kpageflags: fix wrong KPF_THP on non-pmd-mappable compound pages

[PATCH 1/2] mm: Constify folio_order()/folio_test_pmd_mappable()

Posted by ran xiaokai 1 year, 5 months ago

From: Ran Xiaokai <ran.xiaokai@zte.com.cn>

Constify folio_order()/folio_test_pmd_mappable().
No functional changes, just a preparation for the next patch.

Signed-off-by: Ran Xiaokai <ran.xiaokai@zte.com.cn>
---
 include/linux/huge_mm.h | 2 +-
 include/linux/mm.h      | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h
index 2aa986a5cd1b..8d66e4eaa1bc 100644
--- a/include/linux/huge_mm.h
+++ b/include/linux/huge_mm.h
@@ -377,7 +377,7 @@ static inline spinlock_t *pud_trans_huge_lock(pud_t *pud,
  * folio_test_pmd_mappable - Can we map this folio with a PMD?
  * @folio: The folio to test
  */
-static inline bool folio_test_pmd_mappable(struct folio *folio)
+static inline bool folio_test_pmd_mappable(const struct folio *folio)
 {
 	return folio_order(folio) >= HPAGE_PMD_ORDER;
 }
diff --git a/include/linux/mm.h b/include/linux/mm.h
index 9a5652c5fadd..b1c11371a2a3 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -1105,7 +1105,7 @@ static inline unsigned int compound_order(struct page *page)
  *
  * Return: The order of the folio.
  */
-static inline unsigned int folio_order(struct folio *folio)
+static inline unsigned int folio_order(const struct folio *folio)
 {
 	if (!folio_test_large(folio))
 		return 0;
-- 
2.15.2

Re: [PATCH 1/2] mm: Constify folio_order()/folio_test_pmd_mappable()

Posted by Zi Yan 1 year, 5 months ago

On Tue Jun 25, 2024 at 10:49 PM EDT, ran xiaokai wrote:
> From: Ran Xiaokai <ran.xiaokai@zte.com.cn>
>
> Constify folio_order()/folio_test_pmd_mappable().
> No functional changes, just a preparation for the next patch.

What warning/error are you seeing when you just apply patch 2? I wonder why it
did not show up in other places. Thanks.

>
> Signed-off-by: Ran Xiaokai <ran.xiaokai@zte.com.cn>
> ---
>  include/linux/huge_mm.h | 2 +-
>  include/linux/mm.h      | 2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h
> index 2aa986a5cd1b..8d66e4eaa1bc 100644
> --- a/include/linux/huge_mm.h
> +++ b/include/linux/huge_mm.h
> @@ -377,7 +377,7 @@ static inline spinlock_t *pud_trans_huge_lock(pud_t *pud,
>   * folio_test_pmd_mappable - Can we map this folio with a PMD?
>   * @folio: The folio to test
>   */
> -static inline bool folio_test_pmd_mappable(struct folio *folio)
> +static inline bool folio_test_pmd_mappable(const struct folio *folio)
>  {
>  	return folio_order(folio) >= HPAGE_PMD_ORDER;
>  }
> diff --git a/include/linux/mm.h b/include/linux/mm.h
> index 9a5652c5fadd..b1c11371a2a3 100644
> --- a/include/linux/mm.h
> +++ b/include/linux/mm.h
> @@ -1105,7 +1105,7 @@ static inline unsigned int compound_order(struct page *page)
>   *
>   * Return: The order of the folio.
>   */
> -static inline unsigned int folio_order(struct folio *folio)
> +static inline unsigned int folio_order(const struct folio *folio)
>  {
>  	if (!folio_test_large(folio))
>  		return 0;




-- 
Best Regards,
Yan, Zi

Re: [PATCH 1/2] mm: Constify folio_order()/folio_test_pmd_mappable()

Posted by ran xiaokai 1 year, 5 months ago

> On Tue Jun 25, 2024 at 10:49 PM EDT, ran xiaokai wrote:
> > From: Ran Xiaokai <ran.xiaokai@zte.com.cn>
> >
> > Constify folio_order()/folio_test_pmd_mappable().
> > No functional changes, just a preparation for the next patch.
> 
> What warning/error are you seeing when you just apply patch 2? I wonder why it
> did not show up in other places. Thanks.

fs/proc/page.c: In function 'stable_page_flags':
fs/proc/page.c:152:35: warning: passing argument 1 of 'folio_test_pmd_mappable' discards 'const' qualifier from pointer target type [-Wdiscarded-qualifiers]
  152 |  else if (folio_test_pmd_mappable(folio)) {
      |                                   ^~~~~
In file included from include/linux/mm.h:1115,
                 from include/linux/memblock.h:12,
                 from fs/proc/page.c:2:
include/linux/huge_mm.h:380:58: note: expected 'struct folio *' but argument is of type 'const struct folio *'
  380 | static inline bool folio_test_pmd_mappable(struct folio *folio)

u64 stable_page_flags(const struct page *page)
{
	const struct folio *folio; // the const definition causes the warning
	...
}

As almost all the folio_test_XXX(flags) have converted to received
a const parameter, it is Ok to also do this for folio_order()?

> >
> > Signed-off-by: Ran Xiaokai <ran.xiaokai@zte.com.cn>
> > ---
> >  include/linux/huge_mm.h | 2 +-
> >  include/linux/mm.h      | 2 +-
> >  2 files changed, 2 insertions(+), 2 deletions(-)
> >
> > diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h
> > index 2aa986a5cd1b..8d66e4eaa1bc 100644
> > --- a/include/linux/huge_mm.h
> > +++ b/include/linux/huge_mm.h
> > @@ -377,7 +377,7 @@ static inline spinlock_t *pud_trans_huge_lock(pud_t *pud,
> >   * folio_test_pmd_mappable - Can we map this folio with a PMD?
> >   * @folio: The folio to test
> >   */
> > -static inline bool folio_test_pmd_mappable(struct folio *folio)
> > +static inline bool folio_test_pmd_mappable(const struct folio *folio)
> >  {
> >  	return folio_order(folio) >= HPAGE_PMD_ORDER;
> >  }
> > diff --git a/include/linux/mm.h b/include/linux/mm.h
> > index 9a5652c5fadd..b1c11371a2a3 100644
> > --- a/include/linux/mm.h
> > +++ b/include/linux/mm.h
> > @@ -1105,7 +1105,7 @@ static inline unsigned int compound_order(struct page *page)
> >   *
> >   * Return: The order of the folio.
> >   */
> > -static inline unsigned int folio_order(struct folio *folio)
> > +static inline unsigned int folio_order(const struct folio *folio)
> >  {
> >  	if (!folio_test_large(folio))
> >  		return 0;

Re: [PATCH 1/2] mm: Constify folio_order()/folio_test_pmd_mappable()

Posted by Zi Yan 1 year, 5 months ago

On Wed Jun 26, 2024 at 12:30 AM EDT, ran xiaokai wrote:
> > On Tue Jun 25, 2024 at 10:49 PM EDT, ran xiaokai wrote:
> > > From: Ran Xiaokai <ran.xiaokai@zte.com.cn>
> > >
> > > Constify folio_order()/folio_test_pmd_mappable().
> > > No functional changes, just a preparation for the next patch.
> > 
> > What warning/error are you seeing when you just apply patch 2? I wonder why it
> > did not show up in other places. Thanks.
>
> fs/proc/page.c: In function 'stable_page_flags':
> fs/proc/page.c:152:35: warning: passing argument 1 of 'folio_test_pmd_mappable' discards 'const' qualifier from pointer target type [-Wdiscarded-qualifiers]
>   152 |  else if (folio_test_pmd_mappable(folio)) {
>       |                                   ^~~~~
> In file included from include/linux/mm.h:1115,
>                  from include/linux/memblock.h:12,
>                  from fs/proc/page.c:2:
> include/linux/huge_mm.h:380:58: note: expected 'struct folio *' but argument is of type 'const struct folio *'
>   380 | static inline bool folio_test_pmd_mappable(struct folio *folio)
>
> u64 stable_page_flags(const struct page *page)
> {
> 	const struct folio *folio; // the const definition causes the warning
> 	...

Please include the warning in the commit log to explain the change.

> }
>
> As almost all the folio_test_XXX(flags) have converted to received
> a const parameter, it is Ok to also do this for folio_order()?

Yes.