Re: [PATCH 9/9] KVM: x86: smsw emulation is incorrect in 64-bit mode

[Paolo Bonzini <pbonzini@redhat.com>]

Il 05/06/2014 17:04, H. Peter Anvin ha scritto:
> On 06/05/2014 08:02 AM, Nadav Amit wrote:
>>> I'm sorry, I'm missing the place where 64-bit mode is taken into account?
>> It is not, since on 32-bit mode the high-order 16 bits of a register destination are undefined.
>> If I recall correctly, in this case the high-order 16-bits on native
> system actually reflect the high-order 16-bits of CR0.
>
> This sounds like something that really should be verified
> experimentally.  The above claim seems... odd.

Here is a test program:

#if __SIZEOF_LONG__ == 4
#define V "12345678"
#define R "e"
#else
#define V "1234567812345678"
#define R "r"
#endif

#include <stdio.h>
int main()
{
	register volatile unsigned long ecx asm("ecx");

#if __SIZEOF_LONG__ > 4
	asm volatile("mov $0x" V ", %%" R "cx; smswq %%rcx": : :"ecx");
	printf("smswq: %lx\n", ecx);
#endif

	asm volatile("mov $0x" V ", %%" R "cx; smswl %%ecx": : :"ecx");
	printf("smswl: %lx\n", ecx);

	asm volatile("mov $0x" V ", %%" R "cx; smsww %%cx": : :"ecx");
	printf("smsww: %lx\n", ecx);
}

Output in 32-bit mode:
smswq: 80050033
smswl: 12340033

Output in 64-bit mode:
smswq: 80050033
smswl: 80050033
smsww: 1234567812340033

Can you please make a test case for kvm-unit-tests (x86/emulator.c), in 
order to check the validity of the patch?

Paolo