From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1760196AbYEFMjE (ORCPT ); Tue, 6 May 2008 08:39:04 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1754292AbYEFMiv (ORCPT ); Tue, 6 May 2008 08:38:51 -0400 Received: from atrey.karlin.mff.cuni.cz ([195.113.31.123]:55910 "EHLO atrey.karlin.mff.cuni.cz" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754141AbYEFMiu (ORCPT ); Tue, 6 May 2008 08:38:50 -0400 Date: Tue, 6 May 2008 14:39:36 +0200 From: Pavel Machek To: Ingo Molnar Cc: kernel list , kaszak@gmail.com, lcostantino@gmail.com, linux-wireless@vger.kernel.org Subject: Re: w35und: fixed nasty stack overflow, I can actually ssh over this Message-ID: <20080506123936.GA4777@elf.ucw.cz> References: <20080505205617.GD7273@elf.ucw.cz> <20080506123536.GM32591@elte.hu> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20080506123536.GM32591@elte.hu> X-Warning: Reading this can be dangerous to your mental health. User-Agent: Mutt/1.5.17 (2007-11-01) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi! > > Removed nasty stack overflow, and got transmit of big packets to work. > > I can now actually do ssh over softmac-based driver. > > > > Unindent code to make it more readable. > > > --- a/arch/x86/kernel/traps_32.c > > +++ b/arch/x86/kernel/traps_32.c > > @@ -146,11 +146,16 @@ static inline unsigned long print_contex > > unsigned long *stack, unsigned long bp, > > const struct stacktrace_ops *ops, void *data) > > { > > + int max = 5; > > struct stack_frame *frame = (struct stack_frame *)bp; > > > > while (valid_stack_ptr(tinfo, stack, sizeof(*stack))) { > > unsigned long addr; > > > > + if (!max) > > + return; > > + max--; > > + > > hm, you got into an infinite loop here, right? Could you please send > this bit as a separate patch - and i guess it should use > kstack_depth_to_print instead of '5'? No, sorry, this should not have been in this patch. (And no, this is not the overflow I was talking about). What happened to me is that current oops is too long even for vga=1, and I was trying to see where it dies. I was trying to only print first few lines of the backtrace, but then I fix the nasty bug... Pavel -- (english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html