From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Hellwig Subject: Re: [take9 2/2] kevent: poll/select() notifications. Timer notifications. Date: Fri, 18 Aug 2006 11:41:20 +0100 Message-ID: <20060818104120.GA20816@infradead.org> References: <1155536496588@2ka.mipt.ru> <11555364962857@2ka.mipt.ru> <20060816133014.GB32499@infradead.org> <20060816134032.GB4314@2ka.mipt.ru> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Christoph Hellwig , lkml , David Miller , Ulrich Drepper , Andrew Morton , netdev , Zach Brown , tglx@linutronix.de Return-path: Received: from pentafluge.infradead.org ([213.146.154.40]:63172 "EHLO pentafluge.infradead.org") by vger.kernel.org with ESMTP id S1751355AbWHRKmH (ORCPT ); Fri, 18 Aug 2006 06:42:07 -0400 To: Evgeniy Polyakov Content-Disposition: inline In-Reply-To: <20060816134032.GB4314@2ka.mipt.ru> Sender: netdev-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On Wed, Aug 16, 2006 at 05:40:32PM +0400, Evgeniy Polyakov wrote: > > What speaks against a patch the recplaces the epoll core by something that > > build on kevent while still supporting the epoll interface as a compatibility > > shim? > > There is no problem from my side, but epoll and kevent_poll differs on > some aspects, so it can be better to not replace them for a while. Please explain the differences and why they are important. We really shouldn't keep on adding code without beeing able to replace older bits. If there's a really good reason we can keep things separate, but "epoll and kevent_poll differs on some aspects" is not one :)