Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!mips!dimacs.rutgers.edu!rutgers!cmcl2!uupsi!sunic!ugle.unit.no!hanche
From: hanche@imf.unit.no (Harald Hanche-Olsen)
Newsgroups: comp.sys.apollo
Subject: MITs Xapollo (R4) server in gpr mode
Message-ID: <HANCHE.91Mar6215757@hufsa.imf.unit.no>
Date: 6 Mar 91 20:57:57 GMT
Sender: news@ugle.unit.no
Organization: The Norwegian Institute of Technology, Trondheim, Norway.
Lines: 81

In another post, I talked about our problems with the X11R4 Xapollo
server, as compiled from the standared sources and run in mfb mode.
Now I want to talk about problems with the same server in gpr mode:
It seems to work ok except for certain startup problems.

At the heart of the problem are three (undocumented?) calls
gpr_$init_wm(...), gpr_$$inq_rectangle_id(...), aws_$init(...) where
the latter returns the status code:

06090001 (not an resource ID (graphics/IDM/RM resource manager))

In the log file, I find this

----------------
apUnitScreenInit: aws_$init failed. 06090001

Fatal server bug!
no screens found
----------------

After this Xapollo exits, xdm starts a new one, and the process
repeats a number of times -- maybe zero, frequently two or three
times, occasionally up to ten times -- until the call succeeds and the
server comes up.  The relevant code from
mit/server/ddx/apollo/ap_init.c looks like this (with some omissions):

----------------
static Bool
apUnitScreenInit(screenIndex, pScreen, argc, argv)
    int         screenIndex;
    ScreenPtr   pScreen;
    int         argc;   /* these two may NOT be changed */
    char      **argv;
{

        aws_$display_t      scr;

        scr.unit = 0;
        scr.screen = pDisp->display_unit - 1;

/* Initializing a bunch of of data structures, deleted */

/*=>*/  gpr_$init_wm (gpr_$borrow, (short)(pDisp->display_unit),
                      disp, hi_plane, &pDisp->display_bitmap, &status);

        apXSolo = (status.all == status_$ok);
        if (apXSolo)
        {
            gpr_$rect_id_t  rectID;
            
            /* Get rectangle ID from gpr bitmap for root rectangle */
/*=>*/      rectID = gpr_$$inq_rectangle_id (&pDisp->display_bitmap, &status);
            if (status.all != status_$ok)
            {
                xfree (pDisp);
                return FALSE;
            }
            
            /* Init aws */
/*=>*/      (void) aws_$init (scr, rectID, &status);
            if (status.all != status_$ok)
            {
                ErrorF("apUnitScreenInit: aws_$init failed. %08x\n", status.all);
----------------

and this is where we get the error message from.

Now, I cannot find any documentation for the three calls marked /*=>*/ above,
but it seems apparent that aws_$init will not accept the rectID
returned by gpr_$$inq_rectangle_id just before!?

Can anybody tell me where those calls are documented?  Why do they
sometimes fail and sometimes succeed?  If that is the way it has to
be, maybe I could let Xapollo loop on these calls until success,
rather than exiting and forcing xdm to start another Xapollo?  So many
questions - and I have none of the answers.  Suggestions are welcome.

- Harald Hanche-Olsen <hanche@imf.unit.no>
  Division of Mathematical Sciences
  The Norwegian Institute of Technology
  N-7034 Trondheim, NORWAY