[rfc v2][1/6] Distinguish target and "fake" PID values

[rfc v2][1/6] Distinguish target and "fake" PID values

[Ulrich Weigand]

Hello,

this is basically Pedro's patch 03-process_id.diff from:
http://sourceware.org/ml/gdb-patches/2012-01/msg00353.html
to record in `struct inferior' whether the target process id
we have has been faked by GDB or not.

The only difference is that this patch contains no uses of
fake_pid_p yet, those will be introduced later in the series.

Bye,
Ulrich


From: Pedro Alves <palves@redhat.com>

ChangeLog:

	* inferior.h (struct inferior): Add fake_pid_p.
	* inferior.c (exit_inferior_1): Clear fake_pid_p.
	* remote.c (remote_start_remote): Set fake_pid_p if we have to use
	magic_null_ptid since the remote side doesn't provide a real PID.


Index: gdb-head/gdb/inferior.c
===================================================================
--- gdb-head.orig/gdb/inferior.c	2012-01-13 18:44:36.000000000 +0100
+++ gdb-head/gdb/inferior.c	2012-01-13 18:45:55.000000000 +0100
@@ -276,6 +276,7 @@ exit_inferior_1 (struct inferior *inftoe
   observer_notify_inferior_exit (inf);
 
   inf->pid = 0;
+  inf->fake_pid_p = 0;
   if (inf->vfork_parent != NULL)
     {
       inf->vfork_parent->vfork_child = NULL;
Index: gdb-head/gdb/inferior.h
===================================================================
--- gdb-head.orig/gdb/inferior.h	2012-01-13 18:44:36.000000000 +0100
+++ gdb-head/gdb/inferior.h	2012-01-13 18:45:55.000000000 +0100
@@ -421,6 +421,8 @@ struct inferior
   /* Actual target inferior id, usually, a process id.  This matches
      the ptid_t.pid member of threads of this inferior.  */
   int pid;
+  /* True if the PID was actually faked by GDB.  */
+  int fake_pid_p;
 
   /* State of GDB control of inferior process execution.
      See `struct inferior_control_state'.  */
Index: gdb-head/gdb/remote.c
===================================================================
--- gdb-head.orig/gdb/remote.c	2012-01-13 18:44:36.000000000 +0100
+++ gdb-head/gdb/remote.c	2012-01-13 18:45:55.000000000 +0100
@@ -3253,6 +3253,10 @@ remote_start_remote (int from_tty, struc
 
   if (!non_stop)
     {
+      ptid_t ptid;
+      int fake_pid_p = 0;
+      struct inferior *inf;
+
       if (rs->buf[0] == 'W' || rs->buf[0] == 'X')
 	{
 	  if (!extended_p)
@@ -3272,19 +3276,37 @@ remote_start_remote (int from_tty, struc
       /* Let the stub know that we want it to return the thread.  */
       set_continue_thread (minus_one_ptid);
 
-      /* Without this, some commands which require an active target
-	 (such as kill) won't work.  This variable serves (at least)
-	 double duty as both the pid of the target process (if it has
-	 such), and as a flag indicating that a target is active.
-	 These functions should be split out into seperate variables,
-	 especially since GDB will someday have a notion of debugging
-	 several processes.  */
-      inferior_ptid = magic_null_ptid;
+      inferior_ptid = minus_one_ptid;
 
       /* Now, if we have thread information, update inferior_ptid.  */
-      inferior_ptid = remote_current_thread (inferior_ptid);
+      ptid = remote_current_thread (inferior_ptid);
+      if (!ptid_equal (ptid, minus_one_ptid))
+	{
+	  if (ptid_get_pid (ptid) == -1)
+	    {
+	      ptid = ptid_build (ptid_get_pid (magic_null_ptid),
+				 ptid_get_lwp (ptid),
+				 ptid_get_tid (ptid));
+	      fake_pid_p = 1;
+	    }
+
+	  inferior_ptid = ptid;
+	}
+      else
+	{
+	  /* Without this, some commands which require an active
+	     target (such as kill) won't work.  This variable serves
+	     (at least) double duty as both the pid of the target
+	     process (if it has such), and as a flag indicating that a
+	     target is active.  These functions should be split out
+	     into seperate variables, especially since GDB will
+	     someday have a notion of debugging several processes.  */
+	  inferior_ptid = magic_null_ptid;
+	  fake_pid_p = 1;
+	}
 
-      remote_add_inferior (ptid_get_pid (inferior_ptid), -1);
+      inf = remote_add_inferior (ptid_get_pid (inferior_ptid), -1);
+      inf->fake_pid_p = fake_pid_p;
 
       /* Always add the main thread.  */
       add_thread_silent (inferior_ptid);
-- 
  Dr. Ulrich Weigand
  GNU Toolchain for Linux on System z and Cell BE
  Ulrich.Weigand@de.ibm.com

Re: [rfc v2][1/6] Distinguish target and "fake" PID values

[Pedro Alves]

On 01/13/2012 06:13 PM, Ulrich Weigand wrote:
> Hello,
> 
> this is basically Pedro's patch 03-process_id.diff from:
> http://sourceware.org/ml/gdb-patches/2012-01/msg00353.html
> to record in `struct inferior' whether the target process id
> we have has been faked by GDB or not.
> 
> The only difference is that this patch contains no uses of
> fake_pid_p yet, those will be introduced later in the series.

I remembered that extended_remote_create_inferior_1 (for "run")
will need the same treatment.  It's no biggie, since we'll be using
multiprocess extensions in gdbserver.  I'll fix it, but don't let
this block you (unless you want to fix it yourself).

-- 
Pedro Alves

Re: [rfc v2][1/6] Distinguish target and "fake" PID values

[Ulrich Weigand]

Pedro Alves wrote:
> On 01/13/2012 06:13 PM, Ulrich Weigand wrote:
> > Hello,
> > 
> > this is basically Pedro's patch 03-process_id.diff from:
> > http://sourceware.org/ml/gdb-patches/2012-01/msg00353.html
> > to record in `struct inferior' whether the target process id
> > we have has been faked by GDB or not.
> > 
> > The only difference is that this patch contains no uses of
> > fake_pid_p yet, those will be introduced later in the series.
> 
> I remembered that extended_remote_create_inferior_1 (for "run")
> will need the same treatment.  It's no biggie, since we'll be using
> multiprocess extensions in gdbserver.  I'll fix it, but don't let
> this block you (unless you want to fix it yourself).

OK, thanks!

Bye,
Ulrich

-- 
  Dr. Ulrich Weigand
  GNU Toolchain for Linux on System z and Cell BE
  Ulrich.Weigand@de.ibm.com