nothing interesting here

[erikk]

i tested a little bit my system with your working kernel binary.

 

-dual screen on gma950 works with old appleintegratedframebuffer.kext and the new 10.4.8 kext for 3d acceleration. i think there are many improvements also... the gui is much smoother than before.

-pro applications works well (without using old iokit)

-usbaudio make absolutly no problem (first time ever) - the cool thing my access virus doesnt have sync problems here anymore... on my macbookpro i have it

-rosetta feels good... but i never understand how people can scream for new rosetta... i had it on my book already and yes its still a bit faster...

on my core2duo system i can say it rocks... photoshop cs2 works fast, office works fast... sound applications in rosetta are still horror

-for games i can not say much: chess works but at all i feel a better gui performance on both monitors together... thats really great!

-reboot / shutdown and sleepmode works also without problems

 

so i think... 99,9% its finished... its only the hardware information and the about mac window which is not work.

[pion]

Have you repaired permissions?

 

yes these are basic things, i always repair permissions.

[Vitaliy]

-

Edited November 20, 2006 by Vitaliy

[littlepr]

So close but to no avail...

 

What am I doing wrong? I added the karma key into the commpage.c, compiled a kernel following the tutorial:

 

sudo -s

cd /tmp

chmod -R 777 sources

cd sources

make ARCH_CONFIGS=I386 KERNEL_CONFIGS=RELEASE install

 

-chmod and chown applied to newly compiled mach_kernel

 

-Copied to safe place (/backup folder of 10.4.8 running install). Originally a 10.4.7 updated with Jas's sse2 intel combo update.

-Rebooted and repaired permissions in single user mode just to be sure.

- Rebooted and booted up new kernel by manually typing the path:

/backup/mach_kernel -v

 

A lot of the kexts were loaded but a others failed to load. All of the failed ones were in yellow text.

Becaus of this I have not been able to load with the new kernel. I can still boot with the 10.4.4 Kernel since I did not install the compiled kernel in the root path. I'm to witty for that...

 

So what am I missing guys?

[Tom H]

Also I have moved my focus over to using semthex's kernel now as it seems to be more in the spirit of what I think osx86 is all about.

I used Apple's FileMerge (part of Xcode install) to compare today's (26 Oct) versions of semthex's and mikfi's mods. semthex needs to FileMerge from Apple's original source as he made white space (added or removed tabs, spaces and/or blank lines) changes to several files the probably were left over from debugging - just a nit, but it will make future updates a lot easier.

 

semthex added a lot of exported symbols so older kext's will still load => very good idea. The so-called R2D2 hacking code (decodes encrypted files like on a Apple machine), which violates the DCMA in the U.S. if fully coded, appears in both versions, minus the "secret key" numbers that were determined by some smart guys at Hackint0sh.

 

mikfi made the key breakthrough in IORegistryEntry.cpp that replaces EFI, of course it doesn't provide drivers or the same info as Apple's real EFI, so video drivers are still a hard problem; plus Apple only supports a few cards anyway. semthex does similarly and gives mikfi credit - honorable hacking.

semthex corrects me (edit 27 Oct)

Hmm this is just part of truth, since this I was working on this and finished work a day before but rufus and me were unable to find the rigth smbios, which was the key breaktrough, so I gave him credit for this :-). Netherless it was a hard fight and I respect mifki and his work, as I once said I am not intendent to bash anyone especially not skilled people.

The post-10.4.4 kernels need to know the FSB (front side bus) speed, it's 200MHz for most older machines, but semthex only allows three values, while mikfi adds fsb=n where n is the speed in MHz - trivial change, but I like it better than semthex's 3 switch approach (so would overclockers).

 

semthex attacked the non HPET chips and added RTC (real time clock) support as in older kernels. I don't know if HPET still works in his version, but it will run on HPET machines.

 

semthex added, but commented out, my fix (small bow) for Pentium 4 Northwood chips with HTT, SSE2, but no SSE3 returning the number of cores per package as 0. This has to be uncommented or the kernel crashes with a divide by zero exception in cpu_threads.c if running on that CPU - it has no effect on correct CPUID machines.

 

So great job and congrats to mikfi and semthex, now that the heavy lifting is done further tweaks should easy on Intel SSE2 machines. The kernel mods for AMD should not be difficult.

[jbond]

I get a kernel panic if I do update_prebinding -root / -force, in GUI as well as single user mode with Vitality kernel

 

Can anyone help?

 

 

EDIT: New Kernel Fixed Issue.

[BuXb]

Thanks Tom for your comparision and evaluation !

[kerkenez]

i'm posting first time to his forum but i'm reading nearly 10 month

i have no problem by the 10.4.6 goatsex cd, 10.4.8 jas things...

except, you know humbling rosetta

after using clean and unpatched install another part in the same

hd using goatsex cd (yes it is unpatched sys) then 10.4.8 real

combo upd, permission repair, disk diagnostic, all is ok.. finally

i replaced originals, mifki/vitaliy's kernel and AppleSMbios.kext

sudo chown root:wheel for two elements

adding appleazaliaaudio.kext for sound, removing applehda, replacing

ionetworkingfamily.kext with patched one again sudo chown root:wheel

for three

 

result

with the standart settings of mobo bios, it reset continuously just after

loading starts, after changing cpu setingin bios (that i can't remember

now), it gives dns responder errors in verbose mode and than crash dump....

 

some one know about the problem...?

 

thanks all efforts of mifki, vitaliy, shemthex, jass and all the other guys

actually, when i earn extra few bucks i'll donate all of you

 

asus p5ld2-vm se mobo

core2 e6300 overclocked@2445mhz(fsb350 after 357don't work)

gma950 onboard video, this is enough for me i can't play but gma950 can basicly

pci freq 120 (after 120 it lost sata bus, but ide works)

667mhz 2x512 cheap high level stick overclocked@700mhz memory 4-4-4-14-4

160 sata seagate ncq, average hd

80 ide seagate not average ide hd

philips200W6cs-00 quite good monitor

lg4167 dvd-rw

no case fan(disconnected) only in stock cpu fan, ican hear hard drive sound clear

geekbench 252 - 259

xbench 137

[Proteo]

when i earn extra few bucks i'll donate all of you

 

In the meantine, you can go to their respective sites and click some banner ads

[Ezhoon]

Finally! It worked!

 

My mobo is Asus P5LD2.

In the bios ->advanced ->CPU configuration ->Execute Disable Function (=nx)

This must be "Enabled"

 

Once I turned it on, immediate reboot problem just disappeared and

took me right onto desktop, yay!

 

Thanks for your hard work, Vitaliy.

YOU deserve our respect.

[Jeezoflip]

You know...im fine chilling with 10.4.7. I dont see what the big rush to 10.4.8 is. Its not like its a noticeable difference, and right now we will just mess up our computers trying to get this to work. Im waiting untill theres a super stable update for this.

 

EDIT:

 

btw, that is going out to all the users, not the devs. I dont want anyone getting the wrong idea. I love the devs, and am thankful they stick around.

[tokyovigilante]

You know...im fine chilling with 10.4.7. I dont see what the big rush to 10.4.8 is. Its not like its a noticeable difference, and right now we will just mess up our computers trying to get this to work. Im waiting untill theres a super stable update for this.

 

EDIT:

 

btw, that is going out to all the users, not the devs. I dont want anyone getting the wrong idea. I love the devs, and am thankful they stick around.

I realise at this stage the 10.4.8 KERNEL (not userspace) upgrade is tough for most people to get their heads around, but the point of these kernel updates is that we should see big performance upgrades from the 10.4.3/4 kernels which were binary hacked by Maxxuss back in the day. This is a much cleaner and hopefully more stable way to do it. Also, a super stable update only comes when lots of people have tested the pioneering hacks like these.

 

I'm just hanging out for AMD patches for the source, I had a play myself, but I am not a Kernel hacker by any stretch.

 

Thanks to all the developers for making this possible, you guys are champs.

[jbond]

Ok folks, now that I got my test machine working perfectly with 10.4.8 (vitality kernel). I went on to try and update my main hackintosh box (which is on my sig.) I installed the Combo 10.4.8, then I copied the new vitality kernel (one with fsb), and SMBios.kext into its proper location, fixed diskpermissions. and reboot.

 

No problems getting into singleuser (-s) mode. did update_prebinding -root / -force. Worked fine. Then I reboot.

 

Now I have a problem where coreserviced is crashing and respawning? It goes into GUI with blue screen and mouse, then it goes back to black screen with message about coreserviced crashing.

 

Can anyone help figuring this out - what gives?

[T u r b o]

I get the same thing. I tried the binary kernels as well as compiling my own w/ the proper adjustments. On my system coreservicesd crashes and loginwindow goes into a crash loop (alternating blue screen and console)

 

I've got the crash logs if somebody who knows what it means wants to take a look.

[lmaurad]

Ok I'm going a little bit crazy here, maybe someone can help me out.

 

I have compiled the kernel, copied it over to the 2nd OS X unpatched partition.

Fixed Permissions.

I copy AppleSMBIOS.kext and r2d2.kext to the extension folder

fixed Permissions

 

And when i reboot, all the kext begin loading and then it just reboots.

 

Please help.

[abcslayer]

Wow. Wonderful. Apple has shutdown the sourceforge site due to those incredible things. So busy at work and I feel so sorry to know those news late. I make me got the feeling as this time last year when I went online everyday just to update information about new Maxxuss releases.

I think a bigger kernel which has a small portion of code to make it auto configure for SS2, SSE3, NX or non-NX,... is better and suite with the idea: Hackint0sh for almost every PCs.

Cheer.

[mifki]

the sourceforge site has been down for more than one month

[tokyovigilante]

the sourceforge site has been down for more than one month

http://www.opensource.apple.com/darwinsource/tarballs/apsl/

They're a bit obscure but they're there...

[semthex]

"mikfi made the key breakthrough in IORegistryEntry.cpp that replaces EFI, of course it doesn't provide drivers or the same info as Apple's real EFI, so video drivers are still a hard problem; plus Apple only supports a few cards anyway. semthex does similarly and gives mikfi credit - honorable hacking."

 

Hmm this is just part of truth, since this I was working on this and finished work a day before but rufus and me were unable to find the rigth smbios, which was the key breaktrough, so I gave him credit for this :-). Netherless it was a hard fight and I respect mifki and his work, as I once said I am not intendent to bash anyone especially not skilled people.

[giandal]

"mikfi made the key breakthrough in IORegistryEntry.cpp that replaces EFI, of course it doesn't provide drivers or the same info as Apple's real EFI, so video drivers are still a hard problem; plus Apple only supports a few cards anyway. semthex does similarly and gives mikfi credit - honorable hacking."

 

Hmm this is just part of truth, since this I was working on this and finished work a day before but rufus and me were unable to find the rigth smbios, which was the key breaktrough, so I gave him credit for this :-). Netherless it was a hard fight and I respect mifki and his work, as I once said I am not intendent to bash anyone especially not skilled people.

 

I guess I do not understand that completely, but does this mean that at least the graphic cards that are present in the "real macs" work without hacks (like macvidia drivers and co.)???

[semthex]

I guess I do not understand that completely, but does this mean that at least the graphic cards that are present in the "real macs" work without hacks (like macvidia drivers and co.)???

 

I don't understand your question with the quoute :-)

[giandal]

I don't understand your question with the quoute :-)

 

What about it without the quotation marks??

 

The statement by Tom H

 

"mikfi made the key breakthrough in IORegistryEntry.cpp that replaces EFI, of course it doesn't provide drivers or the same info as Apple's real EFI, so video drivers are still a hard problem; plus Apple only supports a few cards anyway."

 

seems to imply that the few cards supported by Apple (including some nVidia ones) may work without the need of different kexts than those already present in the system... is that true or I am reading too much in that sentence?

[rodrigocunha]

I had the same problem. Just copy the IOATAFamily.kext from a previous OSX version.

 

Now it works!!!

 

Thanks prolinesurfer!

 

But a i got a new problem -> NO CI/QE/GL with my GMA900.

 

I removed GMA950 kexts but now i have a black desktop with corrupted windows, texts, etcs.. it´s unusable.

So, i restarted all process.

Now i have a 10.4.8 but no CI/QE/GL.

 

 

OMG, what a pain!

[jbond]

UPDATE - This should interest some of you folks.

 

I took the HDD out of the test machine and put it into my hackintosh (sig), and the machine works flawlessly on first boot. Even About this Mac started working on the hackintosh, which was not working on the test machine. Also the Hardware overview works fine.

 

This experiment leads me to believe that on the same HDD diff mobo have different results. I cannot explain why the About this mac works on one machine (same physical hdd) and not on the other. I think it may have to do with the EFI environment thereoff.

 

My test machine (Asus P5WD2 Premium) has no EFI and a plain BIOS. The hackbox has a EFI based BIOS (Intel 945 GTPLKR).

 

The hackbox can see the RAM FSB as 667 where as the Asus reports Unknown for RAM FSB.

 

My guess is that the hackbox is able to retrieve information from the EFI environment somehow, but when I swap the HDD to the ASUS, it is unable to do that for the ASUS BIOS which is not EFI based.

 

I hope this helps you coding gurus to figure out something that may help us with another big cat coming our way.

[Menno]

Ok folks, now that I got my test machine working perfectly with 10.4.8 (vitality kernel). I went on to try and update my main hackintosh box (which is on my sig.) I installed the Combo 10.4.8, then I copied the new vitality kernel (one with fsb), and SMBios.kext into its proper location, fixed diskpermissions. and reboot.

 

No problems getting into singleuser (-s) mode. did update_prebinding -root / -force. Worked fine. Then I reboot.

 

Now I have a problem where coreserviced is crashing and respawning? It goes into GUI with blue screen and mouse, then it goes back to black screen with message about coreserviced crashing.

 

Can anyone help figuring this out - what gives?

 

I have exactly the same problem, I see you have the same hardware, any progress lately?