----- chao.gao@xxxxxxxxx wrote:
> This patch leverages existing "interrupt" tests. What the patch does
> is L1
> puts L2 in halt state and injects an event to L2. Then check a flag
> to
> identify whether event injection succeeds or not.
>
> Cc: Liran Alon <liran.alon@xxxxxxxxxx>
> Signed-off-by: Chao Gao <chao.gao@xxxxxxxxx>
> ---
> x86/vmx_tests.c | 12 ++++++++++++
> 1 file changed, 12 insertions(+)
>
> diff --git a/x86/vmx_tests.c b/x86/vmx_tests.c
> index 0c3b5a7..3c46d3a 100644
> --- a/x86/vmx_tests.c
> +++ b/x86/vmx_tests.c
> @@ -1584,6 +1584,12 @@ static void interrupt_main(void)
> for (loops = 0; loops < 10000000 && !timer_fired; loops++)
> asm volatile ("nop");
> report("running a guest with interrupt acknowledgement set",
> timer_fired);
> +
> + apic_write(APIC_TMICT, 0);
> + irq_enable();
> + timer_fired = false;
> + vmcall();
> + report("Inject an event to a halted guest", timer_fired);
> }
>
> static int interrupt_exit_handler(void)
> @@ -1615,6 +1621,12 @@ static int interrupt_exit_handler(void)
> case 6:
> vmcs_write(GUEST_ACTV_STATE, ACTV_HLT);
> break;
> +
> + case 8:
> + vmcs_write(GUEST_ACTV_STATE, ACTV_HLT);
> + vmcs_write(ENT_INTR_INFO,
> + TIMER_VECTOR | INTR_INFO_VALID_MASK);
I would specify interrupt-type explicitly rather than implicitly.
Meaning passing here:
(TIMER_VECTOR |
(VMX_INTR_TYPE_EXT_INTR << INTR_INFO_INTR_TYPE_SHIFT) |
INTR_INFO_VALID_MASK)
> + break;
> }
> vmx_inc_test_stage();
> vmcs_write(GUEST_RIP, guest_rip + insn_len);
> --
> 1.9.1
Besides small comment above, looks good to me.
Reviewed-by: Liran Alon <liran.alon@xxxxxxxxxx>
