The WiFi driver crashes on an MT7621 router running OpenWRT21.02rc1 (5 commits further to be exact 6f053e5).
The router used is an InvizBox Go (WiFi only) i.e. MT7621 + MT7603E (not supported by OpenWRT just yet - I'm working on getting there). I don't believe the issue is specific to that hardware though.
The crash seems to happen in AP/STA mode when an AP gets added to the configuration and the STA was connected (I'll add more information if I come across more scenarios). I'm not able to reproduce consistently at the moment.
Once crashed the WiFi stack is not usable anymore until a reboot. Following a reboot, things are back to normal.
The following steps do not necessarily lead to the crash. I expect them to be what leads to it but am still unsure. The stack may show the way to a better reproduction scenario...
On an MT7621 router setup as AP/STA, add a second AP and possibly a third one.
Call /etc/init.d/network reload after each change
Current behaviour
Crash stack is visible in console/dmesg and the STA fails to reconnect (which also leads all APs to become not accessible - expected).
Expected behaviour
No crash
Notes
This crash was also observed on builds before the rc1 tag. I don't know the conditions leading to these crashes.
VincentR:
Summary
The WiFi driver crashes on an MT7621 router running OpenWRT21.02rc1 (5 commits further to be exact 6f053e5).
The router used is an InvizBox Go (WiFi only) i.e. MT7621 + MT7603E (not supported by OpenWRT just yet - I'm working on getting there). I don't believe the issue is specific to that hardware though.
The crash seems to happen in AP/STA mode when an AP gets added to the configuration and the STA was connected (I'll add more information if I come across more scenarios). I'm not able to reproduce consistently at the moment.
Once crashed the WiFi stack is not usable anymore until a reboot. Following a reboot, things are back to normal.
The crash stack is as follow:
[ 726.587920] ------------[ cut here ]------------
[ 726.597570] WARNING: CPU: 0 PID: 1767 at backports-5.10.16-1/net/mac80211/ieee80211_i.h:1468 sta_info_alloc+0x5c4/0x5fc [mac80211]
[ 726.621113] Modules linked in: xt_connlimit pppoe ppp_async nf_conncount iptable_nat xt_state xt_nat xt_helper xt_conntrack xt_connmark xt_connbytes xt_REDIRECT xt_MASQUERADEr
[ 726.621458] algif_skcipher algif_rng algif_hash algif_aead af_alg sha256_generic libsha256 sha1_generic jitterentropy_rng drbg md5 hmac echainiv des_generic libdes cbc authec
[ 726.859235] CPU: 0 PID: 1767 Comm: hostapd Not tainted 5.4.111 #0
[ 726.871379] Stack : 8e375fb4 8007ce5c 80660000 80660d78 806c0000 80660d40 8065fe94 8e257a34
[ 726.888032] 80800000 8dc34188 806aa6e3 805f993c 00000000 00000001 8e2579d8 72c69b9f
[ 726.904659] 00000000 00000000 80830000 00000000 30232031 0000014b 2e352064 31312e34
[ 726.921286] 00000000 00000024 00000000 000d1c63 80000000 806c0000 00000000 8e3075b8
[ 726.937913] 00000009 8e146480 00000005 00000002 00000000 8034dfbc 00000000 80800000
[ 726.954538] ...
[ 726.959397] Call Trace:
[ 726.964284] [<8000b64c>] show_stack+0x30/0x100
[ 726.973153] [<80542370>] dump_stack+0xa4/0xdc
[ 726.981832] [<8002bee0>] __warn+0xc0/0x10c
[ 726.989981] [<8002bf88>] warn_slowpath_fmt+0x5c/0xac
[ 727.000018] [<8e3075b8>] sta_info_alloc+0x5c4/0x5fc [mac80211]
[ 727.011712] [<8e3259b0>] ieee80211_nan_func_match+0x3d88/0x4410 [mac80211]
[ 727.025846] ---[ end trace bc705bf94f0c5c24 ]---
Steps to reproduce
The following steps do not necessarily lead to the crash. I expect them to be what leads to it but am still unsure. The stack may show the way to a better reproduction scenario...
On an MT7621 router setup as AP/STA, add a second AP and possibly a third one.
Call /etc/init.d/network reload after each change
Current behaviour
Crash stack is visible in console/dmesg and the STA fails to reconnect (which also leads all APs to become not accessible - expected).
Expected behaviour
No crash
Notes
This crash was also observed on builds before the rc1 tag. I don't know the conditions leading to these crashes.
I had saved one older crash stack (truncated by console unfortunately) as a reference (early 21.02 branch when I started testing in preparation for release):
[ 1939.972549] ------------[ cut here ]------------
[ 1939.982051] WARNING: CPU: 2 PID: 2079 at backports-5.10.16-1/net/mac80211/ieee80211_i.h:1468 sta_info_alloc+0x5c4/0x]
[ 1940.005526] Modules linked in: xt_connlimit pppoe ppp_async nf_conncount iptable_nat xt_state xt_nat xt_helper xt_cor
[ 1940.005842] algif_skcipher algif_rng algif_hash algif_aead af_alg sha256_generic libsha256 sha1_generic jitterentroc
[ 1940.243641] CPU: 2 PID: 2079 Comm: hostapd Not tainted 5.4.111 #0
[ 1940.255784] Stack : 8df75fb4 8007ce5c 80660000 80660d78 806c0000 80660d40 8065fe94 8dd01a34
[ 1940.272451] 80800000 8fe51fc8 806aa6e3 805f993c 00000002 00000001 8dd019d8 91468ee6
[ 1940.289110] 00000000 00000000 80830000 00000000 30232031 0000012f 2e352064 31312e34
[ 1940.305752] 00000000 00000060 00000000 0003b7b9 80000000 806c0000 00000000 8df075b8
[ 1940.322392] 00000009 8fed4480 00000005 00000002 00000000 8034dfbc 00000008 80800008
[ 1940.339021] ...
[ 1940.343883] Call Trace:
[ 1940.348778] [<8000b64c>] show_stack+0x30/0x100
[ 1940.357652] [<80542370>] dump_stack+0xa4/0xdc
[ 1940.366336] [<8002bee0>] __warn+0xc0/0x10c
[ 1940.374487] [<8002bf88>] warn_slowpath_fmt+0x5c/0xac
[ 1940.384532] [<8df075b8>] sta_info_alloc+0x5c4/0x5fc [mac80211]
[ 1940.396280] [<8df259b0>] ieee80211_nan_func_match+0x3d88/0x4410 [mac80211]
[ 1940.410460] ---[ end trace dda71e821ee728c4 ]---
The text was updated successfully, but these errors were encountered: