From: Oleg Nesterov <oleg@redhat.com>
To: Kefeng Wang <wangkefeng.wang@huawei.com>
Cc: Michal Hocko <mhocko@suse.com>, linux-mm <linux-mm@kvack.org>,
Andrea Arcangeli <aarcange@redhat.com>,
Peter Xu <peterx@redhat.com>, Mike Rapoport <rppt@linux.ibm.com>,
Jann Horn <jannh@google.com>, Jason Gunthorpe <jgg@mellanox.com>,
Andrew Morton <akpm@linux-foundation.org>
Subject: Re: [BUG] kernel BUG at fs/userfaultfd.c:385 after 04f5866e41fb
Date: Wed, 14 Aug 2019 17:41:02 +0200 [thread overview]
Message-ID: <20190814154101.GF11595@redhat.com> (raw)
In-Reply-To: <20190814151049.GD11595@redhat.com>
On 08/14, Oleg Nesterov wrote:
>
> On 08/14, Kefeng Wang wrote:
> >
> > On 2019/8/14 21:53, Michal Hocko wrote:
> > > On Tue 13-08-19 17:08:05, Kefeng Wang wrote:
> > >>
> > >> Syzkaller reproducer:
> > >> # {Threaded:true Collide:true Repeat:false RepeatTimes:0 Procs:1 Sandbox:none Fault:false FaultCall:-1 FaultNth:0 EnableTun:true EnableNetDev:true EnableNetReset:false EnableCgroups:false EnableBinfmtMisc:true EnableCloseFds:true UseTmpDir:true HandleSegv:true Repro:false Trace:false}
> > >> r0 = userfaultfd(0x80800)
> > >> ioctl$UFFDIO_API(r0, 0xc018aa3f, &(0x7f0000000200))
> > >> ioctl$UFFDIO_REGISTER(r0, 0xc020aa00, &(0x7f0000000080)={{&(0x7f0000ff2000/0xe000)=nil, 0xe000}, 0x1})
> > >> ioctl$UFFDIO_COPY(r0, 0xc028aa03, 0x0)
> > >> ioctl$UFFDIO_COPY(r0, 0xc028aa03, &(0x7f0000000000)={&(0x7f0000ffc000/0x3000)=nil, &(0x7f0000ffd000/0x2000)=nil, 0x3000})
> > >> syz_execute_func(&(0x7f00000000c0)="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")
> > >> poll(&(0x7f00000000c0)=[{}], 0x1, 0x0)
> > >
> > > Is there any way to decypher the above?
> >
> > no, I also want to know the way :(
>
> perhaps you can run it under strace?
>
> I am wondering if "goto skip_mm" in userfaultfd_release() is correct...
> shouldn't it clear VM_UFFD_* and reset vm_userfaultfd_ctx.ctx even if
> !mmget_still_valid ?
Heh, I didn't notice you too mentioned userfaultfd_release() in your email.
can you try the patch below?
Oleg.
--- a/fs/userfaultfd.c
+++ b/fs/userfaultfd.c
@@ -880,6 +880,7 @@ static int userfaultfd_release(struct inode *inode, struct file *file)
/* len == 0 means wake all */
struct userfaultfd_wake_range range = { .len = 0, };
unsigned long new_flags;
+ bool xxx;
WRITE_ONCE(ctx->released, true);
@@ -895,8 +896,7 @@ static int userfaultfd_release(struct inode *inode, struct file *file)
* taking the mmap_sem for writing.
*/
down_write(&mm->mmap_sem);
- if (!mmget_still_valid(mm))
- goto skip_mm;
+ xxx = mmget_still_valid(mm);
prev = NULL;
for (vma = mm->mmap; vma; vma = vma->vm_next) {
cond_resched();
@@ -907,19 +907,20 @@ static int userfaultfd_release(struct inode *inode, struct file *file)
continue;
}
new_flags = vma->vm_flags & ~(VM_UFFD_MISSING | VM_UFFD_WP);
- prev = vma_merge(mm, prev, vma->vm_start, vma->vm_end,
- new_flags, vma->anon_vma,
- vma->vm_file, vma->vm_pgoff,
- vma_policy(vma),
- NULL_VM_UFFD_CTX);
- if (prev)
- vma = prev;
- else
- prev = vma;
+ if (xxx) {
+ prev = vma_merge(mm, prev, vma->vm_start, vma->vm_end,
+ new_flags, vma->anon_vma,
+ vma->vm_file, vma->vm_pgoff,
+ vma_policy(vma),
+ NULL_VM_UFFD_CTX);
+ if (prev)
+ vma = prev;
+ else
+ prev = vma;
+ }
vma->vm_flags = new_flags;
vma->vm_userfaultfd_ctx = NULL_VM_UFFD_CTX;
}
-skip_mm:
up_write(&mm->mmap_sem);
mmput(mm);
wakeup:
next prev parent reply other threads:[~2019-08-14 15:41 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-08-13 9:08 [BUG] kernel BUG at fs/userfaultfd.c:385 after 04f5866e41fb Kefeng Wang
2019-08-14 13:53 ` Michal Hocko
2019-08-14 14:45 ` Kefeng Wang
2019-08-14 15:10 ` Oleg Nesterov
2019-08-14 15:41 ` Oleg Nesterov [this message]
2019-08-15 2:21 ` Kefeng Wang
2019-08-15 9:54 ` Oleg Nesterov
2019-08-16 10:37 ` Kefeng Wang
2019-08-19 12:48 ` Oleg Nesterov
2019-08-19 16:05 ` Andrea Arcangeli
2019-08-20 15:59 ` Oleg Nesterov
2019-08-20 16:15 ` Andrea Arcangeli
[not found] ` <73d7b5b1-a88c-5fca-ba16-be214c2524a4@I-love.SAKURA.ne.jp>
2019-08-20 16:09 ` Oleg Nesterov
2019-08-20 16:02 ` [PATCH] userfaultfd_release: always remove uffd flags and clear vm_userfaultfd_ctx Oleg Nesterov
2019-08-20 16:05 ` Andrea Arcangeli
2019-08-21 0:53 ` Kefeng Wang
2019-08-27 16:33 ` [BUG] kernel BUG at fs/userfaultfd.c:385 after 04f5866e41fb Oleg Nesterov
2019-08-27 17:14 ` Andrea Arcangeli
2019-08-28 14:25 ` Oleg Nesterov
2019-08-29 12:05 ` Andrea Arcangeli
2019-08-30 16:49 ` Oleg Nesterov
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=20190814154101.GF11595@redhat.com \
--to=oleg@redhat.com \
--cc=aarcange@redhat.com \
--cc=akpm@linux-foundation.org \
--cc=jannh@google.com \
--cc=jgg@mellanox.com \
--cc=linux-mm@kvack.org \
--cc=mhocko@suse.com \
--cc=peterx@redhat.com \
--cc=rppt@linux.ibm.com \
--cc=wangkefeng.wang@huawei.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).