mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-11-26 12:43:36 +01:00
0c16dcd701
This provides an optimized implementation of SADDO/SSUBO/UADDO/USUBO as well as ADDCARRY/SUBCARRY on top of the new CC implementation. In particular, multi-word arithmetic now uses UADDO/ADDCARRY instead of the old ADDC/ADDE logic, which means we no longer need to use "glue" links for those instructions. This also allows making full use of the memory-based instructions like ALSI, which couldn't be recognized due to limitations in the DAG matcher previously. Also, the llvm.sadd.with.overflow et.al. intrinsincs now expand to directly using the ADD instructions and checking for a CC 3 result. llvm-svn: 331203
117 lines
3.2 KiB
LLVM
117 lines
3.2 KiB
LLVM
; Test 64-bit subtraction in which the second operand is constant.
|
|
;
|
|
; RUN: llc < %s -mtriple=s390x-linux-gnu | FileCheck %s
|
|
|
|
declare i64 @foo()
|
|
|
|
; Check addition of 1.
|
|
define zeroext i1 @f1(i64 %dummy, i64 %a, i64 *%res) {
|
|
; CHECK-LABEL: f1:
|
|
; CHECK: slgfi %r3, 1
|
|
; CHECK-DAG: stg %r3, 0(%r4)
|
|
; CHECK-DAG: ipm [[REG:%r[0-5]]]
|
|
; CHECK-DAG: afi [[REG]], -536870912
|
|
; CHECK-DAG: risbg %r2, [[REG]], 63, 191, 33
|
|
; CHECK: br %r14
|
|
%t = call {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 1)
|
|
%val = extractvalue {i64, i1} %t, 0
|
|
%obit = extractvalue {i64, i1} %t, 1
|
|
store i64 %val, i64 *%res
|
|
ret i1 %obit
|
|
}
|
|
|
|
; Check the high end of the SLGFI range.
|
|
define zeroext i1 @f2(i64 %dummy, i64 %a, i64 *%res) {
|
|
; CHECK-LABEL: f2:
|
|
; CHECK: slgfi %r3, 4294967295
|
|
; CHECK-DAG: stg %r3, 0(%r4)
|
|
; CHECK-DAG: ipm [[REG:%r[0-5]]]
|
|
; CHECK-DAG: afi [[REG]], -536870912
|
|
; CHECK-DAG: risbg %r2, [[REG]], 63, 191, 33
|
|
; CHECK: br %r14
|
|
%t = call {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 4294967295)
|
|
%val = extractvalue {i64, i1} %t, 0
|
|
%obit = extractvalue {i64, i1} %t, 1
|
|
store i64 %val, i64 *%res
|
|
ret i1 %obit
|
|
}
|
|
|
|
; Check the next value up, which must be loaded into a register first.
|
|
define zeroext i1 @f3(i64 %dummy, i64 %a, i64 *%res) {
|
|
; CHECK-LABEL: f3:
|
|
; CHECK: llihl [[REG1:%r[0-9]+]], 1
|
|
; CHECK: slgr %r3, [[REG1]]
|
|
; CHECK-DAG: stg %r3, 0(%r4)
|
|
; CHECK-DAG: ipm [[REG2:%r[0-5]]]
|
|
; CHECK-DAG: afi [[REG]], -536870912
|
|
; CHECK-DAG: risbg %r2, [[REG2]], 63, 191, 33
|
|
; CHECK: br %r14
|
|
%t = call {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 4294967296)
|
|
%val = extractvalue {i64, i1} %t, 0
|
|
%obit = extractvalue {i64, i1} %t, 1
|
|
store i64 %val, i64 *%res
|
|
ret i1 %obit
|
|
}
|
|
|
|
; Likewise for negative values.
|
|
define zeroext i1 @f4(i64 %dummy, i64 %a, i64 *%res) {
|
|
; CHECK-LABEL: f4:
|
|
; CHECK: lghi [[REG1:%r[0-9]+]], -1
|
|
; CHECK: slgr %r3, [[REG1]]
|
|
; CHECK-DAG: stg %r3, 0(%r4)
|
|
; CHECK-DAG: ipm [[REG2:%r[0-5]]]
|
|
; CHECK-DAG: afi [[REG]], -536870912
|
|
; CHECK-DAG: risbg %r2, [[REG2]], 63, 191, 33
|
|
; CHECK: br %r14
|
|
%t = call {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 -1)
|
|
%val = extractvalue {i64, i1} %t, 0
|
|
%obit = extractvalue {i64, i1} %t, 1
|
|
store i64 %val, i64 *%res
|
|
ret i1 %obit
|
|
}
|
|
|
|
; Check using the overflow result for a branch.
|
|
define void @f5(i64 %dummy, i64 %a, i64 *%res) {
|
|
; CHECK-LABEL: f5:
|
|
; CHECK: slgfi %r3, 1
|
|
; CHECK: stg %r3, 0(%r4)
|
|
; CHECK: jgle foo@PLT
|
|
; CHECK: br %r14
|
|
%t = call {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 1)
|
|
%val = extractvalue {i64, i1} %t, 0
|
|
%obit = extractvalue {i64, i1} %t, 1
|
|
store i64 %val, i64 *%res
|
|
br i1 %obit, label %call, label %exit
|
|
|
|
call:
|
|
tail call i64 @foo()
|
|
br label %exit
|
|
|
|
exit:
|
|
ret void
|
|
}
|
|
|
|
; ... and the same with the inverted direction.
|
|
define void @f6(i64 %dummy, i64 %a, i64 *%res) {
|
|
; CHECK-LABEL: f6:
|
|
; CHECK: slgfi %r3, 1
|
|
; CHECK: stg %r3, 0(%r4)
|
|
; CHECK: jgnle foo@PLT
|
|
; CHECK: br %r14
|
|
%t = call {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 1)
|
|
%val = extractvalue {i64, i1} %t, 0
|
|
%obit = extractvalue {i64, i1} %t, 1
|
|
store i64 %val, i64 *%res
|
|
br i1 %obit, label %exit, label %call
|
|
|
|
call:
|
|
tail call i64 @foo()
|
|
br label %exit
|
|
|
|
exit:
|
|
ret void
|
|
}
|
|
|
|
declare {i64, i1} @llvm.usub.with.overflow.i64(i64, i64) nounwind readnone
|
|
|