| To: | Dave Chinner <david@xxxxxxxxxxxxx> |
|---|---|
| Subject: | Re: [PATCH] xfs: fix extent format buffer allocation size |
| From: | Christoph Hellwig <hch@xxxxxxxxxxxxx> |
| Date: | Thu, 31 Mar 2011 02:29:20 -0400 |
| Cc: | Christoph Hellwig <hch@xxxxxxxxxxxxx>, xfs@xxxxxxxxxxx |
| In-reply-to: | <20110331055114.GC30279@dastard> |
| References: | <1301453521-5614-1-git-send-email-david@xxxxxxxxxxxxx> <20110330093333.GA32582@xxxxxxxxxxxxx> <20110331055114.GC30279@dastard> |
| User-agent: | Mutt/1.5.21 (2010-09-15) |
On Thu, Mar 31, 2011 at 04:51:14PM +1100, Dave Chinner wrote: > On Wed, Mar 30, 2011 at 05:33:34AM -0400, Christoph Hellwig wrote: > > > + xfs_bmbt_rec_t *ext_buffer; > > > + > > > + ext_buffer = kmem_alloc(XFS_IFORK_SIZE(ip, whichfork), > > > > If the fork size be the minimum of XFS_IFORK_SIZE and the if_bytes > > value? > > I thought about that, but I don't think it makes any difference. If > there are no delalloc extents, then XFS_IFORK_SIZE and ifp->if_bytes > are identical when the fork is in extent format. If there are > delalloc extents, then XFS_IFORK_SIZE() is the one we want. Hence I > don't think we need to even consider the value of ifp->if_bytes at > all here.... You're right. |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: [PATCH] xfs: fix extent format buffer allocation size, Dave Chinner |
|---|---|
| Next by Date: | Re: [PATCH 4/5] xfs: allow reusing busy extents where safe, Christoph Hellwig |
| Previous by Thread: | Re: [PATCH] xfs: fix extent format buffer allocation size, Dave Chinner |
| Next by Thread: | [PATCH] delete a warning in function xfs_buf_readahead, Pan Weiping |
| Indexes: | [Date] [Thread] [Top] [All Lists] |