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


Groups > linux.kernel > #1610405 > unrolled thread

[PATCH] percpu: drop useless macro PER_CPU_DEF_ATTRIBUTES

Started byCao jin <caoj.fnst@cn.fujitsu.com>
First post2017-03-28 10:30 +0200
Last post2017-03-30 05:30 +0200
Articles 5 — 2 participants

Back to article view | Back to linux.kernel


Contents

  [PATCH] percpu: drop useless macro PER_CPU_DEF_ATTRIBUTES Cao jin <caoj.fnst@cn.fujitsu.com> - 2017-03-28 10:30 +0200
    Re: [PATCH] percpu: drop useless macro PER_CPU_DEF_ATTRIBUTES Tejun Heo <tj@kernel.org> - 2017-03-28 18:10 +0200
      Re: [PATCH] percpu: drop useless macro PER_CPU_DEF_ATTRIBUTES Cao jin <caoj.fnst@cn.fujitsu.com> - 2017-03-29 04:30 +0200
        Re: [PATCH] percpu: drop useless macro PER_CPU_DEF_ATTRIBUTES Tejun Heo <tj@kernel.org> - 2017-03-29 17:30 +0200
          Re: [PATCH] percpu: drop useless macro PER_CPU_DEF_ATTRIBUTES Cao jin <caoj.fnst@cn.fujitsu.com> - 2017-03-30 05:30 +0200

#1610405 — [PATCH] percpu: drop useless macro PER_CPU_DEF_ATTRIBUTES

FromCao jin <caoj.fnst@cn.fujitsu.com>
Date2017-03-28 10:30 +0200
Subject[PATCH] percpu: drop useless macro PER_CPU_DEF_ATTRIBUTES
Message-ID<tpUFP-4R5-3@gated-at.bofh.it>
Signed-off-by: Cao jin <caoj.fnst@cn.fujitsu.com>
---
 include/asm-generic/percpu.h | 4 ----
 include/linux/percpu-defs.h  | 4 ++--
 2 files changed, 2 insertions(+), 6 deletions(-)

diff --git a/include/asm-generic/percpu.h b/include/asm-generic/percpu.h
index 0504ef8..be90c3d 100644
--- a/include/asm-generic/percpu.h
+++ b/include/asm-generic/percpu.h
@@ -61,10 +61,6 @@ extern void setup_per_cpu_areas(void);
 #define PER_CPU_ATTRIBUTES
 #endif
 
-#ifndef PER_CPU_DEF_ATTRIBUTES
-#define PER_CPU_DEF_ATTRIBUTES
-#endif
-
 #define raw_cpu_generic_read(pcp)					\
 ({									\
 	*raw_cpu_ptr(&(pcp));						\
diff --git a/include/linux/percpu-defs.h b/include/linux/percpu-defs.h
index 8f16299..c05cc4d 100644
--- a/include/linux/percpu-defs.h
+++ b/include/linux/percpu-defs.h
@@ -91,7 +91,7 @@
 	extern __PCPU_DUMMY_ATTRS char __pcpu_unique_##name;		\
 	__PCPU_DUMMY_ATTRS char __pcpu_unique_##name;			\
 	extern __PCPU_ATTRS(sec) __typeof__(type) name;			\
-	__PCPU_ATTRS(sec) PER_CPU_DEF_ATTRIBUTES __weak			\
+	__PCPU_ATTRS(sec) __weak					\
 	__typeof__(type) name
 #else
 /*
@@ -101,7 +101,7 @@
 	extern __PCPU_ATTRS(sec) __typeof__(type) name
 
 #define DEFINE_PER_CPU_SECTION(type, name, sec)				\
-	__PCPU_ATTRS(sec) PER_CPU_DEF_ATTRIBUTES			\
+	__PCPU_ATTRS(sec)						\
 	__typeof__(type) name
 #endif
 
-- 
2.1.0

[toc] | [next] | [standalone]


#1611149

FromTejun Heo <tj@kernel.org>
Date2017-03-28 18:10 +0200
Message-ID<tq1R0-1HW-35@gated-at.bofh.it>
In reply to#1610405
On Tue, Mar 28, 2017 at 04:35:07PM +0800, Cao jin wrote:
> Signed-off-by: Cao jin <caoj.fnst@cn.fujitsu.com>

I'm not sure.  I think I used it during development and it's not like
it costs anything to keep.

Thanks.

-- 
tejun

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


#1611513

FromCao jin <caoj.fnst@cn.fujitsu.com>
Date2017-03-29 04:30 +0200
Message-ID<tqbwZ-hk-9@gated-at.bofh.it>
In reply to#1611149

On 03/29/2017 12:00 AM, Tejun Heo wrote:
> On Tue, Mar 28, 2017 at 04:35:07PM +0800, Cao jin wrote:
>> Signed-off-by: Cao jin <caoj.fnst@cn.fujitsu.com>
> 
> I'm not sure.  I think I used it during development and it's not like
> it costs anything to keep.
> 

It easily confusing newbies without any comments for its purpose.

Temporary using during development seems not good to persist. Could I
know its purpose?

-- 
Sincerely,
Cao jin

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


#1612057

FromTejun Heo <tj@kernel.org>
Date2017-03-29 17:30 +0200
Message-ID<tqnHQ-lp-15@gated-at.bofh.it>
In reply to#1611513
On Wed, Mar 29, 2017 at 10:38:17AM +0800, Cao jin wrote:
> On 03/29/2017 12:00 AM, Tejun Heo wrote:
> > On Tue, Mar 28, 2017 at 04:35:07PM +0800, Cao jin wrote:
> >> Signed-off-by: Cao jin <caoj.fnst@cn.fujitsu.com>
> > 
> > I'm not sure.  I think I used it during development and it's not like
> > it costs anything to keep.
> 
> It easily confusing newbies without any comments for its purpose.

percpu variable handling is pretty complex to begin with.  I don't buy
that an empty default attribute macro adds anything noticeable there.

> Temporary using during development seems not good to persist. Could I
> know its purpose?

I don't remember the exact details but at the same time I don't see
the value of the proposed change.  We can apply or not apply the patch
and either way wouldn't make any noticeable difference.  What's the
point?

Thanks.

-- 
tejun

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


#1612522

FromCao jin <caoj.fnst@cn.fujitsu.com>
Date2017-03-30 05:30 +0200
Message-ID<tqyWD-5f-25@gated-at.bofh.it>
In reply to#1612057

On 03/29/2017 11:28 PM, Tejun Heo wrote:
> On Wed, Mar 29, 2017 at 10:38:17AM +0800, Cao jin wrote:
>> On 03/29/2017 12:00 AM, Tejun Heo wrote:
>>> On Tue, Mar 28, 2017 at 04:35:07PM +0800, Cao jin wrote:
>>>> Signed-off-by: Cao jin <caoj.fnst@cn.fujitsu.com>
>>>
>>> I'm not sure.  I think I used it during development and it's not like
>>> it costs anything to keep.
>>
>> It easily confusing newbies without any comments for its purpose.
> 
> percpu variable handling is pretty complex to begin with.  I don't buy
> that an empty default attribute macro adds anything noticeable there.
> 
>> Temporary using during development seems not good to persist. Could I
>> know its purpose?
> 
> I don't remember the exact details but at the same time I don't see
> the value of the proposed change.  We can apply or not apply the patch
> and either way wouldn't make any noticeable difference.  What's the
> point?
> 
> Thanks.
> 

It doesn't matter whether apply or not. I just start to read these code
and want to know what every tiny symbol is for.  Thanks for your time.
-- 
Sincerely,
Cao jin

[toc] | [prev] | [standalone]


Back to top | Article view | linux.kernel


csiph-web