lkml.org 
[lkml]   [2020]   [May]   [26]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
SubjectRe: [PATCH 0/3] make vm_committed_as_batch aware of vm overcommit policy
On Tue, May 26, 2020 at 06:14:13PM -0700, Andi Kleen wrote:
> On Tue, May 26, 2020 at 02:14:59PM -0400, Qian Cai wrote:
> > On Thu, May 21, 2020 at 05:27:26PM -0400, Qian Cai wrote:
> > > On Fri, May 08, 2020 at 03:25:14PM +0800, Feng Tang wrote:
> > > > When checking a performance change for will-it-scale scalability
> > > > mmap test [1], we found very high lock contention for spinlock of
> > > > percpu counter 'vm_committed_as':
> > > >
> > > > 94.14% 0.35% [kernel.kallsyms] [k] _raw_spin_lock_irqsave
> > > > 48.21% _raw_spin_lock_irqsave;percpu_counter_add_batch;__vm_enough_memory;mmap_region;do_mmap;
> > > > 45.91% _raw_spin_lock_irqsave;percpu_counter_add_batch;__do_munmap;
> > > >
> > > > Actually this heavy lock contention is not always necessary. The
> > > > 'vm_committed_as' needs to be very precise when the strict
> > > > OVERCOMMIT_NEVER policy is set, which requires a rather small batch
> > > > number for the percpu counter.
> > > >
> > > > So lift the batch number to 16X for OVERCOMMIT_ALWAYS and
> > > > OVERCOMMIT_GUESS policies, and add a sysctl handler to adjust it
> > > > when the policy is reconfigured.
> > > >
> > > > Benchmark with the same testcase in [1] shows 53% improvement on a
> > > > 8C/16T desktop, and 2097%(20X) on a 4S/72C/144T server. And no change
> > > > for some platforms, due to the test mmap size of the case is bigger
> > > > than the batch number computed, though the patch will help mmap/munmap
> > > > generally.
> > > >
> > > > [1] https://lkml.org/lkml/2020/3/5/57
> > > >
> > > > There are some style complain from checkpatch for patch 3, as it
> > > > follows the similar format of sibling functions
> > > >
> > > > patch1: a cleanup for /proc/meminfo
> > > > patch2: a preparation patch which also improve the accuracy of
> > > > vm_memory_committed
> > > > patch3: the main change
> > > >
> > > > Please help to review, thanks!
> > >
> > > Reverted this series fixed a warning under memory pressue.
> >
> > Andrew, Stephen, can you drop this series?
>
> Hope you're happy now.
>
> The warning is just not needed for the always case. The whole point
> of the patch was to maintain these counters only approximately.
> We could of course use _sum or _sum_positive in the warning, but that would
> just be unnecessary overhead in this moderately hot path.
>
> So this patch should fix it: (untested)
>
> Possibly might need to move it even further down because the patch
> also affected the _GUESS case.

Thanks Andi for the patch. And I agree that it could be after _GUESS
case, as existig check for _GUESS is loose:

if (pages > totalram_pages() + total_swap_pages)
goto error;

Thanks,
Feng

>
>
> diff --git a/mm/util.c b/mm/util.c
> index 988d11e6c17c..fa78f90e29a1 100644
> --- a/mm/util.c
> +++ b/mm/util.c
> @@ -798,10 +798,6 @@ int __vm_enough_memory(struct mm_struct *mm, long pages, int cap_sys_admin)
> {
> long allowed;
>
> - VM_WARN_ONCE(percpu_counter_read(&vm_committed_as) <
> - -(s64)vm_committed_as_batch * num_online_cpus(),
> - "memory commitment underflow");
> -
> vm_acct_memory(pages);
>
> /*
> @@ -810,6 +806,10 @@ int __vm_enough_memory(struct mm_struct *mm, long pages, int cap_sys_admin)
> if (sysctl_overcommit_memory == OVERCOMMIT_ALWAYS)
> return 0;
>
> + VM_WARN_ONCE(percpu_counter_read(&vm_committed_as) <
> + -(s64)vm_committed_as_batch * num_online_cpus(),
> + "memory commitment underflow");
> +
> if (sysctl_overcommit_memory == OVERCOMMIT_GUESS) {
> if (pages > totalram_pages() + total_swap_pages)
> goto error;

\
 
 \ /
  Last update: 2020-05-27 03:44    [W:0.897 / U:0.036 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site