Intel HD Graphics / GMA 5700

[Christoff]

Please Help! i have a Dell 14R N4010 notebook have tried --Piracy Removed-- S3 , partially working (cant get graphics) .Anyone could get Intel HD graphics with Quartz extreme and full resolution? need it for Logic 9 mainly

Specs:

 

QUOTE

Dell 14R touchpad

Dell 14R features a Samsung screen.

camera

two 14R wireless antennas

Core i5 430M processor

N4010 HD 5470 discrete graphics

14R PCH chip

Dell Inspiron 14R Dual DDR3 memory slots (4 GB 1066)

card reader and Bluetooth module

hard disk and CD-ROM SATA Interface

N4010 ATHEROS AR8152 LAN chip

REALTEK ALC269 CODEC chip

N4010 Mini PCIE slot for the installation of wireless LAN equipment

 

Dell Page With drivers http://search.dell.com/results.aspx?spf ... srch=False

 

Please Help!

Thanks!

Christoff

PD:im aNewbie

PD2: Tried Iboothd 3 iso = KP

[Time2Retire]

Hello,

 

May be i put bad kext on upload, so this one works.

 

LegacyIOGraphicsFamily.kext.zip

 

Search AppleDisplay -> IOTest = fred test

iMac_de_fred.zip

 

Fred

I had to change IOProbeScore to 2500 and voilà. That did the trick.

 

I guess that it needs some time to initialize first – the (HP) notebook I am using might be a bit slower than yours. Anyway this works so now we can modify our EDID, fix the checksum, and inject it in a clean way. Without tampering with the original kexts.

 

Thanks again.

 

@yurick and Christoff: Nobody here has a fix. Just wait. Hang in like the rest of us. We're trying as hard as we can, but we're not there yet.

[yurick]

 

........

 

@yurick and Christoff: Nobody here has a fix. Just wait. Hang in like the rest of us. We're trying as hard as we can, but we're not there yet.

 

Well, only on this and hope that the 10.6.6 .. everything will be tip-top. Thank you 

[FredWst]

Hello,

 

I try that, i've boot on linux and get EDID and inject.

On external VGA get one screen split in 2, frequency seems bad or memory framebuffer allocation.

Because with vnc 1024x768 is ok, 1280x1024 strange sreen !

 

So i have to install a fresh OSX, and make other try.

 

Intel bug

http://www.intel.com/support/graphics/sb/cs-028366.htm

 

 

Fred

[Time2Retire]

@Fred,

 

I don't think that we should worry about a Windows only driver bug. Doesn't matter. Won't change a thing for us.

[Christoff]

hi i found this and thought it could help!

 

Quote:

http://www.faq-mac.com/1382/quartz-extreme-casi-todos

Quartz Extreme for (almost) all

 

 

Drafting sent by August 30, 2002 - 1:26.

Want to run Windows or Linux on your Mac?: VMWare Fusion Test.

Extreme Quarktz (only available in Jaguar) only works on some cards very specific and under very special conditions.

but on the forums of Ars Technica seems to have found the solution to the problem.

 

The trick works as follows (Radeon 7000 PCI on beige G3, or Macs with PCI slots only for Video, NOT work on Macs with AGP):

 

Open the configuration file:

 

/ System / Library / Frameworks / ApplicationServices.framework /

 

Versions / A / Frameworks / CoreGraphics.framework / Versions / A /

 

Resources / Configuration.plist

 

find the following lines:

 

 

Change IOAGPDevice by IOPCIDevice

 

In general, this configuration file seems to contain the specifications for the activation of Quartz Extreme. Apple seems to have determined that PCI cards do not work fast enough to run Quartz Extreme on them, but for non-intensive tasks increased speed seems remarkable.

 

The utility Quartz Extreme Check can help you check if the patch has been effective.

[FredWst]

@Fred,

 

I don't think that we should worry about a Windows only driver bug. Doesn't matter. Won't change a thing for us.

 

Ok

 

I saw on another thread that you would like to compile IOGraphics.

 

I've compile IOGraphics rev 373.2 with instruction found here

 

http://www.insanelymac.com/forum/lofiversi...190586-650.html

 

Fred

[Time2Retire]

Ok

 

I saw on another thread that you would like to compile IOGraphics.

 

I've compile IOGraphics rev 373.2 with instruction found here

 

http://www.insanelymac.com/forum/lofiversi...190586-650.html

 

Fred

Oops. I forgot to check. Good thing you did. Thanks!

 

This should help me to understand what is going on. I mean looking at the width of my scrambled screen, it almost looks like a frame buffer initialization problem. It might not fix the problem, but I want to give it a go. Just to be sure that I'm not missing anything,

 

Oh and here's something I found in IOGraphicsLib.c:

[#define kIOFirstBootFlagPath	"/var/db/.com.apple.iokit.graphics"

This is where the "startup-timing" (Data) property is stored. I replaced it with the one from my MacBook Pro and it's there alright, and shows up in IORegistryExplorer, but this does not solve it.

 

Too bad that most code uses: IOFBConnectRef connectRef because I fail to convert a service / connection to this connectRef. And that's why I am stuck with most of my command line tools.

 

Update: I received the OS X 10.6.5 (10H568) upgrade ahead of time. Sweet. Installing right now. Let's see what this brings

 

No change. And it was 10.6.5. Not 10.6.6 Anyway. I was advised to remove all cruft from my DSDT, and that is what I am doing right now. What a fun. 65% done the drain LOL

[riws]

Doesn't the AppleMCCSControl.kext (Monitor Control Command Set?) help us here?

[Time2Retire]

I got an e-mail from someone who claims to have it working by using the old X3100 method (with AppleIntelHDGraphicsFB

.kext removed) but I don't know what this is referring to. And yes he's using a real Intel GMA5700. Anyone tried this already?

[Fiox89]

I got an e-mail from someone who claims to have it working by using the old X3100 method (with AppleIntelHDGraphicsFB

.kext removed) but I don't know what this is referring to. And yes he's using a real Intel GMA5700. Anyone tried this already?

 

No, but it is so difficult to try? and can work with Intel HD?

 

It could be a quick solution

[fmac]

Hi

 

for all here with a Dell E6510 and any other

 

"We are not alone with our problem"

some people on the linux side have the same issue with "core i5 ironelake"

"black screen, only remote display"

Look here : https://bugs.freedesktop.org/show_bug.cgi?id=28070

 

but there are some patches

 

http://lists.freedesktop.org/archives/inte...une/007216.html

http://lists.freedesktop.org/archives/inte...une/007232.html

 

Please take a look at this, maybe there is a solution for us

i dont know, if it is possible to apply the patch "where or how" (chameleon or kext or "I dont know)

 

give it a try

 

Good night (have a long day sine 2.30 am)

[Fiox89]

This can be useful?

 

http://www.insanelymac.com/forum/index.php...215595&st=0

[FredWst]

Hi

 

for all here with a Dell E6510 and any other

 

"We are not alone with our problem"

some people on the linux side have the same issue with "core i5 ironelake"

"black screen, only remote display"

Look here : https://bugs.freedesktop.org/show_bug.cgi?id=28070

 

but there are some patches

 

http://lists.freedesktop.org/archives/inte...une/007216.html

http://lists.freedesktop.org/archives/inte...une/007232.html

 

Please take a look at this, maybe there is a solution for us

i dont know, if it is possible to apply the patch "where or how" (chameleon or kext or "I dont know)

 

give it a try

 

Good night (have a long day sine 2.30 am)

 

Hello,

 

I come from linux world.

And the trouble is quiet different than Mac Osx.

 

Black screen is a trouble of initialization during boot time.

Just boot E6510, connect a screen on vga connector remove it and all is ok.

 

So i forgot to talk about this, but all my test with E6510 was made with external display connected, to be sure that it will not disturb the result.

 

Fred

[Fiox89]

With the portable in my signature, (Asus K52Jc), and whit the ONLY video card Intel HD enaled on the bios, the screen works without problem (but the drivers are not yet good)

(Ubuntu 10.04 and ArchLinux)

 

If the bios are enabled Both video cards, the screen goes black (but not always) and returns to normal after the login (I do the login with closed eyes without seeing the screen)

[Time2Retire]

I removed all brightness related 'junk' from my DSDT. Making it much more fun to read / handle. Then I come here... and have to read comments about Linux? What the BEEP LOL

 

Update: Ok. I'm done with patching Revolution (for now). I'm happy with it as it stands; A much better Geekbench score – gained close to 1000 points. Time to go back to our 'dead' screen problem.

 

Status update: I have a connection with the display. Can initiate sleep (idle) and change the brightness level. Time to dump some values from IOGraphicsFamily.kext Let's see what this brings. But first I want to have a closer look at this IODisplayConnectFlags property and its values. My desktop hack has a value of: 84 41 00 00 (NVIDIA PCI-E CARD) and the internal GPU's on a MacBookPro 00 08 00 00 (INTEL) And looking at some other IOREG dumps... I see both 84 49 00 00 (Internal NVIDIA) and C4 49 00 00 (Internal ATI) for internal GPU's. My bet is that ATI PCI-E cards give C4 49 00 00 but feel free to prove me wrong.

 

A value of 00 00 00 00 suggests that IODisplay.cpp didn't locate the GPU. That it doesn't know if it's an internal or some PCI-E card. So who here has 00 00 00 00? I think that we need to address this first. To make it initialize properly. Should at least show 00 08 00 00 or higher.

 

These values don't change when you switch from Intel to NVDIA and back. Either automatically, or with gfxCardStatus.

 

And what are you findings? LOL Don't tell me that you've been relaxing, while I push hard. No matter what; Even had FOUR games yesterday (won three of them) of which I played three above my pool. Next week I will defend the goal of the seniors (ladies) which is a first time ever for a 15 year old girl!!!!!!!

 

I'm sorry if this sounds harsh to you, but I'm so disappointed that we can't solve this problem. That it takes so long before anything really changes (:

[fmac]

I removed all brightness related 'junk' from my DSDT. Making it much more fun to read / handle. Then I come here... and have to read comments about Linux? What the BEEP LOL

 

Update: Ok. I'm done with patching Revolution (for now). I'm happy with it as it stands; A much better Geekbench score – gained close to 1000 points. Time to go back to our 'dead' screen problem.

 

Status update: I have a connection with the display. Can initiate sleep (idle) and change the brightness level. Time to dump some values from IOGraphicsFamily.kext Let's see what this brings. But first I want to have a closer look at this IODisplayConnectFlags property and its values. My desktop hack has a value of: 84 41 00 00 (NVIDIA PCI-E CARD) and the internal GPU's on a MacBookPro 00 08 00 00 (INTEL) And looking at some other IOREG dumps... I see both 84 49 00 00 (Internal NVIDIA) and C4 49 00 00 (Internal ATI) for internal GPU's. My bet is that ATI PCI-E cards give C4 49 00 00 but feel free to prove me wrong.

 

A value of 00 00 00 00 suggests that IODisplay.cpp didn't locate the GPU. That it doesn't know if it's an internal or some PCI-E card. So who here has 00 00 00 00? I think that we need to address this first. To make it initialize properly. Should at least show 00 08 00 00 or higher.

 

These values don't change when you switch from Intel to NVDIA and back. Either automatically, or with gfxCardStatus.

 

And what are you findings? LOL Don't tell me that you've been relaxing, while I push hard. No matter what; Even had FOUR games yesterday (won three of them) of which I played three above my pool. Next week I will defend the goal of the seniors (ladies) which is a first time ever for a 15 year old girl!!!!!!!

 

I'm sorry if this sounds harsh to you, but I'm so disappointed that we can't solve this problem. That it takes so long before anything really changes (:

 

 

Your not alone

 

Please take a look at this

 

https://github.com/codykrieger/gfxCardStatus

 

look inside switcher.m

 

the main thing there is muxstate, i think this is the state of gmux device you can find on macbooks

 

and i have looked for IODisplayConnectflags my is always 00 00 00 00

with or with out extensions there is no difference

[Time2Retire]

This appears to be VERY important: Who here can boot in any non-VESA resolution above 1024 x 768? And yes I mean get Chameleon (or whatever booter you are using) to boot into a graphics mode above the aforementioned VESA mode (280 / 0x0118). I have to get cracking on a VGA BIOS mod because the framebuffer is stuck at 1024 pixels!!!

 

@fmac: Yeah I've seen it. Even linked to it in my post. I however want a working display, before I start thinking about GPU switching.

 

Update: You can verify what I said by settings "Graphics Mode" in com.apple.Boot.plist to a lower resolution, say 800 x 600 x 32, and still have screen sharing set to 1024 x 768 x 32 – which we all know works just fine – until you do this. This shows us that the framebuffer is using the boot arguments, which you can also see when you modify Chameleon (or in my case Revolution). I'm talking about the following lines in boot2/graphics.c:

		// Update KernBootStruct using info provided by the selected VESA mode.
	bootArgs->Video.v_display  = GRAPHICS_MODE;
	bootArgs->Video.v_width	= minfo.XResolution; // 1920 or 1600
	bootArgs->Video.v_height   = minfo.YResolution; // 1200 or 900
	bootArgs->Video.v_depth	= minfo.BitsPerPixel; // 32
	bootArgs->Video.v_rowBytes = minfo.BytesPerScanline; // 7680 or 6400

The result will be that the kernels graphic throbber (rotating spinner) moves to another screen position. And thus this problem appears to be caused by a faulty Intel HD Graphics ROM.

 

Getting the ROM image was easy, because this can be done with tools like GPU-Z and Everest, but how do we flash a ROM? I'm cool with trying to flash the original ROM. Anyone else here willing to spent some time and find out how this works and what tool(s) we need?

[kizwan]

I think you can load the Intel HD Graphics ROM using Chameleon (VBIOS flag). It kinda VBIOS override (I'm not sure). I don't know whether Revolution still have this feature or not. For integrated GPU (including discrete GPU which is soldered on the motherboard), VBIOS is integrated in system BIOS.

[Time2Retire]

I think you can load the Intel HD Graphics ROM using Chameleon (VBIOS flag). It kinda VBIOS override (I'm not sure). I don't know whether Revolution still have this feature or not. For integrated GPU (including discrete GPU which is soldered on the motherboard), VBIOS is integrated in system BIOS.

Thank you kizwan. I'll investigate this VBIOS load option as soon as possible (it's something completely new to me).

 

I'm off to my homework now

 

Ok. I don't think that the VBIOS option is going to work for us, or I must be reading it completely wrong:

 

#define kVBIOS "VBIOS" /* nvidia.c */

 

And looking at nvidia.c yeah. That sure doesn't look good.

 

We apparently need to use the Insyde BIOS editor called EZH2O Well hello!!!

[eraser0]

hiii

could it be modified to work for gma 4500hd?

 

since they r in the same line i think that the modification could be done

 

 

please report please

 

 

any developer here!!!!

[icy]

Chameleon bootloader since 2.0-RC3-r658 version besides VBIOS option has a VideoROM option.

It can be used like that:

VideoROM=<file> Use an alternate ROM image (default path: /NVIDIA.ROM).

Link to the official chameleon documentation

[Time2Retire]

Chameleon bootloader since 2.0-RC3-r658 version besides VBIOS option has a VideoROM option.

It can be used like that:

VideoROM=<file> Use an alternate ROM image (default path: /NVIDIA.ROM).

Link to the official chameleon documentation

I guess they should update the documentation since grep -r -e VideoROM . in the trunk does not return any hits. Not even a single one.

 

I did however find it in an older version: Chameleon-2.0-RC3-r658/i386/libsaio/nvidia.c

That settles it. NVIDIA only.

[fmac]

Thank you kizwan. I'll investigate this VBIOS load option as soon as possible (it's something completely new to me).

 

I'm off to my homework now

 

Ok. I don't think that the VBIOS option is going to work for us, or I must be reading it completely wrong:

 

#define kVBIOS "VBIOS" /* nvidia.c */

 

And looking at nvidia.c yeah. That sure doesn't look good.

 

We apparently need to use the Insyde BIOS editor called EZH2O Well hello!!!

 

Here it is

 

http://rghost.net/561135

Insyde ExH2O v2.1.0.4

[Time2Retire]

Has anyone else here noticed that certain properties, like IOFBCurrentPixelClock and IOFBCurrentPixelCount aren't changing (getting updated) on resolution changes? And are you folks using OS X 10.6.5 already? Still need cpus=1 and busratio (Apple has yet to release the source code for this update).

 

Note: More information about these properties can be found here.

 

@fmac: Got it. Thanks.