[llvm-commits] [llvm] r164662 - in /llvm/trunk: lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp test/CodeGen/ARM/2012-09-25-InlineAsmScalarToVectorConv2.ll

[Eric Christopher]

>> For this part we probably want to diagnose this in the front end, since we
>> have access to the constraint and the type coming in. No objections
>> to having some information here either, but it's definitely not an ideal
>> location. Maybe even a verifier on IR instead that determines that
>> we'd fail rather than having to go through codegen for it. It couldn't catch
>> cases that are only visibile via optimization, but those should be a
>> relatively small sample size.
>>
> Where in the front-end do you think this should go? It may be a bit too much information for the front-end to have to know -- that it's a vector and that these specific scalar-to-vector conversions are wrong...
>

I'd like to see it in the area where we verify constraints, it might
take some target specific knowledge to know whether or not we can
convert a vector though.

Can you share a small set of C code for this so we can take a look at
whether or not it's reasonable to try to do this in the front end? I
wrote some verification code not too long ago that should still be
around somewhere. We can chat about it.

-eric