InfiniteMac OSx86 - Legacy kernel 10.6.8 with custom dyld and new SSE3 emulator and VIA Nano cpu support

Page 5 of 6

Show 50 post(s) from this thread on one page

InfiniteMac OSx86 (http://infinitemac.com/forum.php)

- 10.X (http://infinitemac.com/forumdisplay.php?f=36)

- - Legacy kernel 10.6.8 with custom dyld and new SSE3 emulator and VIA Nano cpu support (http://infinitemac.com/showthread.php?t=7318)

akimoa

11-09-2011 12:19 AM

here the missing one sorry hope it makes a diffrence(link corrected to)

http://cl.ly/0X0w3s300U002z0W0K3t

andyvand

11-09-2011 12:24 AM

Quote:

Originally Posted by akimoa (Post 55922)

here the missing one sorry hope it makes a diffrence(link corrected to)

http://cl.ly/0X0w3s300U002z0W0K3t

Any report yet on the x86_64 kernels?
I hope at least those work, then I know where to look. :-|

akimoa

11-09-2011 12:26 AM

well on both x86 _64 Chameleon stops rite after i hit enter goes back to command prompt at least here

andyvand

11-09-2011 12:32 AM

Quote:

Originally Posted by akimoa (Post 55924)

well on both x86 _64 Chameleon stops rite after enter goes back to command prompt at least here

I'm almost done compiling the universal (32+64bit) kernel (using Xcode 3.2.6).
Lets hope that at least works... :-/

akimoa

11-09-2011 12:39 AM

will test yours tomorrow im done for today

davisin666

11-09-2011 12:40 AM

Quote:

Originally Posted by andyvand (Post 55925)

I'm almost done compiling the universal (32+64bit) kernel (using Xcode 3.2.6).
Lets hope that at least works... :-/

I'm crossing my fingers :-|

Sn0wStorm

11-09-2011 12:44 AM

Quote:

Originally Posted by akimoa (Post 55924)

well on both x86 _64 Chameleon stops rite after i hit enter goes back to command prompt at least here

Maybe youre using arch=i386 ?
When trying that on the x86_64 Kernels, it goes back to the command promt for me

Otherwise still Instant Reboot :(

andyvand

11-09-2011 12:47 AM

Ok, I finished compiling the noautopatch version using Xcode 3.2.6.
It's built as a ub (i386+x86_64 in one bin).

Download:
https://rapidshare.com/files/2791902...oautopatch.zip

Please test and let me know if it works.

andyvand

11-09-2011 12:58 AM

I tried it on my 32bit Intel netbook.
It doesn't instant reset but it gets stuck on a black screen.
I'm using Cparm's latest bootloader...
I bet that if you test this kernel on a supported Intel, it works...
Could anybody with vanilla cpu test this hypothesis?
Try both my compiled version as well as Akimoa's... :-|

Sn0wStorm

11-09-2011 01:34 AM

doesnt work :( tried arch=i386 with/without -force64 and arch=x86_64
maybe I can test vanilla tomorrow.

m28ew

11-09-2011 02:55 PM

I've tested those new kernels on my AMD P-II x4 965

with flags: arch=i386 ; arch=i386 -force64

there's no KP,
but I got stuck at the "No interval found for . Using 8000000" line

http://www.postimg.com/52000/photo-51473.jpg

with flag: arch=x86_64

I got this KP

http://www.postimg.com/52000/photo-51475.jpg

akimoa

11-09-2011 03:20 PM

and wich kernel did u use for that ? bit more specific if u dont mind:-d

andyvand

11-09-2011 03:43 PM

Quote:

Originally Posted by m28ew (Post 55936)

Can you boot into single user?
Then you could get the dmesg to me... I could use it to look up what goes wrong... :D
A new hope... :-d
EDIT: The second (the kp) is the classic 64bit AMD AppleACPIPlatform.kext panic...

m28ew

11-09-2011 03:48 PM

Quote:

Originally Posted by akimoa (Post 55937)

and wich kernel did u use for that ? bit more specific if u dont mind:-d

I tested all those 5 kernels of yours & Anv's :)
NoAutopatch & Autopatch @ 32 & 64 bits compiled kernels

and I added them all into the iATKOS L1 disc

http://www.postimg.com/52000/photo-51484.jpg

m28ew

11-09-2011 03:58 PM

Quote:

Originally Posted by andyvand (Post 55938)

Can you boot into single user?
Then you could get the dmesg to me... I could use it to look up what goes wrong... :D
A new hope... :-d

unfortunately, with: arch=i386 (-force64) -s
I still got stuck at the same line :'(

http://www.postimg.com/52000/photo-51486.jpg

andyvand

11-09-2011 06:01 PM

Quote:

Originally Posted by m28ew (Post 55940)

unfortunately, with: arch=i386 (-force64) -s
I still got stuck at the same line :'(

http://www.postimg.com/52000/photo-51486.jpg

Try setting a scheduler with one of the following boot arguments.

traditional scheduler: sched=traditional
traditional scheduler with pset runqueue: sched=traditional_with_pset_runqueue
proto scheduler: sched=proto
grrr scheduler: sched=grrr
fixed priority scheduler: sched=fixedpriority
fixed priority scheduler with pset runqueue: sched=fixedpriority_with_pset_runqueue

Only use one and please test all to see if there is a difference.
These control how the processor schedules instructions.

On AMD you may also want to use boot argument npci=0x2000

Still investigating for the interval lines... ;)
EDIT: Oh yeah... are you booting from a DVD? Try using an USB boot... (may be worth trying)
Also grep for the string "No interval found" in all the .kext files with the following command:
grep -E "No interval found" /System/Library/Extensions/*.kext/Contents/MacOS/* /System/Library/Extensions/*.kext/*
Tell me if one matches...

davisin666

11-09-2011 06:28 PM

Pthread support ABORTS when sync kernel primitives misused
No interval found for . Using 8000000

Looks like we need to compare the log with the vanilla dmesg if these messages appears

andyvand

11-09-2011 06:31 PM

Quote:

Originally Posted by davisin666 (Post 55942)

Pthread support ABORTS when sync kernel primitives misused
No interval found for . Using 8000000

Looks like we need to compare the log with the vanilla dmesg if these messages appears

I saw this one appearing on my Intel with DP1 too: Pthread support ABORTS when sync kernel primitives misused
But No interval found for is not part of the kernel... either is coming from an app or coming from a kext...
Still investigating...
EDIT: Can't find it anywhere in DP1... don't have GM so can't check...
Did changing the scheduler change anything?
How about booting the installed product? This is the install DVD, right?
EDIT2: Did you patch /usr/lib/dyld and /usr/lib/libSystem.B.dylib with amd_insn_patcher (the new one for Lion)? It is needed for 32bit support under 64bit...

akimoa

11-09-2011 07:41 PM

K did some intence testing with an installed system (cloned from my MBP when it came out in Juli 10.7.0)

only the kexts i used from SL

Disabler ,NULLPower, FakeSMC, rest is via DSDT

Bootloader i tried were cparm boot(latest) and Chameleon Rev1684

with all kernels and with none i came close to see a verbose mode sadly

tried all scheduler arguments as boot args gave me instant reboots too (on all)

all important bootflags were tested aswell offcourse

:'(

Sn0wStorm

11-09-2011 07:51 PM

1 Attachment(s)

Got one match, 10.7.0 Installer Stick

Code:

grep -E "No interval found" /Volumes/Mac\ OS\ X\ Base\ System/System/Library/Extensions/*.kext/Contents/MacOS/* /Volumes/Mac\ OS\ X\ Base\ System/System/Library/Extensions/*.kext/*



Binary file /Volumes/Mac OS X Base System/System/Library/Extensions/IOHIDFamily.kext/Contents/MacOS/IOHIDFamily matches

For 32 Bits: the new amd_insn_patcher gives me on both dyld and libSystem.B.dylib:

Code:

Code signature removed succesfully (64bit)No patches found, not generating output filePatch report: 0 instructions patched, 208 bad instructions, patches bypassed: NO

andyvand

11-09-2011 11:03 PM

2 Attachment(s)

Fixed amd_insn_patcher... :-|
It's IOHIDFamily that causes the misery allright...
Decompiling + checking...
EDIT: Removed check for interval (which causes the failure).
The problem in this case is VoodooPS2 in fact but the kext I posted should do fine ;) (little workaround with hex editor ;))
Back up your original before testing... :)

davisin666

11-10-2011 12:32 AM

It seems like xmove doesn't work very well with amd, gotta mod iatkosx L1, I will have to buy a blank dvd xddd

akimoa

11-10-2011 07:15 AM

im starting to think something else in my cloned system is bugging the boot ...investigating

$The-King$

11-10-2011 09:54 AM

hi people
i hadnt entered this forum about 4 days
ad got adream that there are alion kernel done :D

so i will do testing but can any one compiles 11.2 kernel as i have 10.7.2
any way will begin with 11.0 kernel

will do tests now :)

m28ew

11-10-2011 10:41 AM

Quote:

Originally Posted by andyvand (Post 55943)

Did changing the scheduler change anything?

still no luck :'(

Quote:

Originally Posted by andyvand (Post 55943)

How about booting the installed product? This is the install DVD, right?

all testings were done by booting the osx from DVD (BSD Root=disk1s3) & HDD (BSD Root=disk0s2)
and they all stuck at the same line

Quote:

Originally Posted by andyvand (Post 55943)

EDIT2: Did you patch /usr/lib/dyld and /usr/lib/libSystem.B.dylib with amd_insn_patcher (the new one for Lion)? It is needed for 32bit support under 64bit...

I haven't patched them yet
could you also share the binary of the amd_insn_patcher ? :)

andyvand

11-10-2011 10:49 AM

Quote:

Originally Posted by m28ew (Post 55954)

still no luck :'(

all testings were done by booting the osx from DVD (BSD Root=disk1s3) & HDD (BSD Root=disk0s2)
and they all stuck at the same line

I haven't patched them yet
could you also share the binary of the amd_insn_patcher ? :)

Should be there under the source... amd_insn_patcher...

m28ew

11-10-2011 10:55 AM

Quote:

Originally Posted by andyvand (Post 55947)

the new patched IOHIDFamily got rid of the "No Interval found for" error

but now I'm getting stuck after the USBF line

http://www.postimg.com/52000/photo-51531.jpg

akimoa

11-10-2011 11:01 AM

here are the compilied ones for 10.7.2 to test

no patch

http://cl.ly/191M3B3j3U0M2i2L1D0R

autopatch

http://cl.ly/2o3z2M1u161N2V3e0p0P

$The-King$

11-10-2011 11:07 AM

Quote:

Originally Posted by akimoa (Post 55957)

here are the compilied ones for 10.7.2 to test

no patch

http://cl.ly/191M3B3j3U0M2i2L1D0R

autopatch

http://cl.ly/2o3z2M1u161N2V3e0p0P

thanks tried anv kernel just auto restart

$The-King$

11-10-2011 11:20 AM

tried four kernel no one fo them work just auto restart :'(

$The-King$

11-10-2011 11:22 AM

i will try on vmware fusion

akimoa

11-10-2011 11:22 AM

@m28ew

Try using patched IOUSB from AnV (SL) might change a thing

and Andy cant we set a debug somewhere ? besides the scheduler aruments ?

m28ew

11-10-2011 11:26 AM

Quote:

Originally Posted by andyvand (Post 55955)

Should be there under the source... amd_insn_patcher...

thanks Andy :)

when I patched the dyld, I got these messages:

Code:

Patching universal binary (2 architectures)

Patching X86_64 part (processor 16777223, architecture 0)

Code signature removed succesfully (64bit)Patch report (1): 0 instructions patched, 208 bad instructions, patches bypassed: NO

Patching I386 part (processor 7, architecture 1)

Code signature removed succesfully (32bit)Patch report (2): 1 instructions patched, 0 bad instructions, patches bypassed: NO

but when I pacthed the libSystem.B.dylib,
I got the "No patches found, not generating output file" message ???

Code:

Patching universal binary (2 architectures)

Patching X86_64 part (processor 16777223, architecture 0)

Code signature removed succesfully (64bit)Patch report (1): 0 instructions patched, 30 bad instructions, patches bypassed: NO

Patching I386 part (processor 7, architecture 1)

Code signature removed succesfully (32bit)Patch report (2): 0 instructions patched, 0 bad instructions, patches bypassed: NO

No patches found, not generating output file

are those reports correct for you ?

btw, is there any difference between the amd_insn_patcher and the amd_insn_patcher_ext ?

wastez

11-10-2011 11:26 AM

Quote:

Originally Posted by $The-King$ (Post 55959)

tried four kernel no one fo them work just auto restart :'(

Akimoa and me are having the same issue. :(
Hope someone will fix that.

@anv
Would it be possible to insert a debug mode?

$The-King$

11-10-2011 11:33 AM

Quote:

Originally Posted by wastez (Post 55963)

Akimoa and me are having the same issue. :(
Hope someone will fix that.

@anv
Would it be possible to insert a debug mode?

:(
downloading fusion and will see if will work or no

i tried kernel on my notebook core i and kernel load

wastez

11-10-2011 11:35 AM

Quote:

Originally Posted by m28ew (Post 55962)

thanks Andy :)

when I patched the dyld, I got these messages:

Code:

Patching universal binary (2 architectures)

Patching X86_64 part (processor 16777223, architecture 0)

Code signature removed succesfully (64bit)Patch report (1): 0 instructions patched, 208 bad instructions, patches bypassed: NO

Patching I386 part (processor 7, architecture 1)

Code signature removed succesfully (32bit)Patch report (2): 1 instructions patched, 0 bad instructions, patches bypassed: NO

but when I pacthed the libSystem.B.dylib,
I got the "No patches found, not generating output file" message ???

Code:

Patching universal binary (2 architectures)

Patching X86_64 part (processor 16777223, architecture 0)

Code signature removed succesfully (64bit)Patch report (1): 0 instructions patched, 30 bad instructions, patches bypassed: NO

Patching I386 part (processor 7, architecture 1)

Code signature removed succesfully (32bit)Patch report (2): 0 instructions patched, 0 bad instructions, patches bypassed: NO

No patches found, not generating output file

are those reports correct for you ?

btw, is there any difference between the amd_insn_patcher and the amd_insn_patcher_ext ?

In the old release the amd_insn_patcher_ext patches up some SSE3 instructions to their SSE2 variants.

I think this release will do the same.

andyvand

11-10-2011 12:18 PM

Quote:

Originally Posted by akimoa (Post 55961)

@m28ew

Try using patched IOUSB from AnV (SL) might change a thing

and Andy cant we set a debug somewhere ? besides the scheduler aruments ?

There is a debug= boot argument (not sure how to use it though).
The other option would be compiling a debug kernel (with make ARCH_CONFIGS="I386 X86_64" KERNEL_CONFIGS=DEBUG install)

akimoa

11-10-2011 12:36 PM

Code:

AkimoAs-MacBook-Pro:xnu-1699.22.73 akimoa$ make ARCH_CONFIGS="I386 X86_64" KERNEL_CONFIGS=DEBUG"install"

/Users/akimoa/xnu-1699.22.73/makedefs/MakeInc.def:164: *** Unsupported KERNEL_CONFIG DEBUGinstall.  Stop.

AkimoAs-MacBook-Pro:xnu-1699.22.73 akimoa$ make ARCH_CONFIGS="I386 X86_64" KERNEL_CONFIGS=DEBUG"RELEASE"

/Users/akimoa/xnu-1699.22.73/makedefs/MakeInc.def:164: *** Unsupported KERNEL_CONFIG DEBUGRELEASE.  Stop.

AkimoAs-MacBook-Pro:xnu-1699.22.73 akimoa$

Not quite he doesnt like that

$The-King$

11-10-2011 01:06 PM

the same problem with vmware :-d

Lion mission impossible :D

m28ew

11-10-2011 01:28 PM

After patching both dyld & libSystem.B.dylib

I got a new KP :'(

http://www.postimg.com/52000/photo-51544.jpg

andyvand

11-10-2011 01:30 PM

Quote:

Originally Posted by akimoa (Post 55969)

Code:

AkimoAs-MacBook-Pro:xnu-1699.22.73 akimoa$ make ARCH_CONFIGS="I386 X86_64" KERNEL_CONFIGS=DEBUG"install"

/Users/akimoa/xnu-1699.22.73/makedefs/MakeInc.def:164: *** Unsupported KERNEL_CONFIG DEBUGinstall.  Stop.

AkimoAs-MacBook-Pro:xnu-1699.22.73 akimoa$ make ARCH_CONFIGS="I386 X86_64" KERNEL_CONFIGS=DEBUG"RELEASE"

/Users/akimoa/xnu-1699.22.73/makedefs/MakeInc.def:164: *** Unsupported KERNEL_CONFIG DEBUGRELEASE.  Stop.

AkimoAs-MacBook-Pro:xnu-1699.22.73 akimoa$

Not quite he doesnt like that

Cut and paste the following:
make ARCH_CONFIGS="I386 X86_64" KERNEL_CONFIGS=DEBUG install
note the space after release... and no ""

davisin666

11-10-2011 01:33 PM

For the usb messages, did you disable the usb legacy support (In BIOS, because it seems that in OHCI, usb 1.1, is the problem)???

If doesn't work, maybe using usb kexts from 10.6.8 can do the trick

m28ew

11-10-2011 01:39 PM

Quote:

Originally Posted by davisin666 (Post 55974)

For the usb messages, did you disable the usb legacy support (In BIOS)???

If doesn't work, maybe using usb kexts from 10.6.8 can do the trick

After patching both dyld & libSystem.B.dylib
the usb error doesn't appear anymore

but I get a new KP :D

http://www.postimg.com/52000/photo-51544.jpg

akimoa

11-10-2011 03:19 PM

here are the debug ones

autopatch

http://cl.ly/28380m2u2f3h0N0a1x1K

no autopatch

http://cl.ly/3x3V2A2a1V1I2g2n133a

hope we can get anything out of these

So test Guys :)

m28ew

11-10-2011 04:17 PM

Quote:

Originally Posted by akimoa (Post 55976)

here are the debug ones

autopatch

http://cl.ly/28380m2u2f3h0N0a1x1K

no autopatch

http://cl.ly/3x3V2A2a1V1I2g2n133a

hope we can get anything out of these

So test Guys :)

still no luck :'(

with your autopatch kernel @ 32-bits

http://www.postimg.com/52000/photo-51563.jpg

with your no autopatch kernel @ 32-bits

http://www.postimg.com/52000/photo-51564.jpg

btw, is there any new boot-flag I should use to enable the debug mode ?

EDIT:

and maybe you should split this thread and create a new one: Legacy Kernel 10.7.x / Darwin 11.x.0 ( Still in Progress... )

Just My Two Cents :)

$The-King$

11-10-2011 04:30 PM

2 Attachment(s)

belong to snow leopard

@ andyvand
since snow cant boot mac on hardware raid

and i got this panic when using force64

http://osx86.co/attachment.php?attac...1&d=1320938470

and here with out using force64

http://osx86.co/attachment.php?attac...1&d=1320938470

this annoying me as i must change data mode to a chi every time i enter mac :(

anther thing can any one help me to learn compiling kernel (tools&steps)

thanks :)

davisin666

11-10-2011 07:46 PM

not_interrupts_enabled

It seems like there is a heavy mess with interrupts and threads

wastez

11-10-2011 08:01 PM

Quote:

Originally Posted by $The-King$ (Post 55978)

Compiling is no problem.
Just use the command make or in case of anv´s kernel the script which he creates to do this. ;)

$The-King$

11-10-2011 09:24 PM

Quote:

Originally Posted by wastez (Post 55981)

Compiling is no problem.
Just use the command make or in case of anv´s kernel the script which he creates to do this. ;)

need i some tools for compiling except xcode ?

wastez

11-11-2011 07:16 AM

Quote:

need i some tools for compiling except xcode ?

No you need only Xcode and the kernel tools.
But don´t forget to patch the libSystemB_debug.dylib and libSystemB_profile.dylib after installing it.