[DebugInfo][DWARF] Emit DW_AT_abstract_origin for concrete/inlined DW_TAG_lexical_blocks #136205
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
dzhidzhoev
requested review from
jmorse,
adrian-prantl,
SLTozer,
ellishg,
dwblaikie and
pogo59
|
@llvm/pr-subscribers-debuginfo Author: Vladislav Dzhidzhoev (dzhidzhoev) ChangesDuring the discussion under #119001, it was noticed that concrete DW_TAG_lexical_blocks should refer to corresponding abstract DW_TAG_lexical_blocks by having DW_AT_abstract_origin, to avoid ambiguity. This behavior is implemented in GCC (https://godbolt.org/z/Khrzdq1Wx), but not in LLVM. Fixes #49297. Full diff: https://github.com/llvm/llvm-project/pull/136205.diff 7 Files Affected:
diff --git a/llvm/lib/CodeGen/AsmPrinter/DwarfCompileUnit.cpp b/llvm/lib/CodeGen/AsmPrinter/DwarfCompileUnit.cpp
index 3939dae81841f..c57722216ea8d 100644
--- a/llvm/lib/CodeGen/AsmPrinter/DwarfCompileUnit.cpp
+++ b/llvm/lib/CodeGen/AsmPrinter/DwarfCompileUnit.cpp
@@ -782,6 +782,8 @@ DIE *DwarfCompileUnit::constructLexicalScopeDIE(LexicalScope *Scope) {
assert(!LexicalBlockDIEs.count(DS) &&
"Concrete out-of-line DIE for this scope exists!");
LexicalBlockDIEs[DS] = ScopeDIE;
+ } else {
+ InlinedLocalScopeDIEs[DS].push_back(ScopeDIE);
}
attachRangesOrLowHighPC(*ScopeDIE, Scope->getRanges());
@@ -1491,6 +1493,20 @@ void DwarfCompileUnit::finishEntityDefinition(const DbgEntity *Entity) {
getDwarfDebug().addAccelName(*this, CUNode->getNameTableKind(), Name, *Die);
}
+void DwarfCompileUnit::attachLexicalScopesAbstractOrigins(
+ LexicalScopes &LScopes) {
+ auto AttachAO = [&](const DILocalScope *LS, DIE *ScopeDIE) {
+ if (auto *AbsLSDie = getAbstractScopeDIEs().lookup(LS))
+ addDIEEntry(*ScopeDIE, dwarf::DW_AT_abstract_origin, *AbsLSDie);
+ };
+
+ for (auto [LScope, ScopeDIE] : LexicalBlockDIEs)
+ AttachAO(LScope, ScopeDIE);
+ for (auto &[LScope, ScopeDIEs] : InlinedLocalScopeDIEs)
+ for (auto *ScopeDIE : ScopeDIEs)
+ AttachAO(LScope, ScopeDIE);
+}
+
DbgEntity *DwarfCompileUnit::getExistingAbstractEntity(const DINode *Node) {
auto &AbstractEntities = getAbstractEntities();
auto I = AbstractEntities.find(Node);
diff --git a/llvm/lib/CodeGen/AsmPrinter/DwarfCompileUnit.h b/llvm/lib/CodeGen/AsmPrinter/DwarfCompileUnit.h
index 104039db03c7c..8a9493d9294cd 100644
--- a/llvm/lib/CodeGen/AsmPrinter/DwarfCompileUnit.h
+++ b/llvm/lib/CodeGen/AsmPrinter/DwarfCompileUnit.h
@@ -82,6 +82,10 @@ class DwarfCompileUnit final : public DwarfUnit {
// List of abstract local scopes (either DISubprogram or DILexicalBlock).
DenseMap<const DILocalScope *, DIE *> AbstractLocalScopeDIEs;
+ // List of inlined lexical block scopes that belong to subprograms within this
+ // CU.
+ DenseMap<const DILocalScope *, SmallVector<DIE *, 2>> InlinedLocalScopeDIEs;
+
DenseMap<const DINode *, std::unique_ptr<DbgEntity>> AbstractEntities;
/// DWO ID for correlating skeleton and split units.
@@ -299,6 +303,7 @@ class DwarfCompileUnit final : public DwarfUnit {
void finishSubprogramDefinition(const DISubprogram *SP);
void finishEntityDefinition(const DbgEntity *Entity);
+ void attachLexicalScopesAbstractOrigins(LexicalScopes &LScopes);
/// Find abstract variable associated with Var.
using InlinedEntity = DbgValueHistoryMap::InlinedEntity;
diff --git a/llvm/lib/CodeGen/AsmPrinter/DwarfDebug.cpp b/llvm/lib/CodeGen/AsmPrinter/DwarfDebug.cpp
index 39f1299a24e81..93f9184c71c7f 100644
--- a/llvm/lib/CodeGen/AsmPrinter/DwarfDebug.cpp
+++ b/llvm/lib/CodeGen/AsmPrinter/DwarfDebug.cpp
@@ -1262,6 +1262,7 @@ void DwarfDebug::finalizeModuleInfo() {
auto &TheCU = *P.second;
if (TheCU.getCUNode()->isDebugDirectivesOnly())
continue;
+ TheCU.attachLexicalScopesAbstractOrigins(LScopes);
// Emit DW_AT_containing_type attribute to connect types with their
// vtable holding type.
TheCU.constructContainingTypeDIEs();
diff --git a/llvm/test/DebugInfo/Generic/inline-scopes.ll b/llvm/test/DebugInfo/Generic/inline-scopes.ll
index 8e7543eb16e69..45ecdd0594f64 100644
--- a/llvm/test/DebugInfo/Generic/inline-scopes.ll
+++ b/llvm/test/DebugInfo/Generic/inline-scopes.ll
@@ -20,16 +20,29 @@
; }
; Ensure that lexical_blocks within inlined_subroutines are preserved/emitted.
+; CHECK: DW_TAG_subprogram
+; CHECK-NEXT: DW_AT_linkage_name ("_Z2f1v")
+; CHECK: [[ADDR1:0x[0-9a-f]+]]: DW_TAG_lexical_block
+; CHECK: DW_TAG_subprogram
+; CHECK-NEXT: DW_AT_linkage_name ("_Z2f2v")
+; CHECK: [[ADDR2:0x[0-9a-f]+]]: DW_TAG_lexical_block
; CHECK: DW_TAG_inlined_subroutine
; CHECK-NOT: DW_TAG
; CHECK-NOT: NULL
-; CHECK: DW_TAG_lexical_block
+; CHECK: DW_TAG_lexical_block
+; CHECK-NOT: {{DW_TAG|NULL}}
+; CHECK: DW_AT_abstract_origin ([[ADDR1]]
; CHECK-NOT: DW_TAG
; CHECK-NOT: NULL
; CHECK: DW_TAG_variable
; Ensure that file changes don't interfere with creating inlined subroutines.
; (see the line directive inside 'f2' in thesource)
; CHECK: DW_TAG_inlined_subroutine
+; CHECK-NOT: {{DW_TAG|NULL}}
+; CHECK: DW_TAG_lexical_block
+; CHECK-NOT: {{DW_TAG|NULL}}
+; CHECK: DW_AT_abstract_origin ([[ADDR2]]
+; CHECK-NOT: {{DW_TAG|NULL}}
; CHECK: DW_TAG_variable
; CHECK-NOT: DW_TAG
; CHECK: DW_AT_abstract_origin
diff --git a/llvm/test/DebugInfo/Generic/lexical-block-abstract-origin.ll b/llvm/test/DebugInfo/Generic/lexical-block-abstract-origin.ll
new file mode 100644
index 0000000000000..cbe2b818fba38
--- /dev/null
+++ b/llvm/test/DebugInfo/Generic/lexical-block-abstract-origin.ll
@@ -0,0 +1,56 @@
+; RUN: %llc_dwarf -filetype=obj -O0 %s -o - | llvm-dwarfdump -debug-info - | FileCheck %s
+
+; Check that DW_AT_abstract_origin is generated for concrete lexical block.
+
+; Generated from:
+; inline __attribute__((always_inline)) int foo(int x) {
+; {
+; int y = x + 5;
+; return y - 10;
+; }
+; }
+;
+; int bar(int x) {
+; int y = foo(7);
+; return y + 8;
+; }
+
+; CHECK: DW_TAG_subprogram
+; CHECK-NEXT: DW_AT_name ("foo")
+; CHECK-NOT: {{DW_TAG|NULL}}
+; CHECK: [[LB:.*]]: DW_TAG_lexical_block
+
+; CHECK: DW_TAG_inlined_subroutine
+; CHECK-NEXT: DW_AT_abstract_origin {{.*}} "foo"
+; CHECK-NOT: {{DW_TAG|NULL}}
+; CHECK: DW_TAG_lexical_block
+; CHECK-NOT: {{DW_TAG|NULL}}
+; CHECK: DW_AT_abstract_origin {{.*}}[[LB]]
+
+target datalayout = "e-m:o-p270:32:32-p271:32:32-p272:64:64-i64:64-i128:128-n32:64-S128-Fn32"
+target triple = "arm64-apple-macosx15.0.0"
+
+define i32 @bar() !dbg !9 {
+entry:
+ %y.i = alloca i32, align 4
+ #dbg_declare(ptr %y.i, !22, !DIExpression(), !24)
+ store i32 0, ptr %y.i, align 4, !dbg !24
+ %1 = load i32, ptr %y.i, align 4
+ ret i32 %1
+}
+
+!llvm.dbg.cu = !{!0}
+!llvm.module.flags = !{!3}
+
+!0 = distinct !DICompileUnit(language: DW_LANG_C11, file: !1, emissionKind: FullDebug)
+!1 = !DIFile(filename: "test.c", directory: "")
+!3 = !{i32 2, !"Debug Info Version", i32 3}
+!9 = distinct !DISubprogram(name: "bar", scope: !1, file: !1, line: 8, type: !10, spFlags: DISPFlagDefinition, unit: !0)
+!10 = !DISubroutineType(types: !13)
+!12 = !DIBasicType(name: "int", size: 32, encoding: DW_ATE_signed)
+!13 = !{}
+!19 = distinct !DISubprogram(name: "foo", scope: !1, file: !1, line: 1, type: !10, scopeLine: 1, spFlags: DISPFlagDefinition, unit: !0)
+!21 = distinct !DILocation(line: 9, column: 11, scope: !9)
+!22 = !DILocalVariable(name: "y", scope: !23, file: !1, line: 3, type: !12)
+!23 = distinct !DILexicalBlock(scope: !19, file: !1, line: 2, column: 3)
+!24 = !DILocation(line: 3, column: 9, scope: !23, inlinedAt: !21)
diff --git a/llvm/test/DebugInfo/X86/lexical-block-file-inline.ll b/llvm/test/DebugInfo/X86/lexical-block-file-inline.ll
index 15d0785bde93d..43c5e7ccef8c6 100644
--- a/llvm/test/DebugInfo/X86/lexical-block-file-inline.ll
+++ b/llvm/test/DebugInfo/X86/lexical-block-file-inline.ll
@@ -28,9 +28,11 @@
; CHECK: DW_TAG_subprogram
; CHECK-NOT: {{DW_TAG|NULL}}
; CHECK: DW_AT_abstract_origin {{.*}} {[[Offset_bar:0x[0-9abcdef]+]]}
-; CHECK-NOT: {{DW_TAG|NULL}}
+; CHECK-NOT: {{DW_TAG|NULL}}
; CHECK: DW_TAG_lexical_block
-; CHECK-NOT: {{DW_TAG|NULL}}
+; CHECK-NOT: {{DW_TAG|NULL}}
+; CHECK: DW_AT_abstract_origin {{.*}}[[Offset_lb:0x[0-9a-f]+]]
+; CHECK-NOT: {{DW_TAG|NULL}}
; CHECK: DW_TAG_variable
;; Abstract "bar" function
@@ -40,7 +42,7 @@
; CHECK-NOT: {{DW_TAG|NULL}}
; CHECK: DW_AT_inline
; CHECK-NOT: {{DW_TAG|NULL}}
-; CHECK: DW_TAG_lexical_block
+; CHECK: [[Offset_lb]]: DW_TAG_lexical_block
; CHECK-NOT: {{DW_TAG|NULL}}
; CHECK: DW_TAG_variable
; CHECK-NOT: {{DW_TAG|NULL}}
@@ -56,8 +58,10 @@
; CHECK-NEXT: DW_AT_abstract_origin {{.*}} {[[Offset_bar]]}
; CHECK-NOT: {{DW_TAG|NULL}}
; CHECK: DW_TAG_lexical_block
-; CHECK-NOT: {{DW_TAG|NULL}}
-; CHECK: DW_TAG_variable
+; CHECK-NOT: {{DW_TAG|NULL}}
+; CHECK: DW_AT_abstract_origin {{.*}}[[Offset_lb]]
+; CHECK-NOT: {{DW_TAG|NULL}}
+; CHECK: DW_TAG_variable
; Function Attrs: alwaysinline nounwind
define i32 @_Z3barv() #0 !dbg !4 {
diff --git a/llvm/test/DebugInfo/X86/missing-abstract-variable.ll b/llvm/test/DebugInfo/X86/missing-abstract-variable.ll
index 572dca2c2cb6c..21363805aff2f 100644
--- a/llvm/test/DebugInfo/X86/missing-abstract-variable.ll
+++ b/llvm/test/DebugInfo/X86/missing-abstract-variable.ll
@@ -37,7 +37,7 @@
; CHECK-NOT: DW_TAG
; CHECK: DW_AT_name ("b")
; CHECK-NOT: {{DW_TAG|NULL}}
-; CHECK: DW_TAG_lexical_block
+; CHECK: [[LB_DECL:.*]]: DW_TAG_lexical_block
; CHECK-NOT: {{DW_TAG|NULL}}
; CHECK: DW_TAG_variable
; CHECK-NOT: DW_TAG
@@ -82,7 +82,9 @@
; CHECK-NOT: {{DW_TAG|NULL}}
; CHECK: DW_TAG_lexical_block
-; CHECK-NOT: {{DW_TAG|NULL}}
+; CHECK-NOT: {{DW_TAG|NULL}}
+; CHECK: DW_AT_abstract_origin {{.*}}[[LB_DECL]]
+; CHECK-NOT: {{DW_TAG|NULL}}
; CHECK: DW_TAG_variable
; CHECK-NOT: DW_TAG
; CHECK: DW_AT_abstract_origin {{.*}} "s"
|
…_TAG_lexical_blocks During the discussion under llvm#119001, it was noticed, that concrete DW_TAG_lexical_blocks should refer to corresponding abstract DW_TAG_lexical_blocks to avoid ambiguity. This behavior is implemented in GCC (https://godbolt.org/z/Khrzdq1Wx), but not in LLVM so far. Fixes llvm#49297.
dzhidzhoev
force-pushed
the
debuginfo-rfc-krisb/lb-abstract-origin/ls
branch
from
cb3f994 to
34da121
Compare
|
I think this is the right direction, and I'm not sure what we'll do about it if we have to do anything, but... could you measure the size impact of this change? (use Bloaty's comparison mode to compare, say, a Release built clang built with a compiler with/without this change) |
Sure! |
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/144/builds/23533 Here is the relevant piece of the build log for the reference |
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/27/builds/9117 Here is the relevant piece of the build log for the reference |
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/46/builds/15718 Here is the relevant piece of the build log for the reference |
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/133/builds/15049 Here is the relevant piece of the build log for the reference |
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/140/builds/21898 Here is the relevant piece of the build log for the reference |
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/154/builds/15279 Here is the relevant piece of the build log for the reference |
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/139/builds/14030 Here is the relevant piece of the build log for the reference |
dwblaikie
added a commit
that referenced
this pull request
Apr 24, 2025
…lined DW_TAG_lexical_blocks" (#137237) Reverts #136205 Breaks buildbots, probably something about needing to restrict the test to running on a specific target or the like - I haven't looked closely. Co-authored-by: Vladislav Dzhidzhoev <[email protected]>
|
Oops, I have (apparently) fixed these failures in 92dc18b. |
llvm-sync bot
pushed a commit
to arm/arm-toolchain
that referenced
this pull request
May 6, 2025
…concrete/inlined DW_TAG_lexical_blocks" (#137237) Reverts llvm/llvm-project#136205 Breaks buildbots, probably something about needing to restrict the test to running on a specific target or the like - I haven't looked closely. Co-authored-by: Vladislav Dzhidzhoev <[email protected]>
IanWood1
pushed a commit
to IanWood1/llvm-project
that referenced
this pull request
May 6, 2025
…_TAG_lexical_blocks (llvm#136205) During the discussion under llvm#119001, it was noticed that concrete DW_TAG_lexical_blocks should refer to corresponding abstract DW_TAG_lexical_blocks by having DW_AT_abstract_origin, to avoid ambiguity. This behavior is implemented in GCC (https://godbolt.org/z/Khrzdq1Wx), but not in LLVM. Fixes llvm#49297.
IanWood1
pushed a commit
to IanWood1/llvm-project
that referenced
this pull request
May 6, 2025
…lined DW_TAG_lexical_blocks" (llvm#137237) Reverts llvm#136205 Breaks buildbots, probably something about needing to restrict the test to running on a specific target or the like - I haven't looked closely. Co-authored-by: Vladislav Dzhidzhoev <[email protected]>
IanWood1
pushed a commit
to IanWood1/llvm-project
that referenced
this pull request
May 6, 2025
…_TAG_lexical_blocks (llvm#136205) During the discussion under llvm#119001, it was noticed that concrete DW_TAG_lexical_blocks should refer to corresponding abstract DW_TAG_lexical_blocks by having DW_AT_abstract_origin, to avoid ambiguity. This behavior is implemented in GCC (https://godbolt.org/z/Khrzdq1Wx), but not in LLVM. Fixes llvm#49297.
IanWood1
pushed a commit
to IanWood1/llvm-project
that referenced
this pull request
May 6, 2025
…lined DW_TAG_lexical_blocks" (llvm#137237) Reverts llvm#136205 Breaks buildbots, probably something about needing to restrict the test to running on a specific target or the like - I haven't looked closely. Co-authored-by: Vladislav Dzhidzhoev <[email protected]>
IanWood1
pushed a commit
to IanWood1/llvm-project
that referenced
this pull request
May 6, 2025
…_TAG_lexical_blocks (llvm#136205) During the discussion under llvm#119001, it was noticed that concrete DW_TAG_lexical_blocks should refer to corresponding abstract DW_TAG_lexical_blocks by having DW_AT_abstract_origin, to avoid ambiguity. This behavior is implemented in GCC (https://godbolt.org/z/Khrzdq1Wx), but not in LLVM. Fixes llvm#49297.
IanWood1
pushed a commit
to IanWood1/llvm-project
that referenced
this pull request
May 6, 2025
…lined DW_TAG_lexical_blocks" (llvm#137237) Reverts llvm#136205 Breaks buildbots, probably something about needing to restrict the test to running on a specific target or the like - I haven't looked closely. Co-authored-by: Vladislav Dzhidzhoev <[email protected]>
Ankur-0429
pushed a commit
to Ankur-0429/llvm-project
that referenced
this pull request
May 9, 2025
…_TAG_lexical_blocks (llvm#136205) During the discussion under llvm#119001, it was noticed that concrete DW_TAG_lexical_blocks should refer to corresponding abstract DW_TAG_lexical_blocks by having DW_AT_abstract_origin, to avoid ambiguity. This behavior is implemented in GCC (https://godbolt.org/z/Khrzdq1Wx), but not in LLVM. Fixes llvm#49297.
Ankur-0429
pushed a commit
to Ankur-0429/llvm-project
that referenced
this pull request
May 9, 2025
…lined DW_TAG_lexical_blocks" (llvm#137237) Reverts llvm#136205 Breaks buildbots, probably something about needing to restrict the test to running on a specific target or the like - I haven't looked closely. Co-authored-by: Vladislav Dzhidzhoev <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
During the discussion under #119001, it was noticed that concrete DW_TAG_lexical_blocks should refer to corresponding abstract DW_TAG_lexical_blocks by having DW_AT_abstract_origin, to avoid ambiguity. This behavior is implemented in GCC (https://godbolt.org/z/Khrzdq1Wx), but not in LLVM.
Fixes #49297.