From mboxrd@z Thu Jan 1 00:00:00 1970 From: Vivek Goyal Subject: Re: [patch 6/7] blk: add plug for blkdev_issue_discard Date: Tue, 13 Mar 2012 13:14:14 -0400 Message-ID: <20120313171414.GJ29169@redhat.com> References: <20120312030412.375458948@fusionio.com> <20120312031305.946211513@fusionio.com> <20120313155120.GG29169@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org To: "Martin K. Petersen" Cc: Shaohua Li , linux-kernel@vger.kernel.org, linux-raid@vger.kernel.org, neilb@suse.de, axboe@kernel.dk List-Id: linux-raid.ids On Tue, Mar 13, 2012 at 01:04:58PM -0400, Martin K. Petersen wrote: > >>>>> "Vivek" == Vivek Goyal writes: > > Vivek> On Mon, Mar 12, 2012 at 11:04:18AM +0800, Shaohua Li wrote: > >> In raid 0 case, a big discard request is divided into several small > >> requests in chunk_size unit. Such requests can be merged in low layer > >> if we have correct plug added. This should improve the performance a > >> little bit. > > Vivek> Martin posted a patch to remove the support for allowing merging > Vivek> of discard requests. But this seems to be a reasonable use case > Vivek> for allowing mering discard requests. CCing Martin. > > Merging discard requests is hard given how we need to prepare the > command payload at the bottom of the stack. The current upstream merge > code pretends to be working but it actually doesn't. That's why I want > it dead and buried. > > I have some changes pending (that I need for the REQ_COPY support) that > will make merging of non-rw requests easier to deal with. But that's a > kernel release cycle away... So first we will get rid of mering discard request and then enable after one release cycle once REQ_COPY support is in? Thanks Vivek