Re: [PATCH] perf: rework memory accounting in perf_mmap()

[Peter Zijlstra]

On Mon, Oct 07, 2019 at 04:31:37PM +0000, Song Liu wrote:
> Hi Peter,
> 
> > On Sep 30, 2019, at 2:02 AM, Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
> > 
> > On Wed, Sep 04, 2019 at 02:46:18PM -0700, Song Liu wrote:
> >> perf_mmap() always increases user->locked_vm. As a result, "extra" could
> >> grow bigger than "user_extra", which doesn't make sense. Here is an
> >> example case:
> >> 
> >> Note: Assume "user_lock_limit" is very small.
> >> | # of perf_mmap calls |vma->vm_mm->pinned_vm|user->locked_vm|
> >> | 0                    | 0                   | 0             |
> >> | 1                    | user_extra          | user_extra    |
> >> | 2                    | 3 * user_extra      | 2 * user_extra|
> >> | 3                    | 6 * user_extra      | 3 * user_extra|
> >> | 4                    | 10 * user_extra     | 4 * user_extra|
> >> 
> >> Fix this by maintaining proper user_extra and extra.
> > 
> > Aah, indeed.
> 
> Thanks for the feedback!
> 
> > 
> > Also, this code is unreadable (which is mostly my own fault I suppose)
> > :/
> 
> How does this patch look to you? Is it ready to merge?

Yes, I queued it, I'll get it into tip soon. It got held up due to me
working on some other patches, sorry about that.