[PATCH] Don't set random_signal for single step breakpoint

[PATCH] Don't set random_signal for single step breakpoint

[Yao Qi]

I happen to see the code, and find the comments are out of date,
because single-step breakpoint is not transparent to the breakpoint
module.  If one thread hits another thread's single-step breakpoint,
random_signal should be zero.  IOW, if random_signal is one, the
thread shouldn't hit any single-step breakpoints.  This patch is to
remove this piece of code.

gdb:

2016-09-26  Yao Qi  <yao.qi@linaro.org>

	* infrun.c (handle_signal_stop): Don't set random_signal
	by ecs->hit_singlestep_breakpoint.
---
 gdb/infrun.c | 6 ------
 1 file changed, 6 deletions(-)

diff --git a/gdb/infrun.c b/gdb/infrun.c
index 648e2b6..b133f04 100644
--- a/gdb/infrun.c
+++ b/gdb/infrun.c
@@ -6056,12 +6056,6 @@ handle_signal_stop (struct execution_control_state *ecs)
     random_signal = !(ecs->event_thread->suspend.stop_signal == GDB_SIGNAL_TRAP
 		      && currently_stepping (ecs->event_thread));
 
-  /* Perhaps the thread hit a single-step breakpoint of _another_
-     thread.  Single-step breakpoints are transparent to the
-     breakpoints module.  */
-  if (random_signal)
-    random_signal = !ecs->hit_singlestep_breakpoint;
-
   /* No?  Perhaps we got a moribund watchpoint.  */
   if (random_signal)
     random_signal = !stopped_by_watchpoint;
-- 
1.9.1

Re: [PATCH] Don't set random_signal for single step breakpoint

[Pedro Alves]

On 09/26/2016 01:01 AM, Yao Qi wrote:
> I happen to see the code, and find the comments are out of date,
> because single-step breakpoint is not transparent to the breakpoint
> module.  If one thread hits another thread's single-step breakpoint,
> random_signal should be zero.  IOW, if random_signal is one, the
> thread shouldn't hit any single-step breakpoints.  This patch is to
> remove this piece of code.

Hmm, not sure.  This hunk does predate the move to have single-step
breakpoints be regular breakpoints, but, OTOH, the single-step breakpoints
are deleted before the bpstat handling takes place:

  /* Pull the single step breakpoints out of the target.  */
  if (ecs->event_thread->suspend.stop_signal == GDB_SIGNAL_TRAP)
    {
...
     /* However, before doing so, if this single-step breakpoint was
	 actually for another thread, set this thread up for moving
	 past it.  */
      if (!thread_has_single_step_breakpoint_here (ecs->event_thread,
						   aspace, pc))
	{
	  if (single_step_breakpoint_inserted_here_p (aspace, pc))
	    {
	      ecs->hit_singlestep_breakpoint = 1;
...
  delete_just_stopped_threads_single_step_breakpoints ();


  bpstat_clear (&ecs->event_thread->control.stop_bpstat);

...
  ecs->event_thread->control.stop_bpstat
    = bpstat_stop_status (get_regcache_aspace (get_current_regcache ()),
			  stop_pc, ecs->ptid, &ecs->ws);

  /* See if the breakpoints module can explain the signal.  */
  random_signal
    = !bpstat_explains_signal (ecs->event_thread->control.stop_bpstat,
			       ecs->event_thread->suspend.stop_signal);


So I wonder whether this:

  /* Maybe this was a trap for a software breakpoint that has since
     been removed.  */
  if (random_signal && target_stopped_by_sw_breakpoint ())
    {

is masking the need for the hunk you're proposing to remove.
What happens if you hack your target to no longer support
target_stopped_by_sw_breakpoint ?

The commit that added the code in question (2adfaa28b5,
eliminate thread-hop code) predates that (the whole
moribund breakpoints -> siginfo.si_code change).

Thanks,
Pedro Alves