Excerpts from Fabiano Rosas's message of January 22, 2022 8:26 am:
> MMIO emulation can fail if the guest uses an instruction that we are
> not prepared to emulate. Since these instructions can be and most
> likely are valid ones, this is (slightly) closer to an access fault
> than to an illegal instruction, so deliver a Data Storage interrupt
> instead of a Program interrupt.
>
> Suggested-by: Nicholas Piggin <npig...@gmail.com>
> Signed-off-by: Fabiano Rosas <faro...@linux.ibm.com>
> ---
> arch/powerpc/kvm/emulate_loadstore.c | 10 +++-------
> arch/powerpc/kvm/powerpc.c | 12 ++++++++++++
> 2 files changed, 15 insertions(+), 7 deletions(-)
>
> diff --git a/arch/powerpc/kvm/emulate_loadstore.c
> b/arch/powerpc/kvm/emulate_loadstore.c
> index 48272a9b9c30..cfc9114b87d0 100644
> --- a/arch/powerpc/kvm/emulate_loadstore.c
> +++ b/arch/powerpc/kvm/emulate_loadstore.c
> @@ -73,7 +73,6 @@ int kvmppc_emulate_loadstore(struct kvm_vcpu *vcpu)
> {
> u32 inst;
> enum emulation_result emulated = EMULATE_FAIL;
> - int advance = 1;
> struct instruction_op op;
>
> /* this default type might be overwritten by subcategories */
> @@ -98,6 +97,8 @@ int kvmppc_emulate_loadstore(struct kvm_vcpu *vcpu)
> int type = op.type & INSTR_TYPE_MASK;
> int size = GETSIZE(op.type);
>
> + vcpu->mmio_is_write = OP_IS_STORE(type);
> +
> switch (type) {
> case LOAD: {
> int instr_byte_swap = op.type & BYTEREV;
> @@ -355,15 +356,10 @@ int kvmppc_emulate_loadstore(struct kvm_vcpu *vcpu)
> }
> }
>
> - if (emulated == EMULATE_FAIL) {
> - advance = 0;
> - kvmppc_core_queue_program(vcpu, 0);
> - }
> -
> trace_kvm_ppc_instr(inst, kvmppc_get_pc(vcpu), emulated);
>
> /* Advance past emulated instruction. */
> - if (advance)
> + if (emulated != EMULATE_FAIL)
> kvmppc_set_pc(vcpu, kvmppc_get_pc(vcpu) + 4);
>
> return emulated;
> diff --git a/arch/powerpc/kvm/powerpc.c b/arch/powerpc/kvm/powerpc.c
> index 214602c58f13..9befb121dddb 100644
> --- a/arch/powerpc/kvm/powerpc.c
> +++ b/arch/powerpc/kvm/powerpc.c
> @@ -305,10 +305,22 @@ int kvmppc_emulate_mmio(struct kvm_vcpu *vcpu)
> case EMULATE_FAIL:
> {
> u32 last_inst;
> + ulong store_bit = DSISR_ISSTORE;
> + ulong cause = DSISR_BADACCESS;
>
> +#ifdef CONFIG_BOOKE
> + store_bit = ESR_ST;
> + cause = 0;
> +#endif
BookE can not cause a bad page fault in the guest with ESR bits AFAIKS,
so it would cause an infinite fault loop here. Maybe stick with the
program interrupt for BookE with a comment about that here.
And if it could use if (IS_ENABLED()) would be good?
Otherwise looks good, it should do the right thing on BookS.
Thanks,
Nick
> kvmppc_get_last_inst(vcpu, INST_GENERIC, &last_inst);
> pr_info_ratelimited("KVM: guest access to device memory using
> unsupported instruction (PID: %d opcode: %#08x)\n",
> current->pid, last_inst);
> +
> + if (vcpu->mmio_is_write)
> + cause |= store_bit;
> +
> + kvmppc_core_queue_data_storage(vcpu, vcpu->arch.vaddr_accessed,
> + cause);
> r = RESUME_GUEST;
> break;
> }
> --
> 2.34.1
>
>
