| View previous topic :: View next topic |
| Author |
Message |
ddc
Guru
Joined: 29 Aug 2006
Posts: 522
|
 Posted: Thu Sep 28, 2006 11:48 am Post subject: sgi indigo2/impact R10K gentoo |
 |
|
hi,
i'have this machine (sgi indigo2/impact R10K, the purple one, equipped with a Solid Impact gfx)
| Code: |
>>hinv
System: IP28
Processor: 175 Mhz R10000, with FPU
Primary I-cache size: 32 Kbytes
Primary D-cache size: 32 Kbytes
Secondary cache size: 1024 Kbytes
Memory size: 320 Mbytes
Graphics: Solid Impact
SCSI Disk: scsi(0)disk(2)
SCSI CDROM: scsi(0)cdrom(4)
Audio: Iris Audio Processor: version A2 revision 1.1.0
|
i installed gentoo 2006.1 to an impact R10K, and i set the PROM-evironment this way
| Code: |
>>printenv
OSLoadPartition=scsi(0)disk(2)rdisk(0)partition(8)
SystemPartition=scsi(0)disk(2)rdisk(0)partition(8)
OSLoadFilename=ip28(gentoo)
OSLoader=sash64
AutoLoad=No
TimeZone=PST8PDT
console=g
diskless=0
dbaud=9600
volume=80
sgilogo=y
autopower=y
netaddr=192.168.0.5
eaddr=08:00:69:0a:3a:9a
boottune=1
cpufreq=174
ConsoleOut=serial(0)
ConsoleIn=serial(0)
|
i configure the last-one 2.6.17.ip28 kernel in these two ways:
(A) without support for EISA-bus and for sgi-framebuffer
(B) with support for EISA-bus, for virtual terminals, for sgi-framebuffer
now, (A) boots very well, while (B) start booting, then loops for ever
see the following:
(A) boots very well
| Code: |
>> boot
1088+54960+336+4320+1216 entry: 0x9000000033ff10c0
ARCLoad version 0.5 (c) 2004-5 Stanislaw Skowronek
-- WARNING --
Your version of GCC does not support nested functions.
Please apply patch or download new version.
Patched 4 occurrences of the wrong trampoline.
-- WARNING --
Loading configuration for 'ip28(gentoo)'...
Loading scsi(0)disk(2)rdisk(0)partition(8)/gentoo...
Reading 4124806 bytes... OK.
Entering kernel.
Linux version 2.6.16.27-sgimips-impact-r10k (root@sakura) (gcc version 3.4.6 (Gentoo 3.4.6)) #6 Wed Sep 27 11:40:43 Local time zone must be set--see zic manu
ARCH: SGI-IP28
PROMLIB: ARC firmware Version 64 Revision 0
CPU revision is: 00000925
FPU revision is: 00000900
Silicon Graphics Indigo2 R10k (IP28) support: (c) 2004 peter fuerst.
MC: SGI memory controller Revision 5
MC: Boardrev. 13, Chiprev. 0
MC: Probing memory configuration:
bank0: 128M @ 20000000
bank1: 128M @ 28000000
bank2: 64M @ 30000000
SGI graphics system @ 0x00000000
Determined physical RAM map:
memory: 0000000014000000 @ 0000000020000000 (usable)
Built 1 zonelists
Kernel command line: root=/dev/sda3 ro gentoo=noudev
Primary instruction cache 32kB, physically tagged, 2-way, linesize 64 bytes.
Primary data cache 32kB, 2-way, linesize 32 bytes.
Unified secondary cache 1024kB 2-way, linesize 128 bytes.
Synthesized TLB refill handler (37 instructions).
Synthesized TLB load handler fastpath (49 instructions).
Synthesized TLB store handler fastpath (49 instructions).
Synthesized TLB modify handler fastpath (48 instructions).
PID hash table entries: 4096 (order: 12, 131072 bytes)
Calibrating system timer... 87500 [175.0000 MHz CPU]
Using 87.500 MHz high precision timer.
Console: colour dummy device 80x25
Dentry cache hash table entries: 131072 (order: 8, 1048576 bytes)
Inode-cache hash table entries: 65536 (order: 7, 524288 bytes)
Memory: 309504k/327680k available (3040k kernel code, 17724k reserved, 800k data, 188k init, 0k highmem)
Mount-cache hash table entries: 256
Checking for 'wait' instruction... unavailable.
Checking for the multiply/shift bug... no.
Checking for the daddi bug... no.
Checking for the daddiu bug... no.
NET: Registered protocol family 16
SCSI subsystem initialized
Installing knfsd (copyright (C) 1996 okir@monad.swb.de).
EFS: 1.0a - http://aeschi.ch.eu.org/efs/
Initializing Cryptographic API
io scheduler noop registered
io scheduler anticipatory registered
io scheduler deadline registered (default)
io scheduler cfq registered
impact_devinit: !gfxaddr
impact: probe of impact.0 failed with error -22
DS1286 Real Time Clock Driver v1.0
indydog: Hardware Watchdog Timer for SGI IP22: 0.3
serio: i8042 AUX port at 0x900000001fbd9843,0x900000001fbd9847 irq 44
serio: i8042 KBD port at 0x900000001fbd9843,0x900000001fbd9847 irq 44
Serial: IP22 Zilog driver (1 chips).
ttyS0 at MMIO 0x200000001fbd9830 (irq = 45) is a IP22-Zilog
Console: ttyS0 (IP22-Zilog)
ttyS1 at MMIO 0x200000001fbd9838 (irq = 45) is a IP22-Zilog
loop: loaded (max 8 devices)
eth0: SGI Seeq8003 08:00:69:0a:3a:9a
wd33c93-0: chip=WD33c93B/13 no_sync=0xff no_dma=0 debug_flags=0x00
setup_args=,,,,,,,,,
Version 1.26 - 22/Feb/2003, Compiled Sep 25 2006 at 23:20:25
wd33c93-1: chip=WD33c93B/13 no_sync=0xff no_dma=0 debug_flags=0x00
setup_args=,,,,,,,,,
Version 1.26 - 22/Feb/2003, Compiled Sep 25 2006 at 23:20:25
scsi0 : SGI WD93
isa bounce pool size: 16 pages
sending SDTR 0103013f0csync_xfer=2c<5> Vendor: IBM Model: DNES-318350 Rev: SA30
Type: Direct-Access ANSI SCSI revision: 03
sending SDTR 0103013f0csync_xfer=2c<5> Vendor: TOSHIBA Model: CD-ROM XM-4101TA Rev: 1084
Type: CD-ROM ANSI SCSI revision: 02
scsi1 : SGI WD93
SCSI device sda: 35843670 512-byte hdwr sectors (18352 MB)
sda: Write Protect is off
SCSI device sda: drive cache: write back
SCSI device sda: 35843670 512-byte hdwr sectors (18352 MB)
sda: Write Protect is off
SCSI device sda: drive cache: write back
sda: sda1 sda2 sda3 sda9 sda11
sd 0:0:2:0: Attached scsi disk sda
sr0: scsi-1 drive
Uniform CD-ROM driver Revision: 3.20
mice: PS/2 mouse device common for all mice
SGI HAL2 revision 0.1.0
NET: Registered protocol family 2
IP route cache hash table entries: 8192 (order: 4, 65536 bytes)
TCP established hash table entries: 32768 (order: 6, 262144 bytes)
TCP bind hash table entries: 32768 (order: 6, 262144 bytes)
TCP: Hash tables configured (established 32768 bind 32768)
TCP reno registered
TCP bic registered
NET: Registered protocol family 1
NET: Registered protocol family 17
kjournald starting. Commit interval 5 seconds
EXT3-fs: mounted filesystem with ordered data mode.
VFS: Mounted root (ext3 filesystem) readonly.
Freeing unused kernel memory: 188k freed
atkbd.c: keyboard reset failed on hpc3ps2/serio1
INIT: version 2.86 booting
atkbd.c: keyboard reset failed on hpc3ps2/serio0
|
(B) start booting, then loops for ever
| Quote: |
>> boot
1088+54960+336+4320+1216 entry: 0x9000000033ff10c0
ARCLoad version 0.5 (c) 2004-5 Stanislaw Skowronek
-- WARNING --
Your version of GCC does not support nested functions.
Please apply patch or download new version.
Patched 4 occurrences of the wrong trampoline.
-- WARNING --
Loading configuration for 'ip28(gentoo)'...
Loading scsi(0)disk(2)rdisk(0)partition( 8 )/gentoo...
Reading 4124806 bytes... OK.
Entering kernel.
---LOOP X EVER---
|
the same happens using the prebuild kernel from http://dev.gentoo.org/~redhatter/mips/sgi/ip28/kernel-images/
i think the problem is related to the gfx solid impact: removing it in (B) cause the system to complete the boot process, keeping the gfx inside EISA bus in (B) cause the system not to complete the boot process: it seems stalled waiting xever for something from the hardware.
do you have any idea of the problem ?
tested kernel 2.6.16, 2.6.17 |
|
| Back to top |
|
 |
Redhatter
Retired Dev
Joined: 20 Sep 2003
Posts: 548
Location: Brisbane, QLD, Australia
|
| Posted: Thu Sep 28, 2006 1:11 pm Post subject: |
|
|
That is unusual. You mention graphics in the EISA bus -- Last I checked, Solid Impact boards plugged into the GIO64 bus ports, not EISA. 
That kernel I built definately does have VT and IMPACT support inbuilt. That said, I'll put up another image in a moment. What does your arc.cf look like?
_________________
Stuart Longland (a.k.a Redhatter, VK4MSL)
I haven't lost my mind - it's backed up on a tape somewhere...
Gentoo/MIPS Cobalt developer, Mozilla herd member. |
|
| Back to top |
|
|
ddc
Guru
Joined: 29 Aug 2006
Posts: 522
|
| Posted: Thu Sep 28, 2006 5:41 pm Post subject: |
|
|
i boot with "ip28(gentoo)" with this arc.cf
| Quote: |
# ARCLoad Configuration
# Some default settings...
append "root=/dev/sda3";
append "ro";
append "gentoo=noudev"; #to exclude udev, static is used
# main definition
ip28 {
# Definition for a gentoo kernel
# Select this by setting OSLoadFilename="ip28(gentoo)"
gentoo {
description "SGI Impact R10000\n\r";
image system "/gentoo";
}
# For debugging a kernel
# Select this by setting OSLoadFilename="ip28(gentoo,debug)"
debug {
description "Debug console";
append "init=/bin/bash";
}
# Select this by setting OSLoadFilename="ip28(gentoo,consoleuart)"
consoleuart
{
description "Console gfx";
append "console=ttyS0,9600";
}
# Select this by setting OSLoadFilename="ip28(gentoo,consolegfx)"
consolegfx
{
description "Console gfx";
append "console=tty0";
}
}
|
|
|
| Back to top |
|
|
Redhatter
Retired Dev
Joined: 20 Sep 2003
Posts: 548
Location: Brisbane, QLD, Australia
|
| Posted: Fri Sep 29, 2006 1:21 am Post subject: |
|
|
That should be fine. So the problem is it loops when using IMPACT console, but not when using serial?
_________________
Stuart Longland (a.k.a Redhatter, VK4MSL)
I haven't lost my mind - it's backed up on a tape somewhere...
Gentoo/MIPS Cobalt developer, Mozilla herd member. |
|
| Back to top |
|
|
ddc
Guru
Joined: 29 Aug 2006
Posts: 522
|
| Posted: Fri Sep 29, 2006 6:35 am Post subject: |
|
|
| Redhatter wrote: | | That should be fine. So the problem is it loops when using IMPACT console, but not when using serial? |
the problem is the that with the gfx-framebuffer enabled in the kernel the machine loops 4 ever.
It loops if the console is UART, it loops if the console is gfx. It always loops for ever causing me to reset and to boot a kernel without the framebuffer support.
I have noted that your indigo2 is a green one, i'm speaking about the purple one, i think it is different in the solid impact, because your kernel loops too, while you say you have booted it with success.
May be we are speaking about 2 different machines: indigo2 green and indigo 2 purple, both are R10K, or i don't know how to explain the difference ... |
|
| Back to top |
|
|
Redhatter
Retired Dev
Joined: 20 Sep 2003
Posts: 548
Location: Brisbane, QLD, Australia
|
| Posted: Fri Sep 29, 2006 10:27 am Post subject: |
|
|
Actually, the teal case is merely cosmetic. It's still an IP28 stuffed in that case. Most R10k Indigo2's are the purple colour you describe though.
Just a question... when you boot from Impact console, is the PROM set up to use Impact or serial console?
The Impact driver in Linux relies on the PROM initialising the graphic board.
_________________
Stuart Longland (a.k.a Redhatter, VK4MSL)
I haven't lost my mind - it's backed up on a tape somewhere...
Gentoo/MIPS Cobalt developer, Mozilla herd member. |
|
| Back to top |
|
|
ddc
Guru
Joined: 29 Aug 2006
Posts: 522
|
| Posted: Fri Sep 29, 2006 3:07 pm Post subject: |
|
|
| Redhatter wrote: | Actually, the teal case is merely cosmetic. It's still an IP28 stuffed in that case. Most R10k Indigo2's are the purple colour you describe though.
Just a question... when you boot from Impact console, is the PROM set up to use Impact or serial console?
The Impact driver in Linux relies on the PROM initialising the graphic board. |
i've done these 2 tests:
A) setenv console g, boot kernel with impact fgx framebuffer, PS/2 mouse & keyboard plugged
B) setenv console d, boot kernel with impact fgx framebuffer, PS/2 mouse & keyboard plugged
in case A and B the kernel halts its booting and loop for evere. I've noticed the monitor has a flashing in case A, nothing happens in case B: blank screen.
Irix boots very well and i can see it on the screen. What could be wrong ?
again, removing the gfx video board form the GIO64bus causes the system to boot fine.
i report gfx solidimpact's chips label:
chip1: SGI/ISD 1994 Vi01 REVA 099-9028-011 83G61HF0008
chip2: SGI/ISD 1994 Vi01 REVA 099-9028-011 83G61HF0008
chip3: ANALOGDEVICES ADV7162KS170 9618 OF56356 |
|
| Back to top |
|
|
Redhatter
Retired Dev
Joined: 20 Sep 2003
Posts: 548
Location: Brisbane, QLD, Australia
|
| Posted: Sat Sep 30, 2006 7:22 am Post subject: |
|
|
I can only think that there's something funny with your framebuffer board that Linux does not like.
What do you mean by "monitor has a flashing"? I've seen some monitors that go into standby mode (flashing their POWER LED) when they detect no signal, or can't sync. Does the monitor have any On-Screen Display menu, and if so, can it tell you the sync-rate of the video signal?
_________________
Stuart Longland (a.k.a Redhatter, VK4MSL)
I haven't lost my mind - it's backed up on a tape somewhere...
Gentoo/MIPS Cobalt developer, Mozilla herd member. |
|
| Back to top |
|
|
ddc
Guru
Joined: 29 Aug 2006
Posts: 522
|
| Posted: Sat Sep 30, 2006 11:26 am Post subject: |
|
|
| Redhatter wrote: | I can only think that there's something funny with your framebuffer board that Linux does not like.
What do you mean by "monitor has a flashing"? I've seen some monitors that go into standby mode (flashing their POWER LED) when they detect no signal, or can't sync. Does the monitor have any On-Screen Display menu, and if so, can it tell you the sync-rate of the video signal? |
it is what happens: i poweron, i see the prom menu, i enter the command menu, i boot gentoo, suddenly the monitor detects no signal, and the system is halted.
The nosync could be a problem of the monitor, but the boot is halted: i mean no serial console, no ssh console, no sound from the harddisk, nothing: the boot is really halted in a loop.
Here, i have 3 solidimpact boards, they are working with IRIX, they are not with gentoo.
Personally i think your SolidImpact board is different from mine, we can explain the monitor flashing (it should be a problem in video sync, Irix works, gentoo doesn't), but if i use your kernel, there is no explanation about what i see on my system: a boot completly halted: i really think is a driver problem.
when i plugged a 3com eisa board on an other indigo2 i experienced slots are not the same: irix had problems if i put this 3com board in the "wrong slot number": EISAslot1 is not the same as EISAslot2, ecc. It seems silly, but slots are numbered and i have experienced this stupid problem. So, which slot have you plugged with the gfx video board ? (yes, i know, i have experienced an EISA slot number, the gfx is GIO64, but ... just to know if this is the silly problem)
Anybody else has experienced this problem, here ?
i'd like to hear from other gentoo-forum-people, expecially from who owns a purple indigo2 impact. |
|
| Back to top |
|
|
ddc
Guru
Joined: 29 Aug 2006
Posts: 522
|
| Posted: Sun Jan 28, 2007 7:30 pm Post subject: |
|
|
my impact is well running, now i'd like to plug an IDE-HD controller in the EISA slot
has anybody expirienced this idea ?
could you suggest me a good IDE-hard-disk-controller-board to plug in the impact-EISA-slot ?
i mean a board supported by the linux IDE/ATA drivers
thank you |
|
| Back to top |
|
|
Redhatter
Retired Dev
Joined: 20 Sep 2003
Posts: 548
Location: Brisbane, QLD, Australia
|
| Posted: Sun Jan 28, 2007 11:17 pm Post subject: |
|
|
| ddc wrote: | my impact is well running, now i'd like to plug an IDE-HD controller in the EISA slot
has anybody expirienced this idea ?
could you suggest me a good IDE-hard-disk-controller-board to plug in the impact-EISA-slot ?
i mean a board supported by the linux IDE/ATA drivers
thank you |
EISA support on IP28 is practically untested. With a few patches (the URLs to which I have misplaced) you can get it to compile. My understanding is that things like DMA don't work properly (sound cards are non-functional for instance) and thus you'd have pathetically slow disk I/O throughput.
_________________
Stuart Longland (a.k.a Redhatter, VK4MSL)
I haven't lost my mind - it's backed up on a tape somewhere...
Gentoo/MIPS Cobalt developer, Mozilla herd member. |
|
| Back to top |
|
|
ddc
Guru
Joined: 29 Aug 2006
Posts: 522
|
| Posted: Tue Feb 06, 2007 8:04 am Post subject: |
|
|
| Redhatter wrote: | | ddc wrote: | my impact is well running, now i'd like to plug an IDE-HD controller in the EISA slot
has anybody expirienced this idea ?
could you suggest me a good IDE-hard-disk-controller-board to plug in the impact-EISA-slot ?
i mean a board supported by the linux IDE/ATA drivers
thank you |
EISA support on IP28 is practically untested. With a few patches (the URLs to which I have misplaced) you can get it to compile. My understanding is that things like DMA don't work properly (sound cards are non-functional for instance) and thus you'd have pathetically slow disk I/O throughput. |
octane2 (on-board) sound card is working 
indigo2 (on-board) sound card is NOT working 
i think i will build (or buy, i see it on ebay-shop @ $100+shipping) an adaptor SCSI<-->ATA
[may it be a good electronic project for my university degree ? scsi are too expensive !!!] |
|
| Back to top |
|
|
Redhatter
Retired Dev
Joined: 20 Sep 2003
Posts: 548
Location: Brisbane, QLD, Australia
|
| Posted: Tue Feb 06, 2007 10:28 am Post subject: |
|
|
Yep... I'm aware of the problems regarding HAL2 on IP28. Incidentally, it's the same result I face on the O2... if they don't crash the machine, they cause severe distortion when playing a waveform.
_________________
Stuart Longland (a.k.a Redhatter, VK4MSL)
I haven't lost my mind - it's backed up on a tape somewhere...
Gentoo/MIPS Cobalt developer, Mozilla herd member. |
|
| Back to top |
|
|
ddc
Guru
Joined: 29 Aug 2006
Posts: 522
|
| Posted: Tue Feb 06, 2007 12:53 pm Post subject: |
|
|
adaptor SCSI<-->ATA
an example, $100 ?
(also there is a guy selling it on ebay/USA )
| Quote: | ACARD AEC-7720U Ultra SCSI-To-IDE Bridge Card
This is a great solution to the cost of SCSI drives. Take your large format (and much less expensive) IDE drives and turn them into a scsi drive with this small adapter. Chain drives together for a full SCSI setup. A great solution for storage of files. The AEC-7720U comes with all the jumpers for setting it up with the SCSI ID that you want. I will also send via email a link to the manual and software for updating firmware, etc. to the winning bidder. Good luck bidding!
OverView
The AEC-7720U Ultra SCSI-IDE bridge is dedicated to general purpose bridge of SCSI (Small Computer System Interface) to IDE (Integrated Device Electronics )interface. With ACARD micro-controller core built-in, it provides more flexibility and stability for highly exploring application. The AEC-7720U supports IDE device which can be easily attached to SCSI bus.
Features
On-chip high speed ACARD RISC micro-controller.
SCSI controller, IDE controller and interfaces.
SCSI host interface synchronous transfer rate up to 20 MB/sec.
On board Flash ROM for easy firmware upgrade.
Full Ultra SCSI target features support.
Specifications
On-chip high speed ACARD RISC micro-controller
SCSI host interface synchronous transfer rate up to 20 MB/sec
On board Flash ROM for easy firmware upgrade
Selectable SCS ID from 0 to 7
Full Ultra SCSI target features support
Devices Compatibility List
Support Hard Disk:
Brand Model
IBM HDD DTLA series, DPTA series, DTTA series, DJNA series
QUANTUM HDD FireBall LM series, LC series, LB series, LD series; FireBall KX series, KA series; FireBall CX series, CR series, EX series; FireBall EL series, AS series
MAXTOR HDD 54098U8, 52049U4, 91366U4, 94098U8, 91020U3, 98196H8 DiamondMax60 series
WD HDD Expert 13BA, Caviar AC14300, WD 300BB
SEAGATE HDD Barracuda IDE series; Barracuda ATA II/III 100 series; ST320430A, ST313620A, ST3240AT, ST38422A, ST36531A
FUJITSU HDD MPF-3204AT, MPD-3173AT, MPD-3084AT |
|
|
| Back to top |
|
|
ddc
Guru
Joined: 29 Aug 2006
Posts: 522
|
| Posted: Tue Feb 06, 2007 1:02 pm Post subject: |
|
|
| Redhatter wrote: | | Yep... I'm aware of the problems regarding HAL2 on IP28. Incidentally, it's the same result I face on the O2... if they don't crash the machine, they cause severe distortion when playing a waveform. |
yes, HAL2-IP28 is terrible ! no sound from it, no love, only noise.
so personally i think i'd better plan to add an I2C-mp3-audio-decodert-chip (<$40) on the IP28-LPT and to buld I2C kernel support
Then i will cat song.mp3 to "/dev/i2c" and the chip will decode stream into audio
gosh, so should have realized a 300Gb-ip28-200Watt-mega-iPOD ?
yes, an alternative use of an obsolete machine: LOL !!!
[silly, mad, but LOL, isn't it ? ] |
|
| Back to top |
|
|
Redhatter
Retired Dev
Joined: 20 Sep 2003
Posts: 548
Location: Brisbane, QLD, Australia
|
| Posted: Mon Feb 12, 2007 1:16 am Post subject: |
|
|
Hardly worth it, when the host CPU can do it just fine. Parallel ports don't have much bandwidth.
On the SCSI->ATA thing... it might work... I've never tried.
_________________
Stuart Longland (a.k.a Redhatter, VK4MSL)
I haven't lost my mind - it's backed up on a tape somewhere...
Gentoo/MIPS Cobalt developer, Mozilla herd member. |
|
| Back to top |
|
|
ddc
Guru
Joined: 29 Aug 2006
Posts: 522
|
| Posted: Wed Mar 14, 2007 3:19 pm Post subject: |
|
|
i'd like to put a 8250-16550 uart chip board on the EISA bus
here i have got a lot of 16450 8bit PC board, i plugged one of them inside the indigo2, i enabled /device-drivers/character-devices/serial-drivers/8250-16550
/bus-optiones/EISA-support
i reboot, the kernel is instable, the indigo2 serial line stops to work, i can't see nothing on the serial console
i used ssh and i reset to
/device-drivers/character-devices/serial-drivers/IP22-zilog-8530-serial-port
the uart card should work on the EISA bus. It is not working.
i wander if we'd better rewrite EISA-driver in polling way 
i realize IRQ and DMA are reasons to failure |
|
| Back to top |
|
|
ddc
Guru
Joined: 29 Aug 2006
Posts: 522
|
| Posted: Sun Apr 13, 2008 7:57 am Post subject: |
|
|
at last, it is well working =P
here I've got this CGA CRT http://www.webalice.it/devilcoder/myphoto/mystudio.jpg (the one on the left, the orange one)
that comes with a wonderful ISA graphic card (it is very hard to find CGA now a day, but i've found it =P)
I'd like to add it into my impact as THE text console =P
I think it is a bit hard to write/adapt the driver in the linux sources =P |
|
| Back to top |
|
|
Redhatter
Retired Dev
Joined: 20 Sep 2003
Posts: 548
Location: Brisbane, QLD, Australia
|
| Posted: Sun Apr 13, 2008 9:08 am Post subject: |
|
|
Little wonder really... CGA cards rely mostly on BIOS routines, or by hooking a specific region of memory, which is directly written to by the OS/applications.
This is why you can't have multiple CGA cards in a machine for instance ... they both try to hook the same BIOS routines, and the same regions of memory.
Now the BIOS is non-existant on a SGI workstation. They have a PROM which pay sweet FA attention to the EISA bus -- that's the operating system's worry -- the PROM is only concerned with the onboard hardware. The memory region used by the CGA adaptor will probably clash with something too -- they were not designed to talk to SGI computers, or vice versa.
Likewise with the serial board -- they were designed with IBM-compatible hardware in mind, which SGI computers are anything but.
_________________
Stuart Longland (a.k.a Redhatter, VK4MSL)
I haven't lost my mind - it's backed up on a tape somewhere...
Gentoo/MIPS Cobalt developer, Mozilla herd member. |
|
| Back to top |
|
|
ddc
Guru
Joined: 29 Aug 2006
Posts: 522
|
| Posted: Sun Apr 13, 2008 11:11 am Post subject: |
|
|
ok, so i'd better build my own CGA adaptor using an CCPLD soldered to an ISA prototype board: in this case this board will be seen by kernel as a normal "memory mapped device" in where the kernel linux has only to move data. What the kernel moves to it will be putted to the video ram so ... displayed on the CRT at last. Easy.
pro:
1) no bion requited
2) handling the graphic board means handling an easy memory mapped device
3) you can allocate the smartest address to the device, so the SGI mon should be happy
4) no DMA, so no futile complicated add on: we are talking about a slow text only console
5) no interrupt design
=P |
|
| Back to top |
|
|
ddc
Guru
Joined: 29 Aug 2006
Posts: 522
|
| Posted: Thu Apr 24, 2008 9:26 pm Post subject: |
|
|
could you suggest me good memory addresses in where i coudl map my own developed device for the indigo2 isa bus ?
is there any documentation about it ? |
|
| Back to top |
|
|
Redhatter
Retired Dev
Joined: 20 Sep 2003
Posts: 548
Location: Brisbane, QLD, Australia
|
| Posted: Thu Apr 24, 2008 10:17 pm Post subject: |
|
|
No idea... you're asking highly technical questions about the hardware implementation of the IP28 that are so far over my head it isn't funny.
We support running Linux on the IP28.
We don't support hardware hacking of the IP28.
_________________
Stuart Longland (a.k.a Redhatter, VK4MSL)
I haven't lost my mind - it's backed up on a tape somewhere...
Gentoo/MIPS Cobalt developer, Mozilla herd member. |
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
Gentoo on Alternative Architectures
