| To: | Christoph Hellwig <hch@xxxxxxxxxxxxx> |
|---|---|
| Subject: | Re: [PATCH] xfs: avoid direct I/O write vs buffered I/O race |
| From: | Alex Elder <aelder@xxxxxxx> |
| Date: | Mon, 12 Sep 2011 13:43:36 -0500 |
| Cc: | <xfs@xxxxxxxxxxx> |
| In-reply-to: | <20110827144253.GA16381@xxxxxxxxxxxxx> |
| References: | <20110827144253.GA16381@xxxxxxxxxxxxx> |
| Reply-to: | <aelder@xxxxxxx> |
On Sat, 2011-08-27 at 10:42 -0400, Christoph Hellwig wrote: > Currently a buffered reader or writer can add pages to the pagecache > while we are waiting for the iolock in xfs_file_dio_aio_write. Prevent > this by re-checking mapping->nrpages after we got the iolock, and if > nessecary upgrade the lock to exclusive mode. To simplify this a bit > only take the ilock inside of xfs_file_aio_write_checks. > > Signed-off-by: Christoph Hellwig <hch@xxxxxx> I agree with Dave, moving the ILOCK acquisition into xfs_file_aio_write_checks() is very nice. In any case the core change looks good. Reviewed-by: Alex Elder <aelder@xxxxxxx> |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: xfs deadlock in stable kernel 3.0.4, Stefan Priebe |
|---|---|
| Next by Date: | Re: [PATCH] xfs: avoid synchronous transactions when deleting attr blocks, Alex Elder |
| Previous by Thread: | [PATCH] xfstests: add new getdents test, Grazvydas Ignotas |
| Next by Thread: | Re: [PATCH] xfs: avoid synchronous transactions when deleting attr blocks, Alex Elder |
| Indexes: | [Date] [Thread] [Top] [All Lists] |