Re: [PATCH] qemu-kvm: remove CPUID host hacks

From: Alexander Graf <agraf@suse.de>
To: Andre Przywara <andre.przywara@amd.com>
Cc: avi@redhat.com, kvm@vger.kernel.org
Subject: Re: [PATCH] qemu-kvm: remove CPUID host hacks
Date: Wed, 1 Jul 2009 09:42:54 +0200	[thread overview]
Message-ID: <6D744C96-C9A2-4430-951A-AB90154F13F3@suse.de> (raw)
In-Reply-To: <1246432947-8693-1-git-send-email-andre.przywara@amd.com>

On 01.07.2009, at 09:22, Andre Przywara wrote:

> KVM provides an in-kernel feature to disable CPUID bits that are not
> present in the current host. So there is no need here to duplicate  
> this
> work. Additionally allows 3DNow! on capable processors, since the
> restriction seems to apply to QEMU/TCG only.
>
> Signed-off-by: Andre Przywara <andre.przywara@amd.com>
> ---
> target-i386/helper.c |   22 +---------------------
> 1 files changed, 1 insertions(+), 21 deletions(-)
>
> This is a rebased version of the qemu.git patch, which caused merge  
> conflicts.
> Please apply!
>
> Regards,
> Andre.
>
> diff --git a/target-i386/helper.c b/target-i386/helper.c
> index d76c224..30e1d77 100644
> --- a/target-i386/helper.c
> +++ b/target-i386/helper.c
> @@ -1645,29 +1645,9 @@ void cpu_x86_cpuid(CPUX86State *env, uint32_t  
> index, uint32_t count,
>         *edx = env->cpuid_ext2_features;
>
>         if (kvm_enabled()) {
> -            uint32_t h_eax, h_edx;
> -
> -            host_cpuid(index, 0, &h_eax, NULL, NULL, &h_edx);
> -
> -            /* disable CPU features that the host does not support */
> -
> -            /* long mode */
> -            if ((h_edx & 0x20000000) == 0 /* || !lm_capable_kernel  
> */)
> -                *edx &= ~0x20000000;
> -            /* syscall */
> -            if ((h_edx & 0x00000800) == 0)
> -                *edx &= ~0x00000800;
> -            /* nx */
> -            if ((h_edx & 0x00100000) == 0)
> -                *edx &= ~0x00100000;
> -
> -            /* disable CPU features that KVM cannot support */
> -
> -            /* svm */
> +            /* disable nested svm if not explicitly requested */
>             if (!kvm_nested)
>                 *ecx &= ~CPUID_EXT3_SVM;
> -            /* 3dnow */
> -            *edx &= ~0xc0000000;

I don't know why this is in, but it's definitely in a kvm_enabled() if  
statement, so it's not because of TCG :-).
Also kvm_nested should control the svm bit in the qemu description  
already. Preferably before the +/-feature bits are interpreted.

... this would lead to total removal of that hack function :-).

Alex