linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: yoma sophian <sophian.yoma@gmail.com>
To: Michal Hocko <mhocko@kernel.org>
Cc: linux-mm@kvack.org
Subject: Re: page order 0 allocation fail but free pages are enough
Date: Mon, 23 May 2016 23:31:09 +0800	[thread overview]
Message-ID: <CADUS3onEpdMF6Pi9-cHkf+hA6bqOc4mkXAci7ikeUhtaELx4WQ@mail.gmail.com> (raw)
In-Reply-To: <20160523144711.GV2278@dhcp22.suse.cz>

hi Michal

2016-05-23 22:47 GMT+08:00, Michal Hocko <mhocko@kernel.org>:
> On Mon 23-05-16 14:47:51, yoma sophian wrote:
>> hi all:
>> I got something wired that
>> 1. in softirq, there is a page order 0 allocation request
>> 2. Normal/High zone are free enough for order 0 page.
>> 3. but somehow kernel return order 0 allocation fail.
>>
>> My kernel version is 3.10 and below is kernel log:
>> from memory info,
>
> Can you reproduce it with the current vanlilla tree?
I think it would be quite hard, since this allocation failuer comes
when a lot of program, such as Youtube, opera, etc. running on ARM
processor at the same time.
Or is there any patch in vanlilla tree I can used for checking?

>
> [...]
>> [   94.586588] ksoftirqd/0: page allocation failure: order:0, mode:0x20
> [...]
>> [   94.865776] Normal free:63768kB min:2000kB low:2500kB high:3000kB
> [...]
>> [ 8606.701343] CompositorTileW: page allocation failure: order:0,
>> mode:0x20
> [...]
>> [ 8606.703590] Normal free:60684kB min:2000kB low:2500kB high:3000kB
>
> This is a lot of free memory to block GFP_ATOMIC. One possible
> explanation would be that this is a race with somebody releasing a lot
I will try to add memory free at buffered_rmqueue like below xxx place
buffered_rmqueue -->
               if (likely(order == 0)) {
                ..................
                if (list_empty(list)) {
                        pcp->count += rmqueue_bulk(zone, 0,
                                        pcp->batch, list,
                                        migratetype, cold);
                        if (unlikely(list_empty(list)))
                                goto failed;    xxxxx ==>  to show
memory free info
                }


> of memory. The free memory is surprisingly similar in both cases.
Would you please give me any clue that free memory silimar hint the
race condition happen?

Many Appreciate your kind suggestion,

--
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:[~2016-05-23 15:31 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-05-23  6:47 yoma sophian
2016-05-23 14:47 ` Michal Hocko
2016-05-23 15:31   ` yoma sophian [this message]
2016-05-23 19:00     ` Michal Hocko
2016-05-24 11:40       ` yoma sophian
2016-05-24 11:50         ` Michal Hocko
2016-05-24 15:06           ` yoma sophian
2016-05-25  2:11             ` yoma sophian
2016-05-25  6:31               ` Michal Hocko

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=CADUS3onEpdMF6Pi9-cHkf+hA6bqOc4mkXAci7ikeUhtaELx4WQ@mail.gmail.com \
    --to=sophian.yoma@gmail.com \
    --cc=linux-mm@kvack.org \
    --cc=mhocko@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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox