[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <b677e859-abad-46b7-8dce-59152d29ea59@redhat.com>
Date: Thu, 4 Sep 2025 08:33:26 +0200
From: David Hildenbrand <david@...hat.com>
To: Andrew Morton <akpm@...ux-foundation.org>,
Stephen Rothwell <sfr@...b.auug.org.au>
Cc: Max Kellermann <max.kellermann@...os.com>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Linux Next Mailing List <linux-next@...r.kernel.org>,
Lorenzo Stoakes <lorenzo.stoakes@...cle.com>,
Vlastimil Babka <vbabka@...e.cz>, Mike Rapoport <rppt@...nel.org>,
Matthew Wilcox <willy@...radead.org>
Subject: Re: linux-next: build warning after merge of the mm-unstable tree
On 04.09.25 04:59, Andrew Morton wrote:
> On Thu, 4 Sep 2025 12:14:38 +1000 Stephen Rothwell <sfr@...b.auug.org.au> wrote:
>
>> After merging the mm-unstable tree, today's linux-next build (arm
>> multi_v7_defconfig) produced this warning:
>>
>> In file included from include/linux/highmem.h:14,
>> from include/linux/bvec.h:10,
>> from include/linux/blk_types.h:10,
>> from include/linux/writeback.h:13,
>> from include/linux/memcontrol.h:23,
>> from include/linux/swap.h:9,
>> from include/linux/suspend.h:5,
>> from arch/arm/mach-highbank/pm.c:9:
>> include/linux/highmem-internal.h: In function 'kmap_local_folio':
>> include/linux/highmem-internal.h:86:46: warning: passing argument 1 of 'folio_page' discards 'const' qualifier from pointer target type [-Wdiscarded-qualifiers]
>> 86 | const struct page *page = folio_page(folio, offset / PAGE_SIZE);
>> | ^~~~~
>> In file included from include/linux/mmzone.h:23,
>> from include/linux/swap.h:7:
>> include/linux/page-flags.h:319:53: note: expected 'struct folio *' but argument is of type 'const struct folio *'
>> 319 | static inline struct page *folio_page(struct folio *folio, unsigned long n)
>
Hm, why did we not detect that in mm-new?
>
> Yeah, thanks, folio_page() is a problem. I'll do the below nasty for
> now, but nasty.
>
> --- a/include/linux/page-flags.h~mm-constify-highmem-related-functions-for-improved-const-correctness-fix
> +++ a/include/linux/page-flags.h
> @@ -316,9 +316,9 @@ static __always_inline unsigned long _co
> * check that the page number lies within @folio; the caller is presumed
> * to have a reference to the page.
> */
> -static inline struct page *folio_page(struct folio *folio, unsigned long n)
> +static inline struct page *folio_page(const struct folio *folio, unsigned long n)
> {
> - return &folio->page + n;
> + return (struct page *)(&folio->page + n);
> }
>
> static __always_inline int PageTail(const struct page *page)
> _
I think the problem appears when nth_page() removal is paired with const-correctness.
Not a fan of going back to a plain macro either. Likely we can handle it similar to page_folio() and
perform the cast cleanly:
diff --git a/include/linux/page-flags.h b/include/linux/page-flags.h
index faf17ca211b4f..dc67fe08e77c8 100644
--- a/include/linux/page-flags.h
+++ b/include/linux/page-flags.h
@@ -307,6 +307,12 @@ static __always_inline unsigned long _compound_head(const struct page *page)
const struct page *: (const struct folio *)_compound_head(p), \
struct page *: (struct folio *)_compound_head(p)))
+static inline const struct page *_folio_page(const struct folio *folio,
+ unsigned long n)
+{
+ return &folio->page + n;
+}
+
/**
* folio_page - Return a page from a folio.
* @folio: The folio.
@@ -316,10 +322,9 @@ static __always_inline unsigned long _compound_head(const struct page *page)
* check that the page number lies within @folio; the caller is presumed
* to have a reference to the page.
*/
-static inline struct page *folio_page(struct folio *folio, unsigned long n)
-{
- return &folio->page + n;
-}
+#define folio_page(f, p) (_Generic((f), \
+ const struct folio *: (const struct page *)_folio_page(f, p), \
+ struct folio *: (struct page *)_folio_page(f, p)))
static __always_inline int PageTail(const struct page *page)
{
--
Cheers
David / dhildenb
Powered by blists - more mailing lists