Re: [PATCH] extend print_fatal_signals for reached RLIMIT_SIGPENDING

[Ingo Molnar <mingo@elte.hu>]

* Naohiro Ooiwa <nooiwa@miraclelinux.com> wrote:

> When the system has too many timers or too many aggregate
> queued signals, the EAGAIN error is returned to application
> from kernel, including timer_create().
> It means that exceeded limit of pending signals at all.
> But we can't imagine it.
> 
> This patch show the message when reached limit of pending signals
> and enabled print_fatal_signals.
> If you see this message and your system behaved unexpectedly,
> you can run following command.
>    # ulimit -i unlimited
> 
> With help from Hiroshi Shimamoto <h-shimamoto@ct.jp.nec.com>.
> 
> 
> Signed-off-by: Naohiro Ooiwa <nooiwa@miraclelinux.com>
> Acked-by: Ingo Molnar <mingo@elte.hu>
> ---
>  Documentation/kernel-parameters.txt |   11 +++++++++--
>  kernel/signal.c                     |   21 ++++++++++++++++++---
>  2 files changed, 27 insertions(+), 5 deletions(-)

Thanks, i've applied your patch to tip:core/signal, for v2.6.33 merge 
(if it passes all tests).

I made a few (very small) changes, see the -tip commit notification 
email in this thread with the final commit:

 - Extended the functions so that we can print which precise signal got 
   dropped - app writers will likely want to know that

 - Changed the message to:

        task/1234: reached RLIMIT_SIGPENDING, dropping signal

   which is slightly more informative.

 - Cleaned up small cleanliness details in surrounding code that caught 
   my eyes.

 - Changed a few variable and function names to be a tiny bit more 
   expressive.

 - Pushed the print_fatal_printks check into the new utility function 
   (print_dropped_signal()), to not clutter __sigqueue_alloc() 
   needlessly.

 - Clarified the commit log message a bit, gave sample output of the new 
   behavior.

Thanks,

	Ingo