public inbox for kvm@vger.kernel.org
 help / color / mirror / Atom feed
From: Paolo Bonzini <pbonzini@redhat.com>
To: "Chang S. Bae" <chang.seok.bae@intel.com>, seanjc@google.com
Cc: kvm@vger.kernel.org, linux-kernel@vger.kernel.org, chao.gao@intel.com
Subject: Re: [PATCH v3 06/20] KVM: x86: Support APX state for XSAVE ABI
Date: Tue, 28 Apr 2026 11:31:47 +0200	[thread overview]
Message-ID: <d998f4f8-0039-43a8-a7a1-4fabeb120d6d@redhat.com> (raw)
In-Reply-To: <20260428050111.39323-7-chang.seok.bae@intel.com>

On 4/28/26 07:00, Chang S. Bae wrote:
> +	/*
> +	 * APX off in XSTATE_BV will guide fpu_copy_uabi_to_guest_fpstate()
> +	 * to avoid from unnecessary handling.
> +	 */
> +	xstate->xsave.header.xfeatures &= ~XFEATURE_MASK_APX;

I would place this statement inside in fpu_copy_uabi_to_guest_fpstate,
and more in general make arch/x86/fpu ignore APX when dealing with
guest_fpstate.  This can be a separate patch like this:

-------------- 8< -------------
From: Paolo Bonzini <pbonzini@redhat.com>
Subject: x86/fpu: ignore APX when copying from/to guest FPU

KVM will store APX extended GPRs directly in the regs[] field of
struct kvm_vcpu.  This is done to make accessors more uniform
between GPRs and EGPRs, and partly because x86 maintainers
expressed the desire to compile Linux with APX enabled sooner
or later; accessing guest EGPRs from KVM's C code would get
in the way.

Therefore, let KVM handle the APX feature on its own when
executing ioctls to get and set the virtual machine's
XSAVE state.

Cc: x86@kernel.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

diff --git a/arch/x86/kernel/fpu/core.c b/arch/x86/kernel/fpu/core.c
index 608983806fd7..c35cff8dbfff 100644
--- a/arch/x86/kernel/fpu/core.c
+++ b/arch/x86/kernel/fpu/core.c
@@ -420,6 +420,8 @@ void fpu_copy_guest_fpstate_to_uabi(struct fpu_guest *gfpu, void *buf,
  	struct membuf mb = { .p = buf, .left = size };
  
  	if (cpu_feature_enabled(X86_FEATURE_XSAVE)) {
+		/* Up to date APX registers are in struct kvm_vcpu anyway.  */
+		xfeatures &= ~XFEATURE_MASK_APX;
  		__copy_xstate_to_uabi_buf(mb, kstate, xfeatures, pkru,
  					  XSTATE_COPY_XSAVE);
  	} else {
@@ -464,6 +465,8 @@ int fpu_copy_uabi_to_guest_fpstate(struct fpu_guest *gfpu, const void *buf,
  	if (!(ustate->xsave.header.xfeatures & XFEATURE_MASK_PKRU))
  		vpkru = NULL;
  
+	/* APX registers are copied to and from struct kvm_vcpu, not the FPU.  */
+	ustate->xsave.header.xfeatures &= ~XFEATURE_MASK_APX;
  	return copy_uabi_from_kernel_to_xstate(kstate, ustate, vpkru);
  }
  EXPORT_SYMBOL_FOR_KVM(fpu_copy_uabi_to_guest_fpstate);


Thanks,

Paolo


  reply	other threads:[~2026-04-28  9:31 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-04-28  5:00 [PATCH v3 00/20] KVM: x86: Enable APX for guests Chang S. Bae
2026-04-28  5:00 ` [PATCH v3 01/20] KVM: VMX: Macrofy 64-bit GPR swapping in __vmx_vcpu_run() Chang S. Bae
2026-04-28  9:03   ` Paolo Bonzini
2026-04-28 20:12     ` Chang S. Bae
2026-04-29  7:50       ` Paolo Bonzini
2026-04-29 17:22         ` Chang S. Bae
2026-04-28  9:09   ` Paolo Bonzini
2026-04-28  5:00 ` [PATCH v3 02/20] KVM: SVM: Macrofy 64-bit GPR swapping in __svm_vcpu_run() Chang S. Bae
2026-04-28  5:00 ` [PATCH v3 03/20] KVM: SEV: Macrofy 64-bit GPR swapping in __svm_sev_es_vcpu_run() Chang S. Bae
2026-04-28  5:00 ` [PATCH v3 04/20] KVM: x86: Extend VCPU registers for EGPRs Chang S. Bae
2026-04-28  5:00 ` [PATCH v3 05/20] KVM: VMX: Save guest EGPRs in VCPU cache Chang S. Bae
2026-04-28  5:00 ` [PATCH v3 06/20] KVM: x86: Support APX state for XSAVE ABI Chang S. Bae
2026-04-28  9:31   ` Paolo Bonzini [this message]
2026-04-28  5:00 ` [PATCH v3 07/20] KVM: VMX: Refactor VMX instruction information access Chang S. Bae
2026-04-28  5:00 ` [PATCH v3 08/20] KVM: VMX: Refactor instruction information decoding Chang S. Bae
2026-04-28  5:01 ` [PATCH v3 09/20] KVM: VMX: Refactor register index retrieval from exit qualification Chang S. Bae
2026-04-28  5:01 ` [PATCH v3 10/20] KVM: VMX: Support instruction information extension Chang S. Bae
2026-04-28  5:01 ` [PATCH v3 11/20] KVM: nVMX: Propagate the extended instruction info field Chang S. Bae
2026-04-28  5:01 ` [PATCH v3 12/20] KVM: x86: Support EGPR accessing and tracking for emulator Chang S. Bae
2026-04-28  5:01 ` [PATCH v3 13/20] KVM: x86: Handle EGPR index and REX2-incompatible opcodes Chang S. Bae
2026-04-28  5:01 ` [PATCH v3 14/20] KVM: x86: Support REX2-prefixed opcode decode Chang S. Bae
2026-04-28  5:01 ` [PATCH v3 15/20] KVM: x86: Reject EVEX-prefixed instructions Chang S. Bae
2026-04-28  5:01 ` [PATCH v3 16/20] KVM: x86: Guard valid XCR0.APX settings Chang S. Bae
2026-04-28  5:01 ` [PATCH v3 17/20] KVM: x86: Expose APX foundation feature to guests Chang S. Bae
2026-04-28  5:01 ` [PATCH v3 18/20] KVM: x86: Expose APX sub-features " Chang S. Bae
2026-04-28  5:01 ` [PATCH v3 19/20] KVM: x86: selftests: Add APX state and ABI test Chang S. Bae
2026-04-28  5:01 ` [PATCH v3 20/20] KVM: x86: selftests: Add APX state handling and XCR0 sanity checks Chang S. Bae

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=d998f4f8-0039-43a8-a7a1-4fabeb120d6d@redhat.com \
    --to=pbonzini@redhat.com \
    --cc=chang.seok.bae@intel.com \
    --cc=chao.gao@intel.com \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=seanjc@google.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