linux-kernel - wifi: mt76: mt7921 Survey is missing noise floor

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Message-ID: <20241013111920.679188-1-martin.blumenstingl@googlemail.com>
Date: Sun, 13 Oct 2024 13:19:20 +0200
From: Martin Blumenstingl <martin.blumenstingl@...glemail.com>
To: linux-wireless@...r.kernel.org
Cc: Felix Fietkau <nbd@....name>,
	Lorenzo Bianconi <lorenzo@...nel.org>,
	Ryder Lee <ryder.lee@...iatek.com>,
	Shayne Chen <shayne.chen@...iatek.com>,
	Sean Wang <sean.wang@...iatek.com>,
	linux-kernel@...r.kernel.org,
	linux-mediatek@...ts.infradead.org,
	Martin Blumenstingl <martin.blumenstingl@...glemail.com>
Subject: wifi: mt76: mt7921 Survey is missing noise floor

Hello,

I'm using an MT7921 PCIe card in AP mode on the 5GHz band.
It works well when I manually set a channel in hostapd.conf.

Setting channel=0 or channel=acs_survey (both have the same meaning)
should make hostapd check for a suitable channel and pick that
automatically.
Unfortunately this does not work as hostapd reports the following (for
all frequencies/channels):
    ACS: Survey for freq 5180 is missing noise floor
    ACS: Survey for freq 5180 is missing noise floor
    ACS: Survey for freq 5180 is missing noise floor
    ACS: Survey for freq 5180 is missing noise floor
    ACS: Survey for freq 5180 is missing noise floor
    ACS: Channel 36 has insufficient survey data

I suspect that this is due to mt792x_phy_update_channel() (from
drivers/net/wireless/mediatek/mt76/mt792x_mac.c) which calls
mt792x_phy_get_nf(). The latter just returns zero - so it seems
reading the noise floor from the hardware/firmware is not
supported yet.

What is needed to implement this? I can test patches or - if
someone describes which registers to program - I can also write
a patch and submit it upstream.

Thank you and best regards,
Martin