This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
[PATCH 2/2] Honour software single step in fallback of displaced stepping
- From: Yao Qi <qiyaoltc at gmail dot com>
- To: gdb-patches at sourceware dot org
- Date: Tue, 7 Apr 2015 16:52:05 +0100
- Subject: [PATCH 2/2] Honour software single step in fallback of displaced stepping
- Authentication-results: sourceware.org; auth=none
- References: <1428421925-18025-1-git-send-email-qiyaoltc at gmail dot com>
From: Yao Qi <yao.qi@linaro.org>
Hi,
When I run gdb.threads/non-stop-fair-events.exp on arm-linux target,
I see the following error in the log,
displaced: breakpoint is gone: Thread 22518, step(1)^M
Sending packet: $vCont;s:p57f3.57f6#9d...
gdb/gdbserver/linux-low.c:3686: A problem internal to GDBserver has been detected.^M
linux_resume_one_lwp_throw: Assertion `step == 0' failed.
GDB sends vCont;s by mistake, and GDBserver fails on assert. GDB
doesn't consider software single step in infrun.c:displaced_step_fixup,
/* Go back to what we were trying to do. */
step = currently_stepping (tp);
if (debug_displaced)
fprintf_unfiltered (gdb_stdlog,
"displaced: breakpoint is gone: %s, step(%d)\n",
target_pid_to_str (tp->ptid), step);
target_resume (ptid, step, GDB_SIGNAL_0);
The patch is to let GDB consider software single step here. It fixes
fails in gdb.threads/non-stop-fair-events.exp on arm.
gdb:
2015-04-02 Yao Qi <yao.qi@linaro.org>
* infrun.c (maybe_software_singlestep): Declare.
(displaced_step_fixup): Call maybe_software_singlestep.
---
gdb/infrun.c | 8 ++++++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/gdb/infrun.c b/gdb/infrun.c
index f5faa0a..f4bbf67 100644
--- a/gdb/infrun.c
+++ b/gdb/infrun.c
@@ -100,6 +100,8 @@ static void insert_step_resume_breakpoint_at_caller (struct frame_info *);
static void insert_longjmp_resume_breakpoint (struct gdbarch *, CORE_ADDR);
+static int maybe_software_singlestep (struct gdbarch *gdbarch, CORE_ADDR pc);
+
/* When set, stop the 'step' command if we enter a function which has
no line number information. The normal behavior is that we step
over such function. */
@@ -1847,6 +1849,7 @@ displaced_step_fixup (ptid_t event_ptid, enum gdb_signal signal)
regcache = get_thread_regcache (ptid);
actual_pc = regcache_read_pc (regcache);
aspace = get_regcache_aspace (regcache);
+ gdbarch = get_regcache_arch (regcache);
if (breakpoint_here_p (aspace, actual_pc))
{
@@ -1857,8 +1860,6 @@ displaced_step_fixup (ptid_t event_ptid, enum gdb_signal signal)
displaced_step_prepare (ptid);
- gdbarch = get_regcache_arch (regcache);
-
if (debug_displaced)
{
CORE_ADDR actual_pc = regcache_read_pc (regcache);
@@ -1891,6 +1892,9 @@ displaced_step_fixup (ptid_t event_ptid, enum gdb_signal signal)
/* Go back to what we were trying to do. */
step = currently_stepping (tp);
+ if (step)
+ step = maybe_software_singlestep (gdbarch, actual_pc);
+
if (debug_displaced)
fprintf_unfiltered (gdb_stdlog,
"displaced: breakpoint is gone: %s, step(%d)\n",
--
1.9.1