Re: [Xen-devel] [PATCH v1] x86/hvm: Generic instruction re-execution mechanism for execute faults

[Razvan Cojocaru <rcojocaru@bitdefender.com>]

On 5/13/19 5:15 PM, Razvan Cojocaru wrote:
> On 5/13/19 5:06 PM, Jan Beulich wrote:
>>>>> On 13.05.19 at 15:58, <rcojocaru@bitdefender.com> wrote:
>>> On 11/27/18 12:49 PM, Razvan Cojocaru wrote:
>>>>> With a sufficiently complete insn emulator, single-stepping should
>>>>> not be needed at all imo. Granted we're not quite there yet with
>>>>> the emulator, but we've made quite a bit of progress. As before,
>>>>> if there are particular instructions you know of that the emulator
>>>>> doesn't handle yet, please keep pointing these out. Last I know
>>>>> were some AVX move instructions, which have long been
>>>>> implemented.
>>>> True, I haven't seen emulator issues in that respect with staging - the
>>>> emulator appears lately to be sufficiently complete. Thank you very 
>>>> much
>>>> for your help and support - we'll definitely point out unsupported
>>>> instructions if we spot some again.
>>>
>>> We've come accross a new instruction that the emulator can't handle in
>>> Xen 4.13-unstable today:
>>>
>>> vpmaddwd xmm4,xmm4,XMMWORD PTR ds:0x513fbb20
>>>
>>> Perhaps there are plans for this to go into the emulator as well?
>>
>> You're kidding? This is already in 4.12.0, and if it weren't I'm sure
>> you're aware there are about 40 more AVX512 patches pending
>> review.
> 
> Right, I did indeed forget about the pending review part, for some 
> reason I was sure they made it in. I've double-checked and we really are 
> using 4.13-unstable - but we've also made changes to the emulator, 
> working on the send-vm-events-from-the-emulator patch, so we'll revert 
> to a pristine staging and retry, there's a chance this happens because 
> of our changes.
> 
> We'll find out what's going on exactly.

I promised I'd return with more details. After some debugging, it 
certainly looks like the emulator returns UNIMPLEMENTED (5):

Mem event emulation failed (5): d5v0 32bit @ 001b:6d96efff -> c5 f9 f5 
05 c0 be ad 6d c5 e1 fe 1d a0 20 af 6d

Looking at the source code, the emulator does appear to support 
vpmaddwd, however only for EVEX:

http://xenbits.xen.org/gitweb/?p=xen.git;a=blob;f=xen/arch/x86/x86_emulate/x86_emulate.c;h=032995ea586aa7dd90a1953b6ded656436652049;hb=refs/heads/staging#l6696

whereas our fail case uses VEX.

This may be in the works in the aforementioned series, but is 
legitimately unsupported in 4.13 staging.


Thanks,
Razvan

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel