[PATCH] Remove non-address bits for longjmp resume breakpoint

[PATCH] Remove non-address bits for longjmp resume breakpoint

[Andreas Arnez]

On 32-bit S390 targets the longjmp target address "naturally" has the
most significant bit set.  That bit indicates the addressing mode and
is not part of the address itself.  Thus, in analogy with similar
cases (like when computing the caller PC in
insert_step_resume_breakpoint_at_caller), this change removes
non-address bits from the longjmp target address before using it as a
breakpoint address.

Note that there are two ways for determining the longjmp target
address: via a probe or via a gdbarch method.  This change only
affects the probe method, because it is assumed that the address
returned by the gdbarch method is usable as-is.

This change was tested together with a patch that enables longjmp
probes in glibc for S/390:

  https://sourceware.org/ml/libc-alpha/2014-10/msg00277.html

gdb/ChangeLog:

	* gdb/infrun.c (process_event_stop_test): Apply
	gdbarch_addr_bits_remove to longjmp resume address.
---
 gdb/infrun.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/gdb/infrun.c b/gdb/infrun.c
index 4681175..42c40c4 100644
--- a/gdb/infrun.c
+++ b/gdb/infrun.c
@@ -4596,7 +4596,10 @@ process_event_stop_test (struct execution_control_state *ecs)
 	     is the third argument to the probe.  */
 	  arg_value = probe_safe_evaluate_at_pc (frame, 2);
 	  if (arg_value)
-	    jmp_buf_pc = value_as_address (arg_value);
+	    {
+	      jmp_buf_pc = value_as_address (arg_value);
+	      jmp_buf_pc = gdbarch_addr_bits_remove (gdbarch, jmp_buf_pc);
+	    }
 	  else if (!gdbarch_get_longjmp_target_p (gdbarch)
 		   || !gdbarch_get_longjmp_target (gdbarch,
 						   frame, &jmp_buf_pc))
-- 
1.8.4.2

Re: [PATCH] Remove non-address bits for longjmp resume breakpoint

[Ulrich Weigand]

Andreas Arnez wrote:

> On 32-bit S390 targets the longjmp target address "naturally" has the
> most significant bit set.  That bit indicates the addressing mode and
> is not part of the address itself.  Thus, in analogy with similar
> cases (like when computing the caller PC in
> insert_step_resume_breakpoint_at_caller), this change removes
> non-address bits from the longjmp target address before using it as a
> breakpoint address.

Makes sense.

> 	* gdb/infrun.c (process_event_stop_test): Apply
> 	gdbarch_addr_bits_remove to longjmp resume address.

This is OK.

Thanks,
Ulrich

-- 
  Dr. Ulrich Weigand
  GNU/Linux compilers and toolchain
  Ulrich.Weigand@de.ibm.com