10.6.5 legacy kernel for INTEL/AMD 32+64bit ready to download by qoopz/nawcom/AnV/BlackKnight/mucha - a few new features

[dickhouse]

Thanks for replying

 

So maybe it's still in 32-bit mode

 

Just to make it sure,

is the 64-bit version of Geekbench able to run on your AMD with "arch=x86_64" ?

 

see my previous post! rollin' back to page 2/3 perhaps.. running geekbench in 64bit mode using arch=i386 -force64 (screenshot will tell you everything), all applications runs natively 64/32bit without single error!

 

no luck using arch=x86_64..

 

but so far i am happy with kernel made by those great guys!

[mac_carol]

see my previous post! rollin' back to page 2/3 perhaps.. running geekbench in 64bit mode using arch=i386 -force64 (screenshot will tell you everything), all applications runs natively 64/32bit without single error!

 

no luck using arch=x86_64..

 

lol... just calm down mate

 

Yes!!! we've already known when booting with "arch=i386 -force64", both 32 & 64 bits apps will work like a charm

 

 

 

But yesterday, we're just discussing about running the osx with "arch=x86_64" on any AMD machines

 

And in conclusion, finally we (or at least me ) had to agree with what nawcom said:

 

Don't boot in x86_64 yet if this is for an AMD processor. Stay with 64 bit userspace for now.

 

but so far i am happy with kernel made by those great guys!

[dickhouse]

@ mac_carol ok! hey mac can i ask you question.. since your rig r similiar to mine, did you get proper sleep on your hackintosh? i have a trouble when i try to wake up my hack1ntosh.. usb froze! are you getting the same like mine! *usb dev draw to much power draw bla..bla..*

 

the funniest things is dat i have unplugged all the usb dev before went to sleep and bang! those popup quiz came out and system hang!

[jeisom]

I got my system to finally boot ( still running 10.6.3) using the kernel pkg from this forum. I have a athlon II 640 with a ASUS M4A78 PRO motherboard. To boot I need to disable ehci usb in the bios and "arch=i386". If I do use ehci the kernel hangs, no KP. if I use -force64 some apps don't run,(Firefox for example). any ideas on the usb2 issue?

[Andy Vandijck]

I got my system to finally boot ( still running 10.6.3) using the kernel pkg from this forum. I have a athlon II 640 with a ASUS M4A78 PRO motherboard. To boot I need to disable ehci usb in the bios and "arch=i386". If I do use ehci the kernel hangs, no KP. if I use -force64 some apps don't run,(Firefox for example). any ideas on the usb2 issue?

Try my custom IOUSBFamily

[jeisom]

Try my custom IOUSBFamily

 

Didn't work. Dies "Attempting to get EHCI controller from bios" I believe.

EDIT: Could this be a bios bug?

 

[Mohamed Khairy]

Didn't work. Dies "Attempting to get EHCI controller from bios" I believe.

EDIT: Could this be a bios bug?

 

 

try update bios and also play around usb setting on bios setting

[jeisom]

try update bios and also play around usb setting on bios setting

 

Had the lastest and even tried downgrading. The only thing I can find that works is disabling EHCI all together.

[jeisom]

Had the lastest and even tried downgrading. The only thing I can find that works is disabling EHCI all together.

Did some more testing. If I set "USB 2.0 Controller Mode" to FullSpeed I have a 50/50 chance of actually get the system up with EHCI controllers running. It appears to stall somewhere else then on the times it fails to boot.

[340frankie]

works good on my phenom II x2 550 at stock frequency/settings , i got 64bit working too.. but when i unlock the 4 cores the system wont boot.. i got an infinite boot sequence but it never reaches the desktop... with 10.6.2 and blackknight 10.2 kernel i was able to use all 4 cores at 3.56 ghz (impressive) but only 32bit apps were working.. so.. any advices??

 

my config:

 

phenom x2 550 3.1 ghz

asrock n68c-s

4gb ddr3 1333

4870 vapor X 1gb ddr5

 

TNX

[M@dMac]

Working great here with arch=i386 -force64, I'm using the one posted by nawcom because it shows V5, hehe, anyhow... just wanted to let you guys know that there was a moment when I could boot with arch=x86_64, then I noticed that the latest release of iasl was doing something weird, it removed the _Yxx from some places, it also removed stuff or modified it, can't tell much because all I did was get into linux to dump a clean copy of the pc's DSDT then used iASLme with the iasl I compiled using kabyl's instructions at the other site, then did a complete fix without doing much hacking, only the errors and warnings, then tested your kernel and it worked fine with either arch=i386 or arch=x86_64 but only with the first one did 32 bit appz run, after a few trials I was noticing a constant lag and it bothered me so much that I left it only with arch=i386, getting no better results at all, then I remembered that HPET has stuff to do with it, fired ioregexp and right away I saw there was no HPET!!! so I went searching my system for an older copy of DSDT, it's worth to keep copies!!! I will publish both pieces so you people get aware of this, be back soon...

 

btw, after I fixed HPET I couldn't boot with arch=x86_64 anymore... so if anyone gets instant reboot after using that key... your HPET is working great!!!

 

 

 

@340frankie look at what @Mohamed Khairy wrote a few post obove: try to update your BIOS to latest version, *BUT* don't play with settings, just see if there is a C states stuff... disable it, also disable legacy usb, leave the rest as is and save your dsdt, try to fix it, see if your system boots w/fixed dsdt, then see if you get HPET working, then see if 4 cores working, we should draw this in google docs draw, I have seen some nice flowcharts from there, anyhow, get back with info about how it went...

[340frankie]

[quote name='M@dMac' date='Dec 13 2010, 04:04 PM' post='1601510'

 

 

@340frankie look at what @Mohamed Khairy wrote a few post obove: try to update your BIOS to latest version, *BUT* don't play with settings, just see if there is a C states stuff... disable it, also disable legacy usb, leave the rest as is and save your dsdt, try to fix it, see if your system boots w/fixed dsdt, then see if you get HPET working, then see if 4 cores working, we should draw this in google docs draw, I have seen some nice flowcharts from there, anyhow, get back with info about how it went...

 

i got a little change today.. at the bootloader i typed in cpus=1 and it boots with the 4 core config.. BUT when it raches the desktop i was only able to see the dock and the hand glass on the right corner of the screen... no bar at all!! and when i try to open the finder it says "could not open the finder with a -0108 or something like that error"

 

do you have some suggests? this will happen only with the 4 cores activated but it works fine with 2...

 

TNX

[gedna]

Starting to get some random Kernel Panics, out of nowhere, using nawcom installer and 32bit flag kernel... (couse 64bit is sooooo glitchy)

[hnak]

After patching dyld, libSystem.B.dylib and QuickTimeComponents, my system works smoothly with arch=i386 -force64 fsb=20000000 with Chamelon rc4. ( rc5r6xx would not boot )

I was not able to use it without specifying maxmem=... before. Thanks.

 

System:

Phenom II X6 1055T

8GB RAM

ASUS M4A78 Plus

Radeon HD4850

[dickhouse]

So quite in here...!

[Azimutz]

Hi guys... Nawcom, Andy, sorry for the long silence and thanks for the great work done, keeping the flame alive

Also thanks for mentioning me i'm just a "carrier" for Qoopz work, which i should have shared with you guys sooner;

anyway, the timing turned out to be fine, i think.

Ok, down to business...

 

Nawcom, i been reading your "calls" for organizing this project and i totally agree; we should get a place to host this and the Voodoo Projects repo seems to me the best place for it atm. I assume you still didn't got a word from Zef!?

 

About the kernel, i have some "thoughts" to share.. for now i'm just going to mention the "removing -force64" thing i mentioned on post #19.

I didn't explained my self properly on that post because i still didn't had the time to look at the code and test the kernel properly.

To explain (and not only for you guys with the knowledge), the patch related with -force64 was introduced on the Voodoo kernel, so we could force some cpu's into Legacy Mode (the mode 32 bit only cpu's run). The -force64 argument is just a way to override that behavior and it's not a native argument, as some might thing.

The problem is that the patch was created on a time were the kernel had only i386 arch and it's inserted on a piece of code that is only executed when a 32 bit only cpu is being used or arch=i386 is specified (#ifdef __i386__).

Keeping the previous info in mind, the behavior of the V5 kernel (which is i386/x86_64) on this matter goes like this:

- if the cpu is identified as 64 bit capable, the kernel will boot x86_64 arch by default, ignoring the patch,

which is fine by me

- if the cpu is identified as 32 bit only or arch=i386 is specified, the kernel will enter Legacy Mode,

which can be done on the vanilla kernel by specifying "arch=i386 -legacy".

- to run the kernel in 32 bit with "64 bit mode" enabled, we need to specify "arch=i386 -force64",

which can be achieved by specifying only arch=i386 on the vanilla kernel.

 

So, do you guys get my point of view on this?

 

See ya later...

[Ghaith]

Hello guys,

 

I am just new posting to this forum but I was reading it since long time...

 

I had installed 10.6.4 on my HP Compaq 6530b Laptop on an external USB drive, that was working great until I upgraded to 10.6.5 when my system got down.

 

After two days of reparations using OS X installed on a VMWare I could repair the files in my hard drive but unfortunately I have lost the mach_kernel, and then I have downloaded the mach_kernel posted in this topic, I unzipped it, copied it to the root on my hard drive and applied chown 0:0 and chmod 755 on the mach_kernel using sudo -s in the terminal and rebooted using the USB drive with -v arch=i386 flags...

 

When start loading the files, it is showing that it is loading the mach_kernel TWICE!! as per the image (1.jpg), I can see from the image (2.jpg) that the Darwin Kernel loaded is 10.5.0 which is correct I assume? when continuing with the boot it is showing messages saying invalid kernel ip, won't attempt.... as per image (4.jpg)

 

If I am booting with only flag -v i am getting the screen (3.jpg)

 

Can you please guys help me out on this.

 

Thanks in advance...

 

Ghaith

[mbembi]

helo everyone, thanks before for andy & nawcom.

 

I had ready installed with legacy_kernel 10.5, without kernel panic, just warning like

"dyld : shared cached file was build against a different libsystem"

My com.apple.boot.plist : arch=x86_64 -notscverify -force64, and i'v got my System Software Overview 64-bit Kernel and Extensions: Yes.

 

The problem is .app crash every time i ran, like firefox, opera, vlc, mplayer etc... , i think may be cause osx86 run on 64 bit. What your suggest kernel for AMD, usualy i used legacy_kernel 10.3 and run good.

 

 

my machine:

AMD X2 Athlon 4200+

Biostar TF7050-M2 (GF7050 PV / NF630)

VGA on board (Geforce 7050)

1Gx2 @6400

[akiNikki]

Perhaps that x86_64 flag you use.. Try using arch=i386 and -force64.

[Nomad_Grabitel]

helo everyone, thanks before for andy & nawcom.

 

I had ready installed with legacy_kernel 10.5, without kernel panic, just warning like

"dyld : shared cached file was build against a different libsystem"

My com.apple.boot.plist : arch=x86_64 -notscverify -force64, and i'v got my System Software Overview 64-bit Kernel and Extensions: Yes.

 

The problem is .app crash every time i ran, like firefox, opera, vlc, mplayer etc... , i think may be cause osx86 run on 64 bit. What your suggest kernel for AMD, usualy i used legacy_kernel 10.3 and run good.

 

 

my machine:

AMD X2 Athlon 4200+

Biostar TF7050-M2 (GF7050 PV / NF630)

VGA on board (Geforce 7050)

1Gx2 @6400

 

Run your apps in x32 mode. Or you can also try amd_inst_patcher.

[Andy Vandijck]

helo everyone, thanks before for andy & nawcom.

 

I had ready installed with legacy_kernel 10.5, without kernel panic, just warning like

"dyld : shared cached file was build against a different libsystem"

My com.apple.boot.plist : arch=x86_64 -notscverify -force64, and i'v got my System Software Overview 64-bit Kernel and Extensions: Yes.

 

The problem is .app crash every time i ran, like firefox, opera, vlc, mplayer etc... , i think may be cause osx86 run on 64 bit. What your suggest kernel for AMD, usualy i used legacy_kernel 10.3 and run good.

 

 

my machine:

AMD X2 Athlon 4200+

Biostar TF7050-M2 (GF7050 PV / NF630)

VGA on board (Geforce 7050)

1Gx2 @6400

1.:

Boot with following boot parameters:

arch=i386 -force64

This will enable the use of the 32bit kernel (in 64bit userspace mode).

I'm still investigating why the 64bit kernel on AMD won't allow 32bit userspace apps to run...

At current with the 64bit kernel on AMD you can only run 64bit apps (AFAIK)

 

2.:

"dyld : shared cached file was build against a different libsystem"

You can fix this from within Terminal with the following command:

sudo update_dyld_shared_cache

This will fix up the shared cache (and thus recache libSystem.B.dylib, which is causing this)

[akiNikki]

How about this anv? I paste it from my console.

 

1/1/11 11:39:32 PM	kernel	[MSR] detected invalid wrmsr([MSR] detected invalid wrmsr([MSR] detected invalid wrmsr(00000287, 00000000:00000287, 00000000:[MSR] detected invalid wrmsr(40000001) at 0x40000001) at 0x00000287, 00000000:2e4486ba
1/1/11 11:39:32 PM	kernel	00000287, 00000000:40000001) at 0x2e4486ba
1/1/11 11:39:32 PM	kernel	2e4486ba
1/1/11 11:39:32 PM	kernel	40000001) at 0x2e4486ba
1/1/11 11:39:32 PM	kernel	[MSR] detected invalid rdmsr([MSR] detected invalid rdmsr([MSR] detected invalid rdmsr(00000288) at 0x00000288) at 0x00000288) at 0x[MSR] detected invalid rdmsr(2e44869a

[Andy Vandijck]

How about this anv? I paste it from my console.

 

1/1/11 11:39:32 PM	kernel	[MSR] detected invalid wrmsr([MSR] detected invalid wrmsr([MSR] detected invalid wrmsr(00000287, 00000000:00000287, 00000000:[MSR] detected invalid wrmsr(40000001) at 0x40000001) at 0x00000287, 00000000:2e4486ba
1/1/11 11:39:32 PM	kernel	00000287, 00000000:40000001) at 0x2e4486ba
1/1/11 11:39:32 PM	kernel	2e4486ba
1/1/11 11:39:32 PM	kernel	40000001) at 0x2e4486ba
1/1/11 11:39:32 PM	kernel	[MSR] detected invalid rdmsr([MSR] detected invalid rdmsr([MSR] detected invalid rdmsr(00000288) at 0x00000288) at 0x00000288) at 0x[MSR] detected invalid rdmsr(2e44869a

That's the invalid rdmsr/wrmsr catcher which handles unsupported wrmsr/rdmsr traps (which I added from Voodoo kernel sources).

Don't worry these get handled by the kernel...

[340frankie]

hi guys.. i can only boot with cpus=1 so by using only 1 core.. how can i resolve this issue? please look at my signature i took my config there..

 

TNX

[Andy Vandijck]

hi guys.. i can only boot with cpus=1 so by using only 1 core.. how can i resolve this issue? please look at my signature i took my config there..

 

TNX

Take a picture when booting verbosely with all cores enabled...