Quick answer

On a GIGABYTE Z690 AORUS Master, macOS wakes right after sleep because Apple's driver writes the wake mask 0x44 to the Aquantia AQC113 firmware, and bit 2 means wake on link up. An OpenCore kernel patch that changes 0x44 to 0x40 in AppleEthernetAquantiaAqtion (Find 8D144183C24441891424, Replace 8D144183C24041891424) stops it. Verified on macOS 15.8.2.

Does your Hackintosh wake by itself seconds after you put it to sleep, or sit in a sleep and wake loop? On the GIGABYTE Z690 AORUS Master running macOS (tested on Sequoia 15.8.2 with OpenCore), the usual cause is the onboard Aquantia AQC113 10GbE port. Apple's network driver hands the NIC firmware a wake-on-link-up flag, and the firmware wakes the whole system. This guide shows how to confirm that this is your problem and how to switch the flag off with one OpenCore kernel patch.

Symptoms

AppleEthernetAquantiaAqtion113 [en0] check wake success 1, reason 3 (Link)

Root cause

This is not an ACPI problem. Before sleep, Apple's driver (AppleEthernetAquantiaAqtion113::handleEnableNicProxy) puts the NIC firmware into sleep-proxy mode. It writes a hardcoded wake mask, 0x44, to the firmware register 0x12028. The bits follow Aquantia's wake_on_lan_s layout:

BitMeaning
0Magic packet
1Pattern
2Link up (the cause)
3Link down
4Ping
5Timer

0x44 is bit 6 plus bit 2. Bit 2 tells the firmware to wake the system whenever the link renegotiates after sleep. Switches and other link partners almost always renegotiate, so the machine wakes every time.

Why the usual fixes do not work

The wake is generated inside the NIC firmware, which nothing at the operating system or ACPI level can reach. Each of these was tried and the machine still woke:

TriedResult
_PRW/GPRW patches, RPxx TPRW SSDTsThe ACPI GPE path was closed, the machine still woke
Clearing PCIe PME_En on the NIC and root port in _PTSStill woke
PCIEXP_WAKE_DIS in PM1_EN, disarming the WAKE# GPIOStill woke
pmset womp 0, BIOS ErPStill woke
ifconfig en0 down before sleepStill woke

The fix: one OpenCore kernel patch

The patch clears bit 2 from the immediate operand, changing 0x44 to 0x40, so the firmware is no longer told to wake on link up. Magic packet and the other wake sources are unchanged. The instruction in the driver looks like this:

lea  edx, [rcx+rax*2]
add  edx, 0x44      ; 83 C2 44  becomes  83 C2 40
mov  [r12], edx

In your config.plist, add this entry under Kernel → Patch:

KeyValue
Archx86_64
CommentAquantia AQC113: disable wake on link-up
Enabledtrue
Identifiercom.apple.driver.AppleEthernetAquantiaAqtion
Find8D144183C24441891424
Replace8D144183C24041891424
Count1
MinKernel24.0.0
MaxKernel24.99.99
Base, Mask, ReplaceMaskLeave empty
How to
  1. Mount your EFI partition and make a backup copy of EFI/OC/config.plist. If you are not sure how to edit the file, see how to edit the OpenCore config.plist.
  2. Open the file in a plist editor and add the entry above under Kernel → Patch.
  3. Save, unmount the EFI partition and restart.

This was verified on macOS 15.8.2 (Darwin 24), where the byte pattern occurs only once in the kext.

Verify the fix

Put the machine to sleep, wait several minutes, then wake it with the keyboard or mouse and read the log:

pmset sleepnow
# wait several minutes, then wake with keyboard or mouse
pmset -g log | grep -E " (Sleep|Wake) " | tail -4
ioreg -lw0 | grep -i '"LastWake'      # Reason should no longer be 3

The machine should now stay asleep until you press a key or move the mouse.

After a macOS update

The patch is tied to the exact bytes in Apple's driver. After every macOS update, check that they are still there. If OpenCore reports the patch as not applied, the code has changed and the patch has to be found again.

Related notes

Is this your problem?

This fix applies if all of the following are true on your Z690 AORUS Master (or another board that uses the AQC113 with Apple's AppleEthernetAquantiaAqtion driver):

If the machine still wakes without the cable, or fails to resume at all, the NIC is not the cause; check the RAM overclock note below. I verified the patch only on the Z690 AORUS Master listed under Tested hardware. Other boards with the same NIC and driver should show the same log line, but I have not tested them.

Wake on LAN still works

The patch clears only the link-up bit. The magic packet bit stays set, so Wake on LAN keeps working as before.

Tested hardware

Frequently Asked Questions

Got questions? We've got answers.

Why does my Hackintosh wake up a few seconds after I put it to sleep?
On boards with the Aquantia AQC113 10GbE NIC, Apple's driver tells the NIC firmware to wake the system on link up. The link renegotiates after sleep, so the machine wakes every time. Unplugging the Ethernet cable stops it.
Do _PRW or GPRW patches fix this?
No. The wake is generated in the NIC firmware, so ACPI patches, clearing PME_En, disabling PCIEXP_WAKE_DIS, pmset womp 0 and BIOS ErP all leave it unchanged.
Does the patch work on every macOS version?
It was verified on macOS 15.8.2 (Darwin 24) and is limited to kernels 24.0.0 to 24.99.99. After each macOS update, check that OpenCore still applies the patch.
What if sleep still fails on later cycles?
Test with XMP off. A separate resume failure on this system came from a 6200 MT/s RAM overclock, not the NIC.
Is this fix only for the GIGABYTE Z690 AORUS Master?
It was tested only on the Z690 AORUS Master with its onboard AQC113C 10GbE. Other boards that use the AQC113 with Apple's AppleEthernetAquantiaAqtion driver should show the same "check wake success 1, reason 3 (Link)" log line, but they are untested.
How do I know the NIC is waking my Hackintosh?
The machine stays asleep with the Ethernet cable unplugged, the kernel log shows reason 3 (Link) for AppleEthernetAquantiaAqtion113, and ioreg reports LastWake Reason = 3.
Does the patch break Wake on LAN?
No. It clears only the link-up bit. The magic packet bit stays set, so Wake on LAN keeps working.