Re: [KJ] question ACPI & CMPXCHG

Re: [KJ] question ACPI & CMPXCHG

[Vicente Feito]

Sorry but... isn't already a dependency?

        depends on (SMP || PREEMPT) && X86_CMPXCHG

Or a dependency where?
Besides, replacing a warning with an error directive... yaks, that hurts 
everybodies feelings isn't?

Vicente
_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
http://lists.osdl.org/mailman/listinfo/kernel-janitors

Re: [KJ] question ACPI & CMPXCHG

[Vicente Feito]

On Sunday 27 February 2005 05:01 pm, you wrote:
> On Sun, Feb 27, 2005 at 01:49:20PM +0000, Vicente Feito wrote:
> > Sorry but... isn't already a dependency?
> >
> >         depends on (SMP || PREEMPT) && X86_CMPXCHG
> >
> > Or a dependency where?
> >...
>
> I see this only as a dependency for HAVE_DEC_LOCK, not for ACPI.
>
> Where did you see this as an ACPI dependency?
Adrian,
No, you're right, it's only a HAVE_DEC_LOCK, I've grepped too fast... my bad

Walter, I agree with putting that in the Kconfig, but I still think an error 
it's more scary than a warning, most people don't care that much about 
wasting some compile time, but they do care (and freaks out) when an error 
appears in the compilation process...
Users won't complain for a warning if the overall works fine, but if you give 
them errors, well, they're not going to be happy... (users/admins/people that 
aren't crazy for optimization)

Vicente
_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
http://lists.osdl.org/mailman/listinfo/kernel-janitors

[KJ] question ACPI & CMPXCHG

[walter harms]

[-- Attachment #1: Type: text/plain, Size: 824 bytes --]

hi list,
when i was looking into arch/i386/kernel/acpi/boot.c

i found:

#ifndef __HAVE_ARCH_CMPXCHG
#warning ACPI uses CMPXCHG, i486 and later hardware
#endif


perhaps its nitpicking but ...
I think the setup should disable ACPI if CMPXCHG does not exist.
second best generate at least an error ! compile time these day is short 
and you have a good change to miss warnings.

re,
	walter


(since i have no idea about the setup, no patch)

re,
	walter


quick fix (better example) , hopyfully someone can find some better text


--- boot.c.org  2005-02-27 17:10:24.000000000 +0100
+++ boot.c      2005-02-27 17:11:58.000000000 +0100
@@ -82,7 +82,7 @@
  #endif

  #ifndef __HAVE_ARCH_CMPXCHG
-#warning ACPI uses CMPXCHG, i486 and later hardware
+#error  ACPI requires CMPXCHG
  #endif

  #define MAX_MADT_ENTRIES       256

[-- Attachment #2: Type: text/plain, Size: 167 bytes --]

_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
http://lists.osdl.org/mailman/listinfo/kernel-janitors

Re: [KJ] question ACPI & CMPXCHG

[Adrian Bunk]

[-- Attachment #1: Type: text/plain, Size: 1319 bytes --]

On Sun, Feb 27, 2005 at 05:17:36PM +0100, walter harms wrote:
> hi list,
> when i was looking into arch/i386/kernel/acpi/boot.c
> 
> i found:
> 
> #ifndef __HAVE_ARCH_CMPXCHG
> #warning ACPI uses CMPXCHG, i486 and later hardware
> #endif
> 
> 
> perhaps its nitpicking but ...
> I think the setup should disable ACPI if CMPXCHG does not exist.
> second best generate at least an error ! compile time these day is short 
> and you have a good change to miss warnings.
>...
> --- boot.c.org  2005-02-27 17:10:24.000000000 +0100
> +++ boot.c      2005-02-27 17:11:58.000000000 +0100
> @@ -82,7 +82,7 @@
>  #endif
> 
>  #ifndef __HAVE_ARCH_CMPXCHG
> -#warning ACPI uses CMPXCHG, i486 and later hardware
> +#error  ACPI requires CMPXCHG
>  #endif
>...

This is wrong - instead of an error this should be put as a dependency 
into the Kconfig file.

But the underlying issue is that the current config semantics is, that 
"386" means a generic kernel. I should make a new version of my patch of 
a new i386 cpu selection for allowing a proper handling of such cases.

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed


[-- Attachment #2: Type: text/plain, Size: 167 bytes --]

_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
http://lists.osdl.org/mailman/listinfo/kernel-janitors

Re: [KJ] question ACPI & CMPXCHG

[Adrian Bunk]

[-- Attachment #1: Type: text/plain, Size: 578 bytes --]

On Sun, Feb 27, 2005 at 01:49:20PM +0000, Vicente Feito wrote:

> Sorry but... isn't already a dependency?
> 
>         depends on (SMP || PREEMPT) && X86_CMPXCHG
> 
> Or a dependency where?
>...

I see this only as a dependency for HAVE_DEC_LOCK, not for ACPI.

Where did you see this as an ACPI dependency?

> Vicente

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed


[-- Attachment #2: Type: text/plain, Size: 167 bytes --]

_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
http://lists.osdl.org/mailman/listinfo/kernel-janitors