R7800 10.4b-ct-9984-fH-012-d6de75e0d SWBA overrun on vdev 0, skipped old beacon

[schwichtgit]

Description of the problem (how to configure, how to reproduce, how often it happens). R7800 multi AP setup using 802.11r.

Software (OS, Firmware version, kernel, driver, etc)

DISTRIB_ID='OpenWrt'
DISTRIB_RELEASE='SNAPSHOT'
DISTRIB_REVISION='r9059-6e104c63d6'
DISTRIB_TARGET='ipq806x/generic'
DISTRIB_ARCH='arm_cortex-a15_neon-vfpv4'
DISTRIB_DESCRIPTION='OpenWrt SNAPSHOT r9059-6e104c63d6'
DISTRIB_TAINTS=''
NAME="OpenWrt"
VERSION="SNAPSHOT"
ID="openwrt"
ID_LIKE="lede openwrt"
PRETTY_NAME="OpenWrt SNAPSHOT"
VERSION_ID="snapshot"
HOME_URL="http://openwrt.org/"
BUG_URL="http://bugs.openwrt.org/"
SUPPORT_URL="http://forum.lede-project.org/"
BUILD_ID="r9059-6e104c63d6"
LEDE_BOARD="ipq806x/generic"
LEDE_ARCH="arm_cortex-a15_neon-vfpv4"
LEDE_TAINTS=""
LEDE_DEVICE_MANUFACTURER="OpenWrt"
LEDE_DEVICE_MANUFACTURER_URL="http://openwrt.org/"
LEDE_DEVICE_PRODUCT="Generic"
LEDE_DEVICE_REVISION="v0"
LEDE_RELEASE="OpenWrt SNAPSHOT r9059-6e104c63d6"

driver: ath10k_pci version: 4.14.93

wpad - 2018-12-02-c2c6c01b-1

Firmware: 10.4b-ct-9984-fH-012-d6de75e0d

Hardware (NIC chipset, platform, etc) R7800

0000:00:00.0 PCI bridge: Qualcomm Device 0101
0000:01:00.0 Network controller: Qualcomm Atheros QCA9984 802.11ac Wave 2 Wireless Network Adapter
0001:00:00.0 PCI bridge: Qualcomm Device 0101
0001:01:00.0 Network controller: Qualcomm Atheros QCA9984 802.11ac Wave 2 Wireless Network Adapter

Logs (dmesg, maybe supplicant and/or hostap) system-SchwichtenburgWrt.log.zip

[schwichtgit]

@greearb will you pick this one up, too ?

[greearb]

All these bugs are belong to me. Do you see similar warnings with stock firmware? And if not, would you like to bisect the issue if I build you an image for each of my commits?

[schwichtgit]

@greearb, on

driver: ath10k_pci
version: 4.14.93
firmware-version: 10.4-3.9.0.2-00021

c38f8d67845d83986d347e6f7436f65b2e0323dad5911dbc96895ce575b451f2  firmware-5.bin_10.4-3.9.0.2-00021

I do not see a single occurrence.

however on

driver: ath10k_pci
version: 4.14.93
firmware-version: 10.4b-ct-9984-fH-012-d6de75e0d

I see

grep -i swba  system-SchwichtenburgWrt.log|wc -l
**915**

I do not mind being helpful to address this issue, as both the CT firmware and this AP R7800 are popular, so I guess many would benefit.

but the backside is, that these victim systems are in real use, so it would be good to have a focused triaging strategy to keep the pain level low on the STA users. Also .. this issue are the "stop ship" reason for me to not apply 10.4b-ct-9984-fH-012-d6de75e0d on the 4 AP 802.11r setup we use in production.

[greearb]

It is liable to be disruptive to users to bisect this problem. Several other users have reported the same basic problem, so maybe one of them will step up with an offer to bisect.

[schwichtgit]

please provide the images ... I see what I can do

[greearb]

Here are images to bisect with:

https://www.candelatech.com/downloads/ath10k-9984-10-4b/ath10k-fw-beta/all_builds-9984b-H-feb-6-2019.tar.gz

[schwichtgit]

@greearb, I see a tar ball with 923 firmwares. what is the process you suggest I should follow?

[pgera]

Binary search would be the best approach. Should take fewer than 10 tests.

[greearb]

Please see section on how to bisect toward the bottom of this page: https://www.candelatech.com/ath10k-bugs.php

[shelterx]

I saw this error too now however it doesn't seem very freqeuent for me so catching this error is next to impossible.

[shelterx]

I can only speak for myself but I haven't seen any SWBA overruns in a long time now.

[greearb]

Closing this one, hopefully it is fixed.