Re: [Patch 0/7] RealMode: Support for big real mode

Re: [Patch 0/7] RealMode: Support for big real mode

[Avi Kivity]

Guillaume Thouvenin wrote:
> Hello, 
>
>   Those patches allow to boot OpenSuse-10.3. The problem with
> Opensuse 10.3 is it uses a version of gfxboot that reads SS after
> switching from real to protected mode, where SS contains an invalid
> value, which VMX does not allow.
>   

Thanks, applied all. I dropped vcpu_sregs_table after all, in favor of 
1:1 mapping of VCPU_SREG_* to the opcode encoding (similar to the 
general purpose registers).

-- 
Do not meddle in the internals of kernels, for they are subtle and quick to panic.

Re: [Patch 0/7] RealMode: Support for big real mode

[Anthony Liguori]

Avi Kivity wrote:
> Guillaume Thouvenin wrote:
>> Hello,
>>   Those patches allow to boot OpenSuse-10.3. The problem with
>> Opensuse 10.3 is it uses a version of gfxboot that reads SS after
>> switching from real to protected mode, where SS contains an invalid
>> value, which VMX does not allow.
>>   
>
> Thanks, applied all. I dropped vcpu_sregs_table after all, in favor of 
> 1:1 mapping of VCPU_SREG_* to the opcode encoding (similar to the 
> general purpose registers).

There still seems to be an issue with this series.  When running:

sudo qemu-system-x86_64 -hda ./10gig.img -snapshot -cdrom 
isos/ubuntu-7.10-server-amd64.iso -boot dc

Occassionally, I'll get:

kvm_run: failed entry, reason 5
kvm_run returned -8

There's nothing in dmesg though.  Can you reproduce this?  If not, I'll 
dig in some more.

Regards,

Anthony Liguori

Re: [Patch 0/7] RealMode: Support for big real mode

[Avi Kivity]

Anthony Liguori wrote:
> Avi Kivity wrote:
>> Guillaume Thouvenin wrote:
>>> Hello,
>>>   Those patches allow to boot OpenSuse-10.3. The problem with
>>> Opensuse 10.3 is it uses a version of gfxboot that reads SS after
>>> switching from real to protected mode, where SS contains an invalid
>>> value, which VMX does not allow.
>>>   
>>
>> Thanks, applied all. I dropped vcpu_sregs_table after all, in favor 
>> of 1:1 mapping of VCPU_SREG_* to the opcode encoding (similar to the 
>> general purpose registers).
>
> There still seems to be an issue with this series.  When running:
>
> sudo qemu-system-x86_64 -hda ./10gig.img -snapshot -cdrom 
> isos/ubuntu-7.10-server-amd64.iso -boot dc
>
> Occassionally, I'll get:
>
> kvm_run: failed entry, reason 5
> kvm_run returned -8

Occasionally, on boot, or after booting has completed?

This isn't a regression if it's on boot only.

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

Re: [Patch 0/7] RealMode: Support for big real mode

[Guillaume Thouvenin]

On Thu, 29 May 2008 12:34:16 +0300
Avi Kivity <avi@qumranet.com> wrote:

> Anthony Liguori wrote:
> > sudo qemu-system-x86_64 -hda ./10gig.img -snapshot -cdrom 
> > isos/ubuntu-7.10-server-amd64.iso -boot dc
> >
> > Occassionally, I'll get:
> >
> > kvm_run: failed entry, reason 5
> > kvm_run returned -8
> 
> Occasionally, on boot, or after booting has completed?
> 
> This isn't a regression if it's on boot only.

It's only on boot, I never had failure after booting has completed. 

Regards,
Guillaume

Re: [Patch 0/7] RealMode: Support for big real mode

[Avi Kivity]

Guillaume Thouvenin wrote:
> On Thu, 29 May 2008 12:34:16 +0300
> Avi Kivity <avi@qumranet.com> wrote:
>
>   
>> Anthony Liguori wrote:
>>     
>>> sudo qemu-system-x86_64 -hda ./10gig.img -snapshot -cdrom 
>>> isos/ubuntu-7.10-server-amd64.iso -boot dc
>>>
>>> Occassionally, I'll get:
>>>
>>> kvm_run: failed entry, reason 5
>>> kvm_run returned -8
>>>       
>> Occasionally, on boot, or after booting has completed?
>>
>> This isn't a regression if it's on boot only.
>>     
>
> It's only on boot, I never had failure after booting has completed. 
>
>   

It may be due to an interrupt being injected when we are emulating.

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

Re: [Patch 0/7] RealMode: Support for big real mode

[Anthony Liguori]

Avi Kivity wrote:
> Guillaume Thouvenin wrote:
>> On Thu, 29 May 2008 12:34:16 +0300
>> Avi Kivity <avi@qumranet.com> wrote:
>>
>>  
>>> Anthony Liguori wrote:
>>>    
>>>> sudo qemu-system-x86_64 -hda ./10gig.img -snapshot -cdrom 
>>>> isos/ubuntu-7.10-server-amd64.iso -boot dc
>>>>
>>>> Occassionally, I'll get:
>>>>
>>>> kvm_run: failed entry, reason 5
>>>> kvm_run returned -8
>>>>       
>>> Occasionally, on boot, or after booting has completed?
>>>
>>> This isn't a regression if it's on boot only.
>>>     
>>
>> It's only on boot, I never had failure after booting has completed.
>>   
>
> It may be due to an interrupt being injected when we are emulating.

Hrm, doesn't appear to be that.  I added some logging and I'm not 
hitting it.

Regards,

Anthony Liguori

Re: [Patch 0/7] RealMode: Support for big real mode

[Anthony Liguori]

Avi Kivity wrote:
> Anthony Liguori wrote:
>> Avi Kivity wrote:
>>> Guillaume Thouvenin wrote:
>>>> Hello,
>>>>   Those patches allow to boot OpenSuse-10.3. The problem with
>>>> Opensuse 10.3 is it uses a version of gfxboot that reads SS after
>>>> switching from real to protected mode, where SS contains an invalid
>>>> value, which VMX does not allow.
>>>>   
>>>
>>> Thanks, applied all. I dropped vcpu_sregs_table after all, in favor 
>>> of 1:1 mapping of VCPU_SREG_* to the opcode encoding (similar to the 
>>> general purpose registers).
>>
>> There still seems to be an issue with this series.  When running:
>>
>> sudo qemu-system-x86_64 -hda ./10gig.img -snapshot -cdrom 
>> isos/ubuntu-7.10-server-amd64.iso -boot dc
>>
>> Occassionally, I'll get:
>>
>> kvm_run: failed entry, reason 5
>> kvm_run returned -8
>
> Occasionally, on boot, or after booting has completed?

I'm pretty sure that it's limited to the functionality added by this code.

> This isn't a regression if it's on boot only.

No, I don't think it's a regression.

Regards,

Anthony Liguori