Hi @Joni174
The issue you are facing with the Intel BE200 blocking the boot process on the AM4 plaform (while working on others platforms) is encountered by a lot of users, myself included. So it seems that it's linked to AGESA Combo-AM4v2PI (at least until 1.2.0.Ca, as I don't have any motherboard running on further release yet), which is the common factor between all those boards.
As nothing is moving for months already (or we don't know yet), I made some extensive testing, and using the COM_DEBUG port, I collected the detailed stream (codes and timestamps) of "Asus Q-Codes" from my TUF GAMING B550-PLUS motherboard to compare between working boots and blocked boots by the Intel BE200.
I added some white lines to align the same blobs of Q-Codes despite slight differences between boots, and placed everything into a spreadsheet to spot differences and timings. See https://www.asusqcodes.com/ for example for Q-Codes meaning.
https://pix-server-sorel.luoss.fr/Manual/Intel-BE200-on-AMD-Combo-AM4v2/boot-compares-v2.ods
- In presence of the blocking Intel BE200 WiFi card, the boot process almost entirely occurs before stopping at the really final steps.
- Looks like the really long time before having the screen turning on is spent waiting between code 07 (AP initialization after microcode loading) to 99 (Super IO Initialization). It seems that AP stands for Application Processor. Both of these 2 Q-Codes (07 and 99) appears several times, not always as slow. At this step, it looks like this is the GPU initialization because during those 20 seconds, the screen is off. Looks like a timeout after something didn't happen or has been ignored/missed.
- It eventually runs aground when going from code A0 to A2 to A0 again: got A0 (IDE initialization is started), A2 (IDE Detect), then nothing more is happening (like IDE detection never ends, or the Detection end signaling is missed/ignored).
I can confirm the BE200 card is working, because it works fine on 1 of my computers (both as WiFi and Bluetooth). Details about it:
- Gigabyte GA-970A-DS3P rev 2.x on BIOS version FD from 2016/02/26 (CPU is AMD FX8350).
- Intel BE200 card is shown as 8086:272b as PCI Vendor:Device ID
- Connected using a PCI-Express X1 <-> M.2 Key E adapter with USB front panel connector cable for Bluetooth
But in 3 of my computers, the UEFI/BIOS get stuck (screen turns on after a very long time, I hear the beep, then nothing more happens). Details about them:
- MSI X470 GAMING PLUS MAX (MS-7B79) on BIOS version 7B79vHK / H.K0 (AGESA ComboAm4v2PI 1.2.0.Ca) (from 2024/07/23). CPU is Ryzen 7 3700X. Sometimes reboot in loop instead. AMD R9 280X GPU
- Asus TUF GAMING B550-PLUS on BIOS version 3607 (AGESA ComboV2PI 1.2.0.Ca) (from 2024/03/18). CPU is Ryzen 5 5500. NVIDIA GTX 950 GPU
- Lenovo Laptop 81W4 15ARE05 on BIOS version DZCN49WW (2023/09/19). CPU is Ryzen 5 4500U
About this Intel® BE200NGW based card, here are the markings
- Intel® Killer™ BE1750x
- Intel® BE200NGW
- Anatel: 06538-23-04423
- T PN: G86C0008A810
- SPS: XXXXXX-XXX XX
- FRU: 5W11H85463
- WFM: 105FADF8D3BA
- BDM: 105FADF8D3BE
- TA: N28353-003
- MM: 99C477
Under the QR Code:
- Scanning the QR Code says 8SSW11H85141T1SS45H065T
- 8SSW11H85141
- T1SS45H065T
- EC: 6653678
- Made in China
Bottom-right corner:
- IC: 1000M-BE200NG
- FCC ID: PD9BE200NG
Obtained through Amazon.fr (ships from Amazon, sold by Amazon).
Already tried:
- Resizable BAR enabled/disabled
- Above 4G decoding enabled/disabled
- USB cable for Bluetooth connected/disconnected
- Every available PCIe port
- Plugging the WiFi card in place of the GPU (and trying to boot without the GPU), as it is connected to the CPU instead of chipset
- Plugging the WiFi card in place of the GPU (and trying to boot with the GPU on another port)
- Hot plugging the WiFi card (at GRUB level depending on which port is used, the screen goes blank or not, but anyway after loading kernel + initramfs, it fails to go further)
- Hot plugging the WiFi card once the OS has booted (but apart from Bluetooth becoming visible, even `echo 1 > /sys/bus/pci/rescan` did nothing into dmesg logs, or into lspci output.
- Reducing the speed of PCIe ports to Gen 1
- Enable/Disable ASPM from BIOS
- Enable/Disable Data Link Feature Exchange from BIOS
- Enable/Disable SR-IOV from BIOS
- Removing M.2 NVMe SSD
- Disabling SuperIO Parallel port and COM port just in case
- Disabling TPM related things
By looking on the Internet about this subject, I found out many others people are facing the same issue
- https://www.reddit.com/r/Amd/comments/1cinpvm/be200_wifi_7_working_on_am2/
- https://community.amd.com/t5/processors/amd-and-wifi-7/td-p/657636
- https://www.tomshardware.com/networking/intel-has-a-new-driver-for-wi-fi-7-adapters-but-they-may-sti...
- https://community.intel.com/t5/Wireless/Laptop-doesn-t-boot-with-BE200/m-p/1571613
And even more discussions about it are available online.
I already shared all that into the following thread, but it looks like Intel won't help: https://community.intel.com/t5/Wireless/Intel-BE200-WiFi-7-M-2-card-causing-boot-failure-on-3/m-p/16...
At this step, difficult to say if the thing from the Intel BE200 that blocks AGESA based motherboard is coming from a broken/wrong approach from the Intel BE200 or from AMD AGESA Combo-AM4v2 (or even an unexpected thing that is nobody's fault), but any progress to debug this situation would be welcome.
