Bugs on MacBook, and proposed code change

Bugs on MacBook, and proposed code change

[Huihong Luo]

I tried to compile kvm on an Intel Mac (MacBook laptop
runing Linux 2.6.17). It does not work, qemu simply
hangs.

I spent sometime to fix the issues, can you correct
the issues? There are to do with the kernel module.
After the following 2 corrections, I can launch
windows XP SP2 on a MacBook.

(1) hvm_main.c

There seems to be extra put_cpu()
static void vcpu_put(struct kvm_vcpu *vcpu)
{
	kvm_arch_ops->vcpu_put(vcpu);
	/* put_cpu(); extra one Huihong */
	mutex_unlock(&vcpu->mutex);
}

(2) in vmx.c:

static __init void hardware_enable(void *garbage)
{
	int cpu = raw_smp_processor_id();
	u64 phys_addr = __pa(per_cpu(vmxarea, cpu));
	u32 eax, edx;

	rdmsr(MSR_IA32_FEATURE_CONTROL, eax, edx);
		
	if (eax & 1) {
		if ((eax & 4) == 0) {
			printk("VMX disabled by Feature Control MSR.\n");
			return;
		}
	}
	else {
		/* enable and lock */
		wrmsr(MSR_IA32_FEATURE_CONTROL, 5, 0);
	}

#if 0 /* Huihong, original */
	if ((old & 5) == 0)
		/* enable and lock */
		wrmsrl(MSR_IA32_FEATURE_CONTROL, old | 5);
#endif

Huihong
	


 
____________________________________________________________________________________
Need a quick answer? Get one in minutes from people who know.
Ask your question on www.Answers.yahoo.com

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

Re: Bugs on MacBook, and proposed code change

[Avi Kivity]

Huihong Luo wrote:
> I tried to compile kvm on an Intel Mac (MacBook laptop
> runing Linux 2.6.17). It does not work, qemu simply
> hangs.
>
> I spent sometime to fix the issues, can you correct
> the issues? There are to do with the kernel module.
> After the following 2 corrections, I can launch
> windows XP SP2 on a MacBook.
>
> (1) hvm_main.c
>
> There seems to be extra put_cpu()
> static void vcpu_put(struct kvm_vcpu *vcpu)
> {
> 	kvm_arch_ops->vcpu_put(vcpu);
> 	/* put_cpu(); extra one Huihong */
> 	mutex_unlock(&vcpu->mutex);
> }
>   

Yup, a bug.  Will fix.


> (2) in vmx.c:
>
> static __init void hardware_enable(void *garbage)
> {
> 	int cpu = raw_smp_processor_id();
> 	u64 phys_addr = __pa(per_cpu(vmxarea, cpu));
> 	u32 eax, edx;
>
> 	rdmsr(MSR_IA32_FEATURE_CONTROL, eax, edx);
> 		
> 	if (eax & 1) {
> 		if ((eax & 4) == 0) {
> 			printk("VMX disabled by Feature Control MSR.\n");
> 			return;
> 		}
> 	}
> 	else {
> 		/* enable and lock */
> 		wrmsr(MSR_IA32_FEATURE_CONTROL, 5, 0);
> 	}
>   

I don't understand this one.

There are 4 possible values of [FEATURE_CONTROL.2, FEATURE_CONTROL.]:

00: disabled, unlocked: handled correctly by the original code
01: disabled, locked: cannot be handled, but should be caught by 
vmx_disabled_by_bios().
10: enabled, unlocked: handled correctly by the original code
11: enabled, locked: handled correctly by the original code

No?

> #if 0 /* Huihong, original */
> 	if ((old & 5) == 0)
> 		/* enable and lock */
> 		wrmsrl(MSR_IA32_FEATURE_CONTROL, old | 5);
> #endif
>
> Huihong
> 	
>
>   

-- 
error compiling committee.c: too many arguments to function


-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV