Repository navigation
Intel 7265 (iwlwifi, firmware 7265D-29): firmware asserts on every ad-hoc start, exclusive mode can't run #16
Description
A card report. The iwlwifi profile line is tested exclusive, so this card runs without opt-in, and every start crashes its firmware.
Setup
- Intel Dual Band Wireless-AC 7265 (
8086:095a, subsystem5110, rev 59),iwlwifi/iwlmvm, firmware29.9ef079ed.0 7265D-29.ucode(the newest the driver requests for this card) - Asus Zenbook UX305CA, Omarchy 4.0.4, kernel
7.2.5-3-omarchy, NetworkManager 1.58.1, wpa_supplicant 2.12, iw 6.17 - omdrop-owl at ee6a57d built with
makepkg -si; pluginnetmojo.omdrop0.16.1 - Profile:
iwlwifi exclusive station no no iw phylists IBSS and "Device supports HT-IBSS"; 2437 MHz is clear, 5220 and 5745 are "No IR"
What happens
The ad-hoc vif comes up beside the station and the join is accepted. About 6 ms later the firmware asserts, mac80211's restart hits the same assert, and the driver reprobes the device:
awdlibss0: Created IBSS using preconfigured BSSID 00:25:00:ff:94:73
iwlwifi 0000:01:00.0: Microcode SW error detected. Restarting 0x2000000.
iwlwifi 0000:01:00.0: FW error in SYNC CMD MAC_CONTEXT_CMD
iwlwifi 0000:01:00.0: 0x0000102C | ADVANCED_SYSASSERT
iwlwifi 0000:01:00.0: 0x0000020A | data1
iwlwifi 0000:01:00.0: Failed to send MAC_CONTEXT_CMD (action:2): -5
ieee80211 phy0: Hardware restart was requested
iwlwifi 0000:01:00.0: Microcode SW error detected. Restarting 0x2000000.
iwlwifi 0000:01:00.0: 0x0000102C | ADVANCED_SYSASSERT
iwlwifi 0000:01:00.0: Device error - reprobe!
iwlwifi 0000:01:00.0 mon0 (unregistering): left promiscuous mode
Three of three starts (twice from tools/test-card.sh, once from the plugin). The call trace at the failing command is iwl_mvm_bss_info_changed -> iwl_mvm_mac_ctxt_cmd_ibss. The reprobe removes awdlibss0 and mon0, and NetworkManager re-associates the new station about 4 s later.
It isn't the helper's parameters
The same sequence as exclusive_up with a plain cell gives the same assert, same code and data1:
iw phy <phy> interface add awdlibss0 type ibss # station down and unmanaged first
ip link set awdlibss0 address <station MAC>; ip link set awdlibss0 up
iw dev awdlibss0 ibss join omdrop 2437 NOHT fixed-freq 02:11:22:33:44:55 beacon-interval 100
So does converting the station in place (iw dev <sta> set type ibss, same join). Two of two. The P2P device was already gone when the cell came up (nl80211: deinit ifname=p2p-dev-... a second earlier). I did not test an older kernel, so I can't say whether this is a firmware limit or an iwlmvm regression.
The helper reported success once
On the first run OWL opened mon0 before the reprobe removed it, so awdl0 existed and start returned 0. For three minutes status said visible while the card was back on the access point:
omdrop-owl card test, 2026-10-11T21:31:11-04:00
card: Network controller: Intel Corporation Wireless 7265 (rev 59)
driver: iwlwifi, kernel 7.2.5-3-omarchy, Omarchy
profile: iwlwifi exclusive station no no
probe: ready
baseline: 10/10 replies
start: ok in 872 ms
phone heard (OWL synced): no, in 180 s
phone -> computer: skipped (omdrop's receiver isn't running for <user>)
Wi-Fi during the window: 0/0 replies, longest gap 0 s
stop: exit 0 in 162 ms
Wi-Fi after teardown: 24/24 replies, longest gap 0 s
When the reprobe wins the race, OWL exits with Could not get link: No such device:
omdrop-owl card test, 2026-10-11T21:32:09-04:00
card: Network controller: Intel Corporation Wireless 7265 (rev 59)
driver: iwlwifi, kernel 7.2.5-3-omarchy, Omarchy
profile: iwlwifi exclusive station no no
probe: ready
baseline: 10/10 replies
start: exit 4 after 12659 ms: 01:32:09Z error: AWDL didn't come up: OWL exited
Suggestions
- List the 7265 as not working, and consider not letting older iwlwifi cards inherit
testedfrom the AX2xx reports.iwlwificovers firmware generations that behave differently here. - After
start, check that the cell andmon0still exist before reporting visible. A device reprobe removes both whileawdl0and the OWL process survive.
Activity
t4t5 commentedon Oct 16, 2026
Thanks for the detailed report. The 7265 is now listed as not working in the README (e6b8cd9). #29 stops the window with radio-lost when mon0 disappears, so a reprobe no longer leaves status saying visible.
You're right that
iwlwificovers firmware generations that behave differently. Profiles only match on the driver for now, so the 7265 and 8260 still start without opt-in. I'll keep this open until older cards fall back to untested.t4t5 commentedon Oct 23, 2026
Owly (0.4.0) now matches cards by PCI ID as well as by driver, and lists the 7265 (
8086:095a) as not working. It won't start on this card unless you setallow_untested = truein/etc/owly/radio.toml, and when it refuses it says why, with a link here. The 8260 and 8265 are listed the same way.Your second point is covered too: Owly ends the window when the radio's interfaces disappear, so a firmware crash can no longer leave it saying the computer is visible.
Closing, since there's nothing left to do on our side. If a newer firmware or kernel stops the crash, please reopen. Thanks again for such a thorough report!