Hi!
> Hi,
>
> function "void tst_mkfs(...)" is used by such flollowing test cases:
>
> testcases/kernel/syscalls/../
> creat06
> fchown04
> fchown04_16
> inotify_03
> ichown03
> ichown03_16
> link08
> linkat02
> mkdir03
> mkdirat02
> mknod07
> mknodat02
> mount01
> mount06
> rename11
> reanameat01
> rmdir02
> umount01
>
> And all of these test cases " tst_sig(NOFORK, DEF_HANDLER, cleanup);" before
> calling tst_mkfs.
> So when the tst_mkfs calls "if (system("mkfs.btrfs 2>&1 | grep '\\-f '
> >/dev/null") == 0) ", the
> SIGCHLD will be caught by parent process, and all of the tests will fail.
> In my opinion, change the tst_mkfs is the better way.
Hmm, we have that trick allready in tst_run_cmd() but that cannot be
used because it fails on non-zero exit value.
What I would like to avoid is spreading signal() calls around the
library code. So the best solution may be to add a tst_system() to
tst_run_cmd.c which would behave like system() but will also swap
the signal handlers (and will include the very same comment explaining
why we do that as we have in the tst_run_cmd.c). Then we can use the
tst_system() internally in the test library.
--
Cyril Hrubis
[email protected]
------------------------------------------------------------------------------
Comprehensive Server Monitoring with Site24x7.
Monitor 10 servers for $9/Month.
Get alerted through email, SMS, voice calls or mobile push notifications.
Take corrective actions from your mobile device.
http://pubads.g.doubleclick.net/gampad/clk?id=154624111&iu=/4140/ostg.clktrk
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list
