On Fri, Nov 01, 2013 at 03:54:25PM +0800, Yuanhan Liu wrote:
> --- a/mm/mmap.c
> +++ b/mm/mmap.c
> @@ -2955,15 +2955,15 @@ static void vm_lock_anon_vma(struct mm_struct *mm,
> struct anon_vma *anon_vma)
>* The LSB of head.next can't change from under us
>* because w
From: Ingo Molnar
This lock basically covers the insertion of avc into rb tree; turn rwsem
to rwlock_t.
With the former per anon_vma lock applied, the performance is nearly
doubled in some heavy fork workload of aim7 on some platforms.
aim7.2000.jobs-per-min
- -
2 matches
Mail list logo
