.bpsync (Synchronize Threads at Breakpoint)

When a thread reaches a breakpoint, the .bpsync command freezes all other threads, until the thread to which the breakpoint applies has stepped through the breakpoint.

.bpsync 1
.bpsync 0
.bpsync 

Parameters

1
Causes all threads to freeze when one thread has reached a breakpoint. After the thread to which the breakpoint applies has stepped through the breakpoint, the other threads are unfrozen.

0
Allows other threads to continue executing when one thread has reached a breakpoint. This is the default behavior.

Environment

Item	Description
Modes	User mode only
Targets	Live, crash dump
Platforms	All

Remarks

With no parameters. the .bpsync command displays the current rule governing breakpoint synchronization behavior.

The .bpsync command applies both to software breakpoints (set with bp, bu, or bm) and to processor breakpoints (set with ba).

If there is a possibility of multiple threads running through the same code, the .bpsync 1 command can be useful for capturing all breakpoint occurrences. Without this command, a breakpoint occurrence could be missed because the first thread to reach the breakpoint always causes the debugger to temporarily remove the breakpoint. In the short period when the breakpoint is removed, other threads could reach the same place in the code and not trigger the breakpoint as intended.

The temporary removal of breakpoints is a normal aspect of debugger operation. When the target reaches a breakpoint and is resumed, the debugger has to remove the breakpoint temporarily so that the target can execute the real code. After the real instruction has been executed, the debugger reinserts the break. To do this, the debugger restores the code (or turns off data breaks), does a single-step, and then puts the break back.

Note that if you use .bpsync 1, there is a risk of deadlocks among the threads that have been frozen.