Skip to contentSkip to conversation

syscall: set FORK in tst_sig for some old tests

PATCH In review Li Wang 4 participants 10 messages Aug 13

lore
Share & tools
lore

Li Wang

author
#0 lore raw

PATCH 1/1 syscall: set FORK in tst_sig for some old tests

From: Li Wang <[email protected]>

Since commit d90d2004657b ("lib: Use backoff polling to wait for loop device nodes") use TST_RETRY_FN_EXP_BACKOFF(), which calls tst_multiply_timeout(). That in turn reaches tst_has_slow_kconfig() -> tst_kconfig_read() -> popen("zcat /proc/config.gz") to read the kernel config.

popen() forks a child, and that child exits the SIGCHLD is reported as an unexpected signal by the tst_sig(NOFORK) handler:

umount2_01 1 TBROK : tst_sig.c:232: unexpected signal SIGCHLD(17) received

Since these tests now genuinely fork (indirectly via popen), declaring FORK matches their actual behavior and lets the handler tolerate the child's SIGCHLD. This is a minimal fix that preserves the retry logic.

These are old-API tests that will eventually be converted to the new tst_test API and this issue disappears entirely. The one-line FORK change is a low-cost bridge until then.

Fixes d90d2004657b ("lib: Use backoff polling to wait for loop device nodes")

---

Notes:
This is a temporary solution for the regression, I will convert these old tests to the new API next few days.

 testcases/kernel/syscalls/linkat/linkat02.c     | 2 +-
 testcases/kernel/syscalls/mknodat/mknodat02.c   | 2 +-
 testcases/kernel/syscalls/rename/rename11.c     | 2 +-
 testcases/kernel/syscalls/renameat/renameat01.c | 2 +-
 testcases/kernel/syscalls/umount2/umount2_01.c  | 2 +-
 5 files changed, 5 insertions(+), 5 deletions(-)
testcases/kernel/syscalls/linkat/linkat02.c +1 -1
index 273587271..6dfa7cfb5 100644
@@ -140,7 +140,7 @@ static void setup(void)
 {
 	tst_require_root();
 
-	tst_sig(NOFORK, DEF_HANDLER, cleanup);
+	tst_sig(FORK, DEF_HANDLER, cleanup);
 
 	tst_tmpdir();
 
testcases/kernel/syscalls/mknodat/mknodat02.c +1 -1
index 2f7c56f00..89cfc5978 100644
@@ -107,7 +107,7 @@ static void setup(void)
 
 	tst_require_root();
 
-	tst_sig(NOFORK, DEF_HANDLER, cleanup);
+	tst_sig(FORK, DEF_HANDLER, cleanup);
 
 	tst_tmpdir();
 
testcases/kernel/syscalls/rename/rename11.c +1 -1
index 8ff9f7d38..26e324d64 100644
@@ -90,7 +90,7 @@ static void setup(void)
 {
 	int i;
 
-	tst_sig(NOFORK, DEF_HANDLER, cleanup);
+	tst_sig(FORK, DEF_HANDLER, cleanup);
 
 	tst_require_root();
 
testcases/kernel/syscalls/renameat/renameat01.c +1 -1
index 72e72b481..f9ed3a9a7 100644
@@ -134,7 +134,7 @@ static void setup(void)
 
 	tst_require_root();
 
-	tst_sig(NOFORK, DEF_HANDLER, cleanup);
+	tst_sig(FORK, DEF_HANDLER, cleanup);
 
 	tst_tmpdir();
 
testcases/kernel/syscalls/umount2/umount2_01.c +1 -1
index 274409ef3..c6496e306 100644
@@ -67,7 +67,7 @@ static void setup(void)
 {
 	tst_require_root();
 
-	tst_sig(NOFORK, DEF_HANDLER, NULL);
+	tst_sig(FORK, DEF_HANDLER, NULL);
 
 	tst_tmpdir();
 
Signature · 5 lines
2.55.0


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp
#1 lore raw

Hi Li,

Quoted 1 line · Li Wang

syscall: set FORK in tst_sig for some old tests

Verdict - Reviewed

---
Note:

The agent can sometimes produce false positives although often its findings are genuine. If you find issues with the review, please comment this email or ignore the suggestions.

Regards,
LTP AI Reviewer

Signature · 1 line
Mailing list info: https://lists.linux.it/listinfo/ltp
Reviewed-by Wake Liu on v1 1/1
#8 lore raw
In reply to Li Wang

Hi Li, Petr, Avinesh,

Thanks Li for the quick analysis and for proposing this patch. I agree that this approach is much better than reverting d90d2004657b.

Best regards,
Wake

Signature · 1 line
Mailing list info: https://lists.linux.it/listinfo/ltp
Reviewed-by Tested-by
#3 lore raw
In reply to Avinesh Kumar via ltp

Hi all,

Quoted 3 lines

Hi Li,

Thank you for addressing this.

Li, thanks for fixing it!

Quoted 4 lines

Petr,
I think we should use this solution instead of reverting the patch.

+1, thanks for a notification

Kind regards,
Petr

Quoted 2 lines

Regards,
Avinesh

Signature · 1 line
Mailing list info: https://lists.linux.it/listinfo/ltp
author
#6 lore raw

Hi Avinesh, Petr,

Quoted 6 lines
Quoted 1 line

Patch merged! Thanks~

Thanks for fixing it!

very nit: there could have been

That's true, apologies for my carelessness.

Signature · 5 lines
Regards,
Li Wang

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp
#7 lore raw

Hi!

Quoted 10 lines · Li Wang

Hi Avinesh, Petr,

Quoted 6 lines
Quoted 1 line

Patch merged! Thanks~

Thanks for fixing it!

very nit: there could have been

That's true, apologies for my carelessness.

No worries at all! I've already credited Claude in my private journal anyway ;)

Signature · 1 line
Mailing list info: https://lists.linux.it/listinfo/ltp