[pups] Overlays?

[sms@moe.2bsd.com (Steven M. Schultz)]

HI!

> From: Jorgen Pehrson <jp at spektr.eu.org>
> I've just installed 2.11BSD on one of my PDPs, from a TK50 tape I just
> re-discovered in the trunk of my car. And now I'm in the process of

	Wow - those TK50s are quite robust to survive being stored in the
	truck of a car.  I've had audio tapes that did not survive ;)

> applying patches. When I installed patch 412 and started to rebuild the
> kernel, ld complains with a "ld: too big for type 431". 
> Does this means that I have to rearrange stuff between the BASE and the
> various OV entries?

	Indeed it does mean exactly that.

> How can I find out which overlay is too big? I've tried to do a:
> 138% root--> size unix.o

	Yep - that is the correct method.

> text    data    bss     dec     hex
> 52352   6928    37622   96902   17a86   total text: 115520
>         overlays: 7680,7232,7808,7744,4864,8576,4736,6848,7680
> 
> How big is too big? And also if none of the overlays above are too big, I
> guess it must be the BASE that is too big? 

	I thought I wrote this up at one time but I may have only thought
	about it ;)

> Is there some sort of documentation anywhere that describe this voodoo
> stuff and black magic a bit? 

	Is it in one an Appendix to the "Setup&Installation" document?   

	In the case above it is OV6 (overlays are numbered from 1) that is
	too big.

	The rules are:

		1) BASE can be 56kb (57344) max

		2) OVerlays can be 8kb (8192) max

		3) There can be no 0 length overlays (except for the very
		    last one).

> Any suggestions on what .o file I should move to what overlay?

	Look at the Makefile - you should see a line that starts "OV6=".  Do
	a "size" on the .o files listed for OV6.   Pick one that will fix
	elsewhere - I'd suggest OV5 since it is only 4864 bytes and has lots
	of room.  You could also move a .o file to the BASE since it has
	adequate room also.

	Steven Schultz
	sms at moe.2bsd.com

Received: (from major at localhost)
	by minnie.cs.adfa.edu.au (8.9.3/8.9.3) id JAA20356
	for pups-liszt; Tue, 1 Aug 2000 09:31:49 +1000 (EST)
	(envelope-from owner-pups at minnie.cs.adfa.edu.au)
Received: from henry.cs.adfa.edu.au (henry.cs.adfa.edu.au [131.236.21.158])
	by minnie.cs.adfa.edu.au (8.9.3/8.9.3) with ESMTP id JAA20352
	for <pups at minnie.cs.adfa.edu.au>; Tue, 1 Aug 2000 09:31:48 +1000 (EST)
	(envelope-from wkt at henry.cs.adfa.edu.au)
Received: (from wkt at localhost)
	by henry.cs.adfa.edu.au (8.9.3/8.9.3) id JAA18497
	for pups at minnie.cs.adfa.edu.au; Tue, 1 Aug 2000 09:29:49 +1000 (EST)
	(envelope-from wkt)
Received: from moe.2bsd.com (0 at MOE.2BSD.COM [206.139.202.200])
	by minnie.cs.adfa.edu.au (8.9.3/8.9.3) with ESMTP id DAA18526
	for <pups at minnie.cs.adfa.edu.au>; Tue, 1 Aug 2000 03:28:25 +1000 (EST)
	(envelope-from sms at moe.2bsd.com)
Received: (from sms at localhost)
	by moe.2bsd.com (8.9.3/8.9.3) id KAA02116;
	Mon, 31 Jul 2000 10:12:43 -0700 (PDT)
Date: Mon, 31 Jul 2000 10:12:43 -0700 (PDT)
From: "Steven M. Schultz" <sms@moe.2bsd.com>
Message-Id: <200007311712.KAA02116 at moe.2bsd.com>
To: pups at minnie.cs.adfa.edu.au, rblair at webteksdesign.com
Subject: Re: [pups] makesimtape.c
Sender: owner-pups at minnie.cs.adfa.edu.au
Precedence: bulk

Hi!

> From: "Ryan Blair" <rblair at webteksdesign.com>
> 
> While reading back through the mailing list messages, trying to find out why
> my tape images never seem to work, I came across a program that I cannot
> seem to find. It was mentioned lately as "makesimtape.c" but I cannot find
> it anywhere in the archives. Anybody have a lead on this?

	Yep - I have a lead (not too surprising since I wrote the program :))

	I think the program's present in the PUPS archive in the top of
	the 2.11 tree area but it may have been relocated or whatever over
	time.

	Here's the program - it should compile on just about anything that
	has the 'mtio' ioctl functions.   It is very similar to 'maketape'
	(intentional since makesimtape.c started out as a copy of maketape.c).

	Steven Schultz
	moe.2bsd.com


--------------------------
/*
 *	@(#)makesimtape.c	2.1 (2.11BSD) 1998/12/31
 *		Hacked 'maketape.c' to write a file in a format suitable for
 *		use with Bob Supnik's PDP-11 simulator (V2.3) emulated tape 
 *		driver.
 *
 * 	NOTE: a PDP-11 has to flip the shorts within the long when writing out
 *	      the record size.  Seems a PDP-11 is neither a little-endian
 *	      machine nor a big-endian one.
 */

#include <stdio.h>
#include <errno.h>
#include <fcntl.h>
#include <stdlib.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/uio.h>

#define MAXB 30

	char	buf[MAXB * 512];
	char	name[50];
	long	recsz, flipped, trl();
	int	blksz;
	int	mt, fd, cnt;
	struct	iovec	iovec[3];
	struct	iovec	tmark[2];
	void	usage();

main(argc, argv)
	int argc;
	char *argv[];
	{
	int i, j = 0, k = 0;
	long zero = 0;
	register char	*outfile = NULL, *infile = NULL;
	FILE *mf;
	struct	stat	st;

	while	((i = getopt(argc, argv, "i:o:")) != EOF)
		{
		switch	(i)
			{
			case	'o':
				outfile = optarg;
				break;
			case	'i':
				infile = optarg;
				break;
			default:
				usage();
				/* NOTREACHED */
			}
		}
	if	(!outfile || !infile)
		usage();
		/* NOTREACHED */
/*
 * Stat the outfile and make sure it either 1) Does not exist, or
 * 2) Exists but is a regular file.
*/
	if	(stat(outfile, &st) != -1 && !(S_ISREG(st.st_mode)))
		errx(1, "outfile must either not exist or be a regular file");
		/* NOTREACHED */

	mt = open(outfile, O_WRONLY|O_CREAT|O_TRUNC, 0600);
	if	(mt < 0)
		err(1, "Can not create %s", outfile);
		/* NOTREACHED */

	mf = fopen(infile, "r");
	if	(!mf)
		err(1, "Can not open %s", infile);
		/* NOTREACHED*/

	tmark[0].iov_len = sizeof (long);
	tmark[0].iov_base = (char *)&zero;

	while	(1)
		{
		if	((i = fscanf(mf, "%s %d", name, &blksz))== EOF)
			exit(0);
		if	(i != 2) {
			fprintf(stderr,"Help! Scanf didn't read 2 things (%d)\n", i);
			exit(1);
			}
		if	(blksz <= 0 || blksz > MAXB)
			{
			fprintf(stderr, "Block size %u is invalid\n", blksz);
			exit(1);
			}
		recsz = blksz * 512;	/* convert to bytes */
		iovec[0].iov_len = sizeof (recsz);
#ifdef	pdp11
		iovec[0].iov_base = (char *)&flipped;
#else
		iovec[0].iov_base = (char *)&recsz;
#endif
		iovec[1].iov_len = (int)recsz;
		iovec[1].iov_base = buf;
		iovec[2].iov_len =  iovec[0].iov_len;
		iovec[2].iov_base = iovec[0].iov_base;

		if	(strcmp(name, "*") == 0)
			{
			if	(writev(mt, tmark, 1) < 0)
				warn(1, "writev of pseudo tapemark failed");
			k++;
			continue;
			}
		fd = open(name, 0);
		if	(fd < 0)
			err(1, "Can't open %s for reading", name);
			/* NOTREACHED */
		printf("%s: block %d, file %d\n", name, j, k);

		/*
		 * we pad the last record with nulls
		 * (instead of the bell std. of padding with trash).
		 * this allows you to access text files on the
		 * tape without garbage at the end of the file.
		 * (note that there is no record length associated
		 *  with tape files)
		 */

		while	((cnt=read(fd, buf, (int)recsz)) == (int)recsz)
			{
			j++;
#ifdef	pdp11
			flipped = trl(recsz);
#endif
			if	(writev(mt, iovec, 3) < 0)
				err(1, "writev #1");
				/* NOTREACHED */
			}
		if	(cnt > 0)
			{
			j++;
			bzero(buf + cnt, (int)recsz - cnt);
#ifdef	pdp11
			flipped = trl(recsz);
#endif
			if	(writev(mt, iovec, 3) < 0)
				err(1, "writev #2");
				/* NOTREACHED */
			}
		close(fd);
		}
/*
 * Write two tape marks to simulate EOT
*/
	writev(mt, tmark, 1);
	writev(mt, tmark, 1);
	}

long
trl(l)
	long	l;
	{
	union	{
		long	l;
		short	s[2];
		} foo;
	register short	x;

	foo.l = l;
	x = foo.s[0];
	foo.s[0] = foo.s[1];
	foo.s[1] = x;
	return(foo.l);
	}

void
usage()
	{
	fprintf(stderr, "usage: makesimtape -o outfilefile -i inputfile\n");
	exit(1);
	}