Difference between revisions of "Flameman/sgi"
< Flameman
(→kernel status) |
(→octane2) |
||
| Line 132: | Line 132: | ||
* Other XIO-based devices like MENET and various Impact addons remain | * Other XIO-based devices like MENET and various Impact addons remain | ||
* untested and are not guaranteed to work. This applies to various digital video conversion boards as well. | * untested and are not guaranteed to work. This applies to various digital video conversion boards as well. | ||
| + | |||
| + | === Xbow === | ||
| + | * [[[http://www.nekochan.net/wiki/XBow]]] | ||
=== kernel status === | === kernel status === | ||
Revision as of 12:57, 24 March 2015
For more interesting projects done by Flameman&Legacy, be sure to check out his project index
Contents
machine
indy
| 8 PinMini DIN (SGI) | DB25 Plug | Name |
|---|---|---|
| 4 | 7 | GND |
| 5 | 2 | TD |
| 3 | 3 | RD |
| 1 | 8 | CD |
| 7 | 4 | DTR |
add on board support
| board over gio32 | kernel support |
|---|---|
| scsi add on | dunno |
| phobos g130 | dunno |
note: if you inclide the gfx framebuffer support in the kernel, then you CAN NOT boot with missing gfx hw board! The kernel is expecting to find it, if it is missing it will panic!
Suggestion: do not include the framebuffer, use uart console only
o2
kernel status
| kernel | note |
|---|---|
| 2.6.37-rc8 | working, under testing, note1 needs to be applied |
- note1
removing - Tickless System (Dynamic Ticks) - High Resolution Timer Support - SCSI generic support - RAID Transport Class - Verbose SCSI error reporting (kernel size +=12K) - SCSI logging facility forcing - No Forced Preemption - SCSI_WAIT_SCAN compiled as builtin (it was an hidden module)
add on board support
| board over pci | kernel support |
|---|---|
| _ | dunno |
Known working PCI cards with IRIX
- SGI PCI Serial I/O a.k.a. PCI SIO UFC, part number 030-1657-003, Provides two RS-232/RS-422 serial ports. The board is slotted for universal voltage, but SGI only provides drivers for Onyx 300/350/3000 systems. The SGI marketing code is PCI-SER-10002.
- SGI PCI Digital Audio Board <also known as the RADical Audio board>. The SGI part number for the example shown is 030-1649-001, it appears in an hinv with PCI vendor ID 0x10a9(SGI) and a device ID of 0x0005(RAD Audio):
- SGI PCI Gigabit Ethernet:, 1000 Base SX (fiber optic interface) and 1000 Base T (copper interface) - shown mounted in an O2k XIO Shoehorn <a PCI-to-XIO slot adapter>: Both are keyed as 'universal' PCI cards to fit in either 3.3 or 5 volt slots.
- SGI O2 10/100BaseT Network Card for O2: SGI marketing code SGI_PCI-ENET-TX-1P; part number 9210105
- SGI USB PCI card, SGI part number 9210286: NEC chipset - Slotted for 3.3 and 5V PCI applications. SGI equipped Onyx 350 InfiniteReality and InfinitePerformance systems with this card as an attachment point for USB keyboards and mice. There are some references on TechPubs indicating Orange Micro as the OEM, while "UP205-0525" screened on the PCB returns Adaptec references. It appears in an hinv with a PCI vendor ID of 0x1033 (NEC) PCI device numbers 0x0035(Dual OHCI controllers plus Single EHCI controller) and 0x00e0(USB 2.0 Host Controller).
- DATAFIRE PCI 1ST ISDN card SGI PN: 9210111 Came inside one O2, SGI part number 9981028 SGI
See more here [1]
impact
add on board support
| board over gio64 / eisa | kernel support |
|---|---|
| _ | dunno |
fuel
linux is not running yet, but OpenBSD is going to run
octane2
Things to keep in mind when building kernel for an SGI Octane
- Impact (MGRAS) console and X driver work, please report any bugs.
- VPro (Odyssey) console works, but no X driver exists yet.
- PCI Card Cages should work for many devices, except certain types like
- PCI-to-PCI bridges (USB hubs, USB flash card readers for example).
- Do not use OHCI-based USB cards in Octane. They're broke on this machine.
- Patches are welcome to fix the issue.
- Equally, UHCI Cards are showing issues in this release, but should still function somewhat. This issue manifests itself when using pl2303 USB->Serial adapters.
- Other XIO-based devices like MENET and various Impact addons remain
- untested and are not guaranteed to work. This applies to various digital video conversion boards as well.
Xbow
- [[[2]]]
kernel status
it has been tested on this machine
>> hinv
System: IP30
Processor: 400 Mhz R12000, 2M secondary cache
Processor: 400 Mhz R12000, 2M secondary cache, (cpu 1)
Memory size: 1536 Mbytes
Network: ef0 ethernet (100/10 base-T)
Audio: RAD Audio Processor
PCI Bus: XTALKPCI (13)
PCI Device: slot 1 vendor 0x1033 part 0x35 rev 67
Graphics: ESI
Graphics: ODYSSEY
Adding/removing ESI, ODYSSEY
| kernel | note |
|---|---|
| 2.6.17 | pretty stable, used in production, 2xR12K and 2Gbyte of ram, ram should be reduced to 1.5Gb in order to use PCI, also IOC-serial-kb must be removed from the kernel, about the ram problem i think the actual bug is in the pdev_to_baddr() function in arch/mips/include/asm/mach-ip30/dma-coherence.h, 32-bit PCI are affected. Funny enough, the BaseIO devices, like the IOC3 or SCSI, are not affected by this, while PCI-USB is affected. |
| 2.6.19 | not stable as 2.6.17, issue when executed on system with 2Gbyte of ram, PCI-cadge not working, issues with SMP |
| 2.6.20.* | may be work, not tested |
| 2.6.25 | busted, stay away |
| 2.6.26 | may be there is a patch, but IOC3 is hardly issued, SMP dead |
| 2.6.29 | bugs in the scheduler, not booting on 2xR12K |
| 3.x.x @https://github.com/jaaron/linux-mips-ip30 | system boots but IOC3 was still dead, as was SMP |
| 2.6.17.4 | 64-bit kernel support | Yes | |
| Memory controller support | Yes | support for 1.5GB | |
| Interrupt support | Yes | working | |
| SMP support | Yes | 2xR12K | |
| Eth0 | Yes | ||
| PCI expansion | Yes | working with USD NEC EHCI | |
| ram rootfs | Yes | uclibc & glibc working | |
| rootfs | Yes | 2006 mips3, mips4 are working | |
| 2.6.29.1 | 64-bit kernel support | Yes | |
| Memory controller support | Yes | support for 1.5GB | |
| Interrupt support | Yes | not pretty working, also scheduling while atomic issue | |
| SMP support | Yes | 2xR12K does not boot, 1xR12K boots | |
| Eth0 | Yes | ||
| PCI expansion | Yes | not tested | |
| ram rootfs | Yes | uclibc & glibc working | |
| rootfs | Yes | 2006 mips3, mips4 are working |
last kernel 2.6.29
Option? 5 Command Monitor. Type "exit" to return to the menu. >> bootp(): Setting $netaddr to 192.168.1.3 (from server ) Obtaining from server 13597184+304112 entry: 0xa800000020009cd0 - dead -
It does not work on 2xR12K@400Mhz, it boots on 1xR12K system
dmesg
last working kernel 2.6.17-r14
kernel 2.6.17-r14, we are stopped at the year 2006, running a 2008 stage3 rootfs
status
| CPU and memory | 64-bit kernel support | Yes | |
| Memory controller support | Yes | support for >1GB added in 2.6.12-R22 | |
| Interrupt support | Yes | ||
| SMP support | Yes | ||
| XIO support | Basic mapping | WP | work in progress - XIO API |
| Device discovery | WP | it basically works, but waits for XIO API | |
| Flow control | WP | soon to come in the new XIO API | |
| PCI support | Basic mapping | Yes | |
| Device discovery | Yes | ||
| 64-bit DMA | Yes | ||
| 32-bit DMA | Yes | ||
| System bridge | Yes | ||
| Additional bridges | Yes | ||
| IOC3 support | Ethernet | Yes | also for IP27 |
| SuperIO meta-device | Yes | also for IP27 | |
| Serial ports | Yes | also for IP27 | |
| RS232/RS422 mode switch | No =( | low priority task, but no real problems | |
| Parallel port | No =( | unavailability of non-standard cable for testing | |
| Real Time Clock | Yes | ||
| Software power down | Yes | ||
| Voltage monitor | No =( | low priority task, not enough information | |
| LEDs | Yes | ||
| PS/2 ports | Yes | ||
| SCSI support | Internal devices | Yes | |
| External devices | Yes | ||
| RAD1 support | Data DMA support | Yes | |
| Status DMA support | Yes | ||
| Analog output | Yes | ||
| Analog input | Yes | ||
| AES output | Yes | ||
| AES input | Yes | ||
| ADAT output | No =( | no test equipment | |
| ADAT input | No =( | no test equipment | |
| ALSA PCM support | Yes | please report bugs! | |
| ALSA mixer support | Yes | ||
| ImpactSR support | Linux console support | Yes | |
| MPlayer output support | Yes | ||
| X Window support | Yes | shadowfb driver | |
| 24-bit depth setup | Yes | ||
| Video mode change | No =( | not enough information | |
| Image DMA | Yes | ||
| Hardware cursor | Yes | ||
| 2D acceleration | Yes | ||
| 3D acceleration | No =( | not enough information | |
| Flow control | No =( | not enough information, lack of XIO flow control | |
| Multiple cards | No =( | lack of test hardware | |
| VPro support | Linux console support | Yes | |
| X Window support | No =( | complexity of accelerated X drivers | |
| 24-bit depth setup | Yes | ||
| 2D acceleration | Yes | ||
| 3D acceleration | Yes | ||
| Image DMA | No =( | it's not really easy |
Other XIO-based devices like MENET and various Impact addons remain untested and are not guaranteed to work. This applies to various digital video conversion boards as well.
PCI Card Cages
- broken OHCI-based USB cards
- partially broken UHCI-based USB cards.
dmesg
add on board support
| board over xio24 / pci cartridge | kernel support |
|---|---|
| v6 over xio24 | 2.6.17: no X11 support, pretty text console |
| impactSR over xio24 | 2.6.17: pretty X11 support, pretty text console |
What is the state of hardware support on this machine?
- V6, V8 gfx: X11 support doesn't exist, and likely won't for a very long while. Stan (the IP30 Port Author) recently got console mode running on this card, but from what I understand, this is a rather complex piece of video hardware. Remote X works, however.
- Impact gfx: X11 is working
- Onboard sound works well enough from what I hear. I tested several MP3s from console via mpg123, and they worked well. A new patch coming out next week (Jul 23-24) will enable support for the optical inputs/outputs for AES (ADAT capabilities will not be supported).
- Onboard Scsi works like a charm. It wasn't entirely usable about 3 weeks ago due to reliance on the old qlogicisp driver, but with some very recent fixes to qla1280, it has replaced qlogicisp. I now run a RAID5 array using 3 50G seagate drives on my Octane, and so far, everything works great (and hdparm reports ~17.4MB/s throughput).
- Onboard Ethernet works fine too. The IOC3 driver was recently re-written to make IOC3 more of a Bus device with peripherals hanging off of it, which makes IOC3 less of a nightmare device from a coder's standpoint (but not by much).
boot examples
dual boot, idea
It should be possible if you set up IRIX as per normal and leave space on the drive for Linux. Install Linux as per the guide except when you get to set up the PROM. you'll want to leave that as is. Then you'd set up arcload to boot up Linux on demand.
- To boot IRIX, just switch the machine on.
- To boot Linux, drop to the monitor prompt and run boot -f arcload.
