[lld/COFF] Fix -start-lib / -end-lib more after reviews.llvm.org/D116434 #124294
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
This is a follow-up to llvm#120452 in a way. Since lld/COFF does not yet insert all defined in an obj file before all undefineds (ELF and MachO do this, see llvm#67445 and things linked from there), it's possible that: 1. We add an obj file a.obj 2. a.obj contains an undefined that's in b.obj, causing b.obj to be added 3. b.obj contains an undefined that's in a part of a.obj that's not yet in the symbol table, causing a recursive load of a.obj, which adds the symbols in there twice, leading to duplicate symbol errors. For normal archives, `ArchiveFile::addMember()` has a `seen` check to prevent this. For start-lib lazy objects, we can just check if the archive is still lazy at the recursive call. This bug is similar to issue llvm#59162. (Eventually, we'll probably want to do what the MachO and ELF ports do.) Includes a test that caused duplicate symbol diagnostics before this code change.
|
@llvm/pr-subscribers-lld-coff Author: Nico Weber (nico) ChangesThis is a follow-up to #120452 in a way. Since lld/COFF does not yet insert all defined in an obj file before all undefineds (ELF and MachO do this, see #67445 and things linked from there), it's possible that:
For normal archives, This bug is similar to issue #59162. (Eventually, we'll probably want to do what the MachO and ELF ports do.) Includes a test that caused duplicate symbol diagnostics before this code change. Full diff: https://github.com/llvm/llvm-project/pull/124294.diff 2 Files Affected:
diff --git a/lld/COFF/SymbolTable.cpp b/lld/COFF/SymbolTable.cpp
index 32ea4a5b2e1fc3..4ade347875fde3 100644
--- a/lld/COFF/SymbolTable.cpp
+++ b/lld/COFF/SymbolTable.cpp
@@ -56,6 +56,8 @@ static void forceLazy(Symbol *s) {
}
case Symbol::Kind::LazyObjectKind: {
InputFile *file = cast<LazyObject>(s)->file;
+ if (!file->lazy)
+ return;
file->lazy = false;
file->symtab.ctx.driver.addFile(file);
break;
diff --git a/lld/test/COFF/start-lib.ll b/lld/test/COFF/start-lib.ll
index a46147f21ccbb3..134cdc2a6e1df6 100644
--- a/lld/test/COFF/start-lib.ll
+++ b/lld/test/COFF/start-lib.ll
@@ -173,3 +173,72 @@ target triple = "x86_64-pc-windows-msvc"
define void @baz() {
ret void
}
+
+
+; Check cycles between symbols in two /start-lib files.
+; If the links succeed and does not emit duplicate symbol diagnostics,
+; that's enough.
+
+; RUN: llc -filetype=obj %t.dir/main3.ll -o %t-main3.obj
+; RUN: llc -filetype=obj %t.dir/cycle1.ll -o %t-cycle1.obj
+; RUN: llc -filetype=obj %t.dir/cycle2.ll -o %t-cycle2.obj
+; RUN: opt -thinlto-bc %t.dir/main3.ll -o %t-main3.bc
+; RUN: opt -thinlto-bc %t.dir/cycle1.ll -o %t-cycle1.bc
+; RUN: opt -thinlto-bc %t.dir/cycle2.ll -o %t-cycle2.bc
+
+; RUN: lld-link -out:%t3.exe -entry:main \
+; RUN: %t-main3.obj %t-cycle1.obj %t-cycle2.obj
+; RUN: lld-link -out:%t3.exe -entry:main \
+; RUN: %t-main3.obj /start-lib %t-cycle1.obj %t-cycle2.obj /end-lib
+; RUN: lld-link -out:%t3.exe -entry:main \
+; RUN: /start-lib %t-cycle1.obj %t-cycle2.obj /end-lib %t-main3.obj
+
+; RUN: lld-link -out:%t3.exe -entry:main \
+; RUN: %t-main3.bc %t-cycle1.bc %t-cycle2.bc
+; RUN: lld-link -out:%t3.exe -entry:main \
+; RUN: %t-main3.bc /start-lib %t-cycle1.bc %t-cycle2.bc /end-lib
+; RUN: lld-link -out:%t3.exe -entry:main \
+; RUN: /start-lib %t-cycle1.bc %t-cycle2.bc /end-lib %t-main3.bc
+
+#--- main3.ll
+
+target datalayout = "e-m:w-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-pc-windows-msvc"
+
+declare void @foo1()
+
+define void @main() {
+ call void () @foo1()
+ ret void
+}
+
+#--- cycle1.ll
+
+target datalayout = "e-m:w-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-pc-windows-msvc"
+
+declare void @bar()
+
+define void @foo1() {
+ ; cycle1.ll pulls in cycle2.ll for bar(), and cycle2.ll then pulls in
+ ; cycle1.ll again for foo2().
+ call void () @bar()
+ ret void
+}
+
+define void @foo2() {
+ ret void
+}
+
+
+#--- cycle2.ll
+
+target datalayout = "e-m:w-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-pc-windows-msvc"
+
+declare void @foo2()
+
+define void @bar() {
+ call void () @foo2()
+ ret void
+}
|
|
@llvm/pr-subscribers-lld Author: Nico Weber (nico) ChangesThis is a follow-up to #120452 in a way. Since lld/COFF does not yet insert all defined in an obj file before all undefineds (ELF and MachO do this, see #67445 and things linked from there), it's possible that:
For normal archives, This bug is similar to issue #59162. (Eventually, we'll probably want to do what the MachO and ELF ports do.) Includes a test that caused duplicate symbol diagnostics before this code change. Full diff: https://github.com/llvm/llvm-project/pull/124294.diff 2 Files Affected:
diff --git a/lld/COFF/SymbolTable.cpp b/lld/COFF/SymbolTable.cpp
index 32ea4a5b2e1fc3..4ade347875fde3 100644
--- a/lld/COFF/SymbolTable.cpp
+++ b/lld/COFF/SymbolTable.cpp
@@ -56,6 +56,8 @@ static void forceLazy(Symbol *s) {
}
case Symbol::Kind::LazyObjectKind: {
InputFile *file = cast<LazyObject>(s)->file;
+ if (!file->lazy)
+ return;
file->lazy = false;
file->symtab.ctx.driver.addFile(file);
break;
diff --git a/lld/test/COFF/start-lib.ll b/lld/test/COFF/start-lib.ll
index a46147f21ccbb3..134cdc2a6e1df6 100644
--- a/lld/test/COFF/start-lib.ll
+++ b/lld/test/COFF/start-lib.ll
@@ -173,3 +173,72 @@ target triple = "x86_64-pc-windows-msvc"
define void @baz() {
ret void
}
+
+
+; Check cycles between symbols in two /start-lib files.
+; If the links succeed and does not emit duplicate symbol diagnostics,
+; that's enough.
+
+; RUN: llc -filetype=obj %t.dir/main3.ll -o %t-main3.obj
+; RUN: llc -filetype=obj %t.dir/cycle1.ll -o %t-cycle1.obj
+; RUN: llc -filetype=obj %t.dir/cycle2.ll -o %t-cycle2.obj
+; RUN: opt -thinlto-bc %t.dir/main3.ll -o %t-main3.bc
+; RUN: opt -thinlto-bc %t.dir/cycle1.ll -o %t-cycle1.bc
+; RUN: opt -thinlto-bc %t.dir/cycle2.ll -o %t-cycle2.bc
+
+; RUN: lld-link -out:%t3.exe -entry:main \
+; RUN: %t-main3.obj %t-cycle1.obj %t-cycle2.obj
+; RUN: lld-link -out:%t3.exe -entry:main \
+; RUN: %t-main3.obj /start-lib %t-cycle1.obj %t-cycle2.obj /end-lib
+; RUN: lld-link -out:%t3.exe -entry:main \
+; RUN: /start-lib %t-cycle1.obj %t-cycle2.obj /end-lib %t-main3.obj
+
+; RUN: lld-link -out:%t3.exe -entry:main \
+; RUN: %t-main3.bc %t-cycle1.bc %t-cycle2.bc
+; RUN: lld-link -out:%t3.exe -entry:main \
+; RUN: %t-main3.bc /start-lib %t-cycle1.bc %t-cycle2.bc /end-lib
+; RUN: lld-link -out:%t3.exe -entry:main \
+; RUN: /start-lib %t-cycle1.bc %t-cycle2.bc /end-lib %t-main3.bc
+
+#--- main3.ll
+
+target datalayout = "e-m:w-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-pc-windows-msvc"
+
+declare void @foo1()
+
+define void @main() {
+ call void () @foo1()
+ ret void
+}
+
+#--- cycle1.ll
+
+target datalayout = "e-m:w-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-pc-windows-msvc"
+
+declare void @bar()
+
+define void @foo1() {
+ ; cycle1.ll pulls in cycle2.ll for bar(), and cycle2.ll then pulls in
+ ; cycle1.ll again for foo2().
+ call void () @bar()
+ ret void
+}
+
+define void @foo2() {
+ ret void
+}
+
+
+#--- cycle2.ll
+
+target datalayout = "e-m:w-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-pc-windows-msvc"
+
+declare void @foo2()
+
+define void @bar() {
+ call void () @foo2()
+ ret void
+}
|
|
@llvm/pr-subscribers-platform-windows Author: Nico Weber (nico) ChangesThis is a follow-up to #120452 in a way. Since lld/COFF does not yet insert all defined in an obj file before all undefineds (ELF and MachO do this, see #67445 and things linked from there), it's possible that:
For normal archives, This bug is similar to issue #59162. (Eventually, we'll probably want to do what the MachO and ELF ports do.) Includes a test that caused duplicate symbol diagnostics before this code change. Full diff: https://github.com/llvm/llvm-project/pull/124294.diff 2 Files Affected:
diff --git a/lld/COFF/SymbolTable.cpp b/lld/COFF/SymbolTable.cpp
index 32ea4a5b2e1fc3..4ade347875fde3 100644
--- a/lld/COFF/SymbolTable.cpp
+++ b/lld/COFF/SymbolTable.cpp
@@ -56,6 +56,8 @@ static void forceLazy(Symbol *s) {
}
case Symbol::Kind::LazyObjectKind: {
InputFile *file = cast<LazyObject>(s)->file;
+ if (!file->lazy)
+ return;
file->lazy = false;
file->symtab.ctx.driver.addFile(file);
break;
diff --git a/lld/test/COFF/start-lib.ll b/lld/test/COFF/start-lib.ll
index a46147f21ccbb3..134cdc2a6e1df6 100644
--- a/lld/test/COFF/start-lib.ll
+++ b/lld/test/COFF/start-lib.ll
@@ -173,3 +173,72 @@ target triple = "x86_64-pc-windows-msvc"
define void @baz() {
ret void
}
+
+
+; Check cycles between symbols in two /start-lib files.
+; If the links succeed and does not emit duplicate symbol diagnostics,
+; that's enough.
+
+; RUN: llc -filetype=obj %t.dir/main3.ll -o %t-main3.obj
+; RUN: llc -filetype=obj %t.dir/cycle1.ll -o %t-cycle1.obj
+; RUN: llc -filetype=obj %t.dir/cycle2.ll -o %t-cycle2.obj
+; RUN: opt -thinlto-bc %t.dir/main3.ll -o %t-main3.bc
+; RUN: opt -thinlto-bc %t.dir/cycle1.ll -o %t-cycle1.bc
+; RUN: opt -thinlto-bc %t.dir/cycle2.ll -o %t-cycle2.bc
+
+; RUN: lld-link -out:%t3.exe -entry:main \
+; RUN: %t-main3.obj %t-cycle1.obj %t-cycle2.obj
+; RUN: lld-link -out:%t3.exe -entry:main \
+; RUN: %t-main3.obj /start-lib %t-cycle1.obj %t-cycle2.obj /end-lib
+; RUN: lld-link -out:%t3.exe -entry:main \
+; RUN: /start-lib %t-cycle1.obj %t-cycle2.obj /end-lib %t-main3.obj
+
+; RUN: lld-link -out:%t3.exe -entry:main \
+; RUN: %t-main3.bc %t-cycle1.bc %t-cycle2.bc
+; RUN: lld-link -out:%t3.exe -entry:main \
+; RUN: %t-main3.bc /start-lib %t-cycle1.bc %t-cycle2.bc /end-lib
+; RUN: lld-link -out:%t3.exe -entry:main \
+; RUN: /start-lib %t-cycle1.bc %t-cycle2.bc /end-lib %t-main3.bc
+
+#--- main3.ll
+
+target datalayout = "e-m:w-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-pc-windows-msvc"
+
+declare void @foo1()
+
+define void @main() {
+ call void () @foo1()
+ ret void
+}
+
+#--- cycle1.ll
+
+target datalayout = "e-m:w-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-pc-windows-msvc"
+
+declare void @bar()
+
+define void @foo1() {
+ ; cycle1.ll pulls in cycle2.ll for bar(), and cycle2.ll then pulls in
+ ; cycle1.ll again for foo2().
+ call void () @bar()
+ ret void
+}
+
+define void @foo2() {
+ ret void
+}
+
+
+#--- cycle2.ll
+
+target datalayout = "e-m:w-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-pc-windows-msvc"
+
+declare void @foo2()
+
+define void @bar() {
+ call void () @foo2()
+ ret void
+}
|
MaskRay
reviewed
Jan 24, 2025
MaskRay
approved these changes
Jan 24, 2025
nico
referenced
this pull request
Jan 24, 2025
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.
This is a follow-up to #120452 in a way.
Since lld/COFF does not yet insert all defined in an obj file before all undefineds (ELF and MachO do this, see #67445 and things linked from there), it's possible that:
For normal archives,
ArchiveFile::addMember()has aseencheck to prevent this. For start-lib lazy objects, we can just check if the archive is still lazy at the recursive call.This bug is similar to issue #59162.
(Eventually, we'll probably want to do what the MachO and ELF ports do.)
Includes a test that caused duplicate symbol diagnostics before this code change.