-
Notifications
You must be signed in to change notification settings - Fork 13.5k
[lld/COFF] Fix -start-lib / -end-lib after reviews.llvm.org/D116434 #120452
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
Conversation
@llvm/pr-subscribers-lld-coff @llvm/pr-subscribers-lld Author: Nico Weber (nico) ChangesThat change forgot to set This is caught by a pretty simple test (included). Seems like nobody uses these flags in practice at the moment. Fixes https://crbug.com/383531486 Full diff: https://github.com/llvm/llvm-project/pull/120452.diff 2 Files Affected:
diff --git a/lld/COFF/SymbolTable.cpp b/lld/COFF/SymbolTable.cpp
index 6b3375e13e8393..b1d375b2265835 100644
--- a/lld/COFF/SymbolTable.cpp
+++ b/lld/COFF/SymbolTable.cpp
@@ -116,6 +116,7 @@ static void forceLazy(Symbol *s) {
}
case Symbol::Kind::LazyObjectKind: {
InputFile *file = cast<LazyObject>(s)->file;
+ file->lazy = false;
file->symtab.addFile(file);
break;
}
diff --git a/lld/test/COFF/start-lib.ll b/lld/test/COFF/start-lib.ll
index f4f49ed4e4f6e6..02bf4c254ee185 100644
--- a/lld/test/COFF/start-lib.ll
+++ b/lld/test/COFF/start-lib.ll
@@ -79,3 +79,78 @@ define i32 @bar() {
!llvm.linker.options = !{!0}
!0 = !{!"/INCLUDE:bar"}
+
+
+; Check that lazy object files trigger loads correctly.
+; If the links succeed, that's enough, no additional tests needed.
+
+; RUN: llc -filetype=obj %t.dir/main2.ll -o %t-main2.obj
+; RUN: llc -filetype=obj %t.dir/foo.ll -o %t-foo.obj
+; RUN: llc -filetype=obj %t.dir/bar.ll -o %t-bar.obj
+; RUN: llc -filetype=obj %t.dir/baz.ll -o %t-baz.obj
+; RUN: opt -thinlto-bc %t.dir/main2.ll -o %t-main2.bc
+; RUN: opt -thinlto-bc %t.dir/foo.ll -o %t-foo.bc
+; RUN: opt -thinlto-bc %t.dir/bar.ll -o %t-bar.bc
+; RUN: opt -thinlto-bc %t.dir/baz.ll -o %t-baz.bc
+
+; RUN: lld-link -out:%t2.exe -entry:main \
+; RUN: %t-main2.obj %t-foo.obj %t-bar.obj %t-baz.obj
+; RUN: lld-link -out:%t2.exe -entry:main \
+; RUN: %t-main2.obj /start-lib %t-foo.obj %t-bar.obj %t-baz.obj /end-lib
+; RUN: lld-link -out:%t2.exe -entry:main \
+; RUN: /start-lib %t-foo.obj %t-bar.obj %t-baz.obj /end-lib %t-main2.obj
+
+; RUN: lld-link -out:%t2.exe -entry:main \
+; RUN: %t-main2.bc %t-foo.bc %t-bar.bc %t-baz.bc
+; RUN: lld-link -out:%t2.exe -entry:main \
+; RUN: %t-main2.bc /start-lib %t-foo.bc %t-bar.bc %t-baz.bc /end-lib
+; RUN: lld-link -out:%t2.exe -entry:main \
+; RUN: /start-lib %t-foo.bc %t-bar.bc %t-baz.bc /end-lib %t-main2.bc
+
+#--- main2.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 @main() {
+ call void () @bar()
+ ret void
+}
+
+
+#--- foo.ll
+
+target datalayout = "e-m:w-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-pc-windows-msvc"
+
+define void @foo() {
+ ret void
+}
+
+
+#--- bar.ll
+
+target datalayout = "e-m:w-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-pc-windows-msvc"
+
+; One undef from the lazy obj file before it, one from the one after it.
+declare void @foo()
+declare void @baz()
+
+define void @bar() {
+ call void () @foo()
+ call void () @baz()
+ ret void
+}
+
+
+#--- baz.ll
+
+target datalayout = "e-m:w-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-pc-windows-msvc"
+
+define void @baz() {
+ ret void
+}
|
✅ With the latest revision this PR passed the undef deprecator. |
It complains about
I suppose I'll make it say "undefined symbol", but that check seems a bit overeager |
That change forgot to set `lazy` to false before calling `addFile()` in `forceLazy()` which caused `addFile()` to parse the file we want to force a load for to be added as a lazy object again instead of adding the file to `ctx.objFileInstances`. This is caught by a pretty simple test (included).
214dcb8
to
511d55a
Compare
…434 (#124294) 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: 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 #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.
That change forgot to set
lazy
to false before callingaddFile()
inforceLazy()
which causedaddFile()
to parse the file we want to force a load for to be added as a lazy object again instead of adding the file toctx.objFileInstances
.This is caught by a pretty simple test (included).
Seems like nobody uses these flags in practice at the moment.
Fixes https://crbug.com/383531486