From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from caramon.arm.linux.org.uk ([217.147.92.249]:4570 "EHLO caramon.arm.linux.org.uk" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1031499AbXD3Jdj (ORCPT ); Mon, 30 Apr 2007 05:33:39 -0400 Date: Mon, 30 Apr 2007 10:33:31 +0100 From: Russell King Subject: Re: utrace comments Message-ID: <20070430093331.GB11345@flint.arm.linux.org.uk> References: <20061127165138.GA2991@lst.de> <20070430040213.BF9901801A4@magilla.sf.frob.com> <20070430090840.GB31397@infradead.org> <20070430091808.GA11345@flint.arm.linux.org.uk> <20070430092200.GA15773@lst.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20070430092200.GA15773@lst.de> Sender: linux-arch-owner@vger.kernel.org To: Christoph Hellwig Cc: Christoph Hellwig , Roland McGrath , linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org List-ID: On Mon, Apr 30, 2007 at 11:22:00AM +0200, Christoph Hellwig wrote: > On Mon, Apr 30, 2007 at 10:18:09AM +0100, Russell King wrote: > > Roland's idea of single-stepping is that it *must* be supported by > > hardware for utrace to use it. There are a number of architectures > > which can only do single-stepping by modifying the text of the > > program being single stepped. ARM is one such example. > > > > As such, even when utrace is complete, some architectures will never > > support in-kernel single step with utrace. I believe Roland's idea > > is to have single step supported on these via some vapourware userspace > > library. > > Does the current arm ptrace code support single stepping in kernelspace? > If yes we absolutely need to continue to support it. single stepping of user space code via standard ptrace calls, yes. > > I'd also like to see utrace become *optional* > > for architectures to support, rather than as it currently stands as > > a *mandatory* requirement when merged. > > No way we'd keep both the old ptrace mess and utrace in the same tree. Given the stated arguments from yourself and Roland, that only leaves one solution to that. I have no real problem with a decision being made to drop kernel-based single stepping _provided_ we have some replacement strategy in place and readily available. At the moment I've not seen such a strategy. I'm not sure if Roland's expecting architecture maintainers to create such a strategy themselves - which would probably turn out to being far worse since you could end up with different implementations for each architecture. -- Russell King Linux kernel 2.6 ARM Linux - http://www.arm.linux.org.uk/ maintainer of: