[llvm] 232e0a0 - [lto] Do not try to internalize symbols with escaped name
via llvm-commits
llvm-commits at lists.llvm.org
Fri Oct 14 13:36:28 PDT 2022
Author: serge-sans-paille
Date: 2022-10-14T22:34:17+02:00
New Revision: 232e0a011e8c07053bdc0156f312046eb09f52b3
URL: https://github.com/llvm/llvm-project/commit/232e0a011e8c07053bdc0156f312046eb09f52b3
DIFF: https://github.com/llvm/llvm-project/commit/232e0a011e8c07053bdc0156f312046eb09f52b3.diff
LOG: [lto] Do not try to internalize symbols with escaped name
Because of LLVM mangling escape sequence (through '\01' prefix), it is possible
for a single symbols two have two different IR representations.
For instance, consider @symbol and @"\01_symbol". On OSX, because of the system
mangling rules, these two IR names point are converted in the same final symbol
upon linkage.
LTO doesn't model this behavior, which may result in symbols being incorrectly
internalized (if all reference use the escaping sequence while the definition
doesn't).
The proper approach is probably to use the mangled name to compute GUID to
avoid the dual representation, but we can also avoid discarding symbols that are
bound to two different IR names. This is an approximation, but it's less
intrusive on the codebase.
Fix #57864
Differential Revision: https://reviews.llvm.org/D135710
Added:
llvm/test/LTO/X86/hidden-escaped-symbols-alt.ll
llvm/test/LTO/X86/hidden-escaped-symbols.ll
llvm/test/ThinLTO/X86/hidden-escaped-symbols-alt.ll
llvm/test/ThinLTO/X86/hidden-escaped-symbols.ll
Modified:
llvm/lib/LTO/LTO.cpp
Removed:
################################################################################
diff --git a/llvm/lib/LTO/LTO.cpp b/llvm/lib/LTO/LTO.cpp
index 0ce8519faa0c1..dc28b681a1515 100644
--- a/llvm/lib/LTO/LTO.cpp
+++ b/llvm/lib/LTO/LTO.cpp
@@ -567,6 +567,22 @@ void LTO::addModuleToGlobalRes(ArrayRef<InputFile::Symbol> Syms,
GlobalRes.IRName = std::string(Sym.getIRName());
}
+ // In rare occasion, the symbol used to initialize GlobalRes has a
diff erent
+ // IRName from the inspected Symbol. This can happen on macOS + iOS, when a
+ // symbol is referenced through its mangled name, say @"\01_symbol" while
+ // the IRName is @symbol (the prefix underscore comes from MachO mangling).
+ // In that case, we have the same actual Symbol that can get two
diff erent
+ // GUID, leading to some invalid internalization. Workaround this by marking
+ // the GlobalRes external.
+
+ // FIXME: instead of this check, it would be desirable to compute GUIDs
+ // based on mangled name, but this requires an access to the Target Triple
+ // and would be relatively invasive on the codebase.
+ if (GlobalRes.IRName != Sym.getIRName()) {
+ GlobalRes.Partition = GlobalResolution::External;
+ GlobalRes.VisibleOutsideSummary = true;
+ }
+
// Set the partition to external if we know it is re-defined by the linker
// with -defsym or -wrap options, used elsewhere, e.g. it is visible to a
// regular object, is referenced from llvm.compiler.used/llvm.used, or was
diff --git a/llvm/test/LTO/X86/hidden-escaped-symbols-alt.ll b/llvm/test/LTO/X86/hidden-escaped-symbols-alt.ll
new file mode 100644
index 0000000000000..ca8e7d8eb2b2e
--- /dev/null
+++ b/llvm/test/LTO/X86/hidden-escaped-symbols-alt.ll
@@ -0,0 +1,41 @@
+; Check interaction between LTO and LLVM mangling escape char, see #57864.
+
+; RUN: split-file %s %t
+; RUN: opt %t/hide-me.ll -o %t/hide-me.bc
+; RUN: opt %t/ref.ll -o %t/ref.bc
+; RUN: llvm-lto2 run \
+; RUN: -r %t/hide-me.bc,_hide_me,p \
+; RUN: -r %t/ref.bc,_main,plx \
+; RUN: -r %t/ref.bc,_hide_me,l \
+; RUN: --select-save-temps=precodegen \
+; RUN: -o %t/out \
+; RUN: %t/hide-me.bc %t/ref.bc
+; RUN: llvm-dis %t/out.0.5.precodegen.bc -o - | FileCheck %s
+
+
+;--- hide-me.ll
+target datalayout = "e-m:o-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-apple-macosx10.7.0"
+
+@"\01_hide_me" = hidden local_unnamed_addr global i8 8, align 1
+
+;--- ref.ll
+target datalayout = "e-m:o-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-apple-macosx10.7.0"
+
+ at hide_me = external local_unnamed_addr global i8
+
+define i8 @main() {
+ %1 = load i8, ptr @hide_me, align 1
+ ret i8 %1
+}
+
+
+; CHECK: @"\01_hide_me" = hidden local_unnamed_addr global i8 8, align 1
+; CHECK: @hide_me = external dso_local local_unnamed_addr global i8
+
+; CHECK: define dso_local i8 @main() local_unnamed_addr #0 {
+; CHECK: %1 = load i8, ptr @hide_me, align 1
+; CHECK: ret i8 %1
+; CHECK: }
+
diff --git a/llvm/test/LTO/X86/hidden-escaped-symbols.ll b/llvm/test/LTO/X86/hidden-escaped-symbols.ll
new file mode 100644
index 0000000000000..aec617e25d3d8
--- /dev/null
+++ b/llvm/test/LTO/X86/hidden-escaped-symbols.ll
@@ -0,0 +1,41 @@
+; Check interaction between LTO and LLVM mangling escape char, see #57864.
+
+; RUN: split-file %s %t
+; RUN: opt %t/hide-me.ll -o %t/hide-me.bc
+; RUN: opt %t/ref.ll -o %t/ref.bc
+; RUN: llvm-lto2 run \
+; RUN: -r %t/hide-me.bc,_hide_me,p \
+; RUN: -r %t/ref.bc,_main,plx \
+; RUN: -r %t/ref.bc,_hide_me,l \
+; RUN: --select-save-temps=precodegen \
+; RUN: -o %t/out \
+; RUN: %t/hide-me.bc %t/ref.bc
+; RUN: llvm-dis %t/out.0.5.precodegen.bc -o - | FileCheck %s
+
+
+;--- hide-me.ll
+target datalayout = "e-m:o-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-apple-macosx10.7.0"
+
+ at hide_me = hidden local_unnamed_addr global i8 8, align 1
+
+;--- ref.ll
+target datalayout = "e-m:o-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-apple-macosx10.7.0"
+
+@"\01_hide_me" = external local_unnamed_addr global i8
+
+define i8 @main() {
+ %1 = load i8, ptr @"\01_hide_me", align 1
+ ret i8 %1
+}
+
+
+; CHECK: @hide_me = hidden local_unnamed_addr global i8 8, align 1
+; CHECK: @"\01_hide_me" = external dso_local local_unnamed_addr global i8
+
+; CHECK: define dso_local i8 @main() local_unnamed_addr #0 {
+; CHECK: %1 = load i8, ptr @"\01_hide_me", align 1
+; CHECK: ret i8 %1
+; CHECK: }
+
diff --git a/llvm/test/ThinLTO/X86/hidden-escaped-symbols-alt.ll b/llvm/test/ThinLTO/X86/hidden-escaped-symbols-alt.ll
new file mode 100644
index 0000000000000..dadd1d434256c
--- /dev/null
+++ b/llvm/test/ThinLTO/X86/hidden-escaped-symbols-alt.ll
@@ -0,0 +1,42 @@
+; Check interaction between LTO and LLVM mangling escape char, see #57864.
+
+; RUN: split-file %s %t
+; RUN: opt -module-summary %t/hide-me.ll -o %t/hide-me.bc
+; RUN: opt -module-summary %t/ref.ll -o %t/ref.bc
+; RUN: llvm-lto2 run \
+; RUN: -r %t/hide-me.bc,_hide_me,p \
+; RUN: -r %t/ref.bc,_main,plx \
+; RUN: -r %t/ref.bc,_hide_me,l \
+; RUN: --select-save-temps=precodegen \
+; RUN: -o %t/out \
+; RUN: %t/hide-me.bc %t/ref.bc
+; RUN: llvm-dis %t/out.1.5.precodegen.bc -o - | FileCheck --check-prefix=CHECK-HIDE %s
+; RUN: llvm-dis %t/out.2.5.precodegen.bc -o - | FileCheck --check-prefix=CHECK-REF %s
+
+
+;--- hide-me.ll
+target datalayout = "e-m:o-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-apple-macosx10.7.0"
+
+@"\01_hide_me" = hidden local_unnamed_addr global i8 8, align 1
+
+;--- ref.ll
+target datalayout = "e-m:o-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-apple-macosx10.7.0"
+
+ at hide_me = external local_unnamed_addr global i8
+
+define i8 @main() {
+ %1 = load i8, ptr @hide_me, align 1
+ ret i8 %1
+}
+
+
+; CHECK-HIDE: @"\01_hide_me" = hidden local_unnamed_addr global i8 8, align 1
+
+; CHECK-REF: @hide_me = external local_unnamed_addr global i8
+; CHECK-REF: define dso_local i8 @main() local_unnamed_addr #0 {
+; CHECK-REF: %1 = load i8, ptr @hide_me, align 1
+; CHECK-REF: ret i8 %1
+; CHECK-REF: }
+
diff --git a/llvm/test/ThinLTO/X86/hidden-escaped-symbols.ll b/llvm/test/ThinLTO/X86/hidden-escaped-symbols.ll
new file mode 100644
index 0000000000000..8d0e22f0fd224
--- /dev/null
+++ b/llvm/test/ThinLTO/X86/hidden-escaped-symbols.ll
@@ -0,0 +1,42 @@
+; Check interaction between LTO and LLVM mangling escape char, see #57864.
+
+; RUN: split-file %s %t
+; RUN: opt -module-summary %t/hide-me.ll -o %t/hide-me.bc
+; RUN: opt -module-summary %t/ref.ll -o %t/ref.bc
+; RUN: llvm-lto2 run \
+; RUN: -r %t/hide-me.bc,_hide_me,p \
+; RUN: -r %t/ref.bc,_main,plx \
+; RUN: -r %t/ref.bc,_hide_me,l \
+; RUN: --select-save-temps=precodegen \
+; RUN: -o %t/out \
+; RUN: %t/hide-me.bc %t/ref.bc
+; RUN: llvm-dis %t/out.1.5.precodegen.bc -o - | FileCheck --check-prefix=CHECK-HIDE %s
+; RUN: llvm-dis %t/out.2.5.precodegen.bc -o - | FileCheck --check-prefix=CHECK-REF %s
+
+
+;--- hide-me.ll
+target datalayout = "e-m:o-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-apple-macosx10.7.0"
+
+ at hide_me = hidden local_unnamed_addr global i8 8, align 1
+
+;--- ref.ll
+target datalayout = "e-m:o-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-apple-macosx10.7.0"
+
+@"\01_hide_me" = external local_unnamed_addr global i8
+
+define i8 @main() {
+ %1 = load i8, ptr @"\01_hide_me", align 1
+ ret i8 %1
+}
+
+
+; CHECK-HIDE: @hide_me = hidden local_unnamed_addr global i8 8, align 1
+
+; CHECK-REF: @"\01_hide_me" = external local_unnamed_addr global i8
+; CHECK-REF: define dso_local i8 @main() local_unnamed_addr #0 {
+; CHECK-REF: %1 = load i8, ptr @"\01_hide_me", align 1
+; CHECK-REF: ret i8 %1
+; CHECK-REF: }
+
More information about the llvm-commits
mailing list