NetBSD Problem Report #59922

From www@netbsd.org  Thu Jan 15 09:23:05 2026
Return-Path: <www@netbsd.org>
Received: from mail.netbsd.org (mail.netbsd.org [199.233.217.200])
	(using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits)
	 key-exchange X25519 server-signature RSA-PSS (2048 bits) server-digest SHA256
	 client-signature RSA-PSS (2048 bits) client-digest SHA256)
	(Client CN "mail.netbsd.org", Issuer "R13" (verified OK))
	by mollari.NetBSD.org (Postfix) with ESMTPS id 2BBDC1A923E
	for <gnats-bugs@gnats.NetBSD.org>; Thu, 15 Jan 2026 09:23:05 +0000 (UTC)
Message-Id: <20260115092303.EE9E71A9242@mollari.NetBSD.org>
Date: Thu, 15 Jan 2026 09:23:03 +0000 (UTC)
From: bartek.krawczyk@gmail.com
Reply-To: bartek.krawczyk@gmail.com
To: gnats-bugs@NetBSD.org
Subject: iwm0 - Intel Dual Band Wireless AC 8265 (rev. 0x78) - stopped working on NetBSD 11.0_BETA
X-Send-Pr-Version: www-1.0
X-From4GNATS: "bartek.krawczyk@gmail.com via gnats" <gnats-admin@NetBSD.org>

>Number:         59922
>Category:       kern
>Synopsis:       iwm0 - Intel Dual Band Wireless AC 8265 (rev. 0x78) - stopped working on NetBSD 11.0_BETA
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Jan 15 09:25:00 +0000 2026
>Last-Modified:  Sun Jan 25 19:40:01 +0000 2026
>Originator:     Bartlomiej Krawczyk
>Release:        11.0_BETA
>Organization:
>Environment:
NetBSD laptop 11.0_BETA NetBSD 11.0_BETA (GENERIC) #0: Thu Jan  8 10:53:08 UTC 2026  mkrepro@mkrepro.NetBSD.org:/usr/src/sys/arch/amd64/compile/GENERIC amd64
>Description:
After upgrading to 11.0_BETA using sysupgrade my iwm0 interface stopped working. I thought it may be a botched upgrade, so did a clean 11.0_BETA install from latest snapshot on 13th of Jan 2026.
In the installer, I could do "ifconfig iwm0 list scan" and it showed networks so I was hopeful. After finishing the installation though, the iwm0 interface stopped working. It was working correctly (albeit slow) on 10.1 release.

laptop# dmesg | grep iwm0
[     1.026015] iwm0 at pci3 dev 0 function 0: Intel Dual Band Wireless AC 8265 (rev. 0x78)
[     1.026015] iwm0: interrupting at msi1 vec 0
[     6.961864] iwm0: autoconfiguration error: failed to read nvm
[     1.038164] iwm0 at pci3 dev 0 function 0: Intel Dual Band Wireless AC 8265 (rev. 0x78)
[     1.038164] iwm0: interrupting at msi1 vec 0
[     4.271387] iwm0: hw rev 0x230, fw ver 22.361476.0, address 40:74:e0:51:38:58
[    19.361390] iwm0: autoconfiguration error: could not init bt coex (error 35)
[    19.421390] iwm0: autoconfiguration error: could not init bt coex (error 35)
[    30.651456] iwm0: autoconfiguration error: could not init bt coex (error 35)
[    30.721457] iwm0: autoconfiguration error: could not init bt coex (error 35)


laptop# ifconfig iwm0 list scan
ifconfig: SIOCS80211: Invalid argument


laptop# ifconfig iwm0
iwm0: flags=0x8803<UP,BROADCAST,SIMPLEX,MULTICAST> mtu 1500
        ssid "" nwkey 65536:"","","",""
        powersave off
        address: 40:74:e0:51:38:58
        media: IEEE802.11 autoselect
        status: no network
        inet6 fe80::4274:e0ff:fe51:3858%iwm0/64 flags 0x8<DETACHED> scopeid 0x3
        inet 169.254.70.202/16 broadcast 169.254.255.255 flags 0x4<DETACHED>
>How-To-Repeat:
Upgrade to 11.0_BETA from 10.1 on a Lenovo ThinkPad T480 using Intel Dual Band Wireless AC 8265 (rev. 0x78) wireless card
>Fix:

>Audit-Trail:
From: mlelstv@serpens.de (Michael van Elst)
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: kern/59922: iwm0 - Intel Dual Band Wireless AC 8265 (rev. 0x78) - stopped working on NetBSD 11.0_BETA
Date: Thu, 15 Jan 2026 10:07:28 -0000 (UTC)

 gnats-admin@NetBSD.org ("bartek.krawczyk@gmail.com via gnats") writes:

 >[     4.271387] iwm0: hw rev 0x230, fw ver 22.361476.0, address 40:74:e0:51:38:58
 >[    19.361390] iwm0: autoconfiguration error: could not init bt coex (error 35)

 >laptop# ifconfig iwm0
 >iwm0: flags=0x8803<UP,BROADCAST,SIMPLEX,MULTICAST> mtu 1500
 >        ssid "" nwkey 65536:"","","",""
 >        powersave off
 >        address: 40:74:e0:51:38:58
 >        media: IEEE802.11 autoselect
 >        status: no network


 There is hardly any difference between netbsd-10 and netbsd-11
 regarding the iwm driver and it is known to be pretty flaky.

 One possible difference is the interaction with dhcpcd. It's
 possible that without dhcpcd the interface associates to the network,
 and when you start dhcpcd aftwards, that it appears to function.
 At least that's one of the failure modes I've seen.

 In netbsd-current I have tried to fix iwm and it now seems to be
 stable for me and some others.

 You could try to boot a current kernel from the daily build
 (you may need to also install the modules set from current).

 Or if you are experienced, you could build a netbsd-11 kernel with
 the updated driver from netbsd-current.


From: "requiem." <rqm@0xde501a7a.net>
To: gnats-bugs@netbsd.org
Cc: kern-bug-people@netbsd.org, netbsd-bugs@netbsd.org,
	bartek.krawczyk@gmail.com
Subject: Re: kern/59922: iwm0 - Intel Dual Band Wireless AC 8265 (rev. 0x78)
 - stopped working on NetBSD 11.0_BETA
Date: Thu, 15 Jan 2026 11:23:38 +0000

 If it provides more context I had a similar issue on i386 with athn0 as wel=
 l as rtwn0. Please see the message here:

 https://marc.info/?l=3Dnetbsd-bugs&m=3D176817698721327&w=3D2

 I haven't found a fix outside of downgrading to 10.1.

 15 Jan 2026 10:10:22 Michael van Elst via gnats <gnats-admin@NetBSD.org>:

 > The following reply was made to PR kern/59922; it has been noted by GNATS=
 .
 >=20
 > From: mlelstv@serpens.de (Michael van Elst)
 > To: gnats-bugs@netbsd.org
 > Cc:
 > Subject: Re: kern/59922: iwm0 - Intel Dual Band Wireless AC 8265 (rev. 0x=
 78) - stopped working on NetBSD 11.0_BETA
 > Date: Thu, 15 Jan 2026 10:07:28 -0000 (UTC)
 >=20
 > gnats-admin@NetBSD.org ("bartek.krawczyk@gmail.com via gnats") writes:
 >=20
 >>[=C2=A0=C2=A0=C2=A0=C2=A0 4.271387] iwm0: hw rev 0x230, fw ver 22.361476.=
 0, address 40:74:e0:51:38:58
 >>[=C2=A0=C2=A0=C2=A0 19.361390] iwm0: autoconfiguration error: could not i=
 nit bt coex (error 35)
 >=20
 >>laptop# ifconfig iwm0
 >>iwm0: flags=3D0x8803<UP,BROADCAST,SIMPLEX,MULTICAST> mtu 1500
 >>=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 ssid "" nwkey 65536:"","","","=
 "
 >>=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 powersave off
 >>=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 address: 40:74:e0:51:38:58
 >>=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 media: IEEE802.11 autoselect
 >>=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 status: no network
 >=20
 >=20
 > There is hardly any difference between netbsd-10 and netbsd-11
 > regarding the iwm driver and it is known to be pretty flaky.
 >=20
 > One possible difference is the interaction with dhcpcd. It's
 > possible that without dhcpcd the interface associates to the network,
 > and when you start dhcpcd aftwards, that it appears to function.
 > At least that's one of the failure modes I've seen.
 >=20
 > In netbsd-current I have tried to fix iwm and it now seems to be
 > stable for me and some others.
 >=20
 > You could try to boot a current kernel from the daily build
 > (you may need to also install the modules set from current).
 >=20
 > Or if you are experienced, you could build a netbsd-11 kernel with
 > the updated driver from netbsd-current.
 >=20
 >=20

From: mlelstv@serpens.de (Michael van Elst)
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: kern/59922: iwm0 - Intel Dual Band Wireless AC 8265 (rev. 0x78) - stopped working on NetBSD 11.0_BETA
Date: Thu, 15 Jan 2026 12:42:16 -0000 (UTC)

 rqm@0xde501a7a.net ("requiem.") writes:

 >If it provides more context I had a similar issue on i386 with athn0 as wel=
 >l as rtwn0. Please see the message here:

 >https://marc.info/?l=3Dnetbsd-bugs&m=3D176817698721327&w=3D2


 It would be interesting, if the interface associates without
 dhcpcd and if it stays connected when then starting dhcpcd
 manually.


From: Bartek Krawczyk <bartek.krawczyk@gmail.com>
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: kern/59922: iwm0 - Intel Dual Band Wireless AC 8265 (rev. 0x78) -
 stopped working on NetBSD 11.0_BETA
Date: Thu, 15 Jan 2026 18:58:01 +0100

 On 1/15/26 11:10 AM, Michael van Elst via gnats wrote:
 >   
 >   There is hardly any difference between netbsd-10 and netbsd-11
 >   regarding the iwm driver and it is known to be pretty flaky.
 >   
 >   One possible difference is the interaction with dhcpcd. It's
 >   possible that without dhcpcd the interface associates to the network,
 >   and when you start dhcpcd aftwards, that it appears to function.
 >   At least that's one of the failure modes I've seen.

 I now disabled dhcpcd on boot, rebooted, and without dhcpcd it's also 
 not working with errors about loading firmware. I remember after my 
 sysupgrade from 10.1 t 11_BETA I did compare hashes of the firmware 
 between both versions and these were identical back then. This seems 
 like a regression? I did a full clean install yesterday.

 [     1.027605] iwm0 at pci3 dev 0 function 0: Intel Dual Band Wireless 
 AC 8265 (rev. 0x78)
 [     1.027605] iwm0: interrupting at msi1 vec 0
 [     5.468614] iwm0: hw rev 0x230, fw ver 22.361476.0, address 
 40:74:e0:51:38:58
 [    18.748664] iwm0: autoconfiguration error: fatal firmware error
 [    19.738678] iwm0: autoconfiguration error: could not load firmware 
 (error 35, ok 1)
 [    19.738678] iwm0: autoconfiguration error: cpu1 status: 0xa5a5a5a0
 [    19.738678] iwm0: autoconfiguration error: cpu2 status: 0xa5a5a5a0
 [    19.738678] iwm0: autoconfiguration error: could not load firmware 
 (error 35)
 [    23.748724] iwm0: autoconfiguration error: could not load firmware 
 (error 35)
 [    42.298930] iwm0: autoconfiguration error: could not init bt coex 
 (error 35)
 [    42.368931] iwm0: autoconfiguration error: could not init bt coex 
 (error 35)


 >   In netbsd-current I have tried to fix iwm and it now seems to be
 >   stable for me and some others.
 >   
 >   You could try to boot a current kernel from the daily build
 >   (you may need to also install the modules set from current).
 >   
 >   Or if you are experienced, you could build a netbsd-11 kernel with
 >   the updated driver from netbsd-current.

 Thx for a hint, I'll try this.

 Regards,
 Bart

From: "requiem." <rqm@0xde501a7a.net>
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: kern/59922: iwm0 - Intel Dual Band Wireless AC 8265 (rev. 0x78)
 - stopped working on NetBSD 11.0_BETA
Date: Fri, 16 Jan 2026 00:08:18 +0000

 On Thu, 15 Jan 2026 12:45:01 +0000 (UTC)
 "Michael van Elst via gnats" <gnats-admin@NetBSD.org> wrote:

 > The following reply was made to PR kern/59922; it has been noted by
 > GNATS.
 > 
 > From: mlelstv@serpens.de (Michael van Elst)
 > To: gnats-bugs@netbsd.org
 > Cc: 
 > Subject: Re: kern/59922: iwm0 - Intel Dual Band Wireless AC 8265
 > (rev. 0x78) - stopped working on NetBSD 11.0_BETA Date: Thu, 15 Jan
 > 2026 12:42:16 -0000 (UTC)
 > 
 >  rqm@0xde501a7a.net ("requiem.") writes:
 >  
 >  >If it provides more context I had a similar issue on i386 with
 >  >athn0 as wel= l as rtwn0. Please see the message here:  
 >  
 >  >https://marc.info/?l=3Dnetbsd-bugs&m=3D176817698721327&w=3D2  
 >  
 >  
 >  It would be interesting, if the interface associates without
 >  dhcpcd and if it stays connected when then starting dhcpcd
 >  manually.

 It does not seem to be able to keep an association. It seems that
 it goes into a 4-way handshake and drops the connection around stage 3.
 Ran wpa_supplicant manually without dhcpcd running.

 I will send some
 extra logs to the original bug report so as not to take this one off
 course, as I was only trying to supply additional info :)

From: Bartek Krawczyk <bartek.krawczyk@gmail.com>
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: kern/59922: iwm0 - Intel Dual Band Wireless AC 8265 (rev. 0x78) -
 stopped working on NetBSD 11.0_BETA
Date: Sat, 17 Jan 2026 17:17:56 +0100

 On 1/15/26 11:10, Michael van Elst via gnats wrote:
 >   Or if you are experienced, you could build a netbsd-11 kernel with
 >   the updated driver from netbsd-current.

 I tried that:

 laptop$ cvs status sys/dev/pci/if_iwm*.*
 ===================================================================
 File: if_iwm.c          Status: Up-to-date

     Working revision:    1.93
     Repository revision: 1.93    /cvsroot/src/sys/dev/pci/if_iwm.c,v
     Commit Identifier:   38b3fwqarjB0jinG
     Sticky Tag:          1.93
     Sticky Date:         (none)
     Sticky Options:      (none)

 ===================================================================
 File: if_iwmreg.h       Status: Up-to-date

     Working revision:    1.15
     Repository revision: 1.15    /cvsroot/src/sys/dev/pci/if_iwmreg.h,v
     Commit Identifier:   6C6fODszsDzcLIkG
     Sticky Tag:          1.15
     Sticky Date:         (none)
     Sticky Options:      (none)

 ===================================================================
 File: if_iwmvar.h       Status: Up-to-date

     Working revision:    1.21
     Repository revision: 1.21    /cvsroot/src/sys/dev/pci/if_iwmvar.h,v
     Commit Identifier:   6C6fODszsDzcLIkG
     Sticky Tag:          1.21
     Sticky Date:         (none)
     Sticky Options:      (none)

 and it doesn't complain about not being able to load firmware anymore, 
 but still doesn't connect:

 [     1.039448] acpiwmi0 at acpi0 (WTBT, PNP0C14-TBFP): ACPI WMI Interface
 [     1.039448] acpiwmibus at acpiwmi0 not configured
 [     1.039448] acpiwmi1 at acpi0 (WMI1, PNP0C14-1): ACPI WMI Interface
 [     1.039448] acpiwmibus at acpiwmi1 not configured
 [     1.039448] acpiwmi2 at acpi0 (WMI2, PNP0C14-2): ACPI WMI Interface
 [     1.039448] acpiwmibus at acpiwmi2 not configured
 [     1.039448] acpiwmi3 at acpi0 (WMI3, PNP0C14-3): ACPI WMI Interface
 [     1.039448] acpiwmibus at acpiwmi3 not configured
 [     1.039448] iwm0 at pci3 dev 0 function 0: Intel Dual Band Wireless 
 AC 8265 (rev. 0x78)
 [     1.039448] iwm0: interrupting at msi2 vec 0
 [     5.369359] iwm0: hw rev 0x230, fw ver 22.361476.0, address 
 40:74:e0:51:38:58
 [    20.489435] iwm0: could not initiate scan (error = 35)
 [    27.449516] iwm0: could not initiate scan (error = 35)
 [    34.389596] iwm0: could not initiate scan (error = 35)
 [    41.339677] iwm0: could not initiate scan (error = 35)
 [    52.289803] iwm0: could not initiate scan (error = 35)
 [    95.950309] iwm0: could not initiate scan (error = 35)
 [   253.142128] iwm0: could not initiate scan (error = 35)

 ifconfig iwm0 list scan works though.

 Regards,
 Bart

From: mlelstv@serpens.de (Michael van Elst)
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: kern/59922: iwm0 - Intel Dual Band Wireless AC 8265 (rev. 0x78) - stopped working on NetBSD 11.0_BETA
Date: Sat, 17 Jan 2026 16:57:27 -0000 (UTC)

 gnats-admin@NetBSD.org ("Bartek Krawczyk via gnats") writes:

 > [    20.489435] iwm0: could not initiate scan (error = 35)

 Do you use dhcpcd on that interface?

 If yes, please try with the following added to dhcpcd.conf.

 interface iwm*
   nolink


From: Bartek Krawczyk <bartek.krawczyk@gmail.com>
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: kern/59922: iwm0 - Intel Dual Band Wireless AC 8265 (rev. 0x78) -
 stopped working on NetBSD 11.0_BETA
Date: Tue, 20 Jan 2026 23:15:15 +0100

 On 1/17/26 18:00, Michael van Elst via gnats wrote:
 >   Do you use dhcpcd on that interface?
 >   
 >   If yes, please try with the following added to dhcpcd.conf.
 >   
 >   interface iwm*
 >     nolink
 >   
 >   

 I can now reliably get "ifconfig iwm0 list scan" I think, but I still 
 get this in dmesg:

 iwm0: autoconfiguration error: could not init bt coex (error 35)

 Also wpa_supplicant is not able to work with iwm0:

 messages:Jan 20 23:14:30 laptop wpa_supplicant[2716]: ioctl[SIOCS80211, 
 op=23, val=0, arg_len=0]: Invalid argument
 messages:Jan 20 23:14:30 laptop wpa_supplicant[2716]: iwm0: 
 CTRL-EVENT-SCAN-FAILED ret=-1 retry=1

 Regards,
 Bart

From: mlelstv@serpens.de (Michael van Elst)
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: kern/59922: iwm0 - Intel Dual Band Wireless AC 8265 (rev. 0x78) - stopped working on NetBSD 11.0_BETA
Date: Wed, 21 Jan 2026 00:17:26 -0000 (UTC)

 gnats-admin@NetBSD.org ("Bartek Krawczyk via gnats") writes:

 > I can now reliably get "ifconfig iwm0 list scan" I think, but I still 
 > get this in dmesg:
 > 
 > iwm0: autoconfiguration error: could not init bt coex (error 35)

 That should mean that the firmware isn't answering (probably because
 it halted for some unknown error condition). The error also breaks
 the interface initialization, so it doesn't go into RUNNING state.


 > Also wpa_supplicant is not able to work with iwm0:
 > 
 > messages:Jan 20 23:14:30 laptop wpa_supplicant[2716]: ioctl[SIOCS80211, 
 > op=23, val=0, arg_len=0]: Invalid argument

 That's a scan request that fails because the interface
 isn't UP and RUNNING.


 You could try to #ifdef out the failing call. It tries to configure
 the chip for Wifi priority over Bluetooth but I doubt that strictly
 necessary for it to function.

 E.g.:

 --- if_iwm.c    21 Dec 2025 16:24:39 -0000      1.93
 +++ if_iwm.c    21 Jan 2026 00:12:30 -0000
 @@ -7651,12 +7651,14 @@ iwm_init_hw(struct iwm_softc *sc)
                 goto err;
         }

 +#if 0
         err = iwm_send_bt_init_conf(sc);
         if (err) {
                 aprint_error_dev(sc->sc_dev,
                     "could not init bt coex (error %d)\n", err);
                 goto err;
         }
 +#endif

         if (isset(sc->sc_enabled_capa,
             IWM_UCODE_TLV_CAPA_SOC_LATENCY_SUPPORT)) {

From: Bartek Krawczyk <bartek.krawczyk@gmail.com>
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: kern/59922: iwm0 - Intel Dual Band Wireless AC 8265 (rev. 0x78) -
 stopped working on NetBSD 11.0_BETA
Date: Sun, 25 Jan 2026 20:36:48 +0100

 On 1/21/26 01:20, Michael van Elst via gnats wrote:
 >   That should mean that the firmware isn't answering (probably because
 >   it halted for some unknown error condition). The error also breaks
 >   the interface initialization, so it doesn't go into RUNNING state.
 >   
 >   
 >   > Also wpa_supplicant is not able to work with iwm0:
 >   >
 >   > messages:Jan 20 23:14:30 laptop wpa_supplicant[2716]: ioctl[SIOCS80211,
 >   > op=23, val=0, arg_len=0]: Invalid argument
 >   
 >   That's a scan request that fails because the interface
 >   isn't UP and RUNNING.
 >   
 >   
 >   You could try to #ifdef out the failing call. It tries to configure
 >   the chip for Wifi priority over Bluetooth but I doubt that strictly
 >   necessary for it to function.
 >   
 >   E.g.:
 >   
 >   --- if_iwm.c    21 Dec 2025 16:24:39 -0000      1.93
 >   +++ if_iwm.c    21 Jan 2026 00:12:30 -0000
 >   @@ -7651,12 +7651,14 @@ iwm_init_hw(struct iwm_softc *sc)
 >                   goto err;
 >           }
 >    
 >   +#if 0
 >           err = iwm_send_bt_init_conf(sc);
 >           if (err) {
 >                   aprint_error_dev(sc->sc_dev,
 >                       "could not init bt coex (error %d)\n", err);
 >                   goto err;
 >           }
 >   +#endif
 >    
 >           if (isset(sc->sc_enabled_capa,
 >               IWM_UCODE_TLV_CAPA_SOC_LATENCY_SUPPORT)) {
 >   

 Thanks for the hints,
 I tried that but unfortunately no change. With or without dhcpcd 
 running, it doesn't connect to ssid:
 [   406.034465] iwm0: could not initiate scan (error = 35)
 [   412.984545] iwm0: could not initiate scan (error = 35)
 [   419.924626] iwm0: could not initiate scan (error = 35)
 [   429.864741] iwm0: could not initiate scan (error = 35)

 ifconfig iwm0 list scan still works. Restarting wpa_supplicant (with or 
 without dhcpcd) doesn't help.

 Regards
 Bart

Home	PR Database Search