On 06/18/2014 11:53 PM, Ren, Qiaowei wrote:
> On 2014-06-19, Borislav Petkov wrote:
>> On Thu, Jun 19, 2014 at 01:13:48AM +0000, Ren, Qiaowei wrote:
>>> On 2014-06-18, Borislav Petkov wrote:
>>>> On Wed, Jun 18, 2014 at 05:44:13PM +0800, Qiaowei Ren wrote:
>>>>
>>>> This whole insn decoding machinery above looks like adapted from
>>>> arch/x86/lib/insn.c. You should merge it with the generic code in
>>>> insn.c instead of homegrowing it here only for the purposes of MPX.
>>>> And if it doesn't work for your needs, you should should extend
>>>> the generic code to do so.
>>
>>> Petkov, as we discussed on initial version of this patchset, general
>>> insn framework didn't work out well and I have tried to use generic
>>> struct insn, insn_field, etc. for obvious benefits.
>>
>> Let me repeat myself: "And if it doesn't work for your needs, you
>> should extend the generic code to do so."
>>
>> We don't do homegrown almost-copies of generic code.
>> 
> I see. If possible, I will be very happy to use or extend generic
> code. But due to extra overhead caused by MPX, I have to use MPX
> specific decoding to do performance optimization.

Could you please support this position with some data?  I'm a bit
skeptical that instruction decoding is going to be a
performance-critical path.

I also don't see the extra field that you talked about in the previous
thread?  What's the extra field?  I see a 'limit' vs. 'length', but you
don't use 'length' at all, so I think you can use it instead, or at
least union it.

I've taken a quick stab at trying to consolidate things.  I think I may
have screwed up this:

	insn->limit = MAX_MPX_INSN_SIZE - bytes;

Qiaowei, is there anything fundamentally broken with what I've got here?