Hi all!
I'm answering this old thread because I had the exact same issue with
High Sierra: "kextd busy timeout IOHDACodecFunction, IOHDACodecFunction, IOHDACodecFunction"...
Admittedly, my Z68 and the SATA I HDD I make my trials on are not the best platform, but I had 7mn of boot in my first trials, including 3mn only because of that issue!
I've used AppleALC.kext + Lilu.kext and had sound perfectly working, but deleting those kexts (and rebuilding cache, of course), I had 3mn less boot time, which confirms it was an audio kext issue.
The weirdest thing is that audio keeps working, except for a small buzz (that I have had for years because my computer is connected to HiFi — the typical "ground loop" issue, see:
https://www.tonymacx86.com/threads/audio-noise-fix-trick.221472/ — so nothing strange about it)
Yes, that's incredible! I've tripled checked my S/L/E/, L/E and Clover kexts folders, no audio kext there except for the unpatched AppleHDA and I still have sound after numerous reboots in Safe Mode, etc.!
Then I've tried the modified VoodooHDA as explained here:
https://www.tonymacx86.com/threads/...o-working-on-nvidia-geforce-210-gt218.161513/ but that didn't change anything except for bringing HDMI Audio — which I don't need.
Then I've tried various things and
ended up with a minimum of 1mn 33s boot time!
Here are the things I've done, YMMV:
- unchecked all the Fixes in the ACPI page of Clover Configurator, plus checked PStates and CStates on the right (gain ~1mn)
- Sys Params page: unchecked NvidiaWeb (as I'm still using the original drivers, for the time being)
- changed SMBIOS for MacPro 5,1: total failure!
- changed SMBIOS for iMacPro 13,3: better!
- and when everything was ok I unchecked -v
About the influence of system definitions
Now, I've booted 3 times in a row and measured the boot time.
- with iMacPro 13,3: 3mn 50s, 1mn 55s and finally 1mn 33s!!!
- with the recommended iMacPro 14,2: 2mn 10s, 2mn 2s and finally 2mn.
Hard to say if iMacPro 13,3 is just totally irregular or best suited for my build...
As these were preliminary tests to see if I make the switch from 10.9 to 10.13, I'll delay applying the update until next year, when I make maybe a Mojave trial — but anyway, I should buy a second SSD for a real side by side comparison with my main OS.
I hope those trials of mine can be of help to someone having the same issue.
EDIT: more reading and more tests... I forgot to mention my checking -disablegfxfirmware boot option, even if theoretically not related to Sandy Bridge, looks it concurred to fastest boot times.
I also ended up turning back to the iMac12,2 Sys Def plus cleaning all caches with Sierra Cache Cleaner for a final 1mn 27s boot time! Yeah.
Before that, I installed Nvidia webdrivers, which really slowed down boot — though I had a slightly better result at Cinebench test... not really worth it.
EDIT 06/2019: I've finally found the culprit for the slow boot and achieved a 22s boot!
It's some issue with recent AppleALC.kext! v1.2.8 is ok. Works faster too in 10.11.
Before that, I've tried modifying a few things:
- I've put the SSDT generated with ssdtPRGen.sh in /EFI/CLOVER/ACPI/patched
- modified config.plist ACPI/Drop Tables with:
• SSDT TableId Cpu0Ist
• SSDT TableId CpuPm
• MATS (mandatory for HiSRA)
- unchecked all the Generate Options
=> all is working! Sound, sleep/wake — tested in 10.9, 10.11 and 10.13. As a side effect, I can now use a more recent version of FakeSMC (6.26.344 while I was quite stuck with 6.17 before)
