]> www.infradead.org Git - users/jedix/linux-maple.git/commit
wifi: rtw88: usb: schedule rx work after everything is set up
authorMarcin Ślusarz <mslusarz@renau.com>
Tue, 28 May 2024 11:02:46 +0000 (13:02 +0200)
committerPing-Ke Shih <pkshih@realtek.com>
Fri, 5 Jul 2024 01:38:18 +0000 (09:38 +0800)
commitadc539784c98a7cc602cbf557debfc2e7b9be8b3
tree4daea226d5762bfc4ad4ca775f5d1b707e7549e8
parent746f4ae52a38459dbed792adb58cadaed370dcc2
wifi: rtw88: usb: schedule rx work after everything is set up

Right now it's possible to hit NULL pointer dereference in
rtw_rx_fill_rx_status on hw object and/or its fields because
initialization routine can start getting USB replies before
rtw_dev is fully setup.

The stack trace looks like this:

rtw_rx_fill_rx_status
rtw8821c_query_rx_desc
rtw_usb_rx_handler
...
queue_work
rtw_usb_read_port_complete
...
usb_submit_urb
rtw_usb_rx_resubmit
rtw_usb_init_rx
rtw_usb_probe

So while we do the async stuff rtw_usb_probe continues and calls
rtw_register_hw, which does all kinds of initialization (e.g.
via ieee80211_register_hw) that rtw_rx_fill_rx_status relies on.

Fix this by moving the first usb_submit_urb after everything
is set up.

For me, this bug manifested as:
[    8.893177] rtw_8821cu 1-1:1.2: band wrong, packet dropped
[    8.910904] rtw_8821cu 1-1:1.2: hw->conf.chandef.chan NULL in rtw_rx_fill_rx_status
because I'm using Larry's backport of rtw88 driver with the NULL
checks in rtw_rx_fill_rx_status.

Link: https://lore.kernel.org/linux-wireless/CA+shoWQ7P49jhQasofDcTdQhiuarPTjYEDa--NiVVx494WcuQw@mail.gmail.com/
Signed-off-by: Marcin Ślusarz <mslusarz@renau.com>
Cc: Tim K <tpkuester@gmail.com>
Cc: Ping-Ke Shih <pkshih@realtek.com>
Cc: Larry Finger <Larry.Finger@lwfinger.net>
Cc: Kalle Valo <kvalo@kernel.org>
Cc: linux-wireless@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20240528110246.477321-1-marcin.slusarz@gmail.com
drivers/net/wireless/realtek/rtw88/usb.c