[PATCH][RESEND] Print a user-friendly message on failed vmentry

[PATCH][RESEND] Print a user-friendly message on failed vmentry

[Mohammed Gamal]

This patch address bug report in https://bugs.launchpad.net/qemu/+bug/530077.

Failed vmentries were handled with handle_unhandled() which prints a rather
unfriendly message to the user. This patch separates handling vmentry failures
from unknown exit reasons and prints a friendly message to the user.

Signed-off-by: Mohammed Gamal <m.gamal005@gmail.com>
---
 qemu-kvm.c |   27 ++++++++++++++++++++++++++-
 1 files changed, 26 insertions(+), 1 deletions(-)

diff --git a/qemu-kvm.c b/qemu-kvm.c
index 9534b31..b8a9278 100644
--- a/qemu-kvm.c
+++ b/qemu-kvm.c
@@ -106,6 +106,31 @@ static int handle_unhandled(uint64_t reason)
     return -EINVAL;
 }
 
+#define VMX_INVALID_GUEST_STATE 0x80000021
+
+static int handle_failed_vmentry(uint64_t reason)
+{
+    fprintf(stderr, "kvm: vm entry failed with error 0x%" PRIx64 "\n\n", reason);
+
+    /* Perhaps we will need to check if this machine is intel since exit reason 0x21 
+       has a different interpretation on SVM */
+    if (reason == VMX_INVALID_GUEST_STATE) {
+        fprintf(stderr, "If you're runnning a guest on an Intel machine without\n");
+        fprintf(stderr, "unrestricted mode support, the failure can be most likely\n");
+        fprintf(stderr, "due to the guest entering an invalid state for Intel VT.\n");
+        fprintf(stderr, "For example, the guest maybe running in big real mode\n");
+        fprintf(stderr, "which is not supported on less recent Intel processors.\n\n");
+        fprintf(stderr, "You may want to try enabling KVM real mode emulation. To\n");
+        fprintf(stderr, "enable it, you can run the following commands as root:\n");
+        fprintf(stderr, "# rmmod kvm_intel\n");
+        fprintf(stderr, "# rmmod kvm\n");
+        fprintf(stderr, "# modprobe kvm_intel emulate_invalid_guest_state=1\n\n");
+        fprintf(stderr, "WARNING: Real mode emulation is still work-in-progress\n");
+        fprintf(stderr, "and thus it is not always guaranteed to work.\n\n");
+    }
+
+    return -EINVAL;
+}
 
 static inline void set_gsi(kvm_context_t kvm, unsigned int gsi)
 {
@@ -586,7 +611,7 @@ int kvm_run(CPUState *env)
             r = handle_unhandled(run->hw.hardware_exit_reason);
             break;
         case KVM_EXIT_FAIL_ENTRY:
-            r = handle_unhandled(run->fail_entry.hardware_entry_failure_reason);
+            r = handle_failed_vmentry(run->fail_entry.hardware_entry_failure_reason);
             break;
         case KVM_EXIT_EXCEPTION:
             fprintf(stderr, "exception %d (%x)\n", run->ex.exception,
-- 
1.7.0.4

Re: [PATCH][RESEND] Print a user-friendly message on failed vmentry

[Avi Kivity]

On 05/31/2010 10:40 PM, Mohammed Gamal wrote:
> This patch address bug report in https://bugs.launchpad.net/qemu/+bug/530077.
>
> Failed vmentries were handled with handle_unhandled() which prints a rather
> unfriendly message to the user. This patch separates handling vmentry failures
> from unknown exit reasons and prints a friendly message to the user.
>
>
>
> +#define VMX_INVALID_GUEST_STATE 0x80000021
> +
> +static int handle_failed_vmentry(uint64_t reason)
> +{
> +    fprintf(stderr, "kvm: vm entry failed with error 0x%" PRIx64 "\n\n", reason);
> +
> +    /* Perhaps we will need to check if this machine is intel since exit reason 0x21
> +       has a different interpretation on SVM */
> +    if (reason == VMX_INVALID_GUEST_STATE) {
> +        fprintf(stderr, "If you're runnning a guest on an Intel machine without\n");
> +        fprintf(stderr, "unrestricted mode support, the failure can be most likely\n");
> +        fprintf(stderr, "due to the guest entering an invalid state for Intel VT.\n");
> +        fprintf(stderr, "For example, the guest maybe running in big real mode\n");
> +        fprintf(stderr, "which is not supported on less recent Intel processors.\n\n");
> +        fprintf(stderr, "You may want to try enabling KVM real mode emulation. To\n");
> +        fprintf(stderr, "enable it, you can run the following commands as root:\n");
> +        fprintf(stderr, "# rmmod kvm_intel\n");
> +        fprintf(stderr, "# rmmod kvm\n");
> +        fprintf(stderr, "# modprobe kvm_intel emulate_invalid_guest_state=1\n\n");
> +        fprintf(stderr, "WARNING: Real mode emulation is still work-in-progress\n");
> +        fprintf(stderr, "and thus it is not always guaranteed to work.\n\n");
> +    }
> +
> +    return -EINVAL;
> +}
>
>    

It's almost guaranteed to fail, isn't it?  Is there any guest which 
fails with emulated_invalid_guest_state=0 but works with e_i_g_s=1?

-- 
error compiling committee.c: too many arguments to function

Re: [PATCH][RESEND] Print a user-friendly message on failed vmentry

[Mohammed Gamal]

On Tue, Jun 1, 2010 at 11:59 AM, Avi Kivity <avi@redhat.com> wrote:
> On 05/31/2010 10:40 PM, Mohammed Gamal wrote:
>>
>> This patch address bug report in
>> https://bugs.launchpad.net/qemu/+bug/530077.
>>
>> Failed vmentries were handled with handle_unhandled() which prints a
>> rather
>> unfriendly message to the user. This patch separates handling vmentry
>> failures
>> from unknown exit reasons and prints a friendly message to the user.
>>
>>
>>
>> +#define VMX_INVALID_GUEST_STATE 0x80000021
>> +
>> +static int handle_failed_vmentry(uint64_t reason)
>> +{
>> +    fprintf(stderr, "kvm: vm entry failed with error 0x%" PRIx64 "\n\n",
>> reason);
>> +
>> +    /* Perhaps we will need to check if this machine is intel since exit
>> reason 0x21
>> +       has a different interpretation on SVM */
>> +    if (reason == VMX_INVALID_GUEST_STATE) {
>> +        fprintf(stderr, "If you're runnning a guest on an Intel machine
>> without\n");
>> +        fprintf(stderr, "unrestricted mode support, the failure can be
>> most likely\n");
>> +        fprintf(stderr, "due to the guest entering an invalid state for
>> Intel VT.\n");
>> +        fprintf(stderr, "For example, the guest maybe running in big real
>> mode\n");
>> +        fprintf(stderr, "which is not supported on less recent Intel
>> processors.\n\n");
>> +        fprintf(stderr, "You may want to try enabling KVM real mode
>> emulation. To\n");
>> +        fprintf(stderr, "enable it, you can run the following commands as
>> root:\n");
>> +        fprintf(stderr, "# rmmod kvm_intel\n");
>> +        fprintf(stderr, "# rmmod kvm\n");
>> +        fprintf(stderr, "# modprobe kvm_intel
>> emulate_invalid_guest_state=1\n\n");
>> +        fprintf(stderr, "WARNING: Real mode emulation is still
>> work-in-progress\n");
>> +        fprintf(stderr, "and thus it is not always guaranteed to
>> work.\n\n");
>> +    }
>> +
>> +    return -EINVAL;
>> +}
>>
>>
>
> It's almost guaranteed to fail, isn't it?  Is there any guest which fails
> with emulated_invalid_guest_state=0 but works with e_i_g_s=1?
>
You're right! Perhaps I should remove the e_i_g_s bit from the
message. What do you think?

> --
> error compiling committee.c: too many arguments to function
>
>

Re: [PATCH][RESEND] Print a user-friendly message on failed vmentry

[Avi Kivity]

On 06/01/2010 01:55 PM, Mohammed Gamal wrote:
> On Tue, Jun 1, 2010 at 11:59 AM, Avi Kivity<avi@redhat.com>  wrote:
>    
>> On 05/31/2010 10:40 PM, Mohammed Gamal wrote:
>>      
>>> This patch address bug report in
>>> https://bugs.launchpad.net/qemu/+bug/530077.
>>>
>>> Failed vmentries were handled with handle_unhandled() which prints a
>>> rather
>>> unfriendly message to the user. This patch separates handling vmentry
>>> failures
>>> from unknown exit reasons and prints a friendly message to the user.
>>>
>>>
>>>
>>> +#define VMX_INVALID_GUEST_STATE 0x80000021
>>> +
>>> +static int handle_failed_vmentry(uint64_t reason)
>>> +{
>>> +    fprintf(stderr, "kvm: vm entry failed with error 0x%" PRIx64 "\n\n",
>>> reason);
>>> +
>>> +    /* Perhaps we will need to check if this machine is intel since exit
>>> reason 0x21
>>> +       has a different interpretation on SVM */
>>> +    if (reason == VMX_INVALID_GUEST_STATE) {
>>> +        fprintf(stderr, "If you're runnning a guest on an Intel machine
>>> without\n");
>>> +        fprintf(stderr, "unrestricted mode support, the failure can be
>>> most likely\n");
>>> +        fprintf(stderr, "due to the guest entering an invalid state for
>>> Intel VT.\n");
>>> +        fprintf(stderr, "For example, the guest maybe running in big real
>>> mode\n");
>>> +        fprintf(stderr, "which is not supported on less recent Intel
>>> processors.\n\n");
>>> +        fprintf(stderr, "You may want to try enabling KVM real mode
>>> emulation. To\n");
>>> +        fprintf(stderr, "enable it, you can run the following commands as
>>> root:\n");
>>> +        fprintf(stderr, "# rmmod kvm_intel\n");
>>> +        fprintf(stderr, "# rmmod kvm\n");
>>> +        fprintf(stderr, "# modprobe kvm_intel
>>> emulate_invalid_guest_state=1\n\n");
>>> +        fprintf(stderr, "WARNING: Real mode emulation is still
>>> work-in-progress\n");
>>> +        fprintf(stderr, "and thus it is not always guaranteed to
>>> work.\n\n");
>>> +    }
>>> +
>>> +    return -EINVAL;
>>> +}
>>>
>>>
>>>        
>> It's almost guaranteed to fail, isn't it?  Is there any guest which fails
>> with emulated_invalid_guest_state=0 but works with e_i_g_s=1?
>>
>>      
> You're right! Perhaps I should remove the e_i_g_s bit from the
> message. What do you think?
>    


Sure.  Better than the current message.

-- 
error compiling committee.c: too many arguments to function