Re: [PATCH] ext3: wait on all pending commits in ext3_sync_fs

[Andrew Morton <akpm@linux-foundation.org>]

On Mon, 3 Nov 2008 16:19:29 -0500
Theodore Tso <tytso@mit.edu> wrote:

> On Mon, Nov 03, 2008 at 01:13:13PM -0800, Andrew Morton wrote:
> > On Mon, 3 Nov 2008 12:58:54 -0800
> > Arthur Jones <ajones@riverbed.com> wrote:
> > 
> > > Hi Andrew, ...
> > > 
> > > On Mon, Nov 03, 2008 at 12:37:50PM -0800, Andrew Morton wrote:
> > > > [...]
> > > > OK.  But still, questions remain.
> > > > 
> > > > - should we clear s_dirt if log_wait_commit() didn't work?
> > > > 
> > > > - ext3_force_commit() clears s_dirt also
> > > > 
> > > > - should ext3_sync_fs() be dropping the ext3_force_commit() return
> > > >   value on the floor?
> 
> Should all the callers of ->sync_fs also be dropping the error returns
> on the floor?

Oh, this is sync_fs.  How meaningful would it be to return an error
from sys_umount()?  Would there be any point in leaving the errored fs
mounted?  Dunno.

But if we're going to drop this error on the floor, we should do
that at a higher level, not on a per-fs basis ;)