[PATCH] mm/mempool: use static key for boot-time debug enablement

[PATCH] mm/mempool: use static key for boot-time debug enablement

[lirongqing]

From: Li RongQing <lirongqing@baidu.com>

Replace the #ifdef CONFIG_SLUB_DEBUG_ON conditional compilation with a
static key (mempool_debug_enabled). This allows enabling mempool debugging
at boot time via:

    mempool_debug

Instead of requiring CONFIG_SLUB_DEBUG_ON at compile time. Benefits:

- Debugging can be enabled without rebuilding the kernel
- Uses standard kernel static_key mechanism with minimal overhead

Suggested-by: Vlastimil Babka (SUSE) <vbabka@kernel.org>
Signed-off-by: Li RongQing <lirongqing@baidu.com>
Cc: Vlastimil Babka <vbabka@kernel.org>
Cc: Harry Yoo <harry@kernel.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Hao Li <hao.li@linux.dev>
Cc: Christoph Lameter <cl@gentwo.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Roman Gushchin <roman.gushchin@linux.dev>
---
 Documentation/admin-guide/kernel-parameters.txt |  5 ++++
 mm/mempool.c                                    | 32 ++++++++++++++++++-------
 2 files changed, 28 insertions(+), 9 deletions(-)

diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt
index 35ed9dc..5a070e6 100644
--- a/Documentation/admin-guide/kernel-parameters.txt
+++ b/Documentation/admin-guide/kernel-parameters.txt
@@ -3998,6 +3998,11 @@ Kernel parameters
 			Note that even when enabled, there are a few cases where
 			the feature is not effective.
 
+	mempool_debug	[MM]
+			Enable mempool debugging. This enables element
+			poison checking when freeing elements back to the
+			pool. Useful for debugging mempool corruption.
+
 	memtest=	[KNL,X86,ARM,M68K,PPC,RISCV,EARLY] Enable memtest
 			Format: <integer>
 			default : 0 <disable>
diff --git a/mm/mempool.c b/mm/mempool.c
index db23e0e..4f429a1 100644
--- a/mm/mempool.c
+++ b/mm/mempool.c
@@ -16,11 +16,28 @@
 #include <linux/export.h>
 #include <linux/mempool.h>
 #include <linux/writeback.h>
+#include <linux/static_key.h>
+#include <linux/init.h>
 #include "slab.h"
 
 static DECLARE_FAULT_ATTR(fail_mempool_alloc);
 static DECLARE_FAULT_ATTR(fail_mempool_alloc_bulk);
 
+/*
+ * Debugging support for mempool using static key.
+ *
+ * This allows enabling mempool debug at boot time via:
+ *   mempool_debug
+ */
+static DEFINE_STATIC_KEY_FALSE(mempool_debug_enabled);
+
+static int __init mempool_debug_setup(char *str)
+{
+	static_branch_enable(&mempool_debug_enabled);
+	return 0;
+}
+early_param("mempool_debug", mempool_debug_setup);
+
 static int __init mempool_faul_inject_init(void)
 {
 	int error;
@@ -37,7 +54,6 @@ static int __init mempool_faul_inject_init(void)
 }
 late_initcall(mempool_faul_inject_init);
 
-#ifdef CONFIG_SLUB_DEBUG_ON
 static void poison_error(struct mempool *pool, void *element, size_t size,
 			 size_t byte)
 {
@@ -73,6 +89,9 @@ static void __check_element(struct mempool *pool, void *element, size_t size)
 
 static void check_element(struct mempool *pool, void *element)
 {
+	if (!static_branch_unlikely(&mempool_debug_enabled))
+		return;
+
 	/* Skip checking: KASAN might save its metadata in the element. */
 	if (kasan_enabled())
 		return;
@@ -112,6 +131,9 @@ static void __poison_element(void *element, size_t size)
 
 static void poison_element(struct mempool *pool, void *element)
 {
+	if (!static_branch_unlikely(&mempool_debug_enabled))
+		return;
+
 	/* Skip poisoning: KASAN might save its metadata in the element. */
 	if (kasan_enabled())
 		return;
@@ -140,14 +162,6 @@ static void poison_element(struct mempool *pool, void *element)
 #endif
 	}
 }
-#else /* CONFIG_SLUB_DEBUG_ON */
-static inline void check_element(struct mempool *pool, void *element)
-{
-}
-static inline void poison_element(struct mempool *pool, void *element)
-{
-}
-#endif /* CONFIG_SLUB_DEBUG_ON */
 
 static __always_inline bool kasan_poison_element(struct mempool *pool,
 		void *element)
-- 
2.9.4

Re: [PATCH] mm/mempool: use static key for boot-time debug enablement

[Usama Arif]

On Wed, 27 May 2026 06:46:34 -0400 lirongqing <lirongqing@baidu.com> wrote:

> From: Li RongQing <lirongqing@baidu.com>
> 
> Replace the #ifdef CONFIG_SLUB_DEBUG_ON conditional compilation with a
> static key (mempool_debug_enabled). This allows enabling mempool debugging
> at boot time via:
> 
>     mempool_debug
> 
> Instead of requiring CONFIG_SLUB_DEBUG_ON at compile time. Benefits:
> 
> - Debugging can be enabled without rebuilding the kernel
> - Uses standard kernel static_key mechanism with minimal overhead
> 
> Suggested-by: Vlastimil Babka (SUSE) <vbabka@kernel.org>
> Signed-off-by: Li RongQing <lirongqing@baidu.com>
> Cc: Vlastimil Babka <vbabka@kernel.org>
> Cc: Harry Yoo <harry@kernel.org>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Cc: Hao Li <hao.li@linux.dev>
> Cc: Christoph Lameter <cl@gentwo.org>
> Cc: David Rientjes <rientjes@google.com>
> Cc: Roman Gushchin <roman.gushchin@linux.dev>
> ---
>  Documentation/admin-guide/kernel-parameters.txt |  5 ++++
>  mm/mempool.c                                    | 32 ++++++++++++++++++-------
>  2 files changed, 28 insertions(+), 9 deletions(-)
> 
> diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt
> index 35ed9dc..5a070e6 100644
> --- a/Documentation/admin-guide/kernel-parameters.txt
> +++ b/Documentation/admin-guide/kernel-parameters.txt
> @@ -3998,6 +3998,11 @@ Kernel parameters
>  			Note that even when enabled, there are a few cases where
>  			the feature is not effective.
>  
> +	mempool_debug	[MM]
> +			Enable mempool debugging. This enables element
> +			poison checking when freeing elements back to the
> +			pool. Useful for debugging mempool corruption.
> +
>  	memtest=	[KNL,X86,ARM,M68K,PPC,RISCV,EARLY] Enable memtest
>  			Format: <integer>
>  			default : 0 <disable>
> diff --git a/mm/mempool.c b/mm/mempool.c
> index db23e0e..4f429a1 100644
> --- a/mm/mempool.c
> +++ b/mm/mempool.c
> @@ -16,11 +16,28 @@
>  #include <linux/export.h>
>  #include <linux/mempool.h>
>  #include <linux/writeback.h>
> +#include <linux/static_key.h>
> +#include <linux/init.h>
>  #include "slab.h"
>  
>  static DECLARE_FAULT_ATTR(fail_mempool_alloc);
>  static DECLARE_FAULT_ATTR(fail_mempool_alloc_bulk);
>  
> +/*
> + * Debugging support for mempool using static key.
> + *
> + * This allows enabling mempool debug at boot time via:
> + *   mempool_debug
> + */
> +static DEFINE_STATIC_KEY_FALSE(mempool_debug_enabled);
> +
> +static int __init mempool_debug_setup(char *str)
> +{
> +	static_branch_enable(&mempool_debug_enabled);
> +	return 0;
> +}
> +early_param("mempool_debug", mempool_debug_setup);
> +

Can static_branch_enable() in mempool_debug_setup() run before
jump_label_init() has set static_key_initialized?

Looking at start_kernel() in init/main.c:

	setup_arch(&command_line);
	mm_core_init_early();
	/* Static keys and static calls are needed by LSMs */
	jump_label_init();
	...
	/* parameters may set static keys */
	parse_early_param();

This will trigger the warning in include/linux/jump_label.h has:

	#define STATIC_KEY_CHECK_USE(key) WARN(!static_key_initialized, \
	    "%s(): static key '%pS' used before call to jump_label_init()", \
	    __func__, (key))


mm/dmapool.c registers an equivalent debug toggle via __setup()
rather than early_param():

	static int __init dmapool_debug_setup(char *str)
	{
		static_branch_enable(&dmapool_debug_enabled);
		return 1;
	}
	__setup("dmapool_debug", dmapool_debug_setup);

I think you can reuse that.

>  static int __init mempool_faul_inject_init(void)
>  {
>  	int error;
> @@ -37,7 +54,6 @@ static int __init mempool_faul_inject_init(void)
>  }
>  late_initcall(mempool_faul_inject_init);
>  
> -#ifdef CONFIG_SLUB_DEBUG_ON
>  static void poison_error(struct mempool *pool, void *element, size_t size,
>  			 size_t byte)
>  {
> @@ -73,6 +89,9 @@ static void __check_element(struct mempool *pool, void *element, size_t size)
>  
>  static void check_element(struct mempool *pool, void *element)
>  {
> +	if (!static_branch_unlikely(&mempool_debug_enabled))
> +		return;
> +
>  	/* Skip checking: KASAN might save its metadata in the element. */
>  	if (kasan_enabled())
>  		return;
> @@ -112,6 +131,9 @@ static void __poison_element(void *element, size_t size)
>  
>  static void poison_element(struct mempool *pool, void *element)
>  {
> +	if (!static_branch_unlikely(&mempool_debug_enabled))
> +		return;
> +

Before this change, building with CONFIG_SLUB_DEBUG_ON=y compiled in
check_element() and poison_element() unconditionally, so the
poisoning and corruption checks ran on every mempool free/alloc.
After this change those checks are gated on the mempool_debug boot
parameter even when CONFIG_SLUB_DEBUG_ON=y.

Existing users who relied on CONFIG_SLUB_DEBUG_ON=y giving them
mempool poison checking will silently lose it on upgrade unless they
also add "mempool_debug" to the command line.

Would it be worth defaulting the static key to true under
CONFIG_SLUB_DEBUG_ON=y, for example:

	#ifdef CONFIG_SLUB_DEBUG_ON
	static DEFINE_STATIC_KEY_TRUE(mempool_debug_enabled);
	#else
	static DEFINE_STATIC_KEY_FALSE(mempool_debug_enabled);
	#endif

so the previous default behaviour is preserved.


>  	/* Skip poisoning: KASAN might save its metadata in the element. */
>  	if (kasan_enabled())
>  		return;
> @@ -140,14 +162,6 @@ static void poison_element(struct mempool *pool, void *element)
>  #endif
>  	}
>  }
> -#else /* CONFIG_SLUB_DEBUG_ON */
> -static inline void check_element(struct mempool *pool, void *element)
> -{
> -}
> -static inline void poison_element(struct mempool *pool, void *element)
> -{
> -}
> -#endif /* CONFIG_SLUB_DEBUG_ON */
>  
>  static __always_inline bool kasan_poison_element(struct mempool *pool,
>  		void *element)
> -- 
> 2.9.4
> 
> 

Re: [PATCH] mm/mempool: use static key for boot-time debug enablement

[Vlastimil Babka (SUSE)]

On 5/27/26 15:03, Usama Arif wrote:
> On Wed, 27 May 2026 06:46:34 -0400 lirongqing <lirongqing@baidu.com> wrote:
>>  static int __init mempool_faul_inject_init(void)
>>  {
>>  	int error;
>> @@ -37,7 +54,6 @@ static int __init mempool_faul_inject_init(void)
>>  }
>>  late_initcall(mempool_faul_inject_init);
>>  
>> -#ifdef CONFIG_SLUB_DEBUG_ON
>>  static void poison_error(struct mempool *pool, void *element, size_t size,
>>  			 size_t byte)
>>  {
>> @@ -73,6 +89,9 @@ static void __check_element(struct mempool *pool, void *element, size_t size)
>>  
>>  static void check_element(struct mempool *pool, void *element)
>>  {
>> +	if (!static_branch_unlikely(&mempool_debug_enabled))
>> +		return;
>> +
>>  	/* Skip checking: KASAN might save its metadata in the element. */
>>  	if (kasan_enabled())
>>  		return;
>> @@ -112,6 +131,9 @@ static void __poison_element(void *element, size_t size)
>>  
>>  static void poison_element(struct mempool *pool, void *element)
>>  {
>> +	if (!static_branch_unlikely(&mempool_debug_enabled))
>> +		return;
>> +
> 
> Before this change, building with CONFIG_SLUB_DEBUG_ON=y compiled in
> check_element() and poison_element() unconditionally, so the
> poisoning and corruption checks ran on every mempool free/alloc.
> After this change those checks are gated on the mempool_debug boot
> parameter even when CONFIG_SLUB_DEBUG_ON=y.

The same would apply to the dmapool_debug [1]

https://lore.kernel.org/all/20260524034015.1830-1-lirongqing@baidu.com/

> Existing users who relied on CONFIG_SLUB_DEBUG_ON=y giving them
> mempool poison checking will silently lose it on upgrade unless they
> also add "mempool_debug" to the command line.
> 
> Would it be worth defaulting the static key to true under
> CONFIG_SLUB_DEBUG_ON=y, for example:
> 
> 	#ifdef CONFIG_SLUB_DEBUG_ON
> 	static DEFINE_STATIC_KEY_TRUE(mempool_debug_enabled);
> 	#else
> 	static DEFINE_STATIC_KEY_FALSE(mempool_debug_enabled);
> 	#endif
> 
> so the previous default behaviour is preserved.

There's DEFINE_STATIC_KEY_MAYBE for this.

But I think nobody will care really. My objection was that this and
dmapool_debug was tied to CONFIG_SLUB_DEBUG_ON in the first place, despite
not being part of slab. I'd rather disconnect it completely.

If testing bots relied on this, we could give them heads up to start using
those boot options. I assume they already use some for stuff that has no
CONFIG_ enablement.

>>  	/* Skip poisoning: KASAN might save its metadata in the element. */
>>  	if (kasan_enabled())
>>  		return;
>> @@ -140,14 +162,6 @@ static void poison_element(struct mempool *pool, void *element)
>>  #endif
>>  	}
>>  }
>> -#else /* CONFIG_SLUB_DEBUG_ON */
>> -static inline void check_element(struct mempool *pool, void *element)
>> -{
>> -}
>> -static inline void poison_element(struct mempool *pool, void *element)
>> -{
>> -}
>> -#endif /* CONFIG_SLUB_DEBUG_ON */
>>  
>>  static __always_inline bool kasan_poison_element(struct mempool *pool,
>>  		void *element)
>> -- 
>> 2.9.4
>> 
>> 

Re: [PATCH] mm/mempool: use static key for boot-time debug enablement

[Andrew Morton]

On Wed, 27 May 2026 06:46:34 -0400 lirongqing <lirongqing@baidu.com> wrote:

> Replace the #ifdef CONFIG_SLUB_DEBUG_ON conditional compilation with a
> static key (mempool_debug_enabled). This allows enabling mempool debugging
> at boot time via:
> 
>     mempool_debug
> 
> Instead of requiring CONFIG_SLUB_DEBUG_ON at compile time. Benefits:

Sashiko is suggesting that we use mempool_debug=<...> here.  Which permits
mempool_debug=n if for some reason the kernel is defaulting to "on".  Which
we might choose to do in the future.   I think that's a little better - do others agree?

Same goes for the new dmapool_debug.

Sashiko asked a second question:
	https://sashiko.dev/#/patchset/20260527104634.2434-1-lirongqing@baidu.com

Re: [PATCH] mm/mempool: use static key for boot-time debug enablement

[Christoph Lameter (Ampere)]

On Wed, 27 May 2026, lirongqing wrote:

> Replace the #ifdef CONFIG_SLUB_DEBUG_ON conditional compilation with a
> static key (mempool_debug_enabled). This allows enabling mempool debugging
> at boot time via:

I am a bit confused here. CONFIG_SLUB_DEBUG_ON sets debugging to be on by
default at compile time. If you do not use CONFIG_SLUB_DEBUG_ON then
debugging code will still be compiled in but is disabled by default.

Debugging can still be enabled by specifying

slub_debug

on the kernel command line. And AFACIT it would be easy
for the mempool subsystem to switch on slub debuggin if needed also.



Why is there a requirement to compile with SLUB_DEBUG_ON?


Oh someone put an #ifdef CONFIG_SLUB_DEBUG_ON in mempool.c. Overloading
the meaning of SLUB_DEBUG_ON as it is used in slub.c with something else.

Someone was assuming that SLUB_DEBUG_ON means the same as SLUB_DEBUG?

Please clean this mess up.

Re: [PATCH] mm/mempool: use static key for boot-time debug enablement

[Matthew Wilcox]

On Wed, May 27, 2026 at 02:29:22PM -0700, Christoph Lameter (Ampere) wrote:
> Oh someone put an #ifdef CONFIG_SLUB_DEBUG_ON in mempool.c. Overloading
> the meaning of SLUB_DEBUG_ON as it is used in slub.c with something else.
> 
> Someone was assuming that SLUB_DEBUG_ON means the same as SLUB_DEBUG?
> 
> Please clean this mess up.

Isn't that what this patch does?

Re: [PATCH] mm/mempool: use static key for boot-time debug enablement

[Christoph Lameter (Ampere)]

On Wed, 27 May 2026, Matthew Wilcox wrote:

> > Please clean this mess up.
>
> Isn't that what this patch does?

Its not marked as fixing something nor as addressing the weirdness of
using CONFIG_SLUB_DEBUG_ON here, A kernel build with CONFIG_SLUB_DEBUG_ON
can still boot without debugging if a certain kernel command line option
is given.

Re: [PATCH] mm/mempool: use static key for boot-time debug enablement

[Matthew Wilcox]

On Wed, May 27, 2026 at 04:06:18PM -0700, Christoph Lameter (Ampere) wrote:
> On Wed, 27 May 2026, Matthew Wilcox wrote:
> 
> > > Please clean this mess up.
> >
> > Isn't that what this patch does?
> 
> Its not marked as fixing something nor as addressing the weirdness of
> using CONFIG_SLUB_DEBUG_ON here, A kernel build with CONFIG_SLUB_DEBUG_ON
> can still boot without debugging if a certain kernel command line option
> is given.

Right, but ... if you look at what the patch _does_, doesn't it do what
you're asking for it to do?