[MC] Fix DWARF file table for files with empty DWARF (#119020) #119229
Conversation
|
Thank you for submitting a Pull Request (PR) to the LLVM Project! This PR will be automatically labeled and the relevant teams will be notified. If you wish to, you can add reviewers by using the "Reviewers" section on this page. If this is not working for you, it is probably because you do not have write permissions for the repository. In which case you can instead tag reviewers by name in a comment by using If you have received no comments on your PR for a week, you can request a review by "ping"ing the PR by adding a comment “Ping”. The common courtesy "ping" rate is once a week. Please remember that you are asking for valuable time from other developers. If you have further questions, they may be answered by the LLVM GitHub User Guide. You can also ask questions in a comment on this PR, on the LLVM Discord or on the forums. |
|
@llvm/pr-subscribers-mc Author: Aleksei Vetrov (noxwell) ChangesUpdate root file in DWARF file/line table as soon as we see the first "#line" directive. This was moved from "enabledGenDwarfForAssembly", which is called right before we emit DWARF information. But if the file is empty or contains expressions that doesn't need DWARF, it is never called, leaving an original root file and not the file in the "#line" directive. Add a test checking for this case. Full diff: https://github.com/llvm/llvm-project/pull/119229.diff 2 Files Affected:
diff --git a/llvm/lib/MC/MCParser/AsmParser.cpp b/llvm/lib/MC/MCParser/AsmParser.cpp
index 3ce45f7d5d67e1..00291147b68682 100644
--- a/llvm/lib/MC/MCParser/AsmParser.cpp
+++ b/llvm/lib/MC/MCParser/AsmParser.cpp
@@ -162,8 +162,8 @@ class AsmParser : public MCAsmParser {
};
CppHashInfoTy CppHashInfo;
- /// The filename from the first cpp hash file line comment, if any.
- StringRef FirstCppHashFilename;
+ /// Have we seen any file line comment.
+ bool HadCppHashFilename = false;
/// List of forward directional labels for diagnosis at the end.
SmallVector<std::tuple<SMLoc, CppHashInfoTy, MCSymbol *>, 4> DirLabels;
@@ -952,12 +952,6 @@ bool AsmParser::enabledGenDwarfForAssembly() {
// the assembler source was produced with debug info already) then emit one
// describing the assembler source file itself.
if (getContext().getGenDwarfFileNumber() == 0) {
- // Use the first #line directive for this, if any. It's preprocessed, so
- // there is no checksum, and of course no source directive.
- if (!FirstCppHashFilename.empty())
- getContext().setMCLineTableRootFile(
- /*CUID=*/0, getContext().getCompilationDir(), FirstCppHashFilename,
- /*Cksum=*/std::nullopt, /*Source=*/std::nullopt);
const MCDwarfFile &RootFile =
getContext().getMCDwarfLineTable(/*CUID=*/0).getRootFile();
getContext().setGenDwarfFileNumber(getStreamer().emitDwarfFileDirective(
@@ -2440,8 +2434,19 @@ bool AsmParser::parseCppHashLineFilenameComment(SMLoc L, bool SaveLocInfo) {
CppHashInfo.Filename = Filename;
CppHashInfo.LineNumber = LineNumber;
CppHashInfo.Buf = CurBuffer;
- if (FirstCppHashFilename.empty())
- FirstCppHashFilename = Filename;
+ if (!HadCppHashFilename) {
+ HadCppHashFilename = true;
+ // If we haven't encountered any .file directives, then the first #line
+ // directive describes the "root" file and directory of the compilation
+ // unit.
+ if (getContext().getGenDwarfFileNumber() == 0) {
+ // It's preprocessed, so there is no checksum, and of course no source
+ // directive.
+ getContext().setMCLineTableRootFile(
+ /*CUID=*/0, getContext().getCompilationDir(), Filename,
+ /*Cksum=*/std::nullopt, /*Source=*/std::nullopt);
+ }
+ }
return false;
}
diff --git a/llvm/test/MC/ELF/debug-hash-file-empty-dwarf.s b/llvm/test/MC/ELF/debug-hash-file-empty-dwarf.s
new file mode 100644
index 00000000000000..bf7f9cb53b5b0e
--- /dev/null
+++ b/llvm/test/MC/ELF/debug-hash-file-empty-dwarf.s
@@ -0,0 +1,26 @@
+// RUN: llvm-mc -triple x86_64-unknown-linux-gnu -filetype obj -g -dwarf-version 5 -o %t %s
+// RUN: llvm-dwarfdump -debug-info -debug-line %t | FileCheck %s
+
+// CHECK-NOT: DW_TAG_
+
+// CHECK: include_directories[ 0] =
+// CHECK-NOT: include_directories[ 1] =
+// CHECK: file_names[ 0]:
+// CHECK-NEXT: name: "/MyTest/Inputs/other.S"
+// CHECK-NEXT: dir_index: 0
+// CHECK-NOT: file_names[ 1]:
+
+// RUN: llvm-mc -triple=x86_64 -filetype=obj -g -dwarf-version=5 -fdebug-prefix-map=/MyTest=/src_root %s -o %t.5.o
+// RUN: llvm-dwarfdump -debug-info -debug-line %t.5.o | FileCheck %s --check-prefixes=MAP
+
+// MAP-NOT: DW_TAG_
+
+// MAP: include_directories[ 0] = "{{.*}}"
+// MAP-NEXT: file_names[ 0]:
+// MAP-NEXT: name: "/src_root/Inputs/other.S"
+// MAP-NEXT: dir_index: 0
+
+# 1 "/MyTest/Inputs/other.S"
+
+.section .data
+ .asciz "data"
|
|
I've noticed that |
noxwell
force-pushed
the
asm-empty-dwarf
branch
from
f9f47d2 to
86a23af
Compare
noxwell
force-pushed
the
asm-empty-dwarf
branch
from
86a23af to
61c261a
Compare
Update root file in DWARF file/line table as soon as we see the first "#line" directive. This was moved from "enabledGenDwarfForAssembly", which is called right before we emit DWARF information. But if the file is empty or contains expressions that doesn't need DWARF, it is never called, leaving an original root file and not the file in the "#line" directive. Add a test checking for this case. Fixes: llvm#119020
noxwell
force-pushed
the
asm-empty-dwarf
branch
from
61c261a to
9b658b7
Compare
IIUC, masm is the "Microsoft Macro Assembler" (https://en.wikipedia.org/wiki/Microsoft_Macro_Assembler). Windows doesn't use ELF for its object file container, and I think DWARF is only supported from ELF and mach-o (darwin-derived platforms). but @rnk or @zmodem could clarify if I'm wrong.
Looks like it.
Maybe @rnk @zmodem @ericastor @nico could help answer that? |
|
It doesn't seem like it would be bad to include this in MasmParser.cpp as well? It is true that it was forked from AsmParser.cpp... but also, I'm not sure if a test case is feasible, since (as mentioned) it's really designed for targeting Windows. I'll let @rnk weigh in on whether it'd be worthwhile. |
dwblaikie
left a comment
dwblaikie
left a comment
There was a problem hiding this comment.
I'm fine with this change as-is. The Masm discussion can continue/be addressed in a separate change if needed.
DWARF is not normally used on Windows, but it can be used (Clang can emit it, and I think it's not uncommon in MinGW builds maybe), and I believe LLDB can consume it on Windows. Using MASM to generate DWARF does seem like a fringe use case though. I don't even know if there are flags to actually do so.
This sounds reasonable to me. |
|
@noxwell Congratulations on having your first Pull Request (PR) merged into the LLVM Project! Your changes will be combined with recent changes from other authors, then tested by our build bots. If there is a problem with a build, you may receive a report in an email or a comment on this PR. Please check whether problems have been caused by your change specifically, as the builds can include changes from many authors. It is not uncommon for your change to be included in a build that fails due to someone else's changes, or infrastructure issues. How to do this, and the rest of the post-merge process, is covered in detail here. If your change does cause a problem, it may be reverted, or you can revert it yourself. This is a normal part of LLVM development. You can fix your changes and open a new PR to merge them again. If you don't get any reports, no action is required from you. Your changes are working as expected, well done! |
|
This causes emitting It makes object files contain the absolute path of build directory for building |
Probably not, but it seems like there were no tests in tree for that! |
Yes, it needs a check |
Update root file in DWARF file/line table as soon as we see the first "#line" directive. This was moved from "enabledGenDwarfForAssembly", which is called right before we emit DWARF information. But if the file is empty or contains expressions that doesn't need DWARF, it is never called, leaving an original root file and not the file in the "#line" directive. Add a test checking for this case. This is reapply of llvm#119229 with the following fix: "MCContext::setMCLineTableRootFile" has the effect of adding ".debug_line" section to the output, even if DWARF generation is disabled. Add a check and a test for this case. Fixes: llvm#119020 Fixes: llvm#119229
Update root file in DWARF file/line table as soon as we see the first "#line" directive. This was moved from "enabledGenDwarfForAssembly", which is called right before we emit DWARF information. But if the file is empty or contains expressions that doesn't need DWARF, it is never called, leaving an original root file and not the file in the "#line" directive. Add a test checking for this case. This is reapply of llvm#119229 with the following fix: "MCContext::setMCLineTableRootFile" has the effect of adding ".debug_line" section to the output, even if DWARF generation is disabled. Add a check and a test for this case. Fixes: llvm#119020 Fixes: llvm#119229
Update root file in DWARF file/line table as soon as we see the first "#line" directive. This was moved from "enabledGenDwarfForAssembly", which is called right before we emit DWARF information. But if the file is empty or contains expressions that doesn't need DWARF, it is never called, leaving an original root file and not the file in the "#line" directive. Add a test checking for this case. This is reapply of #119229 with the following fix: "MCContext::setMCLineTableRootFile" has the effect of adding ".debug_line" section to the output, even if DWARF generation is disabled. Add a check and a test for this case. Fixes: #119020 Fixes: #119229
7 participants
Update root file in DWARF file/line table as soon as we see the first "#line" directive.
This was moved from "enabledGenDwarfForAssembly", which is called right before we emit DWARF information. But if the file is empty or contains expressions that doesn't need DWARF, it is never called, leaving an original root file and not the file in the "#line" directive.
Add a test checking for this case.
Fixes: #119020