From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Hellwig Subject: Re: [PATCH 4/14] avr32: use generic ptrace_resume code Date: Wed, 3 Feb 2010 09:36:11 +0100 Message-ID: <20100203083611.GA29025@lst.de> References: <20100202185755.GA3630@lst.de> <20100202185857.GD3630@lst.de> <20100203084737.6e01247b@siona> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from verein.lst.de ([213.95.11.210]:33845 "EHLO verein.lst.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753102Ab0BCIgk (ORCPT ); Wed, 3 Feb 2010 03:36:40 -0500 Content-Disposition: inline In-Reply-To: <20100203084737.6e01247b@siona> Sender: linux-arch-owner@vger.kernel.org List-ID: To: Haavard Skinnemoen Cc: Christoph Hellwig , roland@redhat.com, oleg@redhat.com, akpm@linux-foundation.org, hskinnemoen@atmel.com, linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org On Wed, Feb 03, 2010 at 08:47:37AM +0530, Haavard Skinnemoen wrote: > Sounds like you found three bugs and fixed two of them. Can you tell me > more about the last one, i.e. what should be done to disable > single-stepping? I don't think the flag clearing on PTRACE_KILL matters as the process is a zombie and almost gone anyway, it's more an artefact of the generic implementation that I wanted to mention. It seems like avr32 implement single stepping by setting the TIF_SINGLE_STEP and TIF_BREAKPOINT thread flags. Simply clearing these flags in user_disable_single_step and calling that function from ptrace_disable (which currently clears them itself) should take care of it.