FW: I think I've found a serious bug in AMD Athlon page_alloc.c routines, where do I mail the developer(s) ?

FW: I think I've found a serious bug in AMD Athlon page_alloc.c routines, where do I mail the developer(s) ?

[David Wilson]

Hi all, howzit going ? ;-)

Maybe I'm being dumb but I've been working with Linux long enough not to
make a silly mistake *I hope*.

I think I've found a serious bug in AMD Athlon page_alloc.c routines in
kernel 2.4.4.
I'm running an AMD athlon 850 running at 100x8.5, everything is setup
correct on the DFI AK75-EC motherboard, if I set the CPU kernel type to 586
everything is 100%, if I use "Athlon" kernel type I get:
kernel BUG at page_alloc.c:73

The system then dies with:
exit mmap: map count is 13
I also get various "unable to handle paging requests" etc. etc.

I've changed RAM, Motherboard etc... still the same.
Also the same system runs linux-2.2.16 100%

I'm not subscribed to linux-kernel@vger.kernel.org, so please cc me in your
reply to the list.
Thanks.

Regards
David Wilson
Technical Support Centre
The S.A Internet
0860 100 869
http://www.sai.co.za

Re: FW: I think I've found a serious bug in AMD Athlon page_alloc.c routines, where do I mail the developer(s) ?

[Mark Hahn]

> I think I've found a serious bug in AMD Athlon page_alloc.c routines in

there's nothing athlon-specific there.

> correct on the DFI AK75-EC motherboard, if I set the CPU kernel type to 586
> everything is 100%, if I use "Athlon" kernel type I get:
> kernel BUG at page_alloc.c:73

when you select athlon at compile time, you're mainly 
getting Arjan's athlon-specific page-clear and -copy functions
(along with some relatively trivial alignment changes).
these functions are ~3x as fast as the generic ones,
and seem to cause dram/cpu-related oopes on some machines.

in short: faster code pushes the hardware past stability.
there's no reason, so far, to think that there's anything 
wrong with the code - Alan had a possible issue with prefetching
and very old Atlons, but the people reporting problems like this
are actually running kt133a and new fsb133 Athlons.

> I've changed RAM, Motherboard etc... still the same.

changed to a non-kt133a board?  how about running fsb and/or dram
at 100, rather than 133?

> Also the same system runs linux-2.2.16 100%

2.2 doesn't have the fast page-clear and -copy code afaik.

afaik, there are *no* problems on kt133 machines,
and haven't heard any pain from people who might have 
Ali Magic1, AMD 760 or KT266 boards, but they're still rare.

RE: FW: I think I've found a serious bug in AMD Athlon page_alloc.c routines, where do I mail the developer(s) ?

[David Wilson]

Hi Mark,

Thanks for getting back to me, yea I'm running my fsb at 100
already....still seems to be the same problem.
I donno, I'm just gonna have to mess around with things till I get it right
;-)

Thanks for the guidance anyways.


Regards
David Wilson
Technical Support Centre
The S.A Internet
0860 100 869
http://www.sai.co.za



-----Original Message-----
From: Mark Hahn [mailto:hahn@coffee.psychology.mcmaster.ca]
Sent: 16 May 2001 01:01
To: David Wilson
Cc: linux-kernel@vger.kernel.org
Subject: Re: FW: I think I've found a serious bug in AMD Athlon
page_alloc.c routines, where do I mail the developer(s) ?


> I think I've found a serious bug in AMD Athlon page_alloc.c routines in

there's nothing athlon-specific there.

> correct on the DFI AK75-EC motherboard, if I set the CPU kernel type to
586
> everything is 100%, if I use "Athlon" kernel type I get:
> kernel BUG at page_alloc.c:73

when you select athlon at compile time, you're mainly
getting Arjan's athlon-specific page-clear and -copy functions
(along with some relatively trivial alignment changes).
these functions are ~3x as fast as the generic ones,
and seem to cause dram/cpu-related oopes on some machines.

in short: faster code pushes the hardware past stability.
there's no reason, so far, to think that there's anything
wrong with the code - Alan had a possible issue with prefetching
and very old Atlons, but the people reporting problems like this
are actually running kt133a and new fsb133 Athlons.

> I've changed RAM, Motherboard etc... still the same.

changed to a non-kt133a board?  how about running fsb and/or dram
at 100, rather than 133?

> Also the same system runs linux-2.2.16 100%

2.2 doesn't have the fast page-clear and -copy code afaik.

afaik, there are *no* problems on kt133 machines,
and haven't heard any pain from people who might have
Ali Magic1, AMD 760 or KT266 boards, but they're still rare.

RE: FW: I think I've found a serious bug in AMD Athlon page_alloc.c routines, where do I mail the developer(s) ?

[David Wilson]

Hi Alan,

Thanks for getting back to me.
I wonder if DFI has a bios or chipset patch available and whether that would
help ?
Maybe disabling the VIA chipset support in the kernel and running generic
drivers would help ?
Thanks.
Keep up the good work anyways !

Regards
David Wilson
Technical Support Centre
The S.A Internet
0860 100 869
http://www.sai.co.za



-----Original Message-----
From: Alan Cox [mailto:alan@lxorguk.ukuu.org.uk]
Sent: 16 May 2001 12:54
To: David Wilson
Subject: Re: FW: I think I've found a serious bug in AMD Athlon
page_alloc.c routines, where do I mail the developer(s) ?


Known funny. Only shows up on via chipset boards. Right now we think but
have
not proven its a hardware flaw somwhere

Re: FW: I think I've found a serious bug in AMD Athlon page_alloc.c routines, where do I mail the developer(s) ?

[Alan Cox]

> I wonder if DFI has a bios or chipset patch available and whether that would
> help ?
> Maybe disabling the VIA chipset support in the kernel and running generic
> drivers would help ?

Play with ideas see what you find out. You might strike lucky. So far nobody
else has

RE: FW: I think I've found a serious bug in AMD Athlon page_alloc.c routines, where do I mail the developer(s) ?

[David Wilson]

Hi Alan,

Mmm no luck so far ;(
I've checked that the bios is the latest available and there doesn't seem to
be any type of chipset software patch or upgrade.
What I have done is tried compiling the kernel for AMD K6/2, K6/3, .....this
seems to work nicely and it's stable.
It seems that I'll have to give the "Athlon" support a miss for now. ;(

How does one go about obtaining a Transmeta chip ? I'm really keen to try it
;-) ...anyways probably just a fantasy for now.

Thanks for the help anyways and keep up the fantastic work !

Regards
David Wilson
Technical Support Centre
The S.A Internet
0860 100 869
http://www.sai.co.za



-----Original Message-----
From: Alan Cox [mailto:alan@lxorguk.ukuu.org.uk]
Sent: 16 May 2001 10:35
To: David Wilson
Cc: Alan Cox; linux-kernel@vger.kernel.org
Subject: Re: FW: I think I've found a serious bug in AMD Athlon
page_alloc.c routines, where do I mail the developer(s) ?


> I wonder if DFI has a bios or chipset patch available and whether that
would
> help ?
> Maybe disabling the VIA chipset support in the kernel and running generic
> drivers would help ?

Play with ideas see what you find out. You might strike lucky. So far nobody
else has