mm/page_alloc: Fix memory accept before watermarks gets initialized

[PATCH] mm/page_alloc: Fix memory accept before watermarks gets initialized

Posted by Kirill A. Shutemov 11 months ago

Watermarks are initialized during the postcore initcall. Until then, all
watermarks are set to zero. This causes cond_accept_memory() to
incorrectly skip memory acceptance because a watermark of 0 is always
met.

To ensure progress, accept one MAX_ORDER page if the watermark is zero.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reported-and-tested-by: Farrah Chen <farrah.chen@intel.com>
---
 mm/page_alloc.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 579789600a3c..4fe93029bcb6 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -7004,7 +7004,7 @@ static inline bool has_unaccepted_memory(void)
 
 static bool cond_accept_memory(struct zone *zone, unsigned int order)
 {
-	long to_accept;
+	long to_accept, wmark;
 	bool ret = false;
 
 	if (!has_unaccepted_memory())
@@ -7013,8 +7013,18 @@ static bool cond_accept_memory(struct zone *zone, unsigned int order)
 	if (list_empty(&zone->unaccepted_pages))
 		return false;
 
+	wmark = promo_wmark_pages(zone);
+
+	/*
+	 * Watermarks have not been initialized yet.
+	 *
+	 * Accepting one MAX_ORDER page to ensure progress.
+	 */
+	if (!wmark)
+		return try_to_accept_memory_one(zone);
+
 	/* How much to accept to get to promo watermark? */
-	to_accept = promo_wmark_pages(zone) -
+	to_accept = wmark -
 		    (zone_page_state(zone, NR_FREE_PAGES) -
 		    __zone_watermark_unusable_free(zone, order, 0) -
 		    zone_page_state(zone, NR_UNACCEPTED));
-- 
2.47.2

Re: [PATCH] mm/page_alloc: Fix memory accept before watermarks gets initialized

Posted by Gupta, Pankaj 11 months ago

> Watermarks are initialized during the postcore initcall. Until then, all
> watermarks are set to zero. This causes cond_accept_memory() to
> incorrectly skip memory acceptance because a watermark of 0 is always
> met.
> 
> To ensure progress, accept one MAX_ORDER page if the watermark is zero.
> 
> Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
> Reported-and-tested-by: Farrah Chen <farrah.chen@intel.com>

Reviewed-by: Pankaj Gupta <pankaj.gupta@amd.com>

Also, did a basic boot test of SNP guest kernel 6.14-rc6 with the patch 
applied on top.

> ---
>   mm/page_alloc.c | 14 ++++++++++++--
>   1 file changed, 12 insertions(+), 2 deletions(-)
> 
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index 579789600a3c..4fe93029bcb6 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -7004,7 +7004,7 @@ static inline bool has_unaccepted_memory(void)
>   
>   static bool cond_accept_memory(struct zone *zone, unsigned int order)
>   {
> -	long to_accept;
> +	long to_accept, wmark;
>   	bool ret = false;
>   
>   	if (!has_unaccepted_memory())
> @@ -7013,8 +7013,18 @@ static bool cond_accept_memory(struct zone *zone, unsigned int order)
>   	if (list_empty(&zone->unaccepted_pages))
>   		return false;
>   
> +	wmark = promo_wmark_pages(zone);
> +
> +	/*
> +	 * Watermarks have not been initialized yet.
> +	 *
> +	 * Accepting one MAX_ORDER page to ensure progress.
> +	 */
> +	if (!wmark)
> +		return try_to_accept_memory_one(zone);
> +
>   	/* How much to accept to get to promo watermark? */
> -	to_accept = promo_wmark_pages(zone) -
> +	to_accept = wmark -
>   		    (zone_page_state(zone, NR_FREE_PAGES) -
>   		    __zone_watermark_unusable_free(zone, order, 0) -
>   		    zone_page_state(zone, NR_UNACCEPTED));

Re: [PATCH] mm/page_alloc: Fix memory accept before watermarks gets initialized

Posted by Vlastimil Babka 11 months ago

On 3/10/25 09:28, Kirill A. Shutemov wrote:
> Watermarks are initialized during the postcore initcall. Until then, all
> watermarks are set to zero. This causes cond_accept_memory() to
> incorrectly skip memory acceptance because a watermark of 0 is always
> met.
> 
> To ensure progress, accept one MAX_ORDER page if the watermark is zero.
> 
> Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
> Reported-and-tested-by: Farrah Chen <farrah.chen@intel.com>

Acked-by: Vlastimil Babka <vbabka@suse.cz>

> ---
>  mm/page_alloc.c | 14 ++++++++++++--
>  1 file changed, 12 insertions(+), 2 deletions(-)
> 
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index 579789600a3c..4fe93029bcb6 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -7004,7 +7004,7 @@ static inline bool has_unaccepted_memory(void)
>  
>  static bool cond_accept_memory(struct zone *zone, unsigned int order)
>  {
> -	long to_accept;
> +	long to_accept, wmark;
>  	bool ret = false;
>  
>  	if (!has_unaccepted_memory())
> @@ -7013,8 +7013,18 @@ static bool cond_accept_memory(struct zone *zone, unsigned int order)
>  	if (list_empty(&zone->unaccepted_pages))
>  		return false;
>  
> +	wmark = promo_wmark_pages(zone);
> +
> +	/*
> +	 * Watermarks have not been initialized yet.
> +	 *
> +	 * Accepting one MAX_ORDER page to ensure progress.
> +	 */
> +	if (!wmark)
> +		return try_to_accept_memory_one(zone);
> +
>  	/* How much to accept to get to promo watermark? */
> -	to_accept = promo_wmark_pages(zone) -
> +	to_accept = wmark -
>  		    (zone_page_state(zone, NR_FREE_PAGES) -
>  		    __zone_watermark_unusable_free(zone, order, 0) -
>  		    zone_page_state(zone, NR_UNACCEPTED));

Re: [PATCH] mm/page_alloc: Fix memory accept before watermarks gets initialized

Posted by Vlastimil Babka 11 months ago

On 3/10/25 09:28, Kirill A. Shutemov wrote:
> Watermarks are initialized during the postcore initcall. Until then, all
> watermarks are set to zero. This causes cond_accept_memory() to
> incorrectly skip memory acceptance because a watermark of 0 is always
> met.

What are the user-visible consequences of that?

> To ensure progress, accept one MAX_ORDER page if the watermark is zero.
> 
> Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
> Reported-and-tested-by: Farrah Chen <farrah.chen@intel.com>

Fixes:, Cc: stable etc?

> ---
>  mm/page_alloc.c | 14 ++++++++++++--
>  1 file changed, 12 insertions(+), 2 deletions(-)
> 
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index 579789600a3c..4fe93029bcb6 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -7004,7 +7004,7 @@ static inline bool has_unaccepted_memory(void)
>  
>  static bool cond_accept_memory(struct zone *zone, unsigned int order)
>  {
> -	long to_accept;
> +	long to_accept, wmark;
>  	bool ret = false;
>  
>  	if (!has_unaccepted_memory())
> @@ -7013,8 +7013,18 @@ static bool cond_accept_memory(struct zone *zone, unsigned int order)
>  	if (list_empty(&zone->unaccepted_pages))
>  		return false;
>  
> +	wmark = promo_wmark_pages(zone);
> +
> +	/*
> +	 * Watermarks have not been initialized yet.
> +	 *
> +	 * Accepting one MAX_ORDER page to ensure progress.
> +	 */
> +	if (!wmark)
> +		return try_to_accept_memory_one(zone);
> +
>  	/* How much to accept to get to promo watermark? */
> -	to_accept = promo_wmark_pages(zone) -
> +	to_accept = wmark -
>  		    (zone_page_state(zone, NR_FREE_PAGES) -
>  		    __zone_watermark_unusable_free(zone, order, 0) -
>  		    zone_page_state(zone, NR_UNACCEPTED));

Re: [PATCH] mm/page_alloc: Fix memory accept before watermarks gets initialized

Posted by Kirill A. Shutemov 11 months ago

On Mon, Mar 10, 2025 at 12:37:25PM +0100, Vlastimil Babka wrote:
> On 3/10/25 09:28, Kirill A. Shutemov wrote:
> > Watermarks are initialized during the postcore initcall. Until then, all
> > watermarks are set to zero. This causes cond_accept_memory() to
> > incorrectly skip memory acceptance because a watermark of 0 is always
> > met.
> 
> What are the user-visible consequences of that?

Premature OOM on boot.

It can be triggered with certain combinations of number of vCPUs and
memory size.

> > To ensure progress, accept one MAX_ORDER page if the watermark is zero.
> > 
> > Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
> > Reported-and-tested-by: Farrah Chen <farrah.chen@intel.com>
> 
> Fixes:, Cc: stable etc?

Fixes: dcdfdd40fa82 ("mm: Add support for unaccepted memory")
Cc: stable@@vger.kernel.org # v6.5+

-- 
  Kiryl Shutsemau / Kirill A. Shutemov