4.8-rc4 spews "BUG: sleeping function called from invalid context at fs/dcache.c:757"

4.8-rc4 spews "BUG: sleeping function called from invalid context at fs/dcache.c:757"

[Takashi Iwai]

Hi,

I noticed that a freshly booted 4.8-rc4 kernel gives the following
kernel messages a few times:

 BUG: sleeping function called from invalid context at fs/dcache.c:757
 in_atomic(): 1, irqs_disabled(): 0, pid: 8431, name: automount
 1 lock held by automount/8431:
  #0:  (&(&sbi->fs_lock)->rlock){+.+...}, at: [<ffffffff81345736>] autofs4_expire_indirect+0x136/0x2d0
 CPU: 4 PID: 8431 Comm: automount Not tainted 4.8.0-rc4-test2+ #210
 Hardware name: Dell Inc. OptiPlex 9010/0M9KCM, BIOS A12 01/10/2013
  0000000000000000 ffff8801843efc30 ffffffff813bbdd3 ffff88019e6500c0
  ffffffff81a330b6 ffff8801843efc58 ffffffff810a3052 ffffffff81a330b6
  00000000000002f5 0000000000000000 ffff8801843efc80 ffffffff810a3149
 Call Trace:
  [<ffffffff813bbdd3>] dump_stack+0x86/0xc3
  [<ffffffff810a3052>] ___might_sleep+0x142/0x1f0
  [<ffffffff810a3149>] __might_sleep+0x49/0x80
  [<ffffffff81260b17>] ? dput+0x37/0x400
  [<ffffffff81260b26>] dput+0x46/0x400
  [<ffffffff8126c96d>] ? lookup_mnt+0x14d/0x1d0
  [<ffffffff8124ff54>] ? follow_down_one+0x14/0x60
  [<ffffffff8124ff67>] follow_down_one+0x27/0x60
  [<ffffffff81344da2>] autofs4_mount_busy+0x32/0x110
  [<ffffffff81345081>] should_expire+0x51/0x3d0
  [<ffffffff81345790>] autofs4_expire_indirect+0x190/0x2d0
  [<ffffffff81345fa0>] ? autofs_dev_ioctl_askumount+0x30/0x30
  [<ffffffff81345bb4>] autofs4_do_expire_multi+0x94/0xb0
  [<ffffffff81345fa0>] ? autofs_dev_ioctl_askumount+0x30/0x30
  [<ffffffff81345fbf>] autofs_dev_ioctl_expire+0x1f/0x30
  [<ffffffff81346679>] _autofs_dev_ioctl+0x2b9/0x350
  [<ffffffff81346723>] autofs_dev_ioctl+0x13/0x20
  [<ffffffff8125af54>] do_vfs_ioctl+0x94/0x690
  [<ffffffff8126820d>] ? __fget+0x10d/0x200
  [<ffffffff81268105>] ? __fget+0x5/0x200
  [<ffffffff8125b5c9>] SyS_ioctl+0x79/0x90
  [<ffffffff8170ef00>] entry_SYSCALL_64_fastpath+0x23/0xc1


I don't remember of a similar stack trace in the past, so if any, it
can be a regression in 4.8 kernel.  But I cannot say it in 100%, as
this looks spontaneous, nor I would be able to reproduce it at the
next boot...


thanks,

Takashi

Re: 4.8-rc4 spews "BUG: sleeping function called from invalid context at fs/dcache.c:757"

[Al Viro]

On Mon, Aug 29, 2016 at 04:35:46PM +0200, Takashi Iwai wrote:

>   [<ffffffff81260b26>] dput+0x46/0x400
	... which should not be called in atomic contexts
>   [<ffffffff8124ff67>] follow_down_one+0x27/0x60
	... and neither should this
>   [<ffffffff81344da2>] autofs4_mount_busy+0x32/0x110
	... nor that (for fsck sake, there's full-blown path_put() in it!)
>   [<ffffffff81345081>] should_expire+0x51/0x3d0
	... so that would better not be called in atomic either (incidentally,
it also calls dput() directly)
>   [<ffffffff81345790>] autofs4_expire_indirect+0x190/0x2d0
	... while here it is called under sbi->fs_lock.

> I don't remember of a similar stack trace in the past, so if any, it
> can be a regression in 4.8 kernel.  But I cannot say it in 100%, as
> this looks spontaneous, nor I would be able to reproduce it at the
> next boot...

It's old; the race is narrow, but it's been there for quite a while, by
the look of it.

Re: 4.8-rc4 spews "BUG: sleeping function called from invalid context at fs/dcache.c:757"

[Ian Kent]

On Mon, 2016-08-29 at 16:18 +0100, Al Viro wrote:
> On Mon, Aug 29, 2016 at 04:35:46PM +0200, Takashi Iwai wrote:
> 
> >   [<ffffffff81260b26>] dput+0x46/0x400
> 	... which should not be called in atomic contexts
> >   [<ffffffff8124ff67>] follow_down_one+0x27/0x60
> 	... and neither should this
> >   [<ffffffff81344da2>] autofs4_mount_busy+0x32/0x110
> 	... nor that (for fsck sake, there's full-blown path_put() in it!)
> >   [<ffffffff81345081>] should_expire+0x51/0x3d0
> 	... so that would better not be called in atomic either (incidentally,
> it also calls dput() directly)
> >   [<ffffffff81345790>] autofs4_expire_indirect+0x190/0x2d0
> 	... while here it is called under sbi->fs_lock.
> 
> > I don't remember of a similar stack trace in the past, so if any, it
> > can be a regression in 4.8 kernel.  But I cannot say it in 100%, as
> > this looks spontaneous, nor I would be able to reproduce it at the
> > next boot...
> 
> It's old; the race is narrow, but it's been there for quite a while, by
> the look of it.

Right, I missed that when the rcu-walk concurrency changes went in, mmm ....

Re: 4.8-rc4 spews "BUG: sleeping function called from invalid context at fs/dcache.c:757"

[Ian Kent]

On Tue, 2016-08-30 at 09:37 +0800, Ian Kent wrote:
> On Mon, 2016-08-29 at 16:18 +0100, Al Viro wrote:
> > On Mon, Aug 29, 2016 at 04:35:46PM +0200, Takashi Iwai wrote:
> > 
> > >   [<ffffffff81260b26>] dput+0x46/0x400
> > 	... which should not be called in atomic contexts
> > >   [<ffffffff8124ff67>] follow_down_one+0x27/0x60
> > 	... and neither should this
> > >   [<ffffffff81344da2>] autofs4_mount_busy+0x32/0x110
> > 	... nor that (for fsck sake, there's full-blown path_put() in it!)
> > >   [<ffffffff81345081>] should_expire+0x51/0x3d0
> > 	... so that would better not be called in atomic either (incidentally,
> > it also calls dput() directly)
> > >   [<ffffffff81345790>] autofs4_expire_indirect+0x190/0x2d0
> > 	... while here it is called under sbi->fs_lock.
> > 
> > > I don't remember of a similar stack trace in the past, so if any, it
> > > can be a regression in 4.8 kernel.  But I cannot say it in 100%, as
> > > this looks spontaneous, nor I would be able to reproduce it at the
> > > next boot...
> > 
> > It's old; the race is narrow, but it's been there for quite a while, by
> > the look of it.
> 
> Right, I missed that when the rcu-walk concurrency changes went in, mmm ....

Umm ... no, the problem has been there much longer than that ...