[llvm-commits] [llvm] r158470 - in /llvm/trunk: lib/Target/Mips/Mips64InstrInfo.td lib/Target/Mips/MipsCondMov.td lib/Target/Mips/MipsInstrFPU.td lib/Target/Mips/MipsInstrInfo.td test/CodeGen/Mips/null.ll

reed kotler rkotler at mips.com
Mon Jun 18 13:53:08 PDT 2012


On 06/16/2012 06:09 PM, Chris Lattner wrote:
> On Jun 14, 2012, at 2:03 PM, Akira Hatanaka wrote:
>
>> Author: ahatanak
>> Date: Thu Jun 14 16:03:23 2012
>> New Revision: 158470
>>
>> URL: http://llvm.org/viewvc/llvm-project?rev=158470&view=rev
>> Log:
>> 1. introduce MipsPat in place of Pat in order to exclude those from
>> being used by Mips16 or Micro Mips
>> 2. clean up a few lines too long encountered
>>
>> Patch by Reed Kotler.
> I'm not familiar with mips encoding, but would it make more sense to name it Mips32Pat to make it explicit that it is not for mips16?
>
> -Chris
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
This is a tricky naming situation. We will think about it some more.

We in general have used "mips" to refer to the standard instruction 
encoding.
This instruction encoding is used for both mips32 and mips64.

Mips16 and Micro Mips are different encodings.

In terms of architecture, there only exists Mips32 and Mips64.

There is no Mip16 architecture, it's just an alternate instruction 
encoding but the core
processor is the same. You can switch between instruction encoding in a 
single program.
I.e. go back and forth between Mips and Mips16 encoding (in either a 
mips32 or mips64 processor).


Reed






More information about the llvm-commits mailing list