Jump to content

Mavericks kernel testing on AMD (formerly Mountain Lion kernel testing on AMD)


theconnactic
 Share

6,414 posts in this topic

Recommended Posts

Hey, mac_carol!

 

I was suggesting you use the latest non-debugging kernel from AnV (test 5) with the TSC module and the flags -enable_ssse3emu cpus=1. :)

 

The fact, i must insist, is that until we get the ssse3emu to properly work, we're stuck. The lack of ssse3 support is the main reason the processes lauched by launchd (forgive me the pun) don't run.

 

Thank you guys for the extensive tests.

Link to comment
Share on other sites

Hey, mac_carol: i was suggesting you use the latest non-debugging kernel from AnV (test 5) with the TSC module and the flags -enable_ssse3emu cpus=1. :)

 

Thank you!

 

 

testing this kernel

 

http://www.insanelym...40#entry1873038

 

 

with flags: cpus=1 -enable_ssse3emu

 

result: can't pass the "Kernel is LP64"

 

 

anv_16.gif

 

 

 

 

 

 

@ALL

 

is there anyone who can provide us any bash scripts for installing:

 

- cparm bootloader & modules,

http://www.insanelym...00#entry1873241

 

- and also AnV's patched launchd ?

http://www.insanelym...60#entry1873131

 

 

please share

 

thanks :)

Link to comment
Share on other sites

Without any mod? Thus without corecrypto... that could explain the assertion failure in itself.

Corecrypto isn't just the hashes sha1 sha256 etc...

It is also a encryption and decryption library...

I think it is also used by dsmos as it also contains aes functions (both sse4 as regular instructions...)

Link to comment
Share on other sites

Andy, i'm up to try to tweak the ssse3 emulator later tonight. I think the more people get involved with a given task, the more quickly the results show up. As i said before, i think we should focus now on get the emulator to work, since is almost certain the lack of ssse3 is preventing the user space to run. SS01 told me yesterday that you're sure the emulator isn't ready for 64-bit usage yet. Can you share your thoughts about it with me?

 

Thank you!

 

P.S.: and thank you, mac_carol, for the tests you did.

Link to comment
Share on other sites

Sorry guys that my test took a little longer.

 

Test with Phenom II X6- Gigabyte GA-990FXA-UD3 with 8 GB (only 8 GB of 16 GB were installed)

booted with cparm bootloader

 

 

boot with ML 10.8 boot flag debug 1 -enable_ssse3emu

Part1 http://www.workupload.com/file/mtXy4BK

Part2 http://www.workupload.com/file/cckrcIs

boot with ML 10.8 boot flag debug 1 without -enable_ssse3emu

Part1 http://www.workupload.com/file/ZjefNbr

Part2 http://www.workupload.com/file/RPlPoB9

Part3 http://www.workupload.com/file/IfM7s4T

 

boot with ML 10.8 boot flag debug 255 -enable_ssse3emu

Part1 http://www.workupload.com/file/sGUIzNl

Part2 http://www.workupload.com/file/UefWfV1

Part3 http://www.workupload.com/file/94URfOG

Part4 http://www.workupload.com/file/pI7qarP

Part5 http://www.workupload.com/file/AuOnpzf

 

 

boot with ML 10.8 boot flag debug 255 without -enable_ssse3emu

Part1 http://www.workupload.com/file/ooCkAvC

Part2 http://www.workupload.com/file/gxuSz0c

Part3 http://www.workupload.com/file/nAPiuqk

Part4 http://www.workupload.com/file/B1VstO7

 

 

boot with ML 10.8.3 boot flag debug 1 without -enable_ssse3emu

 

Part1http://www.workupload.com/file/DHiC0Q5

Part2 http://www.workupload.com/file/xUetmG9

 

boot with ML 10.8.3 boot flag debug 255 without -enable_ssse3emu

 

Part1 http://www.workupload.com/file/aLhEbOO

Part2 http://www.workupload.com/file/TmMAJOR

Link to comment
Share on other sites

Yeah it needs more work.

Afaik Sinetek is working on it and almost has it done but he's still sorting out how to access instructions stored above 0xfff in the set commpage entry ;) I know he'll do a good job on it.

Keep up the good work Sinetek :D

Link to comment
Share on other sites

Andy,

 

This test, however, without the driver adjustments in S / L / E

I still need to install the drivers for my hardware. it was for this test once secondary.

 

Andy I have still a naive question.

debug_kernel created after booting a file in OSX? If so where, can I find this item? the boot over several pages long run I'd like to upload the complete file.

Link to comment
Share on other sites

- and also AnV's patched launchd ?

http://www.insanelym...60#entry1873131

 

please share

thanks :)

 

Include this in his launchd folder... specifically in launchd/Release/. Name it launchdhelper.sh and chmod +x it.

 

***I WILL NOT BE HELD LIABLE IF THIS DOESN'T WORK, DELETES THE CONTENTS OF YOUR HARD DISK, FAILS TO CURE AIDS OR BEGINS WWIII.****

 

#! /bin/bash
if [ $1 = "install" ]; then
	 echo "Where is the Mac volume mounted (  leave blank for '/'   , /Volumes/<drive name> no trailing slash , etc.)?"
	 echo "Type the path and press [Enter]:"
	 read mountpoint
	 echo "Backing up originals..."
	 mv $mountpoint/bin/launchctl $mountpoint/bin/launchctl.orig
	 mv $mountpoint/bin/wait4path $mountpoint/bin/wait4path.orig
	 mv $mountpoint/sbin/launchd $mountpoint/sbin/launchd.orig
	 mv $mountpoint/sbin/SystemStarter $mountpoint/sbin/SystemStarter.orig
	 mv $mountpoint/usr/lib/system/liblaunch.dylib.orig $mountpoint/usr/lib/system/liblaunch.dylib.orig
	 mv $mountpoint/usr/libexec/launchproxy $mountpoint/usr/libexec/launchproxy.orig
	 echo "Installing new launchd..."
	 cp launchctl $mountpoint/bin/
	 cp wait4path $mountpoint/bin/
	 cp launchd $mountpoint/sbin/
	 cp SystemStarter $mountpoint/sbin/
	 cp liblaunch.dylib $mountpoint/usr/lib/system/
	 cp launchproxy $mountpoint/usr/libexec/
elif [ $1 = "remove" ]; then
	 echo "Where is the Mac volume mounted (/, /Volumes/<drive name/, etc.)?"
	 echo "Type the path and press [Enter]:"
	 read mountpoint
	 echo "Deleting launchd..."
	 rm $mountpoint/bin/launchctl
	 rm $mountpoint/bin/wait4path
	 rm $mountpoint/sbin/launchd
	 rm $mountpoint/sbin/SystemStarter
	 rm $mountpoint/usr/lib/system/liblaunch.dylib
	 rm $mountpoint/usr/libexec/launchproxy
	 echo "Restoring original launchd..."
	 mv $mountpoint/bin/launchctl.orig $mountpoint/bin/launchctl
	 mv $mountpoint/bin/wait4path.orig $mountpoint/bin/wait4path
	 mv $mountpoint/sbin/launchd.orig $mountpoint/sbin/launchd
	 mv $mountpoint/sbin/SystemStarter.orig $mountpoint/sbin/SystemStarter
	 mv $mountpoint/usr/lib/system/liblaunch.dylib.orig $mountpoint/usr/lib/system/liblaunch.dylib
	 mv $mountpoint/usr/libexec/launchproxy.orig $mountpoint/usr/libexec/launchproxy
else
echo "Usage launchdhelper install -OR- launchdhelper remove"
fi

Link to comment
Share on other sites

Just a thought: has anybody ever tried booting in single user (-s).

Wondering if that works...

Please test it...

Link to comment
Share on other sites

That script does not exactly work over here...

 

Where is the Mac volume mounted (  /   , /Volumes/<drive name>  , etc.)?
Type the path and press [Enter]:
"/Volumes/AMD Mountain Lion HD/"    
Backing up originals...
usage: mv [-f | -i | -n] [-v] source target
   mv [-f | -i | -n] [-v] source ... directory
usage: mv [-f | -i | -n] [-v] source target
   mv [-f | -i | -n] [-v] source ... directory
usage: mv [-f | -i | -n] [-v] source target
   mv [-f | -i | -n] [-v] source ... directory
usage: mv [-f | -i | -n] [-v] source target
   mv [-f | -i | -n] [-v] source ... directory
usage: mv [-f | -i | -n] [-v] source target
   mv [-f | -i | -n] [-v] source ... directory
usage: mv [-f | -i | -n] [-v] source target
   mv [-f | -i | -n] [-v] source ... directory
Installing new launchd...
usage: cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file target_file
   cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file ... target_directory
usage: cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file target_file
   cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file ... target_directory
usage: cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file target_file
   cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file ... target_directory
usage: cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file target_file
   cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file ... target_directory
usage: cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file target_file
   cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file ... target_directory
usage: cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file target_file
   cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file ... target_directory

Link to comment
Share on other sites

Just a thought: has anybody ever tried booting in single user (-s).

Wondering if that works...

Please test it...

 

I've already tested it, with the same result

Link to comment
Share on other sites

That script does not exactly work over here...

 

Where is the Mac volume mounted ( / , /Volumes/<drive name> , etc.)?
Type the path and press [Enter]:
"/Volumes/AMD Mountain Lion HD/"
Backing up originals...
usage: mv [-f | -i | -n] [-v] source target
 mv [-f | -i | -n] [-v] source ... directory
usage: mv [-f | -i | -n] [-v] source target
 mv [-f | -i | -n] [-v] source ... directory
usage: mv [-f | -i | -n] [-v] source target
 mv [-f | -i | -n] [-v] source ... directory
usage: mv [-f | -i | -n] [-v] source target
 mv [-f | -i | -n] [-v] source ... directory
usage: mv [-f | -i | -n] [-v] source target
 mv [-f | -i | -n] [-v] source ... directory
usage: mv [-f | -i | -n] [-v] source target
 mv [-f | -i | -n] [-v] source ... directory
Installing new launchd...
usage: cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file target_file
 cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file ... target_directory
usage: cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file target_file
 cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file ... target_directory
usage: cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file target_file
 cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file ... target_directory
usage: cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file target_file
 cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file ... target_directory
usage: cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file target_file
 cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file ... target_directory
usage: cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file target_file
 cp [-R [-H | -L | -P]] [-fi | -n] [-apvX] source_file ... target_directory

 

You probably have to escape the spaces in the <volume name>. Don't use the trailing slash either.

 

/Volumes/AMD\ Mountain\ Lion\ HD

Link to comment
Share on other sites

That's what I tried first; it outputs the same. I could just do it manually, but I don't think it's really going to get anywhere.

 

The script works of me in Linux. But my mountpoint is /mnt. I didn't thoroughly test other scenarios. Sure, the script could be improved. Perhaps it won't help the overall cause, but I created it for my own testing anyway.

 

Symlink your volume to a path without spaces and see if that gets you anywhere. And again, user beware.

  • Like 1
Link to comment
Share on other sites

I have now booted with the last kernel Andys. I've adapted the driver in S / L / E . I think that I need to modify the network card driver. otherwise it looks very positive on Phenom II X6.

 

here my test with last Andys Kernel

 

boot with 10.8 with boot flag -enable_ssse3emu

Part1 http://www.workupload.com/file/M4ANTGO

Part2 http://www.workupload.com/file/1ZAhv6v

 

boot with 10.8 with boot flag -s

 

Part1 http://www.workupload.com/file/F8OLV36

Part2 http://www.workupload.com/file/rIeU2MM

Part3 http://www.workupload.com/file/G5p4nsa

 

boot with 10.8.3 with boot flag -s -enable_ssse3emu

 

Part1 http://www.workupload.com/file/H40pu7O

Part2 http://www.workupload.com/file/ICh4sq0

Part3 http://www.workupload.com/file/oIaoMOK

 

What do you think?

 

I have no OSX 10.8.2 to test with Andy's kernel nopla31x18.gif

Link to comment
Share on other sites

The script works of me in Linux. But my mountpoint is /mnt. I didn't thoroughly test other scenarios. Sure, the script could be improved. Perhaps it won't help the overall cause, but I created it for my own testing anyway.

 

Symlink your volume to a path without spaces and see if that gets you anywhere. And again, user beware.

 

Before testing your script, chef, what would be the expected result if everything goes well?

 

That's what I tried first; it outputs the same. I could just do it manually, but I don't think it's really going to get anywhere.

 

I would not say it wouldn't get anywhere, but i think we'll only have a working AMD installation of Mountain Lion when the ssse3 emulator is finished. Sinetek's working on it, so i hope we get good news soon.

 

I have now booted with the last kernel Andys. I've adapted the driver in S / L / E . I think that I need to modify the network card driver. otherwise it looks very positive on Phenom II X6.

 

Thank you for the extensive tests. However, couldn't you post the pics here instead of in a file host?

Link to comment
Share on other sites

Our old friend corecrypto has come back to pay a visit. I have not been using the patched kext during this most recent testing. I've been masking the AMD SSE3 bit in VMware. We've been freezing at LP64. If I remove the mask and boot fips_mode=0, that boot arg is no longer sufficient to continue booting. Observe.

 

fm02.jpg

Link to comment
Share on other sites

 Share

×
×
  • Create New...