From: Eduardo Habkost <ehabkost@redhat.com>
To: Wanpeng Li <kernellwp@gmail.com>
Cc: "Paolo Bonzini" <pbonzini@redhat.com>,
"qemu-devel@nongnu.org Developers" <qemu-devel@nongnu.org>,
kvm <kvm@vger.kernel.org>, "Radim Krčmář" <rkrcmar@redhat.com>
Subject: Re: [Qemu-devel] [PATCH] target-i386: adds PV_DEDICATED hint CPUID feature bit
Date: Fri, 9 Feb 2018 10:46:38 -0200 [thread overview]
Message-ID: <20180209124638.GH13981@localhost.localdomain> (raw)
In-Reply-To: <CANRm+CyJi0r-Gm=ivJDwBhsW1dM5tAhBBx84G+miOQDaF+nLUw@mail.gmail.com>
On Fri, Feb 09, 2018 at 08:24:26PM +0800, Wanpeng Li wrote:
> 2018-02-09 19:46 GMT+08:00 Eduardo Habkost <ehabkost@redhat.com>:
> > On Fri, Feb 09, 2018 at 09:41:41AM +0100, Paolo Bonzini wrote:
> >> On 08/02/2018 19:12, Eduardo Habkost wrote:
> >> > On Thu, Feb 08, 2018 at 05:44:20PM +0800, Wanpeng Li wrote:
> >> >> From: Wanpeng Li <wanpengli@tencent.com>
> >> >>
> >> >> Add PV_DEDICATED hint cpuid feature bit.
> >> >>
> >> > [...]
> >> >> diff --git a/target/i386/cpu.c b/target/i386/cpu.c
> >> >> index d70954b..cf48931 100644
> >> >> --- a/target/i386/cpu.c
> >> >> +++ b/target/i386/cpu.c
> >> >> @@ -4076,6 +4076,9 @@ static int x86_cpu_filter_features(X86CPU *cpu)
> >> >> x86_cpu_get_supported_feature_word(w, false);
> >> >> uint32_t requested_features = env->features[w];
> >> >> env->features[w] &= host_feat;
> >> >> + if (cpu->pv_dedicated && (w == FEAT_KVM)) {
> >> >> + env->features[w] |= CPUID_PV_DEDICATED;
> >> >> + }
> >> >
> >> > Why do this trick here instead of just adding "pv-dedicated" to
> >> > feature_word_info[FEAT_KVM].feat_names?
> >>
> >> I think because it can be added independent of KVM_GET_SUPPORTED_CPUID.
> >
> > If QEMU knows it can be enabled if outside GET_SUPPORTED_CPUID,
> > kvm_arch_get_supported_cpuid() can return it as supported.
>
> The guest fails to detect the new cpuid this time.
This should fix it:
diff --git a/target/i386/kvm.c b/target/i386/kvm.c
index a0aad808bd..44ee524188 100644
--- a/target/i386/kvm.c
+++ b/target/i386/kvm.c
@@ -804,6 +804,7 @@ int kvm_arch_init_vcpu(CPUState *cs)
c = &cpuid_data.entries[cpuid_i++];
c->function = KVM_CPUID_FEATURES | kvm_base;
c->eax = env->features[FEAT_KVM];
+ c->edx = env->features[FEAT_KVM_HINTS];
}
cpu_x86_cpuid(env, 0, 0, &limit, &unused, &unused, &unused);
>
> diff --git a/target/i386/cpu.c b/target/i386/cpu.c
> index d70954b..e2974ad 100644
> --- a/target/i386/cpu.c
> +++ b/target/i386/cpu.c
> @@ -358,6 +358,20 @@ static FeatureWordInfo feature_word_info[FEATURE_WORDS] = {
> .cpuid_eax = KVM_CPUID_FEATURES, .cpuid_reg = R_EAX,
> .tcg_features = TCG_KVM_FEATURES,
> },
> + [FEAT_KVM_HINTS] = {
> + .feat_names = {
> + "hint-dedicated", NULL, NULL, NULL,
> + NULL, NULL, NULL, NULL,
> + NULL, NULL, NULL, NULL,
> + NULL, NULL, NULL, NULL,
> + NULL, NULL, NULL, NULL,
> + NULL, NULL, NULL, NULL,
> + NULL, NULL, NULL, NULL,
> + NULL, NULL, NULL, NULL,
> + },
> + .cpuid_eax = KVM_CPUID_FEATURES, .cpuid_reg = R_EDX,
> + .tcg_features = TCG_KVM_FEATURES,
> + },
> [FEAT_HYPERV_EAX] = {
> .feat_names = {
> NULL /* hv_msr_vp_runtime_access */, NULL /*
> hv_msr_time_refcount_access */,
> diff --git a/target/i386/cpu.h b/target/i386/cpu.h
> index f91e37d..9f73692 100644
> --- a/target/i386/cpu.h
> +++ b/target/i386/cpu.h
> @@ -475,6 +475,7 @@ typedef enum FeatureWord {
> FEAT_8000_0008_EBX, /* CPUID[8000_0008].EBX */
> FEAT_C000_0001_EDX, /* CPUID[C000_0001].EDX */
> FEAT_KVM, /* CPUID[4000_0001].EAX (KVM_CPUID_FEATURES) */
> + FEAT_KVM_HINTS, /* CPUID[4000_0001].EDX */
> FEAT_HYPERV_EAX, /* CPUID[4000_0003].EAX */
> FEAT_HYPERV_EBX, /* CPUID[4000_0003].EBX */
> FEAT_HYPERV_EDX, /* CPUID[4000_0003].EDX */
> @@ -670,6 +671,8 @@ typedef uint32_t FeatureWordArray[FEATURE_WORDS];
> #define CPUID_7_0_EDX_AVX512_4FMAPS (1U << 3) /* AVX512 Multiply
> Accumulation Single Precision */
> #define CPUID_7_0_EDX_SPEC_CTRL (1U << 26) /* Speculation Control */
>
> +#define KVM_HINTS_DEDICATED (1U << 0)
> +
> #define CPUID_8000_0008_EBX_IBPB (1U << 12) /* Indirect Branch
> Prediction Barrier */
>
> #define CPUID_XSAVE_XSAVEOPT (1U << 0)
> diff --git a/target/i386/kvm.c b/target/i386/kvm.c
> index ad4b159..a0aad80 100644
> --- a/target/i386/kvm.c
> +++ b/target/i386/kvm.c
> @@ -383,6 +383,9 @@ uint32_t kvm_arch_get_supported_cpuid(KVMState *s,
> uint32_t function,
> if (!kvm_irqchip_in_kernel()) {
> ret &= ~(1U << KVM_FEATURE_PV_UNHALT);
> }
> + } else if (function == KVM_CPUID_FEATURES && reg == R_EDX) {
> + ret |= KVM_HINTS_DEDICATED;
> + found = 1;
> }
>
> /* fallback for older kernels */
>
>
> Regards,
> Wanpeng Li
--
Eduardo
next prev parent reply other threads:[~2018-02-09 12:46 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-02-08 9:44 [Qemu-devel] [PATCH] target-i386: adds PV_DEDICATED hint CPUID feature bit Wanpeng Li
2018-02-08 18:12 ` Eduardo Habkost
2018-02-09 3:18 ` Wanpeng Li
2018-02-09 8:41 ` Paolo Bonzini
2018-02-09 11:46 ` Eduardo Habkost
2018-02-09 12:24 ` Wanpeng Li
2018-02-09 12:46 ` Paolo Bonzini
2018-02-09 12:46 ` Eduardo Habkost [this message]
2018-02-09 14:26 ` Wanpeng Li
2018-02-09 8:41 ` Paolo Bonzini
2018-02-09 14:27 ` Wanpeng Li
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20180209124638.GH13981@localhost.localdomain \
--to=ehabkost@redhat.com \
--cc=kernellwp@gmail.com \
--cc=kvm@vger.kernel.org \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=rkrcmar@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).