All of lore.kernel.org
 help / color / mirror / Atom feed
From: Nicolas Pitre <nicolas.pitre@linaro.org>
To: Tejun Heo <tj@kernel.org>
Cc: Christoph Lameter <cl@linux.com>,
	linux-mm@kvack.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] mm/percpu.c: use smarter memory allocation for struct pcpu_alloc_info
Date: Tue, 3 Oct 2017 18:29:49 -0400 (EDT)	[thread overview]
Message-ID: <nycvar.YSQ.7.76.1710031731130.5407@knanqh.ubzr> (raw)
In-Reply-To: <20171003210540.GM3301751@devbig577.frc2.facebook.com>

On Tue, 3 Oct 2017, Tejun Heo wrote:

> On Tue, Oct 03, 2017 at 04:57:44PM -0400, Nicolas Pitre wrote:
> > This can be much smaller than a page on very small memory systems. 
> > Always rounding up the size to a page is wasteful in that case, and 
> > required alignment is smaller than the memblock default. Let's round 
> > things up to a page size only when the actual size is >= page size, and 
> > then it makes sense to page-align for a nicer allocation pattern.
> 
> Isn't that a temporary area which gets freed later during boot?

Hmmm...

It may get freed through 3 different paths where 2 of them are error 
paths. What looks like a non-error path is in pcpu_embed_first_chunk() 
called from setup_per_cpu_areas(). But there are two versions of 
setup_per_cpu_areas(): one for SMP and one for !SMP. And the !SMP case 
never calls pcpu_free_alloc_info() currently.

I'm not sure i understand that code fully, but maybe the following patch 
could be a better fit:

----- >8
Subject: [PATCH] percpu: don't forget to free the temporary struct pcpu_alloc_info

Unlike the SMP case, the !SMP case does not free the memory for struct 
pcpu_alloc_info allocated in setup_per_cpu_areas(). And to give it a 
chance of being reused by the page allocator later, align it to a page 
boundary just like its size.

Signed-off-by: Nicolas Pitre <nico@linaro.org>

diff --git a/mm/percpu.c b/mm/percpu.c
index 434844415d..caab63375b 100644
--- a/mm/percpu.c
+++ b/mm/percpu.c
@@ -1416,7 +1416,7 @@ struct pcpu_alloc_info * __init pcpu_alloc_alloc_info(int nr_groups,
 			  __alignof__(ai->groups[0].cpu_map[0]));
 	ai_size = base_size + nr_units * sizeof(ai->groups[0].cpu_map[0]);
 
-	ptr = memblock_virt_alloc_nopanic(PFN_ALIGN(ai_size), 0);
+	ptr = memblock_virt_alloc_nopanic(PFN_ALIGN(ai_size), PAGE_SIZE);
 	if (!ptr)
 		return NULL;
 	ai = ptr;
@@ -2295,6 +2295,7 @@ void __init setup_per_cpu_areas(void)
 
 	if (pcpu_setup_first_chunk(ai, fc) < 0)
 		panic("Failed to initialize percpu areas.");
+	pcpu_free_alloc_info(ai);
 }
 
 #endif	/* CONFIG_SMP */

WARNING: multiple messages have this Message-ID (diff)
From: Nicolas Pitre <nicolas.pitre@linaro.org>
To: Tejun Heo <tj@kernel.org>
Cc: Christoph Lameter <cl@linux.com>,
	linux-mm@kvack.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] mm/percpu.c: use smarter memory allocation for struct pcpu_alloc_info
Date: Tue, 3 Oct 2017 18:29:49 -0400 (EDT)	[thread overview]
Message-ID: <nycvar.YSQ.7.76.1710031731130.5407@knanqh.ubzr> (raw)
In-Reply-To: <20171003210540.GM3301751@devbig577.frc2.facebook.com>

On Tue, 3 Oct 2017, Tejun Heo wrote:

> On Tue, Oct 03, 2017 at 04:57:44PM -0400, Nicolas Pitre wrote:
> > This can be much smaller than a page on very small memory systems. 
> > Always rounding up the size to a page is wasteful in that case, and 
> > required alignment is smaller than the memblock default. Let's round 
> > things up to a page size only when the actual size is >= page size, and 
> > then it makes sense to page-align for a nicer allocation pattern.
> 
> Isn't that a temporary area which gets freed later during boot?

Hmmm...

It may get freed through 3 different paths where 2 of them are error 
paths. What looks like a non-error path is in pcpu_embed_first_chunk() 
called from setup_per_cpu_areas(). But there are two versions of 
setup_per_cpu_areas(): one for SMP and one for !SMP. And the !SMP case 
never calls pcpu_free_alloc_info() currently.

I'm not sure i understand that code fully, but maybe the following patch 
could be a better fit:

----- >8
Subject: [PATCH] percpu: don't forget to free the temporary struct pcpu_alloc_info

Unlike the SMP case, the !SMP case does not free the memory for struct 
pcpu_alloc_info allocated in setup_per_cpu_areas(). And to give it a 
chance of being reused by the page allocator later, align it to a page 
boundary just like its size.

Signed-off-by: Nicolas Pitre <nico@linaro.org>

diff --git a/mm/percpu.c b/mm/percpu.c
index 434844415d..caab63375b 100644
--- a/mm/percpu.c
+++ b/mm/percpu.c
@@ -1416,7 +1416,7 @@ struct pcpu_alloc_info * __init pcpu_alloc_alloc_info(int nr_groups,
 			  __alignof__(ai->groups[0].cpu_map[0]));
 	ai_size = base_size + nr_units * sizeof(ai->groups[0].cpu_map[0]);
 
-	ptr = memblock_virt_alloc_nopanic(PFN_ALIGN(ai_size), 0);
+	ptr = memblock_virt_alloc_nopanic(PFN_ALIGN(ai_size), PAGE_SIZE);
 	if (!ptr)
 		return NULL;
 	ai = ptr;
@@ -2295,6 +2295,7 @@ void __init setup_per_cpu_areas(void)
 
 	if (pcpu_setup_first_chunk(ai, fc) < 0)
 		panic("Failed to initialize percpu areas.");
+	pcpu_free_alloc_info(ai);
 }
 
 #endif	/* CONFIG_SMP */

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  reply	other threads:[~2017-10-03 22:29 UTC|newest]

Thread overview: 58+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-10-03 20:57 [PATCH] mm/percpu.c: use smarter memory allocation for struct pcpu_alloc_info Nicolas Pitre
2017-10-03 20:57 ` Nicolas Pitre
2017-10-03 21:05 ` Tejun Heo
2017-10-03 21:05   ` Tejun Heo
2017-10-03 22:29   ` Nicolas Pitre [this message]
2017-10-03 22:29     ` Nicolas Pitre
2017-10-03 22:36     ` Tejun Heo
2017-10-03 22:36       ` Tejun Heo
2017-10-03 23:48       ` Dennis Zhou
2017-10-03 23:48         ` Dennis Zhou
2017-10-04  0:13         ` Nicolas Pitre
2017-10-04  0:13           ` Nicolas Pitre
2017-10-04 14:15     ` Tejun Heo
2017-10-04 14:15       ` Tejun Heo
2017-11-18 18:25     ` mm/percpu.c: use smarter memory allocation for struct pcpu_alloc_info (crisv32 hang) Guenter Roeck
2017-11-18 18:25       ` Guenter Roeck
2017-11-19 20:36       ` Nicolas Pitre
2017-11-19 20:36         ` Nicolas Pitre
2017-11-20  2:03         ` Guenter Roeck
2017-11-20  2:03           ` Guenter Roeck
2017-11-20  4:08           ` Nicolas Pitre
2017-11-20  4:08             ` Nicolas Pitre
2017-11-20  5:05             ` Guenter Roeck
2017-11-20  5:05               ` Guenter Roeck
2017-11-20 18:18               ` Nicolas Pitre
2017-11-20 18:18                 ` Nicolas Pitre
2017-11-20 18:51                 ` Guenter Roeck
2017-11-20 18:51                   ` Guenter Roeck
2017-11-20 20:21                   ` Nicolas Pitre
2017-11-20 20:21                     ` Nicolas Pitre
2017-11-20 21:11                     ` Guenter Roeck
2017-11-20 21:11                       ` Guenter Roeck
2017-11-21  0:28                       ` Nicolas Pitre
2017-11-21  0:28                         ` Nicolas Pitre
2017-11-21  1:48                         ` Guenter Roeck
2017-11-21  1:48                           ` Guenter Roeck
2017-11-21  3:50                           ` Nicolas Pitre
2017-11-21  3:50                             ` Nicolas Pitre
2017-11-22 15:34                             ` Jesper Nilsson
2017-11-22 15:34                               ` Jesper Nilsson
2017-11-22 20:17                               ` Nicolas Pitre
2017-11-22 20:17                                 ` Nicolas Pitre
2017-11-23  7:56                                 ` Jesper Nilsson
2017-11-23  7:56                                   ` Jesper Nilsson
2017-11-27 19:41       ` Tejun Heo
2017-11-27 19:41         ` Tejun Heo
2017-11-27 20:31         ` Nicolas Pitre
2017-11-27 20:31           ` Nicolas Pitre
2017-11-27 20:33           ` Tejun Heo
2017-11-27 20:33             ` Tejun Heo
2017-11-27 20:51             ` Nicolas Pitre
2017-11-27 20:51               ` Nicolas Pitre
2017-11-27 20:54               ` Tejun Heo
2017-11-27 20:54                 ` Tejun Heo
2017-11-27 21:11                 ` Guenter Roeck
2017-11-27 21:11                   ` Guenter Roeck
2017-11-28  8:19               ` Jesper Nilsson
2017-11-28  8:19                 ` Jesper Nilsson

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=nycvar.YSQ.7.76.1710031731130.5407@knanqh.ubzr \
    --to=nicolas.pitre@linaro.org \
    --cc=cl@linux.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=tj@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.