XFS kernel BUG at fs/buffer.c:470! with 2.6.28.4

[Eric Sandeen]

Alessandro Bono wrote:
> On Sun, 2009-02-08 at 17:42 -0500, Christoph Hellwig wrote:
>> On Sun, Feb 08, 2009 at 11:39:12PM +0100, Alessandro Bono wrote:
>>> On Sun, 2009-02-08 at 17:28 -0500, Christoph Hellwig wrote:
>>>> On Sat, Feb 07, 2009 at 02:06:13PM +0100, Alessandro Bono wrote:
>>>>> Feb  7 12:43:12 champagne kernel: [ 5828.167041] ------------[ cut
>>>>> here ]------------
>>>>> Feb  7 12:43:12 champagne kernel: [ 5828.167048] kernel BUG at
>>>>> fs/buffer.c:470!
>>>> Per
>>>> http://git.kernel.org/?p=linux/kernel/git/hpa/linux-2.6-allstable.git;a=blob;f=fs/buffer.c;h=665d446b25bc034241ef54c3c6b1d239c0ccf0f9;hb=d5b562330ec766292a3ac54ae5e0673610bd5b3d
>>>>
>>>> line 470 in fs/buffer.c of 2.6.28.4 has a comment and no actual code.
>>>>
>>>> What additional patches do you have applied?
>>>>
>>> vanilla kernel
>>> no additional patches at all
>> Well, the 2.6.28.4 clearly doesn't have a bug there.  Can you
>> attach the fs/buffer.c you built the kernel from?
>>
> 
> sure, attached

Well, that seems to not be from the kernel you were running; there is no
BUG() on line 470:

$ cat -n buffer.c | grep -8 "   470"
   462				       bdevname(bh->b_bdev, b));
   463			}
   464			set_bit(AS_EIO, &page->mapping->flags);
   465			set_buffer_write_io_error(bh);
   466			clear_buffer_uptodate(bh);
   467			SetPageError(page);
   468		}
   469	
   470		first = page_buffers(page);
   471		local_irq_save(flags);
   472		bit_spin_lock(BH_Uptodate_Lock, &first->b_state);
   473	
   474		clear_buffer_async_write(bh);
   475		unlock_buffer(bh);
   476		tmp = bh->b_this_page;
   477		while (tmp != bh) {
   478			if (buffer_async_write(tmp)) {