2011-03-16 18:18:43

by Daniel Baluta

Subject: [PATCH] kmemleak: build kmemleak-test.c as a module-only

mm/kmemleak-test.c is used to provide an example of how kmemleak
tool works.

Memory is leaked at module unload-time, so building the test
in kernel (Y) makes the leaks impossible and the test useless.

Qualify DEBUG_KMEMLEAK_TEST config symbol with "depends on m",
to restrict module-only building.

Signed-off-by: Daniel Baluta <[email protected]>
---
lib/Kconfig.debug | 6 ++----
1 files changed, 2 insertions(+), 4 deletions(-)

diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
index 689f311..6e9cacc 100644
--- a/lib/Kconfig.debug
+++ b/lib/Kconfig.debug
@@ -407,11 +407,9 @@ config DEBUG_KMEMLEAK_EARLY_LOG_SIZE

config DEBUG_KMEMLEAK_TEST
tristate "Simple test for the kernel memory leak detector"
- depends on DEBUG_KMEMLEAK
+ depends on DEBUG_KMEMLEAK && m
help
- Say Y or M here to build a test for the kernel memory leak
- detector. This option enables a module that explicitly leaks
- memory.
+ This option enables a module that explicitly leaks memory.

If unsure, say N.

--
1.7.1

2011-03-23 11:18:35

by Daniel Baluta

Subject: Re: [PATCH] kmemleak: build kmemleak-test.c as a module-only

On Wed, Mar 16, 2011 at 8:16 PM, Daniel Baluta <[email protected]> wrote:
> mm/kmemleak-test.c is used to provide an example of how kmemleak
> tool works.
>
> Memory is leaked at module unload-time, so building the test
> in kernel (Y) makes the leaks impossible and the test useless.

Any comments on this?

thanks,
Daniel.

2011-03-23 11:23:18

by Pekka Enberg

Subject: Re: [PATCH] kmemleak: build kmemleak-test.c as a module-only

On Wed, Mar 23, 2011 at 1:18 PM, Daniel Baluta <[email protected]> wrote:
> On Wed, Mar 16, 2011 at 8:16 PM, Daniel Baluta <[email protected]> wrote:
>> mm/kmemleak-test.c is used to provide an example of how kmemleak
>> tool works.
>>
>> Memory is leaked at module unload-time, so building the test
>> in kernel (Y) makes the leaks impossible and the test useless.
>
> Any comments on this?

Looks good to me.

Acked-by: Pekka Enberg <[email protected]>

2011-03-23 11:24:42

by Catalin Marinas

Subject: Re: [PATCH] kmemleak: build kmemleak-test.c as a module-only

On Wed, 2011-03-16 at 18:16 +0000, Daniel Baluta wrote:
> mm/kmemleak-test.c is used to provide an example of how kmemleak
> tool works.
>
> Memory is leaked at module unload-time, so building the test
> in kernel (Y) makes the leaks impossible and the test useless.
>
> Qualify DEBUG_KMEMLEAK_TEST config symbol with "depends on m",
> to restrict module-only building.
>
> Signed-off-by: Daniel Baluta <[email protected]>

Sorry for the delay.

Acked-by: Catalin Marinas <[email protected]>

2011-03-25 14:05:54

by Daniel Baluta

Subject: Re: [PATCH] kmemleak: build kmemleak-test.c as a module-only

On Wed, Mar 23, 2011 at 1:24 PM, Catalin Marinas
<[email protected]> wrote:
> On Wed, 2011-03-16 at 18:16 +0000, Daniel Baluta wrote:
>> mm/kmemleak-test.c is used to provide an example of how kmemleak
>> tool works.
>>
>> Memory is leaked at module unload-time, so building the test
>> in kernel (Y) makes the leaks impossible and the test useless.
>>
>> Qualify DEBUG_KMEMLEAK_TEST config symbol with "depends on m",
>> to restrict module-only building.
>>
>> Signed-off-by: Daniel Baluta <[email protected]>
>
> Sorry for the delay.
>
> Acked-by: Catalin Marinas <[email protected]>

Who should pick this up?

$ scripts/get_maintainer.pl lib/this.patch says:
Andrew Morton <[email protected]> (commit_signer:8/56=14%)
"Paul E. McKenney" <[email protected]> (commit_signer:7/56=12%)
Ingo Molnar <[email protected]> (commit_signer:5/56=9%)
Frederic Weisbecker <[email protected]> (commit_signer:5/56=9%)
Randy Dunlap <[email protected]> (commit_signer:3/56=5%)

Am I missing something?

thanks,
Daniel.

2011-03-30 15:02:07

by Daniel Baluta

Subject: Re: [PATCH] kmemleak: build kmemleak-test.c as a module-only

On Fri, Mar 25, 2011 at 4:05 PM, Daniel Baluta <[email protected]> wrote:
> On Wed, Mar 23, 2011 at 1:24 PM, Catalin Marinas
> <[email protected]> wrote:
>> On Wed, 2011-03-16 at 18:16 +0000, Daniel Baluta wrote:
>>> mm/kmemleak-test.c is used to provide an example of how kmemleak
>>> tool works.
>>>
>>> Memory is leaked at module unload-time, so building the test
>>> in kernel (Y) makes the leaks impossible and the test useless.
>>>
>>> Qualify DEBUG_KMEMLEAK_TEST config symbol with "depends on m",
>>> to restrict module-only building.
>>>
>>> Signed-off-by: Daniel Baluta <[email protected]>
>>
>> Sorry for the delay.
>>
>> Acked-by: Catalin Marinas <[email protected]>
>
> Who should pick this up?
>
> $ scripts/get_maintainer.pl lib/this.patch says:
> Andrew Morton <[email protected]> (commit_signer:8/56=14%)
> "Paul E. McKenney" <[email protected]> (commit_signer:7/56=12%)
> Ingo Molnar <[email protected]> (commit_signer:5/56=9%)
> Frederic Weisbecker <[email protected]> (commit_signer:5/56=9%)
> Randy Dunlap <[email protected]> (commit_signer:3/56=5%)

Catalin, Randy since lib/Kconfig.debug has no official maintainer
I think this should go up through your trees.

thanks,
Daniel.

2011-03-30 15:24:24

by Randy Dunlap

Subject: Re: [PATCH] kmemleak: build kmemleak-test.c as a module-only

On Wed, 30 Mar 2011 18:02:02 +0300 Daniel Baluta wrote:

> On Fri, Mar 25, 2011 at 4:05 PM, Daniel Baluta <[email protected]> wrote:
> > On Wed, Mar 23, 2011 at 1:24 PM, Catalin Marinas
> > <[email protected]> wrote:
> >> On Wed, 2011-03-16 at 18:16 +0000, Daniel Baluta wrote:
> >>> mm/kmemleak-test.c is used to provide an example of how kmemleak
> >>> tool works.
> >>>
> >>> Memory is leaked at module unload-time, so building the test
> >>> in kernel (Y) makes the leaks impossible and the test useless.
> >>>
> >>> Qualify DEBUG_KMEMLEAK_TEST config symbol with "depends on m",
> >>> to restrict module-only building.
> >>>
> >>> Signed-off-by: Daniel Baluta <[email protected]>
> >>
> >> Sorry for the delay.
> >>
> >> Acked-by: Catalin Marinas <[email protected]>
> >
> > Who should pick this up?
> >
> > $ scripts/get_maintainer.pl lib/this.patch says:
> > Andrew Morton <[email protected]> (commit_signer:8/56=14%)
> > "Paul E. McKenney" <[email protected]> (commit_signer:7/56=12%)
> > Ingo Molnar <[email protected]> (commit_signer:5/56=9%)
> > Frederic Weisbecker <[email protected]> (commit_signer:5/56=9%)
> > Randy Dunlap <[email protected]> (commit_signer:3/56=5%)
>
> Catalin, Randy since lib/Kconfig.debug has no official maintainer
> I think this should go up through your trees.

I hadn't seen it until now, but I can merge it.

thanks.

---
~Randy
*** Remember to use Documentation/SubmitChecklist when testing your code ***

2011-03-31 07:44:47

by Daniel Baluta

Subject: Re: [PATCH] kmemleak: build kmemleak-test.c as a module-only

On Wed, Mar 30, 2011 at 6:24 PM, Randy Dunlap <[email protected]> wrote:
> On Wed, 30 Mar 2011 18:02:02 +0300 Daniel Baluta wrote:
>
>> On Fri, Mar 25, 2011 at 4:05 PM, Daniel Baluta <[email protected]> wrote:
>> > On Wed, Mar 23, 2011 at 1:24 PM, Catalin Marinas
>> > <[email protected]> wrote:
>> >> On Wed, 2011-03-16 at 18:16 +0000, Daniel Baluta wrote:
>> >>> mm/kmemleak-test.c is used to provide an example of how kmemleak
>> >>> tool works.
>> >>>
>> >>> Memory is leaked at module unload-time, so building the test
>> >>> in kernel (Y) makes the leaks impossible and the test useless.
>> >>>
>> >>> Qualify DEBUG_KMEMLEAK_TEST config symbol with "depends on m",
>> >>> to restrict module-only building.
>> >>>
>> >>> Signed-off-by: Daniel Baluta <[email protected]>
>> >>
>> >> Sorry for the delay.
>> >>
>> >> Acked-by: Catalin Marinas <[email protected]>
>> >
>> > Who should pick this up?
>> >
>> > $ scripts/get_maintainer.pl lib/this.patch says:
>> > Andrew Morton <[email protected]> (commit_signer:8/56=14%)
>> > "Paul E. McKenney" <[email protected]> (commit_signer:7/56=12%)
>> > Ingo Molnar <[email protected]> (commit_signer:5/56=9%)
>> > Frederic Weisbecker <[email protected]> (commit_signer:5/56=9%)
>> > Randy Dunlap <[email protected]> (commit_signer:3/56=5%)
>>
>> Catalin, Randy since lib/Kconfig.debug has no official maintainer
>> I think this should go up through your trees.
>
> I hadn't seen it until now, but I can merge it.

OK. thanks.