From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753632AbcAOD0j (ORCPT ); Thu, 14 Jan 2016 22:26:39 -0500 Received: from mail-pa0-f46.google.com ([209.85.220.46]:35740 "EHLO mail-pa0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753468AbcAOD0D (ORCPT ); Thu, 14 Jan 2016 22:26:03 -0500 Date: Fri, 15 Jan 2016 12:27:12 +0900 From: Sergey Senozhatsky To: Minchan Kim Cc: Junil Lee , Andrew Morton , ngupta@vflare.org, sergey.senozhatsky.work@gmail.com, linux-mm@kvack.org, linux-kernel@vger.kernel.org Subject: Re: [PATCH] zsmalloc: fix migrate_zspage-zs_free race condition Message-ID: <20160115032712.GC1993@swordfish> References: <1452818184-2994-1-git-send-email-junil0814.lee@lge.com> <20160115023518.GA10843@bbox> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20160115023518.GA10843@bbox> User-Agent: Mutt/1.5.24 (2015-08-30) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Cc Andrew, On (01/15/16 11:35), Minchan Kim wrote: [..] > > Signed-off-by: Junil Lee > > --- > > mm/zsmalloc.c | 1 + > > 1 file changed, 1 insertion(+) > > > > diff --git a/mm/zsmalloc.c b/mm/zsmalloc.c > > index e7414ce..bb459ef 100644 > > --- a/mm/zsmalloc.c > > +++ b/mm/zsmalloc.c > > @@ -1635,6 +1635,7 @@ static int migrate_zspage(struct zs_pool *pool, struct size_class *class, > > free_obj = obj_malloc(d_page, class, handle); > > zs_object_copy(free_obj, used_obj, class); > > index++; > > + free_obj |= BIT(HANDLE_PIN_BIT); > > record_obj(handle, free_obj); > > I think record_obj should store free_obj to *handle with masking off least bit. > IOW, how about this? > > record_obj(handle, obj) > { > *(unsigned long)handle = obj & ~(1< } [just a wild idea] or zs_free() can take spin_lock(&class->lock) earlier, it cannot free the object until the class is locked anyway, and migration is happening with the locked class. extending class->lock scope in zs_free() thus should not affect the perfomance. so it'll be either zs_free() is touching the object or the migration, not both. -ss