[Lldb-commits] [lldb] r295440 - NPL: Fix one more bug in the single step workaround

[Pavel Labath via lldb-commits]

Author: labath
Date: Fri Feb 17 05:48:34 2017
New Revision: 295440

URL: http://llvm.org/viewvc/llvm-project?rev=295440&view=rev
Log:
NPL: Fix one more bug in the single step workaround

In the case we are stepping over the thread creation instruction, we
will end up calling Thread::SingleStep back-to-back twice (because of
the intermediate PTRACE_EVENT_CLONE stop). This will cause the cpu mask
to be set inappropriately (because the old SingleStepCheck object will
be destroyed after we create the new one), and the single-step will
fail.

Before the refactor the code was still incorrect in this case, but in a
different way (the thread was left with the incorrect mask after the
stepping was complete), so this was not easy to spot.

This fixes TestCreateDuringInstructionStep on the affected devices.

Modified:
    lldb/trunk/packages/Python/lldbsuite/test/linux/thread/create_during_instruction_step/TestCreateDuringInstructionStep.py
    lldb/trunk/source/Plugins/Process/Linux/NativeThreadLinux.cpp
    lldb/trunk/source/Plugins/Process/Linux/SingleStepCheck.cpp

Modified: lldb/trunk/packages/Python/lldbsuite/test/linux/thread/create_during_instruction_step/TestCreateDuringInstructionStep.py
URL: http://llvm.org/viewvc/llvm-project/lldb/trunk/packages/Python/lldbsuite/test/linux/thread/create_during_instruction_step/TestCreateDuringInstructionStep.py?rev=295440&r1=295439&r2=295440&view=diff
==============================================================================
--- lldb/trunk/packages/Python/lldbsuite/test/linux/thread/create_during_instruction_step/TestCreateDuringInstructionStep.py (original)
+++ lldb/trunk/packages/Python/lldbsuite/test/linux/thread/create_during_instruction_step/TestCreateDuringInstructionStep.py Fri Feb 17 05:48:34 2017
@@ -16,10 +16,7 @@ from lldbsuite.test import lldbutil
 class CreateDuringInstructionStepTestCase(TestBase):
 
     mydir = TestBase.compute_mydir(__file__)
-
-    def setUp(self):
-        # Call super's setUp().
-        TestBase.setUp(self)
+    NO_DEBUG_INFO_TESTCASE = True
 
     @skipUnlessPlatform(['linux'])
     @expectedFailureAndroid('llvm.org/pr24737', archs=['arm'])

Modified: lldb/trunk/source/Plugins/Process/Linux/NativeThreadLinux.cpp
URL: http://llvm.org/viewvc/llvm-project/lldb/trunk/source/Plugins/Process/Linux/NativeThreadLinux.cpp?rev=295440&r1=295439&r2=295440&view=diff
==============================================================================
--- lldb/trunk/source/Plugins/Process/Linux/NativeThreadLinux.cpp (original)
+++ lldb/trunk/source/Plugins/Process/Linux/NativeThreadLinux.cpp Fri Feb 17 05:48:34 2017
@@ -225,7 +225,13 @@ Error NativeThreadLinux::SingleStep(uint
   MaybeLogStateChange(new_state);
   m_state = new_state;
   m_stop_info.reason = StopReason::eStopReasonNone;
-  m_step_workaround = SingleStepWorkaround::Get(m_tid);
+
+  if(!m_step_workaround) {
+    // If we already hava a workaround inplace, don't reset it. Otherwise, the
+    // destructor of the existing instance will run after the new instance has
+    // fetched the cpu mask, and the thread will end up with the wrong mask.
+    m_step_workaround = SingleStepWorkaround::Get(m_tid);
+  }
 
   intptr_t data = 0;
   if (signo != LLDB_INVALID_SIGNAL_NUMBER)

Modified: lldb/trunk/source/Plugins/Process/Linux/SingleStepCheck.cpp
URL: http://llvm.org/viewvc/llvm-project/lldb/trunk/source/Plugins/Process/Linux/SingleStepCheck.cpp?rev=295440&r1=295439&r2=295440&view=diff
==============================================================================
--- lldb/trunk/source/Plugins/Process/Linux/SingleStepCheck.cpp (original)
+++ lldb/trunk/source/Plugins/Process/Linux/SingleStepCheck.cpp Fri Feb 17 05:48:34 2017
@@ -172,8 +172,9 @@ std::unique_ptr<SingleStepWorkaround> Si
 }
 
 SingleStepWorkaround::~SingleStepWorkaround() {
+  Log *log = ProcessPOSIXLog::GetLogIfAllCategoriesSet(POSIX_LOG_THREAD);
+  LLDB_LOG(log, "Removing workaround");
   if (sched_setaffinity(m_tid, sizeof m_original_set, &m_original_set) != 0) {
-    Log *log(GetLogIfAllCategoriesSet(LIBLLDB_LOG_THREAD));
     LLDB_LOG(log, "Unable to reset cpu affinity for thread {0}: {1}", m_tid,
              Error(errno, eErrorTypePOSIX));
   }