[llvm-dev] error about adding an trinsics

[沈天豪 via llvm-dev]

Thank you very much, especially for copying and testing my code.
I am very surprised that the builtin is recognized. I successfully compile without errors.But when I use the fllowing file to test ,the error occurs.

#include <stdio.h>
int main()
{
  int a;
  int b;
  a=1;b=2;
  c = __builtin_x86_max_qb(a, b);

  return 0;

}

There is a place I don't understand very well, I want to ask you.When did the fatal error you get occur？Compile or test after compiled？May you tell me how you test ?

As for something to connect the intrinsic to the X86ISD::max_qb opcode, do I need to add a DAG?
Thanks again




At 2018-09-17 14:39:13, "Craig Topper" <craig.topper at gmail.com> wrote:

This all looks fine to me. I even pasted into my own repo and the builtin is recognized by my local build of clang. I got this error instead.



fatal error: error in backend: Cannot select: intrinsic %llvm.x86.max.qb



But the code you provide is missing anything to connect the intrinsic to the X86ISD::max_qb opcode so that error is to be expected given what I copied and pasted.


~Craig




On Sun, Sep 16, 2018 at 11:06 PM 沈天豪 via llvm-dev <llvm-dev at lists.llvm.org> wrote:

Hi，every one.
This problem has been bothering me for several days.I really hope that you can help me.
I want to add an trinsics in X86. This trinsics can compare two numbers and return the larger.
There are the changes I do as fllowing.


In /tools/clang/include/clang/Basic/BuiltinsX86.def :

BUILTIN(__builtin_x86_max_qb, "iii", "")

In include/llvm/IR/IntrinsicsX86.td :

 let TargetPrefix = "x86" in {
 def int_x86_max_qb: GCCBuiltin<"__builtin_x86_max_qb">,
  Intrinsic<[llvm_i32_ty], [llvm_i32_ty, llvm_i32_ty], [IntrNoMem]>;
}
In /lib/Target/X86/X86SelLowering.h: add a sdnode
    max_qb,

In /lib/Target/X86/X86SelLowering.cpp:

case X86ISD::max_qb:             return "X86ISD::max_qb";
In /lib/Target/X86/X86InstrInfo.td:
def X86max_qb_flag : SDNode<"X86ISD::max_qb", SDTBinaryArithWithFlags,
                          [SDNPCommutative]>;
In /lib/Target/X86/X86InstrArithmetic.td:
def max_qb : I<0xff,MRMSrcReg, (outs  GR32:$dst), (ins GR32:$src1,GR32:$src2),
             "max_qb\t {$dst, $src1,$src2|$src1,$src2, $dst}", [(set GR32:$dst,EFLAGS,(X86max_qb_flag GR32:$src1, GR32:$src2))]>,
              Sched<[WriteIMul]>, TB, OpSize32 ;

I think it can be work ,at least work as one multiplication(because I use the Sched<[WriteIMul]>).
But there is an error :"error: use of unknown builtin '__builtin_x86_max_qb'". And I don't konw what I should do.
Thanks a lot.






 

_______________________________________________
LLVM Developers mailing list
llvm-dev at lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20180917/9afae774/attachment.html>