KVM: arm64: VHE: Centralize ISBs when returning to host

The VHE hyp code has recently gained a few ISBs. Simplify this to one
unconditional ISB in __kvm_vcpu_run_vhe(), and remove the unnecessary
ISB from the kvm_call_hyp_ret() macro.

While kvm_call_hyp_ret() is also used to invoke
__vgic_v3_get_gic_config(), but no ISB is necessary in that case either.

For the moment, an ISB is left in kvm_call_hyp(), as there are many more
users, and removing the ISB would require a more thorough audit.

Suggested-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Fuad Tabba <tabba@google.com>
Cc: Marc Zyngier <maz@kernel.org>
Cc: Mark Brown <broonie@kernel.org>
Cc: Oliver Upton <oliver.upton@linux.dev>
Cc: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20250617133718.4014181-8-mark.rutland@arm.com
Signed-off-by: Marc Zyngier <maz@kernel.org>
This commit is contained in:
Mark Rutland 2025-06-17 14:37:18 +01:00 committed by Marc Zyngier
parent 3a300a33e4
commit 04c5355b2a
3 changed files with 14 additions and 20 deletions

View file

@ -1289,9 +1289,8 @@ void kvm_arm_resume_guest(struct kvm *kvm);
})
/*
* The couple of isb() below are there to guarantee the same behaviour
* on VHE as on !VHE, where the eret to EL1 acts as a context
* synchronization event.
* The isb() below is there to guarantee the same behaviour on VHE as on !VHE,
* where the eret to EL1 acts as a context synchronization event.
*/
#define kvm_call_hyp(f, ...) \
do { \
@ -1309,7 +1308,6 @@ void kvm_arm_resume_guest(struct kvm *kvm);
\
if (has_vhe()) { \
ret = f(__VA_ARGS__); \
isb(); \
} else { \
ret = kvm_call_hyp_nvhe(f, ##__VA_ARGS__); \
} \

View file

@ -167,9 +167,6 @@ static inline void __debug_switch_to_host_common(struct kvm_vcpu *vcpu)
__debug_save_state(guest_dbg, guest_ctxt);
__debug_restore_state(host_dbg, host_ctxt);
if (has_vhe())
isb();
}
#endif /* __ARM64_KVM_HYP_DEBUG_SR_H__ */

View file

@ -558,10 +558,10 @@ static int __kvm_vcpu_run_vhe(struct kvm_vcpu *vcpu)
host_ctxt = host_data_ptr(host_ctxt);
guest_ctxt = &vcpu->arch.ctxt;
sysreg_save_host_state_vhe(host_ctxt);
fpsimd_lazy_switch_to_guest(vcpu);
sysreg_save_host_state_vhe(host_ctxt);
/*
* Note that ARM erratum 1165522 requires us to configure both stage 1
* and stage 2 translation for the guest context before we clear
@ -586,18 +586,23 @@ static int __kvm_vcpu_run_vhe(struct kvm_vcpu *vcpu)
__deactivate_traps(vcpu);
/* Ensure CPTR trap deactivation has taken effect */
sysreg_restore_host_state_vhe(host_ctxt);
__debug_switch_to_host(vcpu);
/*
* Ensure that all system register writes above have taken effect
* before returning to the host. In VHE mode, CPTR traps for
* FPSIMD/SVE/SME also apply to EL2, so FPSIMD/SVE/SME state must be
* manipulated after the ISB.
*/
isb();
fpsimd_lazy_switch_to_host(vcpu);
sysreg_restore_host_state_vhe(host_ctxt);
if (guest_owns_fp_regs())
__fpsimd_save_fpexc32(vcpu);
__debug_switch_to_host(vcpu);
return exit_code;
}
NOKPROBE_SYMBOL(__kvm_vcpu_run_vhe);
@ -627,12 +632,6 @@ int __kvm_vcpu_run(struct kvm_vcpu *vcpu)
*/
local_daif_restore(DAIF_PROCCTX_NOIRQ);
/*
* When we exit from the guest we change a number of CPU configuration
* parameters, such as traps. We rely on the isb() in kvm_call_hyp*()
* to make sure these changes take effect before running the host or
* additional guests.
*/
return ret;
}