lkml.org 
[lkml]   [2011]   [Jun]   [7]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
DateTue, 7 Jun 2011 17:35:27 -0700
FromAndrew Morton <>
SubjectRe: [PATCH 07/15] writeback: split inode_wb_list_lock into bdi_writeback.list_lock
On Wed, 8 Jun 2011 08:20:57 +0800
Wu Fengguang <fengguang.wu@intel.com> wrote:

> On Wed, Jun 08, 2011 at 07:03:19AM +0800, Andrew Morton wrote:
> > On Wed, 08 Jun 2011 05:32:43 +0800
> > Wu Fengguang <fengguang.wu@intel.com> wrote:
> >
> > > static void bdev_inode_switch_bdi(struct inode *inode,
> > > struct backing_dev_info *dst)
> > > {
> > > - spin_lock(&inode_wb_list_lock);
> > > + struct backing_dev_info *old = inode->i_data.backing_dev_info;
> > > +
> > > + if (unlikely(dst == old)) /* deadlock avoidance */
> > > + return;
> >
> > Why does this occur?
>
> That's a fix from Hugh Dickins:

yes, I remember it. And I remember rubberiness about this at the time ;)

> Yesterday's mmotm hangs at startup, and with lockdep it reports:
> BUG: spinlock recursion on CPU#1, blkid/284 - with bdi_lock_two()
> called from bdev_inode_switch_bdi() in the backtrace. It appears
> that this function is sometimes called with new the same as old.
>
> The problem becomes clear when looking at bdi_lock_two(), which will
> immediately deadlock itself if called with (wb1 == wb2):
>
> void bdi_lock_two(struct bdi_writeback *wb1, struct bdi_writeback *wb2)
> {
> if (wb1 < wb2) {
> spin_lock(&wb1->list_lock);
> spin_lock_nested(&wb2->list_lock, 1);
> } else {
> spin_lock(&wb2->list_lock);
> spin_lock_nested(&wb1->list_lock, 1);
> }
> }

But why are we asking bdev_inode_switch_bdi() to switch an inode to a
bdi where it already resides?



\
 
 \ /
  Last update: 2011-06-08 02:39    [from the cache]
©2003-2010