[PATCH v3] USE_SEIZE: fix detaching from stopped processes

[Dmitry V. Levin]

On Wed, Jun 19, 2013 at 08:17:13PM +0400, Dmitry V. Levin wrote:
> On Wed, Jun 19, 2013 at 03:52:43PM +0200, Denys Vlasenko wrote:
> > V3: split SEIZE/!SEIZE code paths to reduce confusion.
> 
> With your tests/detach* enhancements, I see persistent detach-stopped
> failures on old (!use_seize) kernels:
> 
> $ cat detach-stopped.log 
> FAIL: detach-stopped (exit: 1)
> ==============================
> 
> Process 23456 attached
> --- SIGSTOP {si_signo=SIGSTOP, si_code=SI_USER, si_pid=0, si_uid=0} ---
> --- stopped by SIGSTOP ---
> restart_syscall(<... resuming interrupted call ...>Process 23456 detached
>  <detached ...>
> State:	S (sleeping)
> detach-stopped: failed test: tracee is not group-stopped after detach
> 
> There is no such problem on recent kernels (tested on 3.9.6),
> both use_seize and !use_seize cases work.

Of course, it is not a regression after your patch, this is how things
work on those !use_seize kernels.  If there is no way to workaround this,
I suggest applying your patch.


-- 
ldv
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
URL: <http://lists.strace.io/pipermail/strace-devel/attachments/20130619/30fd4587/attachment.bin>