Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Pedro Alves <palves@redhat.com>
To: gdb-patches@sourceware.org
Subject: [PATCH] Don't switch to the event thread early in non-stop mode.
Date: Tue, 12 Jun 2012 18:44:00 -0000	[thread overview]
Message-ID: <20120612184440.29776.25769.stgit@brno.lan> (raw)

It is largely no longer necessary to special case non-stop mode, WRT
caring about the current context being set (the real "context
switching" context_switch used to exist for, is gone) before calling
handle_inferior event.  Trying to harmonize all-stop and non-stop
modes a bit further, this removes that non-stop special casing.  I
went through handle_inferior_event looking for places that miss
switching to the event thread and made them do so.  The missing
switching was never missed for single-process all-stop, but it's
important for both multi-process all-stop, and non-stop, so we don't
e.g., remove breakpoints in the context of a previously selected
inferior or thread that has already exited, or is presently running.
I actually was going to push a patch that centralized all the
context_switch calls at the top of handle_inferior_event (and I've
mentioned wanted to do it before a few times in the past years), but
in the end decided against, mostly because it felt counter to all the
parameterization in handle_inferior_event that moves away from global
state.

Tested on x86_64 Fedora 16, native and gdbserver, and applied.

2012-06-12  Pedro Alves  <palves@redhat.com>

	* infrun.c (infrun_thread_stop_requested_callback): Don't switch
	threads here.
	(prepare_for_detach): No longer context switch here in non-stop
	mode.
	(fetch_inferior_event): Ditto.
	(handle_inferior_event) <STOP_QUIETLY || NO_STOP_QUIETLY>: Switch
	to the event thread before removing breakpoints.  Switch to the
	event thread before inserting breakpoints and resuming.
	(handle_inferior_event) <TARGET_WAITKIND_SPURIOUS>: Switch to the
	event thread before resuming.
	(handle_inferior_event) <stepping_past_singlestep_breakpoint>:
	Switch to the event thread before removing breakpoints.
---
 gdb/infrun.c |   30 ++++++++++--------------------
 1 files changed, 10 insertions(+), 20 deletions(-)

diff --git a/gdb/infrun.c b/gdb/infrun.c
index b98e379..e36e42e 100644
--- a/gdb/infrun.c
+++ b/gdb/infrun.c
@@ -2437,8 +2437,6 @@ infrun_thread_stop_requested_callback (struct thread_info *info, void *arg)
 
       old_chain = make_cleanup_restore_current_thread ();
 
-      switch_to_thread (info->ptid);
-
       /* Go through handle_inferior_event/normal_stop, so we always
 	 have consistent output as if the stop event had been
 	 reported.  */
@@ -2655,14 +2653,6 @@ prepare_for_detach (void)
       old_chain_2 = make_cleanup (finish_thread_state_cleanup,
 				  &minus_one_ptid);
 
-      /* In non-stop mode, each thread is handled individually.
-	 Switch early, so the global state is set correctly for this
-	 thread.  */
-      if (non_stop
-	  && ecs->ws.kind != TARGET_WAITKIND_EXITED
-	  && ecs->ws.kind != TARGET_WAITKIND_SIGNALLED)
-	context_switch (ecs->ptid);
-
       /* Now figure out what to do with the result of the result.  */
       handle_inferior_event (ecs);
 
@@ -2789,16 +2779,6 @@ fetch_inferior_event (void *client_data)
   if (debug_infrun)
     print_target_wait_results (waiton_ptid, ecs->ptid, &ecs->ws);
 
-  if (non_stop
-      && ecs->ws.kind != TARGET_WAITKIND_IGNORE
-      && ecs->ws.kind != TARGET_WAITKIND_NO_RESUMED
-      && ecs->ws.kind != TARGET_WAITKIND_EXITED
-      && ecs->ws.kind != TARGET_WAITKIND_SIGNALLED)
-    /* In non-stop mode, each thread is handled individually.  Switch
-       early, so the global state is set correctly for this
-       thread.  */
-    context_switch (ecs->ptid);
-
   /* If an error happens while handling the event, propagate GDB's
      knowledge of the executing state to the frontend/user running
      state.  */
@@ -3380,6 +3360,9 @@ handle_inferior_event (struct execution_control_state *ecs)
 	 we're attaching or setting up a remote connection.  */
       if (stop_soon == STOP_QUIETLY || stop_soon == NO_STOP_QUIETLY)
 	{
+	  if (!ptid_equal (ecs->ptid, inferior_ptid))
+	    context_switch (ecs->ptid);
+
 	  /* Loading of shared libraries might have changed breakpoint
 	     addresses.  Make sure new breakpoints are inserted.  */
 	  if (stop_soon == NO_STOP_QUIETLY
@@ -3395,6 +3378,9 @@ handle_inferior_event (struct execution_control_state *ecs)
     case TARGET_WAITKIND_SPURIOUS:
       if (debug_infrun)
         fprintf_unfiltered (gdb_stdlog, "infrun: TARGET_WAITKIND_SPURIOUS\n");
+      if (!ptid_equal (ecs->ptid, inferior_ptid)
+	  && !ptid_equal (ecs->ptid, minus_one_ptid))
+	context_switch (ecs->ptid);
       resume (0, GDB_SIGNAL_0);
       prepare_to_wait (ecs);
       return;
@@ -3769,6 +3755,8 @@ handle_inferior_event (struct execution_control_state *ecs)
 				"infrun: stepping_past_"
 				"singlestep_breakpoint\n");
 	  /* Pull the single step breakpoints out of the target.  */
+	  if (!ptid_equal (ecs->ptid, inferior_ptid))
+	    context_switch (ecs->ptid);
 	  remove_single_step_breakpoints ();
 	  singlestep_breakpoints_inserted_p = 0;
 
@@ -3801,6 +3789,8 @@ handle_inferior_event (struct execution_control_state *ecs)
 	  /* Pull the single step breakpoints out of the target.  */
 	  if (singlestep_breakpoints_inserted_p)
 	    {
+	      if (!ptid_equal (ecs->ptid, inferior_ptid))
+		context_switch (ecs->ptid);
 	      remove_single_step_breakpoints ();
 	      singlestep_breakpoints_inserted_p = 0;
 	    }


             reply	other threads:[~2012-06-12 18:44 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-06-12 18:44 Pedro Alves [this message]
2012-06-13  1:06 ` Yao Qi
2012-06-13  8:52   ` Pedro Alves

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20120612184440.29776.25769.stgit@brno.lan \
    --to=palves@redhat.com \
    --cc=gdb-patches@sourceware.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox