Mirsad Goran Todorovac <mirsad.todorovac(a)alu.unizg.hr> wrote:
> Hi all,
>
> The `make kselftest` hangs in netfilter/nft_trans_stress.sh test when testing 6.2.0-rc2
> (attempted 2 times, waiting half an hour at least):
This script completes in 49 seconds for me.
> # selftests: netfilter: nft_trans_stress.sh
> # PASS: nft add/delete test returned 0
> # PASS: nft reload test returned 0
>
> The test script seems to be stuck in ping flood:
No, it gets stuck in the sub-test after 'PASS: nft reload test returned 0'
Can you send me the output of 'bash -x nft_trans_stress.sh'?
It should tell which command/program isn't making progress.
Hi all,
There is a minor issue that prevents self test net/af_unix to run on my platform:
# ./test_unix_oob
Connect failed: No such file or directory
Terminated
Tracing reveals that bind tried to open a shorter AF_UNIX socket address:
# strace -f ./test_unix_oob
.
.
.
socket(AF_UNIX, SOCK_STREAM, 0) = 3
getpid() = 453059
unlink("unix_oob_453059") = -1 ENOENT (No such file or directory)
bind(3, {sa_family=AF_UNIX, sun_path="unix_oob_453059"}, 110) = 0
pipe2([4, 5], 0) = 0
listen(3, 1) = 0
clone(child_stack=NULL, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0x7fa6a6577a10) = 453060
rt_sigaction(SIGURG, {sa_handler=0x5601e2d014c9, sa_mask=[], sa_flags=SA_RESTORER|SA_RESTART|SA_SIGINFO,
sa_restorer=0x7fa6a623bcf0}, NULL, 8) = 0
write(5, "S", 1) = 1
accept(3, strace: Process 453060 attached
<unfinished ...>
[pid 453060] set_robust_list(0x7fa6a6577a20, 24) = 0
[pid 453060] socket(AF_UNIX, SOCK_STREAM, 0) = 6
[pid 453060] read(4, "S", 5) = 1
[pid 453060] connect(6, {sa_family=AF_UNIX, sun_path="unix_oob_45305"}, 16) = -1 ENOENT (No such file or directory)
.
.
.
NOTE: bind used UNIX_AF addr "unix_oob_453059", while producer tries to connect to "unix_oob_45305".
When pids were up to 5 digits it probably did not manifest, but logically the size of the
consumer_addr is sizeof(struct sockaddr_un).
Please find the patch attached:
Thanks,
Mirsad
------------------------------------------------------------------------------------------------
diff --git a/tools/testing/selftests/net/af_unix/test_unix_oob.c b/tools/testing/selftests/net/af_unix/test_unix_oob.c
index b57e91e1c3f2..7ea733239cd9 100644
--- a/tools/testing/selftests/net/af_unix/test_unix_oob.c
+++ b/tools/testing/selftests/net/af_unix/test_unix_oob.c
@@ -124,7 +124,7 @@ void producer(struct sockaddr_un *consumer_addr)
wait_for_signal(pipefd[0]);
if (connect(cfd, (struct sockaddr *)consumer_addr,
- sizeof(struct sockaddr)) != 0) {
+ sizeof(struct sockaddr_un)) != 0) {
perror("Connect failed");
kill(0, SIGTERM);
exit(1);
--
Mirsad Goran Todorovac
Sistem inženjer
Grafički fakultet | Akademija likovnih umjetnosti
Sveučilište u Zagrebu
--
System engineer
Faculty of Graphic Arts | Academy of Fine Arts
University of Zagreb, Republic of Croatia
The European Union
This series provides a bunch of quick updates which should make the
coverage from pcm-test a bit more useful, it adds some support for
skipping tests when the hardware/driver is unable to support the
requested configuration, support for providing user visible descriptions
and then expands the set of cases we cover to include more sample rates
and channel counts. This should exercise switching between 8kHz and
44.1kHz based rates and ensure that clocking doesn't get confused by
non-stereo channel counts, both of which are I expect common real world
errors, at least for embedded cards.
v4:
- Rebase onto v6.2-rc1.
v3:
- "Rebase" onto Takashi's current tree (with a revert).
- Include Jaroslav's changes to specify all tests in the configuration
file parsing.
- Add a new "description" field to the configuration instead of trying
to name the tests.
- Always run both default and per-system tests, logging our success at
setting the per-system configurations as a separate test since they
shouldn't fail.
v2:
- Rebase onto Takashi's current tree.
- Tweak the buffer sizes for the newly added cases, don't be quite
so ambitious in how big a buffer we request for 96kHz and don't
go quite so small for 8kHz since some devices start hitting lower
limits on period size and struggle to deliver accurate timing.
To: Takashi Iwai <tiwai(a)suse.com>
To: Jaroslav Kysela <perex(a)perex.cz>
To: Shuah Khan <shuah(a)kernel.org>
Cc: alsa-devel(a)alsa-project.org
Cc: linux-kselftest(a)vger.kernel.org
Cc: linux-kernel(a)vger.kernel.org
Signed-off-by: Mark Brown <broonie(a)kernel.org>
---
Jaroslav Kysela (1):
kselftest/alsa: pcm - move more configuration to configuration files
Mark Brown (6):
kselftest/alsa: pcm - Drop recent coverage improvement changes
kselftest/alsa: pcm - Always run the default set of tests
kselftest/alsa: pcm - skip tests when we fail to set params
kselftest/alsa: pcm - Support optional description for tests
kselftest/alsa: pcm - Provide descriptions for the default tests
kselftest/alsa: pcm - Add more coverage by default
tools/testing/selftests/alsa/Makefile | 2 +-
tools/testing/selftests/alsa/alsa-local.h | 3 +
tools/testing/selftests/alsa/conf.c | 26 ++-
.../alsa/conf.d/Lenovo_ThinkPad_P1_Gen2.conf | 43 +++--
tools/testing/selftests/alsa/pcm-test.c | 205 ++++++++++++++-------
tools/testing/selftests/alsa/pcm-test.conf | 63 +++++++
6 files changed, 250 insertions(+), 92 deletions(-)
---
base-commit: 1b929c02afd37871d5afb9d498426f83432e71c2
change-id: 20221208-alsa-pcm-test-hacks-f6c1aa76bd2c
Best regards,
--
Mark Brown <broonie(a)kernel.org>