Skip to content

[X86] Set x87 fld1/fldz pseudo instructions as rematerializable #74592

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Dec 6, 2023
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion llvm/lib/Target/X86/X86InstrFPStack.td
Original file line number Diff line number Diff line change
Expand Up @@ -524,7 +524,7 @@ def XCH_F : FPI<0xD9, MRM1r, (outs), (ins RSTi:$op), "fxch\t$op">;
}

// Floating point constant loads.
let SchedRW = [WriteZero], Uses = [FPCW] in {
let SchedRW = [WriteZero], Uses = [FPCW], isReMaterializable = 1 in {
def LD_Fp032 : FpIf32<(outs RFP32:$dst), (ins), ZeroArgFP,
[(set RFP32:$dst, fpimm0)]>;
def LD_Fp132 : FpIf32<(outs RFP32:$dst), (ins), ZeroArgFP,
Expand Down
6 changes: 6 additions & 0 deletions llvm/lib/Target/X86/X86InstrInfo.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -782,6 +782,12 @@ bool X86InstrInfo::isReallyTriviallyReMaterializable(
break;

case X86::LOAD_STACK_GUARD:
case X86::LD_Fp032:
case X86::LD_Fp064:
case X86::LD_Fp080:
case X86::LD_Fp132:
case X86::LD_Fp164:
case X86::LD_Fp180:
case X86::AVX1_SETALLONES:
case X86::AVX2_SETALLONES:
case X86::AVX512_128_SET0:
Expand Down
10 changes: 3 additions & 7 deletions llvm/test/CodeGen/X86/swifterror.ll
Original file line number Diff line number Diff line change
Expand Up @@ -243,8 +243,6 @@ define float @caller2(ptr %error_ref) {
; CHECK-i386-NEXT: .cfi_offset %edi, -8
; CHECK-i386-NEXT: movl 32(%esp), %esi
; CHECK-i386-NEXT: leal 16(%esp), %edi
; CHECK-i386-NEXT: fld1
; CHECK-i386-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) ## 4-byte Folded Spill
; CHECK-i386-NEXT: LBB2_1: ## %bb_loop
; CHECK-i386-NEXT: ## =>This Inner Loop Header: Depth=1
; CHECK-i386-NEXT: movl $0, 16(%esp)
Expand All @@ -255,7 +253,7 @@ define float @caller2(ptr %error_ref) {
; CHECK-i386-NEXT: jne LBB2_4
; CHECK-i386-NEXT: ## %bb.2: ## %cont
; CHECK-i386-NEXT: ## in Loop: Header=BB2_1 Depth=1
; CHECK-i386-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) ## 4-byte Folded Reload
; CHECK-i386-NEXT: fld1
; CHECK-i386-NEXT: fxch %st(1)
; CHECK-i386-NEXT: fucompp
; CHECK-i386-NEXT: fnstsw %ax
Expand All @@ -270,7 +268,7 @@ define float @caller2(ptr %error_ref) {
; CHECK-i386-NEXT: fstp %st(0)
; CHECK-i386-NEXT: movl %ecx, (%esp)
; CHECK-i386-NEXT: calll _free
; CHECK-i386-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) ## 4-byte Folded Reload
; CHECK-i386-NEXT: fld1
; CHECK-i386-NEXT: addl $20, %esp
; CHECK-i386-NEXT: popl %esi
; CHECK-i386-NEXT: popl %edi
Expand Down Expand Up @@ -470,8 +468,6 @@ define float @foo_loop(ptr swifterror %error_ptr_ref, i32 %cc, float %cc2) {
; CHECK-i386-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) ## 4-byte Folded Spill
; CHECK-i386-NEXT: movl 36(%esp), %esi
; CHECK-i386-NEXT: movl 32(%esp), %edi
; CHECK-i386-NEXT: fld1
; CHECK-i386-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) ## 4-byte Folded Spill
; CHECK-i386-NEXT: LBB4_1: ## %bb_loop
; CHECK-i386-NEXT: ## =>This Inner Loop Header: Depth=1
; CHECK-i386-NEXT: testl %esi, %esi
Expand All @@ -486,7 +482,7 @@ define float @foo_loop(ptr swifterror %error_ptr_ref, i32 %cc, float %cc2) {
; CHECK-i386-NEXT: LBB4_3: ## %bb_cont
; CHECK-i386-NEXT: ## in Loop: Header=BB4_1 Depth=1
; CHECK-i386-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) ## 4-byte Folded Reload
; CHECK-i386-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) ## 4-byte Folded Reload
; CHECK-i386-NEXT: fld1
; CHECK-i386-NEXT: fxch %st(1)
; CHECK-i386-NEXT: fucompp
; CHECK-i386-NEXT: fnstsw %ax
Expand Down