From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1D0Gu2-0001UO-Qx for qemu-devel@nongnu.org; Sun, 13 Feb 2005 05:22:59 -0500 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1D0Gtc-00013V-74 for qemu-devel@nongnu.org; Sun, 13 Feb 2005 05:22:33 -0500 Received: from [199.232.76.173] (helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1D0GtV-0000ZJ-Pe for qemu-devel@nongnu.org; Sun, 13 Feb 2005 05:22:26 -0500 Received: from [65.19.178.186] (helo=pythonhacker.is-a-geek.net) by monty-python.gnu.org with esmtp (Exim 4.34) id 1D0GJZ-0002UJ-AZ for qemu-devel@nongnu.org; Sun, 13 Feb 2005 04:45:17 -0500 Received: from localhost (localhost.localdomain [127.0.0.1]) by pythonhacker.is-a-geek.net (Postfix) with ESMTP id 3CDE1403CF for ; Sun, 13 Feb 2005 04:45:10 -0500 (EST) Received: from pythonhacker.is-a-geek.net ([127.0.0.1]) by localhost (pythonhacker.is-a-geek.net [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 01054-03 for ; Sun, 13 Feb 2005 04:44:22 -0500 (EST) Received: from localhost.localdomain (localhost.localdomain [127.0.0.1]) by pythonhacker.is-a-geek.net (Postfix) with ESMTP id BFEE5403D5 for ; Sun, 13 Feb 2005 04:42:40 -0500 (EST) Subject: Re: [Qemu-devel] KQEMU bus errors From: Darryl Dixon In-Reply-To: <420EE242.2080908@wasp.net.au> References: <1108086522.5629.11.camel@localhost.localdomain> <420EB226.4010301@thechesterfields.org> <41e41e7a0502121817746fc443@mail.gmail.com> <420EE242.2080908@wasp.net.au> Content-Type: multipart/alternative; boundary="=-d2CqNlpZU8jAkfjl/JWw" Date: Sun, 13 Feb 2005 22:42:37 +1300 Message-Id: <1108287757.7619.7.camel@localhost.localdomain> Mime-Version: 1.0 Reply-To: esrever_otua@pythonhacker.is-a-geek.net, qemu-devel@nongnu.org List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org --=-d2CqNlpZU8jAkfjl/JWw Content-Type: text/plain Content-Transfer-Encoding: 7bit Hi, I'm not convinced that this is actually the case. I have tested this on several machines now with several guest OSes and get the same result on all of them (which all previously worked). With the snapshot from 2005-02-09 all works correctly. With CVS from 2005-02-11 onwards I get the error ('Bus error' on host for WinXP guest and 'Protection Fault' inside guest for Win98 guest). I think that perhaps with smaller guest RAM size you are simply hiding the bug. On one machine at least I have 756MB physical RAM and 3GB swap allocated and get the same problem. Most of the references that I can find for a Linux 'Bus error' talk about unaligned memory accesses. I did a diff of the 2005-02-09 tree and the 2005-02-11 tree and noted that most of the *changes* (rather than additions and new things) were relating to the way that memory is handled and marked inside qemu. I think that there is a chance that one of these has introduced the problem. Regards, D On Sun, 2005-02-13 at 09:14 +0400, Brad Campbell wrote: > Hetz Ben Hamo wrote: > > Yes, found the reason... > > > > Test with different ram size - you'll be surprised... > > > I second this one. > I found I had to make sure my tmpfs "partition" was larger than my requested ram size. Problem went > away. > > Brad -- Darryl Dixon --=-d2CqNlpZU8jAkfjl/JWw Content-Type: text/html; charset=utf-8 Content-Transfer-Encoding: quoted-printable Hi,

I'm not convinced that this is actually the case. = I have tested this on several machines now with several guest OSes and get = the same result on all of them (which all previously worked). With th= e snapshot from 2005-02-09 all works correctly. With CVS from 2005-02= -11 onwards I get the error ('Bus error' on host for WinXP guest and 'Prote= ction Fault' inside guest for Win98 guest). I think that perhaps with= smaller guest RAM size you are simply hiding the bug. On one machine= at least I have 756MB physical RAM and 3GB swap allocated and get the same= problem. Most of the references that I can find for a Linux 'Bus err= or' talk about unaligned memory accesses. I did a diff of the 2005-02= -09 tree and the 2005-02-11 tree and noted that most of the *changes* (rath= er than additions and new things) were relating to the way that memory is h= andled and marked inside qemu. I think that there is a chance that on= e of these has introduced the problem.

Regards,
D

On Sun, 2005-02-13 at 09:14 +0400, Brad Campbell wrote:

Hetz Ben Hamo wrote:
> Yes, found the reason...
> 
> Test with different ram size - you'll be surpr=
ised...
> 
I second this one.
I found I had to make sure my tmpfs "partition=
" was larger than my requested ram size. Problem went 
away.

Brad

--
Darryl Dixon <esrever_otua@pythonhacker.is-a-geek.net>

--=-d2CqNlpZU8jAkfjl/JWw--