Contribute
Register

High Sierra on Legacy P55M board

Status
Not open for further replies.
Joined
Jul 6, 2010
Messages
2,573
Motherboard
Asus Z170M-Plus
CPU
i5-6600K
Graphics
R9 280
Mac
  1. MacBook Pro
  2. Mac Pro
Mobile Phone
  1. iOS
Hi guys,

I know I'm late, but I just updated my old GA-P55M-UD2 build, which was rock-stable running El Capitan before, to High Sierra and encountered two major problems. Since this board was quite popular at its time and no one mentioned those problems with this board I'm wondering whether it's just me or if no one cares about this old hardware anymore.

1. Trouble with SATA
After upgrading, my system SSD was detected perfectly fine, but my 2TB Western Digital data drive was detected as "corrupt", and macOS asked my to format my drive. First I thought it was just dying, but later I found similar posts from other users with old chipsets, and it turned out the following patch was necessary to get it working:

Code:
        <key>KextsToPatch</key>
        <array>
            <dict>
                <key>Comment</key>
                <string>fix IO error ICH10 for 10.13, credit SunKi</string>
                <key>Disabled</key>
                <false/>
                <key>Find</key>
                <data>
                RYX2D5XCiciD4P5mhcl4D4TSdQs=
                </data>
                <key>InfoPlistPatch</key>
                <false/>
                <key>MatchOS</key>
                <string>10.13.x</string>
                <key>Name</key>
                <string>AppleAHCIPort</string>
                <key>Replace</key>
                <data>
                iciD4P5mhckPmMFBCMyQkJCQdQs=
                </data>
            </dict>
            <dict>
                <key>Comment</key>
                <string>fix hotplug ICH10 for 10.13, credit SunKi</string>
                <key>Disabled</key>
                <false/>
                <key>Find</key>
                <data>
                icglQGACAD1AIAAAdQyB4b9///+Ji1EBAAA=
                </data>
                <key>InfoPlistPatch</key>
                <false/>
                <key>MatchOS</key>
                <string>10.13.x</string>
                <key>Name</key>
                <string>AppleAHCIPort</string>
                <key>Replace</key>
                <data>
                kJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJA=
                </data>
            </dict>
        </array>

2. Big trouble after sleep
Putting the system to sleep worked perfectly fine, but trying to wake up the machine rendered it DEAD. The lights and fans would turn on for a fraction of a second, just to turn off again. Then the machine remained quiet for a few seconds and the process repeated. The only way to make it POST again (and reach the BIOS splash screen or Clover) was removing the CMOS battery and resetting everything. Never had such a horrible crash before, I actually thought the mainboard was gone.

After repeating the process a few times I finally figured out that it was related to the AppleRTC patch applied by Clover. When disabling the patch, the good old "Invalid checksum" error on boot would return, but the machine was finally able to wake from sleep without bricking itself.

If found some other people with this issue, e.g. here and here, but no one revealed a solution. So I dug a little further and found this old patch which disables writes to the CMOS by AppleRTC completely, instead of just disabling the checksum fixes. Loading the 10.13 AppleRTC into an disassembler and comparing it to the 10.9 one revealed that the patched function was still present, just in a different location and with a different register as argument. So, the correct patch for 10.13 AppleRTC to fully disable it looks like this:

Search: 4189D74189F6
Replace: E9B800000090

Didn't bother to implement it in Clover yet, just tried it with an hex editor a few minutes ago and it works perfectly fine: Sleep/wake works and CMOS checksum error is also gone.

Can someone with this or a similar system confirm this behavior?
 
These were a big help to me.

I ended up replacing AppleAHCIPort with v328 instead of patching above. I used the file and method in this post: https://www.tonymacx86.com/threads/...installer-disk-i-o-error.233104/#post-1592690 .

I think the fix to this problem is to place the Apple AHCIPort.kext v328 in L/E using a kext installer. The idea is it won't be over written with any future updates. There is no need to delete the new one from S/L/E. The problem is the new one doesn't recognise older ssd disks.


I am using the AppleRTC patch you have above for now (just made one right in Clover Configuator using your seach and replace on AppleRTC you have listed). My machine didn't die, but it thought I had set a BIOS password after one test - it reboots immediately unpatched when put to sleep. Luckly there is BIOS reset jumper - the CMOS battery is under my CPU cooler, so a pain to get to.
 
PS the version of Clover that ships with Multibeast for High Sierra (v10.4.0) will not boot my machine - it gets stuck at Loading Operating System. When I upgrade (r4871 currently) it works.
 
Hi guys,

I know I'm late, but I just updated my old GA-P55M-UD2 build, which was rock-stable running El Capitan before, to High Sierra and encountered two major problems. Since this board was quite popular at its time and no one mentioned those problems with this board I'm wondering whether it's just me or if no one cares about this old hardware anymore.

1. Trouble with SATA
After upgrading, my system SSD was detected perfectly fine, but my 2TB Western Digital data drive was detected as "corrupt", and macOS asked my to format my drive. First I thought it was just dying, but later I found similar posts from other users with old chipsets, and it turned out the following patch was necessary to get it working:

Code:
        <key>KextsToPatch</key>
        <array>
            <dict>
                <key>Comment</key>
                <string>fix IO error ICH10 for 10.13, credit SunKi</string>
                <key>Disabled</key>
                <false/>
                <key>Find</key>
                <data>
                RYX2D5XCiciD4P5mhcl4D4TSdQs=
                </data>
                <key>InfoPlistPatch</key>
                <false/>
                <key>MatchOS</key>
                <string>10.13.x</string>
                <key>Name</key>
                <string>AppleAHCIPort</string>
                <key>Replace</key>
                <data>
                iciD4P5mhckPmMFBCMyQkJCQdQs=
                </data>
            </dict>
            <dict>
                <key>Comment</key>
                <string>fix hotplug ICH10 for 10.13, credit SunKi</string>
                <key>Disabled</key>
                <false/>
                <key>Find</key>
                <data>
                icglQGACAD1AIAAAdQyB4b9///+Ji1EBAAA=
                </data>
                <key>InfoPlistPatch</key>
                <false/>
                <key>MatchOS</key>
                <string>10.13.x</string>
                <key>Name</key>
                <string>AppleAHCIPort</string>
                <key>Replace</key>
                <data>
                kJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJA=
                </data>
            </dict>
        </array>

2. Big trouble after sleep
Putting the system to sleep worked perfectly fine, but trying to wake up the machine rendered it DEAD. The lights and fans would turn on for a fraction of a second, just to turn off again. Then the machine remained quiet for a few seconds and the process repeated. The only way to make it POST again (and reach the BIOS splash screen or Clover) was removing the CMOS battery and resetting everything. Never had such a horrible crash before, I actually thought the mainboard was gone.

After repeating the process a few times I finally figured out that it was related to the AppleRTC patch applied by Clover. When disabling the patch, the good old "Invalid checksum" error on boot would return, but the machine was finally able to wake from sleep without bricking itself.

If found some other people with this issue, e.g. here and here, but no one revealed a solution. So I dug a little further and found this old patch which disables writes to the CMOS by AppleRTC completely, instead of just disabling the checksum fixes. Loading the 10.13 AppleRTC into an disassembler and comparing it to the 10.9 one revealed that the patched function was still present, just in a different location and with a different register as argument. So, the correct patch for 10.13 AppleRTC to fully disable it looks like this:

Search: 4189D74189F6
Replace: E9B800000090

Didn't bother to implement it in Clover yet, just tried it with an hex editor a few minutes ago and it works perfectly fine: Sleep/wake works and CMOS checksum error is also gone.

Can someone with this or a similar system confirm this behavior?


Pretty much identical issues except my #1 was message how drive is not initialized (instead of corrupted). I used your code for my conifg file and no message after next boot.

As for #2 what I did was navigate to/System/Library/Extensions/AppleRTC.kext/Contents/MacOS/AppleRTC and opened up this file inhex editor, found 4189D74189F6 and replaced with E9B800000090.Saved this file and tried to sleep and no POST / fans spinning up and going out right after again. So I need help with this. Should I have disabled patch first ?
 
Which macOS are you on? Didn't verify it with anything but 10.13.6, didn't update the machine any further yet.

Did you fix permissions after modifying the file?

I am on 10.13.6. First time using Clover (used Yosemite since it came out on my other hackintosh).
I just booted into it again and started disk utility (not sure if I should have booted through USB for this and run disk utility through there) and run first aid. If this is not right let me know.
Same stupid bootup again.

I must be missing a step somewhere (fixing permissions / disabling patch / something else?).
If I get your kext and replace mine, would that work ?
 
Last edited:
I am spending a little bit more of time with this problem.
My current situation:
1. With original patch the mobs goes to that crazy hiccup state when put in sleep.
2. If I disable original patch and then fix permissions the mobs goes to sleep nicely and it also wakes up nicely but after shutdown it reset CMOS.
3. When I replace value you mentioned (and fix permissions afterwards) and then put PC to sleep the screen and mouse go to sleep but fans do not. When I wake It up, mouse comes back to life but screen does not and I have to shut it down (but no CMOS reset).

So how do I get both positive outcomes from 2 and 3 ? It seems this value stops the sleep but also block the CMOS reset.
 
Status
Not open for further replies.
Back
Top