From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S271727AbTGRHIl (ORCPT ); Fri, 18 Jul 2003 03:08:41 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S271729AbTGRHIl (ORCPT ); Fri, 18 Jul 2003 03:08:41 -0400 Received: from ppp-217-133-42-200.cust-adsl.tiscali.it ([217.133.42.200]:45706 "EHLO dualathlon.random") by vger.kernel.org with ESMTP id S271727AbTGRHIj (ORCPT ); Fri, 18 Jul 2003 03:08:39 -0400 Date: Fri, 18 Jul 2003 09:23:31 +0200 From: Andrea Arcangeli To: linux-kernel@vger.kernel.org Subject: 2.4.22pre6aa2 Message-ID: <20030718072331.GC3928@dualathlon.random> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.4i X-GPG-Key: 1024D/68B9CB43 13D9 8355 295F 4823 7C49 C012 DFA1 686E 68B9 CB43 X-PGP-Key: 1024R/CB4660B9 CC A0 71 81 F4 A0 63 AC C0 4B 81 1D 8C 15 C8 E5 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org URL: http://www.us.kernel.org/pub/linux/kernel/people/andrea/kernels/v2.4/2.4.22pre6aa2.gz changelog diff between 2.4.22pre6aa1 and 2.4.22pre6aa2: Only in 2.4.22pre6aa1: 00_extraversion-26 Only in 2.4.22pre6aa2: 00_extraversion-27 Only in 2.4.22pre6aa1: 9900_aio-21-ppc-1 Only in 2.4.22pre6aa2: 9903_aio-22-ppc-1 Only in 2.4.22pre6aa1: 9999_sched_yield_scale-5 Only in 2.4.22pre6aa2: 9999_sched_yield_scale-6 Rediffed. Only in 2.4.22pre6aa1: 05_vm_23_per-cpu-pages-2 Only in 2.4.22pre6aa2: 05_vm_23_per-cpu-pages-3 Minor cleanup. Only in 2.4.22pre6aa1: 9900_aio-21.gz Only in 2.4.22pre6aa2: 9900_aio-22.gz Add KM_SOFTIRQ0/1 so crypto can compile. Only in 2.4.22pre6aa2: 9902_aio-poll-1 Added the AIO polling functionality. NOTE: the ABI is IOCB_CMD_POLL = 5. IMHO rather than moving poll into aio, aio should be moved on top of epoll. using poll should always be avoided, but apps already uses aio poll, so at least for the short term it'll be useful. the better approch I suggested is to hook the iocb somehow into an epoll fd. that epoll feature is missing right now AFIK, but it shouldn't be hard to add and it will give persistence to the polling too. The efficient point to sleep is epoll, not io_getevents/poll (io_getevents is optimal only for waiting for I/O [be it storage or network or pipe I/O of course]). Only in 2.4.22pre6aa1: 9999900_desktop-2 Only in 2.4.22pre6aa2: 9999900_desktop-3 Convert max-timeslice/min-timeslice to usecs units (the day we'll be interested to schedule more frequently than 1 time per usec, I guess we'll be very happy to break this interface with userspace ;). Only in 2.4.22pre6aa1: 9999900_ecc-20020904-1.gz Only in 2.4.22pre6aa2: 9999900_ecc-20020904-2.gz Merged more recent updates from Chip. Andrea