[Lldb-commits] [PATCH] D73148: [lldb/Value] Report size of Value as size of underlying data buffer

[Vedant Kumar via Phabricator via lldb-commits]

vsk created this revision.
vsk added reviewers: jingham, aprantl.

Value::GetValueByteSize() assumes that the size of a Value is always
equal to the size of the underlying variable's CompilerType. However,
llvm's DWARF generator doesn't appear to create fresh types with new
sizes when emitting fragments / DW_OP_piece. It seems like there's a
faulty assumption in lldb (although it's plausible that llvm is just
generating weird DWARF).

This patch teaches GetValueByteSize() to report the size of a Value as
the size of its data buffer if it's a host address.

I'm not sure whether this is an exhaustive fix. This does fix an ASan
error in lldb seen when debugging a clang binary.

ASan report: F11233245: asan-report-rdar58665925.txt <https://reviews.llvm.org/F11233245>

rdar://58665925


https://reviews.llvm.org/D73148

Files:
  lldb/source/Core/Value.cpp
  lldb/unittests/Expression/DWARFExpressionTest.cpp


Index: lldb/unittests/Expression/DWARFExpressionTest.cpp
===================================================================
--- lldb/unittests/Expression/DWARFExpressionTest.cpp
+++ lldb/unittests/Expression/DWARFExpressionTest.cpp
@@ -360,4 +360,16 @@
       // Note that the "00" should really be "undef", but we can't
       // represent that yet.
       llvm::HasValue(GetScalar(16, 0xff00, true)));
+
+  for (unsigned char ByteSize = 1; ByteSize <= 8; ++ByteSize) {
+    llvm::Expected<Scalar> empty = Evaluate({DW_OP_piece, ByteSize});
+    // Note that the "00" should really be "undef", but we can't
+    // represent that yet.
+    EXPECT_THAT_EXPECTED(empty,
+                         llvm::HasValue(GetScalar(ByteSize * 8, 0, true)));
+
+    Value pieces;
+    ASSERT_EQ(pieces.AppendDataToHostBuffer(empty.get()), ByteSize);
+    ASSERT_EQ(pieces.GetValueByteSize(nullptr, nullptr), ByteSize);
+  }
 }
Index: lldb/source/Core/Value.cpp
===================================================================
--- lldb/source/Core/Value.cpp
+++ lldb/source/Core/Value.cpp
@@ -222,6 +222,11 @@
   case eContextTypeLLDBType: // Type *
   case eContextTypeVariable: // Variable *
   {
+    // The size of this Value may be less than the size of the type of its
+    // source variable due to truncating operations such as DW_OP_piece.
+    if (m_value_type == eValueTypeHostAddress)
+      return GetBuffer().GetByteSize();
+
     auto *scope = exe_ctx ? exe_ctx->GetBestExecutionContextScope() : nullptr;
     if (llvm::Optional<uint64_t> size = GetCompilerType().GetByteSize(scope)) {
       if (error_ptr)


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D73148.239455.patch
Type: text/x-patch
Size: 1613 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/lldb-commits/attachments/20200122/d62adee7/attachment.bin>