Re: [PATCH v4 14/19] KVM: VMX: Dump FRED context in dump_vmcs()

[Xin Li]

On 6/24/2025 9:32 AM, Sean Christopherson wrote:
> > @@ -6519,6 +6521,16 @@ void dump_vmcs(struct kvm_vcpu *vcpu)
> >   	vmx_dump_sel("LDTR:", GUEST_LDTR_SELECTOR);
> >   	vmx_dump_dtsel("IDTR:", GUEST_IDTR_LIMIT);
> >   	vmx_dump_sel("TR:  ", GUEST_TR_SELECTOR);
> > +	if (vmentry_ctl & VM_ENTRY_LOAD_IA32_FRED)
> > +		pr_err("FRED guest: config=0x%016llx, stack_levels=0x%016llx\n"
> > +		       "RSP0=0x%016llx, RSP1=0x%016llx\n"
> > +		       "RSP2=0x%016llx, RSP3=0x%016llx\n",
> > +		       vmcs_read64(GUEST_IA32_FRED_CONFIG),
> > +		       vmcs_read64(GUEST_IA32_FRED_STKLVLS),
> > +		       __rdmsr(MSR_IA32_FRED_RSP0),
>
> There is no guarantee the vCPU's FRED_RSP is loaded in hardware at this point.
> I think you need to use vmx_read_guest_fred_rsp0().

Good catch.

> > +		       vmcs_read64(GUEST_IA32_FRED_RSP1),
> > +		       vmcs_read64(GUEST_IA32_FRED_RSP2),
> > +		       vmcs_read64(GUEST_IA32_FRED_RSP3));
> >   	efer_slot = vmx_find_loadstore_msr_slot(&vmx->msr_autoload.guest, MSR_EFER);
> >   	if (vmentry_ctl & VM_ENTRY_LOAD_IA32_EFER)
> >   		pr_err("EFER= 0x%016llx\n", vmcs_read64(GUEST_IA32_EFER));
> > @@ -6566,6 +6578,16 @@ void dump_vmcs(struct kvm_vcpu *vcpu)
> >   	       vmcs_readl(HOST_TR_BASE));
> >   	pr_err("GDTBase=%016lx IDTBase=%016lx\n",
> >   	       vmcs_readl(HOST_GDTR_BASE), vmcs_readl(HOST_IDTR_BASE));
> > +	if (vmexit_ctl & SECONDARY_VM_EXIT_LOAD_IA32_FRED)
> > +		pr_err("FRED host: config=0x%016llx, stack_levels=0x%016llx\n"
> > +		       "RSP0=0x%016lx, RSP1=0x%016llx\n"
> > +		       "RSP2=0x%016llx, RSP3=0x%016llx\n",
> > +		       vmcs_read64(HOST_IA32_FRED_CONFIG),
> > +		       vmcs_read64(HOST_IA32_FRED_STKLVLS),
> > +		       (unsigned long)task_stack_page(current) + THREAD_SIZE,
>
> Maybe add a helper in arch/x86/include/asm/fred.h to generate the desired RSP0?
> Not sure it's worth doing that just for this code.

It's not just one usage.  I checked with:

    git grep -w task_stack_page | grep THREAD_SIZE | wc -l

And get 25.

However it is used in other architectures, so I'll work it in parallel.
I.e., likely I won't change it in the next iteration.

Thanks!
    Xin