On Mon, 2003-05-19 at 15:21, Christoph Hellwig wrote:
> On Mon, May 19, 2003 at 09:38:49PM +0200, Ingo Juergensmann wrote:
> > 21:32 < ij_> oerx...
> > 21:32 < ij_> xfs in kernel 2.5.69 is not really usuable?
> > 21:32 < hch> it's useable for me at least
> > 21:33 < ij_> kernel oops?
> > 21:33 < hch> as long as you don't use non-standard blocksize or
> > unwritten extents+odirect at least
> > 21:33 < ij_> bad: scheduling while atomic!
> > 21:33 < ij_> Call Trace:
> > 21:33 < ij_> [<c011bba2>] schedule+0x3b2/0x3c0
> > 21:33 < ij_> [<c02098d4>] xfs_iunlock+0x34/0x70
> > 21:33 < ij_> ...
> > 21:33 < hch> ij_: send the full trace to linux-xfs@xxxxxxxxxxx please
> >
> > here it is (attached)
>
> This starts to oops in NFS first (and for a long time). As NFS in 2.5.69
> is know buggy can you reproduce this with the current -bk snapshot?
Plus, there is no call to schedule out of xfs_iunlock.
Steve
--
Steve Lord voice: +1-651-683-3511
Principal Engineer, Filesystem Software email: lord@xxxxxxx
|