[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <aDRnE8B27VgCoJ3N@localhost.localdomain>
Date: Mon, 26 May 2025 15:05:23 +0200
From: Oscar Salvador <osalvador@...e.de>
To: David Hildenbrand <david@...hat.com>
Cc: yangge1116@....com, akpm@...ux-foundation.org, linux-mm@...ck.org,
linux-kernel@...r.kernel.org, 21cnbao@...il.com,
baolin.wang@...ux.alibaba.com, muchun.song@...ux.dev,
liuzixing@...on.cn
Subject: Re: [PATCH] mm/hugetlb: remove redundant folio_test_hugetlb
On Mon, May 26, 2025 at 02:23:58PM +0200, David Hildenbrand wrote:
> Already in 6.15, gah.
>
> Please convert that code to never ever take any hugeglb locks unless we are
> clearly dealing with a hugetlb folios.
I guess we could just do?
diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index 20f08de9e37d..8bd8b950e4cb 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -2911,7 +2914,7 @@ int replace_free_hugepage_folios(unsigned long start_pfn, unsigned long end_pfn)
while (start_pfn < end_pfn) {
folio = pfn_folio(start_pfn);
- if (!folio_ref_count(folio)) {
+ if (!folio_ref_count(folio) && folio_test_hugetlb(folio)) {
ret = alloc_and_dissolve_hugetlb_folio(folio,
&isolate_list);
if (ret)
--
Oscar Salvador
SUSE Labs
Powered by blists - more mailing lists