From mboxrd@z Thu Jan 1 00:00:00 1970 From: Arnd Bergmann Subject: Re: [PATCH 2/6] block: push down BKL into .open and .release Date: Wed, 7 Jul 2010 16:04:03 +0200 Message-ID: <201007071604.04086.arnd@arndb.de> References: <1278193640-24223-1-git-send-email-arnd@arndb.de> <1278193640-24223-3-git-send-email-arnd@arndb.de> <20100707015000.GC2950@infradead.org> Mime-Version: 1.0 Content-Type: Text/Plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20100707015000.GC2950@infradead.org> Sender: linux-kernel-owner@vger.kernel.org To: Christoph Hellwig Cc: Jens Axboe , linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, John Kacur , Frederic Weisbecker , linux-scsi@vger.kernel.org List-Id: linux-scsi@vger.kernel.org On Wednesday 07 July 2010, Christoph Hellwig wrote: > I might be missing something, but this doesn't actually push the > BKL anyway. It adds duplicate BKL locking inside the ->open and > ->release methods without actually removing it in the callers. > > That's a pretty pointless thing to do. If you're not feeling > confident enough about touching block_dev.c I can look into that > part of the pushdown, but this patch on it's own is not actually > useful. My idea was that the patch in its current form would be less controversial if it can't break anything, but you're certainly right that the description doesn't match what the patch does. I'm now integrating the half of patch 5, as you suggested. I would much appreciate if you could verify my assumptions about the blkdev_get/blkdev_put functions in the combined patch. Arnd