| description |
|---|
We'll go through the config.plist sections, one at a time for a Coffee Lake desktop setup. |
I like to start with either the stock config.plist that Clover gives you, or with just a blank canvas. In the next examples, I'll show you how I set things up from scratch; if you start from somewhere else, you may have more things checked/set than I do - but you'll want to follow along with what I do.
I'll also include the raw xml examples too in order to show those that work with a text editor (as I prefer to).
The default Clover settings are pretty overdone and can cause some issues. We'll keep this section fairly minimal, and I'll go through a bit of why we do that for each part as well.
<key>ACPI</key>
<dict>
<key>DSDT</key>
<dict>
<key>Fixes</key>
<dict>
<key>AddMCHC</key>
<true/>
<key>FixHPET</key>
<true/>
<key>FixIPIC</key>
<true/>
<key>FixRTC</key>
<true/>
<key>FixShutdown</key>
<true/>
<key>FixTMR</key>
<true/>
</dict>
<key>Patches</key>
<array>
<dict>
<key>Comment</key>
<string>change XHCI to XHC</string>
<key>Disabled</key>
<false/>
<key>Find</key>
<data>
WEhDSQ==
</data>
<key>Replace</key>
<data>
WEhDXw==
</data>
</dict>
<dict>
<key>Comment</key>
<string>change XHC1 to XHC</string>
<key>Disabled</key>
<false/>
<key>Find</key>
<data>
WEhDMQ==
</data>
<key>Replace</key>
<data>
WEhDXw==
</data>
</dict>
<dict>
<key>Comment</key>
<string>change SAT0 to SATA</string>
<key>Disabled</key>
<false/>
<key>Find</key>
<data>
U0FUMA==
</data>
<key>Replace</key>
<data>
U0FUQQ==
</data>
</dict>
</array>
</dict>
<key>DropTables</key>
<array>
<dict>
<key>Signature</key>
<string>DMAR</string>
</dict>
<dict>
<key>Signature</key>
<string>MATS</string>
</dict>
</array>
<key>FixHeaders</key>
<true/>
<key>SSDT</key>
<dict>
<key>Generate</key>
<dict>
<key>PluginType</key>
<true/>
</dict>
</dict>
</dict>
The first thing we'll go over is the Patches section. This section allows us to dynamically rename parts of the DSDT via Clover. Since we're not running a real mac, and macOS is pretty particular with how things are named, we can make non-destructive changes to keep things mac-friendly. We have three entries here:
- change XHCI to XHC - helps avoid a conflict with built-in USB injectors
- change XHC1 to XHC - helps avoid a conflict with built-in USB injectors
- change SAT0 to SATA - for potential SATA compatibility
You may also need the following Read-Only RTC fix (read about it here) if you have a true 300-series board (non Z370) and you suffer a boot hang around this point:
The raw XML for that specific patch looks like so:
<dict>
<key>Comment</key>
<string>Fix 300-series RTC Bug</string>
<key>Disabled</key>
<false/>
<key>Find</key>
<data>
oAqTU1RBUwE=
</data>
<key>Replace</key>
<data>
oAqRCv8L//8=
</data>
</dict>
If we look then at the Fixes section, we'll see that we have a few things checked (there are 2 pages, so I included 2 screenshots):
- FixShutdown - this can help with some boards that prefer to restart instead of shutdown. Sometimes it can cause shutdown issues on other boards (ironic, right?), so if you have issues shutting down with this enabled, look at disabling it.
- The remaining fixes help avoid IRQ conflicts and etc, and are not known to cause issues. They may not be necessary for all hardware, but do not negatively impact anything if applied.
We touched in gently on DSDT with our Patches section - and this is a a bit of an extension of that. SSDT is like a sub-section of DSDT. The Drop Tables section allows us to omit certain SSDT tables from loading (as I mentioned before, mac and PC DSDT is different, and macOS can be rather picky). The two that I've added are as follows:
- DMAR - this prevents some issues with Vt-d; which is PCI passthrough for VMs, and not very functional (if at all?) on Hackintoshes.
- MATS - with High Sierra on up this table is parsed, and can sometimes contain unprintable characters that can lead to a kernel panic.
The only other things we've done on this page are enable these two checkboxes.
- FixHeaders - this is just a double-up of our MATS table dropping. This checkbox tells Clover to sanitize headers to avoid kernel panics related to unprintable characters.
- PluginType - this injects some DSDT data to get X86PlatformPlugin to load - giving us a leg-up on native CPU power management. This setting only works on Haswell and newer CPUs though.
We don't need to do too much here, but we'll tweak a few things.
<key>Boot</key>
<dict>
<key>Arguments</key>
<string>keepsyms=1 dart=0 debug=0x100 -v</string>
<key>DefaultVolume</key>
<string>LastBootedVolume</string>
<key>Timeout</key>
<integer>5</integer>
</dict>
We have a few boot args set here:
-v- this enables verbose mode, which shows all the behind-the-scenes text that scrolls by as you're booting instead of the Apple logo and progress bar. It's invaluable to any Hackintosher, as it gives you an inside look at the boot process, and can help you identify issues, problem kexts, etc.dart=0- this is just an extra layer of protection against Vt-d issues.debug=0x100- this prevents a reboot on a kernel panic. That way you can (hopefully) glean some useful info and follow the breadcrumbs to get past the issues.keepsyms=1- this is a companion setting todebug=0x100that tells the OS to also print the symbols on a kernel panic. That can give some more helpful insight as to what's causing the panic itself.
These are the only other settings I've updated in this section.
- DefaultBootVolume - this uses NVRAM to remember which volume was last booted by Clover, and auto-select that at the next boot.
- Timeout - this is the number of seconds before the DefaultBootVolume auto-boots. You can set this to
-1to avoid all timeout, or to0to skip the GUI entirely. If set to0, you can press any keys at boot to get the GUI to show back up in case of issues.
Nothing here - just the stock settings. You could adjust this if Clover's scaling needs changes, but I don't mess with it.
Again, nothing here gets changed in most setups I've worked with.
We'll handle some slick property injection for WhateverGreen here, and do some basic audio setup.
<key>Devices</key>
<dict>
<key>Audio</key>
<dict>
<key>Inject</key>
<integer>1</integer>
<key>ResetHDA</key>
<true/>
</dict>
<key>Properties</key>
<dict>
<key>PciRoot(0x0)/Pci(0x2,0x0)</key>
<dict>
<key>AAPL,ig-platform-id</key>
<data>
BwCbPg==
</data>
<key>framebuffer-patch-enable</key>
<data>
AQAAAA==
</data>
<key>framebuffer-stolenmem</key>
<data>
AAAwAQ==
</data>
</dict>
</dict>
<key>USB</key>
<dict>
<key>FixOwnership</key>
<true/>
</dict>
</dict>
This section remains empty for our example setup. In the past, almost-supported iGPUs (like the HD 4400) would get faked to a supported iGPU here, but we'll be using the cleaner Properties section for this.
Under this section, we ensure that Inject and FixOwnership are selected to avoid issues with hanging at a half-printed line somewhere around the Enabling Legacy Matching verbose line. You can also get past that by enabling XHCI Hand Off in BIOS.
Here we set our audio to inject Layout 1 - this may or may not be compatible with your codec, but you can check on AppleALC's Supported Codec Page.
We also enabled ResetHDA which puts the codec back in a neutral state between OS reboots. This prevents some issues with no audio after booting to another OS and then back.
This section is setup via Headkaze's Intel Framebuffer Patching Guide and applies only one actual property to begin, which is the ig-platform-id. The way we get the proper value for this is to look at the ig-platform-id we intend to use, then swap the pairs of hex bytes.
If we think of our ig-plat as 0xAABBCCDD, our swapped version would look like 0xDDCCBBAA.
The two ig-platform-id's we use are as follows:
0x3E9B0007- this is used when the iGPU is used to drive a display07009B3Ewhen hex-swappedBwCbPg==when the hex-swapped version is converted to base64
0x3E920003- this is used when the iGPU is only used for compute tasks, and doesn't drive a display0300923Ewhen hex-swappedAwCSPg==when the hex-swapped version is converted to base64
Worth noting that for 10.12 -> 10.13.5, you would need to fake the iGPU to the same values in the Kaby Lake guide, as this was before native Coffee Lake iGPU showed up.
We also add 2 more properties, framebuffer-patch-enable and framebuffer-stolenmem. The first enables patching via WhateverGreen.kext, and the second sets the min stolen memory to 19MB.
I added another screenshot as well that shows a device-id fake for the i3-8100's UHD 630. This has a different device id than the UHD 630 found on the 8700k, for instance (3e918086 vs 3e928086 ).
For this - we follow a similar procedure as our above ig-platform-id hex swapping - but this time, we only work with the first two pairs of hex bytes. If we think of our device id as 0xAABB0000, our swapped version would look like 0xBBAA0000. We don't do anything with the last 2 pairs of hex bytes.
The device-id fake is setup like so:
0x3e920000- this is the device id for the UHD 630 found on an 8700k923e0000when hex swappedkj4AAA==when the hex-swapped version is converted to base64
If using the raw xml, your Properties would look like this (make sure to still use the appropriate ig-platform-id for your setup):
<key>Properties</key>
<dict>
<key>PciRoot(0x0)/Pci(0x2,0x0)</key>
<dict>
<key>device-id</key>
<data>
kj4AAA==
</data>
<key>AAPL,ig-platform-id</key>
<data>
BwCbPg==
</data>
<key>framebuffer-patch-enable</key>
<data>
AQAAAA==
</data>
<key>framebuffer-stolenmem</key>
<data>
AAAwAQ==
</data>
</dict>
</dict>
I've seen a couple users report a pink tint when using HDMI with the UHD 630 iGPU. I did some experimenting on my own system and was able to work around it a couple different ways.
I saw the issue in a reddit post, and the solution was to apply a custom override for your display to force it to use RGB instead of YCbCr outlined here. I wrote a script that wraps around this method and auto-applies the override when it's completed. This worked fine for me, but didn't feel like a real fix. Which lead me to dive a bit deeper...
I opened up IORegistryExplorer and in the search bar typed IGPU (this is sometimes named GFX0 in ACPI, but Lilu + WhateverGreen should rename it properly) and got the following screen:
Once we've located IGPU in IOReg, we can clear our search - this reveals all the info around the IGPU section while keeping our place:
As you can see in the above screenshot, I had a few different AppleIntelFramebuffer connections listed. I'm looking for the one that's specifically driving my display - which has the AppleDisplay property. In my case, this was AppleIntelFramebuffer@1. With that selected on the left pane, you can find the connector-type property, which was originally set to <00 04 00 00> in my case. The connector type can have a few different values:
<00 04 00 00>- this is DisplayPort<00 08 00 00>- this is HDMI<04 00 00 00>- this is Digital DVI<02 00 00 00>- this is LVDS (for laptops)<01 00 00 00>- this is just a Dummy port
What I noticed in my case was that my HDMI port was listed as a DisplayPort - so I was able to use WhateverGreen's patching abilities to change the connector-type.
Since my incorrect port was located at AppleIntelFramebuffer@1, this is port 1. I needed to enable the port patch in Properties, and then set the connector type to HDMI. I used the following Properties entries for that:
framebuffer-conX-enable = 01000000framebuffer-conX-type = 00080000
I replaced the conX in both patches with con1 to reflect the port that I am changing, then set the values as listed above.
<key>Properties</key>
<dict>
<key>PciRoot(0x0)/Pci(0x2,0x0)</key>
<dict>
<key>AAPL,ig-platform-id</key>
<data>
BwCbPg==
</data>
<key>device-id</key>
<data>
kj4AAA==
</data>
<key>framebuffer-con1-enable</key>
<data>
AQAAAA==
</data>
<key>framebuffer-con1-type</key>
<data>
AAgAAA==
</data>
<key>framebuffer-patch-enable</key>
<data>
AQAAAA==
</data>
<key>framebuffer-stolenmem</key>
<data>
AAAwAQ==
</data>
</dict>
</dict>
This also enabled HDMI audio for me as well.
We have nothing to do here.
<key>GUI</key>
<dict>
<key>Scan</key>
<dict>
<key>Entries</key>
<true/>
<key>Tool</key>
<true/>
</dict>
</dict>
The only settings I've tweaked on this page are the Scan settings. I've selected Custom, then checked everything except Legacy and Kernel. This just omits some of the unbootable entries in Clover to clean up the menu.
I haven't added anything here, but you can hide unwanted volumes here. You can do so by either adding the volume's name, or UUID.
To hide extra APFS entries, add the following to this list:
PrebootVM
To hide all Recovery partitions, add Recovery to the list.
To get the UUID of a drive to hide, you can use the following terminal command:
diskutil info diskXsY | grep -i "Partition UUID" | rev | cut -d' ' -f 1 | rev
Make sure to replace diskXsY with the actual disk number of the volume you'd like to hide.
If you want to test out a new theme (and I suggest you look at clover-next-black), you can add the unzipped theme folder to the /Volumes/EFI/EFI/CLOVER/themes directory, then type the name of the folder in the Theme text field to apply it.
In the past, we'd setup the iGPU here, but since we already did that via Properties in the Devices section, we have nothing to really configure here. NOTE: When Clover detects an Intel iGPU, it automatically enables Intel Injection if the Graphics section doesn't exist in the config.plist. To bypass this, you can explicitly disable injection using the raw XML below, or by clicking the "Inject Intel" button once to check it, and once to uncheck it in CC.
<key>Graphics</key>
<dict>
<key>Inject</key>
<false/>
</dict>
<key>KernelAndKextPatches</key>
<dict>
<key>KernelPm</key>
<true/>
<key>KextsToPatch</key>
<array>
<dict>
<key>Comment</key>
<string>Port limit increase</string>
<key>Disabled</key>
<false/>
<key>Find</key>
<data>
g710////EA==
</data>
<key>InfoPlistPatch</key>
<false/>
<key>MatchOS</key>
<string>10.12.x</string>
<key>Name</key>
<string>com.apple.driver.usb.AppleUSBXHCI</string>
<key>Replace</key>
<data>
g710////Gw==
</data>
</dict>
<dict>
<key>Comment</key>
<string>Port limit increase (RehabMan)</string>
<key>Disabled</key>
<false/>
<key>Find</key>
<data>
g32IDw+DpwQAAA==
</data>
<key>InfoPlistPatch</key>
<false/>
<key>MatchOS</key>
<string>10.13.x</string>
<key>Name</key>
<string>com.apple.driver.usb.AppleUSBXHCI</string>
<key>Replace</key>
<data>
g32ID5CQkJCQkA==
</data>
</dict>
<dict>
<key>Comment</key>
<string>Port limit increase (PMHeart)</string>
<key>Disabled</key>
<true/>
<key>Find</key>
<data>
g/sPD4MDBQAA
</data>
<key>InfoPlistPatch</key>
<false/>
<key>MatchOS</key>
<string>10.14.0</string>
<key>Name</key>
<string>com.apple.driver.usb.AppleUSBXHCI</string>
<key>Replace</key>
<data>
g/sPkJCQkJCQ
</data>
</dict>
<dict>
<key>Comment</key>
<string>Port limit increase (Ricky)</string>
<key>Disabled</key>
<false/>
<key>Find</key>
<data>
g/sPD4OPBAAA
</data>
<key>InfoPlistPatch</key>
<false/>
<key>MatchOS</key>
<string>10.14.x</string>
<key>Name</key>
<string>com.apple.driver.usb.AppleUSBXHCI</string>
<key>Replace</key>
<data>
g/sPkJCQkJCQ
</data>
</dict>
<dict>
<key>Comment</key>
<string>External Icons Patch</string>
<key>Disabled</key>
<false/>
<key>Find</key>
<data>
RXh0ZXJuYWw=
</data>
<key>InfoPlistPatch</key>
<false/>
<key>Name</key>
<string>AppleAHCIPort</string>
<key>Replace</key>
<data>
SW50ZXJuYWw=
</data>
</dict>
</array>
</dict>
In this section, we've enabled a few settings and added some kext patches.
We have a couple checkboxes selected here:
- Apple RTC - this ensures that we don't have a BIOS reset on reboot.
- KernelPM - this setting prevents writing to MSR 0xe2 which can prevent a kernel panic at boot.
We added 4 different kexts to patch here. Three of them are for USB port limit increases, and the last acts as an orange icons fix - when internal drives are hotpluggable, and treated as external drives.
You'll notice that there are MatchOS values set for each of the USB port limit patches. You can remove any of the entries for OS versions you don't intend to run. They won't do any harm being there, but if you want a clean, minimal plist, there's no sense in having them.
<key>RtVariables</key>
<dict>
<key>BooterConfig</key>
<string>0x28</string>
<key>CsrActiveConfig</key>
<string>0x3E7</string>
<key>MLB</key>
<string>C02726902CDH69F1M</string>
<key>ROM</key>
<string>UseMacAddr0</string>
</dict>
<key>SMBIOS</key>
<dict>
<key>BoardSerialNumber</key>
<string>C02726902CDH69F1M</string>
<key>ProductName</key>
<string>iMac18,1</string>
<key>SerialNumber</key>
<string>C02TX0VDH7JY</string>
<key>SmUUID</key>
<string>91492A73-595C-4D97-A6FC-2B5D3ED1B54D</string>
</dict>
For setting up the SMBIOS info, I use acidanthera's macserial application. I wrote a python script that can leverage it as well (and auto-saves to the config.plist when selected). There's plenty of info that's left blank to allow Clover to fill in the blanks; this means that updating Clover will update the info passed, and not require you to also update your config.plist.
For this Coffee Lake example, I chose the iMac18,1 SMBIOS - this is done intentionally for compatibility's sake. There are two main SMBIOS used for Coffee Lake:
- iMac18,1 - this is used for computers utilizing the iGPU for displaying.
- iMac18,3 - this is used for computers using a dGPU for displaying, and an iGPU for compute tasks only.
To get the SMBIOS info generated with macserial, you can run it with the -a argument (which generates serials and board serials for all supported platforms). You can also parse it with grep to limit your search to one SMBIOS type.
With our iMac18,1 example, we would run macserial like so via the terminal:
macserial -a | grep -i iMac18,1
Which would give us output similar to the following:
iMac18,1 | C02T8SZNH7JY | C02707101J9H69F1F
iMac18,1 | C02VXBYDH7JY | C02753100GUH69FCB
iMac18,1 | C02T7RY6H7JY | C02706310GUH69FA8
iMac18,1 | C02VD07ZH7JY | C02737301J9H69FCB
iMac18,1 | C02TQPYPH7JY | C02720802CDH69FAD
iMac18,1 | C02VXYYVH7JY | C02753207CDH69FJC
iMac18,1 | C02VDBZ0H7JY | C02737700QXH69FA8
iMac18,1 | C02VP0H6H7JY | C02746300CDH69FJA
iMac18,1 | C02VL0W9H7JY | C02743303CDH69F8C
iMac18,1 | C02V2NYMH7JY | C02728600J9H69FAD
The order is Product | Serial | Board Serial (MLB)
The iMac18,1 part gets copied to SMBIOS -> Product Name.
The Serial part gets copied to SMBIOS -> Serial Number.
The Board Serial part gets copied to SMBIOS -> Board Serial Number as well as Rt Variables -> MLB.
We can create an SmUUID by running uuidgen in the terminal (or it's auto-generated via my GenSMBIOS script) - and that gets copied to SMBIOS -> SmUUID.
We set Rt Variables -> ROM to UseMacAddr0 which just utilizes our onboard Mac address - this should be unique enough to not conflict with any others.
BooterConfig gets set to 0x28, and CsrActiveConfig is set to 0x3e7 which effectively disables SIP. You can choose a number of other options to enable/disable sections of SIP. Some common ones are as follows:
0x0- SIP completely enabled0x3- Allow unsigned kexts and writing to protected fs locations0x3e7- SIP completely disabled
<key>SystemParameters</key>
<dict>
<key>InjectKexts</key>
<string>Yes</string>
<key>InjectSystemID</key>
<true/>
</dict>
This setting has 3 modes:
Yes- this tells Clover to inject kexts from the EFI regardless.No- this tells Clover not to inject kexts from the EFI.Detect- this has Clover inject kexts only if FakeSMC.kext is not in the kext cache.
We set it to Yes to make sure that all the kexts we added before get injected properly.
This setting tells clover to set the SmUUID as the system-id at boot - which is important for iMessage and such.
At this point, you can do File -> Save to save the config.plist. If you have issues saving directly to the EFI, you can save it on the Desktop, then just copy it over. I'll leave the sample config.plist here too.