[Lldb-commits] [lldb] a80c6c7 - [trace] clear any existing tracing sessions before relaunching the binary

[Walter Erquinigo via lldb-commits]

Author: Walter Erquinigo
Date: 2022-03-21T16:03:37-07:00
New Revision: a80c6c7d36d25999a28cfad32e1f461db95ba4dc

URL: https://github.com/llvm/llvm-project/commit/a80c6c7d36d25999a28cfad32e1f461db95ba4dc
DIFF: https://github.com/llvm/llvm-project/commit/a80c6c7d36d25999a28cfad32e1f461db95ba4dc.diff

LOG: [trace] clear any existing tracing sessions before relaunching the binary

There's a bug caused when a process is relaunched: the target, which
doesn't change, keeps the Trace object from the previous process, which
is already defunct, and causes segmentation faults when it's attempted
to be used.
A fix is to clean up the Trace object when the target is disposing of
the previous process during relaunches.

A way to reproduce this:
```
lldb a.out
b main
r
process trace start
c
r
process trace start
```

Differential Revision: https://reviews.llvm.org/D122176

Added: 
    

Modified: 
    lldb/source/Target/Target.cpp
    lldb/test/API/commands/trace/TestTraceStartStop.py

Removed: 
    


################################################################################
diff  --git a/lldb/source/Target/Target.cpp b/lldb/source/Target/Target.cpp
index 00e9fd1eda571..71991085145ab 100644
--- a/lldb/source/Target/Target.cpp
+++ b/lldb/source/Target/Target.cpp
@@ -185,6 +185,8 @@ void Target::CleanupProcess() {
 
 void Target::DeleteCurrentProcess() {
   if (m_process_sp) {
+    // We dispose any active tracing sessions on the current process
+    m_trace_sp.reset();
     m_section_load_history.Clear();
     if (m_process_sp->IsAlive())
       m_process_sp->Destroy(false);

diff  --git a/lldb/test/API/commands/trace/TestTraceStartStop.py b/lldb/test/API/commands/trace/TestTraceStartStop.py
index 841ca437b29f6..d0d65fdd9cf9c 100644
--- a/lldb/test/API/commands/trace/TestTraceStartStop.py
+++ b/lldb/test/API/commands/trace/TestTraceStartStop.py
@@ -166,3 +166,16 @@ def testStartStopLiveThreads(self):
 
         self.expect("thread trace stop", error=True,
             substrs=["error: Process must be launched"])
+
+        # We should be able to trace the program if we relaunch it
+        # For this, we'll trace starting at a 
diff erent point in the new
+        # process.
+        self.expect("breakpoint disable")
+        self.expect("b main.cpp:4")
+        self.expect("r")
+        self.expect("thread trace start")
+        # We can reconstruct the single instruction executed in the first line
+        self.expect("si")
+        self.expect("thread trace dump instructions -c 1",
+            patterns=[f'''thread #1: tid = .*
+  a.out`main \+ 11 at main.cpp:4'''])