[LoongArch][MC] Add relocation support for fld fst [x]vld [x]vst #133225

wangleiat · 2025-03-27T09:44:56Z

This also fixes errors when using Clang with step-by-step compilation.
Because the optimization will pass relocation information to memory
access instructions. For example:
t.c:

float f = 0.1;
float foo() { return f;}

clang --target=loongarch64 -O2 -c t.c --save-temps

Created using spr 1.3.5-bogner [skip ci]

Created using spr 1.3.5-bogner

llvmbot · 2025-03-27T09:45:36Z

@llvm/pr-subscribers-backend-loongarch

Author: wanglei (wangleiat)

Changes

This also fixes errors when using Clang with step-by-step compilation.
Because the optimization will pass relocation information to memory
access instructions. For example:
t.c:

float f = 0.1;
float foo() { return f;}

clang --target=loongarch64 -O2 -c t.c --save-temps

Full diff: https://github.com/llvm/llvm-project/pull/133225.diff

4 Files Affected:

(modified) llvm/lib/Target/LoongArch/LoongArchFloatInstrFormats.td (+2-2)
(modified) llvm/lib/Target/LoongArch/LoongArchLASXInstrInfo.td (+2-2)
(modified) llvm/lib/Target/LoongArch/LoongArchLSXInstrInfo.td (+2-2)
(modified) llvm/test/MC/LoongArch/Relocations/relocations.s (+18-13)

diff --git a/llvm/lib/Target/LoongArch/LoongArchFloatInstrFormats.td b/llvm/lib/Target/LoongArch/LoongArchFloatInstrFormats.td
index f66f620ca8b26..ce42236895c76 100644
--- a/llvm/lib/Target/LoongArch/LoongArchFloatInstrFormats.td
+++ b/llvm/lib/Target/LoongArch/LoongArchFloatInstrFormats.td
@@ -206,7 +206,7 @@ class FP_LOAD_3R<bits<32> op, RegisterClass rc = FPR32>
     : FPFmtMEM<op, (outs rc:$fd), (ins GPR:$rj, GPR:$rk),
                "$fd, $rj, $rk">;
 class FP_LOAD_2RI12<bits<32> op, RegisterClass rc = FPR32>
-    : FPFmt2RI12<op, (outs rc:$fd), (ins GPR:$rj, simm12:$imm12),
+    : FPFmt2RI12<op, (outs rc:$fd), (ins GPR:$rj, simm12_addlike:$imm12),
                  "$fd, $rj, $imm12">;
 } // hasSideEffects = 0, mayLoad = 1, mayStore = 0
 
@@ -215,7 +215,7 @@ class FP_STORE_3R<bits<32> op, RegisterClass rc = FPR32>
     : FPFmtMEM<op, (outs), (ins rc:$fd, GPR:$rj, GPR:$rk),
                "$fd, $rj, $rk">;
 class FP_STORE_2RI12<bits<32> op, RegisterClass rc = FPR32>
-    : FPFmt2RI12<op, (outs), (ins rc:$fd, GPR:$rj, simm12:$imm12),
+    : FPFmt2RI12<op, (outs), (ins rc:$fd, GPR:$rj, simm12_addlike:$imm12),
                  "$fd, $rj, $imm12">;
 } // hasSideEffects = 0, mayLoad = 0, mayStore = 1
 
diff --git a/llvm/lib/Target/LoongArch/LoongArchLASXInstrInfo.td b/llvm/lib/Target/LoongArch/LoongArchLASXInstrInfo.td
index 24b5ed5a9344f..7022fddf34100 100644
--- a/llvm/lib/Target/LoongArch/LoongArchLASXInstrInfo.td
+++ b/llvm/lib/Target/LoongArch/LoongArchLASXInstrInfo.td
@@ -186,10 +186,10 @@ class LASX2RI10_Load<bits<32> op, Operand ImmOpnd = simm10_lsl2>
 class LASX2RI11_Load<bits<32> op, Operand ImmOpnd = simm11_lsl1>
     : Fmt2RI11_XRI<op, (outs LASX256:$xd), (ins GPR:$rj, ImmOpnd:$imm11),
                   "$xd, $rj, $imm11">;
-class LASX2RI12_Load<bits<32> op, Operand ImmOpnd = simm12>
+class LASX2RI12_Load<bits<32> op, Operand ImmOpnd = simm12_addlike>
     : Fmt2RI12_XRI<op, (outs LASX256:$xd), (ins GPR:$rj, ImmOpnd:$imm12),
                   "$xd, $rj, $imm12">;
-class LASX2RI12_Store<bits<32> op, Operand ImmOpnd = simm12>
+class LASX2RI12_Store<bits<32> op, Operand ImmOpnd = simm12_addlike>
     : Fmt2RI12_XRI<op, (outs), (ins LASX256:$xd, GPR:$rj, ImmOpnd:$imm12),
                   "$xd, $rj, $imm12">;
 
diff --git a/llvm/lib/Target/LoongArch/LoongArchLSXInstrInfo.td b/llvm/lib/Target/LoongArch/LoongArchLSXInstrInfo.td
index d2063a8aaae9b..e37de4f545a2a 100644
--- a/llvm/lib/Target/LoongArch/LoongArchLSXInstrInfo.td
+++ b/llvm/lib/Target/LoongArch/LoongArchLSXInstrInfo.td
@@ -374,10 +374,10 @@ class LSX2RI10_Load<bits<32> op, Operand ImmOpnd = simm10_lsl2>
 class LSX2RI11_Load<bits<32> op, Operand ImmOpnd = simm11_lsl1>
     : Fmt2RI11_VRI<op, (outs LSX128:$vd), (ins GPR:$rj, ImmOpnd:$imm11),
                   "$vd, $rj, $imm11">;
-class LSX2RI12_Load<bits<32> op, Operand ImmOpnd = simm12>
+class LSX2RI12_Load<bits<32> op, Operand ImmOpnd = simm12_addlike>
     : Fmt2RI12_VRI<op, (outs LSX128:$vd), (ins GPR:$rj, ImmOpnd:$imm12),
                   "$vd, $rj, $imm12">;
-class LSX2RI12_Store<bits<32> op, Operand ImmOpnd = simm12>
+class LSX2RI12_Store<bits<32> op, Operand ImmOpnd = simm12_addlike>
     : Fmt2RI12_VRI<op, (outs), (ins LSX128:$vd, GPR:$rj, ImmOpnd:$imm12),
                   "$vd, $rj, $imm12">;
 
diff --git a/llvm/test/MC/LoongArch/Relocations/relocations.s b/llvm/test/MC/LoongArch/Relocations/relocations.s
index 180eeaa3100fa..f91a941295d9e 100644
--- a/llvm/test/MC/LoongArch/Relocations/relocations.s
+++ b/llvm/test/MC/LoongArch/Relocations/relocations.s
@@ -3,9 +3,6 @@
 # RUN: llvm-mc --filetype=obj --triple=loongarch64 < %s \
 # RUN:     | llvm-readobj -r - | FileCheck --check-prefix=RELOC %s
 
-# RUN: not llvm-mc --triple=loongarch64 --defsym=FIXME=1 < %s 2>&1 \
-# RUN:     | FileCheck --check-prefix=ERROR %s
-
 ## Check prefixes:
 ## RELOC - Check the relocation in the object.
 ## FIXUP - Check the fixup on the instruction.
@@ -312,24 +309,32 @@ pcaddi $t1, %desc_pcrel_20(foo)
 # INSTR: pcaddi $t1, %desc_pcrel_20(foo)
 # FIXUP: fixup A - offset: 0, value: %desc_pcrel_20(foo), kind: FK_NONE
 
-.ifdef FIXME
-
 fld.s $ft1, $a0, %pc_lo12(foo)
-# ERROR: :[[#@LINE-1]]:17: error: immediate must be an integer in the range [-2048, 2047]
+# RELOC: R_LARCH_PCALA_LO12 foo 0x0
+# INSTR: fld.s $ft1, $a0, %pc_lo12(foo)
+# FIXUP: fixup A - offset: 0, value: %pc_lo12(foo), kind: FK_NONE
 
 fst.d $ft1, $a0, %pc_lo12(foo)
-# ERROR: :[[#@LINE-1]]:17: error: immediate must be an integer in the range [-2048, 2047]
+# RELOC: R_LARCH_PCALA_LO12 foo 0x0
+# INSTR: fst.d $ft1, $a0, %pc_lo12(foo)
+# FIXUP: fixup A - offset: 0, value: %pc_lo12(foo), kind: FK_NONE
 
 vld $vr9, $a0, %pc_lo12(foo)
-# ERROR: :[[#@LINE-1]]:16: error: immediate must be an integer in the range [-2048, 2047]
+# RELOC: R_LARCH_PCALA_LO12 foo 0x0
+# INSTR: vld $vr9, $a0, %pc_lo12(foo)
+# FIXUP: fixup A - offset: 0, value: %pc_lo12(foo), kind: FK_NONE
 
 vst $vr9, $a0, %pc_lo12(foo)
-# ERROR: :[[#@LINE-1]]:16: error: immediate must be an integer in the range [-2048, 2047]
+# RELOC: R_LARCH_PCALA_LO12 foo 0x0
+# INSTR: vst $vr9, $a0, %pc_lo12(foo)
+# FIXUP: fixup A - offset: 0, value: %pc_lo12(foo), kind: FK_NONE
 
 xvld $xr9, $a0, %pc_lo12(foo)
-# ERROR: :[[#@LINE-1]]:17: error: immediate must be an integer in the range [-2048, 2047]
+# RELOC: R_LARCH_PCALA_LO12 foo 0x0
+# INSTR: xvld $xr9, $a0, %pc_lo12(foo)
+# FIXUP: fixup A - offset: 0, value: %pc_lo12(foo), kind: FK_NONE
 
 xvst $xr9, $a0, %pc_lo12(foo)
-# ERROR: :[[#@LINE-1]]:17: error: immediate must be an integer in the range [-2048, 2047]
-
-.endif
+# RELOC: R_LARCH_PCALA_LO12 foo 0x0
+# INSTR: xvst $xr9, $a0, %pc_lo12(foo)
+# FIXUP: fixup A - offset: 0, value: %pc_lo12(foo), kind: FK_NONE

llvmbot · 2025-03-27T09:45:37Z

@llvm/pr-subscribers-mc

Author: wanglei (wangleiat)

Changes

This also fixes errors when using Clang with step-by-step compilation.
Because the optimization will pass relocation information to memory
access instructions. For example:
t.c:

float f = 0.1;
float foo() { return f;}

clang --target=loongarch64 -O2 -c t.c --save-temps

Full diff: https://github.com/llvm/llvm-project/pull/133225.diff

4 Files Affected:

(modified) llvm/lib/Target/LoongArch/LoongArchFloatInstrFormats.td (+2-2)
(modified) llvm/lib/Target/LoongArch/LoongArchLASXInstrInfo.td (+2-2)
(modified) llvm/lib/Target/LoongArch/LoongArchLSXInstrInfo.td (+2-2)
(modified) llvm/test/MC/LoongArch/Relocations/relocations.s (+18-13)

diff --git a/llvm/lib/Target/LoongArch/LoongArchFloatInstrFormats.td b/llvm/lib/Target/LoongArch/LoongArchFloatInstrFormats.td
index f66f620ca8b26..ce42236895c76 100644
--- a/llvm/lib/Target/LoongArch/LoongArchFloatInstrFormats.td
+++ b/llvm/lib/Target/LoongArch/LoongArchFloatInstrFormats.td
@@ -206,7 +206,7 @@ class FP_LOAD_3R<bits<32> op, RegisterClass rc = FPR32>
     : FPFmtMEM<op, (outs rc:$fd), (ins GPR:$rj, GPR:$rk),
                "$fd, $rj, $rk">;
 class FP_LOAD_2RI12<bits<32> op, RegisterClass rc = FPR32>
-    : FPFmt2RI12<op, (outs rc:$fd), (ins GPR:$rj, simm12:$imm12),
+    : FPFmt2RI12<op, (outs rc:$fd), (ins GPR:$rj, simm12_addlike:$imm12),
                  "$fd, $rj, $imm12">;
 } // hasSideEffects = 0, mayLoad = 1, mayStore = 0
 
@@ -215,7 +215,7 @@ class FP_STORE_3R<bits<32> op, RegisterClass rc = FPR32>
     : FPFmtMEM<op, (outs), (ins rc:$fd, GPR:$rj, GPR:$rk),
                "$fd, $rj, $rk">;
 class FP_STORE_2RI12<bits<32> op, RegisterClass rc = FPR32>
-    : FPFmt2RI12<op, (outs), (ins rc:$fd, GPR:$rj, simm12:$imm12),
+    : FPFmt2RI12<op, (outs), (ins rc:$fd, GPR:$rj, simm12_addlike:$imm12),
                  "$fd, $rj, $imm12">;
 } // hasSideEffects = 0, mayLoad = 0, mayStore = 1
 
diff --git a/llvm/lib/Target/LoongArch/LoongArchLASXInstrInfo.td b/llvm/lib/Target/LoongArch/LoongArchLASXInstrInfo.td
index 24b5ed5a9344f..7022fddf34100 100644
--- a/llvm/lib/Target/LoongArch/LoongArchLASXInstrInfo.td
+++ b/llvm/lib/Target/LoongArch/LoongArchLASXInstrInfo.td
@@ -186,10 +186,10 @@ class LASX2RI10_Load<bits<32> op, Operand ImmOpnd = simm10_lsl2>
 class LASX2RI11_Load<bits<32> op, Operand ImmOpnd = simm11_lsl1>
     : Fmt2RI11_XRI<op, (outs LASX256:$xd), (ins GPR:$rj, ImmOpnd:$imm11),
                   "$xd, $rj, $imm11">;
-class LASX2RI12_Load<bits<32> op, Operand ImmOpnd = simm12>
+class LASX2RI12_Load<bits<32> op, Operand ImmOpnd = simm12_addlike>
     : Fmt2RI12_XRI<op, (outs LASX256:$xd), (ins GPR:$rj, ImmOpnd:$imm12),
                   "$xd, $rj, $imm12">;
-class LASX2RI12_Store<bits<32> op, Operand ImmOpnd = simm12>
+class LASX2RI12_Store<bits<32> op, Operand ImmOpnd = simm12_addlike>
     : Fmt2RI12_XRI<op, (outs), (ins LASX256:$xd, GPR:$rj, ImmOpnd:$imm12),
                   "$xd, $rj, $imm12">;
 
diff --git a/llvm/lib/Target/LoongArch/LoongArchLSXInstrInfo.td b/llvm/lib/Target/LoongArch/LoongArchLSXInstrInfo.td
index d2063a8aaae9b..e37de4f545a2a 100644
--- a/llvm/lib/Target/LoongArch/LoongArchLSXInstrInfo.td
+++ b/llvm/lib/Target/LoongArch/LoongArchLSXInstrInfo.td
@@ -374,10 +374,10 @@ class LSX2RI10_Load<bits<32> op, Operand ImmOpnd = simm10_lsl2>
 class LSX2RI11_Load<bits<32> op, Operand ImmOpnd = simm11_lsl1>
     : Fmt2RI11_VRI<op, (outs LSX128:$vd), (ins GPR:$rj, ImmOpnd:$imm11),
                   "$vd, $rj, $imm11">;
-class LSX2RI12_Load<bits<32> op, Operand ImmOpnd = simm12>
+class LSX2RI12_Load<bits<32> op, Operand ImmOpnd = simm12_addlike>
     : Fmt2RI12_VRI<op, (outs LSX128:$vd), (ins GPR:$rj, ImmOpnd:$imm12),
                   "$vd, $rj, $imm12">;
-class LSX2RI12_Store<bits<32> op, Operand ImmOpnd = simm12>
+class LSX2RI12_Store<bits<32> op, Operand ImmOpnd = simm12_addlike>
     : Fmt2RI12_VRI<op, (outs), (ins LSX128:$vd, GPR:$rj, ImmOpnd:$imm12),
                   "$vd, $rj, $imm12">;
 
diff --git a/llvm/test/MC/LoongArch/Relocations/relocations.s b/llvm/test/MC/LoongArch/Relocations/relocations.s
index 180eeaa3100fa..f91a941295d9e 100644
--- a/llvm/test/MC/LoongArch/Relocations/relocations.s
+++ b/llvm/test/MC/LoongArch/Relocations/relocations.s
@@ -3,9 +3,6 @@
 # RUN: llvm-mc --filetype=obj --triple=loongarch64 < %s \
 # RUN:     | llvm-readobj -r - | FileCheck --check-prefix=RELOC %s
 
-# RUN: not llvm-mc --triple=loongarch64 --defsym=FIXME=1 < %s 2>&1 \
-# RUN:     | FileCheck --check-prefix=ERROR %s
-
 ## Check prefixes:
 ## RELOC - Check the relocation in the object.
 ## FIXUP - Check the fixup on the instruction.
@@ -312,24 +309,32 @@ pcaddi $t1, %desc_pcrel_20(foo)
 # INSTR: pcaddi $t1, %desc_pcrel_20(foo)
 # FIXUP: fixup A - offset: 0, value: %desc_pcrel_20(foo), kind: FK_NONE
 
-.ifdef FIXME
-
 fld.s $ft1, $a0, %pc_lo12(foo)
-# ERROR: :[[#@LINE-1]]:17: error: immediate must be an integer in the range [-2048, 2047]
+# RELOC: R_LARCH_PCALA_LO12 foo 0x0
+# INSTR: fld.s $ft1, $a0, %pc_lo12(foo)
+# FIXUP: fixup A - offset: 0, value: %pc_lo12(foo), kind: FK_NONE
 
 fst.d $ft1, $a0, %pc_lo12(foo)
-# ERROR: :[[#@LINE-1]]:17: error: immediate must be an integer in the range [-2048, 2047]
+# RELOC: R_LARCH_PCALA_LO12 foo 0x0
+# INSTR: fst.d $ft1, $a0, %pc_lo12(foo)
+# FIXUP: fixup A - offset: 0, value: %pc_lo12(foo), kind: FK_NONE
 
 vld $vr9, $a0, %pc_lo12(foo)
-# ERROR: :[[#@LINE-1]]:16: error: immediate must be an integer in the range [-2048, 2047]
+# RELOC: R_LARCH_PCALA_LO12 foo 0x0
+# INSTR: vld $vr9, $a0, %pc_lo12(foo)
+# FIXUP: fixup A - offset: 0, value: %pc_lo12(foo), kind: FK_NONE
 
 vst $vr9, $a0, %pc_lo12(foo)
-# ERROR: :[[#@LINE-1]]:16: error: immediate must be an integer in the range [-2048, 2047]
+# RELOC: R_LARCH_PCALA_LO12 foo 0x0
+# INSTR: vst $vr9, $a0, %pc_lo12(foo)
+# FIXUP: fixup A - offset: 0, value: %pc_lo12(foo), kind: FK_NONE
 
 xvld $xr9, $a0, %pc_lo12(foo)
-# ERROR: :[[#@LINE-1]]:17: error: immediate must be an integer in the range [-2048, 2047]
+# RELOC: R_LARCH_PCALA_LO12 foo 0x0
+# INSTR: xvld $xr9, $a0, %pc_lo12(foo)
+# FIXUP: fixup A - offset: 0, value: %pc_lo12(foo), kind: FK_NONE
 
 xvst $xr9, $a0, %pc_lo12(foo)
-# ERROR: :[[#@LINE-1]]:17: error: immediate must be an integer in the range [-2048, 2047]
-
-.endif
+# RELOC: R_LARCH_PCALA_LO12 foo 0x0
+# INSTR: xvst $xr9, $a0, %pc_lo12(foo)
+# FIXUP: fixup A - offset: 0, value: %pc_lo12(foo), kind: FK_NONE

xen0n

Nice catch!

Created using spr 1.3.5-bogner [skip ci]

Created using spr 1.3.5-bogner

Reviewed By: SixWeining Pull Request: #133224

Created using spr 1.3.5-bogner [skip ci]

Created using spr 1.3.5-bogner

…[x]vst This also fixes errors when using Clang with step-by-step compilation. Because the optimization will pass relocation information to memory access instructions. For example: t.c: ``` float f = 0.1; float foo() { return f;} ``` ``` clang --target=loongarch64 -O2 -c t.c --save-temps ``` Reviewed By: tangaac, SixWeining Pull Request: llvm/llvm-project#133225

SixWeining · 2025-04-01T01:00:10Z

/cherry-pick 725a7b6 d055e58

llvmbot · 2025-04-01T01:06:11Z

/pull-request #133836

Reviewed By: SixWeining Pull Request: llvm#133224 (cherry picked from commit 725a7b6)

This also fixes errors when using Clang with step-by-step compilation. Because the optimization will pass relocation information to memory access instructions. For example: t.c: ``` float f = 0.1; float foo() { return f;} ``` ``` clang --target=loongarch64 -O2 -c t.c --save-temps ``` Reviewed By: tangaac, SixWeining Pull Request: llvm#133225 (cherry picked from commit d055e58)

wangleiat added 2 commits March 27, 2025 17:44

[𝘀𝗽𝗿] changes to main this commit is based on

13751a2

Created using spr 1.3.5-bogner [skip ci]

[𝘀𝗽𝗿] initial version

cfc5a04

Created using spr 1.3.5-bogner

llvmbot added mc Machine (object) code backend:loongarch labels Mar 27, 2025

wangleiat requested review from SixWeining and tangaac March 27, 2025 09:50

xen0n approved these changes Mar 27, 2025

View reviewed changes

wangleiat added 2 commits March 27, 2025 21:01

[𝘀𝗽𝗿] changes introduced through rebase

3532066

Created using spr 1.3.5-bogner [skip ci]

rebase

8f0391f

Created using spr 1.3.5-bogner

SixWeining approved these changes Mar 27, 2025

View reviewed changes

fix test

702b360

Created using spr 1.3.5-bogner

wangleiat added a commit that referenced this pull request Mar 28, 2025

[LoongArch] Pre-commit test for #133225

725a7b6

Reviewed By: SixWeining Pull Request: #133224

wangleiat added 2 commits March 28, 2025 10:22

[𝘀𝗽𝗿] changes introduced through rebase

dbf84fe

Created using spr 1.3.5-bogner [skip ci]

rebase

bd50bee

Created using spr 1.3.5-bogner

tangaac approved these changes Mar 28, 2025

View reviewed changes

wangleiat changed the base branch from users/wangleiat/spr/main.loongarchmc-add-relocation-support-for-fld-fst-xvld-xvst to main March 28, 2025 03:20

wangleiat merged commit d055e58 into main Mar 28, 2025
11 of 17 checks passed

wangleiat deleted the users/wangleiat/spr/loongarchmc-add-relocation-support-for-fld-fst-xvld-xvst branch March 28, 2025 03:20

SixWeining added this to the LLVM 20.X Release milestone Apr 1, 2025

github-project-automation bot added this to LLVM Release Status Apr 1, 2025

github-project-automation bot moved this to Needs Triage in LLVM Release Status Apr 1, 2025

llvmbot moved this from Needs Triage to Done in LLVM Release Status Apr 1, 2025

swift-ci pushed a commit to swiftlang/llvm-project that referenced this pull request Apr 1, 2025

[LoongArch] Pre-commit test for llvm#133225

ba00d9f

Reviewed By: SixWeining Pull Request: llvm#133224 (cherry picked from commit 725a7b6)

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[LoongArch][MC] Add relocation support for fld fst [x]vld [x]vst #133225

[LoongArch][MC] Add relocation support for fld fst [x]vld [x]vst #133225

wangleiat commented Mar 27, 2025

llvmbot commented Mar 27, 2025

llvmbot commented Mar 27, 2025

xen0n left a comment

SixWeining commented Apr 1, 2025

llvmbot commented Apr 1, 2025

[LoongArch][MC] Add relocation support for fld fst [x]vld [x]vst #133225

[LoongArch][MC] Add relocation support for fld fst [x]vld [x]vst #133225

Conversation

wangleiat commented Mar 27, 2025

llvmbot commented Mar 27, 2025

llvmbot commented Mar 27, 2025

xen0n left a comment

Choose a reason for hiding this comment

SixWeining commented Apr 1, 2025

llvmbot commented Apr 1, 2025