Christoph Hellwig wrote: Okay, what about this little patch?: Index: fs/xfs/linux-2.6/xfs_iops.c == RCS file: /cvs/linux-2.6-xfs/fs/xfs/linux-2.6/xfs_iops.c,v retrieving revision 1.212 diff -u -p -r1
Well, it is nessecary now to update i_size. Or rather it was, I think I can get rid of it again after taking care of initialize_vnode. And initialize_vnode. I have a working patch for the latter, but
Here's a new diff, that should get rid of all the warnings except for O_DIRECT. I've also addressed the comments from Dave and Steve. Index: fs/xfs/linux-2.6/xfs_iops.c == RCS file: /cvs/linux-2.6-xf
Author: Miquel van Smoorenburg <miquels@xxxxxxxxxxx>
Date: Thu, 5 Feb 2004 13:30:37 +0100
This patch appears to work fine, but mkdir() still warns about i_size_write() : i_size_write() called without i_sem Call Trace: [<c013baf0>] i_size_write_check+0x61/0x63 [<f8af2cef>] validate_fields+
Well, this one is harmless as the i_size write is on the newly created inode which isn't on the hash yet. Andrew, do you plan to keep the debugging check? If yes I could cludge around it, but I'd pre