[PATCH] SLUB: detach SLUB_DEBUG and SYSFS

[PATCH] SLUB: detach SLUB_DEBUG and SYSFS

[Alexey Dobriyan]

Right now, SYSFS=n means no SLUB debugging, no even basic poisoning,
to hell with tunables.

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
---

 init/Kconfig |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

--- a/init/Kconfig
+++ b/init/Kconfig
@@ -736,7 +736,7 @@ config VM_EVENT_COUNTERS
 config SLUB_DEBUG
 	default y
 	bool "Enable SLUB debugging support" if EMBEDDED
-	depends on SLUB && SYSFS
+	depends on SLUB
 	help
 	  SLUB has extensive debug support features. Disabling these can
 	  result in significant savings in code size. This also disables

Re: [PATCH] SLUB: detach SLUB_DEBUG and SYSFS

[Christoph Lameter]

Alexey Dobriyan wrote:
> Right now, SYSFS=n means no SLUB debugging, no even basic poisoning,
> to hell with tunables.

True. We did this because otherwise we would have complex #ifdeffery in slub.

commit f6acb63508700b5f8cd817082b62c96ba907775e
Author: Christoph Lameter <clameter@sgi.com>
Date:   Tue Apr 29 16:16:06 2008 -0700

    slub: #ifdef simplification

    If we make SLUB_DEBUG depend on SYSFS then we can simplify some
    #ifdefs and avoid others.

    Signed-off-by: Christoph Lameter <clameter@sgi.com>
    Signed-off-by: Pekka Enberg <penberg@cs.helsinki.fi>

Re: [PATCH] SLUB: detach SLUB_DEBUG and SYSFS

[Pekka Enberg]

Alexey Dobriyan wrote:
> Right now, SYSFS=n means no SLUB debugging, no even basic poisoning,
> to hell with tunables.

Applied, thanks!

Re: [PATCH] SLUB: detach SLUB_DEBUG and SYSFS

[Christoph Lameter]

Pekka Enberg wrote:
> Alexey Dobriyan wrote:
>> Right now, SYSFS=n means no SLUB debugging, no even basic poisoning,
>> to hell with tunables.
> 
> Applied, thanks!

Do not apply. This partially reverts an earlier commit and would cause build
issues because the #ifdef parts in slub.c were not reverted.

Re: [PATCH] SLUB: detach SLUB_DEBUG and SYSFS

[Pekka Enberg]

Christoph Lameter wrote:
> Pekka Enberg wrote:
>> Alexey Dobriyan wrote:
>>> Right now, SYSFS=n means no SLUB debugging, no even basic poisoning,
>>> to hell with tunables.
>> Applied, thanks!
> 
> Do not apply. This partially reverts an earlier commit and would cause build
> issues because the #ifdef parts in slub.c were not reverted.

Aww, crap. OK, taking it out.

Re: [PATCH] SLUB: detach SLUB_DEBUG and SYSFS

[Alexey Dobriyan]

On Mon, Aug 04, 2008 at 05:10:37PM +0300, Pekka Enberg wrote:
> Christoph Lameter wrote:
>> Pekka Enberg wrote:
>>> Alexey Dobriyan wrote:
>>>> Right now, SYSFS=n means no SLUB debugging, no even basic poisoning,
>>>> to hell with tunables.
>>> Applied, thanks!
>> Do not apply. This partially reverts an earlier commit and would cause 
>> build
>> issues because the #ifdef parts in slub.c were not reverted.
>
> Aww, crap. OK, taking it out.

OK, I'll test-compile it to death.

For now, do you agree that SYSFS=n users shouldn't be discriminated
against poisoning?

As real world situation, NET_NS feature which is currently in
active development doesn't work with SYSFS, so it was a bit of cold
shower to realise that I did much testing of conntracking in netns
without poisoning. :-(

Re: [PATCH] SLUB: detach SLUB_DEBUG and SYSFS

[Pekka Enberg]

Alexey Dobriyan wrote:
> OK, I'll test-compile it to death.
> 
> For now, do you agree that SYSFS=n users shouldn't be discriminated
> against poisoning?
> 
> As real world situation, NET_NS feature which is currently in
> active development doesn't work with SYSFS, so it was a bit of cold
> shower to realise that I did much testing of conntracking in netns
> without poisoning. :-(

I'm okay with that as long as it doesn't break the build :-)

Re: [PATCH] SLUB: detach SLUB_DEBUG and SYSFS

[Christoph Lameter]

Alexey Dobriyan wrote:
> On Mon, Aug 04, 2008 at 05:10:37PM +0300, Pekka Enberg wrote:
>> Christoph Lameter wrote:
>>> Pekka Enberg wrote:
>>>> Alexey Dobriyan wrote:
>>>>> Right now, SYSFS=n means no SLUB debugging, no even basic poisoning,
>>>>> to hell with tunables.
>>>> Applied, thanks!
>>> Do not apply. This partially reverts an earlier commit and would cause 
>>> build
>>> issues because the #ifdef parts in slub.c were not reverted.
>> Aww, crap. OK, taking it out.
> 
> OK, I'll test-compile it to death.

Not sure what the point would be. If you look at the commit you are reverting
then its clear what configurations will break,

> For now, do you agree that SYSFS=n users shouldn't be discriminated
> against poisoning?

If you want a minimal memory footprint then I thought that also implies that
the debug code would not be compiled in?

> As real world situation, NET_NS feature which is currently in
> active development doesn't work with SYSFS, so it was a bit of cold
> shower to realise that I did much testing of conntracking in netns
> without poisoning. :-(

Ohh.. There are kernel features that conflict with SYSFS?

Re: [PATCH] SLUB: detach SLUB_DEBUG and SYSFS

[Max Krasnyansky]

Alexey Dobriyan wrote:
> Right now, SYSFS=n means no SLUB debugging, no even basic poisoning,
> to hell with tunables.
> 
> Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
> ---
> 
>  init/Kconfig |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> --- a/init/Kconfig
> +++ b/init/Kconfig
> @@ -736,7 +736,7 @@ config VM_EVENT_COUNTERS
>  config SLUB_DEBUG
>  	default y
>  	bool "Enable SLUB debugging support" if EMBEDDED
> -	depends on SLUB && SYSFS
> +	depends on SLUB
>  	help
>  	  SLUB has extensive debug support features. Disabling these can
>  	  result in significant savings in code size. This also disables

While at it, can you guys move this into lib/Kconfig.debug ?

Max