From: Paolo Bonzini <pbonzini@redhat.com>
To: Lara Lazier <laramglazier@gmail.com>, qemu-devel@nongnu.org
Subject: Re: [PATCH v3] target/i386: Added VGIF feature
Date: Mon, 2 Aug 2021 10:25:57 +0200 [thread overview]
Message-ID: <3ff32373-41b7-86b9-d170-a801532242c1@redhat.com> (raw)
In-Reply-To: <20210730070742.9674-1-laramglazier@gmail.com>
On 30/07/21 09:07, Lara Lazier wrote:
> v2->v3: moved guest check into virtual_gif_enabled
>
> VGIF allows STGI and CLGI to execute in guest mode and control virtual
> interrupts in guest mode.
> When the VGIF feature is enabled then:
> * executing STGI in the guest sets bit 9 of the VMCB offset 60h.
> * executing CLGI in the guest clears bit 9 of the VMCB offset 60h.
>
> Signed-off-by: Lara Lazier <laramglazier@gmail.com>
This part looks good, I queued it for QEMU 6.2. However, I am wondering
if STGI/CLGI should also affect the emulation of VIRQ. I am not very
familiar with this part of VGIF because KVM does not need it, but it
makes sense that STGI/CLGI respectively unmask and mask VIRQ just like
STI/CLI do. Can you look up in the manual if there's anything like
this? We can discuss the implementation tomorrow.
Thanks,
Paolo
> ---
> target/i386/cpu.c | 3 ++-
> target/i386/svm.h | 6 ++++++
> target/i386/tcg/sysemu/svm_helper.c | 31 +++++++++++++++++++++++++++--
> 3 files changed, 37 insertions(+), 3 deletions(-)
>
> diff --git a/target/i386/cpu.c b/target/i386/cpu.c
> index edb97ebbbe..71d26cf1bd 100644
> --- a/target/i386/cpu.c
> +++ b/target/i386/cpu.c
> @@ -631,7 +631,8 @@ void x86_cpu_vendor_words2str(char *dst, uint32_t vendor1,
> #define TCG_EXT3_FEATURES (CPUID_EXT3_LAHF_LM | CPUID_EXT3_SVM | \
> CPUID_EXT3_CR8LEG | CPUID_EXT3_ABM | CPUID_EXT3_SSE4A)
> #define TCG_EXT4_FEATURES 0
> -#define TCG_SVM_FEATURES CPUID_SVM_NPT
> +#define TCG_SVM_FEATURES (CPUID_SVM_NPT | CPUID_SVM_VGIF | \
> + CPUID_SVM_SVME_ADDR_CHK)
> #define TCG_KVM_FEATURES 0
> #define TCG_7_0_EBX_FEATURES (CPUID_7_0_EBX_SMEP | CPUID_7_0_EBX_SMAP | \
> CPUID_7_0_EBX_BMI1 | CPUID_7_0_EBX_BMI2 | CPUID_7_0_EBX_ADX | \
> diff --git a/target/i386/svm.h b/target/i386/svm.h
> index adc058dc76..036597a2ff 100644
> --- a/target/i386/svm.h
> +++ b/target/i386/svm.h
> @@ -9,6 +9,12 @@
> #define V_IRQ_SHIFT 8
> #define V_IRQ_MASK (1 << V_IRQ_SHIFT)
>
> +#define V_GIF_ENABLED_SHIFT 25
> +#define V_GIF_ENABLED_MASK (1 << V_GIF_ENABLED_SHIFT)
> +
> +#define V_GIF_SHIFT 9
> +#define V_GIF_MASK (1 << V_GIF_SHIFT)
> +
> #define V_INTR_PRIO_SHIFT 16
> #define V_INTR_PRIO_MASK (0x0f << V_INTR_PRIO_SHIFT)
>
> diff --git a/target/i386/tcg/sysemu/svm_helper.c b/target/i386/tcg/sysemu/svm_helper.c
> index 90a9de30f8..b6ad425cca 100644
> --- a/target/i386/tcg/sysemu/svm_helper.c
> +++ b/target/i386/tcg/sysemu/svm_helper.c
> @@ -110,6 +110,17 @@ static inline bool is_efer_invalid_state (CPUX86State *env)
> return false;
> }
>
> +static inline bool virtual_gif_enabled(CPUX86State *env, uint32_t *int_ctl)
> +{
> + if (likely(env->hflags & HF_GUEST_MASK)) {
> + *int_ctl = x86_ldl_phys(env_cpu(env),
> + env->vm_vmcb + offsetof(struct vmcb, control.int_ctl));
> + return (env->features[FEAT_SVM] & CPUID_SVM_VGIF)
> + && (*int_ctl & V_GIF_ENABLED_MASK);
> + }
> + return false;
> +}
> +
> void helper_vmrun(CPUX86State *env, int aflag, int next_eip_addend)
> {
> CPUState *cs = env_cpu(env);
> @@ -503,13 +514,29 @@ void helper_vmsave(CPUX86State *env, int aflag)
> void helper_stgi(CPUX86State *env)
> {
> cpu_svm_check_intercept_param(env, SVM_EXIT_STGI, 0, GETPC());
> - env->hflags2 |= HF2_GIF_MASK;
> +
> + CPUState *cs = env_cpu(env);
> + uint32_t int_ctl;
> + if (virtual_gif_enabled(env, &int_ctl)) {
> + x86_stl_phys(cs, env->vm_vmcb + offsetof(struct vmcb, control.int_ctl),
> + int_ctl | V_GIF_MASK);
> + } else {
> + env->hflags2 |= HF2_GIF_MASK;
> + }
> }
>
> void helper_clgi(CPUX86State *env)
> {
> cpu_svm_check_intercept_param(env, SVM_EXIT_CLGI, 0, GETPC());
> - env->hflags2 &= ~HF2_GIF_MASK;
> +
> + CPUState *cs = env_cpu(env);
> + uint32_t int_ctl;
> + if (virtual_gif_enabled(env, &int_ctl)) {
> + x86_stl_phys(cs, env->vm_vmcb + offsetof(struct vmcb, control.int_ctl),
> + int_ctl & ~V_GIF_MASK);
> + } else {
> + env->hflags2 &= ~HF2_GIF_MASK;
> + }
> }
>
> bool cpu_svm_has_intercept(CPUX86State *env, uint32_t type)
>
prev parent reply other threads:[~2021-08-02 8:26 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-07-30 7:07 [PATCH v3] target/i386: Added VGIF feature Lara Lazier
2021-08-02 8:25 ` Paolo Bonzini [this message]
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=3ff32373-41b7-86b9-d170-a801532242c1@redhat.com \
--to=pbonzini@redhat.com \
--cc=laramglazier@gmail.com \
--cc=qemu-devel@nongnu.org \
/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).