[PATCH] Serialize the floating point immediate machine operands.

Alex L arphaman at gmail.com
Thu Jul 30 17:41:23 PDT 2015


The updated patch below uses the 'parseConstantValue' function from the
AsmParser
to parse the floating point constants. This radically simplifies the
parsing of floating
point machine operands for the MI parser.
Also, the updated patch can now handle all the floating point types.

Cheers,
Alex

2015-07-30 14:42 GMT-07:00 Alex L <arphaman at gmail.com>:

> Yeah, the IR parser can handle them all of them correctly.
> It would certainly be easier to reuse parseStandaloneConstantValue
> for the floating point machine operands, I'll try to come up with a patch
> that does that.
>
> Alex
>
> 2015-07-30 11:40 GMT-07:00 Duncan P. N. Exon Smith <dexonsmith at apple.com>:
>
>>
>> > On 2015-Jul-29, at 17:23, Alex L <arphaman at gmail.com> wrote:
>> >
>> > Hi,
>> >
>> > The 3 patches below serialize the floating point immediate machine
>> operands
>> > of type half, float, and double.
>> >
>> > The first patch serializes the non-hexadecimal float and double
>> floating point
>> > operands. The second patch serializes the hexadecimal float and double
>> > floating point operands. The third patch serializes the half floating
>> point
>> > operands.
>> >
>> > The x86_fp80, fp128, and ppc_fp128 floating point operands aren't
>> > serialized by these patches.
>> >
>>
>> Do we know how to handle all the float types in IR?  If so, would it
>> be better to reuse `parseStandaloneConstantValue()` for all of these?
>>
>> > Cheers,
>> > Alex
>> >
>> <0001-MIR-Serialization-Initial-serialization-of-floating-.patch><0002-MIR-Serialization-Serialize-the-hexadecimal-floating.patch><0003-MIR-Serialization-Serialize-the-half-floating-point-.patch>
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20150730/dc10f4d1/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-MIR-Serialization-Serialize-the-floating-point-opera.patch
Type: application/octet-stream
Size: 43784 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20150730/dc10f4d1/attachment.obj>


More information about the llvm-commits mailing list