Re: [PATCH 5/9] mm: Turn folio_test_hugetlb into a PageType

From: Matthew Wilcox <willy@infradead.org>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: linux-mm@kvack.org, David Hildenbrand <david@redhat.com>,
	Vlastimil Babka <vbabka@suse.cz>,
	Miaohe Lin <linmiaohe@huawei.com>,
	Muchun Song <muchun.song@linux.dev>,
	Oscar Salvador <osalvador@suse.de>
Subject: Re: [PATCH 5/9] mm: Turn folio_test_hugetlb into a PageType
Date: Mon, 25 Mar 2024 15:14:53 +0000	[thread overview]
Message-ID: <ZgGU7VircjkC0BBY@casper.infradead.org> (raw)
In-Reply-To: <20240321142448.1645400-6-willy@infradead.org>

On Thu, Mar 21, 2024 at 02:24:43PM +0000, Matthew Wilcox (Oracle) wrote:
> The current folio_test_hugetlb() can be fooled by a concurrent folio split
> into returning true for a folio which has never belonged to hugetlbfs.
> This can't happen if the caller holds a refcount on it, but we have a
> few places (memory-failure, compaction, procfs) which do not and should
> not take a speculative reference.
> 
> Since hugetlb pages do not use individual page mapcounts (they are always
> fully mapped and use the entire_mapcount field to record the number
> of mappings), the PageType field is available now that page_mapcount()
> ignores the value in this field.

Update vmcoreinfo:

diff --git a/kernel/vmcore_info.c b/kernel/vmcore_info.c
index f95516cd45bb..41372f5d5c19 100644
--- a/kernel/vmcore_info.c
+++ b/kernel/vmcore_info.c
@@ -205,11 +205,10 @@ static int __init crash_save_vmcoreinfo_init(void)
 	VMCOREINFO_NUMBER(PG_head_mask);
 #define PAGE_BUDDY_MAPCOUNT_VALUE	(~PG_buddy)
 	VMCOREINFO_NUMBER(PAGE_BUDDY_MAPCOUNT_VALUE);
-#ifdef CONFIG_HUGETLB_PAGE
-	VMCOREINFO_NUMBER(PG_hugetlb);
+#define PAGE_HUGETLB_MAPCOUNT_VALUE(	(~PG_hugetlb)
+	VMCOREINFO_NUMBER(PAGE_HUGETLB_MAPCOUNT_VALUE);
 #define PAGE_OFFLINE_MAPCOUNT_VALUE	(~PG_offline)
 	VMCOREINFO_NUMBER(PAGE_OFFLINE_MAPCOUNT_VALUE);
-#endif
 
 #ifdef CONFIG_KALLSYMS
 	VMCOREINFO_SYMBOL(kallsyms_names);

