[PATCH] Longhaul - Protect user from himself

[PATCH] Longhaul - Protect user from himself

[Rafał Bilski]

This patch depens on "Longhaul - Use hardware support".

Older CPU's require transition time >=1ms.
Print message for C7 user (this is based on real story).

Signed-off-by: Rafa³ Bilski <rafalbilski@interia.pl>

---

--- linux-2.6.17-git20/arch/i386/kernel/cpu/cpufreq/longhaul.c.orig	2006-07-03 06:43:54.000000000 +0200
+++ linux-2.6.17-git20/arch/i386/kernel/cpu/cpufreq/longhaul.c	2006-07-03 07:33:38.000000000 +0200
@@ -633,6 +633,13 @@ static int __init longhaul_cpu_init(stru
 		break;
 	};
 
+#if (CONFIG_HZ > 250)
+	if (longhaul_version != TYPE_POWERSAVER) {
+		printk (KERN_INFO PFX "Timer frequency to high. Aborting\n");
+		return -EINVAL;
+	}
+#endif
+
 	ret = longhaul_get_ranges();
 	if (ret != 0)
 		return ret;
@@ -691,8 +698,9 @@ static int __init longhaul_init(void)
 	switch (c->x86_model) {
 	case 6 ... 9:
 		return cpufreq_register_driver(&longhaul_driver);
-	default:
-		printk (KERN_INFO PFX "Unknown VIA CPU. Contact davej@codemonkey.org.uk\n");
+	}
+	if (c->x86_model > 9) {
+		printk(KERN_INFO PFX "VIA C7 (or better) CPU. Use \"ACPI Processor P-States driver\".\n");
 	}
 
 	return -ENODEV;

------------------------------------------------------------------------
Zobacz, gdzie mieszka Twoja dziewczyna!
http://map24.interia.pl

Re: [PATCH] Longhaul - Protect user from himself

[Dave Jones]

On Mon, Jul 03, 2006 at 07:39:08AM +0200, Rafał Bilski wrote:
 > This patch depens on "Longhaul - Use hardware support".
 > 
 > Older CPU's require transition time >=1ms.
 > Print message for C7 user (this is based on real story).
 > 
 > Signed-off-by: Rafał Bilski <rafalbilski@interia.pl>
 > 
 > ---
 > 
 > --- linux-2.6.17-git20/arch/i386/kernel/cpu/cpufreq/longhaul.c.orig	2006-07-03 06:43:54.000000000 +0200
 > +++ linux-2.6.17-git20/arch/i386/kernel/cpu/cpufreq/longhaul.c	2006-07-03 07:33:38.000000000 +0200
 > @@ -633,6 +633,13 @@ static int __init longhaul_cpu_init(stru
 >  		break;
 >  	};
 >  
 > +#if (CONFIG_HZ > 250)
 > +	if (longhaul_version != TYPE_POWERSAVER) {
 > +		printk (KERN_INFO PFX "Timer frequency to high. Aborting\n");
 > +		return -EINVAL;
 > +	}
 > +#endif

This seems wrong.  We should adjust the timings based on HZ to sleep 
for less/more instead.

 > @@ -691,8 +698,9 @@ static int __init longhaul_init(void)
 >  	switch (c->x86_model) {
 >  	case 6 ... 9:
 >  		return cpufreq_register_driver(&longhaul_driver);
 > -	default:
 > -		printk (KERN_INFO PFX "Unknown VIA CPU. Contact davej@codemonkey.org.uk\n");
 > +	}
 > +	if (c->x86_model > 9) {
 > +		printk(KERN_INFO PFX "VIA C7 (or better) CPU. Use \"ACPI Processor P-States driver\".\n");

It can also use speedstep-centrino with some additional patches which I
need to dust off.

		Dave

-- 
http://www.codemonkey.org.uk