This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
[PATCH v2 08/17] btrace: lock-step
- From: Markus Metzger <markus dot t dot metzger at intel dot com>
- To: palves at redhat dot com
- Cc: gdb-patches at sourceware dot org
- Date: Fri, 11 Sep 2015 08:51:29 +0200
- Subject: [PATCH v2 08/17] btrace: lock-step
- Authentication-results: sourceware.org; auth=none
- References: <1441954298-25298-1-git-send-email-markus dot t dot metzger at intel dot com>
Record btrace's to_wait method picks a single thread to step. When passed
minus_one_ptid, it picks the current thread. All other threads remain where
they are.
Change this to step all resumed threads together, one step at a time, until
the first thread reports an event.
We do delay reporting NO_HISTORY events until there are no other events to
report to prevent threads at the end of their execution history from starving
other threads.
We keep threads at the end of their execution history moving and replaying
until we announce their stop in to_wait. This shouldn't really be user-visible
but its a detail worth mentioning.
Since record btrace's to_resume method also picks only a single thread to
resume, there shouldn't be a difference with the current all-stop.
With non-stop or all-stop on top of non-stop, we will see differences. The
behaviour should be more natural as we're moving all threads.
2015-09-11 Markus Metzger <markus.t.metzger@intel.com>
gdb/
* record-btrace.c: Include vec.h.
(record_btrace_find_thread_to_move): Removed.
(btrace_step_no_resumed, btrace_step_again)
(record_btrace_stop_replaying_at_end): New.
(record_btrace_cancel_resume): Call record_btrace_stop_replaying_at_end.
(record_btrace_single_step_forward): Remove calls to
record_btrace_stop_replaying.
(record_btrace_step_thread): Do only one step for BTHR_CONT and
BTHR_RCONT. Keep threads at the end of their history moving.
(record_btrace_wait): Call record_btrace_step_thread for all threads
until one reports an event. Call record_btrace_stop_replaying_at_end
for the eventing thread.
---
gdb/record-btrace.c | 252 ++++++++++++++++++++++++++++++++++++----------------
1 file changed, 175 insertions(+), 77 deletions(-)
diff --git a/gdb/record-btrace.c b/gdb/record-btrace.c
index fcd4351..bda9b7c 100644
--- a/gdb/record-btrace.c
+++ b/gdb/record-btrace.c
@@ -37,6 +37,7 @@
#include "infrun.h"
#include "event-loop.h"
#include "inf-loop.h"
+#include "vec.h"
/* The target_ops of record-btrace. */
static struct target_ops record_btrace_ops;
@@ -1838,6 +1839,26 @@ record_btrace_stop_replaying (struct thread_info *tp)
registers_changed_ptid (tp->ptid);
}
+/* Stop replaying TP if it is at the end of its execution history. */
+
+static void
+record_btrace_stop_replaying_at_end (struct thread_info *tp)
+{
+ struct btrace_insn_iterator *replay, end;
+ struct btrace_thread_info *btinfo;
+
+ btinfo = &tp->btrace;
+ replay = btinfo->replay;
+
+ if (replay == NULL)
+ return;
+
+ btrace_insn_end (&end, btinfo);
+
+ if (btrace_insn_cmp (replay, &end) == 0)
+ record_btrace_stop_replaying (tp);
+}
+
/* The to_resume method of target record-btrace. */
static void
@@ -1910,26 +1931,7 @@ record_btrace_cancel_resume (struct thread_info *tp)
btrace_thread_flag_to_str (flags));
tp->btrace.flags &= ~(BTHR_MOVE | BTHR_STOP);
-}
-
-/* Find a thread to move. */
-
-static struct thread_info *
-record_btrace_find_thread_to_move (ptid_t ptid)
-{
- struct thread_info *tp;
-
- /* First check the parameter thread. */
- tp = find_thread_ptid (ptid);
- if (tp != NULL && (tp->btrace.flags & (BTHR_MOVE | BTHR_STOP)) != 0)
- return tp;
-
- /* Otherwise, find one other thread that has been resumed. */
- ALL_NON_EXITED_THREADS (tp)
- if ((tp->btrace.flags & (BTHR_MOVE | BTHR_STOP)) != 0)
- return tp;
-
- return NULL;
+ record_btrace_stop_replaying_at_end (tp);
}
/* Return a target_waitstatus indicating that we ran out of history. */
@@ -1983,6 +1985,30 @@ btrace_step_spurious (void)
return status;
}
+/* Return a target_waitstatus indicating that the thread was not resumed. */
+
+static struct target_waitstatus
+btrace_step_no_resumed (void)
+{
+ struct target_waitstatus status;
+
+ status.kind = TARGET_WAITKIND_NO_RESUMED;
+
+ return status;
+}
+
+/* Return a target_waitstatus indicating that we should wait again. */
+
+static struct target_waitstatus
+btrace_step_again (void)
+{
+ struct target_waitstatus status;
+
+ status.kind = TARGET_WAITKIND_IGNORE;
+
+ return status;
+}
+
/* Clear the record histories. */
static void
@@ -2047,24 +2073,22 @@ record_btrace_single_step_forward (struct thread_info *tp)
{
unsigned int steps;
+ /* We will bail out here if we continue stepping after reaching the end
+ of the execution history. */
steps = btrace_insn_next (replay, 1);
if (steps == 0)
- {
- record_btrace_stop_replaying (tp);
- return btrace_step_no_history ();
- }
+ return btrace_step_no_history ();
}
while (btrace_insn_get (replay) == NULL);
/* Determine the end of the instruction trace. */
btrace_insn_end (&end, btinfo);
- /* We stop replaying if we reached the end of the trace. */
+ /* The execution trace contains (and ends with) the current instruction.
+ This instruction has not been executed, yet, so the trace really ends
+ one instruction earlier. */
if (btrace_insn_cmp (replay, &end) == 0)
- {
- record_btrace_stop_replaying (tp);
- return btrace_step_no_history ();
- }
+ return btrace_step_no_history ();
return btrace_step_spurious ();
}
@@ -2144,65 +2168,56 @@ record_btrace_step_thread (struct thread_info *tp)
case BTHR_STEP:
status = record_btrace_single_step_forward (tp);
if (status.kind != TARGET_WAITKIND_SPURIOUS)
- return status;
+ break;
return btrace_step_stopped ();
case BTHR_RSTEP:
status = record_btrace_single_step_backward (tp);
if (status.kind != TARGET_WAITKIND_SPURIOUS)
- return status;
+ break;
return btrace_step_stopped ();
case BTHR_CONT:
- for (;;)
- {
- status = record_btrace_single_step_forward (tp);
- if (status.kind != TARGET_WAITKIND_SPURIOUS)
- return status;
-
- if (btinfo->replay != NULL)
- {
- const struct btrace_insn *insn;
-
- insn = btrace_insn_get (btinfo->replay);
- gdb_assert (insn != NULL);
+ status = record_btrace_single_step_forward (tp);
+ if (status.kind != TARGET_WAITKIND_SPURIOUS)
+ break;
- DEBUG ("stepping %d (%s) ... %s", tp->num,
- target_pid_to_str (tp->ptid),
- core_addr_to_string_nz (insn->pc));
- }
- }
+ btinfo->flags |= flags;
+ return btrace_step_again ();
case BTHR_RCONT:
- for (;;)
- {
- const struct btrace_insn *insn;
-
- status = record_btrace_single_step_backward (tp);
- if (status.kind != TARGET_WAITKIND_SPURIOUS)
- return status;
+ status = record_btrace_single_step_backward (tp);
+ if (status.kind != TARGET_WAITKIND_SPURIOUS)
+ break;
- gdb_assert (btinfo->replay != NULL);
+ btinfo->flags |= flags;
+ return btrace_step_again ();
+ }
- insn = btrace_insn_get (btinfo->replay);
- gdb_assert (insn != NULL);
+ /* We keep threads moving at the end of their execution history. The to_wait
+ method will stop the thread for whom the event is reported. */
+ if (status.kind == TARGET_WAITKIND_NO_HISTORY)
+ btinfo->flags |= flags;
- DEBUG ("reverse-stepping %d (%s) ... %s", tp->num,
- target_pid_to_str (tp->ptid),
- core_addr_to_string_nz (insn->pc));
- }
- }
+ return status;
}
+/* A vector of threads. */
+
+typedef struct thread_info * tp_t;
+DEF_VEC_P (tp_t);
+
/* The to_wait method of target record-btrace. */
static ptid_t
record_btrace_wait (struct target_ops *ops, ptid_t ptid,
struct target_waitstatus *status, int options)
{
- struct thread_info *tp, *other;
+ VEC (tp_t) *moving, *no_history;
+ struct thread_info *tp, *eventing;
+ struct cleanup *cleanups = make_cleanup (null_cleanup, NULL);
DEBUG ("wait %s (0x%x)", target_pid_to_str (ptid), options);
@@ -2213,31 +2228,114 @@ record_btrace_wait (struct target_ops *ops, ptid_t ptid,
return ops->to_wait (ops, ptid, status, options);
}
- /* Let's find a thread to move. */
- tp = record_btrace_find_thread_to_move (ptid);
- if (tp == NULL)
+ moving = NULL;
+ no_history = NULL;
+
+ make_cleanup (VEC_cleanup (tp_t), &moving);
+ make_cleanup (VEC_cleanup (tp_t), &no_history);
+
+ /* Keep a work list of moving threads. */
+ ALL_NON_EXITED_THREADS (tp)
+ if (ptid_match (tp->ptid, ptid)
+ && ((tp->btrace.flags & (BTHR_MOVE | BTHR_STOP)) != 0))
+ VEC_safe_push (tp_t, moving, tp);
+
+ if (VEC_empty (tp_t, moving))
+ {
+ *status = btrace_step_no_resumed ();
+
+ DEBUG ("wait ended by %s: %s", target_pid_to_str (null_ptid),
+ target_waitstatus_to_string (status));
+
+ do_cleanups (cleanups);
+ return null_ptid;
+ }
+
+ /* Step moving threads one by one, one step each, until either one thread
+ reports an event or we run out of threads to step.
+
+ When stepping more than one thread, chances are that some threads reach
+ the end of their execution history earlier than others. If we reported
+ this immediately, all-stop on top of non-stop would stop all threads and
+ resume the same threads next time. And we would report the same thread
+ having reached the end of its execution history again.
+
+ In the worst case, this would starve the other threads. But even if other
+ threads would be allowed to make progress, this would result in far too
+ many intermediate stops.
+
+ We therefore delay the reporting of "no execution history" until we have
+ nothing else to report. By this time, all threads should have moved to
+ either the beginning or the end of their execution history. There will
+ be a single user-visible stop. */
+ eventing = NULL;
+ while ((eventing == NULL) && !VEC_empty (tp_t, moving))
+ {
+ unsigned int ix;
+
+ ix = 0;
+ while ((eventing == NULL) && VEC_iterate (tp_t, moving, ix, tp))
+ {
+ *status = record_btrace_step_thread (tp);
+
+ switch (status->kind)
+ {
+ case TARGET_WAITKIND_IGNORE:
+ ix++;
+ break;
+
+ case TARGET_WAITKIND_NO_HISTORY:
+ VEC_safe_push (tp_t, no_history,
+ VEC_ordered_remove (tp_t, moving, ix));
+ break;
+
+ default:
+ eventing = VEC_unordered_remove (tp_t, moving, ix);
+ break;
+ }
+ }
+ }
+
+ if (eventing == NULL)
{
- DEBUG ("wait %s: no thread", target_pid_to_str (ptid));
+ /* We started with at least one moving thread. This thread must have
+ either stopped or reached the end of its execution history.
- status->kind = TARGET_WAITKIND_IGNORE;
- return minus_one_ptid;
+ In the former case, EVENTING must not be NULL.
+ In the latter case, NO_HISTORY must not be empty. */
+ gdb_assert (!VEC_empty (tp_t, no_history));
+
+ /* We kept threads moving at the end of their execution history. Stop
+ EVENTING now that we are going to report its stop. */
+ eventing = VEC_unordered_remove (tp_t, no_history, 0);
+ eventing->btrace.flags &= ~BTHR_MOVE;
+
+ *status = btrace_step_no_history ();
}
- /* We only move a single thread. We're not able to correlate threads. */
- *status = record_btrace_step_thread (tp);
+ gdb_assert (eventing != NULL);
+
+ /* We kept threads replaying at the end of their execution history. Stop
+ replaying EVENTING now that we are going to report its stop. */
+ record_btrace_stop_replaying_at_end (eventing);
/* Stop all other threads. */
if (!target_is_non_stop_p ())
- ALL_NON_EXITED_THREADS (other)
- record_btrace_cancel_resume (other);
+ ALL_NON_EXITED_THREADS (tp)
+ record_btrace_cancel_resume (tp);
/* Start record histories anew from the current position. */
- record_btrace_clear_histories (&tp->btrace);
+ record_btrace_clear_histories (&eventing->btrace);
/* We moved the replay position but did not update registers. */
- registers_changed_ptid (tp->ptid);
+ registers_changed_ptid (eventing->ptid);
+
+ DEBUG ("wait ended by thread %d (%s): %s", eventing->num,
+ target_pid_to_str (eventing->ptid),
+ target_waitstatus_to_string (status));
- return tp->ptid;
+ do_cleanups (cleanups);
+ return eventing->ptid;
}
/* The to_stop method of target record-btrace. */
--
1.8.3.1