From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757877AbbGGPNR (ORCPT ); Tue, 7 Jul 2015 11:13:17 -0400 Received: from mail-pd0-f182.google.com ([209.85.192.182]:35741 "EHLO mail-pd0-f182.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932622AbbGGPMw (ORCPT ); Tue, 7 Jul 2015 11:12:52 -0400 Date: Wed, 8 Jul 2015 00:12:04 +0900 From: Sergey Senozhatsky To: Minchan Kim Cc: Sergey Senozhatsky , Andrew Morton , linux-mm@kvack.org, linux-kernel@vger.kernel.org, Sergey Senozhatsky Subject: Re: [PATCH v6 7/7] zsmalloc: use shrinker to trigger auto-compaction Message-ID: <20150707151204.GE1450@swordfish> References: <1436270221-17844-1-git-send-email-sergey.senozhatsky@gmail.com> <1436270221-17844-8-git-send-email-sergey.senozhatsky@gmail.com> <20150707134445.GD3898@blaptop> <20150707144107.GC1450@swordfish> <20150707150143.GC23003@blaptop> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20150707150143.GC23003@blaptop> User-Agent: Mutt/1.5.23+89 (0255b37be491) (2014-03-12) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On (07/08/15 00:01), Minchan Kim wrote: [..] > > But why would we do this? Yes, it's kinda-sorta bad -- we were not > > able to register zspool shrinker, so there will be no automatic > > compaction... And that's it. > > > > It does not affect zsmalloc/zram functionality by any means. Including > > compaction itself -- user still has a way to compact zspool (manually). > > And in some scenarios user will never even see automatic compaction in > > action (assuming that there is a plenty of RAM available). > > > > Can you explain your decision? > > I don't think it would fail in *real practice*. > Althout it might happen, what does zram could help in that cases? > This argument depends on the current register_shrinker() implementation, should some one add additional return branch there and it's done. > If it were failed, it means there is already little memory on the system > so zram could not be helpful for those environment. > IOW, zram should be enabled earlier. > > If you want it strongly, please reproduce such failing and prove that > zram was helpful for the system. No, thanks. I'll just remove it. -ss