From: James Prestwood <prestwoj@gmail.com>
To: Mauro Condarelli <mc5686@mclink.it>, iwd@lists.linux.dev
Subject: Re: Setup help request
Date: Tue, 2 Jan 2024 06:40:32 -0800 [thread overview]
Message-ID: <bf00447d-9817-4bd9-ba67-149356d6ecd6@gmail.com> (raw)
In-Reply-To: <2aaed79f-9fae-4750-8e85-994a48a45da1@mclink.it>
Hi Mauro,
On 1/2/24 6:07 AM, Mauro Condarelli wrote:
> Thanks James,
> comments inline below.
>
> On 1/2/24 14:01, James Prestwood wrote:
>> Hi Mauro,
>>
>> On 1/1/24 4:34 PM, Mauro Condarelli wrote:
>>> Hi James,
>>> Thanks for the fast reply.
>>>
>>> Comments inline below.
>>>
>>> On 1/2/24 00:03, James Prestwood wrote:
>>>> Hi Mauro,
>>>>
>>>> On 1/1/24 11:26 AM, Mauro Condarelli wrote:
>>>>> Hi,
>>>>> I just installed `IWD version 2.3` on a small MiniPC (Beelink, if
>>>>> it matters) sporting a
>>>>> `Intel Corporation Wireless 3165 (rev 81)` PCI Wireless adapter.
>>>>> O.S. is a fairly up-to-date "Debian GNU/Linux 12 (bookworm)"
>>>>> Installation was done via standard Debian package (`apt install iwd`)
>>>>> ======8<----------- SNIP
>>> I used the following script to reproducibly start IWD and friends:
>>> ---------------------------------
>>> root@lxd:~# cat ./start_iwd.sh
>>> #!/bin/bash
>>>
>>> /usr/libexec/iwd -d >/tmp/iwd.log 2>&1 &
>>> IWD_PID=$!
>>>
>>> sleep 5
>>> iwctl device list
>>> iwctl device wlan0 set-property Mode ap
>>> iwctl ap wlan0 start-profile beelink
>>> iwctl ap wlan0 show
>>>
>>> iwmon --write /tmp/iwmon.pcap >/tmp/iwmon.log 2>&1 &
>>> IWMON_PID=$!
>>>
>>> mc
>>>
>>> kill $IWMON_PID
>>> kill $IWD_PID
>>> ---------------------------------
>>>
>>> While shell was active I:
>>> - opened Fritz! `WLAN` app
>>> - it took ~30s to see a `beelink` network apparently spanning
>>> several channels
>>> - tried to connect from my desktop
>>> - it did see a `beelink` AP a long time (>2m) after `WLAN` app saw
>>> it.
>>> - it tried several times to connect.
>>> - it never succeeded.
>>> - it didn't even ask for a passphrase.
>>> - tried to connect using my Android Phone (the ssame running `WLAN`.
>>> - it never saw a `beelink` AP to attempt connection.
>>>
>>> I attach all logs I took.
>>> I changed password to "temp-password" so it can be shared as needed.
>>
>> I think there is more going on but for starters it appears whatever
>> app your using to scan is sending an invalid probe request frame.
>> This causes IWD to ignore it which seems like the reason you cant
>> even see the AP in the scans. I've attached a patch you can test, but
>> I'm not sure this is something we can really upstream since its out
>> of spec. The concerning IE is Extended capabilities, this is
>> something IWD actually uses and parses, so the fact there is a
>> duplicate and each one is different makes it impossible to know which
>> one we should use. The probe request is including multiple entries
>> for "Extended Capabilities" and "FILS Request Parameters (tag 258)".
>> With the patch applied, and some other modifications to iwmon you can
>> see two entries for these IEs:
>>
>> Extended Capabilities: len 10
>> Capability: bit 2: Extended channel switching
>> Capability: bit 17: WNM-Sleep mode
>> Capability: bit 19: BSS transition
>> Capability: bit 25: SSID list
>> Capability: bit 46: WNM- Notification
>> Capability: bit 62: Opmode Notification
>> 04 00 0a 02 00 40 00 40 80
>> 01 .....@.@..
>> Tag 258: len 2
>> 00
>> 12 ..
>> Extended Capabilities: len 10
>> Capability: bit 17: WNM-Sleep mode
>> Capability: bit 19: BSS transition
>> Capability: bit 25: SSID list
>> Capability: bit 46: WNM- Notification
>> 00 00 0a 02 00 40 00 00 00
>> 01 .....@....
>> Tag 258: len 2
>> 00 ff
>>
>> If this patch doesn't resolve the issue could you re-run the test and
>> start iwmon at the very beginning, prior to starting IWD?
> I'll do it ASAP, but it will take a bit of time since I'll have to
> compile from sources (I installed from standard Debian package).
>
> Notice I'm not using anything "strange" for scans:
> - normal Debian Sid Network Manager applet from desktop (which sees
> "beelink")
> - normal Android phone `WiFi` Settings from phone (MIUI 12.0.3 on
> POCOPHONE F1, if relevant) (does *not* see "beelink")
> Both are working as expected with several other Access Points.
> It seems (to my limited insight) unlikely they're sending invalid
> probe requests.
I'll need to look more at the spec to determine if this is "illegal"
(for lack of a better term) to duplicate these IEs. There are some IEs
that we allow duplicates of (which you can see in that patch). With a
simple search I cant find anything in the spec that disallows duplicates
specifically, but we have this comment in IWD:
"Tag found, make sure no duplicates present unless allowed"
Denis, any idea where that comes from off the top of your head?
Thanks,
James
next prev parent reply other threads:[~2024-01-02 14:40 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-01-01 19:26 Setup help request Mauro Condarelli
2024-01-01 23:03 ` James Prestwood
2024-01-02 0:34 ` Mauro Condarelli
2024-01-02 13:01 ` James Prestwood
2024-01-02 14:07 ` Mauro Condarelli
2024-01-02 14:40 ` James Prestwood [this message]
2024-01-02 17:10 ` Denis Kenzior
2024-01-02 15:11 ` Mauro Condarelli
2024-01-02 15:36 ` James Prestwood
2024-01-02 17:09 ` Mauro Condarelli
2024-01-02 17:30 ` James Prestwood
2024-01-02 17:45 ` Mauro Condarelli
2024-01-02 19:52 ` James Prestwood
2024-01-02 20:30 ` Mauro Condarelli
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=bf00447d-9817-4bd9-ba67-149356d6ecd6@gmail.com \
--to=prestwoj@gmail.com \
--cc=iwd@lists.linux.dev \
--cc=mc5686@mclink.it \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox