Re: iwlwifi devices disappear after suspend on kernel 5.17

["Toke Høiland-Jørgensen" <toke@toke.dk>]

"Grumbach, Emmanuel" <emmanuel.grumbach@intel.com> writes:

> removing Luca
>
> On Wed, 2022-04-20 at 11:49 +0200, Toke Høiland-Jørgensen wrote:
>
>
> <snip>
>
>> > > 
>> > 
>> > So you do get a SAP connection. Weird.
>> > Checking with the CSME guys.
>> > Can you get dynamic debug prints for iwlmei module?
>> > Best would be tracing with -e iwlmei_sap_cmd
>> 
>> Sure! Recorded these while doing a suspend-restore, then restarting iwd:
>> 
>>   kworker/u16:34-34555 [004] 46452.506238: iwlmei_sap_cmd:       sap_cmd Tx: type 1010 len 8 seq
>> 11
>>   kworker/u16:34-34555 [004] 46452.506243: iwlmei_me_msg:        ME message: Tx: type 3 seq 8
>>   kworker/u16:34-34555 [004] 46452.559407: iwlmei_sap_cmd:       sap_cmd Tx: type 1001 len 4 seq
>> 12
>>   kworker/u16:14-33311 [004] 46452.573077: iwlmei_sap_cmd:       sap_cmd Tx: type 1006 len 0 seq
>> 13
>>   kworker/u16:14-33311 [004] 46452.573078: iwlmei_me_msg:        ME message: Tx: type 3 seq 10
>>      kworker/4:0-34305 [004] 46452.706737: iwlmei_me_msg:        ME message: Tx: type 3 seq 11
>>    kworker/u16:8-34462 [000] 46457.939453: iwlmei_me_msg:        ME message: Tx: type 1 seq 1
>>      kworker/4:2-32196 [004] 46457.940910: iwlmei_me_msg:        ME message: Rx: type 2 seq 81
>>      kworker/4:2-32196 [004] 46457.940912: iwlmei_sap_cmd:       sap_cmd Tx: type 1013 len 0 seq 1
>>      kworker/4:2-32196 [004] 46457.940913: iwlmei_me_msg:        ME message: Tx: type 3 seq 2
>>      kworker/4:2-32196 [004] 46457.942056: iwlmei_me_msg:        ME message: Rx: type 3 seq 82
>>      kworker/4:2-32196 [004] 46457.942060: iwlmei_sap_cmd:       sap_cmd Rx: type 502 len 4 seq 1
>>      kworker/4:2-32196 [004] 46458.042205: iwlmei_me_msg:        ME message: Rx: type 3 seq 83
>>      kworker/4:2-32196 [004] 46458.042208: iwlmei_sap_cmd:       sap_cmd Rx: type 500 len 1896 seq
>> 2
>>      kworker/4:2-32196 [004] 46458.042209: iwlmei_sap_cmd:       sap_cmd Rx: type 512 len 64 seq 3
>>      kworker/4:2-32196 [004] 46458.042214: iwlmei_sap_cmd:       sap_cmd Rx: type 502 len 4 seq 4
>>              iwd-35396 [003] 46467.085642: iwlmei_sap_cmd:       sap_cmd Tx: type 1008 len 8 seq 2
>>              iwd-35396 [003] 46467.085645: iwlmei_me_msg:        ME message: Tx: type 3 seq 3
>>              iwd-35396 [003] 46467.102818: iwlmei_sap_cmd:       sap_cmd Tx: type 1008 len 8 seq 3
>>              iwd-35396 [003] 46467.103369: iwlmei_sap_cmd:       sap_cmd Tx: type 1016 len 20 seq
>> 4
>>              iwd-35396 [003] 46467.103842: iwlmei_sap_cmd:       sap_cmd Tx: type 1001 len 4 seq 5
>>              iwd-35396 [003] 46467.104689: iwlmei_sap_cmd:       sap_cmd Tx: type 1002 len 12 seq
>> 6
>
> can you please send the content of tracing unparsed?
> There is more binary data that I can use.

Sure! I'll send you the trace.dat file off-list (it's almost 10 MB).

> I definitely need to talk to the CSME guys, I can see that CSME is
> alive even if you have AMT disabled.

What is this CSME thing, anyway? And does this imply that the AMT is not
really disabling itself (there's also a "permanently disable" option in
BIOS that I have not dared pick..)

> Problem: it's holiday season here.

No worries, I can live with working around the issue in the meantime; I
guess I can get systemd to automatically restart iwd when returning from
suspend if I can just figure out the right incantation :)

-Toke