Re: [PATCH 0/2] sched: Preemption fixlet and cleanup

Re: [PATCH 0/2] sched: Preemption fixlet and cleanup

[Sedat Dilek]

Hi,

I am following the "frequent lockups in 3.18rc4" thread (see [0]).

So, people still work on it, and one fixlet was going towards the sched/x86.
Is there an update of your patchset [1] (especially of patch 2/2 [2]) around?

One good thing in the discussion of [0] would be a short summarize and
config-parameters to test with.

Thanks in advance.

Regards,
- Sedat -

[0] http://marc.info/?t=141600075000006&r=1&w=1
[1] http://marc.info/?l=linux-kernel&m=141861029214340&w=1
[2] http://marc.info/?t=141861036000004&r=1&w=1

P.S.: I have applied the 3 patches of Linux on top of Linux v3.18.1
and testing with that.

Linus Torvalds (3):
      x86: mm: move mmap_sem unlock from mm_fault_error() to caller
      x86: mm: consolidate VM_FAULT_RETRY handling
      x86: mm: fix VM_FAULT_RETRY handling

Re: [PATCH 0/2] sched: Preemption fixlet and cleanup

[Sedat Dilek]

On Thu, Dec 18, 2014 at 1:48 PM, Sedat Dilek <sedat.dilek@gmail.com> wrote:
> Hi,
>
> I am following the "frequent lockups in 3.18rc4" thread (see [0]).
>
> So, people still work on it, and one fixlet was going towards the sched/x86.
> Is there an update of your patchset [1] (especially of patch 2/2 [2]) around?
>
> One good thing in the discussion of [0] would be a short summarize and
> config-parameters to test with.
>
> Thanks in advance.
>
> Regards,
> - Sedat -
>
> [0] http://marc.info/?t=141600075000006&r=1&w=1
> [1] http://marc.info/?l=linux-kernel&m=141861029214340&w=1
> [2] http://marc.info/?t=141861036000004&r=1&w=1
>
> P.S.: I have applied the 3 patches of Linux on top of Linux v3.18.1
> and testing with that.
>
> Linus Torvalds (3):
>       x86: mm: move mmap_sem unlock from mm_fault_error() to caller
>       x86: mm: consolidate VM_FAULT_RETRY handling
>       x86: mm: fix VM_FAULT_RETRY handling

Hi Frederic,

OK, I have seen you pushed a fix called "sched: Fix missing preemption
check in cond_resched()" [1] in your linux-dynticks.git#sched/urgent
branch and Peter took it [2].

Can you please CC me if you have some material for testing.

Thanks.

Regards,
- Sedat -

[1] http://git.kernel.org/cgit/linux/kernel/git/frederic/linux-dynticks.git/commit/?h=sched/urgent&id=4ccbe99852951957419bc616f888e297a478e50b
[2] http://git.kernel.org/cgit/linux/kernel/git/peterz/queue.git/commit/?h=sched/core&id=15789208c0c1a032d7e5392692e25d8c15e501ba

[PATCH 0/2] sched: Preemption fixlet and cleanup

[Frederic Weisbecker]

Here is the reworked fix against the missing preemption check that Linus
reported + the need_resched() loop pulled up to __schedule() callers also
suggested by him.

Thanks.

Frederic Weisbecker (2):
  sched: Fix missing preemption check in cond_resched()
  sched: Pull resched loop to __schedule() callers

 kernel/sched/core.c | 51 ++++++++++++++++++++++++++-------------------------
 1 file changed, 26 insertions(+), 25 deletions(-)

-- 
2.1.3