From mboxrd@z Thu Jan 1 00:00:00 1970 From: Vladislav Bolkhovitin Subject: Re: [PATCH]: Rename REQ_COPY_USER to more descriptive REQ_HAS_TAIL_SPACE_FOR_PADDING Date: Tue, 14 Jul 2009 20:34:43 +0400 Message-ID: <4A5CB3A3.6000407@vlnb.net> References: <4A563259.5000601@vlnb.net> <20090709181713.GQ23611@kernel.dk> <4A56373C.4050409@vlnb.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from moutng.kundenserver.de ([212.227.17.10]:52617 "EHLO moutng.kundenserver.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754906AbZGNQfr (ORCPT ); Tue, 14 Jul 2009 12:35:47 -0400 In-Reply-To: <4A56373C.4050409@vlnb.net> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: Jens Axboe Cc: linux-scsi@vger.kernel.org, linux-kernel@vger.kernel.org, scst-devel@lists.sourceforge.net, Tejun Heo , Boaz Harrosh , James Bottomley , FUJITA Tomonori Vladislav Bolkhovitin, on 07/09/2009 10:30 PM wrote: > Jens Axboe, on 07/09/2009 10:17 PM wrote: >> On Thu, Jul 09 2009, Vladislav Bolkhovitin wrote: >>> Currently names of REQ_COPY_USER and __REQ_COPY_USER constants are confusing, >>> because they actually mean that the buffer for the corresponding requests >>> has space in the tail for padding in case of DMA padding restrictions. >> No, that's not what it means, the fact that there's padding room is a >> side effect of the map type. So I'd suggest adding a comment above that >> if {} in blk_rq_map_sg(), something that should have been there from the >> beginning. > > Can you elaborate a bit more about what REQ_COPY_USER should mean, please? > > As far as I can see from the sources, currently it's used to only to > determine if there is the padding space. Maybe, the original meaning > doesn't make sense anymore? Jens, Sorry for bothering you, but with use of the kernel-generated buffers, like in the patch I sent in http://groups.google.com/group/linux.kernel/browse_thread/thread/b5127eeadbef934e/5da7ec7b424d2f36?lnk=raot name REQ_COPY_USER becomes quite misleading. For instance, for blk_rq_map_kern_sg() usage of REQ_COPY_USER the buffers neither copied, nor from user. Plus, as I wrote above, in 2.6.30 REQ_COPY_USER used only to determine if there is the padding space in the end. Hence, I suggested the rename. Thanks, Vlad