[Lldb-commits] [PATCH] Support evaluation of DWARF expressions setting CFA

[Pavel Labath]

Hi jasonmolenda,

This patch enables evaluation of DWARF expressions setting the CFA during stack unwinding.

This makes TestSigtrampUnwind "almost" pass on linux. I am not enabling the test yet since the
symbol name for the signal trampoline does not get resolved properly due to a different bug, but
apart from that, the backtrace is sane.

I am unsure how this change affects Mac. I think it makes the unwinder prefer the DWARF unwind
plan instead of some custom platform-dependant plan. However, it does not affect the end result
- the stack unwinding works as expected.

http://reviews.llvm.org/D7792

Files:
  source/Plugins/Process/Utility/RegisterContextLLDB.cpp
  source/Symbol/DWARFCallFrameInfo.cpp
  source/Symbol/UnwindPlan.cpp

Index: source/Plugins/Process/Utility/RegisterContextLLDB.cpp
===================================================================
--- source/Plugins/Process/Utility/RegisterContextLLDB.cpp
+++ source/Plugins/Process/Utility/RegisterContextLLDB.cpp
@@ -1743,7 +1743,28 @@
             }
             break;
         }
-    // TODO: handle isDWARFExpression
+    case UnwindPlan::Row::CFAValue::isDWARFExpression:
+        {
+            ExecutionContext exe_ctx(m_thread.shared_from_this());
+            Process *process = exe_ctx.GetProcessPtr();
+            DataExtractor dwarfdata (row->GetCFAValue().GetDWARFExpressionBytes(),
+                                     row->GetCFAValue().GetDWARFExpressionLength(),
+                                     process->GetByteOrder(), process->GetAddressByteSize());
+            ModuleSP opcode_ctx;
+            DWARFExpression dwarfexpr (opcode_ctx, dwarfdata, 0, row->GetCFAValue().GetDWARFExpressionLength());
+            dwarfexpr.SetRegisterKind (row_register_kind);
+            Value result;
+            Error error;
+            if (dwarfexpr.Evaluate (&exe_ctx, NULL, NULL, this, 0, NULL, result, &error))
+            {
+                cfa_value = result.GetScalar().ULongLong();
+
+                UnwindLogMsg ("CFA value set by DWARF expression is 0x%" PRIx64, cfa_value);
+                return true;
+            }
+            UnwindLogMsg ("Failed to set CFA value via DWARF expression: %s", error.AsCString());
+            break;
+        }
     default:
         return false;
     }
Index: source/Symbol/DWARFCallFrameInfo.cpp
===================================================================
--- source/Symbol/DWARFCallFrameInfo.cpp
+++ source/Symbol/DWARFCallFrameInfo.cpp
@@ -748,7 +748,9 @@
                 case DW_CFA_def_cfa_expression  : // 0xF    (CFA Definition Instruction)
                     {
                         size_t block_len = (size_t)m_cfi_data.GetULEB128(&offset);
-                        offset += (uint32_t)block_len;
+                        const uint8_t *block_data =
+                            static_cast<const uint8_t *>(m_cfi_data.GetData(&offset, block_len));
+                        row->GetCFAValue().SetIsDWARFExpression(block_data, block_len);
                     }
                     break;
 
Index: source/Symbol/UnwindPlan.cpp
===================================================================
--- source/Symbol/UnwindPlan.cpp
+++ source/Symbol/UnwindPlan.cpp
@@ -436,8 +436,7 @@
     // a register to use to find the Canonical Frame Address, this is not a valid UnwindPlan.
     // CFA set by a DWARF expression is not currently supported, so ignore that as well.
     if (GetRowAtIndex(0).get() == nullptr ||
-            GetRowAtIndex(0)->GetCFAValue().GetValueType() == Row::CFAValue::unspecified ||
-            GetRowAtIndex(0)->GetCFAValue().GetValueType() == Row::CFAValue::isDWARFExpression)
+            GetRowAtIndex(0)->GetCFAValue().GetValueType() == Row::CFAValue::unspecified)
     {
         Log *log(GetLogIfAllCategoriesSet (LIBLLDB_LOG_UNWIND));
         if (log)

EMAIL PREFERENCES
  http://reviews.llvm.org/settings/panel/emailpreferences/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: D7792.20412.patch
Type: text/x-patch
Size: 3109 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/lldb-commits/attachments/20150220/a553354d/attachment.bin>