From: Tejun Heo Subject: Re: [PATCH v4.4-rc7] sched: isolate task_struct bitfields according to synchronization domains Date: Wed, 30 Dec 2015 15:41:40 -0500 Message-ID: <20151230204140.GA17398@htj.duckdns.org> References: <55FEC685.5010404@oracle.com> <20150921200141.GH13263@mtj.duckdns.org> <20151125144354.GB17308@twins.programming.kicks-ass.net> <20151125150207.GM11639@twins.programming.kicks-ass.net> <20151125174449.GD17308@twins.programming.kicks-ass.net> <20151211162554.GS30240@mtj.duckdns.org> <20151215192245.GK6357@twins.programming.kicks-ass.net> <20151230092337.GD3873@htj.duckdns.org> Mime-Version: 1.0 Return-path: DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=sender:date:from:to:cc:subject:message-id:references:mime-version :content-type:content-disposition:in-reply-to:user-agent; bh=6pR0vetf2W7b8CrAaTpPhJ2D8MoCptySxjL9iTO6dj8=; b=aD5J1oWSuzPYEb75c4gUetLP//gTDbU72WuHv/7/RQm3l41+v+chMQxd4v96yLu+x3 jwuc+bBcutTD3BuXSNEyabBiLzj2+jRwxGg7ltPTrZ/ZmB0s3nepdWlwo8M/w6F3WnBE 7p/sSCfhjWQlqfLxdpFwKjdC26QbfC2L9ERZ8HLpDai4Q7Ptllf73fQOQH+d2SMfV0ox BFV4RpYROCdPt2SR+Be52yAd2usPVoaGN1t5F/4SttoK0FemmxZnmAudBk9NrRwwEiwz /GG0kVTfFBCsjMMkmLsz6e4ZgcQt2XlcRsSEfiBpf+tHlbxDbdA8vIYu9nQvrrzLVH7S Ymww== Content-Disposition: inline In-Reply-To: Sender: cgroups-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-ID: Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Linus Torvalds Cc: Andrey Ryabinin , Ingo Molnar , Sasha Levin , Andrew Morton , Johannes Weiner , Michal Hocko , cgroups , "linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public.gmane.org" , Vladimir Davydov , kernel-team , Dmitry Vyukov , Peter Zijlstra Hello, Linus. On Wed, Dec 30, 2015 at 12:10:12PM -0800, Linus Torvalds wrote: > On Wed, Dec 30, 2015 at 1:23 AM, Tejun Heo wrote: > > > > Peter, I took the patch and changed the bitfields to ulong. > > I wouldn't expect the unsigned long part to matter, except for the > forced split with Right, I was thinking alpha was doing rmw's for things smaller than 64bit. That's 32bit, not 64. > unsigned long :0; > > itself. > > Also, quite frankly, since this is basically very close to other > fields that are *not* unsigned longs, I'd really prefer to not > unnecessarily use a 64-bit field for three bits each. > > So why not just do it with plain unsigned "int", and then maybe just > intersperse them with the other int-sized fields in that neighborhood. > > I'm also wondering if we shouldn't just put the scheduler bits in the > "atomic_flags" thing instead? Sure. -- tejun