Re: [PATCH v2] mm/damon/vaddr: attempt per-vma lock during page table walk

[SeongJae Park <sj@kernel.org>]

On Tue, 12 May 2026 23:15:23 +0800 Kefeng Wang <wangkefeng.wang@huawei.com> wrote:

> Currently, DAMON virtual address operations use mmap_read_lock
> during page table walks, which can cause unnecessary contention
> under high concurrency.
> 
> Introduce damon_va_walk_page_range() to first attempt acquiring a
> per-vma lock. If the VMA is found and the range is fully contained
> within it, the page table walk proceeds with the per-vma lock
> instead of mmap_read_lock.
> 
> This optimization is particularly effective for damon_va_young()
> and damon_va_mkold(), which are frequently called and typically
> operate within a single VMA.

Maybe this is not only making DAMON faster, but avoid DAMON blocking others.

And because we don't have performance measurements, how about making this
sentence less assertive?  E.g., "This optimization is particularly expected to
be effective for ..."

I guess Andrew could make the adjustment when he picks this into mm.git.  If
not I will repost this with the trivial change if you don't mind.

> 
> Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>

Reviewed-by: SeongJae Park <sj@kernel.org>

> ---
> v2: avoid handling VMAs with the VM_PFNMAP for per-vma path, found by
>     Sashiko review.

This patch helped Sashiko finding [1] yet another DAMON bug (impact is minor
imho).  The bug is orthogonal to this patch so there is no blocker for this
patch.

> v1: https://lore.kernel.org/linux-mm/20260511132546.1973270-1-wangkefeng.wang@huawei.com/

[1] https://lore.kernel.org/20260514015053.149396-1-sj@kernel.org


Thanks,
SJ

[...]