English Amiga Board


Go Back   English Amiga Board > Support > support.Hardware

 
 
Thread Tools
Old 12 June 2024, 18:07   #21
pat_ctrl
Registered User
 
Join Date: Jun 2024
Location: London/UK
Posts: 5
I'm keen to help test this and provide feedback etc.
I have an A4000D with Prometheus Firebird and an A1200 with Mediator (original version...)
Also Radeon 9250 and Voodoo 4 cards too (I'm still hoping we can get Voodoo 4 on Prometheus, but not sure what the limitations are there)
pat_ctrl is offline  
Old 12 June 2024, 20:14   #22
Hedeon
Semi-Retired
 
Join Date: Mar 2012
Location: Leiden / The Netherlands
Posts: 2,064
Quote:
Originally Posted by pat_ctrl View Post
(I'm still hoping we can get Voodoo 4 on Prometheus, but not sure what the limitations are there)
Did you ever try it?

Not that is matters, V4 is running in V3 compatibility mode on OS3.
Attached Thumbnails
Click image for larger version

Name:	Voodoo4Prm.jpg
Views:	106
Size:	105.7 KB
ID:	82466  

Last edited by Hedeon; 12 June 2024 at 20:26.
Hedeon is offline  
Old 14 June 2024, 02:02   #23
grelbfarlk
Registered User
 
Join Date: Dec 2015
Location: USA
Posts: 3,003
Just for a test I put a Voodoo 4 on the Prometheus, it starts up just fine, workbench loads with the correct screenmode, however trying to start the Ragnarok, it complains "Error during MMU Set Up of PPC."
Which I would guess this is the problem, the 32MB Voodoo 4 is using up a whole 256MB?
PPCInit a second time and it complains "Could not allocate sufficient PCI memory"
Code:
Internal Configuration Space infos
VendorID : $121a 
DeviceID : $9
ClassRevision : $03000001 
PCI_REVISION_ID $01
Header Type : $00 (0=Normal/1=Bridge/2=Cardbus)
PCI_SUBSYSTEM_ID $ffff
Command : $0003 
I/O space PCI_COMMAND_IO
Memory space PCI_COMMAND_MEMORY
PCI_LATENCY_TIMER : $00
----------------------------------
Pci Device Slot Num : 1, Func Num : 0 
Vendor   : 3dfx Interactive Inc 
Device   : Voodoo5 
VendorID : $121a 
DeviceID : $9
ClassCode: 0x030000
IRQNumber: 5
MemAddr0 : $68000000 - $6fffffff 131072 kbytes
MemAddr1 : $60000000 - $67ffffff 131072 kbytes
IOAddr2  : $7fe01000 - $7fe010ff 256 bytes
RomAddr  : $70010000 - $7001ffff
RomSize	 : 65536 bytes 64 kbytes
No Bus Master Capable
Card PCI_STATUS
Support 66 Mhz PCI 2.1 bus
Accept fast-back to back
What's Enabled in the Card ?
I/O space PCI_COMMAND_IO
Memory space PCI_COMMAND_MEMORY
----------------------------------
Internal Configuration Space infos
VendorID : $1057 
DeviceID : $480b
ClassRevision : $06000002 
PCI_REVISION_ID $02
Header Type : $00 (0=Normal/1=Bridge/2=Cardbus)
PCI_SUBSYSTEM_ID $1057
Command : $0007 
I/O space PCI_COMMAND_IO
Memory space PCI_COMMAND_MEMORY
Bus mastering PCI_COMMAND_MASTER
PCI_LATENCY_TIMER : $00
----------------------------------
Pci Device Slot Num : 2, Func Num : 0 
Vendor   : Motorola 
Device   :  
VendorID : $1057 
DeviceID : $480b
ClassCode: 0x060000
IRQNumber: 0
MemAddr0 : $70003000 - $70003fff 4 kbytes
MemAddr1 : $70004000 - $70004fff 4 kbytes
MemAddr2 : $78000000 - $7fffffff 131072 kbytes
MemAddr3 : $70040000 - $7007ffff 256 kbytes
Bus Master Capable
Card PCI_STATUS
Support 66 Mhz PCI 2.1 bus
Accept fast-back to back
What's Enabled in the Card ?
I/O space PCI_COMMAND_IO
Memory space PCI_COMMAND_MEMORY
Bus mastering PCI_COMMAND_MASTER
2048bytes MEM_PCI DMA Mem Allocated : 61bfd000
16384bytes MEM_PCI DMA Mem Allocated : 61bff000
Free Mem 2048
2040bytes MEM_NONCACHEABLE DMA Mem Allocated : 61bfd000
Free Mem 2040
4096bytes MEM_NONCACHEABLE DMA Mem Allocated : 61c04000
Free Mem 4096
Free Mem 16384
300000 bytes MEM_NONCACHEABLE DMA Mem Allocated : 61bfd000
Free Mem 300000

Last edited by grelbfarlk; 14 June 2024 at 02:43.
grelbfarlk is offline  
Old 14 June 2024, 14:13   #24
Hedeon
Semi-Retired
 
Join Date: Mar 2012
Location: Leiden / The Netherlands
Posts: 2,064
I think we entered the realm of off-topic.
Hedeon is offline  
Old 24 June 2024, 12:02   #25
phx
Natteravn
 
phx's Avatar
 
Join Date: Nov 2009
Location: Herford / Germany
Posts: 2,569
Quote:
Originally Posted by ncafferkey View Post
I have that too
I would be interested in the GREX SDK as well. Was it made available? I don't see it in the Zone.
phx is offline  
Old 02 July 2024, 22:56   #26
steenl
Registered User
 
Join Date: Jan 2020
Location: Herlev/Denmark
Posts: 25
Quote:
Originally Posted by Wrangler View Post
(Yes, I know there is Grex as well but how many of those actually exist...)

I have one
steenl is offline  
Old 04 August 2024, 18:28   #27
Thomas Richter
Registered User
 
Join Date: Jan 2019
Location: Germany
Posts: 3,390
Ok, and the story goes on. I was busy during the last weeks, two business trips, but
also lots of development on the openpci.library development.

Here it is, the next version - please test if possible.

https://eab.abime.net/attachment.php...1&d=1722788498

It extends support to the A1200 mediators - this requires the mmu.library for bankswitching the PCI windows - the Grex, the original Prometheus and the Firestorm. The A4000 Mediator was supported already.

There are still a couple of restrictions you find in the readme, and please also follow the instructions in the readme. It also contains a bit more details on some design choices.

In particular, it contains a "PCIInit" tool that goes into the MMU-Configuration file to cut down the MMU setup time by the help of the openpci.library. This implementation also creates new autoconfig IDs for the bridges and the PCI devices connected to it. Again, more detalis in the readme in the archive.

Have fun!
Attached Files
File Type: lha openpci.lha (10.4 KB, 39 views)
Thomas Richter is offline  
Old 04 August 2024, 18:33   #28
Thomas Richter
Registered User
 
Join Date: Jan 2019
Location: Germany
Posts: 3,390
Quote:
Originally Posted by grelbfarlk View Post
Just for a test I put a Voodoo 4 on the Prometheus, it starts up just fine, workbench loads with the correct screenmode, however trying to start the Ragnarok, it complains "Error during MMU Set Up of PPC."
Which I would guess this is the problem, the 32MB Voodoo 4 is using up a whole 256MB?
I have currently no idea of the PPC and what it requires. However, one particular problem may be that the current openpci.library from my end does not provide mechanisms to cut off memory from the graphics board to act as bounce buffers. You find details on why this is so in the readme in the archive in my post directly above.


In addition, please note that the openpci.library *only* drills holes for PCI hardware it finds connected to the board. Thus, if the PPC or the hardware connected to it requires *additional* room in the 68K address space, this needs to be accquired correctly through the mmu.library. Patching mmu tables manually, or simply assuming that the entire 512MB of the prometheus addess space is accessible is not going to work.
Thomas Richter is offline  
Old 04 August 2024, 21:23   #29
fgh
Registered User
 
Join Date: Dec 2010
Location: Norway
Posts: 841
All pci bridges supported already, great job!

An ‘apocalypse’ (G4/450mHz) is around €100, plus the carrier board. I would donate €50 towards getting you one, I’m sure others would join.
fgh is online now  
Old 04 August 2024, 22:15   #30
Thomas Richter
Registered User
 
Join Date: Jan 2019
Location: Germany
Posts: 3,390
Thanks for the great offer - though there is still a long way ahead of me with this library at the moment. Feedback of course greatly appreciated.
Thomas Richter is offline  
Old Yesterday, 04:49   #31
grelbfarlk
Registered User
 
Join Date: Dec 2015
Location: USA
Posts: 3,003
Quote:
Originally Posted by Thomas Richter View Post
I have currently no idea of the PPC and what it requires. However, one particular problem may be that the current openpci.library from my end does not provide mechanisms to cut off memory from the graphics board to act as bounce buffers. You find details on why this is so in the readme in the archive in my post directly above.


In addition, please note that the openpci.library *only* drills holes for PCI hardware it finds connected to the board. Thus, if the PPC or the hardware connected to it requires *additional* room in the 68K address space, this needs to be accquired correctly through the mmu.library. Patching mmu tables manually, or simply assuming that the entire 512MB of the prometheus addess space is accessible is not going to work.

We split this off into a different thread and Hedeon provided a fix here:
https://eab.abime.net/showthread.php?t=76633&page=277
grelbfarlk is offline  
Old Yesterday, 22:03   #32
Thomas Richter
Registered User
 
Join Date: Jan 2019
Location: Germany
Posts: 3,390
A couple of fixes later....

PCIInit was outright corrupt. Which is strange since I was pretty sure I tested it.
If a MMU-requiring board was used in conjunction with PCIInit, this would not have worked.
If the mmu.library was installed but no MMU was found, the library crashed.

The latter fix I haven't really had time to debug fully, though it might be working.

https://eab.abime.net/attachment.php...1&d=1722888029
Attached Files
File Type: lha openpci.lha (10.5 KB, 12 views)
Thomas Richter is offline  
 


Currently Active Users Viewing This Thread: 2 (1 members and 1 guests)
Docent
Thread Tools

Similar Threads
Thread Thread Starter Forum Replies Last Post
What do you call yourself if you're not a retrogamer Starglider 2 Retrogaming General Discussion 47 06 February 2018 08:15
WANTED: PCI Sound Card VORTEX2 SuperQuard / SQ2500 or Turtlebeach SantaCruz pci card keropi MarketPlace 12 15 March 2017 02:48
Let's call it karma. Or something. trooper Retrogaming General Discussion 10 19 September 2013 23:44
Ports of Call Josh request.Old Rare Games 9 21 February 2010 23:32
Prometheus PCI & Voodoo 3 PCI GFX Card Slayer support.Hardware 21 05 September 2006 10:57

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +2. The time now is 00:18.

Top

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.
Page generated in 0.17885 seconds with 16 queries