From mboxrd@z Thu Jan 1 00:00:00 1970 From: Peter Zijlstra Subject: Re: [PATCH -v6 2/3] lib, Make gen_pool memory allocator lockless Date: Mon, 29 Nov 2010 13:11:37 +0100 Message-ID: <1291032697.32004.28.camel@laptop> References: <1291014201-15513-1-git-send-email-ying.huang@intel.com> <1291014201-15513-3-git-send-email-ying.huang@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 8BIT Return-path: Received: from casper.infradead.org ([85.118.1.10]:55309 "EHLO casper.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752350Ab0K2ML1 convert rfc822-to-8bit (ORCPT ); Mon, 29 Nov 2010 07:11:27 -0500 In-Reply-To: <1291014201-15513-3-git-send-email-ying.huang@intel.com> Sender: linux-acpi-owner@vger.kernel.org List-Id: linux-acpi@vger.kernel.org To: Huang Ying Cc: Len Brown , linux-kernel@vger.kernel.org, Andi Kleen , linux-acpi@vger.kernel.org, Andrew Morton , Linus Torvalds , Ingo Molnar On Mon, 2010-11-29 at 15:03 +0800, Huang Ying wrote: > This version of the gen_pool memory allocator supports lockless > operation. > > This makes it safe to use in NMI handlers and other special > unblockable contexts that could otherwise deadlock on locks. This is > implemented by using atomic operations and retries on any conflicts. > The disadvantage is that there may be livelocks in extreme cases. For > better scalability, one gen_pool allocator can be used for each CPU. > > The lockless operation only works if there is enough memory available. > If new memory is added to the pool a lock has to be still taken. So > any user relying on locklessness has to ensure that sufficient memory > is preallocated. > > The basic atomic operation of this allocator is cmpxchg on long. On > architectures that don't have NMI-safe cmpxchg implementation, a > spin_trylock_irqsave based fallback is used for gen_pool_alloc, so it > can be used in NMI handler safely. But gen_pool_free can not be used > in NMI handler in these architectures, because memory free can not > fail. I still don't see a reason to merge this. It makes the genalloc thing slower for every other user (more LOCK'ed ops) and there is no new user presented in this series.