Separate names with a comma.
Discussion in 'macOS' started by murdercitydevil, Nov 11, 2010.
Title says it all. Was wondering if anyone had any issues.
yup! well, i am speaking in terms of my experience with a vanilla kernel installation (specs in sig). i also have ALC889a audio which is natively supported so i didnt lose my sound after the update!
for much more info visit insanelymac.com or tonymacx86.com
It depends on how many custom kexts you needed to load. I'll be running Carbon Copy Cloner prior to loading 10.6.5 onto my core i7
yep, if you are using the vanilla kernel and arent replacing Apple's kexts, but using separate functionality kexts such as fakesmc, ethernet kexts, audio kexts, IOBlockStorageInjector etc then everything should be fine.
except some audio chips will not work after the update and you will have to rollback AppleHDA.
I just updated my core i7-930 to 10.6.5 and I'm happy to announce that it went seamless and problem free. Phew, I was a little concerned but I was able to from 10.6.0 to 10.6.4 through various updates and never had an issue
When I updated my hackintosh from 10.6.3 to 10.6.4, I had SLOW graphics... like going from nearly 30 FPS in Cinebench to 4 FPS!
I reinstalled 10.6.3 and changed some plists to make my Mac think it's running 10.6.4, and was even able to install the Apple Graphics Update while still retaining the faster graphics speed from 10.6.3.
Any problems with slow graphics on this version?
I upgraded mine with no issues.
Motherboard is a Gigabyte EP45-UD3P.
Graphics, audio and network kexts
I have added some kexts for graphics (for radeon HD5750), is there any problem with this after the update?
should I use multibeast to patch the system or modify something before restart the machine?
Ep45-ud3p updated to 10.6.5 via software update, lost USB functionality, cant click on anything to see if I have sound but yah no mouse or keyboard
^ read http://tonymacx86.blogspot.com/2010/11/mac-os-x-1065-update.html
use use iBoot 2.5.1 to boot your 10.6.5 system and then use latest Multibeast to Rollback to 10.6.2 IOUSBFamily.
looks like there is problems with audio, USB and sleepenabler kexts (if you dont have a DSDT). didnt see these issues havent had an update issue since 10.6.2
If they are Apple shipped .kext that were modified the good chances are yes you will need to make the same changes in them as before if they are updated by this upgrade.
When I was playing around with the new RC5 series of the Chameleon boot loader on my spare partition I use for just this occurrence of upgrades to the OS I saw a fix usb option there perhaps if you track that down and use it at the options prompt (if booting graphically hit the TAB key to get to it otherwise the hit any key option after using the F8 during posting to get to the option loader if no delay set when booting textually) this may solve your problem.
Here is a tip for anyone really use the diskutil program select the restore option choose your main install as source as the destination you can use external USB drive, another hard drive partition then clone your install to the destination install the boot loader onto it. Now you can boot that when upgrades happen do them if there is a mess then you can boot back into your main install fix what is wrong with the testing one and never risk your working install at all..
I got lucky, I hopped on via Screen Sharing, for 20 minutes forgot my Username lol, figured it out, gained control, ran multibeast, easybeast, it restored USB functionality but as I reboot, the system hangs at "verifying DMI Pool Data...." on boot ?
Forgot to mention I have a dual boot on this with MacDrive, so I can access the mac from my windows, I just dont know what to do.
The DMI error usually indicates an incorrect setting in the BIOS/corrupted booting drive. Does the Windows on this machine still boot and have you a fix installed in OS X for the CMOS reset bug that happens with Gigabyte boards?
Upgraded mine this morning via software updates and no problems (vanilla kernel, with just Azalia audio kexts)
I think you only have to beware when the update is going to contain sleep/hibernation fixes, new graphics cards drivers/bug fixes ...
If you're using vanilla kernel or setup your kernel with another name than mach_kernel, it's pretty much always safe to hit the update button.
I don't really see what this has to do with MacRumors, though.
Cool, thanks for the replies guys. I basically have no custom kexts (I opted for a very vanilla install), so I will do a backup and then update.
Updated my EP45-DS3 through Software Update, absolutely no tinkering required.
Chameleon RC4, OpenHaltRestart, FakeSMC 2.5, Radeon HD4870, sound in DSDT.
I had to remove my SleepEnabler.kext before running the update and currently my usb doesn't work for storage devices, but keyboard and mouse work fine. I hope to have the usb fixed after next reboot, will post solution if it works.
Sleep doesn't work with the new SleepEnabler.kext that I found, but maybe another will work.
Edit: fixed the usb issue with the kexts in this post
2nd Edit: Found a working sleep enabler for my 32-bit system: Here
10.6.5 upgrade SleenEnabler.kext Ethernet cmos or bios
The upgrade went well from 10.6.4 to 10.6.5. I decided to add the SleepEnabler.kext as a finishing touch, which resulted in KP. I hadn't read to edit com.apple.Boot.plist (caBp)and under Kernel Flags, to add the string pmVersion=21 although I think pmVersion=0 may also work. I couldn't find a way to edit caBp with Chameleon. So tried mounting the Volume from the SL OS retail disk, and using cat and echo to insert that string. It was easier to reinstall. I nearly lost my cmos, I have a dual bios Gigabyte GA-X58A-UD5.
I had to power off and unplug (or turn off power switch on back of computer) before I got a Cmos screen again. I have to do the same thing, unplug computer, to get a connection to the internet, by making the onboard ethernet work. Since then I've read a guide which said to turn off power and unplug to restore ethernet. I lost connection in both Win7 and SL. I think this may be related to the cmos reset problem for Gigabyte boards. So if you need SleepEnabler.kext edit caBp first before restarting after the update. Or know how to use the Chameleon command line to enter the pmversion=21 switch, to fix the KP mismatch. Also, I had to use MultiBeast to again install kexts, after.
Since you know of the multibeast does it not have the pre-edited DSDT.aml for that board you can install that has the cmos reset bug fixed among other things like the edit for the AppleLPC to load so you having working speed step, if not you can get it at the tonymac site or on insanelymac there is a monster thread for that very board detailing the edits you need to make on your own plus a couple of already made ones to download.
Cantankerous musings re GA-X58A-UD5 DSDT.aml
I am not sure that I have read that thread. I did however read about using iasl on this page, http://www.cantankerouscoder.org/2010/05/05/dsdt-and-kext-information-for-my-hackintosh/
He refused to provide his compiled DSDT.aml for people who asked, who were using an GA-X58A-UD5 with an i7-950 cpu, the same as his setup. He said because you also had to use the right(same as his) release of bios. I think his was F8 and mine is FB. Since that bios is older, I'm not sure his mboard is Revision 2.0 like mine or Revision 1.0.
So I'm reluctant to try these pre-compiled DSTD.aml because they don't provide exact information about their basis. I have read quite a few of TonyMacs threads. And I did try a pre-compiled DSTD.aml for the GA-X58A-UD3(r?) which didn't work. I'm not ready to download Ubuntu and acquire the "diffs" or apply them as he recommends. Also there has been a lot more work done on the GA-EX58-UD? which is an earlier board, and I actually thought that is what I ordered. Also more is that this MultiBeast or maybe it was Efi*.* initial bootloader no longer has the option of selecting UD3 or UD7 and the UD7 I read also works for UD5. There is a thread you may not have read, which goes on for months and over a hundred posts of people who consistently lose their internet connectivity (the self-assigned IP address syndrome), with mostly earlier versions of my board which is new. Some of those people just like me, could not connect to the internet with either Windows or SL. My idea of turning off the power and then the power switch in the back (which I lucked into) was the only consistent solution, the other ideas all resulted in intermittent connectivity. To my mind, that means the drivers or kexts are buggy. I don't know enough about that kind of hardware to speculate. The problem I had with the dual bios was that the first one konked out. The second was started up and failed. I got this sophisticated blue screen eventually which said I had a serious error in the bios. I guess it was corrupted. Anyway it reminded me of system restore which offers 'last known good configuration' and a list of other boots which worked, and the number of times, to choose from. I think the cmos reset problem is in the same ballpark, indirectly related, but I'm not sure exactly how. I think it is really odd that I have to pull the plug to use the ethernet each time I switch between OS's. I'm new to Macs, and my Linux education is rusty, I didn't think about a kernel mismatch. I'm subscribed to insanelymac so will go look for the thread you mentioned. Thanks for your input, I think of you as one of the more erudite mac users.
He is correct there the DSDT you get is specific to the motherboard revision/BIOS contained on that motherboard sometimes it does not change between the BIOS revisions but there can be a considerable difference.
I can understand that but you can use the isal or isalme to decompile the downloaded .aml file to check what changes have been made and looking at one of them files I decompiled for a board I am thinking of upgrading too where they say the CMOS bug is fixed in it it was not they still had the incorrect length there.
You can do it running OS X with I believe it is called the getDSDT.sh, yep just checked an easy find on Google.
The fixes are pretty generic across the line of boards ie. the same fixes are being applied be it a P45 like I have right now or the X58, P55.. so what is being done will still apply to your DSDT if/when you extract it or get one already extracted you should just decompile and look at the thread.
From what I have read the ud7 is just a ud5 with the water block option on it although when searching I have seen mention of 24 phase power on the 7 vs 16 on the 5 so there may be other differences there as well.
I remember a thread from years ago on Linux Questions when I still used it where buddy there claimed that windows screws with the network card. It caused that exact same symptom on Linux at the time now never having had a windows electron pass through my computers in years at that time I had no way to confirm that but others did chime in saying they experienced the same with their dual boots that only a cold boot (ie. disconnect from power then boot) would solve as you describe. Would that be the 8111/8168 thread BTW?
That is exactly the CMOS reset bug it corrupts it then you have to reset/start with known good configuration.
Your welcome and thank you for your kinds words, if you mean the DSDT patching of/for your board then it is in the Tutorials section on the site.
Your suggestion turned out to be terrific! d00d provides a deep Howto and instructions on how to create a dsdt.dsl which seems like the expert way to proceed. He also said,
"There's a Gigabyte bug that can cause a network port to cease functioning.
Shut the machine down, turn off the power supply, press and hold the CMOS button until the light goes out.
Start the machine, press the tab key at the logo screen, then the delete key at the post screen, then in the main menu press the F12 key to select your saved BIOS settings."
SH: I had no logo screen, it was black. That's when I found that I needed to turn off the power supply switch as well as turn off the power button of the computer, just to get to the logo screen which also fixed the network port. I haven't had to hold down the clear CMOS button.
Also d00d's instructions don't require SleepEnabler.kext which, when I installed it, brought the KP about due to kernel mis-match. His instructions about how to create
a working DSDT.aml will work for all kinds of motherboards. He didn't have mine already pre-compiled; the cpu speed, 920, 930, 950, has some impact on the DSDT.aml
I came to realize that speed step was important. I'm not at all sure that the DSDT.aml for 10.6.4 will work unedited for 10.6.5, but MultiBeast basically worked for me. Well, it has this option of Easybeast. Maybe that contains a DSDT.aml file that MultiBeast installs. I didn't use that. I just used the .kext choices, which I sort of understood after a few re-installations. My audio (889) worked and internet connectivity sort of worked, I just forgot if it ain't broke, don't fix it. But speed stepping sounds important.
My first problem with the ethernet was that neither onboard ports, en0 or en1 were recognized. Then I got them recognized, I think with RL8169 kext. The next problem was that they couldn't find or connect to a DHCP server. The message is, "self-assigned IP address, will not connect to the internet". I tried rebooting and resetting the router, switching cables and ports on router, etc. Most of the people in this thread had trouble with Airport. I noticed a few like myself had trouble with ethernet with both windows and SL. The solution for me was to power button off, turn off power switch on back, and then reboot. Now the self-assigned IP address message disappeared and was replaced by the legitimate DHCP settings etc. I have to essentially unplug going from Win to SL or SL to Win to get ethernet working = internet connectivity each time.
So now d00d confirmed that is a Gigabyte bug. I'm either going to fix this with DSDT or I'm going to buy a recommended by d00d Nic card. It is just too primitive. I don't think the 81xx kext fix works either. I wanted to learn, I'm getting my money's worth
I took due note of everything you said.
Good to see it helped I thought it would and had forgot about that network bug part in there that is what is described as a cold boot no power to the machine before you plug it in, clearing the cmos while there is no power wipes the settings to factory default that is why you need to load or re-enter your old ones.
The difference between the chips in your DSDT.aml can be the number of p/c-states it supports and the voltage you want delivered to the chip at the various states this last one can be helpful if over clocking both these factors are specific to each individual chip as Intel uses a variable VID on each chip. Now a DSDT.aml should just work between upgrades it always has for me and I don't think I have ever seen any post otherwise. And yeah speed step is really important if you like your machine running cool now if you don't feel like messing around with the p/c-states in your DSDT.aml then you can go to the OSx86 Project -> New Releases and Updates section to look for the AnVAL boot loader file (in its own massive thread) it allows auto generation of these for you among other things.
BTW I will stress again if you or anyone else reading cares about the install they have on the machine this work is best done on a testing machine/hard drive in same machine/external drive that you have cloned your working install too. There you can play to your hearts content when you have done what you want and got it right then you can duplicate the steps taken on your main install.