Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261281AbVDIElG (ORCPT ); Sat, 9 Apr 2005 00:41:06 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261283AbVDIElG (ORCPT ); Sat, 9 Apr 2005 00:41:06 -0400 Received: from mustang.oldcity.dca.net ([216.158.38.3]:473 "HELO mustang.oldcity.dca.net") by vger.kernel.org with SMTP id S261281AbVDIEjj (ORCPT ); Sat, 9 Apr 2005 00:39:39 -0400 Subject: Re: HELP:porting linux PXA audio driver to RTLinux(RTLinux core driver) From: Lee Revell To: nobin matthew Cc: kernelnewbies@nl.linux.org, linux-arm-kernel@lists.arm.linux.org.uk, linux-kernel@vger.kernel.org, linux-net@vger.kernel.org In-Reply-To: <20050409040327.93029.qmail@web53902.mail.yahoo.com> References: <20050409040327.93029.qmail@web53902.mail.yahoo.com> Content-Type: text/plain Date: Sat, 09 Apr 2005 00:39:38 -0400 Message-Id: <1113021578.5975.56.camel@mindpipe> Mime-Version: 1.0 X-Mailer: Evolution 2.2.1.1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1945 Lines: 48 On Fri, 2005-04-08 at 21:03 -0700, nobin matthew wrote: > Dear Friends, > > I am trying to port Linux PXA audio > driver to RTLinux. I am using pxa-ac7.c and > pxa-audio.c > and eliminated sound_core.c, and i will register two > device /dev/mixer and /dev/dsp to RTLinux kernel. > > The real need is, i wants to generate a sin > wave using audio codec. With in 600us DMA controller > should fill the codec FIFO, if that is not met > distortion will happen. I think normal linux > interrupts and Process scheduling may cause some > problems. > > In porting it seems difficult to port kernel > scheduling , dynamic memory allocation(for DMA) and > synchronization. This is the exact same question you posted to linux-audio-dev. And you'll get the same answer here: Don't waste your time with RTLinux. Use a recent 2.6 kernel with Ingo's realtime-preempt patches. Configure with PREEMPT_RT. This will provide deterministic, hard realtime performance for any RT constraint down to about 50 usecs. RTLinux can meet a ~15 usec RT constraint, the RT preempt kernel will be able to do this once the timer ISR is made preemptible again. RTLinux is an obsolescent product that had its place in the 2.4 era but is being superseded by a solution that's both technically superior and 100% free software. Real time preemption is the future of hard realtime on Linux. All the major real time Linux players have recognized this, even the ones who had competing solutions for 2.4 are porting their work to the 2.6 RT-preempt framework. Besides, RTLinux is a commercial product anyway. If this project requires use of RTLinux for political reasons, call your support rep. Lee - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/