K6 sig11 Bug detection.

K6 sig11 Bug detection.

[James Nord]

Hi,

I have a Debian based system with a custom 2.4.7 kernel compiled with 
gcc version 2.95.4 20010703 (Debian prerelease)
(also saw the same with 2.2.10+ gcc 2.95)

The CPU in the machine is a AMD K6 200MHz, and has 64MB of SDRAM

I reomved the heatsink and the serial is Cxxxxxxxx, however in the 
kernel boot messages I get the following,

Aug 17 09:06:49 phoenix kernel: CPU: Before vendor init, caps: 008001bf 
008005bf 00000000, vendor = 2
Aug 17 09:06:49 phoenix kernel: AMD K6 stepping B detected - <6>K6 BUG 
9016725 20000000 (Report these if test report is incorrect)
Aug 17 09:06:49 phoenix kernel: AMD K6 stepping B detected - probably OK 
(after B9730xxxx).
Aug 17 09:06:49 phoenix kernel: Please see 
http://www.mygale.com/~poulot/k6bug.html
Aug 17 09:06:49 phoenix kernel: CPU: L1 I Cache: 32K (32 bytes/line), D 
cache 32K (32 bytes/line)
Aug 17 09:06:49 phoenix kernel: CPU: After vendor init, caps: 008001bf 
008005bf 00000000 00000000
Aug 17 09:06:49 phoenix kernel: CPU:     After generic, caps: 008001bf 
008005bf 00000000 00000000
Aug 17 09:06:49 phoenix kernel: CPU:             Common caps: 008001bf 
008005bf 00000000 00000000
Aug 17 09:06:49 phoenix kernel: CPU: AMD-K6tm w/ multimedia extensions 
stepping 01

Is the stepping not the first part of the serial? I have 64MB (the 
amount that triggers the bug IIRC) in the system and the the kernel and 
everything else compiles without generating a SIG11.

Is this a false detection or would it be possible that I have a wrongly 
tagged CPU?

Also the link http://www.mygale.com/~poulot/k6bug.html does not exist.

Please CC replies to me as I am not on the list.

Regards,

    /James

-- 
Technology is a word that describes something that doesn't work yet.
	Douglas Adams

Re: K6 sig11 Bug detection.

[Brian Gerst]

James Nord wrote:
> 
> Hi,
> 
> I have a Debian based system with a custom 2.4.7 kernel compiled with
> gcc version 2.95.4 20010703 (Debian prerelease)
> (also saw the same with 2.2.10+ gcc 2.95)
> 
> The CPU in the machine is a AMD K6 200MHz, and has 64MB of SDRAM
> 
> I reomved the heatsink and the serial is Cxxxxxxxx, however in the
> kernel boot messages I get the following,
> 
> Aug 17 09:06:49 phoenix kernel: CPU: Before vendor init, caps: 008001bf
> 008005bf 00000000, vendor = 2
> Aug 17 09:06:49 phoenix kernel: AMD K6 stepping B detected - <6>K6 BUG
> 9016725 20000000 (Report these if test report is incorrect)
> Aug 17 09:06:49 phoenix kernel: AMD K6 stepping B detected - probably OK
> (after B9730xxxx).
> Aug 17 09:06:49 phoenix kernel: Please see
> http://www.mygale.com/~poulot/k6bug.html
> Aug 17 09:06:49 phoenix kernel: CPU: L1 I Cache: 32K (32 bytes/line), D
> cache 32K (32 bytes/line)
> Aug 17 09:06:49 phoenix kernel: CPU: After vendor init, caps: 008001bf
> 008005bf 00000000 00000000
> Aug 17 09:06:49 phoenix kernel: CPU:     After generic, caps: 008001bf
> 008005bf 00000000 00000000
> Aug 17 09:06:49 phoenix kernel: CPU:             Common caps: 008001bf
> 008005bf 00000000 00000000
> Aug 17 09:06:49 phoenix kernel: CPU: AMD-K6tm w/ multimedia extensions
> stepping 01
> 
> Is the stepping not the first part of the serial? I have 64MB (the
> amount that triggers the bug IIRC) in the system and the the kernel and
> everything else compiles without generating a SIG11.
> 
> Is this a false detection or would it be possible that I have a wrongly
> tagged CPU?
> 
> Also the link http://www.mygale.com/~poulot/k6bug.html does not exist.
> 
> Please CC replies to me as I am not on the list.

You have a later B step CPU with that bug fixed.  Otherwise you would
have seen "system stability may be impaired when more than 32 MB are
used." in the logs.  For some reason, AMD didn't bump the stepping when
they fixed that bug.

-- 

						Brian Gerst

[New URL FOR K6 bug] Re: K6 sig11 Bug detection.

[André Dahlqvist]

> Also the link http://www.mygale.com/~poulot/k6bug.html does not exist.

New URL seams to be http://www.multimania.com/poulot/k6bug.html.

Alan, can you correct this in your next -ac?
-- 

André Dahlqvist <andre.dahlqvist@telia.com>