Jump to content

Radeon Xpress 200M

thedopefishlives

By thedopefishlives,
in ATi

 Share
483 posts in this topic

Recommended Posts

dasepsilon
dasepsilon
Posted
Posted
Wow! You works as professinal programmer!

 

indeed, but i am being paid for programming in delphi (you see,

i really do any {censored} for money) and i never programmed hardware.

 

Is it warning or error? Does fglrx work?

 

yes, fglrx seems to work fine, i suppose it is a warning.

i counldnt find informations on that message though.

 

good to see that you're aware of all the linux sources :)

i'd really like to help, but am stupid for hardware programming.

and this notourious notimebug doesn't make it better :P:)

ole2
ole2
Posted
Posted
indeed, but i am being paid for programming in delphi (you see,

i really do any {censored} for money) and i never programmed hardware.

yes, fglrx seems to work fine, i suppose it is a warning.

i counldnt find informations on that message though.

 

good to see that you're aware of all the linux sources :P

i'd really like to help, but am stupid for hardware programming.

and this notourious notimebug doesn't make it better :unsure::wacko:

 

for the participants like dasepsilon

there will be a manual soon, on how to help us with: testing and/or development/research.

Slice
Slice
Posted
Posted

At last I repaired my laptop and continue working.

Yes, ATISlice doesn't resolve the main problem. As I can't boot with F8,size value more then 4Mb so I initialize registers too late. I try to increase priority but not successful. Linux has no such a problem. In real macs the initializing performed by EFI. Thats why some users can boot only on second attempts. First boot - registers initializing. Second boot without hardware reset - using its.

 

I need help. Any ideas on another way to work. Be realistic, I can't rewrite the driver from initial.

dasepsilon
dasepsilon
Posted
Posted
At last I repaired my laptop and continue working.

Yes, ATISlice doesn't resolve the main problem. As I can't boot with F8,size value more then 4Mb so I initialize registers too late. I try to increase priority but not successful. Linux has no such a problem. In real macs the initializing performed by EFI. Thats why some users can boot only on second attempts. First boot - registers initializing. Second boot without hardware reset - using its.

 

I need help. Any ideas on another way to work. Be realistic, I can't rewrite the driver from initial.

 

we all depend on you! :)

 

i hope the "manual" is under way, so that illiterate people like me have

a chance to understand why what needs to happen when.

if it is "just" registers, cant we patch the .kext directly by replacing

superflous parts?

Slice
Slice
Posted
Posted
we all depend on you! :lol:

 

i hope the "manual" is under way, so that illiterate people like me have

a chance to understand why what needs to happen when.

if it is "just" registers, cant we patch the .kext directly by replacing

superflous parts?

I can patch any part of any kext. But I can't find what in particular.

 

As I know koverg patch make any Radeon as RV380 chip. It is not good.

I propose you to try other patch in which assumed that DevID=5955 is chip RV420.

Else ideas?

Slice
Slice
Posted
Posted
sorry, i misunderstood you. i thought you said that you cannot

influence the "order" of the kext-loading, but the registers are more or less ok.

I can influence more or less but I found that GL.bundle uses previuos value so my ATISlice loaded too late.

Slice
Slice
Posted
Posted

I have a great advance! :(

OpenGL Driver Monitor no more crashes. Mouse no more hangs at dock.

But without AGP driver I have texture memory=0 so no 3D.

A key in IOPCIFamily that might be version 1.7 but not 2.1. Why? Because of quite different interrupt resolver.

I use MacOSX 10.4.6

IOPCIFamily 1.6 (from apple sources v32). v1.7 differ for multiprocessor support

http://forum.insanelymac.com/index.php?act...st&id=20059

ATI1Slice.kext 1.0.6 - name changed for boot priority

ATINDRV.kext v1.4.26 - don't work but influence in strange way

Callisto, CallistoHAL 003fixed - yes, now it works with the configuration

ATIRadeon8500 patched for my 9000IGP, you might use appropriate driver.

GA.plugin corrected for CallistoFB

IONDRVSupport.kext 1.4.3

 

It is not final solution for me because I need AGP driver compatible with the configuration.

Your tests?

ATI1Slice106.kext.zip

Slice
Slice
Posted
Posted

Where are the testers? Where are your reports? :jerry:

I found final solution. :rolleyes:

First of all you need to replace IOPCIFamily.kext to Chun-Nan version

http://forum.insanelymac.com/index.php?showtopic=81048

Next step depends on your testing because I have no such cards.

My Radeon9000IGP works with my patched 8500.kext.

Full report will be soon at new thread.

dasepsilon
dasepsilon
Posted
Posted

i'd like to test but i'm quite irritated where to get what version of why.

is callisto needed or not? i've killed it. what patches apply to .GA?

where do i get IONDrv 1.4.3? if you have spare time left, a .zip on

rapidshare or whatever would do magic for us humble users...

nonetheless, i'll try to replace as much as i can do by hand, but time is

short as always.

Slice
Slice
Posted
Posted
i'd like to test but i'm quite irritated where to get what version of why.

is callisto needed or not? i've killed it. what patches apply to .GA?

where do i get IONDrv 1.4.3? if you have spare time left, a .zip on

rapidshare or whatever would do magic for us humble users...

nonetheless, i'll try to replace as much as i can do by hand, but time is

short as always.

Look here to find ATI packages with old IONDRV and OpenGL.framework

http://www.insanelymac.com/lofiversion/index.php/t36886.html

9700 patched I upload in the thread before.

Or you can use Koverg patch (discussed before)

Or use non-patched 9700 - it is not good, but...

Callisto is not needed because it performs other function. If you want you can install it or no. Same link.

dasepsilon
dasepsilon
Posted
Posted

mmh, i havent have all needed kexts yet,

but the driver at least reports QE/CI and 64mb vram

i do not have the 20s-wait-bug, but the overall

performance is slideshow-ish. i will continue trying

on sunday. nonetheless THANKS slice, you're getting

very near :unsure: and special thanks for pointing out the

pc-card driver-- i bough new wlan pc-card yesterday,

and thanks to the pccard driver it was instantanously

recognized as airport! :D

GoldenChaos
GoldenChaos
Posted
Posted

My system is so wonderfully stable right now, not even the prospect of QE/CI is going to make me fiddle with it. If reports are good from a few people, I'll try the next advancement. It sounds like it's really getting close, and I'm astounded - just earlier you were talking summer 2008!

 

Something to consider, once you know you have a working solution, is to make a basic installer with all the necessary kexts to make this easy for all x200m users.

 

One question, however: Do you actually have 2D and 3D acceleration working (with some sort of performance boost), or is it just saying that it has QE/CI and not actually functioning with it?

BlackCH
BlackCH
Posted
Posted

Testing on a Radeon mobility 7000. your kext is loading (I put my device ID on the plist), but still cannot see any performance boost. The card in not CI capable; and not shure about QE, but at least I could have GL acceleration if I get it working... actually I dont know if this driver is meant to work with the 7000.

When booting on verbose I can see that your kext is loading but theres nothing in the system.log file

Anyway; I think I will reinstall the 10.4.6 for testing, right now I have 10.4.11 and when trying to load the Atiradeon.kext I get the superclass changed? error

 

wonderfull job; thanx

DocDoom
DocDoom
Posted
Posted

i tried this combos so far:

 

MAC OS 10.5.1 ToH

Callisto 03 fixed

ATI1Slice.kext (changed PCIIDPrimaryMatch to 5975, done chmod,chown)

IOPCIFamily from Chun Nan (done chmod,chown)

 

result: booting with callisto loaded correctly but only with 8MB VRAM instead of 16MB, ATI1Slice is loaded but no QE/CI, resolution change working

 

MAC OS 10.5.1 Toh

Callisto 03 fixed

ATI1Slice.kext (changed PCIIDPrimirayMatch to 5975 done chmod,chown)

IOPCIFamily from Chun Nan (done chmod, chown)

ATIRadeon9700.kext from your last post (edited PCIIDPrimaryMatch to 5975 myself, done chmod, chown)

 

result: booting into graphic mode, graphic faluires all the time, scrolling is nearly impossible, only 8MB VRAM

 

MAC OS 10.5.1 Toh

ATI1Slice.kext (changed PCIIDPrimirayMatch to 5975 done chmod,chown)

IOPCIFamily from Chun Nan (done chmod, chown)

ATIRadeon9700.kext from your last post (edited PCIIDPrimaryMatch to 5975 myself, done chmod, chown)

 

result: booting, no QE/CI, Radeon9700.kext loaded, ATI1Slice.kext loaded. same result as using no graphicdriver at all. 128MB VRAM no resolution change.

 

 

Greetings

Doc

ole2
ole2
Posted
Posted
I've tried the following combination:

1. OS X 10.4.6 JaS

2. Latest Chun-Nan's IOPCIFamily.kext

3. ATIRadeon9700.kext from post #181

4. ATI1Slice.kext ver.1.06

5. ATISlice.kext ver.1.05

... and got kernel panic. What's wrong?

 

Here are screenshots.

 

dvavd, try following sequence of steps:

 

step1: download IOPCIFamily.kext redeveloped by Chun-Nan from here:

http://forum.insanelymac.com/index.php?showtopic=81048

install his kexts by replacing originals (don't forget to backup them first)

restart and make sure you have no kernel panic

and check for side effects, describe them here in public thread

 

only if step1 passed OK proceed with step2

 

step2: download ATI1Slice.kext developed by Slice from here:

http://forum.insanelymac.com/index.php?sho...st&p=582921

apply extracted PCIID to ATI1Slice.kext

keep Chun-Nan's IOPCIFamily.kext

restart and check for side effects, describe them here in public thread

BigBoi
BigBoi
Posted
Posted

Here we go. My video card is onboard ATI X1270. Device ID is 791f.

 

1st test (Leopard instance (Zephyroth AMD))

 

Step 2 (Dropped ID into ATI 2000 kext, loaded kext and rebooted):

 

Some strange behavior. When I move windows, they become transparent. I still can't change my resolution nor do I have QE/CI.

 

Step 3 (Installed IOPCIFamily.kext, loaded and rebooted):

 

Same behavior as listed above.

 

Step 4 (Installed ATISlice.kext, loaded and rebooted):

 

Same as before, except I have more in the system profiler. (See screenshot)

 

 

2nd Test (Tiger 10.4.9 Uphuck Universal)

 

Step 2:

 

There wasn't an ATIRadeonX2000.kext, so I copied the one from Leopard. Loaded it. Rebooted, nothing seemed to have changed at all.

 

Step 3:

 

Same. Nothing changed

 

Step 4:

 

Only change is the system profiler view. A little more info, same as the screenshot below.

 

 

Let me know if I can do anything else.

 

My box:

 

http://support.gateway.com/s/Mobile/2007/T...14770Rsp2.shtml

post-157865-1200974202_thumb.png

dvavd
dvavd
Posted
Posted
dvavd, try following sequence of steps:

 

step1: download IOPCIFamily.kext redeveloped by Chun-Nan from here:

http://forum.insanelymac.com/index.php?showtopic=81048

install his kexts by replacing originals (don't forget to backup them first)

restart and make sure you have no kernel panic

and check for side effects, describe them here in public thread

 

only if step1 passed OK proceed with step2

 

step2: download ATI1Slice.kext developed by Slice from here:

http://forum.insanelymac.com/index.php?sho...st&p=582921

apply extracted PCIID to ATI1Slice.kext

keep Chun-Nan's IOPCIFamily.kext

restart and check for side effects, describe them here in public thread

there are some changes in I/O Registry after changing ATIPCIFamily.kext to Chun-Nan's one. i don't understand any of them though. just noticed that version of the original kext in my installation was 1.8, not 2.2 as in Chun-Nan's. anyway i listed all the changes below:

1)in the IOPCI2PCIBridge nub a new property shows up – "IOPCIConfigured" = Yes

2)display driver has property "IOPCIResourced" = Yes

3)"AAPL,iokit-ndrv" property of display driver changed from "`Mo7" to 0x605d8838

4)property "reg" of the same driver expanded from 00280100 to 00280100000000000000000000000000000000001028014200...

 

there are also changes in I/O Registry when ATI1Slice.kext is loaded in addition to Chun-Nan's ATIPCIFamily. again, here's the list:

1)new properties of display driver:

"AAPL,gart-width" = <" ">

"fboffset" = <00000200>

"VRAM,memsize" = <00000008>

"AAPL,gray-page" = <01000000>

"GARTCacheSize" = <01000000>

"device_type" = <"display">

"model" = <"ATI Radeon XPRESS 200M (RS400) (PCIE)">

"F8,size" = <00003000>

(obviously all this additions are provided by ATI1Slice)

2)display driver's "AAPL,iokit-ndrv" changed to 0x607dd22f

3)ATI1Slice object appears

4)"IOFBMemorySize" of IONDRVFramebuffer becomes 0x8000000 (was 0x10000000)

 

maybe i missed some details so i attach here excerpts from ioreg outputs.

__1040___1088___1093___1080___1074_.zip

ole2
ole2
Posted
Posted
Here we go. My video card is onboard ATI X1270. Device ID is 791f.

 

1st test (Leopard instance (Zephyroth AMD))

 

My box:

 

http://support.gateway.com/s/Mobile/2007/T...14770Rsp2.shtml

 

hello BigBoy

 

thanks for preliminary report (we might need to get deeper)

however, please stay on your thread, dedicated to RS600 chipsets

http://forum.insanelymac.com/index.php?sho...483&hl=1250

 

this thread is about RS400 chipsets, which are very different

BigBoi
BigBoi
Posted
Posted
hello BigBoy

 

thanks for preliminary report (we might need to get deeper)

however, please stay on your thread, dedicated to RS600 chipsets

http://forum.insanelymac.com/index.php?sho...483&hl=1250

 

this thread is about RS400 chipsets, which are very different

 

Sorry about that Oleg. I've posted into that thread. Let me know if you need any more info from me.

 

Regards,

 Share
Go to topic listing
×
×
  • Create New...