On 5/5/20 9:49 AM, Sven Eckelmann wrote: > On Tuesday, 5 May 2020 09:01:34 CEST Sven Eckelmann wrote: >> On Tuesday, 5 May 2020 01:46:12 CEST Rajkumar Manoharan wrote: >> [...] >>> IIRC this was fixed a while ago by below patch. Somehow it never landed >>> in ath.git. >>> Simple one line change is enough. >>> >>> https://patchwork.kernel.org/patch/10550707/ >> Because it doesn't work for everything. Remember that 10.2.4.x overflows all >> the time (14-30s) because it used only 31 bit for the counters. >> >> But feel free to point me to the firmware version which fixed this. > See also https://patchwork.kernel.org/patch/9701459/ > > Kind regards, > Sven This patch already fixes the problem for me. I tested it on QCA988X hw with firmware 10.2.4. [ 10.350919] ath10k_pci 0000:04:00.0: qca988x hw2.0 target 0x4100016c chip_id 0x043222ff sub 0000:0000 [ 10.350930] ath10k_pci 0000:04:00.0: kconfig debug 1 debugfs 1 tracing 1 dfs 0 testmode 0 [ 10.351803] ath10k_pci 0000:04:00.0: firmware ver 10.2.4-1.0-00047 api 5 features no-p2p,raw-mode,mfp,allows-mesh-bcast crc32 35bd9258 [ 10.385617] ath10k_pci 0000:04:00.0: board_file api 1 bmi_id N/A crc32 bebc7c08 [ 11.536818] ath10k_pci 0000:04:00.0: htt-ver 2.1 wmi-op 5 htt-op 2 cal otp max-sta 128 raw 0 hwcrypto 1 I also did not see the 31 bit overflow after a small amount of seconds. Survey data from wlp4s0 frequency: 2412 MHz [in use] noise: -65 dBm channel active time: 5370225 ms channel busy time: 924199 ms channel receive time: 140 ms channel transmit time: 0 ms Kind regards, Markus
