| To: | Barry Naujok <bnaujok@xxxxxxx> |
|---|---|
| Subject: | Re: REVIEW: Fix xfs_check SEGV when encountering an unreadable block |
| From: | Christoph Hellwig <hch@xxxxxxxxxxxxx> |
| Date: | Tue, 26 Aug 2008 15:13:44 -0400 |
| Cc: | "xfs@xxxxxxxxxxx" <xfs@xxxxxxxxxxx> |
| In-reply-to: | <op.udjgr6qc3jf8g2@pc-bnaujok.melbourne.sgi.com> |
| References: | <op.udjgr6qc3jf8g2@pc-bnaujok.melbourne.sgi.com> |
| Sender: | xfs-bounce@xxxxxxxxxxx |
| User-agent: | Mutt/1.5.18 (2008-05-17) |
On Mon, Jun 30, 2008 at 11:51:32AM +1000, Barry Naujok wrote: > xfs_check (xfs_db "check" command) internally uses a stack for I/Os > it reads/writes. In the check command, there are a few places where > the I/O stack is pushed, a read is issued and the read fails but > does not pop the stack location back. > > In nested uses of this stack, the caller then accesses this un-popped > block which the data pointer is "NULL" causing a SEGV. > > I've checked all calls to push_cur()/set_cur() to make sure all > failures call pop_cur(). Looks good to me. |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: REVIEW: xfs_metadump improvements, Christoph Hellwig |
|---|---|
| Next by Date: | Re: REVIEW: Remove convert_extent function in xfs_repair, Christoph Hellwig |
| Previous by Thread: | Re: REVIEW: Fix xfs_check SEGV when encountering an unreadable block, Barry Naujok |
| Next by Thread: | TAKE 983809 - Un-popped buffers for failed reads in xfs_db can SEGV, Barry Naujok |
| Indexes: | [Date] [Thread] [Top] [All Lists] |