From: Mel Gorman <mgorman@suse.de>
To: Jovi Zhangwei <jovi@cloudflare.com>
Cc: linux-kernel@vger.kernel.org, sasha.levin@oracle.com,
n-horiguchi@ah.jp.nec.com, akpm@linux-foundation.org,
Hugh Dickins <hughd@google.com>,
linux-mm@kvack.org, vbabka@suse.cz, rientjes@google.com
Subject: Re: kernel bug(VM_BUG_ON_PAGE) with 3.18.13 in mm/migrate.c
Date: Tue, 2 Jun 2015 08:19:41 +0100 [thread overview]
Message-ID: <20150602071941.GB26425@suse.de> (raw)
In-Reply-To: <CABPcSq+5SR0vqs6fGOwKJ0AZMiLSDQ6Rsevi2wB4YgZPJ9iadg@mail.gmail.com>
On Thu, May 28, 2015 at 11:38:36AM -0700, Jovi Zhangwei wrote:
> Hi Mel,
>
> On Thu, May 28, 2015 at 5:00 AM, Mel Gorman <mgorman@suse.de> wrote:
> > On Wed, May 27, 2015 at 11:05:33AM -0700, Jovi Zhangwei wrote:
> >> Hi,
> >>
> >> I got below kernel bug error in our 3.18.13 stable kernel.
> >> "kernel BUG at mm/migrate.c:1661!"
> >>
> >> Source code:
> >>
> >> 1657 static int numamigrate_isolate_page(pg_data_t *pgdat, struct page *page)
> >> 1658 {
> >> 1659 int page_lru;
> >> 1660
> >> 1661 VM_BUG_ON_PAGE(compound_order(page) &&
> >> !PageTransHuge(page), page);
> >>
> >> It's easy to trigger the error by run tcpdump in our system.(not sure
> >> it will easily be reproduced in another system)
> >> "sudo tcpdump -i bond0.100 'tcp port 4242' -c 100000000000 -w 4242.pcap"
> >>
> >> Any comments for this bug would be great appreciated. thanks.
> >>
> >
> > What sort of compound page is it? What sort of VMA is it in? hugetlbfs
> > pages should never be tagged for NUMA migrate and never enter this
> > path. Transparent huge pages are handled properly so I'm wondering
> > exactly what type of compound page this is and what mapped it into
> > userspace.
> >
> Thanks for your reply.
>
> After reading net/packet/af_packet.c:alloc_one_pg_vec_page, I found
> there indeed have compound page maped into userspace.
>
Ok, it's clear now. Thanks very much.
> I sent a patch for this issue(you may received it), but not sure it's
> right to fix,
> feel free to update it or use your own patch.
>
It avoids the problem but it's not the best fix because a lot of useless
overhead has been incurred for a page that can never be migrated. Can you
try the following instead please?
---8<---
sched, numa: Do not hint for NUMA balancing on VM_MIXEDMAP mappings
Jovi Zhangwei reported the following problem
Below kernel vm bug can be triggered by tcpdump which mmaped a lot of pages
with GFP_COMP flag.
[Mon May 25 05:29:33 2015] page:ffffea0015414000 count:66 mapcount:1 mapping: (null) index:0x0
[Mon May 25 05:29:33 2015] flags: 0x20047580004000(head)
[Mon May 25 05:29:33 2015] page dumped because: VM_BUG_ON_PAGE(compound_order(page) && !PageTransHuge(page))
[Mon May 25 05:29:33 2015] ------------[ cut here ]------------
[Mon May 25 05:29:33 2015] kernel BUG at mm/migrate.c:1661!
[Mon May 25 05:29:33 2015] invalid opcode: 0000 [#1] SMP
Compound pages cannot be migrated and it was not expected that such pages
be marked for NUMA balancing. This did not take into account that drivers
such as net/packet/af_packet.c may insert compound pages into userspace
with vm_insert_page. This patch tells the NUMA balancing protection scanner
to skip all VM_MIXEDMAP mappings which avoids the possibility that compound
pages are marked for migration.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reported-by: Jovi Zhangwei <jovi@cloudflare.com>
---
kernel/sched/fair.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 241213be507c..486d00c408b0 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -2166,7 +2166,7 @@ void task_numa_work(struct callback_head *work)
}
for (; vma; vma = vma->vm_next) {
if (!vma_migratable(vma) || !vma_policy_mof(vma) ||
- is_vm_hugetlb_page(vma)) {
+ is_vm_hugetlb_page(vma) || (vma->vm_flags & VM_MIXEDMAP)) {
continue;
}
next prev parent reply other threads:[~2015-06-02 7:19 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-05-27 18:05 kernel bug(VM_BUG_ON_PAGE) with 3.18.13 in mm/migrate.c Jovi Zhangwei
2015-05-27 18:42 ` Jovi Zhangwei
2015-05-28 12:00 ` Mel Gorman
2015-05-28 18:38 ` Jovi Zhangwei
2015-05-31 1:39 ` Jovi Zhangwei
2015-06-02 7:19 ` Mel Gorman [this message]
-- strict thread matches above, loose matches on Subject: below --
2015-05-26 23:44 Jovi Zhangwei
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=20150602071941.GB26425@suse.de \
--to=mgorman@suse.de \
--cc=akpm@linux-foundation.org \
--cc=hughd@google.com \
--cc=jovi@cloudflare.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=n-horiguchi@ah.jp.nec.com \
--cc=rientjes@google.com \
--cc=sasha.levin@oracle.com \
--cc=vbabka@suse.cz \
/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).