networkstation-developers — Porting Linux and related applications to the NetworkStation

[nwst-dev] Audio driver again

[Santiago G. M. <ma...@ay...>]

This morning I got it to load with the help of Jochen patch, so the driver
now loads ok with the line I had put here on my previous message.

I have run aumix and seems to work, one weird thing I saw with aumix was
that there was no main mixer (Volume they call it) but I had the synth, pcm,
... maybe it is like that, maybe it needs working :-?

The problem however is that if you try to play digital audio (mp3 with
mpg123 for example) the machine completely hangs, if any of you has any
ideas on how to test this, where to look for this info, ... it will be
welcome.

As soon as we get this working I'd say that the other thing we are missing
are more colours (initializing the S3) and we would mostly have it, needing
only some more polishing ;-)

Regards!

Re: [nwst-dev] Audio on the NS1000 locks (help)

[Jochen R. <jo...@pa...>]

> 
> I have traced the problem when issuing this line:
> 
> modprobe cs4232 io=0x830 irq=10 dma=0 dma2=1
> 
> till this line of code:
> 
> if (request_irq(devc->irq, adintr, 0, devc->name, (void *)my_dev) < 0)
>
> It doesn't come back from the request_irq call, as it starts executing
> adintr again and again in a loop without stopping.
>

Perhaps the interrupt controller is configured for level-sensitive (PCI)
interrupt on irq 10, instead of edge-triggered (ISA).

In arch/ppc/kernel/prep_pci.c, change line 962:

                        irq_edge_mask_hi = 0xA4; /* irq's 10, 13, 15 level-triggered */

to this:

#ifdef CONFIG_NS1000
                        irq_edge_mask_hi = 0xA0; /* irq's 13, 15 level-triggered */
#else
                        irq_edge_mask_hi = 0xA4; /* irq's 10, 13, 15 level-triggered */
#endif

Hope this works.

I will try to put in a dynamic detection for the NS1000 at some point to avoid all
the #ifdef stuff.

Jochen

Re: [nwst-dev] Audio on the NS1000

[Jochen R. <jo...@pa...>]

> > Sorry to be so silent, ive found the serial-bug in my Netstation.... one 
> > of voltage regulator is slowly died... melting a some chips around.. so 
> > till i don't fix it or find another one i can only listen to your good 
> > news ;))

Do you mean one of the bigger regulators with the heat sinks? Then the whole
machine should not work any longer, I think. There is another 7805 (surface
mount) which I think is used for analog supply only. Then there is a 7812
with small heat sink. I thought the 12V was needed for the smartcard, but it
could be for the serial port as well. The serial port uses a maxim part, so
it should work on 5 volts -- strange.

> 
> Sorry to hear this :-( You know that the Serial is no longer needed. I
> suppose that this has really broken things for your NS :-(
> 
> > "wss0 at isa0 port 0x82c-0x833 irq 10 drq 0 drq2 1: CS4231A"
> 
> Yes, this pretty much looks like what I had found brute forcing it. I had
> left an script testing io ports and irqs and found this morning that it had
> stopped with io=0x830 saying that the irq was not ok, so I just had to test
> the irq and came to the result that the io is 0x830 and the irq is 10, then
> I added the dma info you got from the BSD wich took me to this line:
> 
> modprobe cs4232 io=0x830 irq=10 dma=0 dma2=1
 
How come the io port range on the BSD setup starts 4 bytes lower than 0x830?
Maybe one of the other small chips on the board implements the PnP stuff for
the CS4231A?

> The problem with this is that the NS hangs when I do this :-(

Does it hang when you do the modprobe command, or when you try to play back?

> Any ideas on why this can be happening?

The only other device in the box that uses ISA-style DMA is the floppy disk
controller. I don't have access to hardware right now, or I would try to configure
a kernel with floppy driver. The National Semiconductor SuperIO chip contains a
floppy interface, and that may be bumbling around on dma 0 unless configured.

Regards,
Jochen

[nwst-dev] Audio on the NS1000 locks (help)

[Santiago G. M. <ma...@ay...>]

I have traced the problem when issuing this line:

modprobe cs4232 io=0x830 irq=10 dma=0 dma2=1

till this line of code:

if (request_irq(devc->irq, adintr, 0, devc->name, (void *)my_dev) < 0)

It doesn't come back from the request_irq call, as it starts executing
adintr again and again in a loop without stopping.

Ideas on this?

Regards!

Re: [nwst-dev] Audio on the NS1000

[Santiago G. M. <ma...@ay...>]

> Sorry to be so silent, ive found the serial-bug in my Netstation.... one 
> of voltage regulator is slowly died... melting a some chips around.. so 
> till i don't fix it or find another one i can only listen to your good 
> news ;))

Sorry to hear this :-( You know that the Serial is no longer needed. I
suppose that this has really broken things for your NS :-(

> "wss0 at isa0 port 0x82c-0x833 irq 10 drq 0 drq2 1: CS4231A"

Yes, this pretty much looks like what I had found brute forcing it. I had
left an script testing io ports and irqs and found this morning that it had
stopped with io=0x830 saying that the irq was not ok, so I just had to test
the irq and came to the result that the io is 0x830 and the irq is 10, then
I added the dma info you got from the BSD wich took me to this line:

modprobe cs4232 io=0x830 irq=10 dma=0 dma2=1

The problem with this is that the NS hangs when I do this :-(

Any ideas on why this can be happening?

Are we missing some inits?

Regards!

Re: [nwst-dev] Audio on the NS1000

[Federico F. <ff...@cy...>]

Hi,
Sorry to be so silent, ive found the serial-bug in my Netstation.... one 
of voltage regulator is slowly died... melting a some chips around.. so 
till i don't fix it or find another one i can only listen to your good 
news ;))
Anyway i saved a dmesg log when the original BSD was runnig ... in which 
u can find the parameter U need.
"wss0 at isa0 port 0x82c-0x833 irq 10 drq 0 drq2 1: CS4231A"
A "sad" regards
Federico

----dmesg---------------------------------------------------------------
 > CPU: PowerPC 603ev (Revision 201)
 > CPU: L2 Cache not fitted
 > real memory  = 33554432 (32768K bytes)
 > avail memory = 28393472 (27728K bytes)
 > using 409 buffers containing 1675264 bytes of memory
 > mainbus0 (root)
 > pci0 at mainbus0
 > pci0: i/o enabled, memory enabled
 > pchb0 at pci0 dev 0 function 0
 > pchb0: vendor 0x1014 product 0x0037 (rev. 0x02)
 > pcib0 at pci0 dev 1 function 0
 > pcib0: vendor 0x1014 product 0x000a (rev. 0x03)
 > lep0 at pci0 dev 2 function 0: PCnet-PCI Ethernet
 > lep0: No serial EEPROM
 > lep0: interrupting at irq 15
 > lep0: ether addr 00:00:a7:23:01:64
 > lep0: 32 receive buffers, 32 transmit buffers
 > icsphy0 at lep0 phy 0: ICS1890 10/100 media interface, rev. 3
 > icsphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto
 > OUI 0x00001a model 0x0001 rev 1 at lep0 phy 31 not configured
 > pcic0 at pci0 dev 3 function 0
 > pcic0: vendor 0x104c product 0xac15 (rev. 0x01)
 > pcic0: can't map i/o space
 > vendor 0x104c product 0xac15 (CardBus bridge, revision 0x01) at pci0 dev
 > 3 function 1 not configured
 > vgafb0 at pci0 dev 4 function 0: 1024 x 384, 8bpp
 > wsdisplay0 at vgafb0: console (std, vt100 emulation)
 > isa0 at pcib0
 > com0 at isa0 port 0x3f8-0x3ff irq 4: ns16550a, working fifo
 > com5 at isa0 port 0x2f8-0x2ff irq 3: ns16550a, working fifo
 > lpt0 at isa0 port 0x378-0x37b irq 7
 > wss0 at isa0 port 0x82c-0x833 irq 10 drq 0 drq2 1: CS4231A
 > audio0 at wss0: full duplex
 > pckbc0 at isa0 port 0x60-0x64
 > pckbd0 at pckbc0 (kbd slot)
 > pckbc0: using irq 1 for kbd slot
 > wskbd0 at pckbd0: console keyboard
 > pms0 at pckbc0 (aux slot)
 > pckbc0: using irq 12 for aux slot
 > wsmouse0 at pms0
 > wslpn0 at pms0
 > pcppi0 at isa0 port 0x61
 > sysbeep0 at pcppi0
 > ppi_attached = 1
 > atnvram0 at isa0 port 0x804-0xe69
 > nvram0 at atnvram0
 > low_ram->boot_source = 2
 > biomask 0 netmask 8000 ttymask 9082
 > rfs driver level: 19991105
 > unknown device major 0xfff00
 > Trying device ppp1
 > unknown device major 0xfff00
 > Trying device ppp0
 > Trying device lep0
 > RFS root not mounted. errno=53 (10.0.0.52:/usr/netstation-ppc/ppc)
 > arp info overwritten for 10.0.0.52 by 00:60:08:0e:86:02
 > nfs_boot: timeout...
 > nfs_boot: timeout...
 > nfs_boot: timeout...
 > nfs_boot: mountd error=60
 > nfs_boot: mountd `10.0.0.52:/usr/netstation-ppc/ppc', error=60
 > no file system for lep0
 > Cannot mount root device lep0, error = 79
 > unknown device major 0xfff00
 > Trying device ppp1
 > unknown device major 0xfff00
 > Trying device ppp0
 > Trying device lep0
 > arp info overwritten for 10.0.0.52 by 01:a0:24:ea:3d:9a
 > RFS root not mounted. errno=53 (10.0.0.52:/usr/netstation-ppc/ppc)
 > arp info overwritten for 10.0.0.52 by 01:a0:24:ea:3d:9a
 > root on 10.0.0.52:/usr/netstation-ppc/ppc
 > root file system type: nfs
 > arp info overwritten for 10.0.0.52 by 00:60:08:0e:86:02
 > arp info overwritten for 10.0.0.52 by 01:a0:24:ea:3d:9a
 > ----dmesg---------------------------------------------------------------



Santiago Garcia Mantinan wrote:

>Hi!
>
>I'm trying to work out how to make this thing work, we have a CS4231A-KQ.
>
>The first problem would be which driver to use for this chip, the thing is
>that searching at google, this chip doesn't seem to be alone on a card, the
>cards usually have a DSP or some other things as for the CS4231 seems just
>to be a codec :-? Depending on the card you need one driver or another, so I
>really don't know which one to use.
>
>Jochen sugests to use the CS4232 driver, but you can read on it:
>
> * The low level driver for Crystal CS4232 based cards. The CS4232 is
> * a PnP compatible chip which contains a CS4231A codec, SB emulation,
> * a MPU401 compatible MIDI port, joystick and synthesizer and IDE CD-ROM 
> * interfaces. This is just a temporary driver until full PnP support
> * gets implemented. Just the WSS codec, FM synth and the MIDI ports are
> * supported. Other interfaces are left uninitialized.
>
>So, the CS4232 seems to be a chip that has a CS4231A inside, ok, maybe this
>one works, it seems to be the best one at least looking on the kernel, don't
>know for ALSA. I know it compiles, that's a good start ;-)
>
>Now, the second problem is that this driver needs the IO, IRQ and DMA for
>the card, and I don't know how or where to get this info from, so...
>is there anybody that may guess this? maybe one can use NetBSD for getting
>this values? (Berto, you have that BSD thing running there, can you try to
>find this values?)
>
>Well, hope you can help a bit with this, as for me, having sound and 16 bpp
>are the things we are missing the most.
>
>Regards!
>
>
>
>-------------------------------------------------------
>This SF.net email is sponsored by: Get the new Palm Tungsten T 
>handheld. Power & Color in a compact size! 
>http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0002en
>_______________________________________________
>Networkstation-developers mailing list
>Net...@li...
>https://lists.sourceforge.net/lists/listinfo/networkstation-developers
>

[nwst-dev] Audio on the NS1000

[Santiago G. M. <ma...@ay...>]

Hi!

I'm trying to work out how to make this thing work, we have a CS4231A-KQ.

The first problem would be which driver to use for this chip, the thing is
that searching at google, this chip doesn't seem to be alone on a card, the
cards usually have a DSP or some other things as for the CS4231 seems just
to be a codec :-? Depending on the card you need one driver or another, so I
really don't know which one to use.

Jochen sugests to use the CS4232 driver, but you can read on it:

 * The low level driver for Crystal CS4232 based cards. The CS4232 is
 * a PnP compatible chip which contains a CS4231A codec, SB emulation,
 * a MPU401 compatible MIDI port, joystick and synthesizer and IDE CD-ROM 
 * interfaces. This is just a temporary driver until full PnP support
 * gets implemented. Just the WSS codec, FM synth and the MIDI ports are
 * supported. Other interfaces are left uninitialized.

So, the CS4232 seems to be a chip that has a CS4231A inside, ok, maybe this
one works, it seems to be the best one at least looking on the kernel, don't
know for ALSA. I know it compiles, that's a good start ;-)

Now, the second problem is that this driver needs the IO, IRQ and DMA for
the card, and I don't know how or where to get this info from, so...
is there anybody that may guess this? maybe one can use NetBSD for getting
this values? (Berto, you have that BSD thing running there, can you try to
find this values?)

Well, hope you can help a bit with this, as for me, having sound and 16 bpp
are the things we are missing the most.

Regards!

Re: [nwst-dev] kernel patch and changes to my Debian basic root

[Santiago G. M. <ma...@ay...>]

> I'd like to make some changes:

Well, in fact I did some of them already and they look great.

I have fixed the df issue, I didn't think it was a real problem, but hey,
looks great fixed ;-)

> 1- Right now as /dev must be writable I'm using devfs, wich implies having
> devfs support on the kernel and then having devfsd running, I was thinking
> in using a normal /dev but putting it on the tmpfs we already have compiled
> in and mounted.

Ok removed this and I use the tmpfs instead, it is just 60k on tmpfs so
looks like the way to go ;-)

> 2- I'd like to remove that telnetd and thus inetd to free up some mem.

removed on the new woody-ppc-basic.tbz

> 3- What about not running kerneld ans syslogd?

This I still run them, in fact I had to make a little hack because of a
"bug" on syslog. Opinions?

> 4- Run just 4 gettys, I don't think this will free up anything, but having 6
> seems too much to me.

Still not done, opinions?

The redo script is upgraded to cope with the no devfs thing and the df and
all that.

Well, that is all for today, regards!

[nwst-dev] kernel patch and changes to my Debian basic root

[Santiago G. M. <ma...@ay...>]

Hi!

I've found a "bug" on last Jochen patch, the console thinks it is at 80x25
when we are at 128x48 thus the scrolling doesn't work and also clear and
other stuff doesn't work ok, this patch on top of Jochen's 3rd version
solves that:

--- linux-2.4.19-orig/drivers/video/s3trio.c.orig	Thu Nov 28 16:29:09 2002
+++ linux-2.4.19/drivers/video/s3trio.c	Fri Nov 29 09:28:48 2002
@@ -474,8 +474,8 @@
 	outw(SRC_BACKGROUND | MIX_SRC,  0xb6e8);/* direct color*/
 	outw(MF_T_CLIP | 0, 0xbee8 );     /* clip virtual area  */
 	outw(MF_L_CLIP | 0, 0xbee8 );
-	outw(MF_R_CLIP | (640 - 1), 0xbee8);
-	outw(MF_B_CLIP | (480 - 1),  0xbee8);
+	outw(MF_R_CLIP | (1024 - 1), 0xbee8);
+	outw(MF_B_CLIP | (768 - 1),  0xbee8);
 	Trio_WaitQueue(0x0400);
 	outw(0xffff,  0xaae8);       /* Enable all planes */
 	outw(0xffff, 0xaae8);       /* Enable all planes */
@@ -740,7 +740,7 @@
 
     s3trio_setcolreg(255, 56, 100, 160, 0, NULL /* not used */);
     s3trio_setcolreg(254, 0, 0, 0, 0, NULL /* not used */);
-	memset((char *)s3trio_base, 0, 640*480);
+/*	memset((char *)s3trio_base, 0, 640*480);*/
 	memset((char *)s3trio_base, 0, fb_var.xres * fb_var.yres * 1);
 
 #if 0
@@ -978,7 +978,8 @@
 
     s3trio_setcolreg(255, 56, 100, 160, 0, NULL /* not used */);
     s3trio_setcolreg(254, 0, 0, 0, 0, NULL /* not used */);
-    memset((char *)s3trio_base, 0, 640*480);
+/*    memset((char *)s3trio_base, 0, 640*480);*/
+    memset((char *)s3trio_base, 0, fb_var.xres * fb_var.yres * 1);
 
 #if 0
     Trio_RectFill(0, 0, 90, 90, 7, 1);

About the Debian basic root fs...

I'd like to make some changes:

1- Right now as /dev must be writable I'm using devfs, wich implies having
devfs support on the kernel and then having devfsd running, I was thinking
in using a normal /dev but putting it on the tmpfs we already have compiled
in and mounted.

2- I'd like to remove that telnetd and thus inetd to free up some mem.

3- What about not running kerneld ans syslogd?

4- Run just 4 gettys, I don't think this will free up anything, but having 6
seems too much to me.

What do you think about all this?

I have upgraded my redo script which now copies the /root dir to the tmpfs
so that it is writable, the Debian stuff (the old 3.3.6 stuff had the config
broken), also the kernel and the config file for it and a bloated filesystem
with a lot of stuff like java2 mozilla, swfdec, and have setup a domain name
for putting this on a more normal place: networkstation.aytolacoruna.es

Hope you like all this ;-)

I'll keep on working on a good setup here, I'll try to see what I can do
with the sound, I do not have info on the S3, otherwise I'll be focusing on
getting the S3 to work on 16bpp :-(

Regards!

Re: Re: [nwst-dev] Re: Video on the ns1000

[Santiago G. M. <ma...@ay...>]

> let me do the easy work ;-) You should add the "rw" to the list of kernel-parameter and you're done :-)

:-)

The thing is that I don't want it rw, in fact, my exports looks like this:

/NetworkStationV2	(ro,insecure,all_squash)

The original NetBSD stuff was also on ro, witch seems good, there is no
problem on a rw system if you only have one NC against it, but having
several... and that without having to account the security of a NFS rw
server ;-)

Regards...

Re: [nwst-dev] Re: Video on the ns1000

[Kurt Z. <Kur...@we...>]

Jochen,

"Jochen Roth" <jo...@pa...> schrieb am 28.11.02 22:23:29:
> Here is my current command line:
> 
> console=ttyS0,9600 root=/dev/nfs rw nfsroot=/home/jochen/export/basic ip=dhcp

that are basicly the same parameters I use...

> I think there's something the suse rc scripts do differently. Check out
> the df error message below. Probably /etc/mtab not updated after boot.

hm, the only thing I found is that SuSE deletes the /etc/mtab before mounting the other filesystems...
In my case mount and /proc/mount are looking alright, have a look:

ibm:~ # mount 
/dev/nfs on / type nfs (rw)
proc on /proc type proc (rw)
devpts on /dev/pts type devpts (rw,gid=5,mode=0620)
haegar:/home on /home type nfs (rw,rsize=8192,wsize=8192,hard,intr,addr=10.10.1.1)
helga:/var/spool/mail on /var/spool/mail type nfs (rw,rsize=8192,wsize=8192,hard,intr,addr=10.10.1.100)
ibm:~ # cat /proc/mounts 
rootfs / rootfs rw 0 0
/dev/root / nfs rw,v2,rsize=8192,wsize=8192,hard,udp,nolock,addr=10.10.1.1 0 0
proc /proc proc rw 0 0
devpts /dev/pts devpts rw 0 0
haegar:/home /home nfs rw,v2,rsize=8192,wsize=8192,hard,intr,udp,lock,addr=haegar 0 0
helga:/var/spool/mail /var/spool/mail nfs rw,v2,rsize=8192,wsize=8192,hard,intr,udp,lock,addr=helga 0 0
ibm:~ # df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/nfs              6.8G  6.7G   88M  99% /
haegar:/home          6.8G  5.7G  714M  89% /home
helga:/var/spool/mail 1.1G  940M  144M  87% /var/spool/mail

> Kurt, any way you can post your root file system as a tarball, or step-by-step
> instructions for install-handicapped people like myself?

I think, that I could write an instruction at the weekend. But if you can give me a ftp-account where I can put my  filesystem, I can upload it tomorrow at work. If you are interested, contact me! It's about 80MB zipped...

ciao
Kurt
______________________________________________________________________________
Die Kurier-SMS von WEB.DE FreeMail, eine Benachrichtigung 
der modernen Art! http://freemail.web.de/features/?mc=021175

Re: Re: [nwst-dev] Re: Video on the ns1000

[Jochen R. <jo...@pa...>]

Kurt,

> Santiago Garcia Mantinan <ma...@ay...> schrieb am 28.11.02 20:12:25:
> > > One weird thing with the nfs root fs -- the mount command does not show any
> > > of the mounts, i.e. rootfs, /proc, /dev, etc. ???? The nfs root fs works
> > > very nicely otherwise.
> > 
> > Yes, this is because the root fs is read only (at least I have it like that)
> > and /etc/mtab doesn't show a thing about actual mounts, I'll try to get this
> > a little bit better, and also to fix the lot of noisy errors coming on the
> > screen, but after we get all hard working, I prefer to try to get hard
> > working first. ;-)
> 
> let me do the easy work ;-) You should add the "rw" to the list of kernel-parameter and you're done :-)

Here is my current command line:

console=ttyS0,9600 root=/dev/nfs rw nfsroot=/home/jochen/export/basic ip=dhcp

And here is what I export on the nfs server:

[jochen@mops backup]$ exportfs -v
/home/jochen/export/basic
                10.4.4.0/24(rw,async,wdelay,no_root_squash)
/home/jochen/export/xfb
                10.4.4.0/24(rw,async,wdelay,no_root_squash)

The directories basic and xfb contain what Santiago put up in tarballs
on http://prohibido.aytolacoruna.es/nc/ .

 
> btw: I got a SuSE 7.0 PPC running on the NS1000 :-)))
> ... OK, still without X, but that's the hard e.g. YOUR part ;-))))

I think there's something the suse rc scripts do differently. Check out
the df error message below. Probably /etc/mtab not updated after boot.

debian:~# mount             
debian:~# cat /proc/mounts
rootfs / rootfs rw 0 0
/dev/root / nfs rw,v2,rsize=4096,wsize=4096,hard,udp,nolock,addr=10.4.4.11 0 0
none /dev devfs rw 0 0
proc /proc proc rw 0 0
debian:~# df
df: cannot read table of mounted filesystems
debian:~# mount -r /dev/hda1 hd
debian:~# mount
/dev/hda1 on /root/hd type ext2 (ro)
debian:~# df
Filesystem           1k-blocks      Used Available Use% Mounted on
/dev/hda1              1587396    302712   1204048  21% /root/hd
debian:~# 

Anyway, I am sure we will figure it out.

Kurt, any way you can post your root file system as a tarball, or step-by-step
instructions for install-handicapped people like myself?

Jochen

Re: Re: [nwst-dev] Re: Video on the ns1000

[Kurt Z. <Kur...@we...>]

Hi all!

Santiago Garcia Mantinan <ma...@ay...> schrieb am 28.11.02 20:12:25:
> > One weird thing with the nfs root fs -- the mount command does not show any
> > of the mounts, i.e. rootfs, /proc, /dev, etc. ???? The nfs root fs works
> > very nicely otherwise.
> 
> Yes, this is because the root fs is read only (at least I have it like that)
> and /etc/mtab doesn't show a thing about actual mounts, I'll try to get this
> a little bit better, and also to fix the lot of noisy errors coming on the
> screen, but after we get all hard working, I prefer to try to get hard
> working first. ;-)

let me do the easy work ;-) You should add the "rw" to the list of kernel-parameter and you're done :-)

btw: I got a SuSE 7.0 PPC running on the NS1000 :-)))
... OK, still without X, but that's the hard e.g. YOUR part ;-))))

ciao
Kurt
______________________________________________________________________________
Finger wund tippen beim SMS Schreiben? Schluss damit! 
Web.de FreeMail - http://freemail.web.de/features/?mc=021148

Re: [nwst-dev] Re: Video on the ns1000

[Santiago G. M. <ma...@ay...>]

I have just had lunch, yes, at 17:30 :-(

I have been till late at work trying to get java to work on mozilla on the
NC, but I could not do it (Igalia guys reading this, you could have a look
at this) the Blackdown-1.3.1-02b-FCS machine didn't seem to run ok on powerpc
:-( I have not been able to find IBM's JVM for powerpc (that sounds weird to
me) and blackdown 1.1 machine doesn't seem to have a plugin for Mozilla.

Anyway, I was testing this because I had patched the kernel to get the
64Megs used and so I had memory for a jvm.

> I uploaded a newer patch that fixes the memory size problem and the slow
> timer. This version also knows how to output the kernel command line

Heh heh, rules, I left the ppc compiling it and has finished by now, I'll
test tomorrow.

> prompt to the vga. The boot graphics driver uses whatever is configured
> in the firmware for the x-server. You need to disable CONFIG_VGA_CONSOLE
> for the built-in fb driver to work.

Umm, is that because of the command line output to the vga? Because I was
having 1024x768 with CONFIG_VGA_CONSOLE enabled and having a 128x48 console
without any problem. I'll try to have a look at the code tomorrow, I didn't
have time today.

> As you said, this driver does not init
> the VGA registers, so it will only run at the resolution set by the
> firmware.

Umm, when are we gonna have mode switching on the FB? I mean, like being
able to use 16bpp or something like that :-) Did you have any luck with your
info for setting the modes?

> Can you also post what you added to the fb driver? Is that the init code
> from cyberfb.c?

Nope, I don't have any init code, I just played with the mode settings, I
calculated the values from the XFree settings for 1024x768x85 ISO stuff.

Here is my patch:

diff -r -u linux-2.4.19.orig/drivers/video/s3trio.c linux-2.4.19/drivers/video/s3trio.c
--- linux-2.4.19.orig/drivers/video/s3trio.c	Thu Nov 28 18:34:19 2002
+++ linux-2.4.19/drivers/video/s3trio.c	Wed Nov 27 10:32:51 2002
@@ -72,12 +72,23 @@
 
 static struct fb_fix_screeninfo fb_fix;
 static struct fb_var_screeninfo fb_var = {
+		1024, 768, 1024, 768, 0, 0, 8, 0,
+		{0, 8, 0}, {0, 8, 0}, {0, 8, 0}, {0, 0, 0},
+		/* 1024x768, 85 Hz,    pixclock (ns) lm   rm  um  lm hs  vs */
+		0, 0, -1, -1, FB_ACCELF_TEXT, 10580, 208, 48, 36, 1, 96, 3,
+		FB_SYNC_COMP_HIGH_ACT|FB_SYNC_VERT_HIGH_ACT, 
+		FB_VMODE_NONINTERLACED
+};
+
+/*
+static struct fb_var_screeninfo fb_var = {
 		640, 480, 640, 480, 0, 0, 8, 0,
 		{0, 8, 0}, {0, 8, 0}, {0, 8, 0}, {0, 0, 0},
 		0, 0, -1, -1, FB_ACCELF_TEXT, 39722, 40, 24, 32, 11, 96, 2,
 		FB_SYNC_COMP_HIGH_ACT|FB_SYNC_VERT_HIGH_ACT, 
 		FB_VMODE_NONINTERLACED
 };
+*/
 
 static struct fb_var_screeninfo default_var_640x480 = {
 		640, 480, 640, 480, 0, 0, 8, 0,
@@ -331,8 +342,8 @@
 	outw(SRC_BACKGROUND | MIX_SRC,  0xb6e8);/* direct color*/
 	outw(MF_T_CLIP | 0, 0xbee8 );     /* clip virtual area  */
 	outw(MF_L_CLIP | 0, 0xbee8 );
-	outw(MF_R_CLIP | (640 - 1), 0xbee8);
-	outw(MF_B_CLIP | (480 - 1),  0xbee8);
+	outw(MF_R_CLIP | (1024 - 1), 0xbee8);
+	outw(MF_B_CLIP | (768 - 1),  0xbee8);
 	Trio_WaitQueue(0x0400);
 	outw(0xffff,  0xaae8);       /* Enable all planes */
 	outw(0xffff, 0xaae8);       /* Enable all planes */
@@ -570,7 +581,7 @@
 
     s3trio_setcolreg(255, 56, 100, 160, 0, NULL /* not used */);
     s3trio_setcolreg(254, 0, 0, 0, 0, NULL /* not used */);
-    memset((char *)s3trio_base, 0, 640*480);
+    memset((char *)s3trio_base, 0, 1024*768);
 
 #if 0
     Trio_RectFill(0, 0, 90, 90, 7, 1);
@@ -804,7 +815,7 @@
 
     s3trio_setcolreg(255, 56, 100, 160, 0, NULL /* not used */);
     s3trio_setcolreg(254, 0, 0, 0, 0, NULL /* not used */);
-    memset((char *)s3trio_base, 0, 640*480);
+    memset((char *)s3trio_base, 0, 1024*768);
 
 #if 0
     Trio_RectFill(0, 0, 90, 90, 7, 1);


> Don't know about alsa. The kernel contains an OSS driver for the Chrystal
> Semi CS4232, but the PReP code in arch/ppc/prep_setup.c is broken. The
> broken code does nothing more than load the driver with the proper settings
> for the machine, i.e. io, interrupt, dma - I think. You should be able to
> enable modules for the kernel, compile cs4232 as module, and insmod with
> proper command line options.

I think I tried this but it didn't compile, I'll try again.

> One weird thing with the nfs root fs -- the mount command does not show any
> of the mounts, i.e. rootfs, /proc, /dev, etc. ???? The nfs root fs works
> very nicely otherwise.

Yes, this is because the root fs is read only (at least I have it like that)
and /etc/mtab doesn't show a thing about actual mounts, I'll try to get this
a little bit better, and also to fix the lot of noisy errors coming on the
screen, but after we get all hard working, I prefer to try to get hard
working first. ;-)

Well, let's see if we can get the FB to work on 16 bits and that sound and
we are almost done ;-)

Regards!

Re: [nwst-dev] Re: Video on the ns1000

[Jochen R. <jo...@pa...>]

Hi Santiago,

> And I have good news, I have X running here using the 3.3.6 FB server, I
> have changed your FB driver (the one that doesn't init the vga) and I got
> the console and the X to work at 1024x768 8 bits.

Very cool!

> The two main things I'm caring for are getting 16 bits to work and also get
> it to recognice the 64Megs of ram (currently it only detects 32Mb here)

I uploaded a newer patch that fixes the memory size problem and the slow
timer. This version also knows how to output the kernel command line
prompt to the vga. The boot graphics driver uses whatever is configured
in the firmware for the x-server. You need to disable CONFIG_VGA_CONSOLE
for the built-in fb driver to work. As you said, this driver does not init
the VGA registers, so it will only run at the resolution set by the
firmware.

> I'll try to post a compressed tarball of the filesystems once I get some
> little things working here and clean up all this ;-)

Can you also post what you added to the fb driver? Is that the init code
from cyberfb.c?

> One thing, do you think we could try to get the sound card working by just
> compiling alsa or is there any support on the kernel still needed?

Don't know about alsa. The kernel contains an OSS driver for the Chrystal
Semi CS4232, but the PReP code in arch/ppc/prep_setup.c is broken. The
broken code does nothing more than load the driver with the proper settings
for the machine, i.e. io, interrupt, dma - I think. You should be able to
enable modules for the kernel, compile cs4232 as module, and insmod with
proper command line options.

One weird thing with the nfs root fs -- the mount command does not show any
of the mounts, i.e. rootfs, /proc, /dev, etc. ???? The nfs root fs works
very nicely otherwise.

Jochen

Re: [nwst-dev] Re: Video on the ns1000

[Santiago G. M. <ma...@ay...>]

> > Modeline "640x480xmio"   25.175  640 664 760 800  480 491 493 525
> 
> Does the 'xmio' specify color depth? If so, that may be the problem. The
> current code only supports 8bits/pixel color-mapped modes.

Nope, it was to make sure it selected this special mode and not the other
640x480 modes, I was running 8 bits.

> I just got more information on the graphics chip from one of the guys
> on the fbdev mailing list. That should help fixing mode setup on the

Great.

> have 800x600 and 1024x768 working here, except that strange things
> happen on the last line of the console. Probably something broken in
> the scrolling code or accelerator. I will post a newer patch as soon
> as I have something working.

I have been working on all this also.

And I have good news, I have X running here using the 3.3.6 FB server, I
have changed your FB driver (the one that doesn't init the vga) and I got
the console and the X to work at 1024x768 8 bits.

I'm running opera, mozilla, ... on the machine without problems.

The two main things I'm caring for are getting 16 bits to work and also get
it to recognice the 64Megs of ram (currently it only detects 32Mb here)

I'll try to post a compressed tarball of the filesystems once I get some
little things working here and clean up all this ;-)

One thing, do you think we could try to get the sound card working by just
compiling alsa or is there any support on the kernel still needed?

Regards...

[nwst-dev] Re: Video on the ns1000

[Jochen R. <jo...@pa...>]

Hi,

> it looks great to see that penguin on the screen ;-)
doesn't it?

> Modeline "640x480xmio"   25.175  640 664 760 800  480 491 493 525

Does the 'xmio' specify color depth? If so, that may be the problem. The
current code only supports 8bits/pixel color-mapped modes.

I just got more information on the graphics chip from one of the guys
on the fbdev mailing list. That should help fixing mode setup on the
have 800x600 and 1024x768 working here, except that strange things
happen on the last line of the console. Probably something broken in
the scrolling code or accelerator. I will post a newer patch as soon
as I have something working.

Jochen

[nwst-dev] Re: Video on the ns1000

[Santiago G. M. <ma...@ay...>]

Hi, I'm back.

> I posted a newer patch on sourceforge which allows to run the S3 chip at
> 640x480-8, or 80x30. I need to add support for multiple modes, and
> perhaps a couple other items. The boot stage does not know how to talk
> to the S3 chip, so that still goes over the serial port for now.

Hi!

I have been testing the patch and have built some kernels with it using my
previous config but adding the FB stuff, console works ok, now I get all the
messages on it and in the terminal, with all the Debian messages on the
console already, it looks great to see that penguin on the screen ;-)

What I have built is as always on prohibido.aytolacoruna.es/nc/ (BTW, I have
a SF account, mantinan, so if you want I could be uploading stuff there,
there is no problem for me to continue doing it here, though), I have moved
the old kernel and xfree log to old/ and the new stuff I have put there
looks like this:

   851869 Nov 26 14:44 kernel.fb	(kernel with framebuffer support)
    12548 Nov 26 15:02 config.fb	(the config I used for this kernel)
  1443987 Nov 26 15:10 fb.log		(log and strace of X using fb server)
  1273904 Nov 26 15:10 s3.log		(log and strace of X using s3 server)
 30223414 Nov 26 15:11 woody-xfb.tbz	(new debian base with lots of X stuff
					 and strace)

> I think that XFree86 should work in frame buffer mode, but I could be
> wrong. I have done very little testing with this driver so far.

My tests show that the fb is not working, at least it doesn't show anything
on the screen, wich becomes black after you try to start X. I thought that
maybe this was a problem with the modeline that X was trying to use so I
tried to calculate a modeline equivalent of the one we use on the FB, the
result was this line:

Modeline "640x480xmio"   25.175  640 664 760 800  480 491 493 525

and this is the output that the X server showed:

(--) FBDEV(0): Virtual size is 640x480 (pitch 640)
(**) FBDEV(0): Mode "640x480xmio": 25.2 MHz (scaled from 0.0 MHz), 31.5 kHz, 59.9 Hz
(II) FBDEV(0): Modeline "640x480xmio"   25.18  640 664 760 800  480 491 493 525

this should be at the log of the fb X test, however screen was still black.


I had reported a bug on the S3 server on this cards to the XFree guys, still
got no answer to it :-(

> Maybe 3.3.6 works in framebuffer mode, I know fairly little about XFree86
> inner workings.

Umm I tried 4.2.1, I'll try to test 3.3.6 tomorrow.

Well, that is all I could do today, basically gatter info to see if anybody
can diagnose this ;-)

Hope this helps a bit, work is really promissing, seing Tux here by my side
on the NC next to the one I'm using really looks great, we only need a
little push more.

Regards...

[nwst-dev] (no subject)

[C. C. <jmc...@ig...>]

Hi Jochen,

Can you give just a little information about how to cross-compile the
kernel using an i388 architecture? I'm getting mad using a 100MHz NC.

Yesterday we have been working with the NC, running a NetBSD 1.4, I
think it worked almost perfectly using the NCOS kernel and its Xserver.
And then using chroot at the base directory of the netbsd base
distribution. Then we started fwwm95 and played mp3 with mpg123.

Thanks,
--=20
Jos=E9 Mar=EDa Casanova Crespo
Ingeniero en Inform=E1tica		=09
mailto:jmc...@ig...
IGALIA S.L. http://www.igalia.com

[nwst-dev] Re: Video on the ns1000

[Santiago G. M. <ma...@ay...>]

> I posted a newer patch on sourceforge which allows to run the S3 chip at
> 640x480-8, or 80x30. I need to add support for multiple modes, and
> perhaps a couple other items. The boot stage does not know how to talk
> to the S3 chip, so that still goes over the serial port for now.

This are great news anyway, really a great work that you are doing.

I do not have the time to test all this now as I'm leaving for Madrid, and
then I'll be on vacation for a week and I won't have access to the hardware,
but I have posted all the xfree stuff I have compiled for Debian on the same
url I used before for the rest of the stuff
(http://prohibido.aytolacoruna.es/nc/) this includes the full xfree 4.2.1
stuff compiled for woody and also basic xfree 3.3.6 for woody, wich has the
xserver-fbdev that you wanted to test.

There seems to be a problem with XFree and this cards, as I couldn't get
4.2.1 to run on a GX version of the S3 64V2 on a Pentium based machine, so
I'm filling a bug on this now. The 3.3.6 version of Xfree had support for
this on the svga server, but it didn't compile at first try here and I don't
have time to see why, however fixing 4.2.1 seems a better way to go for me,
maybe the cvs could help with this, we'll see what the X boys say about
this.

Well, I'll see what you all can do and try to help you when I'm back.

See you all and good luck with this.

Regards.

Re: [nwst-dev] Re: Linux Kernel 2.4.19 running on the 8632-A53

[Jochen R. <jo...@pa...>]

Federico,

I am running 3.1.something firmware on the boxes here.

IBM had the firmware updates somewhere, perhaps in one of the NSM
packages. I have them somewhere here, but I would have to search...
I also think that, at least for now, the NSM is the only way to
update the flash.

It is quite possible that the earlier firmware sets some things up
too differently.

Jochen

> 
> Jochen,
> No my NS do not have the L2 cache on board...  
> Actually its configuration is : Firmware version 2.9.7 - Video RAM 2M - 
> Sys RAM 32 M.
> Is it similar to your?
> Federico
> 
> Jochen Roth wrote:
> 
> >Federico,
> >
> >Does your NS have the L2 cache module? I have one here with it, but I
> >have not tried it yet. If it does have it, try pulling the module.
> >
> >What version firmware do you have?
> >
> >Jochen
> >
> >>Hi guys
> >>I know it seems an idiot question, but im going to throw out the window 
> >>my Ns!
> >>Im not able to configure a minicom session to get readable datas from 
> >>the NS serial console.
> >>May be im missing stg....the cable i use is a null modem cable (coming 
> >>up from a CISCO box), i used it also to configure a radio bridge and it 
> >>works.
> >>During boot up i have a full working session, but after serial setting 
> >>up i get anly rotten lines:
> >>" VFS: Mounted root (ext2 filesystem) readonly."
> >>" Freeing unused kernel memory: 124k init 44k pmac 8k chrp 4k openfirmware "
> >>" se_rriael cled"
> >>" rrtinanls dvuh: CREADccutisnt!t .v"
> >>" eexxeccutisn.!."
> >>" e`EXT2-fs warning: checktime reached, runnin reached, running e2fsck 
> >>is rslooogd"
> >>" syrslooog diini"
> >>" "
> >>" TTestg d03niiionTn 00.0dduh: CREAD not set!"
> >>" "
> >>" (ndnoninee)) l "
> >>This is my status line for minicom: "|  9600 8N1 | NOR | Minicom 1.83.1 
> >>| VT102 |" and after login this is the prompt ive "ssh-22..0.0# "
> >>Just to be sure ive also tried an Hyperterminal session... with same 
> >>results... 
> >>Do U have any tips?
> >>Thanks
> >>Federico
> >>
> >>
> >>Jochen Roth wrote:
> >>
> >>>Everyone,
> >>>
> >>>I finally found time to pick up my lose ends here and posted the current
> >>>state of my work to http://sourceforge.net/projects/networkstation/ .
> >>>
> >>>Check out the patches and the compiled kernel and let me know if
> >>>that works for you. There is also a list of problems in the file BUGS
> >>>in the patches and kernel tarballs.
> >>>
> >>>I can tar up the cross compiler I am using and upload that as well, if
> >>>it helps.
> >>>
> >>>The kernels I looked at came with a driver for the S3 trio to run it
> >>>as a frame buffer device, but they all assumed that there was some Open
> >>>Firmware-type information available for the device. The NCD firmware
> >>>provides neither that nor PReP residual data, so there is a need for
> >>>some work there.
> >>>
> >>>I soldered a debug connector into my NS1000 which allowed me to use a
> >>>jtag-based debugger. If you have one available, like the Abatron BDI
> >>>2000 or a visionprobe, then I highly recommend that.
> >>>
> >>>I also managed to add a couple of parts to get the IDE interface working,
> >>>basically a couple of octal buffers and the IDE connector. I will try
> >>>to take some photos of that and post them.
> >>>
> >>>Regards,
> >>>Jochen
> >>>
> >>>
> >>
> >

[nwst-dev] Video on the ns1000

[Jochen R. <jo...@pa...>]

Hi Santiago,

I posted a newer patch on sourceforge which allows to run the S3 chip at
640x480-8, or 80x30. I need to add support for multiple modes, and
perhaps a couple other items. The boot stage does not know how to talk
to the S3 chip, so that still goes over the serial port for now.

I think that XFree86 should work in frame buffer mode, but I could be
wrong. I have done very little testing with this driver so far.

> The filesystem:
> What I've been trying to do is setup a Debian to get it exported through
> NFS, the NFS is ro, so it gives some errors, but seems to work all the well
> that can be expected from this development we are trying to make.

Great, I will take a look at that.

> Starting X:
> Well, we have Linux running, and the pci bus seems to work ok for the
> network card, so, even thought the kernel seems not to be able to init the
> video card, I thought that maybe X could do it, I was going to go for X
> 3.3.6, but there is no compiled version of that on woody, also I seem to
> recall that our S3 is not supported on X 4.1.0, so I had to go for X 4.2.1
> wich is on unstable, tomorrow I'll try to compile X 4.2.1 for woody.

Maybe 3.3.6 works in framebuffer mode, I know fairly little about XFree86
inner workings.

I looked at your log file and had an idea why X may not work.

The NS1000 is essentially a PReP machine, which is what the 82660 host
bridge supports. PReP machines use a weird memory map, in that the addresses
seen by the PCI bus are not the same as those seen by the CPU. I am talking
about physical addresses here, so phys/virt translation comes on top of
that.

A good description of the PReP map is in the Motorola MPC106 UM page 3-2,
or http://e-www.motorola.com/brdata/PDFDB/docs/MPC106UM.pdf page 96.

The S3 chip is mapped on PCI address 0, which is physical 60x bus address
0xc0000000. This has nothing to do with the virtual address 0xc0000000
used as the kernel base.

Now when you do a cat /proc/pci on the ns1000, you see the S3 chip listed
with a 0xc0000000 base address. This is because somewhere in the PReP PCI
config cycle code conversion the address is translated.

XFree86 4.2, otoh, seems to read the value un-translated, sees the 0, and
decides that the device is not mapped, i.e. not usable.

So one option would be to try setpci and change the bar0 from 0x0 to e.g.
0x08000000. The ethernet chip is mapped at 0x04000000 pci or 0xc4000000 cpu
physical. Note: /proc/pci seems to not get updated when pci registers are
changed. I think /proc/bus/pci/00/0e.0 does change dynamically.

sh-2.04# cat /proc/pci
PCI devices found:
  Bus  0, device  11, function  0:
    ISA bridge: IBM Fire Coral (rev 3).
      I/O at 0x100020 [0x100027].
  Bus  0, device  12, function  0:
    Ethernet controller: Advanced Micro Devices [AMD] 79c970 [PCnet LANCE] (rev 37).
      IRQ 15.
      Master Capable.  Latency=128.  Min Gnt=6.Max Lat=255.
      I/O at 0x100000 [0x10001f].
      Non-prefetchable 32 bit memory at 0xc4000000 [0xc400001f].
  Bus  0, device  14, function  0:
    VGA compatible controller: S3 Inc. 86c775/86c785 [Trio 64V2/DX or /GX] (rev 20).
      IRQ 9.
      Non-prefetchable 32 bit memory at 0xc0000000 [0xc3ffffff].
sh-2.04# 
sh-2.04# hexdump /proc/bus/pci/00/0e.0
0000000 3353 0189 0300 0002 1400 0003 0000 0000
0000010 0000 0000 0000 0000 0000 0000 0000 0000
*
0000030 0000 0c00 0000 0000 0000 0000 0001 0000
0000040 0000 0000 0000 0000 0000 0000 0000 0000
*
0000100
sh-2.04# 

PCI bar0 is at offset 0x10. Offset 0x30 is the ROM address, 0x000c0000 in
this case.  the hexdump output is little endian 16-bit words printed as
big endian, I think.

I will put some code into the boot stage to remap the S3 chip in the next
patch.

Jochen

[nwst-dev] Re: Linux Kernel 2.4.19 running on the 8632-A53

[Federico F. <ff...@cy...>]

Jochen,
No my NS do not have the L2 cache on board...  
Actually its configuration is : Firmware version 2.9.7 - Video RAM 2M - 
Sys RAM 32 M.
Is it similar to your?
Federico

Jochen Roth wrote:

>Federico,
>
>Does your NS have the L2 cache module? I have one here with it, but I
>have not tried it yet. If it does have it, try pulling the module.
>
>What version firmware do you have?
>
>Jochen
>
>>Hi guys
>>I know it seems an idiot question, but im going to throw out the window 
>>my Ns!
>>Im not able to configure a minicom session to get readable datas from 
>>the NS serial console.
>>May be im missing stg....the cable i use is a null modem cable (coming 
>>up from a CISCO box), i used it also to configure a radio bridge and it 
>>works.
>>During boot up i have a full working session, but after serial setting 
>>up i get anly rotten lines:
>>" VFS: Mounted root (ext2 filesystem) readonly."
>>" Freeing unused kernel memory: 124k init 44k pmac 8k chrp 4k openfirmware "
>>" se_rriael cled"
>>" rrtinanls dvuh: CREADccutisnt!t .v"
>>" eexxeccutisn.!."
>>" e`EXT2-fs warning: checktime reached, runnin reached, running e2fsck 
>>is rslooogd"
>>" syrslooog diini"
>>" "
>>" TTestg d03niiionTn 00.0dduh: CREAD not set!"
>>" "
>>" (ndnoninee)) l "
>>This is my status line for minicom: "|  9600 8N1 | NOR | Minicom 1.83.1 
>>| VT102 |" and after login this is the prompt ive "ssh-22..0.0# "
>>Just to be sure ive also tried an Hyperterminal session... with same 
>>results... 
>>Do U have any tips?
>>Thanks
>>Federico
>>
>>
>>Jochen Roth wrote:
>>
>>>Everyone,
>>>
>>>I finally found time to pick up my lose ends here and posted the current
>>>state of my work to http://sourceforge.net/projects/networkstation/ .
>>>
>>>Check out the patches and the compiled kernel and let me know if
>>>that works for you. There is also a list of problems in the file BUGS
>>>in the patches and kernel tarballs.
>>>
>>>I can tar up the cross compiler I am using and upload that as well, if
>>>it helps.
>>>
>>>The kernels I looked at came with a driver for the S3 trio to run it
>>>as a frame buffer device, but they all assumed that there was some Open
>>>Firmware-type information available for the device. The NCD firmware
>>>provides neither that nor PReP residual data, so there is a need for
>>>some work there.
>>>
>>>I soldered a debug connector into my NS1000 which allowed me to use a
>>>jtag-based debugger. If you have one available, like the Abatron BDI
>>>2000 or a visionprobe, then I highly recommend that.
>>>
>>>I also managed to add a couple of parts to get the IDE interface working,
>>>basically a couple of octal buffers and the IDE connector. I will try
>>>to take some photos of that and post them.
>>>
>>>Regards,
>>>Jochen
>>>
>>>
>>
>
>
>------------------------------------------------
>Jochen Roth              jochen at panix dot com
>

[Networkstation-developers] ping

[Jochen R. <jo...@pa...>]

Hello,

I set up a mailing list via sourceforge.net and added the
addresses of all people I thought interested in the matter.
I hope this is ok with everyone.

To post, send email to:

  net...@li...

To unsubscribe or change options, send email to:

  net...@li...


Regards,
Jochen