[Openchrome-users] Support for S3 Graphics DeltaChrome, GammaChrome and ChromeS20 generation

[Xavier Bachelot]

Luc Verhaegen wrote:
> On Tue, Mar 20, 2007 at 12:10:00PM +0100, Stefan Bassing wrote:
>> Hello!
>>
>> Thanks for your reply. I was just making those lspci outputs. The vendor
>> and device IDs you have discovered are correct. Nonetheless, S3 Graphics
>> is the developer of the graphic cores of the IGPs from VIA and K8M890 is
>> just the integrated version of DeltaChrome S8 oder GammaChrome S18
>> (which is in fact identical to ChromeS27). So in my somehow simple
>> thoughts, it could work if the openchrome driver is somehow (I think you
>> know a lot more about this than me) modified that it accepts the vendor
>> and device IDs for the K8M890 part of the driver. As I said before, I'm
>> really no expert in programming, especially with Linux, but could it be
>> possible to make some experimental versions of that specific driver
>> parts for testing?
>>
>> Docs from S3 Graphics are as non-existant as from VIA. Same policy... If
>> it would help you or another developer, I would send you my DeltaChrome
>> S8 card for devloping a working 2D driver!
>>
>> Anyway here the outputs from lspci -vn (made with Knoppix 5.1):
>> DeltaChrome S8, AGP 8x, 256 MB RAM:
>> 01:00.0 0300: 5333:8e00 (rev 01) (prog-if 00 [VGA])
>>        Subsystem: 5333:8e00
>>        Flags: bus master, 66MHz, medium devsel, latency 32, IRQ 11
>>        Memory at fde00000 (32-bit, non-prefetchable) [size=512K]
>>        Memory at e0000000 (32-bit, prefetchable) [size=256M]
>>        [virtual] Expansion ROM at fde80000 [disabled] [size=64K]
>>        Capabilities: [dc] Power Management version 2
>>        Capabilities: [80] AGP version 3.0
>>        Capabilities: [c0] Message Signalled Interrupts: Mask- 64bit-
>> Queue=0/0 Enable-
>>
>> Chrome S27, PCI-Express, 128 MB RAM:
>> 02:00.0 0300: 5333:8e48 (rev 01) (prog-if 00 [VGA])
>>        Subsystem: 5333:0122
>>        Flags: bus master, fast devsel, latency 0, IRQ 10
>>        Memory at fe980000 (32-bit, non-prefetchable) [size=512K]
>>        Memory at d0000000 (32-bit, prefetchable) [size=256M]
>>        Expansion ROM at fe970000 [disabled] [size=64K]
>>        Capabilities: [dc] Power Management version 2
>>        Capabilities: [40] Express Legacy Endpoint IRQ 0
>>        Capabilities: [c0] Message Signalled Interrupts: Mask- 64bit-
>> Queue=0/0 Enable-
>>
>> Perhaps you could do something with this...
>>
>> Please let me know if I could help in any way!
>>
>> greetings
>>
>> Stefan
>>
>> P.S.: Sorry for sending this first directly! I was a bit confused with 
>> the handling of the mailing list.
>>
> Feel free to hand me some rom dumps: look in 
> http://unichrome.cvs.sourceforge.net/unichrome/utils/, there's a rather 
> simplistic dumper there.
> 
> I knew from BIOS images that k8m890 was highly compatible (basic init 
> and modesetting only of course), days before VIA finally released code. 
> I doubt that this will be repeatable with the discrete cards. My gut 
> feeling is that even basic init and modesetting will be different.
> 
> So do get me the images, i'll have a quick look to see whether this is a 
> full image. If not, i will, at some point, provide you with a pci 
> aperture based rom dumper, as i've been thinking about doing this for a 
> longer while.
> 
> I will not suddenly pool a massive amount of time in this. Even though 
> i'm, by now, very experienced disassembling VIAs roms, this always takes 
> a lot of time. There will hopefully be a direct relation between the 
> roms for all these cards.
> 
> So i will not go and do this right now, but if i don't get those roms, 
> there is absolutely no chance that anyone will go and verify these 
> things. Well, in a meaningful manner that is.
> 
> Luc Verhaegen.
> http://unichrome.sf.net/
> 

Does this mean you found a way around your usual -ENOHW or you're just 
feeling alone in unichrome lists ?

May I again repeat the split between unichrome and openchrome is a 
massive waste of energy ? May I also repeat that mpeg2 hw decoding is 
not just chrome for at lot of ppl out there ? And may I also repeat that 
bad hack to support hardware is still better than no support at all, at 
least temporary until someone do The Right Thing (w/o breaking all the 
not so right things, obviously) ?

I really don't understand what you're trying to achieve...

X.