From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tejun Heo Subject: Re: [PATCH 01/17] Add ida and idr helper routines. Date: Tue, 15 Sep 2015 14:41:57 -0400 Message-ID: <20150915184157.GA495@htj.duckdns.org> References: <915ec9ff5e9cc1fae0b36bf7d4c4cb115439e15d.1442263512.git.lduncan@suse.com> <20150915182755.GA31484@htj.duckdns.org> <1442342322.2177.13.camel@HansenPartnership.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <1442342322.2177.13.camel@HansenPartnership.com> Sender: linux-kernel-owner@vger.kernel.org To: James Bottomley Cc: Lee Duncan , linux-scsi@vger.kernel.org, linux-kernel@vger.kernel.org, Hannes Reinecke , Johannes Thumshirn , Christoph Hellwig List-Id: linux-scsi@vger.kernel.org Hello, On Tue, Sep 15, 2015 at 11:38:42AM -0700, James Bottomley wrote: > For most of the SCSI stuff, yes. I'm less sure about the sd numbers. > They go up very high and get hammered a lot during system bring up and > hot plug. I think having their own lock rather than wrapping everything > around simple_ida_lock makes more sense here just because the system is > heavily contended on getting indexes at bring up. > > To continue the thought, why not move simple_ida_lock into struct ida so > we don't have to worry about the contention and can sue ida_simple_... > everywhere? We sure can do that if necessary but I'm rather doubtful that even with sd number hammering this is likely to be a problem. Let's convert the users to the simple interface and make the lock per-ida if we actually see contention on the lock. Thanks. -- tejun