Yang, Sheng wrote:
> I modified the patch as you suggest.
> For any failure in this due to hardware error, I replace all the error
> return code with -EIO.
>
> Thanks.
> ---
> From: Sheng Yang <[EMAIL PROTECTED]>
> Date: Tue, 31 Jul 2007 17:23:20 +0800
> Subject: [PATCH] Add cpu consistence check i
I modified the patch as you suggest.
For any failure in this due to hardware error, I replace all the error
return code with -EIO.
Thanks.
---
From: Sheng Yang <[EMAIL PROTECTED]>
Date: Tue, 31 Jul 2007 17:23:20 +0800
Subject: [PATCH] Add cpu consistence check in vmx.
All the physical CPUs on the
Yang, Sheng wrote:
>>> + return -1;
>>>
>>>
>> -1 is -EPERM. We need a real, more suitable, error code here.
>>
>
> How about using -EINVAL? But the comment of it is "Invalid argument"...
> I can't find more one here...
>
>
EINVAL usually means a user-supplied argument is
>> +return -1;
>>
>
>-1 is -EPERM. We need a real, more suitable, error code here.
How about using -EINVAL? But the comment of it is "Invalid argument"...
I can't find more one here...
>Also, having a single function either construct vmcs_config or verify,
>depending on whether it is
>-Original Message-
>From: Avi Kivity [mailto:[EMAIL PROTECTED]
>Sent: Tuesday, July 31, 2007 4:10 PM
>To: Li, Xin B
>Cc: Yang, Sheng; kvm-devel@lists.sourceforge.net
>Subject: Re: [kvm-devel] [PATCH] Add cpu consistence check in vmx.
>
>Li, Xin B wrote:
>>
Li, Xin B wrote:
>> btw, what about cpu hotplug? we need to deal with that too. Do we
>> error out and refuse to enable the cpu if it isn't compatible enough?
>>
>
> I think we shouldn't enable it, or the code becomes to complex.
> -xin
>
>
What do you mean exactly?
Don't enable cp
>
>btw, what about cpu hotplug? we need to deal with that too. Do we
>error out and refuse to enable the cpu if it isn't compatible enough?
I think we shouldn't enable it, or the code becomes to complex.
-xin
>
>--
>error compiling committee.c: too many arguments to function
>
>
>---
Yang, Sheng wrote:
> Thank you for point out my fault.
>
> Here is a modified version which is clearer. And I have tested it with
> version d9feefe(for the latest git repository broken).
>
I recommend building kvm.git, not the external module. kvm.git is not
broken at the moment.
> All the p
Thank you for point out my fault.
Here is a modified version which is clearer. And I have tested it with
version d9feefe(for the latest git repository broken).
Thanks
Yang, Sheng
--
From: Sheng Yang <[EMAIL PROTECTED]>
Date: Tue, 31 Jul 2007 10:21:32 +0800
Subject: [PATCH] Add cpu consistence che
Yang, Sheng wrote:
> All the physical CPUs on the board should support the same VMX feature
> set.
> The hardware_enable() do the per-cpu check now.
> In case of vmx/svm enabling failure, transfer a variable in
> hardware_enable()
> for error report.
>
>
> static void hardware_disable(void *junk
All the physical CPUs on the board should support the same VMX feature
set.
The hardware_enable() do the per-cpu check now.
In case of vmx/svm enabling failure, transfer a variable in
hardware_enable()
for error report.
Signed-off-by: Sheng Yang <[EMAIL PROTECTED]>
---
drivers/kvm/kvm_main.c |
11 matches
Mail list logo