View Full Version : Legacy kernel 10.4.0 updated
andyvand
09-14-2010, 10:23 PM
I finished the modifications to the kernel.
It now builds for both 32/64bit easily.
The patch is also here, as well as the binaries
Update: I fixed the patcher and updated it to be able to recognize any Intel instruction.
Since we don't have an auto-patching dyld yet, it's only for application binaries + dyld qua patching but it's definitely good progress.
I also created an auto-patching cpuid + sysenter tool which can operate on any Mach-O file, even directly on universal binaries.
There is also an extended tool included for patching certain sse3 instructions along with cpuid + sysenter.
Update (V4): I worked on the patcher so that it can decrypt and patch encrypted binaries (works, shows failures untill DSMOS has arrived though)
Kernel patcher (V3): Updated code to solve minor issues and not to generate output files if there is no patching done
Kernel patcher (V4): Updated code to solve no output file generated on UB if x86_64 part had no patches needed but i386 part did
Update (V5): Fixed CoreSymbolication support and fixed backtraces
Patch (V5):
http://rapidshare.com/files/423081298/legacy_kernel-1504.7.4_updated5.diff.zip
Kernel Binaries (V5):
http://rapidshare.com/files/423083429/legacy_kernel_10.4.0_rel5.zip
AMD Instruction Patcher (V4):
http://rapidshare.com/files/421925937/amd_insn_patcher_V4.zip
eMatoS
09-14-2010, 11:43 PM
Testing it now, I had to boot with arch=i386 and patcher_opts=2 like before otherwise kp or instant reboot. Seems to work fine for now, I've seen a new message in the console, I guess it's related to the kernel or sysenter patch but I don't know for real.
com.stclairsoft.DefaultFolderX.helper sysctlbyname_with_pid: sysctl failed:Invalid argument
com.stclairsoft.DefaultFolderX.helper proc_native: sysctlbyname_with_pid failed:Invalid argument
Overall the system feels snappier (advanced tech term hehe) but perhaps it's a placebo effect.
Thank you again Andy for your work
$The-King$
09-15-2010, 05:26 AM
thanks this work my quicktime but it crashes with mp4 and safari when open vedio on it with quick time
thanks andy for all your work
jackokk
09-15-2010, 05:30 AM
Thank you very much for your hard work.
davisin666
09-15-2010, 10:16 PM
Gracias amigo, no me iba a dar la paja de parchar xDDD
Like for eMatoS:
without arch=i386 and patcher_opts=2: reboot
with arch=i386 but without patcher_opts: kernel panic (see attachment but it's similar to those I've reported before)
with both it boots and somewhat works but still not really usable.
Unlike eMatoS I still have the same problems I wrote about here:
http://osx86.net/f36/qoopz-anv-xnu-10-4-0-rel4-t6491/post50208.html
(graphics problems, and missing dates and times) now with a 10.6.3 install on an AMD Athlon64 3200+
Imkantus
09-16-2010, 01:03 AM
sysenter patch but I don't know for real.
Does SYSENTER patch work via Kernel this time? (going to try later by myself...)
andyvand
09-16-2010, 09:35 AM
thanks this work my quicktime but it crashes with mp4 and safari when open vedio on it with quick time
thanks andy for all your work
try running them in 32 bit.
it is ssse3 related
andyvand
09-17-2010, 02:20 PM
Try the latest kernel on, lots of updates and it works now like a charm ;)
andyvand
09-17-2010, 02:21 PM
Does SYSENTER patch work via Kernel this time? (going to try later by myself...)
It does now for dyld.
libSystem.B(_debug/_profile).dylib won't get patched as I haven't been able to build a stable custom dyld yet.
There is however now a tool for AMD which can do the hard work for you...
RayFlower
09-17-2010, 03:28 PM
Thanks for yet another kernel andy!
I still have the problems i had before with steam and abobe cs3 pse and cs5 and some au plugins though.
Just out of curiosity, does this make any sense to you?:
/Users/my_user/Library/Application\ Support/Steam/SteamApps/
[email protected]/half-life\ 2/hl2_osx ; exit;
dyld: Library not loaded: @rpath/libtier0.dylib
Referenced from: /Users/my_user/Library/Application Support/Steam/SteamApps/
[email protected]/half-life 2/hl2_osx
Reason: image not found
Trace/BPT trap
logout
$The-King$
09-18-2010, 03:57 AM
thanks andy for yours hard work
after tests with new clean install
i must use sysenter patcher if not many programs crashes
i put amd instruction patcher on usr/bin but i can't see any benefit for it
no more patcher_opts=2
on 32bit mode every things work good except sound volume dont change and and quicktime crashes when use subtitles
on 64bit mode sound volume change work but qiucktime totally not work
thanks andy keep your good working
Imkantus
09-18-2010, 01:11 PM
i put amd instruction patcher on usr/bin but i can't see any benefit for it
The patcher is made to patch out bad instructions from single binary files - not made for patching complete folders...
For example:
sudo -s
amd_insn_patcher /usr/lib/libSystem.B.dylib /usr/lib/libSystem.B.dylib.patched
cp /usr/lib/libSystem.B.dylib.patched /usr/lib/libSystem.B.dylib
Patcher is working great and Kernel is able to boot without patcher_opts=2 now, thanks Andy. ;) I am going to do further testing once I am back from work this evening. :cool:
$The-King$
09-18-2010, 02:19 PM
The patcher is made to patch out bad instructions from single binary files - not made for patching complete folders...
For example:
sudo -s
amd_insn_patcher /usr/lib/libSystem.B.dylib /usr/lib/libSystem.B.dylib.patched
cp /usr/lib/libSystem.B.dylib.patched /usr/lib/libSystem.B.dylib
Patcher is working great and Kernel is able to boot without patcher_opts=2 now, thanks Andy. ;) I am going to do further testing once I am back from work this evening. :cool:
which files is needed to be patched with this tool
and is there need for sysenter patcher ?
thanks
diskeeper
09-18-2010, 02:38 PM
Hey Andy! Thanks for your work!
Installed your updated kernel and VmWare Fusion stopped working. With the 10.4.0 rel 4 kernel everything good.
http://img819.imageshack.us/img819/9760/kpernyfot20100918153445.png
Sep 18 15:34:37.735: vmx| Log for VMware Fusion pid=287 version=3.1.1 build=build-282344 option=Release
Sep 18 15:34:37.735: vmx| The process is 64-bit.
Sep 18 15:34:37.735: vmx| Host codepage=UTF-8 encoding=UTF-8
Sep 18 15:34:37.735: vmx| Hostname=diskeeper-Mac-Pro.local
Sep 18 15:34:37.735: vmx| IP=::1 (lo0)
Sep 18 15:34:37.735: vmx| IP=fe80::1 (lo0)
Sep 18 15:34:37.735: vmx| IP=127.0.0.1 (lo0)
Sep 18 15:34:37.735: vmx| IP=fe80::21e:8cff:fec9:3a4d (en0)
Sep 18 15:34:37.735: vmx| IP=192.168.2.101 (en0)
Sep 18 15:34:37.735: vmx| IP=192.168.211.1 (vmnet1)
Sep 18 15:34:37.735: vmx| IP=192.168.82.1 (vmnet8)
Sep 18 15:34:37.735: vmx| Command line: "/Library/Application Support/VMware Fusion/vmware-vmx" "-E" "en" "-D" "w+gVbXFDnoYDPQAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=" "-s" "vmx.stdio.keep=TRUE" "-#" "product=1;name=VMware Fusion;version=3.1.1;buildnumber=282344;licensenam e=VMware Fusion for Mac OS;licenseversion=3.0 build-282344;" "-@" "pipe=/var/folders/Bb/Bb5kr8hCFQq0qHcq9UoweU+++TI/-Tmp-//vmware-diskeeper/vmx23433769ca3d0533;readyEvent=31" "/Volumes/OSX/XP.vmwarevm/XP.vmx"
Sep 18 15:34:37.735: vmx| Msg_SetLocale: HostLocale=UTF-8 UserLocale=en
Sep 18 15:34:37.735: vmx| Msg_Reset:
Sep 18 15:34:37.735: vmx| [msg.dictionary.load.openFailed] Cannot open file "/Library/Application Support/VMware Fusion/messages/en/vmware.vmsg": No such file or directory.
Sep 18 15:34:37.735: vmx| ----------------------------------------
Sep 18 15:34:37.755: vmx| Ready event: 31
Sep 18 15:34:37.785: vmx| UI Connecting to pipe '/var/folders/Bb/Bb5kr8hCFQq0qHcq9UoweU+++TI/-Tmp-//vmware-diskeeper/vmx23433769ca3d0533' with user '(null)'
Sep 18 15:34:37.785: vmx| VMXVmdb: Local connection timeout: 60000 ms.
Sep 18 15:34:37.786: vmx| UUID: Unable to open /dev/mem: No such file or directory
Sep 18 15:34:37.787: vmx| /Volumes/OSX/XP.vmwarevm/XP.vmx: Setup symlink /var/run/vmware/e155f65b6d40c0055927d152657b9538 -> /var/run/vmware/diskeeper_501/1284816877755659_287
Sep 18 15:34:37.787: vmx| Vix: [140735088876576 mainDispatch.c:374]: VMAutomation: Initializing VMAutomation.
Sep 18 15:34:37.788: vmx| Vix: [140735088876576 mainDispatch.c:397]: VMAutomation: Detected the VM is not managed
Sep 18 15:34:37.788: vmx| Vix: [140735088876576 mainDispatch.c:658]: VMAutomationOpenListenerSocket() listening
Sep 18 15:34:37.788: vmx| Sig_Init already initialized
Sep 18 15:34:37.790: vmx| Vix: [140735088876576 mainDispatch.c:3661]: VMAutomation_ReportPowerOpFinished: statevar=0, newAppState=1870, success=1
Sep 18 15:34:37.790: vmx| Transitioned vmx/execState/val to poweredOff
Sep 18 15:34:37.790: vmx| Vix: [140735088876576 mainDispatch.c:3661]: VMAutomation_ReportPowerOpFinished: statevar=1, newAppState=1873, success=1
Sep 18 15:34:37.790: vmx| Vix: [140735088876576 mainDispatch.c:3661]: VMAutomation_ReportPowerOpFinished: statevar=2, newAppState=1877, success=1
Sep 18 15:34:37.790: vmx| Vix: [140735088876576 mainDispatch.c:3661]: VMAutomation_ReportPowerOpFinished: statevar=3, newAppState=1881, success=1
Sep 18 15:34:37.790: vmx| Panic: dropping lock (was bug 49968)
Sep 18 15:34:37.790: vmx| NOT_IMPLEMENTED bora/vmx/main/vmmonPosix.c:312
Sep 18 15:34:37.791: vmx| Core dump limit is 0 KB.
Sep 18 15:34:37.920: vmx| Child process 288 failed to dump core (status 0x6).
Sep 18 15:34:37.921: vmx| Backtrace:
Sep 18 15:34:37.923: vmx| Backtrace[0] 00007fff5fbfef60 rip=00000001003fdb05 rbx=0000000000000000 rbp=00007fff5fbfef90 r12=00000001003fd560 r13=00007fff5fbfef6c r14=00007fff5fbff500 r15=00007fff5fbff6e0
Sep 18 15:34:37.923: vmx| Backtrace[1] 00007fff5fbfefa0 rip=0000000100019df0 rbx=0000000000000000 rbp=00007fff5fbff480 r12=00007fff5fbfefa0 r13=00007fff5fbff4c0 r14=00007fff5fbff500 r15=00007fff5fbff6e0
Sep 18 15:34:37.923: vmx| Backtrace[2] 00007fff5fbff490 rip=000000010003e011 rbx=0000000000000001 rbp=00007fff5fbff490 r12=00007fff5fbff4a0 r13=00007fff5fbff4c0 r14=00007fff5fbff500 r15=00007fff5fbff6e0
Sep 18 15:34:37.923: vmx| Backtrace[3] 00007fff5fbff4a0 rip=000000010003e969 rbx=0000000000000001 rbp=00007fff5fbff4f0 r12=00007fff5fbff4a0 r13=00007fff5fbff4c0 r14=00007fff5fbff500 r15=00007fff5fbff6e0
Sep 18 15:34:37.923: vmx| Backtrace[4] 00007fff5fbff500 rip=000000010003f0ee rbx=00007fff5fbff500 rbp=00007fff5fbff680 r12=000000000000000a r13=0000000000000004 r14=0000000000000000 r15=00007fff5fbff6e0
Sep 18 15:34:37.923: vmx| Backtrace[5] 00007fff5fbff690 rip=00000001000039bb rbx=00000001006931d0 rbp=00007fff5fbff6b0 r12=00000001006931c0 r13=0000000100693020 r14=00007fff5fbff6e0 r15=00007fff5fbff6e0
Sep 18 15:34:37.923: vmx| Backtrace[6] 00007fff5fbff6c0 rip=0000000100019b39 rbx=0000000113801000 rbp=00007fff5fbff6c0 r12=0000000000000001 r13=0000000000000003 r14=00007fff5fbff6e0 r15=00007fff5fbff6e0
Sep 18 15:34:37.923: vmx| Backtrace[7] 00007fff5fbff6d0 rip=0000000100003079 rbx=0000000113801000 rbp=00007fff5fbff750 r12=0000000000000001 r13=0000000000000003 r14=00007fff5fbff6e0 r15=00007fff5fbff6e0
Sep 18 15:34:37.923: vmx| Backtrace[8] 00007fff5fbff760 rip=0000000100003857 rbx=0000000000000003 rbp=00007fff5fbff890 r12=0000000000000003 r13=00007fff5fbff770 r14=000000000000000c r15=0000000000000000
Sep 18 15:34:37.924: vmx| SymBacktrace[0] 00007fff5fbfef60 rip=00000001003fdb05 in function (null) in object /Library/Application Support/VMware Fusion/vmware-vmx loaded at 0000000100000000
Sep 18 15:34:37.924: vmx| SymBacktrace[1] 00007fff5fbfefa0 rip=0000000100019df0 in function (null) in object /Library/Application Support/VMware Fusion/vmware-vmx loaded at 0000000100000000
Sep 18 15:34:37.924: vmx| SymBacktrace[2] 00007fff5fbff490 rip=000000010003e011 in function (null) in object /Library/Application Support/VMware Fusion/vmware-vmx loaded at 0000000100000000
Sep 18 15:34:37.924: vmx| SymBacktrace[3] 00007fff5fbff4a0 rip=000000010003e969 in function (null) in object /Library/Application Support/VMware Fusion/vmware-vmx loaded at 0000000100000000
Sep 18 15:34:37.924: vmx| SymBacktrace[4] 00007fff5fbff500 rip=000000010003f0ee in function (null) in object /Library/Application Support/VMware Fusion/vmware-vmx loaded at 0000000100000000
Sep 18 15:34:37.924: vmx| SymBacktrace[5] 00007fff5fbff690 rip=00000001000039bb in function (null) in object /Library/Application Support/VMware Fusion/vmware-vmx loaded at 0000000100000000
Sep 18 15:34:37.924: vmx| SymBacktrace[6] 00007fff5fbff6c0 rip=0000000100019b39 in function (null) in object /Library/Application Support/VMware Fusion/vmware-vmx loaded at 0000000100000000
Sep 18 15:34:37.924: vmx| SymBacktrace[7] 00007fff5fbff6d0 rip=0000000100003079 in function (null) in object /Library/Application Support/VMware Fusion/vmware-vmx loaded at 0000000100000000
Sep 18 15:34:37.924: vmx| SymBacktrace[8] 00007fff5fbff760 rip=0000000100003857 in function (null) in object /Library/Application Support/VMware Fusion/vmware-vmx loaded at 0000000100000000
Sep 18 15:34:37.925: vmx| Msg_Post: Error
Sep 18 15:34:37.925: vmx| [msg.log.error.unrecoverable] VMware Fusion unrecoverable error: (vmx)
Sep 18 15:34:37.925: vmx| NOT_IMPLEMENTED bora/vmx/main/vmmonPosix.c:312
Sep 18 15:34:37.925: vmx| [msg.panic.haveLog] A log file is available in "/var/folders/Bb/Bb5kr8hCFQq0qHcq9UoweU+++TI/-Tmp-//vmware-diskeeper/vmware-diskeeper-287.log". [msg.panic.requestSupport.withLog] Please request support and include the contents of the log file. [msg.panic.requestSupport.vmSupport.apple.menu]
Sep 18 15:34:37.925: vmx| To collect data to submit to VMware support, choose "Collect Support Information" from the Help menu.
Sep 18 15:34:37.925: vmx| You can also run "/Library/Application Support/VMware Fusion/vm-support.tool" directly.
Sep 18 15:34:37.925: vmx| [msg.panic.response] We will respond on the basis of your support entitlement.
Sep 18 15:34:37.925: vmx| ----------------------------------------
opiniatre
09-18-2010, 06:22 PM
Wow it does feel snappier compare to legacy 10.4.0 kernel.
Vmware not working for me too parallel is.
Great work ! thanks
Imkantus
09-18-2010, 11:46 PM
which files is needed to be patched with this tool
and is there need for sysenter patcher ?
thanks
Well the tool patches out sysenter traps as well as it is able to patch cpuid checks into interrupt 0xfb, like we already have been doing using good old maxxus patcher.
I would recommend you to use the kernel in order with -force64 bootflag, as the 64-Bit code of OS X doesn't contain any cpuid trouble, what is making our lifes much easier and working great thanks to Andys hard work. :cool:
On my testing installation I have actually patched only these files and no issues (mp4 Playback and DivX De- & Encoder Plugins in QTX, iTunes etc all working just great):
/usr/lib/libSystem.B.dylib
/System/Library/Frameworks/CoreServices.framework/Versions/A/Frameworks/CarbonCore.framework/Versions/A/CarbonCore
/System/Library/QuickTime/QuickTimeComponents.component/Contents/MacOS/QuickTimeComponents
/System/Library/QuickTime/AppleProResDecoder.component/Contents/MacOS/AppleProResDecoder
If Developer Tools are installed libSystem.B_debug.dylib & libSystem.B_profile.dylib need to be patched as well.
The two Quiktime folder ones might not be required, but I patched them as I am still having QT7 in use - I haven't tried without patching those, yet. (going to try that later)
OK. Those two files have to be patched furthermore QTX needs to be executed in 32-Bit mode for mp4 playback to work - I haven't seen that the coresponding checkbox was still set true.
It does now for dyld.
libSystem.B(_debug/_profile).dylib won't get patched as I haven't been able to build a stable custom dyld yet.
There is however now a tool for AMD which can do the hard work for you...Anyway this is the best Snow Leopard XNU release for the AMD folks out there by now.
You have done a realy great job to the point and I hope you'll keep it going. ;)
$The-King$
09-19-2010, 01:18 AM
when patching with amd ins patcher
it says bus error
RayFlower
09-19-2010, 02:52 AM
when patching with amd ins patcher
it says bus error
Same here:S
fumoboy007
09-19-2010, 04:53 AM
I tried using arch=x86_64 and -force64 together but it just won't boot into 64-bit mode, it always goes into 32-bit mode.
eMatoS
09-19-2010, 06:18 AM
Andy, why is it that with the latest kernel I can't get chameleon to take the arch=x86_64 command? If I boot with Wait=Yes arch=x86_64 -force64 when Chameleon stops before the kernel loads it always says Loading i386 System or something similar, not X86_64. Maybe it's just cosmetic? Don't know. Perhaps you forced the loader to boot with the i386 command...
@Fumoboy007, don't pay attention to what System Profiler (and many other apps) says, you should check in activity monitor for 64bit apps.
Anyway, great work again Andy, the system's been very stable, it works better than in 32bit though I have to force some apps to run in 32bit but that's no big deal. Thank you...
fumoboy007
09-19-2010, 07:01 AM
@Fumoboy007, don't pay attention to what System Profiler (and many other apps) says, you should check in activity monitor for 64bit apps.
But it seems to be loading 32-bit kexts though…so…iunno. In the kernel logs it says "64 bit mode enabled" but it also says "Darwin Kernel Version 10.4.0: vr 17 sep 2010 14:40:38 CEST; annappirtrvh v4 :xnu-1504.7.4-legacy/BUILD/obj/RELEASE_I386"…i386?
I'm using the binary that andy uploaded…maybe it wasn't compiled for 64-bit? hehe..just a thought.
fumoboy007
09-19-2010, 07:06 AM
maybe it wasn't compiled for 64-bit?
WOAH…im actually right…i just did a "lipo -info" on that file and it's only i386!
fumoboy007
09-19-2010, 07:13 AM
WOAH…im actually right…i just did a "lipo -info" on that file and it's only i386!
…and since I don't know how to apply a patch…could someone upload the hybrid version of the kernel please =).
andyvand
09-19-2010, 03:08 PM
when patching with amd ins patcher
it says bus error
With v after the 2 binaries it works but I've made a V2.
This release also strips code signatures from the binaries it patches (even on universal binaries).
andyvand
09-19-2010, 03:10 PM
I tried using arch=x86_64 and -force64 together but it just won't boot into 64-bit mode, it always goes into 32-bit mode.
I'll make (compile) a 32+64bit kernel soon ;)
$The-King$
09-19-2010, 08:16 PM
thanks andy
hope next one will be more featured and x86-64 compile work because i love to work with x86-64 kernel
Imkantus
09-19-2010, 10:42 PM
WOAH…im actually right…i just did a "lipo -info" on that file and it's only i386!You should read more precisely, no one here has ever said that the released binaries would be universal. Andy only wrote:
It now builds for both 32/64bit easily.And as I tried myself, I can confirm that it does. Just that the Kernel in BUILD/obj/RELEASE_X86_64 doesn't work nor it's possible to merge both builds into one universal bin file. ;)
But as eMatoS wrote using 32-Bit Kernel does not mean that 64-Bit support doesn't work. Good explanation is avaible at http://www.macworld.com/article/142379/2009/08/snow_leopard_64_bit.html
What would be the advantage you promise yourself from running 64-Bit Kernel? :) Even real Macs don't use this by default...
PS: You might want to use http://osx86.net/images/infinitemacpro/buttons/edit.png next time...
@Andy
Do you have any idea how I could fix IPC for Quiktime to load 32-Bit Codecs & Safari to work with Flash Player? ;)
eMatoS
09-19-2010, 10:54 PM
Imkantus, what is the problem with Flash Player and Safari? Have you tried the beta build? It runs in 64bit now. You can download it from the Adobe Labs.
Imkantus
09-19-2010, 11:27 PM
Thx, just installing latest flash player plugin (no preview release) did the trick for Safari... sometimes solutions are even closer as expected. :D
fumoboy007
09-19-2010, 11:51 PM
What would be the advantage you promise yourself from running 64-Bit Kernel? :) Even real Macs don't use this by default...
iunno…are my EXC_I386_DIV crashes related?
RayFlower
09-20-2010, 12:49 AM
Got a geekbench error with this kernel:
http://img696.imageshack.us/img696/7541/screenshot20100920at121.png
$The-King$
09-21-2010, 03:08 PM
the quick time crashes with latest kernel and patcher
hope it is solved in new kernel and arch x86-64 work also
thanks andy for your time spending it on programing this kernel
$The-King$
09-24-2010, 06:45 AM
anther problem with this kernel is that vmware and parallel software cant be setuped it says it isnt intel cpu
but with anv 10.4 rel4 works fine with this software
thanks andy
andyvand
09-24-2010, 02:57 PM
There is another update to the kernel.
This time it can auto patch encrypted binaries (decrypting them to the new dsmos_page_transform, I figured out how it works).
Ignore the failures at beginning boot, they are because it waits for "DSMOS has arrived".
It will continue to load the binaries untill they are succesfully loaded :)
All that is left now is getting a stable custom dyld built (for library + framework auto patching) and then we will have everything where we need it ;)
I could however use some help on this because the issue (I think) is that the coresymbolication functions aren't available.
$The-King$
09-24-2010, 03:19 PM
thanks andy you are very good i love your work :)
will test it ;)
with out you what will we do :D
$The-King$
09-24-2010, 03:56 PM
after first test
x86-64
not work dont know what is problem with x86-64 bulid
fumoboy007
09-24-2010, 04:44 PM
Great work so far, Andy! There is one more issue which you should add to your list and that is 64-bit. When I use the kernel flag -force64, it just causes my computer to restart.
gedna
09-24-2010, 05:20 PM
new kernel work for me and with -force64 now, can some one tell me how to apply those patches? thanks
RayFlower
09-24-2010, 05:43 PM
The kernel patches?
You wont need them unless you plan to modify the source yourself, i haven't done it in osx for ages but check out this if you're interested.
http://shantonu.blogspot.com/2009/09/mac-os-x-kernel-source-xnu-has-been.html
And please read this:
http://linux.die.net/man/1/patch
There is no sense in going trough this unless you know what you're doing and actually commit some changes.
gedna
09-24-2010, 05:45 PM
what about amd_ins_patcher. ? is it necessary ?
fumoboy007
09-24-2010, 10:00 PM
Great work so far, Andy! There is one more issue which you should add to your list and that is 64-bit. When I use the kernel flag -force64, it just causes my computer to restart.
Also…the original legacy kernel allowed me to boot into 64-bit albeit without the ability to run 32-bit apps.
fumoboy007
09-25-2010, 12:17 AM
This happens sometimes:
Process: Safari [885]
Path: /Applications/Safari.app/Contents/MacOS/Safari
Identifier: com.apple.Safari
Version: 5.0.2 (6533.18.5)
Build Info: WebBrowser-75331805~1
Code Type: X86 (Native)
Parent Process: launchd [104]
Date/Time: 2010-09-24 19:15:01.943 -0400
OS Version: Mac OS X 10.6.4 (10F569)
Report Version: 6
Interval Since Last Report: 536347 sec
Crashes Since Last Report: 88
Per-App Interval Since Last Report: 280553 sec
Per-App Crashes Since Last Report: 8
Anonymous UUID: E4732136-7A6B-4049-B557-9CD6BCAE719B
Exception Type: EXC_BAD_ACCESS (SIGSEGV)
Exception Codes: KERN_INVALID_ADDRESS at 0x00000000676966ee
davisin666
09-25-2010, 01:52 AM
With your latest kernel (v4) iTunes 10 doesn't work, FAIL
24-09-10 20:55:52 kernel CODE SIGNING: cs_invalid_page(0x8fe2b000): p=339[iTunes] clearing CS_VALID
24-09-10 20:55:53 com.apple.launchd.peruser.501[104] ([0x0-0x38038].com.apple.iTunes[339]) Exited: Killed
eMatoS
09-25-2010, 02:42 AM
Itunes 10 is working here, at first the system felt slow but maybe it was busy with something cause it runs fine now. This time I needed arch=i386 to boot, otherwise I would get an instant reboot.
Andy, I haven't seen thoes errors before DSMOS you mention in the first post (I guess you have to un an unpatched system to see them), also what does these messages mean:
24/09/10 19:22:43 kernel warning: kernel was not built with custom dyld blob
24/09/10 19:22:43 kernel warning: using standard dyld with patcher enabled
24/09/10 22:25:15 kernel MPlayer OSX Exte (map: 0x4a35520) triggered DYLD shared region unnest for map: 0x4a35520, region 0x7fff85000000->0x7fff85200000. While not abnormal for debuggers, this increases system memory footprint until the target exits.
Another thing I've noticed is that "About this Mac" says Unknown CPU since V3 of your kernel, though I'm not sure if it's your kernel or Chameleon as I've been testing their latest RC5 release.
davisin666
09-25-2010, 04:33 AM
I'll stay in V3 XD
$The-King$
09-25-2010, 05:16 AM
hi people who use updated legacy kernel is vmware fusion allowed to be setuped or say non intel cpu
http://img251.imageshack.us/img251/2204/vmware.png
thanks
anv kernel rev4 vmware work perfect but with legacy not work
fxwizrd
09-25-2010, 04:59 PM
Thank you Andy for the V4. Everything seems to work fine here.
Though I am still not able to boot with arch=x86_64 kernel flag - getting into reboot loop.
Also I had to patch Firefox as it was randomly crashing. After patching with your AMD patcher it now seems to be stable.
Also I have a question: how do I determine which files need patching? I previously used Marvin AMD utility, which scanned all files in a specified directory and created a patcher package. Unfortunately Marvin utility does not work in 64-bit mode...
Thank you,
FX
andyvand
09-25-2010, 08:34 PM
Another thing I've noticed is that "About this Mac" says Unknown CPU since V3 of your kernel, though I'm not sure if it's your kernel or Chameleon as I've been testing their latest RC5 release.
It's probably Chameleon, try my branch as it has a CPU auto-detect fix.
andyvand
09-25-2010, 08:36 PM
hi people who use updated legacy kernel is vmware fusion allowed to be setuped or say non intel cpu
http://img251.imageshack.us/img251/2204/vmware.png
thanks
anv kernel rev4 vmware work perfect but with legacy not work
Try setting the patcher to forcing for every app it encounters with:
patcher_opts=50
=FORCE+CPUID+SYSENTER
=2+16+32
andyvand
09-25-2010, 08:42 PM
dont know what is problem with x86-64 build
Yeah, neither do I.
I suspect they removed some code and that is the reason that it doesn't work.
$The-King$
09-27-2010, 04:44 AM
Yeah, neither do I.
I suspect they removed some code and that is the reason that it doesn't work.
with out patching any thing by tools
on 32bit
nearly kernel works good just few very few problems you can say fly patching work ;)
on 64bit
all media programs crashes fly patching not works good :(
if you fixed 64bit fly patching may arch x86-64 works :D
andyvand
09-28-2010, 01:28 PM
with out patching any thing by tools
on 32bit
nearly kernel works good just few very few problems you can say fly patching work ;)
on 64bit
all media programs crashes fly patching not works good :(
if you fixed 64bit fly patching may arch x86-64 works :D
AFAIK on the fly patching for x86_64 works just fine but the media programs (QuickTime in 64bit) crash because there are no Supplemental SSE3 instructions available.
The only solution is running it in 32bit mode (Get Info, check "run in 32bit")
andyvand
09-28-2010, 01:52 PM
New version of the AMD instruction patcher.
This version fixes issue of no output file generation on universal binaries when i386 part needed patching but x86_64 part didn't
snovich
09-28-2010, 05:22 PM
Everything working here for me. ;) I can use Logic 9 64 bit with 32 bit plugins, like on my macbook!!! I'm booting mac os with:
arch=i386 -force64 -maxmem=4096 busratio=20
I'm using andy kernel v4 and AMD sysenter trap patch (10.6.4)(I patched sysenter_patch with maxxuss patcher gui)
ANDY THANK YOU VERY MUCH FOR ALL YOUR HARD WORK!!!:D
I have Amd Phenom x4 965 Processor
http://dl.dropbox.com/u/4815884/Screen%20shot%202010-09-28%20at%206.05.10%20PM.png
cynargo
09-28-2010, 05:44 PM
AFAIK on the fly patching for x86_64 works just fine but the media programs (QuickTime in 64bit) crash because there are no Supplemental SSE3 instructions available.
The only solution is running it in 32bit mode (Get Info, check "run in 32bit")
That's a solution, but some apps are crashing anyway eg. iTunes while importing h.264 video. Even though it's x86 app, it crashes due to lack of SSSE3 on my CPU if the kernel is in x64 mode. It doesn't happen if i run kernel in x86 mode.
The same thing with valve's games - they crash in x64 on intro movies, even if the game's binary itself is 32-bit.
snovich
09-28-2010, 07:43 PM
I tested all application i have installed on 10.6.4 with andy kernel -force64 and all work good, like with arch=i386 only. Quicktime works in 64bit at my computer. You can view image at my first post.
$The-King$
09-28-2010, 09:02 PM
I tested all application i have installed on 10.6.4 with andy kernel -force64 and all work good, like with arch=i386 only. Quicktime works in 64bit at my computer. You can view image at my first post.
which files you patched
snovich
09-28-2010, 10:01 PM
1. boot in 32 bit mode with arch=i386, with Legacy kernel 10.4.0 updated(v4).
2. download AMD sysenter trap patch (10.6.4) (http://OSx86.co/f36/amd-sysenter-trap-patch-10-6-4-t6498/)
3. I used maxxuss pather gui and patched sysenter_patch.txt
(Patcher patched just 2 files of 4)
4. restart, boot with -force64 & cross your fingers:)
kabuu
09-29-2010, 12:06 PM
I gave boot With Verbose mode, appears these errors
"Decryption hasn't succeeded yet (probably will retry)
patch_exec_obj_segment failed
map_segment failed"
This is normal ?
I used Legacy Kernel v4 and instruction patcher
booting with "arch=i386 -force64"
Thank you Andy for his Hard work
Sorry for my English
snovich
09-29-2010, 02:37 PM
I think that's normal with updated kernel.
gedna
09-29-2010, 02:53 PM
how to make parralers work, its says its not an Intel cpu, can any one help?
fumoboy007
09-29-2010, 03:47 PM
how to make parralers work, its says its not an Intel cpu, can any one help?
1. Turn on virtualization support in your BIOS.
2. Make the Parallels dmg read/write and mount it.
3. Open Terminal and navigate to the mounted dmg.
4. cd Install.mpkg/Contents
5. open -a TextEdit distribution.dist
6. Find "var cpu_check".
7. Set value to 0.
8. Save and close.
9. Run Parallels Desktop 6 installer.
10. If you're using a serial that you got from online, do not register or we're all effed.
eMatoS
09-29-2010, 09:49 PM
Hey Andy, I've gone back to AnVal beta2 (chameleon) and my CPU is still not recognized in "About this Mac". Will try an older version but I'm sure it worked fine with that one.
EDIT: already solved, it seems that there was a problem with my rig being identified as an iMac 11,1 (smbios.plist) cause setting it to MacMini2,1 got the cpu back to the usual Core Solo (in About this Mac) and now even FakeSMC with the ITEController and Nvidia plugins are reading all the temp and fans correctly. Don't know why but it used to work fine before, I haven't touched the smbios.plist in quite some time and only recently I've noticed the cpu thing. Anyways, it's all right now, sorry for the trouble.
kabuu
09-30-2010, 11:31 PM
I'm having problems with a few applications as "iTunes, Firefox, Cod4, Transmission".
sometimes latch the System ..
appears on the console Failed to create CSSymbolicatorRef for Call of Duty 4 Multiplayer
some solution ?
sorry for my english
jackokk
10-01-2010, 06:13 AM
For your information,I get the error message from CleanMyMac program before access to it:
" We apologize, but CleanMyMac couldn't define your Mac's architecture. Probably, you do not have /usr/bin/arch program installed, and this is why Universal Binary plugin is unable to find anything. "
Thank you very much for your hard work, Andy!!!
davisin666
10-01-2010, 09:45 PM
Thanks Andy for the kernel, now Front Row works!
I have used patcher_opts=50 arch=i386 and -force64
Great job!!!
fxwizrd
10-02-2010, 05:01 PM
Hello all,
I am getting spontaneous crashes time to time, with applications unexpectedly quitting one by one eventually resulting in total freeze of the system. It happens some time (usually hours) after system has been started. Today it crashed when the following applications where running:
Transmission
Firefox
Mail
These messages appear in logs:
10/2/10 6:28:03 PM mDNSResponder[20] mDNSPlatformRawTime went backwards by 2 ticks; setting correction factor to 3332901547
10/2/10 6:28:03 PM mDNSResponder[20] mDNSPlatformRawTime: last_mach_absolute_time 27C0809807F57AED
10/2/10 6:28:03 PM mDNSResponder[20] mDNSPlatformRawTime: this_mach_absolute_time 27A6131300000000
10/2/10 6:28:03 PM mDNSResponder[20] mDNSPlatformRawTime went backwards by 1 ticks; setting correction factor to 3332901548
10/2/10 6:28:19 PM com.apple.launchd.peruser.501[2139] ([0x0-0x2e02e].org.mozilla.firefox[3034]) Exited with exit code: 3
10/2/10 6:28:29 PM ReportCrash[3054] Failed to create CSSymbolicatorRef for fontd
10/2/10 6:28:29 PM com.apple.launchd.peruser.501[2139] (com.apple.fontd[2149]) Job appears to have crashed: Floating point exception
10/2/10 6:28:29 PM ReportCrash[3054] Saved crash report for fontd[2149] version ??? (???) to /Users/mrx/Library/Logs/DiagnosticReports/fontd_2010-10-02-182829_MrXs-Mac-Pro.crash
10/2/10 6:29:23 PM ReportCrash[3054] Failed to create CSSymbolicatorRef for Transmission
10/2/10 6:29:23 PM ReportCrash[3054] Saved crash report for Transmission[2916] version 2.04 (11151) to /Users/mrx/Library/Logs/DiagnosticReports/Transmission_2010-10-02-182923_MrXs-Mac-Pro.crash
10/2/10 6:29:23 PM com.apple.launchd.peruser.501[2139] ([0x0-0x20020].org.m0k.transmission[2916]) Job appears to have crashed: Floating point exception
Has anyone experienced similar behavior? Any suggestions what could be done to fix this?
My system:
MacOS SL 10.6.4 (Retail modified), kernel AnV 10.4.0v4 Chameleon V2.0RC4, PC_EFI v10.6
Mobo: Asus A8N-SLI
CPU: AMD Athlon64 X2 3800+
Video: Geforce 7900GT PCIe 256MB (vanilla kexts + GraphicsEnabler boot option)
LAN: onboard (nForceLAN 0.64.5 kext)
Thanks in advance!
FX
fumoboy007
10-02-2010, 07:29 PM
I get this KP sometimes:
http://img337.imageshack.us/img337/8021/pa024080.jpg
Edit: Nevermind..it was because I was using an older AppleNForceATA.
scmsystm
10-02-2010, 11:54 PM
Hi, what's requirement to able boot with patcher_opts=50?
I'm only able boot with separate patcher_opts,
2, 16, 32, 64, 128
If I boot with patcher_opts=18 its got KP, also with patcher_opts=34
Instant reboot if using patcher_opts=48 and patcher_opts=50
Some universal applications with non intel 64 supported sometimes got random crash after running within 2 hours.
Transmissions, Speed Download, Firefox.
btw I'm using old machine, BIOSTAR NF61S-M2 TE, AMD Athlon 64 4400+ on 2.7 GHz :'(
andyvand
10-04-2010, 04:27 PM
I gave boot With Verbose mode, appears these errors
This is normal ?
I used Legacy Kernel v4 and instruction patcher
booting with "arch=i386 -force64"
Thank you Andy for his Hard work
Sorry for my English
Yes, look more closely, it gives these from "Waiting for DSMOS"
untill "DSMOS has arrived" message, it's because it waits for the decryptor
andyvand
10-04-2010, 04:37 PM
Again an update.
This time I fixed CoreSymbolication support and fixing that issue also fixed backtracing issues.
Debugging is thus available :)
Normally no more CS errors thus ;)
I decided to abandon 64 bit kernel for now (untill I can figure out what's wrong with it)
snovich
10-04-2010, 08:03 PM
Rapidshare:
Too%20many%20users%20downloading%20from%20this%20s erver%20right%20now.%20Please%20try%20again%20late r...
:)
fumoboy007
10-05-2010, 01:11 AM
there are no Supplemental SSE3 instructions available.
The only solution is running it in 32bit mode (Get Info, check "run in 32bit")
You can't just make an SSSE3 emulator?
RayFlower
10-05-2010, 04:57 AM
Thanks for another update, now the backtrace works so it gives a very verbose crash report, for the sake of not flooding this post I pastebin'd it:
Steam
http://pastebin.com/dQRYMwtd
And this is the crash report i get when i try to use/validate my precious IK multimedia plugins with logic:
http://pastebin.com/7dBdF2hb
Again thanks for the update:)
andyvand
10-05-2010, 02:32 PM
Thanks for another update, now the backtrace works so it gives a very verbose crash report, for the sake of not flooding this post I pastebin'd it:
Steam
http://pastebin.com/dQRYMwtd
And this is the crash report i get when i try to use/validate my precious IK multimedia plugins with logic:
http://pastebin.com/7dBdF2hb
Again thanks for the update:)
In case of the plugins it seems something to have to do with some functions of the dyld.
In case of steam it is the uninstalling of the exception handler that goes wrong (thread 7)
Thread 7 Crashed:
0 crashhandler.dylib 0x1b4c6299 google_breakpad::ExceptionHandler::UninstallHandle r(bool) + 133
1 crashhandler.dylib 0x1b4c702f google_breakpad::ExceptionHandler::WaitForMessage( void*) + 381
2 libSystem.B.dylib 0x91c4981d _pthread_start + 345
3 libSystem.B.dylib 0x91c496a2 thread_start + 3
$The-King$
10-05-2010, 03:37 PM
link for rel5 doesnot work
eMatoS
10-14-2010, 12:15 AM
I have to run Pixelmator in 32bit for get it working, this is what I get when creating a new file in 64bit: http://pastebin.com/DYYvRMRk
@eMatoS: line 23: vPermuteChannels_ARGB8888_SSSE3, looks like it's trying to use SSSE3 which as Andy said before is not yet emulated. Maybe this is why.
<p>
For me Snow Leopard still does not work at all even in 32 bit, I get the same KP I posted already above (see post #6) even with rel5 (no more debugging info either, only hex numbers that do not make sense to me). Any idea? Andy where are you?
opiniatre
10-21-2010, 03:09 PM
On my phenom II black edition (multiplier not locked) i'm not able to change multiplier, it freezes at boot. At default multiplier it runs smoothly.
I tried changing the bus ratio but no luck. Is it due to chameleon ?
Thanks
RayFlower
10-21-2010, 08:18 PM
Just saw this when i booted now:
warning: kernel was not built with custom dyld blob
warning: using standard dyld with patcher enabled
RayFlower
11-11-2010, 08:58 PM
uhm the rapidshare link is dead:S could you please upload it again? :)
Imkantus
11-11-2010, 09:59 PM
uhm the rapidshare link is dead:S could you please upload it again? :)
It's not dead rapidshit is just unavaible for non-paying-users sometimes...
Mirror:
http://www.megaupload.com/?d=JI92IXMQ
http://www.mediafire.com/?a7u2d2wcq5i4chd
RayFlower
11-11-2010, 10:32 PM
Thanks! got stuck with v4 since i forgot to store it propperly
ManuB
11-12-2010, 12:31 PM
It's not dead rapidshit is just unavaible for non-paying-users sometimes...
Mirror:
http://www.megaupload.com/?d=JI92IXMQ
http://www.mediafire.com/?a7u2d2wcq5i4chd
__________________________________________________ _______________________
Thank you! :)