This is the mail archive of the
gdb-cvs@sourceware.org
mailing list for the GDB project.
[binutils-gdb] Misc switch_back_to_stepped_thread cleanups
- From: Pedro Alves <palves at sourceware dot org>
- To: gdb-cvs at sourceware dot org
- Date: 7 Aug 2015 16:45:19 -0000
- Subject: [binutils-gdb] Misc switch_back_to_stepped_thread cleanups
https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;h=1afd5965eda86caed3af7f23fd1f8802831360b6
commit 1afd5965eda86caed3af7f23fd1f8802831360b6
Author: Pedro Alves <palves@redhat.com>
Date: Fri Aug 7 17:23:58 2015 +0100
Misc switch_back_to_stepped_thread cleanups
Several misc cleanups that prepare the tail end of this function, the
part that actually re-resumes the stepped thread.
The most non-obvious would be the currently_stepping change, I guess.
That's because it isn't ever correct to pass step=1 to target_resume
on software single-step targets, and currently_stepping works at a
conceptual higher level, it returns step=true even on software step
targets. It doesn't really matter on hardware step targets, as the
breakpoint will be hit immediately, but it's just wrong on software
step targets. I tested it against my x86 software single-step branch,
and it indeed fixes failed assertions (that catch spurious
PTRACE_SINGLESTEP requests) there.
gdb/ChangeLog:
2015-08-07 Pedro Alves <palves@redhat.com>
* infrun.c (switch_back_to_stepped_thread): Use ecs->ptid instead
of inferior_ptid. If the stepped thread vanished, return 0
instead of resuming here. Use reset_ecs. Print the prev_pc and
the current stop_pc in log message. Clear trap_expected if the
thread advanced. Don't pass currently_stepping to
do_target_resume.
Diff:
---
gdb/ChangeLog | 9 +++++++++
gdb/infrun.c | 33 ++++++++++++++++++---------------
2 files changed, 27 insertions(+), 15 deletions(-)
diff --git a/gdb/ChangeLog b/gdb/ChangeLog
index 3a6668a..bd9f969 100644
--- a/gdb/ChangeLog
+++ b/gdb/ChangeLog
@@ -1,5 +1,14 @@
2015-08-07 Pedro Alves <palves@redhat.com>
+ * infrun.c (switch_back_to_stepped_thread): Use ecs->ptid instead
+ of inferior_ptid. If the stepped thread vanished, return 0
+ instead of resuming here. Use reset_ecs. Print the prev_pc and
+ the current stop_pc in log message. Clear trap_expected if the
+ thread advanced. Don't pass currently_stepping to
+ do_target_resume.
+
+2015-08-07 Pedro Alves <palves@redhat.com>
+
* gdbthread.h (struct thread_info) <prev_pc>: Extend comment.
* infrun.c (struct execution_control_state): Move higher up in the
file.
diff --git a/gdb/infrun.c b/gdb/infrun.c
index 94075ce..455fc2d 100644
--- a/gdb/infrun.c
+++ b/gdb/infrun.c
@@ -5717,7 +5717,7 @@ switch_back_to_stepped_thread (struct execution_control_state *ecs)
{
/* Ignore threads of processes we're not resuming. */
if (!sched_multi
- && ptid_get_pid (tp->ptid) != ptid_get_pid (inferior_ptid))
+ && ptid_get_pid (tp->ptid) != ptid_get_pid (ecs->ptid))
continue;
/* When stepping over a breakpoint, we lock all threads
@@ -5781,19 +5781,17 @@ switch_back_to_stepped_thread (struct execution_control_state *ecs)
"stepped thread, it has vanished\n");
delete_thread (tp->ptid);
- keep_going (ecs);
- return 1;
+ return 0;
}
if (debug_infrun)
fprintf_unfiltered (gdb_stdlog,
"infrun: switching back to stepped thread\n");
- ecs->event_thread = tp;
- ecs->ptid = tp->ptid;
- context_switch (ecs->ptid);
+ reset_ecs (ecs, tp);
+ switch_to_thread (tp->ptid);
- stop_pc = regcache_read_pc (get_thread_regcache (ecs->ptid));
+ stop_pc = regcache_read_pc (get_thread_regcache (tp->ptid));
frame = get_current_frame ();
gdbarch = get_frame_arch (frame);
@@ -5817,23 +5815,28 @@ switch_back_to_stepped_thread (struct execution_control_state *ecs)
if (debug_infrun)
fprintf_unfiltered (gdb_stdlog,
- "infrun: expected thread advanced also\n");
+ "infrun: expected thread advanced also "
+ "(%s -> %s)\n",
+ paddress (target_gdbarch (), tp->prev_pc),
+ paddress (target_gdbarch (), stop_pc));
/* Clear the info of the previous step-over, as it's no
- longer valid. It's what keep_going would do too, if
- we called it. Must do this before trying to insert
- the sss breakpoint, otherwise if we were previously
- trying to step over this exact address in another
- thread, the breakpoint ends up not installed. */
+ longer valid (if the thread was trying to step over a
+ breakpoint, it has already succeeded). It's what
+ keep_going would do too, if we called it. Do this
+ before trying to insert the sss breakpoint, otherwise
+ if we were previously trying to step over this exact
+ address in another thread, the breakpoint is
+ skipped. */
clear_step_over_info ();
+ tp->control.trap_expected = 0;
insert_single_step_breakpoint (get_frame_arch (frame),
get_frame_address_space (frame),
stop_pc);
resume_ptid = user_visible_resume_ptid (tp->control.stepping_command);
- do_target_resume (resume_ptid,
- currently_stepping (tp), GDB_SIGNAL_0);
+ do_target_resume (resume_ptid, 0, GDB_SIGNAL_0);
prepare_to_wait (ecs);
}
else