Re: [PATCH] ext4: Fix delalloc sync hang with journal lock inversion

["Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>]

On Thu, May 22, 2008 at 11:10:17AM -0700, Mingming wrote:
> 
> On Wed, 2008-05-21 at 23:14 +0530, Aneesh Kumar K.V wrote:
> > Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
> > ---
> >  fs/ext4/inode.c |   10 +++++++---
> >  1 files changed, 7 insertions(+), 3 deletions(-)
> > 
> > diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
> > index 46cc610..076d00f 100644
> > --- a/fs/ext4/inode.c
> > +++ b/fs/ext4/inode.c
> > @@ -1571,13 +1571,17 @@ static int ext4_da_writepages(struct address_space *mapping,
> >  		 */
> >  		if (wbc->nr_to_write > EXT4_MAX_WRITEBACK_PAGES)
> >  			wbc->nr_to_write = EXT4_MAX_WRITEBACK_PAGES;
> > -		to_write -= wbc->nr_to_write;
> > 
> > +		to_write -= wbc->nr_to_write;
> >  		ret = mpage_da_writepages(mapping, wbc, ext4_da_get_block_write);
> >  		ext4_journal_stop(handle);
> > -		to_write +=wbc->nr_to_write;
> > +		if (wbc->nr_to_write) {
> > +			/* We failed to write what we requested for */
> > +			to_write += wbc->nr_to_write;
> > +			break;
> > +		}
> 
> Not sure about the break here...
> > +		wbc->nr_to_write = to_write;
> 
> Looks right. thanks.
> 
> >  	}
> > -
> >  out_writepages:
> >  	wbc->nr_to_write = to_write;
> >  	wbc->range_cyclic = range_cyclic;
> 

The call chain that made me look at this was 

#0  ext4_da_writepages (mapping=0xc76dc244, wbc=0xc790bf70) at fs/ext4/inode.c:1557
#1  0xc0150176 in do_writepages (mapping=0xc76dc244, wbc=0xc790bf70) at mm/page-writeback.c:1004
#2  0xc0180fe6 in __writeback_single_inode (inode=0xc76dc11c, wbc=0xc790bf70) at fs/fs-writeback.c:285
#3  0xc018146c in sync_sb_inodes (sb=0xc7abac00, wbc=0xc790bf70) at fs/fs-writeback.c:502
#4  0xc0181701 in writeback_inodes (wbc=0xc790bf70) at fs/fs-writeback.c:570
#5  0xc01509f8 in background_writeout (_min_pages=<value optimized out>) at mm/page-writeback.c:639
#6  0xc0150f57 in pdflush (dummy=<value optimized out>) at mm/pdflush.c:127
#7  0xc01324af in kthread (_create=<value optimized out>) at kernel/kthread.c:79
#8  0xc0104633 in kernel_thread_helper () at include/asm/string_32.h:238

ext4_da_writepages gets called with nr_to_write MAX_WRITEBACK_PAGES. the
file size is only 4K. ie there is only one page to write. With these
value we get stuck in the above loop because to_write will never decrement
below 1023.


-aneesh