Re: [9fans] VGA and laptops

Re: [9fans] VGA and laptops

[jmk]

saroj@bear.com:
	Which laptops are safe bets as of today for installing Plan 9?
	I know that lab guys use Thinkpad 600E, but the successor series
	600X laptops have problems. (Some guy posted a message, but
	nobody replied to that message).

I believe we did get the 600X to work, we had the discussion offline.
However, I think even the 600X series has been superseded.

The most recent laptop we've used is the Toshiba Portégé 3440CT. It has
a couple of annoying features (no PS/2 port on the case, don't know how
to make all 4 mouse buttons on the keypad work) but is otherwise very
nice.

There is also the COMPAQ Armada E500, a more conventional 3-spindle
machine.

Re: [9fans] VGA and laptops

[Christopher Nielsen]

On Tue, 26 Sep 2000 jmk@plan9.bell-labs.com wrote:

> saroj@bear.com:
> 	Which laptops are safe bets as of today for installing Plan 9?
> 	I know that lab guys use Thinkpad 600E, but the successor series
> 	600X laptops have problems. (Some guy posted a message, but
> 	nobody replied to that message).
> 
> I believe we did get the 600X to work, we had the discussion offline.
> However, I think even the 600X series has been superseded.

We got it mostly working (thanks for all the help!) :-)

It just needs a few more tweaks that I never got around to finishing. Life
has been too demanding of my time; not enough time for the coding I'd
rather be doing.

-- 
Christopher Nielsen
cnielsen@pobox.com

Re: [9fans] VGA and laptops

[Latchesar Ionkov]

On Tue, Sep 26, 2000 at 09:23:01AM -0400, jmk@plan9.bell-labs.com said:
> saroj@bear.com:
> 	Which laptops are safe bets as of today for installing Plan 9?
> 	I know that lab guys use Thinkpad 600E, but the successor series
> 	600X laptops have problems. (Some guy posted a message, but
> 	nobody replied to that message).
> 
> I believe we did get the 600X to work, we had the discussion offline.
> However, I think even the 600X series has been superseded.
> 
> The most recent laptop we've used is the Toshiba Portégé 3440CT. It has
> a couple of annoying features (no PS/2 port on the case, don't know how
> to make all 4 mouse buttons on the keypad work) but is otherwise very
> nice.

Small buttons of the keypad mouse work as the wheel button of PS/2
Intellimouse. The mouse sends 4 instead of 3 bytes of info to the aux port.
The last byte contains 1 (0x01) or -1 (0xFF) if the left or right button is
pressed. The most annoying feature is that these buttons have autorepeat.
They send events every 150ms (400ms between first and second event) while
they are pressed.

I modified the kernel I am running to map these two buttons to button 2 and
created a program (stealing some code from pipefile.c) to replace button 2
presses that are within 150ms with one press and one release. Using the
small buttons as button 2 is still annoying but is better than nothing.

If my changes are of interest for anyone, I'll clean them up and post them.

Thanks,
	Lucho

Re: [9fans] VGA and laptops

[rob pike]

I would like to see the code that enables on the buttons on the intellimouse.

Thanks.

-rob

Re: [9fans] VGA and laptops

[Latchesar Ionkov]

[-- Attachment #1: Type: text/plain, Size: 362 bytes --]

On Fri, Sep 29, 2000 at 12:01:58AM -0400, rob pike said:
> I would like to see the code that enables on the buttons on the intellimouse.

This is the code that I wrote to make the small buttons of the mouse of
Toshiba Portege 3480CT to work as button 2. It is not tested anywhere else.

mouse.c.diff is patch for /sys/src/9/pc/mouse.c.

Thanks,
	Lucho

[-- Attachment #2: mouse.c.diff --]
[-- Type: text/plain, Size: 426 bytes --]

24a25
> static int intellimouse;
60c61
< 	static short msg[3];
---
> 	static short msg[4];
63c64
< 	int buttons, dx, dy;
---
> 	int buttons, dx, dy, dz;
72c73,76
< 	if(++nb == 3){
---
> 	if (++nb >= 3) {
> 		if ((intellimouse) && (nb < 4))
> 			return;
> 
79a84,91
> 
> 		if (intellimouse) {
> 			dz = (char) msg[3];
> 
> 			if (dz) 
> 				buttons |= 2;
> 		}
> 
104d115
< static int intellimouse;

[-- Attachment #3: tms.c --]
[-- Type: text/plain, Size: 2392 bytes --]

#include <u.h>
#include <libc.h>
#include <thread.h>

#define TEMP "/n/temp"
#define STACKSIZE 4096

struct {
	QLock;
	int mfd;
	int rfd, wfd;
	QLock prs;
	int pressed;
	int b;
	int x;
	int y;
	long t;
	vlong nsec;
} b2;

static void tproc(void *);
static void wproc(void *);

void 
threadmain(int argc, char *argv[])
{
	char *file = "/dev/mouse";
	int n, b, x, y;
	long t;
	char buf[50];

	b2.mfd = open("/dev/mouse", ORDWR|OCEXEC);
	if (b2.mfd < 0)
		sysfatal("open /dev/mouse: %r");

	if (bind("#|", TEMP, MREPL) < 0)
		sysfatal("bind pipe %s: %r", TEMP);

	if (bind(TEMP "/data", "/dev/mouse", MREPL) < 0)
		sysfatal("bind %s /dev/mouse: %r", TEMP "/data");

	b2.rfd = open(TEMP "/data1", OREAD);
	if (b2.rfd < 0)
		sysfatal("open %s: %r", TEMP "/data1");

	b2.wfd = open(TEMP "/data1", OWRITE);
	if (b2.wfd < 0) 
		sysfatal("open %s: %r", TEMP "/data1");

	qlock(&b2.prs);

	proccreate(wproc, 0, STACKSIZE);
	proccreate(tproc, 0, STACKSIZE);

	while (1) {
		n = read(b2.mfd, buf, 49);
		x = atoi(buf + 1 + 0*12);
		y = atoi(buf + 1 + 1*12);
		b = atoi(buf + 1 + 2*12);
		t = atoi(buf + 1 + 3*12);

		qlock(&b2);
		if (b & 0x02) {
			if (b2.pressed == 0) {
				qunlock(&b2.prs);
				b2.pressed = 1;
			}

			b2.t = t;
			b2.x = x;
			b2.y = y;
			b2.b = b;
			b2.nsec = nsec();
		}

		if (b2.pressed)
			b |= 0x02;

		sprint(buf, "m%11d %11d %11d %11lud", x, y, b, t);
		write(b2.wfd, buf, 49);
		qunlock(&b2);
	}	
}

static void
wproc(void *)
{
	int n;
	char buf[50];

	while (1) {
		n = read(b2.rfd, buf, sizeof(buf));
		if (n < 0)
			break;

		write(b2.mfd, buf, n);
	}
}

static void
tproc(void *)
{
	vlong t1, t2;
	char buf[50];

	while (1) {
		qlock(&b2.prs);

		qlock(&b2);
		sprint(buf, "m%11d %11d %11d %11lud", b2.x, b2.y, b2.b | 0x02, b2.t);
		write(b2.wfd, buf, 49);
		qunlock(&b2);

		/* delay between first and second event (400ms) */
		sleep(400);

		qlock(&b2);
		b2.nsec = nsec();
		qunlock(&b2);

		while (1) {
			/* delay between events (150ms) */
			sleep(80);
			qlock(&b2);
			t2 = nsec();
			if ((t2 - b2.nsec) > 150000000L) {
				b2.nsec = t2;
				b2.pressed = 0;
				sprint(buf, "m%11d %11d %11d %11lud", b2.x, b2.y, b2.b & ~0x02, b2.t);
				write(b2.wfd, buf, 49);
				qunlock(&b2);
				break;
			}
			qunlock(&b2);
		}
	}
}

Re: [9fans] VGA and laptops

[jmk]

	On Tue, Sep 26, 2000 at 09:25:45AM -0400, jmk@plan9.bell-labs.com said:
	> Oops. Another problem with the Toshiba Portégé 3440CT is it has a USB
	> floppy so booting the Plan 9 distribution 'straight' would be problematic.

	What is the best way to install it in this case? My lack of knowledge forced
	me to do some ugly hacks in the kernel and format command to do the job. I
	suspect there is much easier way.

		Lucho

It's all ugly. Because the BIOS can make the USB floppy available via
the INT13 interface I made a version of the bootstrap that had the necessary
plan9.ini info wired in (i.e. which ether to use). Once that's loaded
it doesn't need the floppy any more and can boot a kernel over the ethernet.
After that it's easy (provided you have a Plan 9 fileserver to talk to).

--jim

Re: [9fans] VGA and laptops

[nigel]

> I do have a question though, wasn't VESA compliance supposed to do away with
> these problems

It does, if you feel like you can put your trust in someone's BIOS.... Making a call
into the BIOS is hairy in that you have no way to make it adhere to the particular
set of rules which your operating system makes on drivers, e.g. time spent with
interrupts off, how long spent spinning waiting for VBE etc..

As a result, one can imagine using it as a means to get the right screen mode at
the start, but you would never think of making calls to it for accelerated drawing.
So you can get a frame buffer and then bit bash, which I accept is a whole lot
better than nothing.

However, if you want an accelerated driver, then there are typically interrelationships
between the screen mode, and the blitter, so VESA is not much help. Putting it
another way, to be able to accelerate a chipset, you would probably need
to understand every setting that VESA makes; if so, then why use it?

VNH's native Inferno uses VESA calls to initialise the screen mode.

Re: [9fans] VGA and laptops

[nigel]

[-- Attachment #1: Type: text/plain, Size: 473 bytes --]

I'm pretty sure I got a cl-gd7543 going once before by simply adding the
right chip magic number. It was on a Texas Instruments laptop of about
3-4 years ago vintage. The gl7543 is the Alpine I think. I had several of
the manuals somewhere, but threw them away just after selling the laptop.

If all you have to do is add a chip number then, I'm sorry, it might feel like
it, but you're not in VGA Hell! Just somewhere nearby, where the weather is
more pleasant.


[-- Attachment #2: Type: message/rfc822, Size: 2033 bytes --]

From: "Matt" <matt@proweb.co.uk>
To: <9fans@cse.psu.edu>
Subject: Re: [9fans] VGA and laptops
Date: Tue, 26 Sep 2000 19:05:54 +0100
Message-ID: <007d01c027e4$6a6b0840$0301a8c0@freeze>

dont get a compaq lte 5300

:)

it's got a floppy, oh yes

but vga hell

Re: [9fans] VGA and laptops

[Matt]

> I'm pretty sure I got a cl-gd7543 going once before by simply adding the
> right chip magic number. It was on a Texas Instruments laptop of about
> 3-4 years ago vintage. The gl7543 is the Alpine I think. I had several of
> the manuals somewhere, but threw them away just after selling the laptop.

well the pdf for the chipset is on the Cirrus site
I've got a copy of it at :
http://www.proweb.co.uk/~matt/plan9/gd7543db.pdf

the magic number reported by aux/vga is 0x30

I have no means of sorting it out myself

> If all you have to do is add a chip number then, I'm sorry, it might feel
like
> it, but you're not in VGA Hell! Just somewhere nearby, where the weather
is
> more pleasant.

ooh I might get a slight tan. Is being closer more or less hellish.
I bought the thing just for plan9 too.
it's got the right soundcard
the right pcmcia controller


jmc has been helping me

he's already had to make me a custom .9gz file for the install
cos the laptop thinks it's having sanity problems which may or may not be
true

quoting jim

"it's broken in the sense that it persistently reports there is a hardware
error. from the earlier trace MCA 0011c160 MCT 00000009
means there was probably a parity error on either a bus or memory
transaction "


I do have a question though, wasn't VESA compliance supposed to do away with
these problems
I downloaded the QNX demo disk and that gave me a choice of 5 or 6 screen
modes which all worked.


I don't pretend to know anything about VESA compliance other than I remember
it bringing true colour cards to the early windows world and helped some
game writers get around the drivers problem.

Matt

Re: [9fans] VGA and laptops

[jmk]

Oops. Another problem with the Toshiba Portégé 3440CT is it has a USB
floppy so booting the Plan 9 distribution 'straight' would be problematic.

Re: [9fans] VGA and laptops

[Latchesar Ionkov]

On Tue, Sep 26, 2000 at 09:25:45AM -0400, jmk@plan9.bell-labs.com said:
> Oops. Another problem with the Toshiba Portégé 3440CT is it has a USB
> floppy so booting the Plan 9 distribution 'straight' would be problematic.

What is the best way to install it in this case? My lack of knowledge forced
me to do some ugly hacks in the kernel and format command to do the job. I
suspect there is much easier way.

	Lucho

Re: [9fans] VGA and laptops

[Matt]

dont get a compaq lte 5300

:)

it's got a floppy, oh yes

but vga hell

[9fans] VGA and laptops

[saroj]

Hi,

Which laptops are safe bets as of today for installing Plan 9?
I know that lab guys use Thinkpad 600E, but the successor series
600X laptops have problems. (Some guy posted a message, but
nobody replied to that message).

Thank you,
Saroj Mahapatra


Sent via Deja.com http://www.deja.com/
Before you buy.

Re: [9fans] VGA and laptops

[Latchesar Ionkov]

On Tue, Sep 26, 2000 at 09:13:59AM +0000, saroj@bear.com said:
> Hi,
> 
> Which laptops are safe bets as of today for installing Plan 9?
> I know that lab guys use Thinkpad 600E, but the successor series
> 600X laptops have problems. (Some guy posted a message, but
> nobody replied to that message).

Plan9 works on Toshiba 3480CT (S3 Savage IX-MX chip), though there are some
problems with suspend/resume. You'll have tough time installing it on the
laptop though (only USB floppy supported).

	Lucho

Re: [9fans] VGA and laptops

[Axel Belinfante]

Toshiba Tecra 8000 seems to work ok.
(apart for the audio -- it has a yamaha OPL3-SA3 which according to the
 documentation of the tecra should be soundblaster pro compatible;
 a minor amount of fiddling was not enough to get it working)
(on a off topic note: is the xircom rem56g-10bt pcmcia ethernet/modem
 supported? If so, what are the magic config words to make it work?)

Regards,
Axel.


> Which laptops are safe bets as of today for installing Plan 9?
> I know that lab guys use Thinkpad 600E, but the successor series
> 600X laptops have problems. (Some guy posted a message, but
> nobody replied to that message).