746c682c4a
According to the offical LoongArch reference manual, the 32-bit
LoongArch is divied into two variants: the Reduced version (LA32R) and
Standard version (LA32S). LA32S extends LA32R by adding additional
instructions, and the 64-bit version (LA64) fully includes the LA32S
instruction set.
This patch introduces a new target feature `32s` for the LoongArch
backend, enabling support for instructions specific to the LA32S
variant.
The LA32S exntension includes the following additional instructions:
- ALSL.W
- {AND,OR}N
- B{EQ,NE}Z
- BITREV.{4B,W}
- BSTR{INS,PICK}.W
- BYTEPICK.W
- CL{O,Z}.W
- CPUCFG
- CT{O,Z}.W
- EXT.W,{B,H}
- F{LD,ST}X.{D,S}
- MASK{EQ,NE}Z
- PC{ADDI,ALAU12I}
- REVB.2H
- ROTR{I},W
Additionally, LA32R defines three new instruction aliases:
- RDCNTID.W RJ => RDTIMEL.W ZERO, RJ
- RDCNTVH.W RD => RDTIMEH.W RD, ZERO
- RDCNTVL.W RD => RDTIMEL.W RD, ZERO
124 lines
3.6 KiB
LLVM
124 lines
3.6 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc --mtriple=loongarch32 --mattr=+32s,+f,-d < %s | FileCheck %s --check-prefix=LA32F
|
|
; RUN: llc --mtriple=loongarch32 --mattr=+32s,+d < %s | FileCheck %s --check-prefix=LA32D
|
|
; RUN: llc --mtriple=loongarch64 --mattr=+f,-d < %s | FileCheck %s --check-prefix=LA64F
|
|
; RUN: llc --mtriple=loongarch64 --mattr=+d < %s | FileCheck %s --check-prefix=LA64D
|
|
|
|
declare float @llvm.copysign.f32(float, float)
|
|
declare double @llvm.copysign.f64(double, double)
|
|
|
|
define float @fcopysign_s(float %a, float %b) nounwind {
|
|
; LA32F-LABEL: fcopysign_s:
|
|
; LA32F: # %bb.0:
|
|
; LA32F-NEXT: fcopysign.s $fa0, $fa0, $fa1
|
|
; LA32F-NEXT: ret
|
|
;
|
|
; LA32D-LABEL: fcopysign_s:
|
|
; LA32D: # %bb.0:
|
|
; LA32D-NEXT: fcopysign.s $fa0, $fa0, $fa1
|
|
; LA32D-NEXT: ret
|
|
;
|
|
; LA64F-LABEL: fcopysign_s:
|
|
; LA64F: # %bb.0:
|
|
; LA64F-NEXT: fcopysign.s $fa0, $fa0, $fa1
|
|
; LA64F-NEXT: ret
|
|
;
|
|
; LA64D-LABEL: fcopysign_s:
|
|
; LA64D: # %bb.0:
|
|
; LA64D-NEXT: fcopysign.s $fa0, $fa0, $fa1
|
|
; LA64D-NEXT: ret
|
|
%1 = call float @llvm.copysign.f32(float %a, float %b)
|
|
ret float %1
|
|
}
|
|
|
|
define double @fcopysign_d(double %a, double %b) nounwind {
|
|
; LA32F-LABEL: fcopysign_d:
|
|
; LA32F: # %bb.0:
|
|
; LA32F-NEXT: srli.w $a2, $a3, 31
|
|
; LA32F-NEXT: bstrins.w $a1, $a2, 31, 31
|
|
; LA32F-NEXT: ret
|
|
;
|
|
; LA32D-LABEL: fcopysign_d:
|
|
; LA32D: # %bb.0:
|
|
; LA32D-NEXT: fcopysign.d $fa0, $fa0, $fa1
|
|
; LA32D-NEXT: ret
|
|
;
|
|
; LA64F-LABEL: fcopysign_d:
|
|
; LA64F: # %bb.0:
|
|
; LA64F-NEXT: srli.d $a1, $a1, 63
|
|
; LA64F-NEXT: bstrins.d $a0, $a1, 63, 63
|
|
; LA64F-NEXT: ret
|
|
;
|
|
; LA64D-LABEL: fcopysign_d:
|
|
; LA64D: # %bb.0:
|
|
; LA64D-NEXT: fcopysign.d $fa0, $fa0, $fa1
|
|
; LA64D-NEXT: ret
|
|
%1 = call double @llvm.copysign.f64(double %a, double %b)
|
|
ret double %1
|
|
}
|
|
|
|
define double @fold_promote_d_s(double %a, float %b) nounwind {
|
|
; LA32F-LABEL: fold_promote_d_s:
|
|
; LA32F: # %bb.0:
|
|
; LA32F-NEXT: movfr2gr.s $a2, $fa0
|
|
; LA32F-NEXT: srli.w $a2, $a2, 31
|
|
; LA32F-NEXT: bstrins.w $a1, $a2, 31, 31
|
|
; LA32F-NEXT: ret
|
|
;
|
|
; LA32D-LABEL: fold_promote_d_s:
|
|
; LA32D: # %bb.0:
|
|
; LA32D-NEXT: fcvt.d.s $fa1, $fa1
|
|
; LA32D-NEXT: fcopysign.d $fa0, $fa0, $fa1
|
|
; LA32D-NEXT: ret
|
|
;
|
|
; LA64F-LABEL: fold_promote_d_s:
|
|
; LA64F: # %bb.0:
|
|
; LA64F-NEXT: movfr2gr.s $a1, $fa0
|
|
; LA64F-NEXT: lu12i.w $a2, -524288
|
|
; LA64F-NEXT: lu32i.d $a2, 0
|
|
; LA64F-NEXT: and $a1, $a1, $a2
|
|
; LA64F-NEXT: slli.d $a1, $a1, 32
|
|
; LA64F-NEXT: bstrins.d $a1, $a0, 62, 0
|
|
; LA64F-NEXT: move $a0, $a1
|
|
; LA64F-NEXT: ret
|
|
;
|
|
; LA64D-LABEL: fold_promote_d_s:
|
|
; LA64D: # %bb.0:
|
|
; LA64D-NEXT: fcvt.d.s $fa1, $fa1
|
|
; LA64D-NEXT: fcopysign.d $fa0, $fa0, $fa1
|
|
; LA64D-NEXT: ret
|
|
%c = fpext float %b to double
|
|
%t = call double @llvm.copysign.f64(double %a, double %c)
|
|
ret double %t
|
|
}
|
|
|
|
define float @fold_demote_s_d(float %a, double %b) nounwind {
|
|
; LA32F-LABEL: fold_demote_s_d:
|
|
; LA32F: # %bb.0:
|
|
; LA32F-NEXT: movgr2fr.w $fa1, $a1
|
|
; LA32F-NEXT: fcopysign.s $fa0, $fa0, $fa1
|
|
; LA32F-NEXT: ret
|
|
;
|
|
; LA32D-LABEL: fold_demote_s_d:
|
|
; LA32D: # %bb.0:
|
|
; LA32D-NEXT: fcvt.s.d $fa1, $fa1
|
|
; LA32D-NEXT: fcopysign.s $fa0, $fa0, $fa1
|
|
; LA32D-NEXT: ret
|
|
;
|
|
; LA64F-LABEL: fold_demote_s_d:
|
|
; LA64F: # %bb.0:
|
|
; LA64F-NEXT: srli.d $a0, $a0, 32
|
|
; LA64F-NEXT: movgr2fr.w $fa1, $a0
|
|
; LA64F-NEXT: fcopysign.s $fa0, $fa0, $fa1
|
|
; LA64F-NEXT: ret
|
|
;
|
|
; LA64D-LABEL: fold_demote_s_d:
|
|
; LA64D: # %bb.0:
|
|
; LA64D-NEXT: fcvt.s.d $fa1, $fa1
|
|
; LA64D-NEXT: fcopysign.s $fa0, $fa0, $fa1
|
|
; LA64D-NEXT: ret
|
|
%c = fptrunc double %b to float
|
|
%t = call float @llvm.copysign.f32(float %a, float %c)
|
|
ret float %t
|
|
}
|