Re: [RFC PATCH 0/3] irq: Fix stack overflow due to softirq called on current stack

[Benjamin Herrenschmidt <benh@kernel.crashing.org>]

On Thu, 2013-09-05 at 17:33 +0200, Frederic Weisbecker wrote:
> Hi,
> 
> This series is a proposition to fix the crash reported here: http://lkml.kernel.org/r/1378330796.4321.50.camel%40pasglop
> And it has the upside to also consolidate a bit the arch do_softirq overriden implementation.
> 
> Only tested in x86-64 for now.

Thanks. I'm off for a few days, Paul can you give that a spin ? My
understanding is that the crash was fairly reproduceable... which makes
me wonder what are the chances to get RH to pick that series up for
RHEL7...

Cheers,
Ben.

> Thanks.
> 
> Frederic Weisbecker (3):
>   irq: Consolidate do_softirq() arch overriden implementations
>   irq: Execute softirq on its own stack on irq exit
>   irq: Comment on the use of inline stack for ksoftirq
> 
>  arch/metag/kernel/irq.c    |   56 +++++++++++++++++------------------------
>  arch/parisc/kernel/irq.c   |   17 +-----------
>  arch/powerpc/kernel/irq.c  |   17 +-----------
>  arch/s390/kernel/irq.c     |   52 +++++++++++++++----------------------
>  arch/sh/kernel/irq.c       |   60 ++++++++++++++++++-------------------------
>  arch/sparc/kernel/irq_64.c |   31 +++++++---------------
>  arch/x86/kernel/irq_32.c   |   34 +++++++++----------------
>  arch/x86/kernel/irq_64.c   |   18 ++-----------
>  include/linux/interrupt.h  |   11 ++++++++
>  kernel/softirq.c           |   10 +++----
>  10 files changed, 112 insertions(+), 194 deletions(-)
>