From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Disseldorp Subject: Re: [PATCH 2/2] rbd: get rid of rbd_mapping::read_only Date: Wed, 8 Nov 2017 19:32:42 +0100 Message-ID: <20171108193242.7b3b16c5@suse.de> References: <1509983370-1853-1-git-send-email-idryomov@gmail.com> <1509983370-1853-2-git-send-email-idryomov@gmail.com> <20171107142627.0d66f97b@suse.de> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Return-path: Received: from mx2.suse.de ([195.135.220.15]:37336 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751304AbdKHScn (ORCPT ); Wed, 8 Nov 2017 13:32:43 -0500 In-Reply-To: Sender: ceph-devel-owner@vger.kernel.org List-ID: To: Ilya Dryomov Cc: Ceph Development On Tue, 7 Nov 2017 17:56:51 +0100, Ilya Dryomov wrote: > Heh, apparently I only tested writes and didn't do discards... Writes > go through blkdev_write_iter() but zeroout/discards seem to rely solely > on FMODE_WRITE checks, which the block layer doesn't enforce. > > I'd rather try to fix it in the block layer before working around it in > krbd. I'll work up a patch tomorrow. FWIW I've pushed a blktests regression test for this to the wip_ro_get_set_discard branch of my repo at: https://github.com/ddiss/blktests.git Cheers, David