Avoid to use devices running MediaTek chipsets!

[ZerBea]

During my latest tests, I noticed that the mt76 driver is still a mess on latest Linux longterm kernel!

$ uname -r
6.12.20-1-lts

To make sure that the bugs are not related to hcxdumptool, I used Linux iw wireless configuration tool
https://wireless.docs.kernel.org/en/latest/en/users/documentation/iw.html

The driver promises things it doesn't keep:

it reports faulty TX power:

$ iw dev
phy#0
    Interface wlp48s0f4u2u3
        ifindex 3
        wdev 0x1
        addr 00:c0:ca:b1:0a:87
        type managed
        txpower 6.00 dBm
        multicast TXQ:
            qsz-byt    qsz-pkt    flows    drops    marks    overlmt    hashcol    tx-bytes    tx-packets
            0    0    0    0    0    0    0    0        0

channel is locked in active monitor mode:

driver accepts to set a channel
driver reports that the interface is on the new channel
but interface still stays on channel 1 (confirmed by radiotap header)

$ iw list
...
    Supported commands:
    ...
     * set_channel
...
    Device supports active monitor (which will ACK incoming frames)


$ sudo iw wlp48s0f4u2u3 set monitor active
$ sudo ip link set wlp48s0f4u2u3 up
$ sudo iw dev wlp48s0f4u2u3 set channel 11
$ iw dev
phy#1
    Interface wlp48s0f4u2u3
        ifindex 4
        wdev 0x100000001
        addr 00:c0:ca:b1:0a:87
        type monitor
        channel 11 (2462 MHz), width: 20 MHz (no HT), center1: 2462 MHz
        txpower 6.00 dBm
        multicast TXQ:
            qsz-byt    qsz-pkt    flows    drops    marks    overlmt    hashcol    tx-bytes    tx-packets
            0    0    0    0    0    0    0    0        0
$ tshark -i wlp48s0f4u2u3 -T fields -e  radiotap.channel.freq
Capturing on 'wlp48s0f4u2u3'
2412
2412
2412
2412

[strasharo]

Which adapter was this test on?

[ZerBea]

ALFA AWUS036ACHM
driver mt76x0u
ID 0e8d:7610 MediaTek Inc. WiFi

TP-Link Archer T2UH
driver mt76x0u
ID 148f:761a Ralink Technology, Corp. MT7610U ("Archer T2U" 2.4G+5G WLAN Adapter

ASUS USB AC51
driver mt76x0u
ID 0b05:17d1 ASUSTek Computer, Inc. AC51 802.11a/b/g/n/ac Wireless Adapter [Mediatek MT7610U]

ALFA AWUS036AXML (active monitor mode completely broken)
driver mt7921u
ID 0e8d:7961 MediaTek Inc. Wireless_Device

[strasharo]

Has this propagated to OpenWRT?

[ZerBea]

Not by me.
Since my last issue report was ignored:
openwrt/mt76#839
I don't create upstream bug reports anymore.

[strasharo]

I mean if the bug is there as well, I also decided to not contribute anymore since they asked me to supply my real name for a simple PR. 👍

[ZerBea]

According to OpenWRT guide line the real name is mandatory only for PRs:
https://openwrt.org/submitting-patches
This is in accordance with the Linux Kernel patches guide.

It is not mandatory to use your real name, when opening an issue report.

After release of Linux Kernel 6.14.1 I'll do the test again. According to the Linux Wireless Mailing List there are a lot of changes on mt76. Let's see if the changes are improvements or regressions.