[PATCH] D79967: Fix debug info for NoDebug attr

[David Blaikie via Phabricator via cfe-commits]

dblaikie added a subscriber: vsk.
dblaikie added a comment.

Could you check the commit history for this feature and rope in some folks who added the function declaration work (it's for debug call sites) - maybe @vsk is the right person, or knows who is, to check this is the right fix for it/doesn't adversely affect the feature this code was added to implement.



================
Comment at: clang/test/CodeGen/nodebug-attr.c:5-6
+
+// CHECK-NOT: define {{.*}}@foo{{.*}}!dbg
+// CHECK-LABEL: define {{.*}}@foo
+// CHECK-NOT: ret {{.*}}!dbg
----------------
Does this test fail when the bug is present? I'd have thought not - my understanding was that CHECK-LABEL is found first, then CHECK-NOT is tested between labels/other checks, so it wouldn't find @foo.*!dbg anyway.

I think maybe it'd be better to tighten up the CHECK-LABEL to include "#0 {" at the end and a comment saying how that CHECK part ensures there's no !dbg attached to it.


================
Comment at: clang/test/CodeGen/nodebug-attr.c:8-15
+__attribute__((nodebug)) void foo(int *a) {
+  *a = 1;
+}
+
+// CHECK-LABEL: define {{.*}}@bar{{.*}}!dbg
+void bar(int *a) {
+  foo(a);
----------------
It looks like this test case currently crashes LLVM:

h$ clang++-tot test.cpp -g -c -O3 && llvm-dwarfdump-tot test.o
PLEASE submit a bug report to https://bugs.llvm.org/ and include the crash backtrace, preprocessed source, and associated run script.
Stack dump:
0.      Program arguments: clang++-tot test.cpp -g -c -O3 
1.      <eof> parser at end of file
2.      Code generation
3.      Running pass 'Function Pass Manager' on module 'test.cpp'.
4.      Running pass 'Debug Variable Analysis' on function '@_Z3fooPi'
 #0 0x0000000006b49927 llvm::sys::PrintStackTrace(llvm::raw_ostream&) /usr/local/google/home/blaikie/dev/llvm/src/llvm/lib/Support/Unix/Signals.inc:564:11
 #1 0x0000000006b49ac9 PrintStackTraceSignalHandler(void*) /usr/local/google/home/blaikie/dev/llvm/src/llvm/lib/Support/Unix/Signals.inc:625:1
 #2 0x0000000006b482db llvm::sys::RunSignalHandlers() /usr/local/google/home/blaikie/dev/llvm/src/llvm/lib/Support/Signals.cpp:67:5
 #3 0x0000000006b4921e llvm::sys::CleanupOnSignal(unsigned long) /usr/local/google/home/blaikie/dev/llvm/src/llvm/lib/Support/Unix/Signals.inc:362:1
 #4 0x0000000006a7cae8 (anonymous namespace)::CrashRecoveryContextImpl::HandleCrash(int, unsigned long) /usr/local/google/home/blaikie/dev/llvm/src/llvm/lib/Support/CrashRecoveryContext.cpp:77:20
 #5 0x0000000006a7cd6e CrashRecoverySignalHandler(int) /usr/local/google/home/blaikie/dev/llvm/src/llvm/lib/Support/CrashRecoveryContext.cpp:383:1
 #6 0x00007fc197167520 __restore_rt (/lib/x86_64-linux-gnu/libpthread.so.0+0x13520)
 #7 0x0000000005a309ac llvm::DICompileUnit::getEmissionKind() const /usr/local/google/home/blaikie/dev/llvm/src/llvm/include/llvm/IR/DebugInfoMetadata.h:1272:31
 #8 0x0000000005a4a107 llvm::LexicalScopes::initialize(llvm::MachineFunction const&) /usr/local/google/home/blaikie/dev/llvm/src/llvm/lib/CodeGen/LexicalScopes.cpp:53:70
 #9 0x0000000005e128dd (anonymous namespace)::LDVImpl::computeIntervals() /usr/local/google/home/blaikie/dev/llvm/src/llvm/lib/CodeGen/LiveDebugVariables.cpp:972:17
#10 0x0000000005e11284 (anonymous namespace)::LDVImpl::runOnMachineFunction(llvm::MachineFunction&) /usr/local/google/home/blaikie/dev/llvm/src/llvm/lib/CodeGen/LiveDebugVariables.cpp:0:3
#11 0x0000000005e10fe4 llvm::LiveDebugVariables::runOnMachineFunction(llvm::MachineFunction&) /usr/local/google/home/blaikie/dev/llvm/src/llvm/lib/CodeGen/LiveDebugVariables.cpp:1014:3


Is that crash something this patch is intended to address, or unrelated? I guess it's intended to address that problem - because it's a DISubprogram for 'foo' with no DICompileUnit attachment that causes the crash later on.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D79967/new/

https://reviews.llvm.org/D79967