| To: | Brian Foster <bfoster@xxxxxxxxxx> |
|---|---|
| Subject: | Re: [PATCH 4/4] xfs: rewrite and optimize the delalloc write path |
| From: | Christoph Hellwig <hch@xxxxxx> |
| Date: | Tue, 30 Aug 2016 16:44:26 +0200 |
| Cc: | Christoph Hellwig <hch@xxxxxx>, xfs@xxxxxxxxxxx |
| Delivered-to: | xfs@xxxxxxxxxxx |
| In-reply-to: | <20160826160753.GD17728@xxxxxxxxxxxxxxx> |
| References: | <1471816273-28940-1-git-send-email-hch@xxxxxx> <1471816273-28940-5-git-send-email-hch@xxxxxx> <20160825143708.GD25041@xxxxxxxxxxxxxxx> <20160826143344.GB21535@xxxxxx> <20160826160339.GC17728@xxxxxxxxxxxxxxx> <20160826160753.GD17728@xxxxxxxxxxxxxxx> |
| User-agent: | Mutt/1.5.17 (2007-11-01) |
On Fri, Aug 26, 2016 at 12:07:53PM -0400, Brian Foster wrote: > > Not quite sure I follow the last bit, but I don't necessarily think the > > whole thing has to be boxed into a helper to clean it up. E.g., I'd do > > something like the appended diff (compile tested only). > > > > ... and if the function signature is really an issue, trade off idx & > prev for a conditional base preallocation size (applies on top of the > previous diff): These two patches together look pretty reasonable. I'll retest with it include and will pick it up for the next version if it works. |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: [PATCH 3/4] xfs: make xfs_inode_set_eofblocks_tag cheaper for the common case, Christoph Hellwig |
|---|---|
| Next by Date: | Re: BUG: Internal error xfs_trans_cancel at line 984 of file fs/xfs/xfs_trans.c, Eryu Guan |
| Previous by Thread: | Re: [PATCH 4/4] xfs: rewrite and optimize the delalloc write path, Brian Foster |
| Next by Thread: | Re: [PATCH 4/4] xfs: rewrite and optimize the delalloc write path, Brian Foster |
| Indexes: | [Date] [Thread] [Top] [All Lists] |