Posted: Tue Oct 13, 2020 16:14 Post subject: Crazy high TX error rates
I've got Firmware: DD-WRT v3.0-r44483 std (10/02/20) running on 2 Netgear boxes, one r7000p, and one r6400v2. I'm seeing REALLY high eth TX error rates on both. Where should I look first to diagnose this?
Both boxes have been up for about 3 days, and have the following error counts in ifconfig:
Both are running as APs inside my home network, WAN connection "Disabled", WAN port assigned to switch, DHCP servers disabled, Operating mode "Router".
r6400v2 has only one LAN connection, in port 2, connected to an upstream gig switch.
r7000p has ports W, 1, 2, 3 connected. 4 is down. All ports are assigned to the LAN bridge.
Wifi is enabled on both, although I currently have turned off wl0 on r6400v2 trying to troubleshoot this problem.
My network is flat. Everything is in 192.168.1.0/24
r6400v2 is simple, single store-bought cat 5e cable to linksys gigabit switch, then upstream to larger gigabit switch, then upstream to cable modem. total cable distance from r6400v2 port 2 through 2 switches to cable modem is about 50 feet.
r7000p is more of a workhorse but same basic setup. Handmade cable from r7000p port (can't recall which one) to wall jack up through attic to larger gigabit switch, then upstream to cable modem. Similar overall cable length.
I don't know how to map from physical ports to eth1, eth2 ports. I'd appreciate some help. However, based on where the error counts show up in the web UI, those errors are coming from the wifi side not the ethernet side.
current r6400v2 after upgrading to DD-WRT v3.0-r44538 std (10/13/20) and running for just a few minutes is
Wireless Packet Info (wl0)
Received (RX)3260 OK, no error
Transmitted (TX)11299 OK, 12369 errors
Wireless Packet Info (wl1)
Received (RX)5082 OK, 12 errors
Transmitted (TX) 34203 OK, 3608 errors
Code:
clients
MAC Address Radioname Interface Uptime TX Rate RX Rate Info Signal Noise SNR Signal Quality
B8:27:EB:CC:D4:C0 wl0 0:09:41 72M 65M HT20SGI -52 -90 38 96%
F0:D5:BF:01:3B:F0 wl1 0:11:38 400M 400M VHT80SGI -53 -92 39 94%
I've always had high tx errors on wl0 which only seems to happen when using my laptop with an intel adapter, in total I have 21 devices connected through wifi and as long as the laptop isn't connected the tx errors stay between 3-15. what is your percentage bar reading? Mine is 100% with the worst i've seen being 99%
Posted: Thu Oct 15, 2020 14:00 Post subject: kernel.log for anyone interested
I'm seeing "INFO: rcu_sched self-detected stall on CPU" and stack traces in the kernel.log of r6400v2. I've posted https://0x0.st/iGkd.log here in case somebody wants to take a look.