selftests/bpf: adding delay in socketmap_listen to reduce flakyness
Subsystems (3)
PATCH 1/1 selftests/bpf: adding delay in socketmap_listen to reduce flakyness
This patch adds a 1ms delay to reduce flakyness of the test.
--- .../selftests/bpf/prog_tests/sockmap_listen.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-)
tools/testing/selftests/bpf/prog_tests/sockmap_listen.c +9 -3
@@ -1603,8 +1603,10 @@ static void unix_redir_to_connected(int sotype, int sock_mapfd, again: n =read (mode ==REDIR_INGRESS ? p0 : c0, &b,1 );if (n <0 ) { -if (errno ==EAGAIN && retries--) +if (errno ==EAGAIN && retries--) { +usleep (1000 ); goto again; + }FAIL_ERRNO ("%s: read" , log_prefix); }if (n ==0 )
@@ -1776,8 +1778,10 @@ static void udp_redir_to_connected(int family, int sock_mapfd, int verd_mapfd, again: n =read (mode ==REDIR_INGRESS ? p0 : c0, &b,1 );if (n <0 ) { -if (errno ==EAGAIN && retries--) +if (errno ==EAGAIN && retries--) { +usleep (1000 ); goto again; + }FAIL_ERRNO ("%s: read" , log_prefix); }if (n ==0 )
@@ -1869,8 +1873,10 @@ static void inet_unix_redir_to_connected(int family, int type, int sock_mapfd, again: n =read (mode ==REDIR_INGRESS ? p0 : c0, &b,1 );if (n <0 ) { -if (errno ==EAGAIN && retries--) +if (errno ==EAGAIN && retries--) { +usleep (1000 ); goto again; + }FAIL_ERRNO ("%s: read" , log_prefix); }if (n ==0 )
Signature · 1 line
2.30.2
Quoted 5 lines · Yucong Sun
This patch adds a 1ms delay to reduce flakyness of the test.
Signed-off-by Yucong Sun <[email protected]>---
Any reasons to not implement exponential back-off, like we did for test_maps?
Quoted 46 lines
.../selftests/bpf/prog_tests/sockmap_listen.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-)tools/testing/selftests/bpf/prog_tests/sockmap_listen.c +9
-4@@ -1603,8 +1603,10 @@ static void unix_redir_to_connected(int sotype, int sock_mapfd, again: n =read (mode ==REDIR_INGRESS ? p0 : c0, &b,1 );if (n <0 ) { -if (errno ==EAGAIN && retries--) +if (errno ==EAGAIN && retries--) { +usleep (1000 ); goto again; + }FAIL_ERRNO ("%s: read" , log_prefix); }if (n ==0 )@@ -1776,8 +1778,10 @@ static void udp_redir_to_connected(int family, int sock_mapfd, int verd_mapfd, again: n =read (mode ==REDIR_INGRESS ? p0 : c0, &b,1 );if (n <0 ) { -if (errno ==EAGAIN && retries--) +if (errno ==EAGAIN && retries--) { +usleep (1000 ); goto again; + }FAIL_ERRNO ("%s: read" , log_prefix); }if (n ==0 )@@ -1869,8 +1873,10 @@ static void inet_unix_redir_to_connected(int family, int type, int sock_mapfd, again: n =read (mode ==REDIR_INGRESS ? p0 : c0, &b,1 );if (n <0 ) { -if (errno ==EAGAIN && retries--) +if (errno ==EAGAIN && retries--) { +usleep (1000 ); goto again; + }FAIL_ERRNO ("%s: read" , log_prefix); }if (n ==0 ) --2.30.2
Quoted 9 lines · Andrii Nakryiko
Quoted 5 lines · Yucong Sun
This patch adds a 1ms delay to reduce flakyness of the test.
Signed-off-by Yucong Sun <[email protected]>---
Any reasons to not implement exponential back-off, like we did for
test_maps?
for simplicity, since there are no contention involved here I figured we don’t need random delay and back-offs.
(sorry for resending, I was fooled by the mobile gmail client that it doesn't do plain text).
Quoted 47 lines
Quoted 47 lines
.../selftests/bpf/prog_tests/sockmap_listen.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-)tools/testing/selftests/bpf/prog_tests/sockmap_listen.c +9
-4@@ -1603,8 +1603,10 @@ static void unix_redir_to_connected(int sotype, int sock_mapfd, again: n =read (mode ==REDIR_INGRESS ? p0 : c0, &b,1 );if (n <0 ) { -if (errno ==EAGAIN && retries--) +if (errno ==EAGAIN && retries--) { +usleep (1000 ); goto again; + }FAIL_ERRNO ("%s: read" , log_prefix); }if (n ==0 )@@ -1776,8 +1778,10 @@ static void udp_redir_to_connected(int family, int sock_mapfd, int verd_mapfd, again: n =read (mode ==REDIR_INGRESS ? p0 : c0, &b,1 );if (n <0 ) { -if (errno ==EAGAIN && retries--) +if (errno ==EAGAIN && retries--) { +usleep (1000 ); goto again; + }FAIL_ERRNO ("%s: read" , log_prefix); }if (n ==0 )@@ -1869,8 +1873,10 @@ static void inet_unix_redir_to_connected(int family, int type, int sock_mapfd, again: n =read (mode ==REDIR_INGRESS ? p0 : c0, &b,1 );if (n <0 ) { -if (errno ==EAGAIN && retries--) +if (errno ==EAGAIN && retries--) { +usleep (1000 ); goto again; + }FAIL_ERRNO ("%s: read" , log_prefix); }if (n ==0 ) --2.30.2
Quoted 8 lines · On Thu, Aug 19, 2021 at 12:16 PM [email protected] <[email protected]>…
Quoted 5 lines · Andrii Nakryiko
Quoted 3 lines · Yucong Sun
This patch adds a 1ms delay to reduce flakyness of the test.
Quoted 7 lines
Quoted 4 lines
Quoted 2 lines
Signed-off-by Yucong Sun <[email protected]>---
Any reasons to not implement exponential back-off, like we did for
test_maps?for simplicity, since there are no contention involved here I figured we don’t need random delay and back-offs.
Alright, it's an improvement anyways. Let's see if it still causes problem. Applied to bpf-next, thanks!
Quoted 4 lines
(sorry for resending, I was fooled by the mobile gmail client that it doesn't do plain text).
AFAIK, mobile client doesn't support plain text
Quoted 48 lines
Quoted 48 lines
Quoted 48 lines
.../selftests/bpf/prog_tests/sockmap_listen.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-)tools/testing/selftests/bpf/prog_tests/sockmap_listen.c +9
-4@@ -1603,8 +1603,10 @@ static void unix_redir_to_connected(int sotype, int sock_mapfd, again: n =read (mode ==REDIR_INGRESS ? p0 : c0, &b,1 );if (n <0 ) { -if (errno ==EAGAIN && retries--) +if (errno ==EAGAIN && retries--) { +usleep (1000 ); goto again; + }FAIL_ERRNO ("%s: read" , log_prefix); }if (n ==0 )@@ -1776,8 +1778,10 @@ static void udp_redir_to_connected(int family, int sock_mapfd, int verd_mapfd, again: n =read (mode ==REDIR_INGRESS ? p0 : c0, &b,1 );if (n <0 ) { -if (errno ==EAGAIN && retries--) +if (errno ==EAGAIN && retries--) { +usleep (1000 ); goto again; + }FAIL_ERRNO ("%s: read" , log_prefix); }if (n ==0 )@@ -1869,8 +1873,10 @@ static void inet_unix_redir_to_connected(int family, int type, int sock_mapfd, again: n =read (mode ==REDIR_INGRESS ? p0 : c0, &b,1 );if (n <0 ) { -if (errno ==EAGAIN && retries--) +if (errno ==EAGAIN && retries--) { +usleep (1000 ); goto again; + }FAIL_ERRNO ("%s: read" , log_prefix); }if (n ==0 ) --2.30.2
This patch was applied to bpf/bpf-next.git (refs/heads/master):
Hello:
This patch was applied to bpf/bpf-next.git (refs/heads/master):
Quoted 6 lines · On Thu, 19 Aug 2021 09:36:09 -0700 you wrote:
This patch adds a 1ms delay to reduce flakyness of the test.
Signed-off-by Yucong Sun <[email protected]>--- .../selftests/bpf/prog_tests/sockmap_listen.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-)
Here is the summary with links:
- [bpf-next] selftests/bpf: adding delay in socketmap_listen to reduce flakyness
https://git.kernel.org/bpf/bpf-next/c/3666b167ea68
You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html