On AMD, the state of the VMCB is undefined after a shutdown VMEXIT. KVM takes a very conservative approach to that and resets the guest altogether when that happens. This causes the set_memory_region_test to fail because the RIP is 0xfff0 (the reset vector). Restrict the RIP test to KVM_EXIT_INTERNAL_ERROR in order to fix this. Signed-off-by: Paolo Bonzini <pbonzini@xxxxxxxxxx> --- .../testing/selftests/kvm/set_memory_region_test.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) diff --git a/tools/testing/selftests/kvm/set_memory_region_test.c b/tools/testing/selftests/kvm/set_memory_region_test.c index 260e638826dc..b3ece55a2da6 100644 --- a/tools/testing/selftests/kvm/set_memory_region_test.c +++ b/tools/testing/selftests/kvm/set_memory_region_test.c @@ -287,10 +287,15 @@ static void test_delete_memory_region(void) vcpu_regs_get(vm, VCPU_ID, ®s); - TEST_ASSERT(regs.rip >= final_rip_start && - regs.rip < final_rip_end, - "Bad rip, expected 0x%lx - 0x%lx, got 0x%llx\n", - final_rip_start, final_rip_end, regs.rip); + /* + * On AMD, after KVM_EXIT_SHUTDOWN the VMCB has been reinitialized already, + * so the instruction pointer would point to the reset vector. + */ + if (run->exit_reason == KVM_EXIT_INTERNAL_ERROR) + TEST_ASSERT(regs.rip >= final_rip_start && + regs.rip < final_rip_end, + "Bad rip, expected 0x%lx - 0x%lx, got 0x%llx\n", + final_rip_start, final_rip_end, regs.rip); kvm_vm_free(vm); } -- 2.18.2
