Re: [RFC] "single step" atomic instruction sequences as a whole on PPC

["Ulrich Weigand" <uweigand@de.ibm.com>]

Hello Emi-san,

> When that multiple SIGTRAP events occured, GDB selects one event and
> cancels the other if the cause of SIGTRAP is a breakpoint hit, or just
> leave it pended.  The procedure will be done by `cancel_breakpoints_callback' 
> in linux-nat.c.  And the pended events will be deteced and noticed
> when the next time the target resumes.  
> 
> The probrem is that GDB doesn't check if the breakpoint is inserted
> for software single stepping when cancelling the trap event: when the
> event occured by a software single step breakpoint is not selected,
> GDB would not cancel it but leave it pended.  

This is cancel_breakpoints_callback in linux-nat.c, right?
 
> When the next time the target resumes, GDB restores the pended event.
> But if you have removed the watchpoint that the target get stopped by
> before resuming, GDB can never decide the cause of SIGTRAP anymore.
> The session log above shows the phenomenon.  

This looks just like a problem we fixed for the combined debugger;
cancel_breakpoints_callback should cancel SIGTRAP events caused by
software single-step breakpoints just the same as those caused by
other breakpoints.

Can you try whether the following patch fixes the problem for you?

Bye,
Ulrich


ChangeLog:

	* breakpoint.c (single_step_breakpoint_inserted_here_p): New function.
	* breakpoint.h (single_step_breakpoint_inserted_here_p): New prototype.
	* linux-nat.c (cancel_breakpoints_callback): Treat software single-
	step breakpoints the same way as regular software breakpoints.

diff -urNp gdb-orig/gdb/breakpoint.c gdb-head/gdb/breakpoint.c
--- gdb-orig/gdb/breakpoint.c	2007-05-08 14:37:32.000000000 +0200
+++ gdb-head/gdb/breakpoint.c	2007-05-10 22:34:36.776868666 +0200
@@ -7951,6 +7951,21 @@ remove_single_step_breakpoints (void)
     }
 }
 
+int
+single_step_breakpoint_inserted_here_p (CORE_ADDR pc)
+{
+  int i;
+
+  for (i = 0; i < 2; i++)
+    {
+      struct bp_target_info *bp_tgt = single_step_breakpoints[i];
+      if (bp_tgt && bp_tgt->placed_address == pc)
+	return 1;
+    }
+
+  return 0;
+}
+
 \f
 /* This help string is used for the break, hbreak, tbreak and thbreak commands.
    It is defined as a macro to prevent duplication.
diff -urNp gdb-orig/gdb/breakpoint.h gdb-head/gdb/breakpoint.h
--- gdb-orig/gdb/breakpoint.h	2007-04-13 19:13:45.000000000 +0200
+++ gdb-head/gdb/breakpoint.h	2007-05-10 22:34:36.829861036 +0200
@@ -832,6 +832,8 @@ extern int remove_hw_watchpoints (void);
    twice before remove is called.  */
 extern void insert_single_step_breakpoint (CORE_ADDR);
 extern void remove_single_step_breakpoints (void);
+extern int single_step_breakpoint_inserted_here_p (CORE_ADDR pc);
+
 
 /* Manage manual breakpoints, separate from the normal chain of
    breakpoints.  These functions are used in murky target-specific
diff -urNp gdb-orig/gdb/linux-nat.c gdb-head/gdb/linux-nat.c
--- gdb-orig/gdb/linux-nat.c	2007-05-07 02:18:12.000000000 +0200
+++ gdb-head/gdb/linux-nat.c	2007-05-10 22:34:36.860856573 +0200
@@ -1765,21 +1765,26 @@ cancel_breakpoints_callback (struct lwp_
      tripped on it.  */
 
   if (lp->status != 0
-      && WIFSTOPPED (lp->status) && WSTOPSIG (lp->status) == SIGTRAP
-      && breakpoint_inserted_here_p (read_pc_pid (lp->ptid) -
-				     DECR_PC_AFTER_BREAK))
+      && WIFSTOPPED (lp->status) && WSTOPSIG (lp->status) == SIGTRAP)
     {
-      if (debug_linux_nat)
-	fprintf_unfiltered (gdb_stdlog,
-			    "CBC: Push back breakpoint for %s\n",
-			    target_pid_to_str (lp->ptid));
+      CORE_ADDR break_pc;
+      break_pc = read_pc_pid (lp->ptid) - DECR_PC_AFTER_BREAK;
 
-      /* Back up the PC if necessary.  */
-      if (DECR_PC_AFTER_BREAK)
-	write_pc_pid (read_pc_pid (lp->ptid) - DECR_PC_AFTER_BREAK, lp->ptid);
+      if (breakpoint_inserted_here_p (break_pc)
+	  || single_step_breakpoint_inserted_here_p (break_pc))
+	{
+	  if (debug_linux_nat)
+	    fprintf_unfiltered (gdb_stdlog,
+				"CBC: Push back breakpoint for %s\n",
+				target_pid_to_str (lp->ptid));
 
-      /* Throw away the SIGTRAP.  */
-      lp->status = 0;
+	  /* Back up the PC if necessary.  */
+	  if (DECR_PC_AFTER_BREAK)
+	    write_pc_pid (break_pc, lp->ptid);
+
+	  /* Throw away the SIGTRAP.  */
+	  lp->status = 0;
+	}
     }
 
   return 0;


-- 
  Dr. Ulrich Weigand
  GNU Toolchain for Linux on System z and Cell BE
  Ulrich.Weigand@de.ibm.com