public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* 2.4.19-pre10-ac1: Hardcoded cpu_khz in powernow-k6.c
@ 2002-06-05  0:52 Adrian Bunk
  2002-06-05  1:02 ` Dave Jones
  0 siblings, 1 reply; 3+ messages in thread
From: Adrian Bunk @ 2002-06-05  0:52 UTC (permalink / raw)
  To: Dave Jones; +Cc: linux-kernel

Hi Dave,

while reading through powernow-k6.c in 2.4.19-pre10-ac1 I found the
following that seems to be a bug:

  static unsigned long cpu_khz=350000;

Not every K6-2/3 runs at 350 MHz...


cu
Adrian

-- 

You only think this is a free country. Like the US the UK spends a lot of
time explaining its a free country because its a police state.
								Alan Cox




^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: 2.4.19-pre10-ac1: Hardcoded cpu_khz in powernow-k6.c
  2002-06-05  0:52 2.4.19-pre10-ac1: Hardcoded cpu_khz in powernow-k6.c Adrian Bunk
@ 2002-06-05  1:02 ` Dave Jones
  2002-06-05  6:04   ` Adrian Bunk
  0 siblings, 1 reply; 3+ messages in thread
From: Dave Jones @ 2002-06-05  1:02 UTC (permalink / raw)
  To: Adrian Bunk; +Cc: linux-kernel, arjanv

On Wed, Jun 05, 2002 at 02:52:15AM +0200, Adrian Bunk wrote:
 > Hi Dave,
 > 
 > while reading through powernow-k6.c in 2.4.19-pre10-ac1 I found the
 > following that seems to be a bug:
 > 
 >   static unsigned long cpu_khz=350000;
 > 
 > Not every K6-2/3 runs at 350 MHz...

iirc, there aren't any MSRs[*] on the K6-2 where we can read
the current FSB.  I think 350MHz was used as it was probably
the slowest K6-2 to be found at the time.  You can override
it with boot time arguments.

    Dave.

[*] The K6 style powernow was reverse engineered, as there were
no publically available documents explaining it. All we can
do is scale multipliers. No voltage scaling, no FSB decoding.

-- 
| Dave Jones.        http://www.codemonkey.org.uk
| SuSE Labs

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: 2.4.19-pre10-ac1: Hardcoded cpu_khz in powernow-k6.c
  2002-06-05  1:02 ` Dave Jones
@ 2002-06-05  6:04   ` Adrian Bunk
  0 siblings, 0 replies; 3+ messages in thread
From: Adrian Bunk @ 2002-06-05  6:04 UTC (permalink / raw)
  To: Dave Jones; +Cc: linux-kernel, arjanv

On Wed, 5 Jun 2002, Dave Jones wrote:

>  > while reading through powernow-k6.c in 2.4.19-pre10-ac1 I found the
>  > following that seems to be a bug:
>  >
>  >   static unsigned long cpu_khz=350000;
>  >
>  > Not every K6-2/3 runs at 350 MHz...
>
> iirc, there aren't any MSRs[*] on the K6-2 where we can read
> the current FSB.  I think 350MHz was used as it was probably
> the slowest K6-2 to be found at the time.  You can override
> it with boot time arguments.

Really? From reading the code I have the impression that cpu_khz holds
more or less the information a "cat /proc/cpuinfo | grep MHz" gives for a
CPU at it's "normal" speed. The only place where cpu_khz is used is in

  busfreq = cpu_khz / get_cpu_multiplier() / 1000;

and this gives IMHO a wrong result if the "normal" speed of your CPU is
different from 350 MHz.

>     Dave.

cu
Adrian

BTW: 350MHz was never the slowest K6-2, there are K6-2 @ 300 MHz.

-- 

You only think this is a free country. Like the US the UK spends a lot of
time explaining its a free country because its a police state.
								Alan Cox


^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2002-06-05  6:04 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2002-06-05  0:52 2.4.19-pre10-ac1: Hardcoded cpu_khz in powernow-k6.c Adrian Bunk
2002-06-05  1:02 ` Dave Jones
2002-06-05  6:04   ` Adrian Bunk

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox