On 4/29/2025 6:06 PM, Zhao Liu wrote:
On Thu, Apr 24, 2025 at 01:51:55PM +0800, Xiaoyao Li wrote:
Date: Thu, 24 Apr 2025 13:51:55 +0800
From: Xiaoyao Li <xiaoyao...@intel.com>
Subject: Re: [PATCH v8 31/55] i386/cpu: introduce
x86_confidential_guest_cpu_instance_init()
Hi Paolo,
On 4/1/2025 9:01 PM, Xiaoyao Li wrote:
...
diff --git a/target/i386/cpu.c b/target/i386/cpu.c
index add6430f7edd..5c69d1489365 100644
--- a/target/i386/cpu.c
+++ b/target/i386/cpu.c
@@ -36,6 +36,7 @@
#include "hw/qdev-properties.h"
#include "hw/i386/topology.h"
#ifndef CONFIG_USER_ONLY
+#include "confidential-guest.h"
#include "system/reset.h"
#include "qapi/qapi-commands-machine-target.h"
#include "exec/address-spaces.h"
@@ -8504,6 +8505,15 @@ static void x86_cpu_post_initfn(Object *obj)
}
accel_cpu_instance_init(CPU(obj));
+
+#ifndef CONFIG_USER_ONLY
+ MachineState *ms = MACHINE(object_dynamic_cast(qdev_get_machine(),
+ TYPE_MACHINE));
It leads to
qemu-system-x86_64: ../hw/core/qdev.c:824: qdev_get_machine: Assertion
`dev' failed.
Aborted (core dumped)
for the case of "-cpu help" due to the assert(dev) in qdev_get_machine().
How do you want to resolve it? I can think of two:
1. remove the assert() in qdev_get_machine(). or
2. drop the callback introduce by this patch. Instead just do
if (is_tdx_vm()) {
tdx_cpu_instance_init();
}
Sorry I missed this mail when review this patch.
What about checking `current_machine`?
@@ -8541,10 +8541,8 @@ static void x86_cpu_post_initfn(Object *obj)
accel_cpu_instance_init(CPU(obj));
#ifndef CONFIG_USER_ONLY
- MachineState *ms = MACHINE(object_dynamic_cast(qdev_get_machine(),
- TYPE_MACHINE));
- if (ms && ms->cgs) {
-
x86_confidential_guest_cpu_instance_init(X86_CONFIDENTIAL_GUEST(ms->cgs),
+ if (current_machine && current_machine->cgs) {
+
x86_confidential_guest_cpu_instance_init(X86_CONFIDENTIAL_GUEST(current_machine->cgs),
(CPU(obj)));
}
#endif
---
"-cpu help" is processed before machine creation. The cpu-core
(cpu_core_instance_init) also checks current_machine to avoid similar
issue.
This is a really good suggestion! I'll take it.
Thanks!
Regards,
Zhao
