diff --git a/tools/testing/selftests/bpf/prog_tests/sockmap_listen.c b/tools/testing/selftests/bpf/prog_tests/sockmap_listen.c
> index 6a5df28f9a3d..5c5979046523 100644
> --- a/tools/testing/selftests/bpf/prog_tests/sockmap_listen.c
> +++ b/tools/testing/selftests/bpf/prog_tests/sockmap_listen.c
> @@ -949,6 +949,7 @@ static void redir_to_connected(int family, int sotype, int sock_mapfd,
> int err, n;
> u32 key;
> char b;
> + int retries = 100;
>
> zero_verdict_count(verd_mapfd);
>
> @@ -1001,10 +1002,15 @@ static void redir_to_connected(int family, int sotype, int sock_mapfd,
> goto close_peer1;
> if (pass != 1)
> FAIL("%s: want pass count 1, have %d", log_prefix, pass);
> -
> +again:
> n = read(c0, &b, 1);
> - if (n < 0)
> + if (n < 0) {
> + if (errno == EAGAIN && retries--) {
TCP was fixed differently in
commit 30b4cb36b111 ("selftests/bpf: Fix spurious failures in accept
due to EAGAIN").
Would a similar approach work here?
