Re: BUG: unable to handle kernel paging request

[Stephen Wilson]

On Mon, Mar 28, 2011 at 12:00:44AM +0800, AmÃrico Wang wrote:
> On Sun, Mar 27, 2011 at 11:37 PM, Linus Torvalds
> <torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
> > Al, this smells like your /proc cleanups/fixes...
> >
> > On Sun, Mar 27, 2011 at 2:16 AM, Anca Emanuel <anca.emanuel@xxxxxxxxx> wrote:
> >> Hi, I'm using latest kernel git.
> >>
> >>
> >> [15117.080119] BUG: unable to handle kernel paging request at fffffffffffffff3
> >
> > That's "-13" (possibly -EACCES)
> >
> >> [15117.080152] IP: [<ffffffff811b4989>] vma_stop+0x19/0x40
> >
> > .. and the code disassembles to
> >
> >  0:  55           Âpush  %rbp
> >  1:  48 89 e5        Âmov  Â%rsp,%rbp
> >  4:  53           Âpush  %rbx
> >  5:  48 83 ec 08       sub  Â$0x8,%rsp
> >  9:  0f 1f 44 00 00     Ânopl  0x0(%rax,%rax,1)
> >  e:  48 85 f6        Âtest  %rsi,%rsi
> > Â11:  74 1a          je   0x2d
> > Â13:  48 39 77 10       cmp  Â%rsi,0x10(%rdi)
> > Â17:  74 14          je   0x2d
> > Â19:  8b 1e          mov  Â(%rsi),%ebx
> > Â1b:  48 8d 7b 60       lea  Â0x60(%rbx),%rdi
> > Â1f:  e8 eb 2f ed ff     Âcallq Âup_read
> >
> > where that instruction at 0x19 is the access "mm = vma->vm_mm". So
> > it's vma that is -EPERM.
> >
> > I bet it's due to commit ec6fd8a4355c ("report errors in /proc/*/*map*
> > sanely"), which replaces NULL with various ERR_PTR() cases.
> >
> 
> Exactly... should be fixed by something like:
> 
> diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c
> index 7c708a4..6b82632 100644
> --- a/fs/proc/task_mmu.c
> +++ b/fs/proc/task_mmu.c
> @@ -90,7 +90,7 @@ static void pad_len_spaces(struct seq_file *m, int len)
> 
>  static void vma_stop(struct proc_maps_private *priv, struct
> vm_area_struct *vma)
>  {
> -       if (vma && vma != priv->tail_vma) {
> +       if (vma && !IS_ERR(vma) && vma != priv->tail_vma) {
>                 struct mm_struct *mm = vma->vm_mm;
>                 up_read(&mm->mmap_sem);
>                 mmput(mm);

FWIW, that looks like the right fix to me.

Also CC'ing Tony Luck as he reported what appears to be the same issue
on the 25'th.

-- 
steve

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/