[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH 1/3] x86: add support for computing the instruction length

On 09/09/14 17:25, Jan Beulich wrote:
>>>> On 09.09.14 at 18:01, <mdontu@xxxxxxxxxxxxxxx> wrote:
>> I've opted to send a new mail so I can remove Masami from CC, as he's
>> probably not interested in the rest of the conversation.
>> Right now we have two patches which work around x86/emulator
>> limitations:
>>  * one computes the instruction length;
>>  * the other uses single stepping to jump over unsupported instructions;
>> Adding support for the complete x86(_64) instruction set to the
>> existent emulator in Xen would make those two unneeded and while I
>> would like to try my hand at it, I'm not sure the effort would be pay
>> off. Not to mention that I would very much like to _somehow_ catch the
>> 4.5 deadline. I wonder if it's possible to do this in iterations: take
>> this (or a decent derivation of it) in, while RÄzvan and I work on doing
>> a better work for 4.6. Am I pushing it? :-)
> Personally I don't think this makes sense to push for 4.5, but in the
> end it'll be Konrad's call. We already have enough other half-way
> reviewed patch series that need finalizing, so I don't think this series
> (which was posted just once many weeks ago) is a candidate.
> Furthermore I'm rather unconvinced of this being code useful to
> other than just your product. And finally, we had (with other
> submitters) some bad experience in the past taking what they
> promised they would clean up later.

I would agree that pushing for 4.5 is unlikely to happen.  I would
suggest that your efforts would be be better spent getting a high
quality series for 4.6 sorted.

As for the instruction length itself, that could be sorted in a more
generic way.

One issue with the current x86_emulate() is that it mixes instruction
decode with instruction emulation.  The AMD SVM code already has a
partial instruction length checker for certain cases where hardware
support is lacking.

It seems plausible to split instruction decode and instruction emulate
into two distinct steps.  This would allow easier unit testing of each
step (always a good thing), and for the instruction decode to be used
independently of emulation.


Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.