linux-kernel - Re: BIOS bug, APIC version if 0 for CPU#0

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <4821089D.9080901@googlemail.com>
Date:	Wed, 07 May 2008 03:40:45 +0200
From:	Gabriel C <nix.or.die@...glemail.com>
To:	"Rafael J. Wysocki" <rjw@...k.pl>
CC:	Justin Mattock <justinmattock@...il.com>,
	Kasper Sandberg <lkml@...anurb.dk>,
	Jacek Luczak <difrost.kernel@...il.com>,
	Linux Kernel <linux-kernel@...r.kernel.org>
Subject: Re: BIOS bug, APIC version if 0 for CPU#0

Rafael J. Wysocki wrote:
> On Tuesday, 22 of April 2008, Justin Mattock wrote:
>> On Tue, Apr 22, 2008 at 1:49 PM, Kasper Sandberg <lkml@...anurb.dk> wrote:
>>> On Tue, 2008-04-22 at 06:10 +0000, Justin Mattock wrote:
>>>  > On Tue, Apr 22, 2008 at 5:21 AM, Jacek Luczak <difrost.kernel@...il.com> wrote:
>>>  > > Hi,
>>>  > >
>>>  > >
>>>  > >  > Hello;
>>>  > >  > with the latest git on a macbook the system can barely boot without
>>>  > >  > freezing, a work around for me at leat is either disable smp, or once
>>>  > >  > booted  turn on glxgears,or compiling the kernel i.g. when using the
>>>  > >  > processors I'm not receiving a freeze, only when everything starts
>>>  > >  > lowering down. in dmesg I'm noticing BIOS bug, APIC version is 0 for
>>>  > >  > CPU#0! fixing up to 0x10. (tell you're hw vendor) could this be the
>>>  > >  > reason for the system freezes?,
>>>  > >
>>>  > >  I don't think so. On my machine there's no freezes with latest git and:
>>>  > >  ACPI: LAPIC (acpi_id[0x00] lapic_id[0x00] enabled)
>>>  > >  BIOS bug, APIC version is 0 for CPU#0! fixing up to 0x10. (tell your hw vendor)
>>>  > >  ACPI: LAPIC (acpi_id[0x01] lapic_id[0x01] enabled)
>>>  > >  BIOS bug, APIC version is 0 for CPU#0! fixing up to 0x10. (tell your hw vendor)
>>>  > >
>>>  > >  I've got exactly the same messages.
>>>  >
>>>  > Are you using a macbook aswell?
>>>  >
>>>  > >
>>>  > >
>>>  > >  > also I didnt have a camera handy on me
>>>  > >  > but the screen spit out = Bug spinlock lockup on CPU#0
>>>  > >  > swapper/0. c17fa7c0 PID 0, comm: swapper not tainted.
>>>  > >
>>>  > >  Can you try to write down more of those bug messages?
>>>  >
>>>  > Yeah,
>>>  >
>>>  > >
>>>  > >
>>>  > >  > I dont have a problem disabling smp, except when I cat /proc/cpuinfo I
>>>  > >  > don't see anything about my cores.
>>>  > >
>>>  > >  What exactly you expect to see in /proc/cpuinfp while smp is disabled?
>>>  >
>>>  > under smp I see cpu cores. disabling smp I dont see cpu cores, under
>>>  > cat /proc/cpuinfo.
>>>
>>>  well.. SMP is kindof support for multiple processors, be it in one
>>>  socket, or multiple sockets.. without SMP you will only be able to use 1
>>>  of your X cores.
>> Yeah disabling smp is not a good idea on a macbook pro, performace is horrible.
>>
>>>  >
>>>  >
>>>  > >
>>>  > >
>>>  > >  > I know this is a stupid question
>>>  > >  > but will it hurt the system if smp is disabled
>>>  > >  > for the macbook pro ati chipset.?
>>>  > >
>>>  > >  No it won't.
>>>  > >
>>>  > >  -Jacek
>>>  > >
>>>  >
>>>  > Cool as long as the system Is o.k. disabling smp.
>>>  It is.. but with the limitations outlined above.
>>>  >   What about x86: PAT
>>>  > support is disabled. how do I enable this option?
>>>  > regards;
>>>  >
>>>
>>>
>> Thanks for the feedback, well I guess this would be either a
>> regression or a bug. I think the best would be to have somebody else
>> with a macbook pro 15' ati chipset confirm this freezeing as well,
>> before going out and jumping the gun, and filing a bug report.
>> regards;
> 
> What's the current status of this issue?

I reported this issue too but got ignored .. No one of the x86 maintainer did answer my questions.

http://www.ussg.iu.edu/hypermail/linux/kernel/0805.0/1258.html

However it does not do any bad on my boxes ( so far I can tell ) is just plain annoying and wrong some who since
is *only* occurs on 32bit kernels. HW is broken or even isn't.

It seems to me this issue is related to the move of generic_processor_info() and fiends to apic_32.c ( but I'm not really sure ) ,
all I'm sure about this message was not in 2.6.25 nor in 2.6.24.

> 
> Rafael

Regards,

Gabriel

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/