From mboxrd@z Thu Jan 1 00:00:00 1970 From: Hannes Reinecke Subject: Re: [PATCH 4/7] dm mpath: remove process_queued_ios() Date: Tue, 11 Feb 2014 19:03:13 +0100 Message-ID: <52FA65E1.7070602@suse.de> References: <1391511280-29325-1-git-send-email-hare@suse.de> <1391511280-29325-5-git-send-email-hare@suse.de> <11AF7C027C4C02408624617A498607840105F76A@BPXM12GP.gisp.nec.co.jp> <52F0CF91.60302@suse.de> <20140210133046.GA19209@redhat.com> <52F9F18A.9000309@suse.de> <20140211155552.GB27812@redhat.com> Reply-To: device-mapper development Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1"; Format="flowed" Content-Transfer-Encoding: quoted-printable Return-path: In-Reply-To: <20140211155552.GB27812@redhat.com> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: dm-devel-bounces@redhat.com Errors-To: dm-devel-bounces@redhat.com To: Mike Snitzer Cc: Junichi Nomura , "dm-devel@redhat.com" , Alasdair Kergon List-Id: dm-devel.ids On 02/11/2014 04:55 PM, Mike Snitzer wrote: > On Tue, Feb 11 2014 at 4:46am -0500, > Hannes Reinecke wrote: > >> On 02/10/2014 02:30 PM, Mike Snitzer wrote: >>> >>> Just to verify, this seems to be the only outstanding question for this >>> patchset? >>> >>> What value are you using for HZ? If this portion of the change does >>> turn out to be meaningul: Rather than tieing to HZ should we just use an >>> explicitly non-zero value for __pg_init_all_paths()'s @min_delay? >>> >> The actual amount here is irrelevant, as long as it's non-zero. >> It's just there to force execution of the work item off the current >> thread. > > I'm aware we just need a non-zero value. My concern, as originally > raised by Junichi in an earlier reply when you had it as HZ/50, is that > the value could be 0 if HZ is really small. While unlikely I see no > point allowing the variable nature of HZ compromise passing a non-zero > value here. Best to just be explicit by passing 1 or something. > > All said, the question of why this is actually needed remains. I trust > you're working on answering that via reproducer (by not forcing the use > of workqueue context)? > Precisely. But as this is a bit hard to trigger it might take some time. (you'll only be hitting this issue if you have to retry = scsi_dh_activate, so you'll need to trigger this somehow). I hope to get it done this week. Is there any deadline which I might miss with that? Cheers, Hannes -- = Dr. Hannes Reinecke zSeries & Storage hare@suse.de +49 911 74053 688 SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 N=FCrnberg GF: J. Hawn, J. Guild, F. Imend=F6rffer, HRB 16746 (AG N=FCrnberg)