[Lldb-commits] [PATCH] D18692: Fix a cornercase in breakpoint reporting
Pavel Labath via lldb-commits
lldb-commits at lists.llvm.org
Fri Apr 1 05:27:21 PDT 2016
labath created this revision.
labath added a reviewer: clayborg.
labath added a subscriber: lldb-commits.
This resolves a similar problem as D16720 (which handled the case when we single-step onto a
breakpoint), but this one deals with involutary stops: when we stop a thread (e.g. because
another thread has hit a breakpont and we are doing a full stop), we can end up stopping it right
before it executes a breakpoint instruction. In this case, the stop reason will be empty, but we
will still step over the breakpoint when do the next resume, thereby missing a breakpoint hit.
I have observed this happening in TestConcurrentEvents, but I have no idea how to reproduce this
behavior more reliably.
@@ -2089,6 +2089,23 @@
handled = true;
+ addr_t pc = thread_sp->GetRegisterContext()->GetPC();
+ lldb::BreakpointSiteSP bp_site_sp =
+ // If the current pc is a breakpoint site then the StopInfo should be set to Breakpoint
+ // even though the remote stub did not set it as such. This can happen when
+ // the thread is involuntarily interrupted (e.g. due to stops on other
+ // threads) just as it is about to execute the breakpoint instruction.
+ if (bp_site_sp && bp_site_sp->ValidForThisThread(thread_sp.get()))
+ StopInfo::CreateStopReasonWithBreakpointSiteID(*thread_sp, bp_site_sp->GetID()));
+ handled = true;
if (!handled && signo && did_exec == false)
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 1532 bytes
Desc: not available
More information about the lldb-commits