From mboxrd@z Thu Jan 1 00:00:00 1970 From: Brad Campbell Subject: Issue with bcache on top of RAID10 Date: Sun, 22 Apr 2012 19:07:01 +0800 Message-ID: <4F93E655.4080002@fnarfbargle.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: Sender: linux-bcache-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: linux-bcache-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-Id: linux-bcache@vger.kernel.org I've just fired up my test server with the latest git code, and I'm seeing this in the logs Apr 22 18:54:29 test kernel: [ 169.015498] md/raid10:md10: make_request bug: can't convert block across chunks or bigger than 512k 260098047 4 Apr 22 18:54:30 test kernel: [ 170.015440] md/raid10:md10: make_request bug: can't convert block across chunks or bigger than 512k 260098047 4 Apr 22 18:54:31 test kernel: [ 171.015455] md/raid10:md10: make_request bug: can't convert block across chunks or bigger than 512k 260098047 4 Apr 22 18:54:32 test kernel: [ 172.015486] md/raid10:md10: make_request bug: can't convert block across chunks or bigger than 512k 260098047 4 It's triggering this code in /device/md/raid10.c in make_request() /* Sanity check -- queue functions should prevent this happening */ if (bio->bi_vcnt != 1 || bio->bi_idx != 0) goto bad_map; As I didn't see this with plain ext4 on the md device, but I do see it with ext4 on bcache0 on the md device I can only assume something funky is happening in bcache. Happy to do anything to help. Regards, Brad