Re: ashmem_shrink with long term stable kernel [3.0.36]

Re: ashmem_shrink with long term stable kernel [3.0.36]

[Konstantin Khlebnikov]

Akhilesh Kumar wrote:
> Hi All,
>
> During mm performance testing sometimes we observed below kernel messages
>
> shrink_slab: ashmem_shrink+0x0/0x114 negative objects to delete nr=-2133936901
> shrink_slab: ashmem_shrink+0x0/0x114 negative objects to delete nr=-2139256767
> shrink_slab: ashmem_shrink+0x0/0x114 negative objects to delete nr=-2079333971
> shrink_slab: ashmem_shrink+0x0/0x114 negative objects to delete nr=-2096156269
> shrink_slab: ashmem_shrink+0x0/0x114 negative objects to delete nr=-20658392
>
>   After debugging is we fount below patch mm/vmscan
> http://git.kernel.org/?p=linux/kernel/git/stable/linux-stable.git;a=commitdiff;h=635697c663f38106063d5659f0cf2e45afcd4bb5
> Since patch fix critical issue and same is not integrated with long term stable kernel (3.0.36)
> and  we are using below patch with long term stable kernel (3.0.36) is there any side effects ?

Nothing special, your patch should work fine.

> @@ -248,10 +248,12 @@ unsigned long shrink_slab(struct shrink_control *shrink,
>
>          list_for_each_entry(shrinker, &shrinker_list, list) {
>                  unsigned long long delta;
> -               unsigned long total_scan;
> -               unsigned long max_pass;
> +               long total_scan;
> +               long max_pass;
>
>                  max_pass = do_shrinker_shrink(shrinker, shrink, 0);
> +               if (max_pass <= 0)
> +                       continue;
>                  delta = (4 * nr_pages_scanned) / shrinker->seeks;
>                  delta *= max_pass;
>                  do_div(delta, lru_pages + 1);
> --
> Please review and share ur comments.
> Thanks,
> Akhilesh