emulation failed but !mmio_needed? rip 10000 fc 0f 01 15

emulation failed but !mmio_needed? rip 10000 fc 0f 01 15

[Neo Jia]

hi,

I happened to get a "emulation fail" when running the following command:

System environment: Intel Core 2 Duo (E6600) x86_64 Fedora 8 (2.6.23.1-49.fc8).

qemu-img create -f qcow debian-testing.img 10G

sudo qemu-system-x86_64 -cdrom
/home/cjia/download/debian-testing-i386-netinst.iso -hda
debian-testing.img -boot d -m 1024

Everything works fine until it prompts me that the installation is
complete and need reboot.

> sudo qemu-system-x86_64 -cdrom /home/cjia/download/debian-testing-i386-netinst.iso -hda debian-testing.img -boot d -m 1024
exception 13 (0)
rax 0000000000000100 rbx 0000000000000100 rcx 000000000000ffff rdx
0000000000000600
rsi 0000000000000000 rdi 0000000000000000 rsp 0000000000000000 rbp
0000000000000000
r8  0000000000000000 r9  0000000000000000 r10 0000000000000000 r11
0000000000000000
r12 0000000000000000 r13 0000000000000000 r14 0000000000000000 r15
0000000000000000
rip 0000000000010000 rflags 00033003
cs f000 (000f0000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
ds 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
es 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
ss 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
fs 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
gs 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
tr 0080 (40c00000/00002088 p 1 dpl 0 db 0 s 0 type b l 0 g 0 avl 0)
ldt 0000 (00000000/0000ffff p 1 dpl 0 db 0 s 0 type 2 l 0 g 0 avl 0)
gdt 0/ffff
idt 0/ffff
cr0 60000010 cr2 0 cr3 0 cr4 0 cr8 0 efer 0
code: 00 00 00 00 ea 5b e0 00 f0 31 30 2f 32 38 2f 30 37 00 fc f5 -->
fc 0f 01 15 92 a0 31 00 b8 18 00 00 00 8e d8 8e c0 8e e0 8e e8 31 c0
bf 00 30 39 00 b9 7c
Aborted

---> dmesg:

kvm: emulating exchange as write
emulation failed but !mmio_needed? rip 10000 fc 0f 01 15    // I
searched through the source code but not find a place to generate "but
!mmio_needed?" string ...

The version of kvm I used for kvm module is
"694401697ccd822bb08019731c3ee1bb34323d8e" and the kvm-userspace is
"a57b838b49bc4e4e7439b18d0323385d53e41c7f".
// Not sure if this is the right way to show the version I am using.
If not, please correct me, this is the first time I am using git...

Anyway, I really want to get some comments about how to trace this
problem and would be very glad to dig into it!

Thanks,
Neo

-- 
I would remember that if researchers were not ambitious
probably today we haven't the technology we are using!

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

Re: emulation failed but !mmio_needed? rip 10000 fc 0f 01 15

[Sheng Yang]

On Saturday 24 November 2007 07:23:20 Neo Jia wrote:
> hi,
>
> I happened to get a "emulation fail" when running the following command:
>
> System environment: Intel Core 2 Duo (E6600) x86_64 Fedora 8
> (2.6.23.1-49.fc8).
>
> qemu-img create -f qcow debian-testing.img 10G
>
> sudo qemu-system-x86_64 -cdrom
> /home/cjia/download/debian-testing-i386-netinst.iso -hda
> debian-testing.img -boot d -m 1024
>
> Everything works fine until it prompts me that the installation is
> complete and need reboot.
>
> > sudo qemu-system-x86_64 -cdrom
> > /home/cjia/download/debian-testing-i386-netinst.iso -hda
> > debian-testing.img -boot d -m 1024
>
> exception 13 (0)
> rax 0000000000000100 rbx 0000000000000100 rcx 000000000000ffff rdx
> 0000000000000600
> rsi 0000000000000000 rdi 0000000000000000 rsp 0000000000000000 rbp
> 0000000000000000
> r8  0000000000000000 r9  0000000000000000 r10 0000000000000000 r11
> 0000000000000000
> r12 0000000000000000 r13 0000000000000000 r14 0000000000000000 r15
> 0000000000000000
> rip 0000000000010000 rflags 00033003
> cs f000 (000f0000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
> ds 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
> es 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
> ss 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
> fs 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
> gs 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
> tr 0080 (40c00000/00002088 p 1 dpl 0 db 0 s 0 type b l 0 g 0 avl 0)
> ldt 0000 (00000000/0000ffff p 1 dpl 0 db 0 s 0 type 2 l 0 g 0 avl 0)
> gdt 0/ffff
> idt 0/ffff
> cr0 60000010 cr2 0 cr3 0 cr4 0 cr8 0 efer 0
> code: 00 00 00 00 ea 5b e0 00 f0 31 30 2f 32 38 2f 30 37 00 fc f5 -->
> fc 0f 01 15 92 a0 31 00 b8 18 00 00 00 8e d8 8e c0 8e e0 8e e8 31 c0
> bf 00 30 39 00 b9 7c
> Aborted
>
> ---> dmesg:
>
> kvm: emulating exchange as write
> emulation failed but !mmio_needed? rip 10000 fc 0f 01 15    // I
> searched through the source code but not find a place to generate "but
> !mmio_needed?" string ...

It due to one instruction haven't been emulated: 0xfc(one byte instruction). 
Look it up in SDM Vol.2B's opcode map, you can find that's "CLD". And look at 
what is CLD in SDM Vol.2, you will find "CLD - Clear direction flag". So we 
miss this in x86_emulate.c.

>
> The version of kvm I used for kvm module is
> "694401697ccd822bb08019731c3ee1bb34323d8e" and the kvm-userspace is
> "a57b838b49bc4e4e7439b18d0323385d53e41c7f".
> // Not sure if this is the right way to show the version I am using.
> If not, please correct me, this is the first time I am using git...

Yeah, that's the right ones. :)

>
> Anyway, I really want to get some comments about how to trace this
> problem and would be very glad to dig into it!

Here is the chance! Add this instruction to our x86 emulator. You can take the 
implement of "sti" and "cli" as a reference.

Good luck!

>
> Thanks,
> Neo



-- 
Thanks
Yang, Sheng

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

Re: emulation failed but !mmio_needed? rip 10000 fc 0f 01 15

[Avi Kivity]

Sheng Yang wrote:
> On Saturday 24 November 2007 07:23:20 Neo Jia wrote:
>   
>> hi,
>>
>> I happened to get a "emulation fail" when running the following command:
>>
>> System environment: Intel Core 2 Duo (E6600) x86_64 Fedora 8
>> (2.6.23.1-49.fc8).
>>
>> qemu-img create -f qcow debian-testing.img 10G
>>
>> sudo qemu-system-x86_64 -cdrom
>> /home/cjia/download/debian-testing-i386-netinst.iso -hda
>> debian-testing.img -boot d -m 1024
>>
>> Everything works fine until it prompts me that the installation is
>> complete and need reboot.
>>
>>     
>>> sudo qemu-system-x86_64 -cdrom
>>> /home/cjia/download/debian-testing-i386-netinst.iso -hda
>>> debian-testing.img -boot d -m 1024
>>>       
>> exception 13 (0)
>> rax 0000000000000100 rbx 0000000000000100 rcx 000000000000ffff rdx
>> 0000000000000600
>> rsi 0000000000000000 rdi 0000000000000000 rsp 0000000000000000 rbp
>> 0000000000000000
>> r8  0000000000000000 r9  0000000000000000 r10 0000000000000000 r11
>> 0000000000000000
>> r12 0000000000000000 r13 0000000000000000 r14 0000000000000000 r15
>> 0000000000000000
>> rip 0000000000010000 rflags 00033003
>> cs f000 (000f0000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
>> ds 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
>> es 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
>> ss 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
>> fs 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
>> gs 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
>> tr 0080 (40c00000/00002088 p 1 dpl 0 db 0 s 0 type b l 0 g 0 avl 0)
>> ldt 0000 (00000000/0000ffff p 1 dpl 0 db 0 s 0 type 2 l 0 g 0 avl 0)
>> gdt 0/ffff
>> idt 0/ffff
>> cr0 60000010 cr2 0 cr3 0 cr4 0 cr8 0 efer 0
>> code: 00 00 00 00 ea 5b e0 00 f0 31 30 2f 32 38 2f 30 37 00 fc f5 -->
>> fc 0f 01 15 92 a0 31 00 b8 18 00 00 00 8e d8 8e c0 8e e0 8e e8 31 c0
>> bf 00 30 39 00 b9 7c
>> Aborted
>>
>> ---> dmesg:
>>
>> kvm: emulating exchange as write
>> emulation failed but !mmio_needed? rip 10000 fc 0f 01 15    // I
>> searched through the source code but not find a place to generate "but
>> !mmio_needed?" string ...
>>     
>
>   

That's been removed some time ago (and replaced by another string). See 
054b1369679fb97582fc77f25a700d4290ff3e89.


> It due to one instruction haven't been emulated: 0xfc(one byte instruction). 
> Look it up in SDM Vol.2B's opcode map, you can find that's "CLD". And look at 
> what is CLD in SDM Vol.2, you will find "CLD - Clear direction flag". So we 
> miss this in x86_emulate.c.
>
>   
>> The version of kvm I used for kvm module is
>> "694401697ccd822bb08019731c3ee1bb34323d8e" and the kvm-userspace is
>> "a57b838b49bc4e4e7439b18d0323385d53e41c7f".
>> // Not sure if this is the right way to show the version I am using.
>> If not, please correct me, this is the first time I am using git...
>>     
>
> Yeah, that's the right ones. :)
>
>   
>> Anyway, I really want to get some comments about how to trace this
>> problem and would be very glad to dig into it!
>>     
>
> Here is the chance! Add this instruction to our x86 emulator. You can take the 
> implement of "sti" and "cli" as a reference.

While adding the instrucion is helpful, something else has gone wrong 
here.  cs:ip == f000:10000.  You can see the far jump at f000:fff0 
(opcode ea) -- kvm just skipped over it somehow.

Ah, it's been fixed already: see 
c408e4e8d9045d53c1d82c622a5756febd051ef9.  I need to backport it to 2.6.23.

-- 
Any sufficiently difficult bug is indistinguishable from a feature.


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

Re: emulation failed but !mmio_needed? rip 10000 fc 0f 01 15

[Neo Jia]

On Nov 23, 2007 8:21 PM, Sheng Yang <sheng.yang-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org> wrote:
>
> On Saturday 24 November 2007 07:23:20 Neo Jia wrote:
> > hi,
> >
> > I happened to get a "emulation fail" when running the following command:
> >
> > System environment: Intel Core 2 Duo (E6600) x86_64 Fedora 8
> > (2.6.23.1-49.fc8).
> >
> > qemu-img create -f qcow debian-testing.img 10G
> >
> > sudo qemu-system-x86_64 -cdrom
> > /home/cjia/download/debian-testing-i386-netinst.iso -hda
> > debian-testing.img -boot d -m 1024
> >
> > Everything works fine until it prompts me that the installation is
> > complete and need reboot.
> >
> > > sudo qemu-system-x86_64 -cdrom
> > > /home/cjia/download/debian-testing-i386-netinst.iso -hda
> > > debian-testing.img -boot d -m 1024
> >
> > exception 13 (0)
> > rax 0000000000000100 rbx 0000000000000100 rcx 000000000000ffff rdx
> > 0000000000000600
> > rsi 0000000000000000 rdi 0000000000000000 rsp 0000000000000000 rbp
> > 0000000000000000
> > r8  0000000000000000 r9  0000000000000000 r10 0000000000000000 r11
> > 0000000000000000
> > r12 0000000000000000 r13 0000000000000000 r14 0000000000000000 r15
> > 0000000000000000
> > rip 0000000000010000 rflags 00033003
> > cs f000 (000f0000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
> > ds 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
> > es 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
> > ss 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
> > fs 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
> > gs 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
> > tr 0080 (40c00000/00002088 p 1 dpl 0 db 0 s 0 type b l 0 g 0 avl 0)
> > ldt 0000 (00000000/0000ffff p 1 dpl 0 db 0 s 0 type 2 l 0 g 0 avl 0)
> > gdt 0/ffff
> > idt 0/ffff
> > cr0 60000010 cr2 0 cr3 0 cr4 0 cr8 0 efer 0
> > code: 00 00 00 00 ea 5b e0 00 f0 31 30 2f 32 38 2f 30 37 00 fc f5 -->
> > fc 0f 01 15 92 a0 31 00 b8 18 00 00 00 8e d8 8e c0 8e e0 8e e8 31 c0
> > bf 00 30 39 00 b9 7c
> > Aborted
> >
> > ---> dmesg:
> >
> > kvm: emulating exchange as write
> > emulation failed but !mmio_needed? rip 10000 fc 0f 01 15    // I
> > searched through the source code but not find a place to generate "but
> > !mmio_needed?" string ...
>
> It due to one instruction haven't been emulated: 0xfc(one byte instruction).
> Look it up in SDM Vol.2B's opcode map, you can find that's "CLD". And look at
> what is CLD in SDM Vol.2, you will find "CLD - Clear direction flag". So we
> miss this in x86_emulate.c.

Could you show me how to decode the log message "rip 10000 fc 0f 01 15". How
can you determine that it is a one byte instruction?

Thanks,
Neo


>
> >
> > The version of kvm I used for kvm module is
> > "694401697ccd822bb08019731c3ee1bb34323d8e" and the kvm-userspace is
> > "a57b838b49bc4e4e7439b18d0323385d53e41c7f".
> > // Not sure if this is the right way to show the version I am using.
> > If not, please correct me, this is the first time I am using git...
>
> Yeah, that's the right ones. :)
>
> >
> > Anyway, I really want to get some comments about how to trace this
> > problem and would be very glad to dig into it!
>
> Here is the chance! Add this instruction to our x86 emulator. You can take the
> implement of "sti" and "cli" as a reference.
>
> Good luck!
>
> >
> > Thanks,
> > Neo
>
>
>
> --
> Thanks
> Yang, Sheng
>



-- 
I would remember that if researchers were not ambitious
probably today we haven't the technology we are using!

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

Re: emulation failed but !mmio_needed? rip 10000 fc 0f 01 15

[Avi Kivity]

Neo Jia wrote:
> The version of kvm I used for kvm module is
> "694401697ccd822bb08019731c3ee1bb34323d8e" and the kvm-userspace is
> "a57b838b49bc4e4e7439b18d0323385d53e41c7f".
>   

These are very recent versions, but the nature of the problem leads me 
to expect you are using the host's kvm modules.  Can you verify?

-- 
Any sufficiently difficult bug is indistinguishable from a feature.


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

Re: emulation failed but !mmio_needed? rip 10000 fc 0f 01 15

[Neo Jia]

On Nov 24, 2007 12:00 AM, Avi Kivity <avi-atKUWr5tajBWk0Htik3J/w@public.gmane.org> wrote:
> Neo Jia wrote:
> > The version of kvm I used for kvm module is
> > "694401697ccd822bb08019731c3ee1bb34323d8e" and the kvm-userspace is
> > "a57b838b49bc4e4e7439b18d0323385d53e41c7f".
> >
>
> These are very recent versions, but the nature of the problem leads me
> to expect you are using the host's kvm modules.  Can you verify?

Yes, I found that also and doing the verification now.

Thanks,
Neo

>
>
> --
> Any sufficiently difficult bug is indistinguishable from a feature.
>
>



-- 
I would remember that if researchers were not ambitious
probably today we haven't the technology we are using!

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

Re: emulation failed but !mmio_needed? rip 10000 fc 0f 01 15

[Neo Jia]

On Nov 24, 2007 12:00 AM, Avi Kivity <avi-atKUWr5tajBWk0Htik3J/w@public.gmane.org> wrote:
> Neo Jia wrote:
> > The version of kvm I used for kvm module is
> > "694401697ccd822bb08019731c3ee1bb34323d8e" and the kvm-userspace is
> > "a57b838b49bc4e4e7439b18d0323385d53e41c7f".
> >
>
> These are very recent versions, but the nature of the problem leads me
> to expect you are using the host's kvm modules.  Can you verify?

Yes. Just verified that I ran the modules from kernel "2.6.23.1-49.fc8".

But according to Sheng's comments, I also don't see the "cld"
instruction in the latest version.

Why does kvm not crash?

Thanks,
Neo

>
>
> --
> Any sufficiently difficult bug is indistinguishable from a feature.
>
>



-- 
I would remember that if researchers were not ambitious
probably today we haven't the technology we are using!

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

Re: emulation failed but !mmio_needed? rip 10000 fc 0f 01 15

[Avi Kivity]

Neo Jia wrote:
> On Nov 24, 2007 12:00 AM, Avi Kivity <avi-atKUWr5tajBWk0Htik3J/w@public.gmane.org> wrote:
>   
>> Neo Jia wrote:
>>     
>>> The version of kvm I used for kvm module is
>>> "694401697ccd822bb08019731c3ee1bb34323d8e" and the kvm-userspace is
>>> "a57b838b49bc4e4e7439b18d0323385d53e41c7f".
>>>
>>>       
>> These are very recent versions, but the nature of the problem leads me
>> to expect you are using the host's kvm modules.  Can you verify?
>>     
>
> Yes. Just verified that I ran the modules from kernel "2.6.23.1-49.fc8".
>
> But according to Sheng's comments, I also don't see the "cld"
> instruction in the latest version.
>
> Why does kvm not crash?
>
>   

With the reset fix, kvm never sees the cld instruction.  It executes the 
jmp instruction at f000:fff0 and continues processing ROM code.

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


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

Re: emulation failed but !mmio_needed? rip 10000 fc 0f 01 15

[Neo Jia]

Avi Kivity wrote:
> Neo Jia wrote:
>> On Nov 24, 2007 12:00 AM, Avi Kivity <avi-atKUWr5tajBWk0Htik3J/w@public.gmane.org> wrote:
>>  
>>> Neo Jia wrote:
>>>    
>>>> The version of kvm I used for kvm module is
>>>> "694401697ccd822bb08019731c3ee1bb34323d8e" and the kvm-userspace is
>>>> "a57b838b49bc4e4e7439b18d0323385d53e41c7f".
>>>>
>>>>       
>>> These are very recent versions, but the nature of the problem leads me
>>> to expect you are using the host's kvm modules.  Can you verify?
>>>     
>>
>> Yes. Just verified that I ran the modules from kernel "2.6.23.1-49.fc8".
>>
>> But according to Sheng's comments, I also don't see the "cld"
>> instruction in the latest version.
>>
>> Why does kvm not crash?
>>
>>   
>
> With the reset fix, kvm never sees the cld instruction.  It executes 
> the jmp instruction at f000:fff0 and continues processing ROM code.
>
Do we need to add a "default" in x86 emulator switch statement?

Thanks,
Neo

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

Re: emulation failed but !mmio_needed? rip 10000 fc 0f 01 15

[Avi Kivity]

Neo Jia wrote:
> Avi Kivity wrote:
>   
>> Neo Jia wrote:
>>     
>>> On Nov 24, 2007 12:00 AM, Avi Kivity <avi-atKUWr5tajBWk0Htik3J/w@public.gmane.org> wrote:
>>>  
>>>       
>>>> Neo Jia wrote:
>>>>    
>>>>         
>>>>> The version of kvm I used for kvm module is
>>>>> "694401697ccd822bb08019731c3ee1bb34323d8e" and the kvm-userspace is
>>>>> "a57b838b49bc4e4e7439b18d0323385d53e41c7f".
>>>>>
>>>>>       
>>>>>           
>>>> These are very recent versions, but the nature of the problem leads me
>>>> to expect you are using the host's kvm modules.  Can you verify?
>>>>     
>>>>         
>>> Yes. Just verified that I ran the modules from kernel "2.6.23.1-49.fc8".
>>>
>>> But according to Sheng's comments, I also don't see the "cld"
>>> instruction in the latest version.
>>>
>>> Why does kvm not crash?
>>>
>>>   
>>>       
>> With the reset fix, kvm never sees the cld instruction.  It executes 
>> the jmp instruction at f000:fff0 and continues processing ROM code.
>>
>>     
> Do we need to add a "default" in x86 emulator switch statement?
>
>   

Take a look at the code.  That path is already covered.


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


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

Re: emulation failed but !mmio_needed? rip 10000 fc 0f 01 15

[Neo Jia]

On Nov 26, 2007 2:16 AM, Avi Kivity <avi-atKUWr5tajBWk0Htik3J/w@public.gmane.org> wrote:
>
> Neo Jia wrote:
> > Avi Kivity wrote:
> >
> >> Neo Jia wrote:
> >>
> >>> On Nov 24, 2007 12:00 AM, Avi Kivity <avi-atKUWr5tajBWk0Htik3J/w@public.gmane.org> wrote:
> >>>
> >>>
> >>>> Neo Jia wrote:
> >>>>
> >>>>
> >>>>> The version of kvm I used for kvm module is
> >>>>> "694401697ccd822bb08019731c3ee1bb34323d8e" and the kvm-userspace is
> >>>>> "a57b838b49bc4e4e7439b18d0323385d53e41c7f".
> >>>>>
> >>>>>
> >>>>>
> >>>> These are very recent versions, but the nature of the problem leads me
> >>>> to expect you are using the host's kvm modules.  Can you verify?
> >>>>
> >>>>
> >>> Yes. Just verified that I ran the modules from kernel "2.6.23.1-49.fc8".
> >>>
> >>> But according to Sheng's comments, I also don't see the "cld"
> >>> instruction in the latest version.
> >>>
> >>> Why does kvm not crash?
> >>>
> >>>
> >>>
> >> With the reset fix, kvm never sees the cld instruction.  It executes
> >> the jmp instruction at f000:fff0 and continues processing ROM code.
> >>
> >>
> > Do we need to add a "default" in x86 emulator switch statement?
> >
> >
>
> Take a look at the code.  That path is already covered.

Avi,

I just checkout the latest kvm.git. Please correct me if I am wrong.

I don't see the explicit "default" in swtich (c->b) statement but I
found a goto writeback.

Is that one?

Thanks,
Neo


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



-- 
I would remember that if researchers were not ambitious
probably today we haven't the technology we are using!

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

Re: emulation failed but !mmio_needed? rip 10000 fc 0f 01 15

[Avi Kivity]

Neo Jia wrote:
>>>
>>> Do we need to add a "default" in x86 emulator switch statement?
>>>
>>>
>>>       
>> Take a look at the code.  That path is already covered.
>>     
>
> Avi,
>
> I just checkout the latest kvm.git. Please correct me if I am wrong.
>
> I don't see the explicit "default" in swtich (c->b) statement but I
> found a goto writeback.
>   

If the insn decode flag word c->d (derived from opcode_table) is zero, 
we return early.  See the test just after the label 'done_prefixes'.

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


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/