Re: problem with nilfs_cleanerd

[Ryusuke Konishi <ryusuke-sG5X7nlA6pw@public.gmane.org>]

On Wed, 07 Jul 2010 08:25:22 +0200, Łukasz_Wójcicki wrote:
> Does this patch eliminate the problem?

No, it just detects a likely cause.
I think the root cause exists somewhere else.

> Unfortunately I can not generate this error again because after
> re-formatting the memory (mkfs.nilfs2) the problem disappeared.

All right, I'll try to fix the problem based on the review of btree
code, or at least will include a patch to help debugging this.

Thanks,
Ryusuke Konishi

> Dnia 2010-07-07, śro o godzinie 15:02 +0900, Ryusuke Konishi pisze:
> > Hi,
> > 
> > On Tue, 06 Jul 2010 11:02:41 +0200, Łukasz_Wójcicki wrote:
> > > I'am using kernel 2.6.20 with nilf_utils-2-0-18 and nilfs kernel module
> > > 2-0-20 .
> > > 
> > > <1>Unable to handle kernel NULL pointer dereference at virtual address
> > > 00000000
> > > <1>pgd = c4eb0000
> > > <1>[00000000] *pgd=c4e2d031, *pte=00000000, *ppte=00000000
> > > <4>Internal error: Oops: 17 [#1]
> > > <4>Modules linked in:
> > > <4>CPU: 0
> > > <4>PC is at nilfs_btree_lookup_dirty_buffers+0x10c/0x1ec
> > > <4>LR is at 0xc4c99b24
> > <snip>
> > 
> > Could you try the following patch to narrow down the cause?
> > 
> > There is a function lacking range check in the btree code.  This patch
> > will catch the defect if the oops comes from a range error.
> > 
> > I'm now reviewing the btree code to find out root cause of the range
> > error.
> > 
> > Thanks,
> > Ryusuke Konishi
> > ---
> > 
> > diff --git a/fs/btree.c b/fs/btree.c
> > index beb11fe..9622ff4 100644
> > --- a/fs/btree.c
> > +++ b/fs/btree.c
> > @@ -2007,6 +2007,15 @@ static void nilfs_btree_add_dirty_buffer(struct nilfs_btree *btree,
> >  	node = (struct nilfs_btree_node *)bh->b_data;
> >  	key = nilfs_btree_node_get_key(node, 0);
> >  	level = nilfs_btree_node_get_level(node);
> > +	if (level < NILFS_BTREE_LEVEL_NODE_MIN ||
> > +	    level >= NILFS_BTREE_LEVEL_MAX) {
> > +		dump_stack();
> > +		printk(KERN_CRIT
> > +		       "%s: invalid btree level: %d (ino=%lu)\n", __func__,
> > +		       level, NILFS_BMAP_I(&btree->bt_bmap)->vfs_inode.i_ino);
> > +		return;
> > +	}
> > +
> >  	list_for_each(head, &lists[level]) {
> >  		cbh = list_entry(head, struct buffer_head, b_assoc_buffers);
> >  		cnode = (struct nilfs_btree_node *)cbh->b_data;
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-nilfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html