Newsgroups: comp.sys.apple2.programmer
Path: news.weeg.uiowa.edu!news.uiowa.edu!uunet!math.fu-berlin.de!informatik.tu-muenchen.de!hphalle6!behrenss
From: behrenss@Informatik.TU-Muenchen.DE (Soenke Behrens)
Subject: Using SRAM while developing Firmware (a "guide")
Sender: news@Informatik.TU-Muenchen.DE (USENET Newssystem)
Organization: Technische Universitaet Muenchen, Germany
Date: Mon, 11 Oct 1993 10:51:28 GMT
Message-ID: <behrenss.750336688@hphalle6>
Lines: 206

Hallo,
 
here's the promised little "essay" on using SRAM instead of EPROM
while developing a card.
 
I assume that you have a pretty good knowledge of the GS's hardware
and bus, otherwise you wouldn't need this info.
 
It is really quite simple: An SRAM is pin-compatible with an EPROM of the
same size. Pin 27, however, is A14 (or /PGM for smaller sizes) on the EPROM
and /WE on the SRAM. Thus, to be able to use either EPROM or SRAM have
a jumper that connects Pin 27 to your EPROM bank-switching device in one
position and to a logic that serves /WE in the other.
This has one limitation: If you actually _use_ the space addressed through
A14, you will have to do some more work. In that case, connect the signal
from the bank-switching logic (A14) to Pin 1 of the SRAM. If you use
a 27C512, I can't help you ... there are no 64k*8 SRAMs to my knowledge.

As far as the /WE logic is concerned: All it has to do is to trap write
cycles to Cs00 and C800 and pull /WE low on these. This fits easily into
the GAL that already serves /OE, /CE and A8 through A10 on the EPROM.
If you do not use a GAL for that purpose, consider doing so. You can
do the same things with TTL logic, but it needs more space and chips and
is therefore more expensive.
 
Appended is the loader I use for my current hardware project. It's
an EXE that loads an SRAM in a slot with information from a file.
 
Assumptions this code makes:
- The C800 space in the EPROM (from 2nd 256-byte "page" on) is switched in
  at C900 through CFBF (it's that way on my card since it makes address
  decoding easier).
 
Limitations this code has:
- Bank switching on the EPROM has not been "implemented" since my firmware is
  so small right now it's not necessary.
 
These assumption and limitations are easily removed/replaced, though.
 
Have fun developing
 
Soenke

========== cut here to get C code ========================

#include <stdlib.h>
#include <stdio.h>
#include <assert.h>

int
main (int argc, char *argv[])
{
	size_t fsize;
	int slot;
	int i;
	FILE *fp;
	unsigned char *buffer;
	volatile unsigned char *writeto;
	unsigned char *freebuf;

	if (argc != 3)
	{
		fprintf(stderr,"Usage: %s <file> <slot>\n",argv[0]);
		exit (-1);
	}

	slot = atoi (argv[2]);

	if (slot < 1 || slot > 7)
	{
		fprintf(stderr,"%s: <slot> parameter was invalid. It has to"
			" be an integer between 1 and 7.\n",argv[0]);
		fprintf(stderr,"It was: %s\n",argv[1]);
		exit (-1);
	}

	fp = fopen(argv[1],"rb");
	if (fp == NULL)
	{
		fprintf(stderr,"%s: Unable to open file %s\n",argv[0],argv[1]);
		perror(argv[1]);
		exit (-1);
	}

	fseek(fp, 0, SEEK_END);
	fsize = ftell(fp);
	fseek(fp, 0, SEEK_SET);

/* Take out if you bank-switch !!! */

	if (fsize > 0x07BF)
	{
		fprintf(stderr,"I'm terribly sorry, this version of %s does"
			" not support bank-switched EPROMs. Your input\n"
			"file is too large.\n",argv[0]);
		if (fclose (fp) == EOF)
		{
			fprintf(stderr,"Also, the file %s could not be "
			"properly closed. Its contents may be lost\n",argv[1]);
		}
		free (buffer);
		exit (-1);
	}

	buffer = malloc (fsize);

	if (buffer == NULL)
	{
		fprintf(stderr,"%s: Unable to allocate memory for file"
			" buffer\n", argv[0]);
		exit (-1);
	}

	fread ((void *) buffer, 1UL, fsize, fp);

	if (ferror(fp))
	{
		fprintf(stderr,"%s: An error occured while trying to read"
			" from %s\n",argv[0],argv[1]);
		perror(argv[1]);
		if (fclose (fp) == EOF)
		{
			fprintf(stderr,"Also, the file could not be properly"
				" closed. Its contents may be lost\n");
		}
		exit (-1);
	}

	/*
	 * Now that I finally have the contents of my file, let's write them
	 * to the slot SRAM.
	 */

	/*
	 * First, check if the slot is switched in at all.
	 */

	{
		volatile unsigned char *slotreg = (unsigned char *) 0x00C02D;
		unsigned char mask = 0x01 << slot;

		if ((*slotreg & mask) == 0)
		{ /* Slot is set to internal */
			fprintf (stderr,"%s: Slot %d is set to internal."
				" Please change and reboot\n",argv[0],slot);
			if (fclose (fp) == EOF)
			{
				fprintf(stderr,"Also, the file %s could not"
				" be properly closed. Its contents may be"
				" lost\n",argv[1]);
			}
			free (buffer);
			exit (-1);
		}
	}

	/* It apparently is ... so let's continue with writing the stuff
	 * out actually.
	 */

	freebuf = buffer;

	/* Cs00 space first */
	
	writeto = (unsigned char *) 0x00C000 + (0x0100 * slot);

	for (i = 0; i < 0x100; i++)
	{ /* write out the first 256 bytes */
		*writeto++ = *buffer++;
	}

	/* Continue with C900 to CFBF space */

	writeto = (unsigned char *) 0x00C900;

	for (i = 0,fsize -= 0x100; i < fsize; i++)
	{
		assert (writeto < (unsigned char *) 0xCFC0);
		assert (writeto >= (unsigned char *) 0xC900);
		*writeto++ = *buffer++;
	}

	/* I'm done ... clean up and exit */

	buffer = freebuf;

	if (fclose (fp) == EOF)
	{
		fprintf(stderr,"%s: The file %s could not be properly"
			" closed. Its contents may be lost\n",argv[0],argv[1]);
		free (buffer);
		exit (-1);
	}

	free (buffer);

	printf("File successfully loaded into firmware space of slot %d\n",
		slot);
	return (0);
}

-- 
===========================================================================
More than silver, more than gold, I need   !    behrenss@informatik.
electricity. I need it for my dreams.      !    tu-muenchen.de
(RACTER)                                   !  FIDO: 2:246/13 Soenke Behrens
Newsgroups: comp.sys.apple2.programmer
Path: news.weeg.uiowa.edu!news.uiowa.edu!hobbes.physics.uiowa.edu!math.ohio-state.edu!cs.utexas.edu!uunet!math.fu-berlin.de!informatik.tu-muenchen.de!hphalle6!behrenss
From: behrenss@Informatik.TU-Muenchen.DE (Soenke Behrens)
Subject: Re: Using SRAM while developing Firmware (a "guide")
References: <behrenss.750336688@hphalle6>
Sender: news@Informatik.TU-Muenchen.DE (USENET Newssystem)
Organization: Technische Universitaet Muenchen, Germany
Date: Tue, 12 Oct 1993 00:43:29 GMT
Message-ID: <behrenss.750386609@hphalle6>
Lines: 65

behrenss@Informatik.TU-Muenchen.DE (Soenke Behrens) writes:

>Hallo,

Again, just goes to show how easy it is to mess up real bad
if you're not careful ... yeah right.

I forgot something important without which the whole scheme won't
work.



>	/* It apparently is ... so let's continue with writing the stuff
>	 * out actually.
>	 */

>	freebuf = buffer;

Insert some code here ....:

	asm
	{
		sei
		sta >$CFFE
	}

Hope that works now, haven't tested the sta yet actually ....

>	/* Cs00 space first */
>	
>	writeto = (unsigned char *) 0x00C000 + (0x0100 * slot);

>	for (i = 0; i < 0x100; i++)
>	{ /* write out the first 256 bytes */
>		*writeto++ = *buffer++;
>	}

>	/* Continue with C900 to CFBF space */

>	writeto = (unsigned char *) 0x00C900;

>	for (i = 0,fsize -= 0x100; i < fsize; i++)
>	{
>		assert (writeto < (unsigned char *) 0xCFC0);
>		assert (writeto >= (unsigned char *) 0xC900);
>		*writeto++ = *buffer++;
>	}

And here too:

	asm
	{
		cli
	}


Sorry for that. W/O disabling interrupts and switching in your ROM
space, you get into a real mess ....

Soenke
-- 
===========================================================================
More than silver, more than gold, I need   !    behrenss@informatik.
electricity. I need it for my dreams.      !    tu-muenchen.de
(RACTER)                                   !  FIDO: 2:246/13 Soenke Behrens