From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S262267AbVBSXdj (ORCPT ); Sat, 19 Feb 2005 18:33:39 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S262275AbVBSXdi (ORCPT ); Sat, 19 Feb 2005 18:33:38 -0500 Received: from gate.crashing.org ([63.228.1.57]:29631 "EHLO gate.crashing.org") by vger.kernel.org with ESMTP id S262267AbVBSXdh (ORCPT ); Sat, 19 Feb 2005 18:33:37 -0500 Subject: Re: Current bk on ppc32: kernel text corruption From: Benjamin Herrenschmidt To: Linus Torvalds Cc: Linux Kernel list , Paul Mackerras , "Antonino A. Daplas" In-Reply-To: <1108710554.5587.1.camel@gaston> References: <1108710554.5587.1.camel@gaston> Content-Type: text/plain Date: Sun, 20 Feb 2005 10:32:12 +1100 Message-Id: <1108855933.5584.2.camel@gaston> Mime-Version: 1.0 X-Mailer: Evolution 2.0.3 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Fri, 2005-02-18 at 18:09 +1100, Benjamin Herrenschmidt wrote: > Ok, we may not be over with memory corruption bugs yet. ppc64 now seem > stable running LTP overnight, but my laptop has a page of kernel .text > replaced with zero's as soon as I launch X (and just X, no need to launc > the whole desktop environment). > > I suspect remap_pfn_range() but I haven't checked yet. Ok, I found it. It's a bug that was there forever it seems, where radeonfb accel ops may corrupt kernel memory (via bus master from the radeon chip) when switching back from X, due to X remapping some areas of the card vs. fbcon calling some radeonfb accel ops before radeonfb has a chance to re-initialize the chip (when switching back from X). I'm cooking a workaround patch for 2.6.11, will be ready as soon as I have tested. Proper fix is to make sure fbcon never calls any fbdev accel op before the chip has been properly restored (by a set_var call), but that is beyond the scope of 2.6.11 I suppose... Ben.