Encountering an extremely strange issue that is confusing me, and curious if anyone has any ideas.
On my FW13 (AI 7 350), my AX210 network card enters an unrecoverable fault state on every single resume from s2idle. The device remains visible to the PCI bus and addressable, but it stays in D3cold forever and ignores any attempts by the intel wifi driver to reinitialize it. The issue does not occur on the supported RZ616/Mediatek card, but this card does not work on my network which is why I’m using the AX210. I’ve never had issues with intel cards on linux before, which is why this is driving me crazy. Issue only began after upgrading mainboards; card is known-good!
What I’ve already tried:
- updating firmware.. on latest packaged by Arch, which is 89.735b75a4.0 ty-a0-gf-a0-89.ucode
- d3cold_allowed=0; this does nothing, it seems that the firmware forces all devices into d3cold on suspend
- turning off PCIe ASPM entirely; this outright breaks suspend/s2idle; the system just turns off instead of going into idle
- adding a suspend hook to unload iwlwifi before suspend; does nothing, which suggests to me that it’s something at a firmware level and not the driver level
- trying to nudge the card back aware after resume by removing it from the bus and rescanning; this causes it to just vanish entirely and never reappear on the bus until a complete power cycle (shutdown then power up, a warm reboot also resolves it)
- Windows; the issue remains present though I don’t know how to get any real logs out of Windows as I don’t use it regularly
- All the power saving options on iwlwifi.. current cocktail I’m using is amdsu_size=3, power_save=0, s0i3_disable=0
Symptoms:
- A seemingly-related spam of ‘GPIO 8 is active: 0x18141b00’
- A massive amount of iwlwifi log output
- the driver after resume gets stuck in a cycle of trying to reinitialize the card, which locks up the entire system for minutes after resume until it eventually gives up (after ~4m or so, in my experience)
- The abovementioned issue actually prevents complete recovery from s2idle - the system only fully resumes a couple minutes after something actually triggers return, during this period there is the abovementioned log spam. SysRq is ignored too (I suspect the keyboard is not powered on during this log spam)
- The system will recover from iwlwifi mess (I assume it gives up?) after precisely 4m30s, according amd-s2idle
THE WEIRDEST PART (to me at least): bluetooth is fine. Even after suspend and the five minute hang, and the card being just gone entirely from the PCI bus, the bluetooth segment of the card is still visible (and working) to bluetoothctl. ???
It seems that a nearly identical bug has been reported and fixed a couple times upstream in the kernel, but I have those patches.. I’m not confident enough in my troubleshooting abilities to call it a regression though, so I come to the community for help! I’ve seen other people on here with eerily similar issues that were resolved by updating Intel firmware. I’ve tried that, to no avail… hopefully someone’s got an idea. This is driving me crazy.
Thx for any ideas, y’all!
dmesg output: https://paste.c-net.org/WhichFreddy
Which Linux distro are you using? Arch, but I have also tested with the latest version of Ubuntu to similar results
Which release version?
Arch rolling release kept up-to-date
Which kernel are you using?
7.0.12
Which BIOS version are you using?
3.05
Which Framework Laptop 13 model are you using? (AMD Ryzen™ AI 300 Series, AMD Ryzen™ 7040 Series, Intel® Core™ Ultra Series 1, 13th Gen Intel® Core™ , 12th Gen Intel® Core™, 11th Gen Intel® Core™)
AI 300