[PATCH] x86: enable CPUID on Cyrix cpus with CPUID disabled

[PATCH] x86: enable CPUID on Cyrix cpus with CPUID disabled

[Krzysztof Helt]

From: Krzysztof Helt <krzysztof.h1@wp.pl>

The macros setCx86_old() and getCx86_old() do not
work as a comment says in the processor-cyrix.h.

Use working inline functions instead to enable
the CPUID instriction on Cyrix cpus. 

Tested on IBM 6x86MX cpu with disabled the
CPUID instruction.

Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl>
---

This patch is against the linux-next tree (20080819)
as it contains the following patch:

x86, cyrix: debug
Signed-off-by: Ingo Molnar <mingo@elte.hu>

http://git.kernel.org/?p=linux/kernel/git/sfr/linux-next.git;a=commit;h=026e2c05ef58ef413e2d52696f125d5ea1aa8bce

The linux-2.6-tip tree has not the Ingo's patch so it 
does not require my patch.

I am confused. Is the x86 detection code for Cyrix
cpus is right in the linux-next or in the linux-2.6-tip?
Should I prepare patches against the linux-2.6-tip
or linux-next?

diff -urp linux-orig/arch/x86/kernel/cpu/cyrix.c linux-2.6.27/arch/x86/kernel/cpu/cyrix.c
--- linux-orig/arch/x86/kernel/cpu/cyrix.c	2008-09-29 00:16:31.000000000 +0200
+++ linux-2.6.27/arch/x86/kernel/cpu/cyrix.c	2008-09-29 17:03:52.563864398 +0200
@@ -428,7 +428,7 @@ static void __cpuinit cyrix_identify(str
 			local_irq_save(flags);
 			ccr3 = getCx86(CX86_CCR3);
 			setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10);       /* enable MAPEN  */
-			setCx86_old(CX86_CCR4, getCx86_old(CX86_CCR4) | 0x80);  /* enable cpuid  */
+			setCx86(CX86_CCR4, getCx86(CX86_CCR4) | 0x80);  /* enable cpuid  */
 			setCx86(CX86_CCR3, ccr3);                       /* disable MAPEN */
 			local_irq_restore(flags);
 		}

----------------------------------------------------------------------
Dzwon taniej na zagraniczne komorki!
Sprawdz >> http://link.interia.pl/f1f26 

Re: [PATCH] x86: enable CPUID on Cyrix cpus with CPUID disabled

[Yinghai Lu]

On Mon, Sep 29, 2008 at 10:24 AM, Krzysztof Helt <krzysztof.h1@poczta.fm> wrote:
> From: Krzysztof Helt <krzysztof.h1@wp.pl>
>
> The macros setCx86_old() and getCx86_old() do not
> work as a comment says in the processor-cyrix.h.
>
> Use working inline functions instead to enable
> the CPUID instriction on Cyrix cpus.
>
> Tested on IBM 6x86MX cpu with disabled the
> CPUID instruction.
>
> Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl>
> ---
>
> This patch is against the linux-next tree (20080819)
> as it contains the following patch:
>
> x86, cyrix: debug
> Signed-off-by: Ingo Molnar <mingo@elte.hu>
>
> http://git.kernel.org/?p=linux/kernel/git/sfr/linux-next.git;a=commit;h=026e2c05ef58ef413e2d52696f125d5ea1aa8bce
>
> The linux-2.6-tip tree has not the Ingo's patch so it
> does not require my patch.
>
> I am confused. Is the x86 detection code for Cyrix
> cpus is right in the linux-next or in the linux-2.6-tip?
> Should I prepare patches against the linux-2.6-tip
> or linux-next?

linux-2.6-tip.

YH

Re: [PATCH] x86: enable CPUID on Cyrix cpus with CPUID disabled

[Ingo Molnar]

* Krzysztof Helt <krzysztof.h1@poczta.fm> wrote:

> From: Krzysztof Helt <krzysztof.h1@wp.pl>
> 
> The macros setCx86_old() and getCx86_old() do not
> work as a comment says in the processor-cyrix.h.
> 
> Use working inline functions instead to enable
> the CPUID instriction on Cyrix cpus. 
> 
> Tested on IBM 6x86MX cpu with disabled the
> CPUID instruction.
> 
> Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl>
> ---
> 
> This patch is against the linux-next tree (20080819)
> as it contains the following patch:
> 
> x86, cyrix: debug
> Signed-off-by: Ingo Molnar <mingo@elte.hu>
> 
> http://git.kernel.org/?p=linux/kernel/git/sfr/linux-next.git;a=commit;h=026e2c05ef58ef413e2d52696f125d5ea1aa8bce
> 
> The linux-2.6-tip tree has not the Ingo's patch so it 
> does not require my patch.
> 
> I am confused. Is the x86 detection code for Cyrix
> cpus is right in the linux-next or in the linux-2.6-tip?
> Should I prepare patches against the linux-2.6-tip
> or linux-next?

now that the bug is correctly understood and fixed, could you please 
send a patch that cleans it all up and gets rid of the setCx86_old() / 
setCx86() API distinction? I.e. revert commit 026e2c0 ("x86, cyrix: 
debug"). Against tip/master please.

note that this was all done in connection to:

 commit c6744955d0ec0cb485c28c51eeb7185e260f6172
 Author: Samuel Sieb <samuel@sieb.net>
 Date:   Wed Aug 6 22:06:29 2008 -0700

    x86: fix "kernel won't boot on a Cyrix MediaGXm (Geode)"

	Ingo

Re: [PATCH] x86: enable CPUID on Cyrix cpus with CPUID disabled

[krzysztof.h1]

> 
> * Krzysztof Helt <krzysztof.h1@poczta.fm> wrote:
> 
> > From: Krzysztof Helt <krzysztof.h1@wp.pl>
> > 
> > The macros setCx86_old() and getCx86_old() do not
> > work as a comment says in the processor-cyrix.h.
> > 
> > Use working inline functions instead to enable
> > the CPUID instriction on Cyrix cpus. 
> > 
> > Tested on IBM 6x86MX cpu with disabled the
> > CPUID instruction.
> > 
> > Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl>
> > ---
> > 
> > This patch is against the linux-next tree (20080819)
> > as it contains the following patch:
> > 
> > x86, cyrix: debug
> > Signed-off-by: Ingo Molnar <mingo@elte.hu>
> > 
> >
> http://git.kernel.org/?p=linux/kernel/git/sfr/linux-next.git;a=commit;h=026e2c05ef58ef413e2d52696f125d5ea1aa8bce
> > 
> > The linux-2.6-tip tree has not the Ingo's patch so it 
> > does not require my patch.
> > 
> > I am confused. Is the x86 detection code for Cyrix
> > cpus is right in the linux-next or in the linux-2.6-tip?
> > Should I prepare patches against the linux-2.6-tip
> > or linux-next?
> 
> now that the bug is correctly understood and fixed, could you please 
> send a patch that cleans it all up and gets rid of the setCx86_old() / 
> setCx86() API distinction? I.e. revert commit 026e2c0 ("x86, cyrix: 
> debug"). Against tip/master please.
> 
> note that this was all done in connection to:
> 
>  commit c6744955d0ec0cb485c28c51eeb7185e260f6172
>  Author: Samuel Sieb <samuel@sieb.net>
>  Date:   Wed Aug 6 22:06:29 2008 -0700
> 
>     x86: fix "kernel won't boot on a Cyrix MediaGXm (Geode)"
> 

I know that it was connected with the Geode booting, but I read the cyrix.c file and most paths are Geode specific and I won't be able to test them. I tested that reverting the commit 026e2c0 ("x86, cyrix: debug") does not break anything on 6x86MX cpu.

If the tip/master is a correct tree, please revert the commit 026e2c0 in the linux-next and drop this patch. This way everything is correct.

Regards,
Krzysztof

----------------------------------------------------------------------
Dzwon taniej na zagraniczne komorki!
Sprawdz >> http://link.interia.pl/f1f26