From: Sean Christopherson <sean.j.christopherson@intel.com>
To: Paolo Bonzini <pbonzini@redhat.com>
Cc: linux-kernel@vger.kernel.org, kvm@vger.kernel.org
Subject: Re: [PATCH] KVM: VMX: simplify vmx_prepare_switch_to_{guest,host}
Date: Fri, 7 Jun 2019 10:37:10 -0700 [thread overview]
Message-ID: <20190607173710.GG9083@linux.intel.com> (raw)
In-Reply-To: <1559927301-8124-1-git-send-email-pbonzini@redhat.com>
On Fri, Jun 07, 2019 at 07:08:21PM +0200, Paolo Bonzini wrote:
> vmx->loaded_cpu_state can only be NULL or equal to vmx->loaded_vmcs,
> so change it to a bool. Because the direction of the bool is
> now the opposite of vmx->guest_msrs_dirty, change the direction of
> vmx->guest_msrs_dirty so that they match.
>
> Finally, do not imply that MSRs have to be reloaded when
> vmx->guest_sregs_loaded is false; instead, set vmx->guest_msrs_loaded
> to false explicitly in vmx_prepare_switch_to_host.
>
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
...
> @@ -1165,13 +1163,15 @@ static void vmx_prepare_switch_to_host(struct vcpu_vmx *vmx)
> wrmsrl(MSR_KERNEL_GS_BASE, vmx->msr_host_kernel_gs_base);
> #endif
> load_fixmap_gdt(raw_smp_processor_id());
> + vmx->guest_sregs_loaded = false;
> + vmx->guest_msrs_loaded = false;
> }
>
> #ifdef CONFIG_X86_64
> static u64 vmx_read_guest_kernel_gs_base(struct vcpu_vmx *vmx)
> {
> preempt_disable();
> - if (vmx->loaded_cpu_state)
> + if (vmx->guest_sregs_loaded)
> rdmsrl(MSR_KERNEL_GS_BASE, vmx->msr_guest_kernel_gs_base);
This is the hiccup with naming it sregs_loaded. The split bools is also
kinda wonky since the 32->64 case is a one-off scenario. I think a
cleaner solution would be to remove guest_msrs_dirty and refresh the MSRs
directly from setup_msrs(). Then loaded_cpu_state -> loaded_guest_state
can be a straight conversion from loaded_vmcs -> bool. I'll send patches.
> preempt_enable();
> return vmx->msr_guest_kernel_gs_base;
> @@ -1180,7 +1180,7 @@ static u64 vmx_read_guest_kernel_gs_base(struct vcpu_vmx *vmx)
> static void vmx_write_guest_kernel_gs_base(struct vcpu_vmx *vmx, u64 data)
> {
> preempt_disable();
> - if (vmx->loaded_cpu_state)
> + if (vmx->guest_sregs_loaded)
Same issue here, one would expect this to check guest_msrs_loaded.
> wrmsrl(MSR_KERNEL_GS_BASE, data);
> preempt_enable();
> vmx->msr_guest_kernel_gs_base = data;
> @@ -1583,7 +1583,7 @@ static void setup_msrs(struct vcpu_vmx *vmx)
> move_msr_up(vmx, index, save_nmsrs++);
>
> vmx->save_nmsrs = save_nmsrs;
> - vmx->guest_msrs_dirty = true;
> + vmx->guest_msrs_loaded = false;
>
> if (cpu_has_vmx_msr_bitmap())
> vmx_update_msr_bitmap(&vmx->vcpu);
> diff --git a/arch/x86/kvm/vmx/vmx.h b/arch/x86/kvm/vmx/vmx.h
> index ed65999b07a8..fc369473f9df 100644
> --- a/arch/x86/kvm/vmx/vmx.h
> +++ b/arch/x86/kvm/vmx/vmx.h
> @@ -187,13 +187,23 @@ struct vcpu_vmx {
> struct kvm_vcpu vcpu;
> u8 fail;
> u8 msr_bitmap_mode;
> +
> + /*
> + * If true, host state has been stored in vmx->loaded_vmcs for
> + * the CPU registers that only need to be switched when transitioning
> + * to/from the kernel, and the registers have been loaded with guest
> + * values. If false, host state is loaded in the CPU registers
> + * and vmx->loaded_vmcs->host_state is invalid.
> + */
> + bool guest_sregs_loaded;
> +
> u32 exit_intr_info;
> u32 idt_vectoring_info;
> ulong rflags;
> struct shared_msr_entry *guest_msrs;
> int nmsrs;
> int save_nmsrs;
> - bool guest_msrs_dirty;
> + bool guest_msrs_loaded;
> #ifdef CONFIG_X86_64
> u64 msr_host_kernel_gs_base;
> u64 msr_guest_kernel_gs_base;
> @@ -208,14 +218,10 @@ struct vcpu_vmx {
> /*
> * loaded_vmcs points to the VMCS currently used in this vcpu. For a
> * non-nested (L1) guest, it always points to vmcs01. For a nested
> - * guest (L2), it points to a different VMCS. loaded_cpu_state points
> - * to the VMCS whose state is loaded into the CPU registers that only
> - * need to be switched when transitioning to/from the kernel; a NULL
> - * value indicates that host state is loaded.
> + * guest (L2), it points to a different VMCS.
> */
> struct loaded_vmcs vmcs01;
> struct loaded_vmcs *loaded_vmcs;
> - struct loaded_vmcs *loaded_cpu_state;
>
> struct msr_autoload {
> struct vmx_msrs guest;
> --
> 1.8.3.1
>
next prev parent reply other threads:[~2019-06-07 17:37 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-06-07 17:08 [PATCH] KVM: VMX: simplify vmx_prepare_switch_to_{guest,host} Paolo Bonzini
2019-06-07 17:37 ` Sean Christopherson [this message]
2019-06-07 17:47 ` Sean Christopherson
2019-06-07 18:13 ` Paolo Bonzini
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=20190607173710.GG9083@linux.intel.com \
--to=sean.j.christopherson@intel.com \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=pbonzini@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