Intel Wi-Fi 6 AX200 frequent disconnects on 5GHz/AX networks with "Microcode SW error" (2.4GHz works fine)

Ask Fedora
, , , , , ,
1

Hello,

I’m experiencing persistent connection drops with my Intel Wi-Fi 6 AX200 card, but only when connected to 5GHz (Wi-Fi 5/AC and Wi-Fi 6/AX) networks. The connection is perfectly stable on 2.4GHz networks. The disconnects on 5GHz happen frequently, especially under load, and the kernel log shows Queue X is stuck followed by a Microcode SW error detected, forcing a hardware restart.

Hardware and Software Environment

  • Motherboard: ASUSTeK COMPUTER INC. PRIME A320M-K
  • BIOS: American Megatrends Inc. Version 6253 (Release Date: 09/10/2025)
  • CPU: AMD Ryzen 5 5500
  • Wi-Fi Adapter:
    07:00.0 Network controller [0280]: Intel Corporation Wi-Fi 6 AX200 [8086:2723] (rev 1a)
    Subsystem: Intel Corporation Wi-Fi 6 AX200NGW [8086:0084]
    Kernel driver in use: iwlwifi
  • OS: Fedora 43 Silverblue (Pre-release)
  • Kernel: 6.17.1-300.fc43.x86_64
  • Kernel Parameters currently active:
    rd.luks.uuid=... rhgb quiet ... nvidia-drm.modeset=1 iwlwifi.11n_disable=8 power_scheme=1 ...
  • Router Model: TP-Link AX1800

Steps to Reproduce

  1. Connect to a 5GHz Wi-Fi network (either 802.11ac or 802.11ax).
  2. Start any network-intensive task (e.g., download a large file, stream a 4K video).
  3. Within minutes, the Wi-Fi connection drops and the kernel logs the error.
  4. The connection remains stable indefinitely on the 2.4GHz network from the same router.

Troubleshooting Steps Attempted

I have tried an extensive list of potential solutions without success:

  • Upgraded from Fedora 42 to the pre-release Fedora 43 to get the latest kernel and firmware. This updated the firmware from 77.864baa2e.0 to 77.6eaf654b.0, but the issue persists.
  • Updated the motherboard BIOS to the latest version (6253).
  • Tested multiple kernel parameters for the iwlwifi module, including 11n_disable=8, bt_coex_active=0, power_save=0, uapsd_disable=1, and pcie_aspm=off. None solved the issue. The only parameter that makes the connection stable is 11n_disable=1, but it drops the speed to an unusable ~25 Mbps.
  • Confirmed that the linux-firmware package in Fedora provides the same firmware file version as the latest one on git.kernel.org.
  • Modified router settings: disabled AX mode (forcing AC), disabled OFDMA, set channel width to 80MHz, and tried fixed non-DFS channels (36, 40, 44, 48).

Kernel Logs

Here is the relevant dmesg output captured at the moment of failure on the latest Fedora 43 kernel and firmware:

[  +0,130404] wlp7s0: authenticate with 40:ae:30:79:57:35 (local address=f6:d1:df:82:89:a8)
[  +0,000743] wlp7s0: send auth to 40:ae:30:79:57:35 (try 1/3)
[  +0,030916] wlp7s0: authenticated
[  +0,000929] wlp7s0: associate with 40:ae:30:79:57:35 (try 1/3)
[  +0,009590] wlp7s0: RX AssocResp from 40:ae:30:79:57:35 (capab=0x1811 status=0 aid=20)
[  +0,006116] wlp7s0: associated
[  +0,037783] wlp7s0: Limiting TX power to 30 (30 - 0) dBm as advertised by 40:ae:30:79:57:35
[ +11,610853] iwlwifi 0000:07:00.0: Queue 3 is stuck 0 17
[  +0,000107] iwlwifi 0000:07:00.0: Microcode SW error detected. Restarting 0x0.
[  +0,000105] iwlwifi 0000:07:00.0: Start IWL Error Log Dump:
[  +0,000002] iwlwifi 0000:07:00.0: Transport status: 0x0000004A, valid: 6
[  +0,000003] iwlwifi 0000:07:00.0: Loaded firmware version: 77.6eaf654b.0 cc-a0-77.ucode
[  +0,000003] iwlwifi 0000:07:00.0: 0x00000084 | NMI_INTERRUPT_UNKNOWN        
[  +0,000002] iwlwifi 0000:07:00.0: 0x00802AF0 | trm_hw_status0
[  +0,000002] iwlwifi 0000:07:00.0: 0x00000000 | trm_hw_status1
[  +0,000002] iwlwifi 0000:07:00.0: 0x004F8F22 | branchlink2
[  +0,000002] iwlwifi 0000:07:00.0: 0x004EEF72 | interruptlink1
[  +0,000002] iwlwifi 0000:07:00.0: 0x004EEF72 | interruptlink2
... (and the rest of the error dump) ...
[  +0,670484] iwlwifi 0000:07:00.0: Device error - SW reset
[  +0,000159] ieee80211 phy0: Hardware restart was requested
[  +0,231198] iwlwifi 0000:07:00.0: restart completed

Has anyone else on Fedora 42/43 with an AX200 card encountered this? I’ve tried extensive troubleshooting and plan to file a formal bug report, but wanted to check with the community first for any other ideas. Thanks!