On Wed, Aug 19, 2026 at 05:32:25PM +0100, Leonardo Bras wrote: > On Wed, Aug 12, 2026 at 08:12:02PM +0100, Mark Brown wrote:
> > In order to allow guests to use GCS we also need to configure
> > HCRX_EL2.GCSEn, if this is not set GCS instructions will be noops and
> > CHKFEAT will report GCS as disabled.
> It is zero on reset, and keeping it in zero disables GCS in EL0&EL1, so
> unless we are in EL2&0 (HCR_EL2.{E2H, TGE} is {1, 1}), we need to enable it
> so EL1&0 (guests) can have access to it.
Right.
> > @@ -77,6 +80,8 @@ static void __sysreg_save_vel2_state(struct kvm_vcpu
> > *vcpu)
> > __vcpu_assign_sys_reg(vcpu, SP_EL2, read_sysreg(sp_el1));
> > __vcpu_assign_sys_reg(vcpu, ELR_EL2, read_sysreg_el1(SYS_ELR));
> > __vcpu_assign_sys_reg(vcpu, SPSR_EL2, read_sysreg_el1(SYS_SPSR));
> > + if (ctxt_has_gcs(&vcpu->arch.ctxt))
> > + __vcpu_assign_sys_reg(vcpu, GCSPR_EL2,
> > read_sysreg_el1(SYS_GCSPR));
> Out of curiosity here,
> Why does the saving of GCSCR_EL2 depends on ctxt_has_tcrx() (and E2H set),
> and saving SYS_GCSPR does not? ...
This is just because there isn't a preexisting ctxt_has_tcrx() check
already there, FEAT_GCS architecturally depends on FEAT_TCRX and there
was a request to make this explicit in the code to try to optimise
things a bit. The compiler should skip over both blocks at once if TCRX
isn't there rather than having two separate tests or static branches. I
didn't add new checks where there were none since I expect that to be
unhelpful for code generation, you'd get the reverse situation and emit
two checks.
signature.asc
Description: PGP signature
