2009-11-06 11:43:47

by Alan

Subject: [PATCH] slqb: A not quite so trivial warning

mm/slqb.c:2773: warning: initialization from incompatible pointer type

Variable scope...

Signed-off-by: Alan Cox <[email protected]>
---

mm/slqb.c | 8 ++++----
1 files changed, 4 insertions(+), 4 deletions(-)


diff --git a/mm/slqb.c b/mm/slqb.c
index e745d9a..fddcf85 100644
--- a/mm/slqb.c
+++ b/mm/slqb.c
@@ -2770,16 +2770,16 @@ static DEFINE_PER_CPU(struct delayed_work, cache_trim_work);

static void __cpuinit start_cpu_timer(int cpu)
{
- struct delayed_work *cache_trim_work = &per_cpu(cache_trim_work, cpu);
+ struct delayed_work *lcache_trim_work = &per_cpu(cache_trim_work, cpu);

/*
* When this gets called from do_initcalls via cpucache_init(),
* init_workqueues() has already run, so keventd will be setup
* at that time.
*/
- if (keventd_up() && cache_trim_work->work.func == NULL) {
- INIT_DELAYED_WORK(cache_trim_work, cache_trim_worker);
- schedule_delayed_work_on(cpu, cache_trim_work,
+ if (keventd_up() && lcache_trim_work->work.func == NULL) {
+ INIT_DELAYED_WORK(lcache_trim_work, cache_trim_worker);
+ schedule_delayed_work_on(cpu, lcache_trim_work,
__round_jiffies_relative(HZ, cpu));
}
}

2009-11-08 11:39:17

by Pekka Enberg

Subject: Re: [PATCH] slqb: A not quite so trivial warning

Hi Alan,

Alan Cox wrote:
> mm/slqb.c:2773: warning: initialization from incompatible pointer type
>
> Variable scope...
>
> Signed-off-by: Alan Cox <[email protected]>

I don't see this warning in my tree. Is this coming from linux-next? Did
someone change DEFINE_PER_CPU not to prefix the variable name with
"per_cpu__"?

Pekka

> ---
>
> mm/slqb.c | 8 ++++----
> 1 files changed, 4 insertions(+), 4 deletions(-)
>
>
> diff --git a/mm/slqb.c b/mm/slqb.c
> index e745d9a..fddcf85 100644
> --- a/mm/slqb.c
> +++ b/mm/slqb.c
> @@ -2770,16 +2770,16 @@ static DEFINE_PER_CPU(struct delayed_work, cache_trim_work);
>
> static void __cpuinit start_cpu_timer(int cpu)
> {
> - struct delayed_work *cache_trim_work = &per_cpu(cache_trim_work, cpu);
> + struct delayed_work *lcache_trim_work = &per_cpu(cache_trim_work, cpu);
>
> /*
> * When this gets called from do_initcalls via cpucache_init(),
> * init_workqueues() has already run, so keventd will be setup
> * at that time.
> */
> - if (keventd_up() && cache_trim_work->work.func == NULL) {
> - INIT_DELAYED_WORK(cache_trim_work, cache_trim_worker);
> - schedule_delayed_work_on(cpu, cache_trim_work,
> + if (keventd_up() && lcache_trim_work->work.func == NULL) {
> + INIT_DELAYED_WORK(lcache_trim_work, cache_trim_worker);
> + schedule_delayed_work_on(cpu, lcache_trim_work,
> __round_jiffies_relative(HZ, cpu));
> }
> }
>

2009-11-08 11:57:31

by Alan

Subject: Re: [PATCH] slqb: A not quite so trivial warning

On Sun, 08 Nov 2009 13:39:13 +0200
Pekka Enberg <[email protected]> wrote:

> Hi Alan,
>
> Alan Cox wrote:
> > mm/slqb.c:2773: warning: initialization from incompatible pointer type
> >
> > Variable scope...
> >
> > Signed-off-by: Alan Cox <[email protected]>
>
> I don't see this warning in my tree. Is this coming from linux-next? Did
> someone change DEFINE_PER_CPU not to prefix the variable name with
> "per_cpu__"?

I see it from -next on an x86-32 build

2009-11-08 12:04:35

by Tejun Heo

Subject: Re: [PATCH] slqb: A not quite so trivial warning

Hello,

Pekka Enberg wrote:
> Alan Cox wrote:
>> mm/slqb.c:2773: warning: initialization from incompatible pointer type
>>
>> Variable scope...
>>
>> Signed-off-by: Alan Cox <[email protected]>
>
> I don't see this warning in my tree. Is this coming from linux-next? Did
> someone change DEFINE_PER_CPU not to prefix the variable name with
> "per_cpu__"?

Yeap, the prefix is dropped in the latest percpu tree. Patch already
pending.

http://patchwork.kernel.org/patch/56948/

Thanks.

--
tejun

2009-11-08 12:29:41

by Pekka Enberg

Subject: Re: [PATCH] slqb: A not quite so trivial warning

Tejun Heo wrote:
> Hello,
>
> Pekka Enberg wrote:
>> Alan Cox wrote:
>>> mm/slqb.c:2773: warning: initialization from incompatible pointer type
>>>
>>> Variable scope...
>>>
>>> Signed-off-by: Alan Cox <[email protected]>
>> I don't see this warning in my tree. Is this coming from linux-next? Did
>> someone change DEFINE_PER_CPU not to prefix the variable name with
>> "per_cpu__"?
>
> Yeap, the prefix is dropped in the latest percpu tree. Patch already
> pending.
>
> http://patchwork.kernel.org/patch/56948/
>
> Thanks.

OK, I applied the above patch. Please CC me on patches that have the
magic letters S, L, and B in the future.

Pekka