WiFi hang requiring hard reset (occurred 2x)

MediaTek WiFi mt7921e (MT7921/Filogic 330) misbehaves: the WiFi stays "connected" but the device loses internet access; toggling WiFi has no effect; and a normal reboot hangs/freezes during shutdown — never completes. The only way out is a HARD RESET (hold power button). If left alone, it can escalate into a kernel panic. Happened 2 times.

Symptoms:

1. WiFi shows connected but no internet access while using the system
2. WiFi on/off toggle has no effect (chip is hung)
3. Restart/reboot via the normal method freezes / hangs during shutdown — never reaches login
4. Hard reset (hold the power button) → recovers, WiFi works again
5. Without a hard reset it can escalate into an actual kernel panic

Evidence (logs from boot before the event, -b1):

mt7921e 0000:03:00.0: Timeout for driver own       (repeats 09:45–09:47; x26)

mt7921e 0000:03:00.0: driver own failed           (repeats 09:45–09:47; x30)

mt7921e 0000:03:00.0: chip reset failed           (x2)

mt7921e 0000:03:00.0: Message 00020003 (seq N) timeout   (repeats; x10)

Hardware:

• 03:00.0 Network controller [0280]: MEDIATEK Corp. MT7921 802.11ax PCIe Wireless Network Adapter [Filogic 330] [14c3:7961]

Software:

• Kernel: 7.2.4.p03.31-4.fc44.x86_64
• Driver: mt7921e
• Firmware: linux-firmware-20260910-1.fc44 (MT7961 firmware present)

Alright thanks will take look at this make sure we are not missing anything on the images for the chip

Possible lead — likely upstream MediaTek mt7921e driver bug

While digging, this doesn't look RakuOS-specific — it's a known upstream kernel/driver issue with MT7921/Filogic 330 chips.

To confirm matching cases: lspci -nn shows 14c3:7961 (MT7921), and the crashed boot's journal repeats driver own failed / Timeout for driver own / chip reset failed — exactly matching known open reports upstream.

References:

• Kernel Bugzilla #215391 — same issue since 2021
• Kernel Bugzilla #220353 — mt7921e driver own failed
• linux-mediatek mailing list (2026) — identical symptoms; earlier patch changed a panic into a hang, so root cause is still open
• Arch forum #312249 — here the culprit was the linux-firmware-mediatek package (firmware, not kernel)
• Arch wiki/forums document mt7921e.disable_aspm=Y as an official MediaTek module param

Caveat: reports suggest the ASPM workaround is sometimes incomplete on newer 7.x kernels — so it may narrow down the cause rather than fully fix it. Happy to provide more logs if helpful.

============================================

New Report  — WiFi fully hangs (driver mt7921e)

============================================

• Symptom

WiFi dies on its own, cannot connect, restarting the service doesn't help, the chip becomes unresponsive → requires a force reboot.

• How to reproduce (verification commands)

1. List boot sessions to find the one before the forced reboot

journalctl --list-boots

→  IDX   BOOT ID                          FIRST                LAST

-2   15ef5dae…  Fri 2026-09-18 16:33:23  Fri 2026-09-18 16:35:58   (clean)

-1   b16f2cce…  Sat 2026-09-19 09:11:15  Sat 2026-09-19 10:26:30   ← HANG (this boot)

0   961f4df9…  Sat 2026-09-19 10:27:41  Sat 2026-09-19 10:51:02   (current, after force reboot)

2. Count hang events in that boot

journalctl -b -1 --no-pager | grep -cE "driver own failed|chip reset failed"

→ result: 96 error lines (severe hang)

3. Show the timeline

journalctl -b -1 --no-pager | grep -E "mt7921e.*(timeout|fail|reset|error)"

Log evidence (boot 19/09, hang started 10:21:04)

09:11:29 kernel: mt7921e ...: WM Firmware Version: ____010000, Build Time: 20260224110949

10:21:04 kernel: mt7921e 0000:03:00.0: driver own failed

10:21:05 kernel: mt7921e 0000:03:00.0: Timeout for driver own

10:21:24 kernel: mt7921e 0000:03:00.0: chip reset failed

10:21:27 kernel: mt7921e 0000:03:00.0: Message 00020001 (seq 10) timeout

10:26:14 kernel: mt7921e 0000:03:00.0: Message 00002ced (seq 4) timeout

10:26:25 kernel: mt7921e 0000:03:00.0: chip reset failed

10:26:30 kernel: mt7921e 0000:03:00.0: Message 00020001 (seq 6) timeout

(Total 96 lines: driver own failed + Timeout for driver own + Message ... timeout.)

• Analysis

1. No user-space trigger: the 60 seconds before the hang are silent (no power/PCIe/ACPI/NetworkManager errors).
2. driver own failed = the driver cannot take control of the chip over PCIe; chip reset failed = not even the chip reset works → the chip is fully hung (firmware/hardware hang).
3. disable_aspm=1 was already active (in /etc/modprobe.d/mt7921e-aspm.conf) when the hang occurred → not ASPM-related, not user configuration related.
4. Firmware loaded: "WM Firmware ____010000 (Build 20260224)" from /lib/firmware/mediatek/WIFI_RAM_CODE_MT7961_1.bin.xz, shipped by linux-firmware-20260910 (newer than the log's build timestamp) → the hang is not caused by a stale/failed firmware load.

• Conclusion

Known upstream mt7921e driver bug — same "Timeout for driver own"/"driver own failed"/"chip reset failed" pattern reported and still open since 2021 (Kernel Bugzilla #215391 & #220353, linux-mediatek 2026 thread). Firmware verified OK (all required binaries present, clean load, ~1h runtime before the hang; a Fedora 43 user with identical firmware still hangs). Not a driver–firmware mismatch — a runtime firmware hang in the mt7921e driver. → Needs handling in the base kernel (P03).