Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1316318 > unrolled thread

[PATCH] mm/debug_pagealloc: Ask users for default setting of debug_pagealloc

Back to article view | Back to linux.kernel

Contents

  [PATCH] mm/debug_pagealloc: Ask users for default setting of debug_pagealloc Christian Borntraeger <borntraeger@de.ibm.com> - 2016-01-25 10:20 +0100
    Re: [PATCH] mm/debug_pagealloc: Ask users for default setting of  debug_pagealloc Christian Borntraeger <borntraeger@de.ibm.com> - 2016-01-25 10:50 +0100
      Re: [PATCH] mm/debug_pagealloc: Ask users for default setting of  debug_pagealloc Heiko Carstens <heiko.carstens@de.ibm.com> - 2016-01-25 11:10 +0100
        Re: [PATCH] mm/debug_pagealloc: Ask users for default setting of  debug_pagealloc Christian Borntraeger <borntraeger@de.ibm.com> - 2016-01-25 11:10 +0100
          Re: [PATCH] mm/debug_pagealloc: Ask users for default setting of  debug_pagealloc Christian Borntraeger <borntraeger@de.ibm.com> - 2016-01-25 12:20 +0100
    Re: [PATCH] mm/debug_pagealloc: Ask users for default setting of  debug_pagealloc Heiko Carstens <heiko.carstens@de.ibm.com> - 2016-01-25 10:50 +0100

#1316318 — [PATCH] mm/debug_pagealloc: Ask users for default setting of debug_pagealloc

since commit 031bc5743f158 ("mm/debug-pagealloc: make debug-pagealloc
boottime configurable") CONFIG_DEBUG_PAGEALLOC is by default a no-op.

This resulted in several unnoticed bugs, e.g.

https://lkml.kernel.org/g/<569F5E29.3090107@de.ibm.com>
or
https://lkml.kernel.org/g/<56A20F30.4050705@de.ibm.com>

as this behaviour change was not even documented in Kconfig.

Let's provide a new Kconfig symbol that allows to change the default
back to enabled, e.g. for debug kernels. This also makes the change
obvious to kernel packagers.

Let's also change the Kconfig description for CONFIG_DEBUG_PAGEALLOC,
to indicate that it is ok to enable this by default.

Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
---
 mm/Kconfig.debug | 17 +++++++++++++++++
 mm/page_alloc.c  |  6 +++++-
 2 files changed, 22 insertions(+), 1 deletion(-)

diff --git a/mm/Kconfig.debug b/mm/Kconfig.debug
index 957d3da..4cf1212 100644
--- a/mm/Kconfig.debug
+++ b/mm/Kconfig.debug
@@ -26,5 +26,22 @@ config DEBUG_PAGEALLOC
 	  that would result in incorrect warnings of memory corruption after
 	  a resume because free pages are not saved to the suspend image.
 
+	  By default this option will be almost for free and can be activated
+	  in distribution kernels. The overhead and the debugging can be enabled
+	  by DEBUG_PAGEALLOC_ENABLE_DEFAULT or the debug_pagealloc command line
+	  parameter.
+
+	  If unsure say yes.
+
+config DEBUG_PAGEALLOC_ENABLE_DEFAULT
+	bool "Enable debug page memory allocations by default?"
+        default off
+        depends on DEBUG_PAGEALLOC
+        ---help---
+	  Enable debug page memory allocations by default? This value
+	  can be overridden by debug_pagealloc=off|on.
+
+	  If unsure say no.
+
 config PAGE_POISONING
 	bool
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 9d666df..933def7 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -479,7 +479,8 @@ void prep_compound_page(struct page *page, unsigned int order)
 
 #ifdef CONFIG_DEBUG_PAGEALLOC
 unsigned int _debug_guardpage_minorder;
-bool _debug_pagealloc_enabled __read_mostly;
+bool _debug_pagealloc_enabled __read_mostly
+			= IS_ENABLED(CONFIG_DEBUG_PAGEALLOC_ENABLE_DEFAULT);
 bool _debug_guardpage_enabled __read_mostly;
 
 static int __init early_debug_pagealloc(char *buf)
@@ -490,6 +491,9 @@ static int __init early_debug_pagealloc(char *buf)
 	if (strcmp(buf, "on") == 0)
 		_debug_pagealloc_enabled = true;
 
+	if (strcmp(buf, "off") == 0)
+		_debug_pagealloc_enabled = false;
+
 	return 0;
 }
 early_param("debug_pagealloc", early_debug_pagealloc);
-- 
2.3.0

[toc] | [next] | [standalone]

#1316336 — Re: [PATCH] mm/debug_pagealloc: Ask users for default setting of debug_pagealloc

On 01/25/2016 10:41 AM, Heiko Carstens wrote:
> On Mon, Jan 25, 2016 at 10:19:48AM +0100, Christian Borntraeger wrote:
>> since commit 031bc5743f158 ("mm/debug-pagealloc: make debug-pagealloc
>> boottime configurable") CONFIG_DEBUG_PAGEALLOC is by default a no-op.
>>
>> This resulted in several unnoticed bugs, e.g.
>>
>> https://lkml.kernel.org/g/<569F5E29.3090107@de.ibm.com>
>> or
>> https://lkml.kernel.org/g/<56A20F30.4050705@de.ibm.com>
>>
>> as this behaviour change was not even documented in Kconfig.
>>
>> Let's provide a new Kconfig symbol that allows to change the default
>> back to enabled, e.g. for debug kernels. This also makes the change
>> obvious to kernel packagers.
>>
>> Let's also change the Kconfig description for CONFIG_DEBUG_PAGEALLOC,
>> to indicate that it is ok to enable this by default.
>>
>> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
>> Cc: Andrew Morton <akpm@linux-foundation.org>
>> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
>> ---
>>  mm/Kconfig.debug | 17 +++++++++++++++++
>>  mm/page_alloc.c  |  6 +++++-
>>  2 files changed, 22 insertions(+), 1 deletion(-)
>>
>> diff --git a/mm/Kconfig.debug b/mm/Kconfig.debug
>> index 957d3da..4cf1212 100644
>> --- a/mm/Kconfig.debug
>> +++ b/mm/Kconfig.debug
>> @@ -26,5 +26,22 @@ config DEBUG_PAGEALLOC
>>  	  that would result in incorrect warnings of memory corruption after
>>  	  a resume because free pages are not saved to the suspend image.
>>
>> +	  By default this option will be almost for free and can be activated
>> +	  in distribution kernels. The overhead and the debugging can be enabled
>> +	  by DEBUG_PAGEALLOC_ENABLE_DEFAULT or the debug_pagealloc command line
>> +	  parameter.
> 
> Sorry, but it's not almost for free and should not be used by distribution
> kernels. If we have DEBUG_PAGEALLOC enabled, at least on s390 we will not
> make use of 2GB and 1MB pagetable entries for the identy mapping anymore.
> Instead we will only use 4K mappings.

Hmmm, can we change these code areas to use debug_pagealloc_enabled? I guess
this evaluated too late?


> I assume this is true for all architectures since freeing pages can happen
> in any context and therefore we can't allocate memory in order to split
> page tables.
> 
> So enabling this will cost memory and put more pressure on the TLB.

So I will change the description and drop the "if unsure" statement.

Christian

[toc] | [prev] | [next] | [standalone]

#1316379 — Re: [PATCH] mm/debug_pagealloc: Ask users for default setting of debug_pagealloc

On Mon, Jan 25, 2016 at 10:45:50AM +0100, Christian Borntraeger wrote:
> >> +	  By default this option will be almost for free and can be activated
> >> +	  in distribution kernels. The overhead and the debugging can be enabled
> >> +	  by DEBUG_PAGEALLOC_ENABLE_DEFAULT or the debug_pagealloc command line
> >> +	  parameter.
> > 
> > Sorry, but it's not almost for free and should not be used by distribution
> > kernels. If we have DEBUG_PAGEALLOC enabled, at least on s390 we will not
> > make use of 2GB and 1MB pagetable entries for the identy mapping anymore.
> > Instead we will only use 4K mappings.
> 
> Hmmm, can we change these code areas to use debug_pagealloc_enabled? I guess
> this evaluated too late?

Yes, that should be possible. "debug_pagealloc" is an early_param, which
will be evaluated before we call paging_init() (both in
arch/s390/kernel/setup.c).

So it looks like this can be trivially changed. (replace the ifdefs in
arch/s390/mm/vmem.c with debug_pagealloc_enabled()).

> > I assume this is true for all architectures since freeing pages can happen
> > in any context and therefore we can't allocate memory in order to split
> > page tables.
> > 
> > So enabling this will cost memory and put more pressure on the TLB.
> 
> So I will change the description and drop the "if unsure" statement.

Well, given that we can change it like above... I don't care anymore ;)

[toc] | [prev] | [next] | [standalone]

#1316385 — Re: [PATCH] mm/debug_pagealloc: Ask users for default setting of debug_pagealloc

On 01/25/2016 11:02 AM, Heiko Carstens wrote:
> On Mon, Jan 25, 2016 at 10:45:50AM +0100, Christian Borntraeger wrote:
>>>> +	  By default this option will be almost for free and can be activated
>>>> +	  in distribution kernels. The overhead and the debugging can be enabled
>>>> +	  by DEBUG_PAGEALLOC_ENABLE_DEFAULT or the debug_pagealloc command line
>>>> +	  parameter.
>>>
>>> Sorry, but it's not almost for free and should not be used by distribution
>>> kernels. If we have DEBUG_PAGEALLOC enabled, at least on s390 we will not
>>> make use of 2GB and 1MB pagetable entries for the identy mapping anymore.
>>> Instead we will only use 4K mappings.
>>
>> Hmmm, can we change these code areas to use debug_pagealloc_enabled? I guess
>> this evaluated too late?
> 
> Yes, that should be possible. "debug_pagealloc" is an early_param, which
> will be evaluated before we call paging_init() (both in
> arch/s390/kernel/setup.c).
> 
> So it looks like this can be trivially changed. (replace the ifdefs in
> arch/s390/mm/vmem.c with debug_pagealloc_enabled()).
> 
>>> I assume this is true for all architectures since freeing pages can happen
>>> in any context and therefore we can't allocate memory in order to split
>>> page tables.
>>>
>>> So enabling this will cost memory and put more pressure on the TLB.
>>
>> So I will change the description and drop the "if unsure" statement.
> 
> Well, given that we can change it like above... I don't care anymore ;)

Ok, I will give it a try, and come back with a rewording or an s390 patch.

[toc] | [prev] | [next] | [standalone]

#1316493 — Re: [PATCH] mm/debug_pagealloc: Ask users for default setting of debug_pagealloc

On 01/25/2016 11:07 AM, Christian Borntraeger wrote:
> On 01/25/2016 11:02 AM, Heiko Carstens wrote:
>> On Mon, Jan 25, 2016 at 10:45:50AM +0100, Christian Borntraeger wrote:
>>>>> +	  By default this option will be almost for free and can be activated
>>>>> +	  in distribution kernels. The overhead and the debugging can be enabled
>>>>> +	  by DEBUG_PAGEALLOC_ENABLE_DEFAULT or the debug_pagealloc command line
>>>>> +	  parameter.
>>>>
>>>> Sorry, but it's not almost for free and should not be used by distribution
>>>> kernels. If we have DEBUG_PAGEALLOC enabled, at least on s390 we will not
>>>> make use of 2GB and 1MB pagetable entries for the identy mapping anymore.
>>>> Instead we will only use 4K mappings.
>>>
>>> Hmmm, can we change these code areas to use debug_pagealloc_enabled? I guess
>>> this evaluated too late?
>>
>> Yes, that should be possible. "debug_pagealloc" is an early_param, which
>> will be evaluated before we call paging_init() (both in
>> arch/s390/kernel/setup.c).
>>
>> So it looks like this can be trivially changed. (replace the ifdefs in
>> arch/s390/mm/vmem.c with debug_pagealloc_enabled()).
>>
>>>> I assume this is true for all architectures since freeing pages can happen
>>>> in any context and therefore we can't allocate memory in order to split
>>>> page tables.
>>>>
>>>> So enabling this will cost memory and put more pressure on the TLB.
>>>
>>> So I will change the description and drop the "if unsure" statement.
>>
>> Well, given that we can change it like above... I don't care anymore ;)
> 
> Ok, I will give it a try, and come back with a rewording or an s390 patch.

I have a patch for x86 and s390. powerpc should also be possible.

Now it seems that sparc already defines the TSB very early in head.S. 
Unless we find a solution for sparc to use debug_pagealloc_enabled()
I will modify the patch description and resend the patch.

[toc] | [prev] | [next] | [standalone]

#1316340 — Re: [PATCH] mm/debug_pagealloc: Ask users for default setting of debug_pagealloc

On Mon, Jan 25, 2016 at 10:19:48AM +0100, Christian Borntraeger wrote:
> since commit 031bc5743f158 ("mm/debug-pagealloc: make debug-pagealloc
> boottime configurable") CONFIG_DEBUG_PAGEALLOC is by default a no-op.
> 
> This resulted in several unnoticed bugs, e.g.
> 
> https://lkml.kernel.org/g/<569F5E29.3090107@de.ibm.com>
> or
> https://lkml.kernel.org/g/<56A20F30.4050705@de.ibm.com>
> 
> as this behaviour change was not even documented in Kconfig.
> 
> Let's provide a new Kconfig symbol that allows to change the default
> back to enabled, e.g. for debug kernels. This also makes the change
> obvious to kernel packagers.
> 
> Let's also change the Kconfig description for CONFIG_DEBUG_PAGEALLOC,
> to indicate that it is ok to enable this by default.
> 
> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
> ---
>  mm/Kconfig.debug | 17 +++++++++++++++++
>  mm/page_alloc.c  |  6 +++++-
>  2 files changed, 22 insertions(+), 1 deletion(-)
> 
> diff --git a/mm/Kconfig.debug b/mm/Kconfig.debug
> index 957d3da..4cf1212 100644
> --- a/mm/Kconfig.debug
> +++ b/mm/Kconfig.debug
> @@ -26,5 +26,22 @@ config DEBUG_PAGEALLOC
>  	  that would result in incorrect warnings of memory corruption after
>  	  a resume because free pages are not saved to the suspend image.
> 
> +	  By default this option will be almost for free and can be activated
> +	  in distribution kernels. The overhead and the debugging can be enabled
> +	  by DEBUG_PAGEALLOC_ENABLE_DEFAULT or the debug_pagealloc command line
> +	  parameter.

Sorry, but it's not almost for free and should not be used by distribution
kernels. If we have DEBUG_PAGEALLOC enabled, at least on s390 we will not
make use of 2GB and 1MB pagetable entries for the identy mapping anymore.
Instead we will only use 4K mappings.

I assume this is true for all architectures since freeing pages can happen
in any context and therefore we can't allocate memory in order to split
page tables.

So enabling this will cost memory and put more pressure on the TLB.

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web