From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1759634AbbA0UOP (ORCPT ); Tue, 27 Jan 2015 15:14:15 -0500 Received: from aserp1040.oracle.com ([141.146.126.69]:32334 "EHLO aserp1040.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1759589AbbA0UOM (ORCPT ); Tue, 27 Jan 2015 15:14:12 -0500 Message-ID: <54C7F180.1030505@oracle.com> Date: Tue, 27 Jan 2015 15:13:52 -0500 From: Sasha Levin User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.3.0 MIME-Version: 1.0 To: Joonsoo Kim CC: linux-kernel@vger.kernel.org, m.szyprowski@samsung.com, akpm@linux-foundation.org, lauraa@codeaurora.org Subject: Re: [PATCH v2 3/3] mm: cma: release trigger References: <1422282365-20015-1-git-send-email-sasha.levin@oracle.com> <1422282365-20015-4-git-send-email-sasha.levin@oracle.com> <20150127081012.GD11358@js1304-P5Q-DELUXE> <54C7D801.3070502@oracle.com> In-Reply-To: <54C7D801.3070502@oracle.com> Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 7bit X-Source-IP: ucsinet21.oracle.com [156.151.31.93] Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 01/27/2015 01:25 PM, Sasha Levin wrote: > On 01/27/2015 03:10 AM, Joonsoo Kim wrote: >>> >> + if (mem->n <= count) { >>>> >> > + cma_release(cma, mem->p, mem->n); >>>> >> > + count -= mem->n; >>>> >> > + kfree(mem); >>>> >> > + } else { >>>> >> > + cma_release(cma, mem->p, count); >>>> >> > + mem->p += count; >>>> >> > + mem->n -= count; >>>> >> > + count = 0; >>>> >> > + cma_add_to_cma_mem_list(mem); >>>> >> > + } >>>> >> > + } >> > If order_per_bit is not 0 and count used in cma_release() is >> > different with the count used in cma_alloc(), problem could >> > occur since bitmap management code can't handle that situation. >> > >> > Could we just disable this case in this testing module? > How should it behave then? Just free a max of 'count' pages and > stop beforehand if we're going to go over it? Actually, Can I just check for order_per_bit == 0 and execute it then? I don't want to avoid testing these paths. Thanks, Sasha