[HELP] macOS 15.6 installer stops at IOPCIConfigurator::configure() - Ryzen 7 7435H laptop, RTX 4060 as the only GPU (no iGPU)
Hi everyone,
I have been trying to install macOS 15.6 (Sequoia) on a MECHREVO laptop and I am stuck at PCI configuration. I have read the Dortania troubleshooting page and tried every booter quirk combination I could find. I would really appreciate a pointer before I give up on this machine.
HARDWARE
SOFTWARE
SYMPTOM
Verbose boot goes through ACPI tables, AppleACPICPU 1..16, IOAPIC, RTC, then:
[ PCI configuration begin ]
IOPCIConfigurator::configure() PCI configuration PCI0
and then the machine either hangs or resets. There is no kernel panic, and the next boot logs "OC: Panic log does not exist".
Booting with -x is not usable here: boot.efi stops before the kernel with
AAPL: #[EB|STOP] 0x16
AAPL: Displaying boot failure graphic for 10 seconds...
EVIDENCE (from OpenCore SysReport)
1) The GPU aperture is above 4 GB:
GOP INFO 1920(2048)x1080 fmt 1
GOP FB 0x7C00000000[0x870000]
2) PCI0._CRS is computed at runtime and ends up empty. The DSDT uses a template and then fills it from firmware variables:
Name (CRS2, ResourceTemplate () { ... 8A 2B 00 00 0C 03 00 ... })
Method (_CRS, 0, Serialized) {
...
Store (BRB_MIN0, MIN0)
Store (BRL_LEN0, LEN0)
...
}
The 64-bit (0x8A) address-space descriptors come out as min=0 max=0 len=0, i.e. PCI0 advertises no 64-bit MMIO window.
3) The firmware's own OS check is newer than SSDT-XOSI covers. Firmware SSDT-7 (32 KB) tests _OSI for "Windows 2016/2017/2018/2019/2020/2021" and "Windows 11", while OpenCore's SSDT-XOSI.aml only returns true up to "Windows 2015". So even with the _OSI to XOSI rename the firmware takes its non-Windows branch.
4) OCABC: MAT support is 1; Above 4G appears enabled (framebuffer above 4 GB).
WHAT I HAVE TRIED
Background: with the 26.6 installer the system died at EXITBS with no output at all; moving to 15.6 got a kernel running. Removing cpus=1 got past an early SMBus panic (CR2 = 0xFED80260). npci=0x3000 then moved the stop point to PCI configuration, where it now stays.
MY QUESTIONS
1) Is there a supported way to give PCI0 a valid 64-bit MMIO window when the firmware computes _CRS at runtime (an SSDT that overrides _CRS, or patching the _INI that sets BRB_MIN0 / BRL_LEN0 / PMIN)?
2) Are there known workarounds for AMD laptops where the root bridge reports an empty 64-bit window and the dGPU BAR lives above 4 GB?
3) Would a different SMBIOS (iMacPro1,1 / MacPro7,1) or another npci value change how IOPCIConfigurator handles an empty window?
Attached: SysReport (all ACPI tables including DSDT and SSDTs, plus PCI/CPU/GOP info), config.plist, and the OpenCore logs of the last few boots.
Thanks for reading.
Hi everyone,
I have been trying to install macOS 15.6 (Sequoia) on a MECHREVO laptop and I am stuck at PCI configuration. I have read the Dortania troubleshooting page and tried every booter quirk combination I could find. I would really appreciate a pointer before I give up on this machine.
HARDWARE
- Laptop: MECHREVO JIAOLONG15K Series GM5BG0E
- CPU: AMD Ryzen 7 7435H (Rembrandt-R, Zen 3+, 8C/16T, family 19h model 44h) - this SKU has NO integrated GPU
- dGPU: NVIDIA RTX 4060 Laptop (10DE:28A0) - the only display device in the system
- RAM: 16 GB, NVMe: Kingston OM8PGP4512N-A0
- BIOS: AMI 2.22.0054 (2024). There is no Above 4G / CSM / IOMMU / Resizable-BAR toggle exposed in the setup menu (only Device Control options), so BIOS-level workarounds are not available.
SOFTWARE
- OpenCore DBG-109-2026-10-23 (1.0.9 debug build)
- Installer: macOS 15.6 BaseSystem (Darwin 24.6.0) via com.apple.recovery.boot/BaseSystem.dmg on USB
- SMBIOS MacBookPro16,2
- AMD_Vanilla kernel patches: 27 entries, all report Success for 24.6.0 in the log
- Kexts: Lilu, VirtualSMC, WhateverGreen, AppleALC, AppleMCEReporterDisabler, VoodooI2C (+HID), VoodooPS2, USBToolBox/UTBDefault, NVMeFix, RealtekRTL8111, itlwm, ECEnabler, SMCBatteryManager, SMCLightSensor, BrightnessKeys, IntelBluetoothFirmware/Patcher, BlueToolFixup, RestrictEvents, AMFIPass
- boot-args: -v keepsyms=1 debug=0x100 -vi2c-force-polling npci=0x3000
SYMPTOM
Verbose boot goes through ACPI tables, AppleACPICPU 1..16, IOAPIC, RTC, then:
[ PCI configuration begin ]
IOPCIConfigurator::configure() PCI configuration PCI0
and then the machine either hangs or resets. There is no kernel panic, and the next boot logs "OC: Panic log does not exist".
Booting with -x is not usable here: boot.efi stops before the kernel with
AAPL: #[EB|STOP] 0x16
AAPL: Displaying boot failure graphic for 10 seconds...
EVIDENCE (from OpenCore SysReport)
1) The GPU aperture is above 4 GB:
GOP INFO 1920(2048)x1080 fmt 1
GOP FB 0x7C00000000[0x870000]
2) PCI0._CRS is computed at runtime and ends up empty. The DSDT uses a template and then fills it from firmware variables:
Name (CRS2, ResourceTemplate () { ... 8A 2B 00 00 0C 03 00 ... })
Method (_CRS, 0, Serialized) {
...
Store (BRB_MIN0, MIN0)
Store (BRL_LEN0, LEN0)
...
}
The 64-bit (0x8A) address-space descriptors come out as min=0 max=0 len=0, i.e. PCI0 advertises no 64-bit MMIO window.
3) The firmware's own OS check is newer than SSDT-XOSI covers. Firmware SSDT-7 (32 KB) tests _OSI for "Windows 2016/2017/2018/2019/2020/2021" and "Windows 11", while OpenCore's SSDT-XOSI.aml only returns true up to "Windows 2015". So even with the _OSI to XOSI rename the firmware takes its non-Windows branch.
4) OCABC: MAT support is 1; Above 4G appears enabled (framebuffer above 4 GB).
WHAT I HAVE TRIED
- npci=0x2000 : stops earlier, right after the RTC lines
- npci=0x3000 : stops at IOPCIConfigurator::configure()
- ResizeAppleGpuBars = 0 and -1 : no change
- DevirtualiseMmio on / off : no change
- ProtectMemoryRegions = true : no change
- Reset NVRAM before each attempt : no change
- -x (safe mode) : boot.efi STOP 0x16, kernel never starts
- ACPI string patches "Windows 2021/2020/2019/2018/2017/2016" -> "Windows 2015" (so XOSI answers true) : patches applied, no change
Background: with the 26.6 installer the system died at EXITBS with no output at all; moving to 15.6 got a kernel running. Removing cpus=1 got past an early SMBus panic (CR2 = 0xFED80260). npci=0x3000 then moved the stop point to PCI configuration, where it now stays.
MY QUESTIONS
1) Is there a supported way to give PCI0 a valid 64-bit MMIO window when the firmware computes _CRS at runtime (an SSDT that overrides _CRS, or patching the _INI that sets BRB_MIN0 / BRL_LEN0 / PMIN)?
2) Are there known workarounds for AMD laptops where the root bridge reports an empty 64-bit window and the dGPU BAR lives above 4 GB?
3) Would a different SMBIOS (iMacPro1,1 / MacPro7,1) or another npci value change how IOPCIConfigurator handles an empty window?
Attached: SysReport (all ACPI tables including DSDT and SSDTs, plus PCI/CPU/GOP info), config.plist, and the OpenCore logs of the last few boots.
Thanks for reading.