shylie/llvm-project
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
llvm-project/llvm/test/CodeGen/Xtensa/mul.ll
Andrei Safronov 4154ada1d4
[Xtensa] Implement Xtensa Floating Point Option. (#136086) 
Implement Xtensa FP Option instructions and lowering 
of the base FP operations with tests. Implement UR registers parsing.
 Fix loading from constant pool callee, basic block, globaladdress and
jumptable addresses. Also fixed potential memory leakage when several
similar XtensaConstantPoolValue objects are created Fix lowering i32 immediate.
2025-06-23 01:18:04 +03:00

1544 lines
50 KiB
LLVM
Raw Permalink Blame History

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
; RUN: llc -mtriple=xtensa -verify-machineinstrs < %s \
; RUN: | FileCheck -check-prefix=XTENSA %s
; RUN: llc -mtriple=xtensa -mattr=+mul32high,mul32 -verify-machineinstrs < %s \
; RUN: | FileCheck -check-prefix=XTENSA-MUL %s
define signext i32 @square(i32 %a) nounwind {
; XTENSA-LABEL: square:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: l32r a8, .LCPI0_0
; XTENSA-NEXT: or a3, a2, a2
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: square:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: mull a2, a2, a2
; XTENSA-MUL-NEXT: ret
%1 = mul i32 %a, %a
ret i32 %1
}
define signext i32 @mul(i32 %a, i32 %b) nounwind {
; XTENSA-LABEL: mul:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: l32r a8, .LCPI1_0
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: mul:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: mull a2, a2, a3
; XTENSA-MUL-NEXT: ret
%1 = mul i32 %a, %b
ret i32 %1
}
define signext i32 @mul_constant(i32 %a) nounwind {
; XTENSA-LABEL: mul_constant:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: movi a3, 5
; XTENSA-NEXT: l32r a8, .LCPI2_0
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: mul_constant:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: movi a8, 5
; XTENSA-MUL-NEXT: mull a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i32 %a, 5
ret i32 %1
}
define i32 @mul_pow2(i32 %a) nounwind {
; XTENSA-LABEL: mul_pow2:
; XTENSA: # %bb.0:
; XTENSA-NEXT: slli a2, a2, 3
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: mul_pow2:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: slli a2, a2, 3
; XTENSA-MUL-NEXT: ret
%1 = mul i32 %a, 8
ret i32 %1
}
define i64 @mul64(i64 %a, i64 %b) nounwind {
; XTENSA-LABEL: mul64:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: l32r a8, .LCPI4_0
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: mul64:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: mull a8, a2, a5
; XTENSA-MUL-NEXT: muluh a9, a2, a4
; XTENSA-MUL-NEXT: add a8, a9, a8
; XTENSA-MUL-NEXT: mull a9, a3, a4
; XTENSA-MUL-NEXT: add a3, a8, a9
; XTENSA-MUL-NEXT: mull a2, a2, a4
; XTENSA-MUL-NEXT: ret
%1 = mul i64 %a, %b
ret i64 %1
}
define i64 @mul64_constant(i64 %a) nounwind {
; XTENSA-LABEL: mul64_constant:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: movi a4, 5
; XTENSA-NEXT: movi a5, 0
; XTENSA-NEXT: l32r a8, .LCPI5_0
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: mul64_constant:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: movi a8, 5
; XTENSA-MUL-NEXT: mull a9, a3, a8
; XTENSA-MUL-NEXT: muluh a10, a2, a8
; XTENSA-MUL-NEXT: add a3, a10, a9
; XTENSA-MUL-NEXT: mull a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i64 %a, 5
ret i64 %1
}
define i32 @mulhs(i32 %a, i32 %b) nounwind {
; XTENSA-LABEL: mulhs:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: or a4, a3, a3
; XTENSA-NEXT: srai a3, a2, 31
; XTENSA-NEXT: srai a5, a4, 31
; XTENSA-NEXT: l32r a8, .LCPI6_0
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: or a2, a3, a3
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: mulhs:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: mulsh a2, a2, a3
; XTENSA-MUL-NEXT: ret
%1 = sext i32 %a to i64
%2 = sext i32 %b to i64
%3 = mul i64 %1, %2
%4 = lshr i64 %3, 32
%5 = trunc i64 %4 to i32
ret i32 %5
}
define i32 @mulhs_positive_constant(i32 %a) nounwind {
; XTENSA-LABEL: mulhs_positive_constant:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: srai a3, a2, 31
; XTENSA-NEXT: movi a4, 5
; XTENSA-NEXT: movi a5, 0
; XTENSA-NEXT: l32r a8, .LCPI7_0
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: or a2, a3, a3
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: mulhs_positive_constant:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: movi a8, 5
; XTENSA-MUL-NEXT: mulsh a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = sext i32 %a to i64
%2 = mul i64 %1, 5
%3 = lshr i64 %2, 32
%4 = trunc i64 %3 to i32
ret i32 %4
}
define i32 @mulhs_negative_constant(i32 %a) nounwind {
; XTENSA-LABEL: mulhs_negative_constant:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: srai a3, a2, 31
; XTENSA-NEXT: movi a4, -5
; XTENSA-NEXT: movi a5, -1
; XTENSA-NEXT: l32r a8, .LCPI8_0
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: or a2, a3, a3
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: mulhs_negative_constant:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: movi a8, -5
; XTENSA-MUL-NEXT: mulsh a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = sext i32 %a to i64
%2 = mul i64 %1, -5
%3 = lshr i64 %2, 32
%4 = trunc i64 %3 to i32
ret i32 %4
}
define zeroext i32 @mulhu(i32 zeroext %a, i32 zeroext %b) nounwind {
; XTENSA-LABEL: mulhu:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: or a4, a3, a3
; XTENSA-NEXT: movi a3, 0
; XTENSA-NEXT: l32r a8, .LCPI9_0
; XTENSA-NEXT: or a5, a3, a3
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: or a2, a3, a3
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: mulhu:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: muluh a2, a2, a3
; XTENSA-MUL-NEXT: ret
%1 = zext i32 %a to i64
%2 = zext i32 %b to i64
%3 = mul i64 %1, %2
%4 = lshr i64 %3, 32
%5 = trunc i64 %4 to i32
ret i32 %5
}
define i32 @mulhsu(i32 %a, i32 %b) nounwind {
; XTENSA-LABEL: mulhsu:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: or a4, a3, a3
; XTENSA-NEXT: srai a5, a4, 31
; XTENSA-NEXT: movi a3, 0
; XTENSA-NEXT: l32r a8, .LCPI10_0
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: or a2, a3, a3
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: mulhsu:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: muluh a8, a2, a3
; XTENSA-MUL-NEXT: srai a9, a3, 31
; XTENSA-MUL-NEXT: mull a9, a2, a9
; XTENSA-MUL-NEXT: add a2, a8, a9
; XTENSA-MUL-NEXT: ret
%1 = zext i32 %a to i64
%2 = sext i32 %b to i64
%3 = mul i64 %1, %2
%4 = lshr i64 %3, 32
%5 = trunc i64 %4 to i32
ret i32 %5
}
define i32 @mulhu_constant(i32 %a) nounwind {
; XTENSA-LABEL: mulhu_constant:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: movi a4, 5
; XTENSA-NEXT: movi a3, 0
; XTENSA-NEXT: l32r a8, .LCPI11_0
; XTENSA-NEXT: or a5, a3, a3
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: or a2, a3, a3
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: mulhu_constant:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: movi a8, 5
; XTENSA-MUL-NEXT: muluh a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = zext i32 %a to i64
%2 = mul i64 %1, 5
%3 = lshr i64 %2, 32
%4 = trunc i64 %3 to i32
ret i32 %4
}
define i32 @muli32_p65(i32 %a) nounwind {
; XTENSA-LABEL: muli32_p65:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: movi a3, 65
; XTENSA-NEXT: l32r a8, .LCPI12_0
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muli32_p65:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: movi a8, 65
; XTENSA-MUL-NEXT: mull a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i32 %a, 65
ret i32 %1
}
define i32 @muli32_p63(i32 %a) nounwind {
; XTENSA-LABEL: muli32_p63:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: movi a3, 63
; XTENSA-NEXT: l32r a8, .LCPI13_0
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muli32_p63:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: movi a8, 63
; XTENSA-MUL-NEXT: mull a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i32 %a, 63
ret i32 %1
}
define i64 @muli64_p65(i64 %a) nounwind {
; XTENSA-LABEL: muli64_p65:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: movi a4, 65
; XTENSA-NEXT: movi a5, 0
; XTENSA-NEXT: l32r a8, .LCPI14_0
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muli64_p65:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: movi a8, 65
; XTENSA-MUL-NEXT: mull a9, a3, a8
; XTENSA-MUL-NEXT: muluh a10, a2, a8
; XTENSA-MUL-NEXT: add a3, a10, a9
; XTENSA-MUL-NEXT: mull a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i64 %a, 65
ret i64 %1
}
define i64 @muli64_p63(i64 %a) nounwind {
; XTENSA-LABEL: muli64_p63:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: movi a4, 63
; XTENSA-NEXT: movi a5, 0
; XTENSA-NEXT: l32r a8, .LCPI15_0
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muli64_p63:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: movi a8, 63
; XTENSA-MUL-NEXT: mull a9, a3, a8
; XTENSA-MUL-NEXT: muluh a10, a2, a8
; XTENSA-MUL-NEXT: add a3, a10, a9
; XTENSA-MUL-NEXT: mull a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i64 %a, 63
ret i64 %1
}
define i32 @muli32_m63(i32 %a) nounwind {
; XTENSA-LABEL: muli32_m63:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: movi a3, -63
; XTENSA-NEXT: l32r a8, .LCPI16_0
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muli32_m63:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: movi a8, -63
; XTENSA-MUL-NEXT: mull a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i32 %a, -63
ret i32 %1
}
define i32 @muli32_m65(i32 %a) nounwind {
; XTENSA-LABEL: muli32_m65:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: movi a3, -65
; XTENSA-NEXT: l32r a8, .LCPI17_0
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muli32_m65:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: movi a8, -65
; XTENSA-MUL-NEXT: mull a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i32 %a, -65
ret i32 %1
}
define i64 @muli64_m63(i64 %a) nounwind {
; XTENSA-LABEL: muli64_m63:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: movi a4, -63
; XTENSA-NEXT: movi a5, -1
; XTENSA-NEXT: l32r a8, .LCPI18_0
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muli64_m63:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: movi a8, -63
; XTENSA-MUL-NEXT: mull a9, a3, a8
; XTENSA-MUL-NEXT: muluh a10, a2, a8
; XTENSA-MUL-NEXT: sub a10, a10, a2
; XTENSA-MUL-NEXT: add a3, a10, a9
; XTENSA-MUL-NEXT: mull a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i64 %a, -63
ret i64 %1
}
define i64 @muli64_m65(i64 %a) nounwind {
; XTENSA-LABEL: muli64_m65:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: movi a4, -65
; XTENSA-NEXT: movi a5, -1
; XTENSA-NEXT: l32r a8, .LCPI19_0
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muli64_m65:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: movi a8, -65
; XTENSA-MUL-NEXT: mull a9, a3, a8
; XTENSA-MUL-NEXT: muluh a10, a2, a8
; XTENSA-MUL-NEXT: sub a10, a10, a2
; XTENSA-MUL-NEXT: add a3, a10, a9
; XTENSA-MUL-NEXT: mull a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i64 %a, -65
ret i64 %1
}
define i32 @muli32_p384(i32 %a) nounwind {
; XTENSA-LABEL: muli32_p384:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: movi a3, 384
; XTENSA-NEXT: l32r a8, .LCPI20_0
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muli32_p384:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: movi a8, 384
; XTENSA-MUL-NEXT: mull a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i32 %a, 384
ret i32 %1
}
define i32 @muli32_p12288(i32 %a) nounwind {
; XTENSA-LABEL: muli32_p12288:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: l32r a3, .LCPI21_0
; XTENSA-NEXT: l32r a8, .LCPI21_1
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muli32_p12288:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: l32r a8, .LCPI21_0
; XTENSA-MUL-NEXT: mull a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i32 %a, 12288
ret i32 %1
}
define i32 @muli32_p4352(i32 %a) nounwind {
; XTENSA-LABEL: muli32_p4352:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: l32r a3, .LCPI22_0
; XTENSA-NEXT: l32r a8, .LCPI22_1
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muli32_p4352:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: l32r a8, .LCPI22_0
; XTENSA-MUL-NEXT: mull a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i32 %a, 4352
ret i32 %1
}
define i32 @muli32_p3840(i32 %a) nounwind {
; XTENSA-LABEL: muli32_p3840:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: l32r a3, .LCPI23_0
; XTENSA-NEXT: l32r a8, .LCPI23_1
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muli32_p3840:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: l32r a8, .LCPI23_0
; XTENSA-MUL-NEXT: mull a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i32 %a, 3840
ret i32 %1
}
define i32 @muli32_m3840(i32 %a) nounwind {
; XTENSA-LABEL: muli32_m3840:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: l32r a3, .LCPI24_0
; XTENSA-NEXT: l32r a8, .LCPI24_1
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muli32_m3840:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: l32r a8, .LCPI24_0
; XTENSA-MUL-NEXT: mull a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i32 %a, -3840
ret i32 %1
}
define i32 @muli32_m4352(i32 %a) nounwind {
; XTENSA-LABEL: muli32_m4352:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: l32r a3, .LCPI25_0
; XTENSA-NEXT: l32r a8, .LCPI25_1
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muli32_m4352:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: l32r a8, .LCPI25_0
; XTENSA-MUL-NEXT: mull a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i32 %a, -4352
ret i32 %1
}
define i64 @muli64_p4352(i64 %a) nounwind {
; XTENSA-LABEL: muli64_p4352:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: l32r a4, .LCPI26_0
; XTENSA-NEXT: movi a5, 0
; XTENSA-NEXT: l32r a8, .LCPI26_1
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muli64_p4352:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: l32r a8, .LCPI26_0
; XTENSA-MUL-NEXT: mull a9, a3, a8
; XTENSA-MUL-NEXT: muluh a10, a2, a8
; XTENSA-MUL-NEXT: add a3, a10, a9
; XTENSA-MUL-NEXT: mull a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i64 %a, 4352
ret i64 %1
}
define i64 @muli64_p3840(i64 %a) nounwind {
; XTENSA-LABEL: muli64_p3840:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: l32r a4, .LCPI27_0
; XTENSA-NEXT: movi a5, 0
; XTENSA-NEXT: l32r a8, .LCPI27_1
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muli64_p3840:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: l32r a8, .LCPI27_0
; XTENSA-MUL-NEXT: mull a9, a3, a8
; XTENSA-MUL-NEXT: muluh a10, a2, a8
; XTENSA-MUL-NEXT: add a3, a10, a9
; XTENSA-MUL-NEXT: mull a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i64 %a, 3840
ret i64 %1
}
define i64 @muli64_m4352(i64 %a) nounwind {
; XTENSA-LABEL: muli64_m4352:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: l32r a4, .LCPI28_0
; XTENSA-NEXT: movi a5, -1
; XTENSA-NEXT: l32r a8, .LCPI28_1
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muli64_m4352:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: l32r a8, .LCPI28_0
; XTENSA-MUL-NEXT: mull a9, a3, a8
; XTENSA-MUL-NEXT: muluh a10, a2, a8
; XTENSA-MUL-NEXT: sub a10, a10, a2
; XTENSA-MUL-NEXT: add a3, a10, a9
; XTENSA-MUL-NEXT: mull a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i64 %a, -4352
ret i64 %1
}
define i64 @muli64_m3840(i64 %a) nounwind {
; XTENSA-LABEL: muli64_m3840:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: l32r a4, .LCPI29_0
; XTENSA-NEXT: movi a5, -1
; XTENSA-NEXT: l32r a8, .LCPI29_1
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a0, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 16
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muli64_m3840:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: l32r a8, .LCPI29_0
; XTENSA-MUL-NEXT: mull a9, a3, a8
; XTENSA-MUL-NEXT: muluh a10, a2, a8
; XTENSA-MUL-NEXT: sub a10, a10, a2
; XTENSA-MUL-NEXT: add a3, a10, a9
; XTENSA-MUL-NEXT: mull a2, a2, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i64 %a, -3840
ret i64 %1
}
define i128 @muli128_m3840(i128 %a) nounwind {
; XTENSA-LABEL: muli128_m3840:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -64
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 60 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a12, a1, 56 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a13, a1, 52 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a14, a1, 48 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a15, a1, 44 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a5, a1, 20 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a4, a1, 16 # 4-byte Folded Spill
; XTENSA-NEXT: or a13, a3, a3
; XTENSA-NEXT: l32r a15, .LCPI30_0
; XTENSA-NEXT: movi a12, 0
; XTENSA-NEXT: l32r a8, .LCPI30_1
; XTENSA-NEXT: s32i a2, a1, 36 # 4-byte Folded Spill
; XTENSA-NEXT: or a3, a12, a12
; XTENSA-NEXT: or a4, a15, a15
; XTENSA-NEXT: or a5, a12, a12
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: s32i a2, a1, 28 # 4-byte Folded Spill
; XTENSA-NEXT: or a14, a3, a3
; XTENSA-NEXT: l32r a8, .LCPI30_2
; XTENSA-NEXT: s32i a13, a1, 40 # 4-byte Folded Spill
; XTENSA-NEXT: or a2, a13, a13
; XTENSA-NEXT: or a3, a12, a12
; XTENSA-NEXT: s32i a15, a1, 12 # 4-byte Folded Spill
; XTENSA-NEXT: or a4, a15, a15
; XTENSA-NEXT: or a5, a12, a12
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: add a13, a2, a14
; XTENSA-NEXT: movi a15, 1
; XTENSA-NEXT: or a8, a15, a15
; XTENSA-NEXT: bltu a13, a2, .LBB30_2
; XTENSA-NEXT: # %bb.1:
; XTENSA-NEXT: or a8, a12, a12
; XTENSA-NEXT: .LBB30_2:
; XTENSA-NEXT: add a8, a3, a8
; XTENSA-NEXT: s32i a8, a1, 32 # 4-byte Folded Spill
; XTENSA-NEXT: movi a14, -1
; XTENSA-NEXT: l32r a8, .LCPI30_3
; XTENSA-NEXT: l32i a2, a1, 36 # 4-byte Folded Reload
; XTENSA-NEXT: or a3, a12, a12
; XTENSA-NEXT: or a4, a14, a14
; XTENSA-NEXT: or a5, a12, a12
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: add a9, a2, a13
; XTENSA-NEXT: or a8, a15, a15
; XTENSA-NEXT: s32i a9, a1, 24 # 4-byte Folded Spill
; XTENSA-NEXT: bltu a9, a2, .LBB30_4
; XTENSA-NEXT: # %bb.3:
; XTENSA-NEXT: or a8, a12, a12
; XTENSA-NEXT: .LBB30_4:
; XTENSA-NEXT: add a8, a3, a8
; XTENSA-NEXT: l32i a9, a1, 32 # 4-byte Folded Reload
; XTENSA-NEXT: add a13, a9, a8
; XTENSA-NEXT: l32r a8, .LCPI30_4
; XTENSA-NEXT: l32i a2, a1, 40 # 4-byte Folded Reload
; XTENSA-NEXT: or a3, a12, a12
; XTENSA-NEXT: or a4, a14, a14
; XTENSA-NEXT: or a5, a12, a12
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: s32i a3, a1, 4 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a13, a1, 8 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a2, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: add a13, a2, a13
; XTENSA-NEXT: l32r a8, .LCPI30_5
; XTENSA-NEXT: l32i a2, a1, 16 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a3, a1, 20 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a4, a1, 12 # 4-byte Folded Reload
; XTENSA-NEXT: or a5, a14, a14
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: s32i a2, a1, 16 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a3, a1, 20 # 4-byte Folded Spill
; XTENSA-NEXT: l32r a8, .LCPI30_6
; XTENSA-NEXT: l32i a2, a1, 36 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a3, a1, 40 # 4-byte Folded Reload
; XTENSA-NEXT: or a4, a14, a14
; XTENSA-NEXT: or a5, a14, a14
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a8, a1, 16 # 4-byte Folded Reload
; XTENSA-NEXT: add a9, a2, a8
; XTENSA-NEXT: add a4, a13, a9
; XTENSA-NEXT: or a8, a15, a15
; XTENSA-NEXT: bltu a4, a13, .LBB30_6
; XTENSA-NEXT: # %bb.5:
; XTENSA-NEXT: or a8, a12, a12
; XTENSA-NEXT: .LBB30_6:
; XTENSA-NEXT: or a10, a15, a15
; XTENSA-NEXT: l32i a11, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: bltu a13, a11, .LBB30_8
; XTENSA-NEXT: # %bb.7:
; XTENSA-NEXT: or a10, a12, a12
; XTENSA-NEXT: .LBB30_8:
; XTENSA-NEXT: or a11, a15, a15
; XTENSA-NEXT: l32i a7, a1, 32 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a6, a1, 8 # 4-byte Folded Reload
; XTENSA-NEXT: bltu a6, a7, .LBB30_10
; XTENSA-NEXT: # %bb.9:
; XTENSA-NEXT: or a11, a12, a12
; XTENSA-NEXT: .LBB30_10:
; XTENSA-NEXT: l32i a7, a1, 4 # 4-byte Folded Reload
; XTENSA-NEXT: add a11, a7, a11
; XTENSA-NEXT: add a10, a11, a10
; XTENSA-NEXT: bltu a9, a2, .LBB30_12
; XTENSA-NEXT: # %bb.11:
; XTENSA-NEXT: or a15, a12, a12
; XTENSA-NEXT: .LBB30_12:
; XTENSA-NEXT: l32i a9, a1, 20 # 4-byte Folded Reload
; XTENSA-NEXT: add a9, a3, a9
; XTENSA-NEXT: add a9, a9, a15
; XTENSA-NEXT: add a9, a10, a9
; XTENSA-NEXT: add a5, a9, a8
; XTENSA-NEXT: l32i a2, a1, 28 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a3, a1, 24 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a15, a1, 44 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a14, a1, 48 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a13, a1, 52 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a12, a1, 56 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a0, a1, 60 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 64
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muli128_m3840:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: addi a8, a1, -48
; XTENSA-MUL-NEXT: or a1, a8, a8
; XTENSA-MUL-NEXT: s32i a12, a1, 32 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: s32i a13, a1, 28 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: s32i a14, a1, 24 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: s32i a15, a1, 20 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: s32i a5, a1, 16 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: l32r a10, .LCPI30_0
; XTENSA-MUL-NEXT: muluh a8, a2, a10
; XTENSA-MUL-NEXT: mull a7, a3, a10
; XTENSA-MUL-NEXT: add a8, a7, a8
; XTENSA-MUL-NEXT: movi a5, 0
; XTENSA-MUL-NEXT: movi a11, 1
; XTENSA-MUL-NEXT: or a9, a11, a11
; XTENSA-MUL-NEXT: bltu a8, a7, .LBB30_2
; XTENSA-MUL-NEXT: # %bb.1:
; XTENSA-MUL-NEXT: or a9, a5, a5
; XTENSA-MUL-NEXT: .LBB30_2:
; XTENSA-MUL-NEXT: muluh a7, a3, a10
; XTENSA-MUL-NEXT: add a15, a7, a9
; XTENSA-MUL-NEXT: sub a8, a8, a2
; XTENSA-MUL-NEXT: neg a7, a2
; XTENSA-MUL-NEXT: or a9, a11, a11
; XTENSA-MUL-NEXT: s32i a8, a1, 12 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: s32i a7, a1, 8 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: bltu a8, a7, .LBB30_4
; XTENSA-MUL-NEXT: # %bb.3:
; XTENSA-MUL-NEXT: or a9, a5, a5
; XTENSA-MUL-NEXT: .LBB30_4:
; XTENSA-MUL-NEXT: movi a8, -1
; XTENSA-MUL-NEXT: muluh a13, a2, a8
; XTENSA-MUL-NEXT: add a9, a13, a9
; XTENSA-MUL-NEXT: add a12, a15, a9
; XTENSA-MUL-NEXT: sub a6, a12, a3
; XTENSA-MUL-NEXT: mull a9, a4, a10
; XTENSA-MUL-NEXT: sub a14, a9, a2
; XTENSA-MUL-NEXT: add a9, a6, a14
; XTENSA-MUL-NEXT: or a7, a11, a11
; XTENSA-MUL-NEXT: s32i a9, a1, 4 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: bltu a9, a6, .LBB30_6
; XTENSA-MUL-NEXT: # %bb.5:
; XTENSA-MUL-NEXT: or a7, a5, a5
; XTENSA-MUL-NEXT: .LBB30_6:
; XTENSA-MUL-NEXT: s32i a7, a1, 0 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: neg a9, a3
; XTENSA-MUL-NEXT: or a7, a11, a11
; XTENSA-MUL-NEXT: bltu a6, a9, .LBB30_8
; XTENSA-MUL-NEXT: # %bb.7:
; XTENSA-MUL-NEXT: or a7, a5, a5
; XTENSA-MUL-NEXT: .LBB30_8:
; XTENSA-MUL-NEXT: or a6, a11, a11
; XTENSA-MUL-NEXT: bltu a12, a15, .LBB30_10
; XTENSA-MUL-NEXT: # %bb.9:
; XTENSA-MUL-NEXT: or a6, a5, a5
; XTENSA-MUL-NEXT: .LBB30_10:
; XTENSA-MUL-NEXT: muluh a8, a3, a8
; XTENSA-MUL-NEXT: add a8, a8, a6
; XTENSA-MUL-NEXT: add a8, a8, a7
; XTENSA-MUL-NEXT: l32i a9, a1, 16 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: mull a9, a9, a10
; XTENSA-MUL-NEXT: muluh a7, a4, a10
; XTENSA-MUL-NEXT: sub a7, a7, a4
; XTENSA-MUL-NEXT: add a9, a7, a9
; XTENSA-MUL-NEXT: sub a7, a13, a2
; XTENSA-MUL-NEXT: sub a7, a7, a3
; XTENSA-MUL-NEXT: add a7, a7, a9
; XTENSA-MUL-NEXT: l32i a9, a1, 8 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: bltu a14, a9, .LBB30_12
; XTENSA-MUL-NEXT: # %bb.11:
; XTENSA-MUL-NEXT: or a11, a5, a5
; XTENSA-MUL-NEXT: .LBB30_12:
; XTENSA-MUL-NEXT: add a9, a7, a11
; XTENSA-MUL-NEXT: add a8, a8, a9
; XTENSA-MUL-NEXT: l32i a9, a1, 0 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: add a5, a8, a9
; XTENSA-MUL-NEXT: mull a2, a2, a10
; XTENSA-MUL-NEXT: l32i a3, a1, 12 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: l32i a4, a1, 4 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: l32i a15, a1, 20 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: l32i a14, a1, 24 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: l32i a13, a1, 28 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: l32i a12, a1, 32 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: addi a8, a1, 48
; XTENSA-MUL-NEXT: or a1, a8, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i128 %a, -3840
ret i128 %1
}
define i128 @muli128_m63(i128 %a) nounwind {
; XTENSA-LABEL: muli128_m63:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -64
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 60 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a12, a1, 56 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a13, a1, 52 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a14, a1, 48 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a15, a1, 44 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a5, a1, 20 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a4, a1, 16 # 4-byte Folded Spill
; XTENSA-NEXT: or a13, a3, a3
; XTENSA-NEXT: movi a15, -63
; XTENSA-NEXT: movi a12, 0
; XTENSA-NEXT: l32r a8, .LCPI31_0
; XTENSA-NEXT: s32i a2, a1, 36 # 4-byte Folded Spill
; XTENSA-NEXT: or a3, a12, a12
; XTENSA-NEXT: or a4, a15, a15
; XTENSA-NEXT: or a5, a12, a12
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: s32i a2, a1, 28 # 4-byte Folded Spill
; XTENSA-NEXT: or a14, a3, a3
; XTENSA-NEXT: l32r a8, .LCPI31_1
; XTENSA-NEXT: s32i a13, a1, 40 # 4-byte Folded Spill
; XTENSA-NEXT: or a2, a13, a13
; XTENSA-NEXT: or a3, a12, a12
; XTENSA-NEXT: s32i a15, a1, 12 # 4-byte Folded Spill
; XTENSA-NEXT: or a4, a15, a15
; XTENSA-NEXT: or a5, a12, a12
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: add a13, a2, a14
; XTENSA-NEXT: movi a15, 1
; XTENSA-NEXT: or a8, a15, a15
; XTENSA-NEXT: bltu a13, a2, .LBB31_2
; XTENSA-NEXT: # %bb.1:
; XTENSA-NEXT: or a8, a12, a12
; XTENSA-NEXT: .LBB31_2:
; XTENSA-NEXT: add a8, a3, a8
; XTENSA-NEXT: s32i a8, a1, 32 # 4-byte Folded Spill
; XTENSA-NEXT: movi a14, -1
; XTENSA-NEXT: l32r a8, .LCPI31_2
; XTENSA-NEXT: l32i a2, a1, 36 # 4-byte Folded Reload
; XTENSA-NEXT: or a3, a12, a12
; XTENSA-NEXT: or a4, a14, a14
; XTENSA-NEXT: or a5, a12, a12
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: add a9, a2, a13
; XTENSA-NEXT: or a8, a15, a15
; XTENSA-NEXT: s32i a9, a1, 24 # 4-byte Folded Spill
; XTENSA-NEXT: bltu a9, a2, .LBB31_4
; XTENSA-NEXT: # %bb.3:
; XTENSA-NEXT: or a8, a12, a12
; XTENSA-NEXT: .LBB31_4:
; XTENSA-NEXT: add a8, a3, a8
; XTENSA-NEXT: l32i a9, a1, 32 # 4-byte Folded Reload
; XTENSA-NEXT: add a13, a9, a8
; XTENSA-NEXT: l32r a8, .LCPI31_3
; XTENSA-NEXT: l32i a2, a1, 40 # 4-byte Folded Reload
; XTENSA-NEXT: or a3, a12, a12
; XTENSA-NEXT: or a4, a14, a14
; XTENSA-NEXT: or a5, a12, a12
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: s32i a3, a1, 4 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a13, a1, 8 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a2, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: add a13, a2, a13
; XTENSA-NEXT: l32r a8, .LCPI31_4
; XTENSA-NEXT: l32i a2, a1, 16 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a3, a1, 20 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a4, a1, 12 # 4-byte Folded Reload
; XTENSA-NEXT: or a5, a14, a14
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: s32i a2, a1, 16 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a3, a1, 20 # 4-byte Folded Spill
; XTENSA-NEXT: l32r a8, .LCPI31_5
; XTENSA-NEXT: l32i a2, a1, 36 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a3, a1, 40 # 4-byte Folded Reload
; XTENSA-NEXT: or a4, a14, a14
; XTENSA-NEXT: or a5, a14, a14
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: l32i a8, a1, 16 # 4-byte Folded Reload
; XTENSA-NEXT: add a9, a2, a8
; XTENSA-NEXT: add a4, a13, a9
; XTENSA-NEXT: or a8, a15, a15
; XTENSA-NEXT: bltu a4, a13, .LBB31_6
; XTENSA-NEXT: # %bb.5:
; XTENSA-NEXT: or a8, a12, a12
; XTENSA-NEXT: .LBB31_6:
; XTENSA-NEXT: or a10, a15, a15
; XTENSA-NEXT: l32i a11, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: bltu a13, a11, .LBB31_8
; XTENSA-NEXT: # %bb.7:
; XTENSA-NEXT: or a10, a12, a12
; XTENSA-NEXT: .LBB31_8:
; XTENSA-NEXT: or a11, a15, a15
; XTENSA-NEXT: l32i a7, a1, 32 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a6, a1, 8 # 4-byte Folded Reload
; XTENSA-NEXT: bltu a6, a7, .LBB31_10
; XTENSA-NEXT: # %bb.9:
; XTENSA-NEXT: or a11, a12, a12
; XTENSA-NEXT: .LBB31_10:
; XTENSA-NEXT: l32i a7, a1, 4 # 4-byte Folded Reload
; XTENSA-NEXT: add a11, a7, a11
; XTENSA-NEXT: add a10, a11, a10
; XTENSA-NEXT: bltu a9, a2, .LBB31_12
; XTENSA-NEXT: # %bb.11:
; XTENSA-NEXT: or a15, a12, a12
; XTENSA-NEXT: .LBB31_12:
; XTENSA-NEXT: l32i a9, a1, 20 # 4-byte Folded Reload
; XTENSA-NEXT: add a9, a3, a9
; XTENSA-NEXT: add a9, a9, a15
; XTENSA-NEXT: add a9, a10, a9
; XTENSA-NEXT: add a5, a9, a8
; XTENSA-NEXT: l32i a2, a1, 28 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a3, a1, 24 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a15, a1, 44 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a14, a1, 48 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a13, a1, 52 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a12, a1, 56 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a0, a1, 60 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 64
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muli128_m63:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: addi a8, a1, -48
; XTENSA-MUL-NEXT: or a1, a8, a8
; XTENSA-MUL-NEXT: s32i a12, a1, 32 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: s32i a13, a1, 28 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: s32i a14, a1, 24 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: s32i a15, a1, 20 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: s32i a5, a1, 16 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: movi a10, -63
; XTENSA-MUL-NEXT: muluh a8, a2, a10
; XTENSA-MUL-NEXT: mull a7, a3, a10
; XTENSA-MUL-NEXT: add a8, a7, a8
; XTENSA-MUL-NEXT: movi a5, 0
; XTENSA-MUL-NEXT: movi a11, 1
; XTENSA-MUL-NEXT: or a9, a11, a11
; XTENSA-MUL-NEXT: bltu a8, a7, .LBB31_2
; XTENSA-MUL-NEXT: # %bb.1:
; XTENSA-MUL-NEXT: or a9, a5, a5
; XTENSA-MUL-NEXT: .LBB31_2:
; XTENSA-MUL-NEXT: muluh a7, a3, a10
; XTENSA-MUL-NEXT: add a15, a7, a9
; XTENSA-MUL-NEXT: sub a8, a8, a2
; XTENSA-MUL-NEXT: neg a7, a2
; XTENSA-MUL-NEXT: or a9, a11, a11
; XTENSA-MUL-NEXT: s32i a8, a1, 12 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: s32i a7, a1, 8 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: bltu a8, a7, .LBB31_4
; XTENSA-MUL-NEXT: # %bb.3:
; XTENSA-MUL-NEXT: or a9, a5, a5
; XTENSA-MUL-NEXT: .LBB31_4:
; XTENSA-MUL-NEXT: movi a8, -1
; XTENSA-MUL-NEXT: muluh a13, a2, a8
; XTENSA-MUL-NEXT: add a9, a13, a9
; XTENSA-MUL-NEXT: add a12, a15, a9
; XTENSA-MUL-NEXT: sub a6, a12, a3
; XTENSA-MUL-NEXT: mull a9, a4, a10
; XTENSA-MUL-NEXT: sub a14, a9, a2
; XTENSA-MUL-NEXT: add a9, a6, a14
; XTENSA-MUL-NEXT: or a7, a11, a11
; XTENSA-MUL-NEXT: s32i a9, a1, 4 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: bltu a9, a6, .LBB31_6
; XTENSA-MUL-NEXT: # %bb.5:
; XTENSA-MUL-NEXT: or a7, a5, a5
; XTENSA-MUL-NEXT: .LBB31_6:
; XTENSA-MUL-NEXT: s32i a7, a1, 0 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: neg a9, a3
; XTENSA-MUL-NEXT: or a7, a11, a11
; XTENSA-MUL-NEXT: bltu a6, a9, .LBB31_8
; XTENSA-MUL-NEXT: # %bb.7:
; XTENSA-MUL-NEXT: or a7, a5, a5
; XTENSA-MUL-NEXT: .LBB31_8:
; XTENSA-MUL-NEXT: or a6, a11, a11
; XTENSA-MUL-NEXT: bltu a12, a15, .LBB31_10
; XTENSA-MUL-NEXT: # %bb.9:
; XTENSA-MUL-NEXT: or a6, a5, a5
; XTENSA-MUL-NEXT: .LBB31_10:
; XTENSA-MUL-NEXT: muluh a8, a3, a8
; XTENSA-MUL-NEXT: add a8, a8, a6
; XTENSA-MUL-NEXT: add a8, a8, a7
; XTENSA-MUL-NEXT: l32i a9, a1, 16 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: mull a9, a9, a10
; XTENSA-MUL-NEXT: muluh a7, a4, a10
; XTENSA-MUL-NEXT: sub a7, a7, a4
; XTENSA-MUL-NEXT: add a9, a7, a9
; XTENSA-MUL-NEXT: sub a7, a13, a2
; XTENSA-MUL-NEXT: sub a7, a7, a3
; XTENSA-MUL-NEXT: add a7, a7, a9
; XTENSA-MUL-NEXT: l32i a9, a1, 8 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: bltu a14, a9, .LBB31_12
; XTENSA-MUL-NEXT: # %bb.11:
; XTENSA-MUL-NEXT: or a11, a5, a5
; XTENSA-MUL-NEXT: .LBB31_12:
; XTENSA-MUL-NEXT: add a9, a7, a11
; XTENSA-MUL-NEXT: add a8, a8, a9
; XTENSA-MUL-NEXT: l32i a9, a1, 0 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: add a5, a8, a9
; XTENSA-MUL-NEXT: mull a2, a2, a10
; XTENSA-MUL-NEXT: l32i a3, a1, 12 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: l32i a4, a1, 4 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: l32i a15, a1, 20 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: l32i a14, a1, 24 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: l32i a13, a1, 28 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: l32i a12, a1, 32 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: addi a8, a1, 48
; XTENSA-MUL-NEXT: or a1, a8, a8
; XTENSA-MUL-NEXT: ret
%1 = mul i128 %a, -63
ret i128 %1
}
define i64 @mulhsu_i64(i64 %a, i64 %b) nounwind {
; XTENSA-LABEL: mulhsu_i64:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addi a8, a1, -64
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: s32i a0, a1, 52 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a12, a1, 48 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a13, a1, 44 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a14, a1, 40 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a15, a1, 36 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a5, a1, 28 # 4-byte Folded Spill
; XTENSA-NEXT: or a13, a4, a4
; XTENSA-NEXT: or a15, a3, a3
; XTENSA-NEXT: movi a12, 0
; XTENSA-NEXT: l32r a8, .LCPI32_0
; XTENSA-NEXT: s32i a2, a1, 32 # 4-byte Folded Spill
; XTENSA-NEXT: or a3, a12, a12
; XTENSA-NEXT: or a5, a12, a12
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: or a14, a3, a3
; XTENSA-NEXT: l32r a8, .LCPI32_1
; XTENSA-NEXT: s32i a15, a1, 20 # 4-byte Folded Spill
; XTENSA-NEXT: or a2, a15, a15
; XTENSA-NEXT: or a3, a12, a12
; XTENSA-NEXT: s32i a13, a1, 16 # 4-byte Folded Spill
; XTENSA-NEXT: or a4, a13, a13
; XTENSA-NEXT: or a5, a12, a12
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: add a14, a2, a14
; XTENSA-NEXT: movi a15, 1
; XTENSA-NEXT: or a8, a15, a15
; XTENSA-NEXT: bltu a14, a2, .LBB32_2
; XTENSA-NEXT: # %bb.1:
; XTENSA-NEXT: or a8, a12, a12
; XTENSA-NEXT: .LBB32_2:
; XTENSA-NEXT: add a8, a3, a8
; XTENSA-NEXT: s32i a8, a1, 24 # 4-byte Folded Spill
; XTENSA-NEXT: l32r a8, .LCPI32_2
; XTENSA-NEXT: l32i a2, a1, 32 # 4-byte Folded Reload
; XTENSA-NEXT: or a3, a12, a12
; XTENSA-NEXT: l32i a4, a1, 28 # 4-byte Folded Reload
; XTENSA-NEXT: or a5, a12, a12
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: add a9, a2, a14
; XTENSA-NEXT: or a8, a15, a15
; XTENSA-NEXT: bltu a9, a2, .LBB32_4
; XTENSA-NEXT: # %bb.3:
; XTENSA-NEXT: or a8, a12, a12
; XTENSA-NEXT: .LBB32_4:
; XTENSA-NEXT: add a8, a3, a8
; XTENSA-NEXT: l32i a9, a1, 24 # 4-byte Folded Reload
; XTENSA-NEXT: add a14, a9, a8
; XTENSA-NEXT: l32r a8, .LCPI32_3
; XTENSA-NEXT: l32i a2, a1, 20 # 4-byte Folded Reload
; XTENSA-NEXT: or a3, a12, a12
; XTENSA-NEXT: l32i a13, a1, 28 # 4-byte Folded Reload
; XTENSA-NEXT: or a4, a13, a13
; XTENSA-NEXT: or a5, a12, a12
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: s32i a3, a1, 8 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a14, a1, 12 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a2, a1, 4 # 4-byte Folded Spill
; XTENSA-NEXT: add a14, a2, a14
; XTENSA-NEXT: l32r a8, .LCPI32_4
; XTENSA-NEXT: l32i a2, a1, 16 # 4-byte Folded Reload
; XTENSA-NEXT: or a3, a13, a13
; XTENSA-NEXT: or a4, a12, a12
; XTENSA-NEXT: or a5, a12, a12
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: s32i a2, a1, 0 # 4-byte Folded Spill
; XTENSA-NEXT: s32i a3, a1, 16 # 4-byte Folded Spill
; XTENSA-NEXT: srai a2, a13, 31
; XTENSA-NEXT: l32r a8, .LCPI32_5
; XTENSA-NEXT: or a3, a2, a2
; XTENSA-NEXT: l32i a4, a1, 32 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a5, a1, 20 # 4-byte Folded Reload
; XTENSA-NEXT: callx0 a8
; XTENSA-NEXT: or a8, a2, a2
; XTENSA-NEXT: l32i a9, a1, 0 # 4-byte Folded Reload
; XTENSA-NEXT: add a10, a8, a9
; XTENSA-NEXT: add a2, a14, a10
; XTENSA-NEXT: or a9, a15, a15
; XTENSA-NEXT: bltu a2, a14, .LBB32_6
; XTENSA-NEXT: # %bb.5:
; XTENSA-NEXT: or a9, a12, a12
; XTENSA-NEXT: .LBB32_6:
; XTENSA-NEXT: or a11, a15, a15
; XTENSA-NEXT: l32i a7, a1, 4 # 4-byte Folded Reload
; XTENSA-NEXT: bltu a14, a7, .LBB32_8
; XTENSA-NEXT: # %bb.7:
; XTENSA-NEXT: or a11, a12, a12
; XTENSA-NEXT: .LBB32_8:
; XTENSA-NEXT: or a7, a15, a15
; XTENSA-NEXT: l32i a6, a1, 24 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a5, a1, 12 # 4-byte Folded Reload
; XTENSA-NEXT: bltu a5, a6, .LBB32_10
; XTENSA-NEXT: # %bb.9:
; XTENSA-NEXT: or a7, a12, a12
; XTENSA-NEXT: .LBB32_10:
; XTENSA-NEXT: l32i a6, a1, 8 # 4-byte Folded Reload
; XTENSA-NEXT: add a7, a6, a7
; XTENSA-NEXT: add a11, a7, a11
; XTENSA-NEXT: bltu a10, a8, .LBB32_12
; XTENSA-NEXT: # %bb.11:
; XTENSA-NEXT: or a15, a12, a12
; XTENSA-NEXT: .LBB32_12:
; XTENSA-NEXT: l32i a8, a1, 16 # 4-byte Folded Reload
; XTENSA-NEXT: add a8, a3, a8
; XTENSA-NEXT: add a8, a8, a15
; XTENSA-NEXT: add a8, a11, a8
; XTENSA-NEXT: add a3, a8, a9
; XTENSA-NEXT: l32i a15, a1, 36 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a14, a1, 40 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a13, a1, 44 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a12, a1, 48 # 4-byte Folded Reload
; XTENSA-NEXT: l32i a0, a1, 52 # 4-byte Folded Reload
; XTENSA-NEXT: addi a8, a1, 64
; XTENSA-NEXT: or a1, a8, a8
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: mulhsu_i64:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: addi a8, a1, -16
; XTENSA-MUL-NEXT: or a1, a8, a8
; XTENSA-MUL-NEXT: s32i a12, a1, 12 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: s32i a13, a1, 8 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: s32i a14, a1, 4 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: s32i a15, a1, 0 # 4-byte Folded Spill
; XTENSA-MUL-NEXT: muluh a8, a2, a4
; XTENSA-MUL-NEXT: mull a7, a3, a4
; XTENSA-MUL-NEXT: add a8, a7, a8
; XTENSA-MUL-NEXT: movi a10, 0
; XTENSA-MUL-NEXT: movi a9, 1
; XTENSA-MUL-NEXT: or a11, a9, a9
; XTENSA-MUL-NEXT: bltu a8, a7, .LBB32_2
; XTENSA-MUL-NEXT: # %bb.1:
; XTENSA-MUL-NEXT: or a11, a10, a10
; XTENSA-MUL-NEXT: .LBB32_2:
; XTENSA-MUL-NEXT: muluh a7, a3, a4
; XTENSA-MUL-NEXT: add a11, a7, a11
; XTENSA-MUL-NEXT: mull a7, a2, a5
; XTENSA-MUL-NEXT: add a6, a7, a8
; XTENSA-MUL-NEXT: or a8, a9, a9
; XTENSA-MUL-NEXT: bltu a6, a7, .LBB32_4
; XTENSA-MUL-NEXT: # %bb.3:
; XTENSA-MUL-NEXT: or a8, a10, a10
; XTENSA-MUL-NEXT: .LBB32_4:
; XTENSA-MUL-NEXT: muluh a7, a2, a5
; XTENSA-MUL-NEXT: add a8, a7, a8
; XTENSA-MUL-NEXT: add a12, a11, a8
; XTENSA-MUL-NEXT: mull a14, a3, a5
; XTENSA-MUL-NEXT: add a15, a14, a12
; XTENSA-MUL-NEXT: srai a6, a5, 31
; XTENSA-MUL-NEXT: mull a7, a6, a2
; XTENSA-MUL-NEXT: add a8, a15, a7
; XTENSA-MUL-NEXT: or a4, a9, a9
; XTENSA-MUL-NEXT: bgeu a8, a15, .LBB32_9
; XTENSA-MUL-NEXT: # %bb.5:
; XTENSA-MUL-NEXT: or a13, a9, a9
; XTENSA-MUL-NEXT: bgeu a15, a14, .LBB32_10
; XTENSA-MUL-NEXT: .LBB32_6:
; XTENSA-MUL-NEXT: bltu a12, a11, .LBB32_8
; XTENSA-MUL-NEXT: .LBB32_7:
; XTENSA-MUL-NEXT: or a9, a10, a10
; XTENSA-MUL-NEXT: .LBB32_8:
; XTENSA-MUL-NEXT: muluh a10, a3, a5
; XTENSA-MUL-NEXT: add a9, a10, a9
; XTENSA-MUL-NEXT: add a9, a9, a13
; XTENSA-MUL-NEXT: mull a10, a6, a3
; XTENSA-MUL-NEXT: muluh a11, a6, a2
; XTENSA-MUL-NEXT: add a10, a11, a10
; XTENSA-MUL-NEXT: add a10, a10, a7
; XTENSA-MUL-NEXT: add a9, a9, a10
; XTENSA-MUL-NEXT: add a3, a9, a4
; XTENSA-MUL-NEXT: or a2, a8, a8
; XTENSA-MUL-NEXT: l32i a15, a1, 0 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: l32i a14, a1, 4 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: l32i a13, a1, 8 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: l32i a12, a1, 12 # 4-byte Folded Reload
; XTENSA-MUL-NEXT: addi a8, a1, 16
; XTENSA-MUL-NEXT: or a1, a8, a8
; XTENSA-MUL-NEXT: ret
; XTENSA-MUL-NEXT: .LBB32_9:
; XTENSA-MUL-NEXT: or a4, a10, a10
; XTENSA-MUL-NEXT: or a13, a9, a9
; XTENSA-MUL-NEXT: bltu a15, a14, .LBB32_6
; XTENSA-MUL-NEXT: .LBB32_10:
; XTENSA-MUL-NEXT: or a13, a10, a10
; XTENSA-MUL-NEXT: bgeu a12, a11, .LBB32_7
; XTENSA-MUL-NEXT: j .LBB32_8
%1 = zext i64 %a to i128
%2 = sext i64 %b to i128
%3 = mul i128 %1, %2
%4 = lshr i128 %3, 64
%5 = trunc i128 %4 to i64
ret i64 %5
}
define i8 @muladd_demand(i8 %x, i8 %y) nounwind {
; XTENSA-LABEL: muladd_demand:
; XTENSA: # %bb.0:
; XTENSA-NEXT: slli a8, a2, 1
; XTENSA-NEXT: sub a8, a3, a8
; XTENSA-NEXT: movi a9, 15
; XTENSA-NEXT: and a2, a8, a9
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muladd_demand:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: slli a8, a2, 1
; XTENSA-MUL-NEXT: sub a8, a3, a8
; XTENSA-MUL-NEXT: movi a9, 15
; XTENSA-MUL-NEXT: and a2, a8, a9
; XTENSA-MUL-NEXT: ret
%m = mul i8 %x, 14
%a = add i8 %y, %m
%r = and i8 %a, 15
ret i8 %r
}
define i8 @mulsub_demand(i8 %x, i8 %y) nounwind {
; XTENSA-LABEL: mulsub_demand:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addx2 a8, a2, a3
; XTENSA-NEXT: movi a9, 15
; XTENSA-NEXT: and a2, a8, a9
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: mulsub_demand:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: addx2 a8, a2, a3
; XTENSA-MUL-NEXT: movi a9, 15
; XTENSA-MUL-NEXT: and a2, a8, a9
; XTENSA-MUL-NEXT: ret
%m = mul i8 %x, 14
%a = sub i8 %y, %m
%r = and i8 %a, 15
ret i8 %r
}
define i8 @muladd_demand_2(i8 %x, i8 %y) nounwind {
; XTENSA-LABEL: muladd_demand_2:
; XTENSA: # %bb.0:
; XTENSA-NEXT: slli a8, a2, 1
; XTENSA-NEXT: sub a8, a3, a8
; XTENSA-NEXT: movi a9, -16
; XTENSA-NEXT: or a2, a8, a9
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: muladd_demand_2:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: slli a8, a2, 1
; XTENSA-MUL-NEXT: sub a8, a3, a8
; XTENSA-MUL-NEXT: movi a9, -16
; XTENSA-MUL-NEXT: or a2, a8, a9
; XTENSA-MUL-NEXT: ret
%m = mul i8 %x, 14
%a = add i8 %y, %m
%r = or i8 %a, 240
ret i8 %r
}
define i8 @mulsub_demand_2(i8 %x, i8 %y) nounwind {
; XTENSA-LABEL: mulsub_demand_2:
; XTENSA: # %bb.0:
; XTENSA-NEXT: addx2 a8, a2, a3
; XTENSA-NEXT: movi a9, -16
; XTENSA-NEXT: or a2, a8, a9
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: mulsub_demand_2:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: addx2 a8, a2, a3
; XTENSA-MUL-NEXT: movi a9, -16
; XTENSA-MUL-NEXT: or a2, a8, a9
; XTENSA-MUL-NEXT: ret
%m = mul i8 %x, 14
%a = sub i8 %y, %m
%r = or i8 %a, 240
ret i8 %r
}
define signext i32 @mul_imm_2(i32 %a) nounwind {
; XTENSA-LABEL: mul_imm_2:
; XTENSA: # %bb.0:
; XTENSA-NEXT: slli a2, a2, 1
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: mul_imm_2:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: slli a2, a2, 1
; XTENSA-MUL-NEXT: ret
%1 = mul i32 %a, 2
ret i32 %1
}
define signext i32 @mul_imm_1024(i32 %a) nounwind {
; XTENSA-LABEL: mul_imm_1024:
; XTENSA: # %bb.0:
; XTENSA-NEXT: slli a2, a2, 10
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: mul_imm_1024:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: slli a2, a2, 10
; XTENSA-MUL-NEXT: ret
%1 = mul i32 %a, 1024
ret i32 %1
}
define signext i32 @mul_imm_16384(i32 %a) nounwind {
; XTENSA-LABEL: mul_imm_16384:
; XTENSA: # %bb.0:
; XTENSA-NEXT: slli a2, a2, 14
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: mul_imm_16384:
; XTENSA-MUL: # %bb.0:
; XTENSA-MUL-NEXT: slli a2, a2, 14
; XTENSA-MUL-NEXT: ret
%1 = mul i32 %a, 16384
ret i32 %1
}
define <4 x i32> @mul_vec_splat_constant(<4 x i32> %a) {
; XTENSA-LABEL: mul_vec_splat_constant:
; XTENSA: .cfi_startproc
; XTENSA-NEXT: # %bb.0:
; XTENSA-NEXT: slli a2, a2, 2
; XTENSA-NEXT: slli a3, a3, 2
; XTENSA-NEXT: slli a4, a4, 2
; XTENSA-NEXT: slli a5, a5, 2
; XTENSA-NEXT: ret
;
; XTENSA-MUL-LABEL: mul_vec_splat_constant:
; XTENSA-MUL: .cfi_startproc
; XTENSA-MUL-NEXT: # %bb.0:
; XTENSA-MUL-NEXT: slli a2, a2, 2
; XTENSA-MUL-NEXT: slli a3, a3, 2
; XTENSA-MUL-NEXT: slli a4, a4, 2
; XTENSA-MUL-NEXT: slli a5, a5, 2
; XTENSA-MUL-NEXT: ret
%mul = mul <4 x i32> %a, <i32 4, i32 4, i32 4, i32 4>
ret <4 x i32> %mul
}
Reference in New Issue View Git Blame Copy Permalink