| To: | Dave Chinner <david@xxxxxxxxxxxxx> |
|---|---|
| Subject: | Re: [PATCH 4/9] repair: detect and correct CRC errors in directory blocks |
| From: | Christoph Hellwig <hch@xxxxxxxxxxxxx> |
| Date: | Mon, 21 Apr 2014 00:08:24 -0700 |
| Cc: | xfs@xxxxxxxxxxx |
| Delivered-to: | xfs@xxxxxxxxxxx |
| In-reply-to: | <1397550301-31883-5-git-send-email-david@xxxxxxxxxxxxx> |
| References: | <1397550301-31883-1-git-send-email-david@xxxxxxxxxxxxx> <1397550301-31883-5-git-send-email-david@xxxxxxxxxxxxx> |
| User-agent: | Mutt/1.5.21 (2010-09-15) |
On Tue, Apr 15, 2014 at 06:24:56PM +1000, Dave Chinner wrote: > + * Need to handle CRC and validation errors specially here. If > + * there is a validator error, re-read without the verifier so > + * that we get a buffer we can check and repair. Re-attach the > + * ops to the buffer after the read so that when it is rewritten > + * the CRC is recalculated. > + * > + * Returns a positive error on failure, 0 for success, and negative > + * error if a verifier error occurred and we reread the block without > + * the verifier successfully. This comment doesn't come clear to using up the 80 character allowance :) I have to say the positive and negative errno convention is a bit confusing to me. I'd rather always return 0 if we could read it, and pass a bool *crc_error instead that indicates if it needs fixing. |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: [PATCH 2/9] db: verify buffer on type change, Christoph Hellwig |
|---|---|
| Next by Date: | Re: [PATCH 5/9] repair: detect CRC errors in AG headers, Christoph Hellwig |
| Previous by Thread: | [PATCH 4/9] repair: detect and correct CRC errors in directory blocks, Dave Chinner |
| Next by Thread: | Seeking XFS tuning advice for PostgreSQL on SATA SSDs/Linux-md, Johannes Truschnigg |
| Indexes: | [Date] [Thread] [Top] [All Lists] |