xfs
[Top] [All Lists]

RE: TAKE - pagebuf

To: "'Nathan Scott'" <nathans@xxxxxxx>, "HABBINGA,ERIK (HP-Loveland,ex1)" <erik.habbinga@xxxxxx>
Subject: RE: TAKE - pagebuf
From: "HABBINGA,ERIK (HP-Loveland,ex1)" <erik.habbinga@xxxxxx>
Date: Tue, 11 Mar 2003 12:17:22 -0500
Cc: linux-xfs@xxxxxxxxxxx
Sender: linux-xfs-bounce@xxxxxxxxxxx
Nathan,
   When I pull out this particular take, everything works fine.  The volumes
that fail to initialize quotas are on LVM devices, not standard scsi
devices.  The scsi devices mount correctly.

I'm not consciously changing the pagesize and blocksize, so I assume I'm
using the defaults.

I'll work on adding the printks/debugging enablement later on today.

Thanks,
Erik

> -----Original Message-----
> From: Nathan Scott [mailto:nathans@xxxxxxx]
> Sent: Monday, March 10, 2003 3:45 PM
> To: HABBINGA,ERIK (HP-Loveland,ex1)
> Cc: linux-xfs@xxxxxxxxxxx
> Subject: Re: TAKE - pagebuf
> 
> 
> hi Erik,
> 
> On Mon, Mar 10, 2003 at 03:09:47PM -0500, HABBINGA,ERIK 
> (HP-Loveland,ex1) wrote:
> > When applying this patch from March 3rd, creating a file 
> system, and then
> > unmounting and remounting xfs filesystems gives me the 
> following error:
> > 
> > Mar 10 13:04:07 localhost kernel: XFS mounting filesystem lvm(58,1)
> > Mar 10 13:04:07 localhost kernel: Filesystem "lvm(58,1)": Failed to
> > initialize disk quotas.
> > 
> > We're using the -o rw,nodev,noatime,quota mount options.  
> Mounting without
> > the quota option doesn't generate the error.  Using a CVS 
> take from March
> > 6th does the same thing, but dumps a stack trace as well.
> > 
> > Any ideas?
> > 
> 
> Hmm... not really.  I just tried a mkfs & mount with those options
> and it worked correctly for me.  A few suggestions - try adding a
> few printk's in xfs_qm_mount_quotas and friends, to pinpoint where
> the error is really coming from (there are several potential error
> sources in that routine).  You might also want to try turning XFS
> debugging on, see if you trip an assert earlier on.
> 
> Also what pagesize and blocksize are you using here?  If you back
> this mod out (and just this mod), does it start working again?
> 
> thanks.
> 
> > 
> > Remove an unused parameter from pagebuf_lookup.  Ensure we 
> don't mark
> > a page uptodate in the bsize==pgsize case when we didn't do 
> IO to the
> > entire page.
> > 
> > 
> > Date:  Mon Mar  3 18:17:57 PST 2003
> > Workarea:  snort.melbourne.sgi.com:/home/nathans/pagebuf
> > 
> > The following file(s) were checked into:
> >   bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs
> > 
> > 
> > Modid:  2.4.x-xfs:slinx:140809a
> > linux/fs/xfs/pagebuf/page_buf.c - 1.100
> >     - Remove an unused parameter from pagebuf_lookup.  
> Ensure we don't
> > mark
> >       a page uptodate in the bsize==pgsize case when we 
> didn't do IO to
> > the
> >       entire page.
> > 
> > linux/fs/xfs/pagebuf/page_buf.h - 1.59
> > linux/fs/xfs/linux/xfs_aops.c - 1.19
> >     - Remove an unused parameter from pagebuf_lookup.
> > 
> > 
> > 
> > 
> 
> -- 
> Nathan
> 


<Prev in Thread] Current Thread [Next in Thread>