| To: | Christoph Hellwig <hch@xxxxxxxxxxxxx> |
|---|---|
| Subject: | Re: [PATCH 1/4] libxfs: handle read errors in libxfs_trans_read_buf |
| From: | Alex Elder <aelder@xxxxxxx> |
| Date: | Wed, 21 Sep 2011 15:13:26 -0500 |
| Cc: | <xfs@xxxxxxxxxxx> |
| In-reply-to: | <20110921194936.GA20400@xxxxxxxxxxxxx> |
| References: | <20110920215917.508421839@xxxxxxxxxxxxxxxxxxxxxx> <20110920220030.204339904@xxxxxxxxxxxxxxxxxxxxxx> <1316630707.5872.13.camel@doink> <20110921194936.GA20400@xxxxxxxxxxxxx> |
| Reply-to: | <aelder@xxxxxxx> |
On Wed, 2011-09-21 at 15:49 -0400, Christoph Hellwig wrote: > On Wed, Sep 21, 2011 at 01:45:07PM -0500, Alex Elder wrote: > > Probably wouldn't hurt to initialize b_fsprivate2 even > > if the transaction pointer is null. Looks good though. > > It would - given that another transaction could be set in there. > Fairly unlikely in userspace, but quite possible in the kernel. > You're right. -Alex |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: [PATCH v2] xfsdump: call mlog_exit in content_stream_restore, Alex Elder |
|---|---|
| Next by Date: | Re: [PATCH] xfstests: simplify TRIM_OFF_LEN() in "ltp/fsx.c", Alex Elder |
| Previous by Thread: | Re: [PATCH 1/4] libxfs: handle read errors in libxfs_trans_read_buf, Christoph Hellwig |
| Next by Thread: | [PATCH 3/4] libxfs: handle short reads in libxfs_readbufr, Christoph Hellwig |
| Indexes: | [Date] [Thread] [Top] [All Lists] |