[BOLT] Don't emit invalid (gdb-breaking) address ranges in gdb-index

[itrofimow]

Empty address map ranges in gdb-index break gdb, so don't emit them

[llvmbot]

@llvm/pr-subscribers-bolt

Author: None (itrofimow)

Changes

Empty address map ranges in gdb-index break gdb, so don't emit them

---

Full diff: https://github.com/llvm/llvm-project/pull/151923.diff

1 Files Affected:

• (modified) bolt/lib/Core/GDBIndex.cpp (+19-5)

diff --git a/bolt/lib/Core/GDBIndex.cpp b/bolt/lib/Core/GDBIndex.cpp
index c7fb4889646b4..5406f0c4e38d2 100644
--- a/bolt/lib/Core/GDBIndex.cpp
+++ b/bolt/lib/Core/GDBIndex.cpp
@@ -99,10 +99,19 @@ void GDBIndex::updateGdbIndexSection(
   Data += SymbolTableOffset - CUTypesOffset;
 
   // Calculate the size of the new address table.
+  const auto IsValidAddressRange = [](const DebugAddressRange &Range) {
+    return Range.HighPC > Range.LowPC;
+  };
+
   uint32_t NewAddressTableSize = 0;
   for (const auto &CURangesPair : ARangesSectionWriter.getCUAddressRanges()) {
     const SmallVector<DebugAddressRange, 2> &Ranges = CURangesPair.second;
-    NewAddressTableSize += Ranges.size() * 20;
+    NewAddressTableSize +=
+        llvm::count_if(Ranges,
+                       [&IsValidAddressRange](const DebugAddressRange &Range) {
+                         return IsValidAddressRange(Range);
+                       }) *
+        20;
   }
 
   // Difference between old and new table (and section) sizes.
@@ -163,10 +172,15 @@ void GDBIndex::updateGdbIndexSection(
     const uint32_t CUIndex = OffsetToIndexMap[CURangesPair.first];
     const DebugAddressRangesVector &Ranges = CURangesPair.second;
     for (const DebugAddressRange &Range : Ranges) {
-      write64le(Buffer, Range.LowPC);
-      write64le(Buffer + 8, Range.HighPC);
-      write32le(Buffer + 16, CUIndex);
-      Buffer += 20;
+      // Don't emit ranges that break gdb,
+      // https://sourceware.org/bugzilla/show_bug.cgi?id=33247.
+      // We've seen [0, 0) ranges here, for instance.
+      if (IsValidAddressRange(Range)) {
+        write64le(Buffer, Range.LowPC);
+        write64le(Buffer + 8, Range.HighPC);
+        write32le(Buffer + 16, CUIndex);
+        Buffer += 20;
+      }
     }
   }
 

[maksfb]

Looks good - thanks for the fix! Can you add a test case?

[itrofimow]

Thanks!
Thing is, I have no clue why such ranges are present in the first place, I just know that this reproduces on a lot of binaries of mine.

So this PR is me fixing the symptoms, and I have no idea what exactly leads to BOLT generating such ranges, nor how to fix the root-cause; thus I don't know how to set up a test for this

[maksfb]

Okay. This change looks straightforward and it only affects gdb-index. Thanks for the fix!

[itrofimow]

Thanks! Could you please merge this PR for me?