From: wang Yu <yuwang668899@gmail.com>
To: mhocko@suse.com, penguin-kernel@i-love.sakura.ne.jp, linux-mm@kvack.org
Cc: chenggang.qcg@alibaba-inc.com, yuwang.yuwang@alibaba-inc.com
Subject: [PATCH] mm,page_alloc: softlockup on warn_alloc on
Date: Fri, 15 Sep 2017 17:58:49 +0800 [thread overview]
Message-ID: <20170915095849.9927-1-yuwang668899@gmail.com> (raw)
From: "yuwang.yuwang" <yuwang.yuwang@alibaba-inc.com>
I found a softlockup when running some stress testcase in 4.9.x,
but i think the mainline have the same problem.
call trace:
[365724.502896] NMI watchdog: BUG: soft lockup - CPU#31 stuck for 22s!
[jbd2/sda3-8:1164]
...
...
[365724.503258] Call Trace:
[365724.503260] [<ffffffff811ace5f>] warn_alloc+0x13f/0x170
[365724.503264] [<ffffffff811ad8c2>] __alloc_pages_slowpath+0x9b2/0xc10
[365724.503265] [<ffffffff811add43>] __alloc_pages_nodemask+0x223/0x2a0
[365724.503268] [<ffffffff811fe838>] alloc_pages_current+0x88/0x120
[365724.503270] [<ffffffff811a3644>] __page_cache_alloc+0xb4/0xc0
[365724.503272] [<ffffffff811a49e9>] pagecache_get_page+0x59/0x230
[365724.503275] [<ffffffff8126b2db>] __getblk_gfp+0xfb/0x2f0
[365724.503281] [<ffffffffa00f9cee>]
jbd2_journal_get_descriptor_buffer+0x5e/0xe0 [jbd2]
[365724.503286] [<ffffffffa00f2a01>]
jbd2_journal_commit_transaction+0x901/0x1880 [jbd2]
[365724.503291] [<ffffffff8102d6a5>] ? __switch_to+0x215/0x730
[365724.503294] [<ffffffff810f962d>] ? lock_timer_base+0x7d/0xa0
[365724.503298] [<ffffffffa00f7cda>] kjournald2+0xca/0x260 [jbd2]
[365724.503300] [<ffffffff810cfb00>] ? prepare_to_wait_event+0xf0/0xf0
[365724.503304] [<ffffffffa00f7c10>] ? commit_timeout+0x10/0x10 [jbd2]
[365724.503307] [<ffffffff810a8d66>] kthread+0xe6/0x100
[365724.503309] [<ffffffff810a8c80>] ? kthread_park+0x60/0x60
[365724.503313] [<ffffffff816f3795>] ret_from_fork+0x25/0x30
we can limit the warn_alloc caller to workaround it.
__alloc_pages_slowpath only call once warn_alloc each time.
Signed-off-by: yuwang.yuwang <yuwang.yuwang@alibaba-inc.com>
Suggested-by: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
---
mm/page_alloc.c | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 2abf8d5..8b86686 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -3525,6 +3525,7 @@ bool gfp_pfmemalloc_allowed(gfp_t gfp_mask)
unsigned long alloc_start = jiffies;
unsigned int stall_timeout = 10 * HZ;
unsigned int cpuset_mems_cookie;
+ static unsigned long stall_warn_lock;
/*
* In the slowpath, we sanity check order to avoid ever trying to
@@ -3698,11 +3699,13 @@ bool gfp_pfmemalloc_allowed(gfp_t gfp_mask)
goto nopage;
/* Make sure we know about allocations which stall for too long */
- if (time_after(jiffies, alloc_start + stall_timeout)) {
+ if (time_after(jiffies, alloc_start + stall_timeout) &&
+ !test_and_set_bit_lock(0, &stall_warn_lock)) {
warn_alloc(gfp_mask,
"page allocation stalls for %ums, order:%u",
jiffies_to_msecs(jiffies-alloc_start), order);
- stall_timeout += 10 * HZ;
+ stall_timeout = jiffies - alloc_start + 10 * HZ;
+ clear_bit_unlock(0, &stall_warn_lock);
}
if (should_reclaim_retry(gfp_mask, order, ac, alloc_flags,
--
1.8.3.1
--
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>
next reply other threads:[~2017-09-15 9:59 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-09-15 9:58 wang Yu [this message]
2017-09-15 10:39 ` [PATCH] mm,page_alloc: softlockup on warn_alloc on Michal Hocko
2017-09-15 11:38 ` Tetsuo Handa
2017-09-15 12:00 ` Michal Hocko
2017-09-15 12:09 ` Tetsuo Handa
2017-09-15 12:14 ` Michal Hocko
2017-09-15 14:12 ` Tetsuo Handa
2017-09-15 14:23 ` Michal Hocko
2017-09-24 1:56 ` Tetsuo Handa
2017-09-15 14:37 ` Johannes Weiner
2017-09-15 15:23 ` Tetsuo Handa
2017-09-15 18:44 ` Johannes Weiner
2017-09-16 0:25 ` Tetsuo Handa
2017-09-18 6:05 ` Michal Hocko
2017-09-18 6:31 ` Tetsuo Handa
2017-09-18 6:43 ` Michal Hocko
2017-09-16 4:12 ` Tetsuo Handa
2017-10-11 11:14 ` Tetsuo Handa
2017-10-18 10:54 ` Tetsuo Handa
2017-09-18 6:03 ` 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=20170915095849.9927-1-yuwang668899@gmail.com \
--to=yuwang668899@gmail.com \
--cc=chenggang.qcg@alibaba-inc.com \
--cc=linux-mm@kvack.org \
--cc=mhocko@suse.com \
--cc=penguin-kernel@i-love.sakura.ne.jp \
--cc=yuwang.yuwang@alibaba-inc.com \
/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;
as well as URLs for NNTP newsgroup(s).