A printk() was previously introduced that is passing the log level as the first argument
instead of prefixing the format with the level. Given this code path, this is causing a
kernel fault and CPU reboot without any kernel panic/stack trace, since it's crashing
inside of printk().
CPU:0 ts:206871944795 ffffffc008dcf828 ffffffc008dfe914 cleanup_module [batman_adv] <- cfg80211_vendor_cmd_reply+0x7ff4/0xa064 [cfg80211]
CPU:0 ts:206871944800 ffffffc0108e69d0 ffffffc008dcf80c printk <- cleanup_module+0xb22c/0xa20 [batman_adv]
CPU:0 ts:206871944802 ffffffc0101173c0 ffffffc0108e6a08 vprintk_func <- printk+0x60/0x6c
CPU:0 ts:206871944806 ffffffc010115e44 ffffffc0101160ec vprintk_emit <- vprintk_default+0x4c/0x60
CPU:0 ts:206871944809 ffffffc0101172d0 ffffffc010115e88 __printk_safe_enter <- vprintk_emit+0x84/0x29c
CPU:0 ts:206871944812 ffffffc010115c20 ffffffc010115ec4 vprintk_store <- vprintk_emit+0xc0/0x29c
CPU:0 ts:206871944816 ffffffc0100a53a4 ffffffc010080fb4 do_translation_fault <- do_mem_abort+0x54/0xb0
CPU:0 ts:206871944819 ffffffc0100a4eb4 ffffffc0100a5448 do_page_fault <- do_translation_fault+0xc8/0xe0
CPU:0 ts:206871944821 ffffffc0100a4d08 ffffffc0100a5038 __do_kernel_fault <- do_page_fault+0x1a8/0x4f0
CPU:0 ts:206871944837 ffffffc0100a4c88 ffffffc0100a4df0 die_kernel_fault <- __do_kernel_fault+0x110/0x1b0
After applying the fix, we can see this code path is being hit:
[26799.175166] cfg80211_calculate_bitrate_he: invalid rate->nss: 0
This still doesn't fix the original issue triggering this code path, which is why a nss value
of 0 is being reported.
Fixes: eb9cbaec7 ("ipq807x: Shorten the kernel backtrace warning msg for ieee80211_bss_get_elem")
Fixes: WIFI-14677
Signed-off-by: Paul White <paul@shasta.cloud>
Fix the case where firmware crashes when STA sends AUTH with same
MAC address to multiple SSIDs on the same radio.
Fixes: WIFI-13276
Signed-off-by: Arif Alam <arif.alam@netexperience.com>
When using psk2-radius in combination with enhanced MPSK,
the passphrase was not properly propagated to user scripts
via the ucode interface, because the PSK field was not set
in the connected station context.
This patch fixes that by copying the passphrase into the
psk field.
Signed-off-by: Marek Kwaczynski <marek@shasta.cloud>
For the Edgecore EAP-105 platform, configure the 2.4GHz, 5GHz,
and 6GHz WLAN interfaces to use MAC addresses derived
from the base WAN MAC address.
Fixes: WIFI-14624
Signed-off-by: Marek Kwaczynski <marek@shasta.cloud>
* While Roaming the AP couldn't find the wildcard R0KH and R1KH ids,
which are required by Fast transistion.
* Issue caused by the placement of conf parser in the invalid location.
Fixes: WIFI-14544
Signed-off-by: Marek Kwaczynski <marek@shasta.cloud>
9710867 (HEAD -> main, origin/main, origin/HEAD) make the MTU configurable on GRE tunnels
4dd0904 SSH IdleTimeout can be configured from JSON config
6faaa1f HaLow: Extend ucentral schema & state for HaLow
Signed-off-by: John Crispin <john@phrozen.org>
6faaa1f HaLow: Extend ucentral schema & state for HaLow
aa9cac5 dhcp_inject: Use same keyword "dhcp-inject" in config file
Signed-off-by: John Crispin <john@phrozen.org>
1. porting MorseMicro HaLow driver to support HaLow on EAP112
2. Only support FCC regulation because of hardware limitation
3. Add /etc/init.d/halow-gpio-reset to initialize HaLow chip in early stage
4. Add /etc/uci-defaults/aaa-fix-phy0-to-morse to correct the default uci for HaLow radio.
Signed-off-by: Ian Chen <ian77_chen@accton.com>
* add "limit_rtlphy_10g_ablity" in DTS , no side effect on other product.
* disable 10G capability if DTS defined limit_rtlphy_10g_ablity , no side effect on other product.
* revert the last 0006-qca-ssdk-Fix-10G-rtl-phy-driver-for-c45-mdio-read-wr.patch and based on 0005 patch.
Fixes: WIFI-14567
Signed-off-by: Ken Shi <xshi@actiontec.com>
Added check when parsing ssid info retrieved from iwinfo.
Program will exit if expected interface count and iwinfo entry count mismatch.
Fixes: WIFI-14564
Signed-off-by: alex18_huang <alex18_huang@accton.com>
The main reason was adding missed functionality for Radius
configuration which caused rejecting WiFi clients on
authentication level.
Still some changes not included:
* AFC,
* hs20_release,
* multiple_bssid,
* he_co_locate
Fixes: WIFI-14459
Signed-off-by: Marek Kwaczynski <marek@shasta.cloud>
- avoid calculated Channel Utilization value to be 'infinity' from cycle_count_delta being '0'
Fixes: WIFI-14536
Signed-off-by: Tanya Singh <tanya_singh@accton.com>
Program sometimes did not setup tc qdisc on some interfaces when creating SSID with dhcpinject enabled initially.
Added delayed startup.
Fixes: WIFI-14522
Signed-off-by: alex18_huang <alex18_huang@accton.com>
The bug was in the upstream hostapd codebase, backport/rebase the fix
--> ba150059d1ec964add8f29eb2c92dd6dfde97308
Fixes: WIFI-14508
Signed-off-by: John Crispin <john@phrozen.org>
