| To: | Christoph Hellwig <hch@xxxxxxxxxxxxx> |
|---|---|
| Subject: | Re: [PATCH 2/2 v2] hook up fiemap & associated formatter |
| From: | Eric Sandeen <sandeen@xxxxxxxxxxx> |
| Date: | Sat, 25 Oct 2008 23:38:17 -0500 |
| Cc: | xfs-oss <xfs@xxxxxxxxxxx> |
| In-reply-to: | <20081025202657.GG28002@xxxxxxxxxxxxx> |
| References: | <49031E81.7040807@xxxxxxxxxxx> <4903207B.6030009@xxxxxxxxxxx> <20081025202657.GG28002@xxxxxxxxxxxxx> |
| User-agent: | Thunderbird 2.0.0.17 (Macintosh/20080914) |
Christoph Hellwig wrote: >> +#define XFS_FIEMAP_FLAGS (FIEMAP_FLAG_SYNC|FIEMAP_FLAG_XATTR) > > Why do we need FIEMAP_FLAG_SYNC in here? That flag is implemented in > the VFS and should be handled there. (Probably more a question for > the generic implementation than your patch) well, sorta, it calls filemap_write_and_wait but no lock is held to keep that consistent when the filesystem gets to it... IMHO it *should* probably be up to the fs to handle the sync flag. But anyway, for it all to hang together I think we need to tell it that we understand this flag. -Eric >> + >> +STATIC int xfs_fiemap_format( > > STATIC int > xfs_fiemap_format( > > > > Buit the meat of the patch looks good to me. > > |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: deadlock with latest xfs, Dave Chinner |
|---|---|
| Next by Date: | Re: linux-next: left over things in linux-next after 2.6.28-c1, Christoph Hellwig |
| Previous by Thread: | Re: [PATCH 2/2 v2] hook up fiemap & associated formatter, Christoph Hellwig |
| Next by Thread: | Re: [PATCH 0/2 v2] hook xfs to fiemap ioctl, Christoph Hellwig |
| Indexes: | [Date] [Thread] [Top] [All Lists] |