GMA X4500

[Trauma!]

I tried it, but doesn't seem to work. What I get in my log that disturbs me is:

Jul  3 17:44:34 localhost kernel[0]: Natit: Starting.
Jul  3 17:44:34 localhost kernel[0]: Chosen model Unknown
Jul  3 17:44:34 localhost kernel[0]: Natit: Required dictionaries not found

 

Did you maybe forget something? Or do I have to add something to the plist?

 

Give this new Natit a try

 

Natit.kext.zip

 

Analysis of BSD source completed.

 

I will post the details soon. My quick tease: there are signs of hope, but maybe we need a few other things to be done. We will likely have to get help from paulicat, or some of the XLabs guys who are working on alternate display drivers, perhaps Slice or netkas.

 

I need a break, back later.

 

-u

 

Great ! Do you think we should give up X3100 kext hacking track ? Or maybe we need an injector kext to recover informations from VGA BIOS ?

 

Cya.

[sqn816]

anyone getting mouse tearing/artifacts?

[WinSucks_4ever]

We can't answer you that yet, We don't have QE/CI working yet. This is more of an experimental thread. If you need information, read the thread from the beginning, or go to the genius bar. If we get an update about the workings of the project, and if that update is more or less reproducable for everybody, we'll update that thread. We can't do wonders at the moment

 

(Bottomline: Be patient )

 

 

ok thanks for the quick reply!

[enzobelmont]

Trauma: your kexts works the same as Tienneke's ones...

 

thanks for the effort, and keep the good work!

[Tienneke]

Give this new Natit a try

 

Natit.kext.zip

Great ! Do you think we should give up X3100 kext hacking track ? Or maybe we need an injector kext to recover informations from VGA BIOS ?

 

Cya.

 

 

I'm still trying to figure out how to get it to work. The injector seems to work properly, in my logs I get this:

 

Jul  3 20:44:07 localhost kernel[0]: Natit: Starting.
Jul  3 20:44:07 localhost kernel[0]: Chosen model Unknown
Jul  3 20:44:07 localhost kernel[0]: Natit: Setting AAPL,HasPanel=<data not shown>
Jul  3 20:44:07 localhost kernel[0]: Natit: Setting AAPL,NumDisplays=<data not shown>
Jul  3 20:44:08 localhost kernel[0]: Natit: Setting AAPL,NumFramebuffers=<data not shown>
Jul  3 20:44:08 localhost kernel[0]: Natit: Setting AAPL,SelfRefreshSupported=<data not shown>
Jul  3 20:44:09 localhost kernel[0]: Natit: Setting AAPL01,DataJustify=<data not shown>
Jul  3 20:44:09 localhost kernel[0]: Natit: Setting AAPL01,Depth=<data not shown>
Jul  3 20:44:09 localhost kernel[0]: Natit: Setting AAPL01,Pipe=<data not shown>
Jul  3 20:44:09 localhost kernel[0]: Natit: Setting built-in=<data not shown>

 

But I still just get to a normal blue screen, with my cursor on it, but no desktop loading (both internal screen and external through the VGA connector (that's the only connector I have on my laptop, but the chipset should also allow HDMI and DVI if I understood well..)).

I'm now trying to get a ioreg, but I don't seem to be able to connect to my laptop at the moment.

 

It seems that the AppleIntelGMAX3100.kext loads, but can't give more info :s

log.zip

[Trauma!]

Tienneke,

 

It seems natit won't fix anything.... We can give a last try with a new DSDT i've made.

I have added the maximum injections i know on X3100.....

 

 

 

Let us now if it's change something.

[Tienneke]

I'm afraid we'll get stuck at framebuffer only for the moment.. I'll give it a try again, but it seems there is something we don't know about yet.. And I know to little about programming to help writing (Well, I could try, but it'll take years, since I'm a complete noob...)

 

Report: Same result, except the pixels on my screen are like 4 times as big as the should be.. The resolution seems to be correct though (or at least the aspect ratio is correct..)

 

What exactly did you change in this DSDT?

[enzobelmont]

How do I do to get Quartz Extreme

please

 

please... can you see the effort of everybody here?

 

you are not helping us that way.

[Trauma!]

I'm afraid we'll get stuck at framebuffer only for the moment.. I'll give it a try again, but it seems there is something we don't know about yet.. And I know to little about programming to help writing (Well, I could try, but it'll take years, since I'm a complete noob...)

 

Report: Same result, except the pixels on my screen are like 4 times as big as the should be.. The resolution seems to be correct though (or at least the aspect ratio is correct..)

 

What exactly did you change in this DSDT?

 

I agree with you, and i hope coldequation can make something happen ! There so much laptop on the market with this chipset inside, it would great to make this stuff works with OS X. I'm not a dev too, as you've probably noticed yet, it's just a kind of hobby for me.

 

About DSDT decompile it an you will see the changes, i have added EFI info from X3100 ioreg dumps, to much i guess

(like pixel format etc)

 

Cya.

[Tienneke]

I agree with you, and i hope coldequation can make something happen ! There so much laptop on the market with this chipset inside, it would great to make this stuff works with OS X. I'm not a dev too, as you've probably noticed yet, it's just a kind of hobby for me.

 

About DSDT decompile it an you will see the changes, i have added EFI info from X3100 ioreg dumps, to much i guess

(like pixel format etc)

 

Cya.

 

I did notice that but hey, every help is welcome.. We did already do a huge advancement by having the framebuffer working properly, it's already kinda usable..

 

I'll try to be active as long as possible, but it seems we'll need some help from the big mac guru's, those who already have written drivers before..

 

Anybody able to get those on the line?

[enzobelmont]

Trauma! can you tell me how to use dsdt.aml file (sorry)

[coldequation]

Edit: Summary so far & ANALYSIS OF BSD SOURCE:

 

Edit: Action items are highlighted with ===> ACTION:

 

Hi all:

 

Summary:

=======

Here's a quick summary to catch up all who don't want/need to read all the recent posts.

 

My goal was to see if the X3100 kexts could drive the 4500 (4-series) chips. I believe this to be important because with Apple moving to Nvidia, the majority of budget Windows laptops now use Intel 4-series chips. The net result is that the OSx86 world is moving steadily towards custom desktops. Philosophically, I have little problem with that as Apple makes excellent laptops (I have one), whereas their desktop offerings are not competitive for power users. But Apple's laptops remain pricey for many users, especially those outside the USA.

 

In my quick reading of the Intel chipset manuals, it seemed that the 4-series was basically compatible with earlier Intel chipsets and graphics cores. I suggested an experiment that Tienneke, Trauma, mindphazer, enzobelmont and others have tried. Basically we hexedited device ids into the AppleIntelGMAX3100 and AppleIntelGMAX3100FB kexts. We had a short term positive result in that the 4-series can now be used at native resolutions without resorting to grub-tricks, or special Chameleon VESA patches.

 

EDIT: However, the goal of a fully QE/CI supported 4-series has remained elusive. The basic problem is that the AppleIntelGMAX3100 is not getting loaded. The AppleIntelGMAX3100FB does load. Trauma has tried some further experiments with DSDT injection, but so far, that hasn't worked either.

 

Today, I examined the BSD changes (thanks Martin255) that added 4-series support to find evidence for or against my hunch. Since the Intel Linux drivers and BSD sources are similar, it's possible that even Apple's OS X (another Unix derivative) might follow suit. I've managed to examine all of the BSD changes that Martin pointed me to, as well as some others in the tree.

 

Note: I *did not* examine any fixes after the patch which enabled 4-series support. I do think that will have to be done, but since the fixes came from Intel, maybe not.

 

CAVEAT: I am not a graphics programmer. So could someone check my logic/conclusions below?

 

RESULTS:

======

 

I *do* believe that we can make the X3100 work for the 4500 (4-series) chips, *if* we can control some addresses that are controlled by the kext. But to determine for sure, we will need help from anyone who has experience with video drivers (perhaps the XLabs guys, paulicat, netkas, Slice or others I don't know). I've tried to indicate what I consider to be *significant* problems, and *non-significant* issues.

 

===> ACTION: Can someone forward/send/PM this post to those guys?

 

DETAILS:

=======

 

Feel free to skip this part if you have no interest. I've tried to document my assumptions and conclusions, so others can verify or refute if they get a chance.

 

1. The defines:

+#define PCI_CHIP_IGD_E_G_BRIDGE 0x2E00
  +#define PCI_CHIP_G45_G_BRIDGE	0x2E20
  +#define PCI_CHIP_Q45_G_BRIDGE	0x2E10

 

It looks like subtracting 2 from the device id gets you the device id of the chipset "bridge". On both the MacBook X3100 and the HackBook X3100, there is a device in ioreg called MCHC ("acpi-path" = "IOACPIPlane:/_SB/PCI0/MCHC@0") that is registered against it. The BSD source defines new constants for the 4-series bridges, but I don't know how significant that is, because on my GMA950, I just have a device called "pci8086,27a0@0" and my display functions (albeit with artifacts).

 

CONCLUSION: *not significant* (although it might be nice to add the MCHC to one's DSDT, so the ioreg output looks "pretty.")

 

2. This lines:

#define IS_I965G(pI810) (DEVICE_ID(pI810->PciInfo) == PCI_CHIP_I965_G || 
   DEVICE_ID(pI810->PciInfo) == PCI_CHIP_G35_G || DEVICE_ID(pI810->PciInfo) == PCI_CHIP_I965_Q || 
   DEVICE_ID(pI810->PciInfo) == PCI_CHIP_I946_GZ || DEVICE_ID(pI810->PciInfo) == PCI_CHIP_I965_GM || 
   DEVICE_ID(pI810->PciInfo) == PCI_CHIP_I965_GME || IS_IGD_GM(pI810) || IS_G4X(pI810))

CONCLUSION: *not significant* Code that uses IS_I965G (Desktop chipsets) considers the G4X series to be just a variant of the 965.

 

This is great! It means that for most purposes the 4-series (G4X) is compatible.

 

3. Seems that the differences between the various 4500 versions (with the notable exception of the laptop/mobile GMA 4500MHD) are pretty much ignored:

 

PCI_CHIP_IGD_GM/IS_IGD_GM (0x2a42) is the Cantiga with GMA 4500MHD graphics core

PCI_CHIP_IGD_E_G (0x2E02) is probably the G41/G43 chipset with GMA X4500 graphics core

PCI_CHIP_G45_G (0x2E22) is the G45 chipset with GMA X4500HD graphics core

PCI_CHIP_Q45_G (0x2E12) is the Q45 chipset with GMA 4500 graphics core

 

The line:

#define IS_G4X(pI810) (DEVICE_ID(pI810->PciInfo) == PCI_CHIP_IGD_E_G || 
   DEVICE_ID(pI810->PciInfo) == PCI_CHIP_G45_G || 
   DEVICE_ID(pI810->PciInfo) == PCI_CHIP_Q45_G)

CONCLUSION: *not significant* The Desktop GMA 4500, GMA X4500HD, and GMA X4500 or GMA X4500HD are treated effectively the same in the code. This means there should not be much variation between the desktop versions of the 4-series graphics core.

 

4. OK, to figure out the Laptop/Mobile GMA4500MHD we need to look at a previous commit to the one Martin sent. Looking at this: http://cgit.freedesktop.org/~anholt/xf86-v...3fbce21492a7c32

 

CONCLUSION: *not significant* the changes in adding IS_IGD_GM (GMA4500MHD) were very minor. This implies that a laptop with GMA4500MHD *could* work with very minor editing (device id hexediting of the laptop X3100 kexts), assuming Apple hasn't done anything weird in the X3100 kexts. This may also explain why people struggle getting desktop 3100s or 3500s to work. But see below:

 

4. This line:

 

+/* mark chipsets without overlay hw */
  +#define OVERLAY_NOEXIST(pI810) (IS_IGD_GM(pI810) || IS_G4X(pI810))

means that the GMA 4500 (Q45), X4500HD (G45), and X4500 (G41/G43) all do not have overlay hardware. More on that later:

 

5. This change:

 

/* chipsets require graphics mem for hardware status page */
  -#define HWS_NEED_GFX(pI810) (IS_G33CLASS(pI810) || IS_IGD_GM(pI810))
  +#define HWS_NEED_GFX(pI810) (IS_G33CLASS(pI810) || IS_IGD_GM(pI810) || IS_G4X(pI810))
  +/* chipsets require status page in non stolen memory */
  +#define HWS_NEED_NONSTOLEN(pI810) (IS_IGD_GM(pI810) || IS_G4X(pI810))

Means that hardware status (HWS) requires graphics memory. But all the Desktop GMA 3100 (G33/Q35/Q33) do the same. But we don't know what the mobile GMA X3100 (PCI_CHIP_I965_GM and PCI_CHIP_I965_GME) do.

 

CONCLUSION 1: *not significant* Desktop 3100 and 4-series act the same WRT to HWS (hardware status).

 

	/* new chipsets need non-stolen status page */
   if (HWS_NEED_GFX(pI830) && HWS_NEED_NONSTOLEN(pI830))
	   mmsize -= HWSTATUS_PAGE_SIZE;

CONCLUSION 2: Desktop 3100 and 4-series act the same WRT to mmsize. (Bad coding here: HWS_NEED_NONSTOLEN is a subset of HWS_NEED_GFX)

 

CONCLUSION 3: *significant* This could be an ISSUE.

 

===> ACTION: Can anyone confirm if Apple implemented only the X3100 mobile chips? If they implemented Desktop 3100 (on say the iMac), then we may be okay.

 

If Apple implemented only 3100 mobile, the implication is that laptop X3100 (PCI_CHIP_I965_GM || PCI_CHIP_I965_GME) *does not need a non-stolen* status page. To fix this, we may need to decrease the amount of memory for the driver by 4KB.

 

6. This change:

 

+/* chipsets require status page in non stolen memory */
  +#define HWS_NEED_NONSTOLEN(pI810) (IS_IGD_GM(pI810) || IS_G4X(pI810))

could be significant, since it is ONLY defined for the 4-series (mobile or desktop) but let's see:

 

+ if (HWS_NEED_NONSTOLEN(pI830))
  flags |= NEED_NON_STOLEN;

CONCLUSION: *significant* The BSD driver allocates a separate, fixed hardware status page of 4KB for the 4-series. But perhaps it will work with stolen memory?

 

7. OVERLAY_NOEXIST:

 

CONCLUSION: *significant* The 4-series (desktop & mobile) *do not* allocate any space in the kernel for overlay registers. It's possible that the driver will just fall back to not supporting overlay if we use the X3100 kexts to drive 4-series hardware.

 

8. These changes:

 

+#define INTEL_GMCH_GMS_STOLEN_96M (0xa << 4)
  +#define INTEL_GMCH_GMS_STOLEN_160M (0xb << 4)
  +#define INTEL_GMCH_GMS_STOLEN_224M (0xc << 4)
  +#define INTEL_GMCH_GMS_STOLEN_352M (0xd << 4)

are used to setup the The Graphics Translation Table (GTT). For the G33, 965 or older chipsets, the GTT is stolen from your memory, and may or not be larger than the aperture. But:

 

   /* The stolen memory has the GTT at the top, and the 4KB popup below that.
   * Everything else can be freely used by the graphics driver.
   */
  range = gtt_size + 4; // - I think this will be 8 KB ???

  /* new 4 series hardware has separate GTT stolen with GFX stolen */
  if (IS_G4X(pI830))
	  range = 0;

CONCLUSION: *significant* Your stolen memory's size is decreased by your GTT + 4 KB popup, whatever that is. But the 4-series has it's own stolen GTT, so its effective stolen memory size is not decreased by 8KB. Not sure how to fix this.

 

9. This change:

 

	if (IS_I965G(pI830)) 
   {
	   if (IS_IGD_GM(pI830) || IS_G4X(pI830)) {
			gttaddr = pI830->MMIOAddr + MB(2);
			pI830->GTTMapSize = MB(2);
		} else {
			gttaddr = pI830->MMIOAddr + KB(512);
			pI830->GTTMapSize = KB(512);
		}
   }

CONCLUSION: *significant* It affects the address (location) and size of the GTT. The 4-series will be different from all previous Intel chips.

 

===> ACTION: Can we do anything with this?

 

10. This change:

 

	/* Begin the long sequence of commands needed to set up the 3D
	* rendering pipe
	*/
  ...
   if (IS_IGD_GM(pI830) || IS_G4X(pI830))
	   OUT_BATCH(NEW_PIPELINE_SELECT | PIPELINE_SELECT_3D);
   else
	   OUT_BATCH(BRW_PIPELINE_SELECT | PIPELINE_SELECT_3D);

 

Affects the 3D pipeline.

 

+#define BRW_3D(Pipeline,Opcode,Subopcode) (CMD_3D | \
  + ((Pipeline) << 27) | \
  + ((Opcode) << 24) | \
  + ((Subopcode) << 16))

  +#define BRW_PIPELINE_SELECT BRW_3D(0, 1, 4)
  +#define NEW_PIPELINE_SELECT BRW_3D(1, 1, 4)

 

CONCLUSION: *significant* For 3D graphics, the 4-series use a new pipeline (pipeline 1) while the older cards (including the X3100) use pipeline 0.

 

Final thoughts:

==========

 

Given that the ATI drivers seem to use different pipelines, memory sizes, registers, etc, and IOKit being very object-oriented, we may be able to trick the 3100 drivers by pretending to be a different 3100, albeit with different GTT location, overlays, pipelines etc. Or maybe not. But,

 

Edit: We need to know what's in the X3100 MacBook (and any Apple desktop 3100). So...

 

===> ACTION: Anyone who knows the EFI Utils, getting strings from EFI, etc. We need to know what (if any) specific keys and values exist for the X3100.

 

Edit: The ball's in the community's court now.

 

==coldequation==

 

-u

[nomi81]

Hello there,

 

I've tried the kexts with latitude e6400 but i've faced all the time same problem of blue screen before my desktop comes to alive.

 

i've tested it with 10.5.6 and 10.5.7 but with no luck.

 

We really need of some help here from all OSX 86 gurus....

 

Please listen to the public.

[deathanarchist2]

I have a GMA x4500 mhd is there anything i can do to help and does that post mean im more likely to be supported in the near future?

edit it works for dection of card but i have major artifacts

[Tienneke]

I have a GMA x4500 mhd is there anything i can do to help and does that post mean im more likely to be supported in the near future?

edit it works for dection of card but i have major artifacts

 

Did you use the custom DSDT.aml that trauma made for me? Try the previous version maybe, cause with this version I have the same problem, with the previous version it works better for me (then again, that DSDT was made for my hardware, so it would be strange if it wouldn't give some kind of problem on your pc..)

 

Do you get to your desktop in normal boot, and do you have qe/ci?

[WinSucks_4ever]

Hi guys I have a question!

Is there anyway to force the translucid menubar in Leopard without QE/CI

Please a tweak or something

[Trauma!]

Coldequation,

 

Thanks for this analisys ! What i can tell is Apple has only used X3100 in Macbook3,1. See attached documents about MacBook and is X3100 :

 

- DSDT injections to match original MacBook3,1 IOreg.

- MacBook3,1 IOreg save.

- MacBook3,1 DSDT and SSDT Dumps.

- The script to quickly patch X3100 driver to X4500MHD ID's.

- A set of patched drivers, for testing purpose.

- A Natit to works with 950 / X3100 and hopefully X4500.

 

My thought, if the framebuffer runs fine without injection we may not need to use injection at all.

 

 

[coldequation]

Critical question is whether the framebuffer is loading. <= Title

 

Bonjour Trauma, Comment ça va?

 

Thanks for the thanks and thanks for the files you posted. I will take a look.

 

I believe the critical question is to determine whether the framebuffer is loading.

 

Tienneve, enzobelmont: is there any way you can ssh into your box and tell? Also, did either of you find out whether AppleIntegratedFrameBuffer is interfering with the AppleGMAX3100FB loading? Earlier, I suggested deleting/moving it to see if it makes a difference.

 

My analysis leads me to believe that we should first concentrate on the GMA4500MHD, as it is the most compatible with the X3100.

 

Also, if would be good if someone could find an EFI strings dump from a real X3100 Mac, and turn it into a plist for us to study. We could then see if there is a key to alter the pipeline, overlay, and memory regions. We may end up finally hexediting those into the kext if there is no way from the AAPL/EFI keys.

 

Back later,

-u

Coldequation,

 

Thanks for this analisys ! What i can tell is Apple has only used X3100 in Macbook3,1. See attached documents about MacBook and is X3100 :

 

- DSDT injections to match original MacBook3,1 IOreg.

- MacBook3,1 IOreg save.

- MacBook3,1 DSDT and SSDT Dumps.

- The script to quickly patch X3100 driver to X4500MHD ID's.

- A set of patched drivers, for testing purpose.

- A Natit to works with 950 / X3100 and hopefully X4500.

 

My thought, if the framebuffer runs fine without injection we may not need to use injection at all.

 

[WinSucks_4ever]

Coldequation,

 

Thanks for this analisys ! What i can tell is Apple has only used X3100 in Macbook3,1. See attached documents about MacBook and is X3100 :

 

- DSDT injections to match original MacBook3,1 IOreg.

- MacBook3,1 IOreg save.

- MacBook3,1 DSDT and SSDT Dumps.

- The script to quickly patch X3100 driver to X4500MHD ID's.

- A set of patched drivers, for testing purpose.

- A Natit to works with 950 / X3100 and hopefully X4500.

 

My thought, if the framebuffer runs fine without injection we may not need to use injection at all.

 

Attached File Archive.zip ( 840.82K ) Number of downloads: 9

 

Tell me if that works!

I m affraid to test because If my system breaks I need to install Mac OS X againt!

 

how do I do to delete a kext using -s command

and /sbin........

 

I want to delete natit.kext

 

please guys I need the anwsore today!!!!!!!!

[coldequation]

Your enthusiasm is noted but, <== TITLE

 

Rubenmarques,

 

Your enthusiasm is noted, but until we get some detailed and repeatable instructions, I would suggest waiting.

 

And that advice is to all who consider themselves noobs. I'm not being pejorative as we were all noobs once. But right now, we need to have some volunteers who can do some basic stuff: research, testing, idea generation, provide moral support , and communicate with the Elder Gods of the OSX86 universe. So if any noob-oriented questions are ignored, trust that we are focusing down on the problem.

 

Check back every few days. I'm sure if something works, you'll see a title of "WORKING!!!" or "HEAVENS HAVE OPENED!!!!!!"

 

P.S. don't take this in a bad way. We all have only so much bandwidth.

 

Thanks,

-u

 

how do I do to delete a kext using -s command

and /sbin........

I want to delete natit.kext

 

Type -s when you boot the machine. That puts you into single-user mode and does not run the desktop.

 

At that point, it gives you two commands to type to check and mount the filesystem. I don't remember them from memory, but the first is fsck, and the second is a mount command, mounting the / directory.

 

IIRC, you should be root at that point, so you can:

cd /System/Library/Extensions

rm -rf natit.kext

 

be careful with the rm command! One way to check that you don't delete anything else is when you start to type nat, hit the TAB key and the shell will complete the filename for you. Also, remember that filenames are case-sensitive, so make sure which it is, natit.kext or Natit.kext.

 

You can use the ls command to check, before removing with rm; ls lists the files in the directory.

 

Once you have removed natit, then reboot with the -f option to get rid of your kext cache (it reloads all of them).

[WinSucks_4ever]

Your enthusiasm is noted but, <== TITLE

 

Rubenmarques,

 

Your enthusiasm is noted, but until we get some detailed and repeatable instructions, I would suggest waiting.

 

And that advice is to all who consider themselves noobs. I'm not being pejorative as we were all noobs once. But right now, we need to have some volunteers who can do some basic stuff: research, testing, idea generation, provide moral support , and communicate with the Elder Gods of the OSX86 universe. So if any noob-oriented questions are ignored, trust that we are focusing down on the problem.

 

Check back every few days. I'm sure if something works, you'll see a title of "WORKING!!!" or "HEAVENS HAVE OPENED!!!!!!"

 

P.S. don't take this in a bad way. We all have only so much bandwidth.

 

Thanks,

-u

 

 

 

Type -s when you boot the machine. That puts you into single-user mode and does not run the desktop.

 

At that point, it gives you two commands to type to check and mount the filesystem. I don't remember them from memory, but the first is fsck, and the second is a mount command, mounting the / directory.

 

IIRC, you should be root at that point, so you can:

cd /System/Library/Extensions

rm -rf natit.kext

 

be careful with the rm command! One way to check that you don't delete anything else is when you start to type nat, hit the TAB key and the shell will complete the filename for you. Also, remember that filenames are case-sensitive, so make sure which it is, natit.kext or Natit.kext.

 

You can use the ls command to check, before removing with rm; ls lists the files in the directory.

 

Once you have removed natit, then reboot with the -f option to get rid of your kext cache (it reloads all of them).

[Trauma!]

Here you are coldequation

 

ça va bien merci !

 

Exctrated from the ioreg i've posted earlier.

 

 

 

 

Cheers.

 

And please people stop spamming the thread with inappropriate questions ! We are looking for a solution to make X4500MHD works fully.... be sure you'll know it when it will work fine.

 

@ rubenmarques : If you don't know how to use this forum... please clean your mistakes at least

 

@ Tienneke : Try to add that inside natit plist in the "Unknow" dictionary (<dict>)

 

<key>revision-id</key>
<data>
AwAAAA==
</data>

[ADN]

Mount your file system in read/write then :

 

sudo rm -R /System/Library/Extensions/2A42AppleIntelGMAX3100.kext

sudo reboot

 

Use -f for first boot.

 

Should be enough.

 

If not :

 

sudo rm -R /System/Library/Extensions/2A42AppleIntelGMAX3100FB.kext

sudo rm -R /System/Library/Extensions/Natit.kext

sudo reboot

 

Did you see natit outputs during the boot ?

 

Hi,

 

I try all case but its not work. Maybe I will reinstall and test your driver .

 

Thanks

[enzobelmont]

Tienneve, enzobelmont: is there any way you can ssh into your box and tell? Also, did either of you find out whether AppleIntegratedFrameBuffer is interfering with the AppleGMAX3100FBloading? Earlier, I suggested deleting/moving it to see if it makes a difference.

 

-u

 

i've deleted 2A42AppleGMAX3100FB and my system do not load the GUI...

[Trauma!]

PLEASE DO NOT INSTALL THIS EXCEPT IF YOU DO KNOW WHAT YOU DO !!!! PLEASE

 

THERE IS 99% THIS WILL MESS UP YOUR DISPLAY.

 

All others a welcome to test it freely of course

 

Maybe i can create an IRC channel so we can discuss live about GMA X4500 debugging ?