Confirmed. r12194 has no internet access also in my router.
@Bartvz I am sending you the logs for debugging.
Thank you!
Confirmed. r12194 has no internet access also in my router.
@Bartvz I am sending you the logs for debugging.
Thank you!
Mushoz, can you share your network configuration. I want apply it to my xiaomi mir3g router. I also suppose there is bug in the switch of mir3g in all recent version of the openwrt.
Guys. The "no internet" issue is a DHCP issue in reality. The solution is very easy. Make sure that you edit the "Resolve file" according to the following:
Go to Network --> DHCP and DNS --> Resolve and Hosts files and edit the "Resolve file" part to this: /tmp/resolv.conf.auto
Then hit Save and Apply, and reboot the router. This fixed it on my end.
Recently I bought a smart switch, with which the errors still emerge (probably less frequent) without the disable flow control patch.
[253784.502614] WARNING: CPU: 3 PID: 0 at net/sched/sch_generic.c:320 dev_watchdog+0x1ac/0x324
[253784.519261] NETDEV WATCHDOG: eth0 (mtk_soc_eth): transmit queue 0 timed out
Happened again. Are we sure the patch got applied this time?
I can confirm this patch works, since the crash only completely disappeared with it applied.
You can always use the force install flag, should work fine as long as there is no major kernel revision
Heh, really should have noted that in the release notes
Again, you are correct. Tired due to work and forgot to do a make clean, doh
You are absolutely correct and it probably is snake oil but it looks like it works for some people. Such a cosmetic change doesn't "hurt".
@ All, new build, r12215 with the cosmetic fix applied
This looks promising, could you post your config please?
In latest build Cake is not available anymore, moving back to fq codel + simple qos
According to the github, OpenWRT master's 4.19 kernel will have new drivers of MT7621 backported from 5.4 or 5.5 and ready for test.
Is it possible to make such a build with the new driver?
Edit: I asked this because of the interrupt errs suffering:
root@OpenWrt:~# cat /proc/interrupts
CPU0 CPU1 CPU2 CPU3
8: 23742611 23742585 23742551 23742578 MIPS GIC Local 1 timer
9: 169213 0 0 0 MIPS GIC 63 IPI call
10: 0 53682005 0 0 MIPS GIC 64 IPI call
11: 0 0 52102628 0 MIPS GIC 65 IPI call
12: 0 0 0 53240053 MIPS GIC 66 IPI call
13: 4314414 0 0 0 MIPS GIC 67 IPI resched
14: 0 5016773 0 0 MIPS GIC 68 IPI resched
15: 0 0 893050 0 MIPS GIC 69 IPI resched
16: 0 0 0 1027280 MIPS GIC 70 IPI resched
19: 14 0 0 0 MIPS GIC 33 ttyS0
20: 0 0 0 0 MIPS GIC 29 xhci-hcd:usb1
21: 417335224 0 0 0 MIPS GIC 10 1e100000.ethernet
22: 2 0 0 0 MIPS GIC 30 gsw
23: 277 0 6651569 0 MIPS GIC 11 mt76x2e
24: 21390323 0 0 0 MIPS GIC 31 mt76x2e
26: 0 0 0 0 GPIO 7 keys
27: 0 0 0 0 GPIO 18 keys
ERR: 1199
in more than 2 days running:
root@OpenWrt:~# uptime
13:24:16 up 2 days, 17:58, load average: 0.00, 0.01, 0.00
I have no clue how these errs came from, and if they're just harmless.
I have the same kernel error again after 5 days.
Can you try this modified patch instead of the one we have now?
It removes the if statement, so on all mt7621 devices it will turn off flow control. If it does not work with this, then this is not the solution. But it worth a try.
Nice. FYI, the patch also made it to master. Gonna start testing it today.
MOD: it did not helped, same error message, within 24 hours...
Is it just me or some else also experienced this? The interrupt Errs seem greatly reduced with latest mt76 code upgrade.
CPU0 CPU1 CPU2 CPU3
8: 11438439 11438367 11438417 11438407 MIPS GIC Local 1 timer
9: 65180 0 0 0 MIPS GIC 63 IPI call
10: 0 21816266 0 0 MIPS GIC 64 IPI call
11: 0 0 21809977 0 MIPS GIC 65 IPI call
12: 0 0 0 20825530 MIPS GIC 66 IPI call
13: 290556 0 0 0 MIPS GIC 67 IPI resched
14: 0 1585346 0 0 MIPS GIC 68 IPI resched
15: 0 0 3819879 0 MIPS GIC 69 IPI resched
16: 0 0 0 430114 MIPS GIC 70 IPI resched
19: 14 0 0 0 MIPS GIC 33 ttyS0
20: 0 0 0 0 MIPS GIC 29 xhci-hcd:usb1
21: 174998675 0 0 0 MIPS GIC 10 1e100000.ethernet
22: 2 0 0 0 MIPS GIC 30 gsw
23: 2 0 2782385 0 MIPS GIC 11 mt76x2e
24: 9815114 0 0 0 MIPS GIC 31 mt76x2e
26: 0 0 0 0 GPIO 7 keys
27: 0 0 0 0 GPIO 18 keys
ERR: 172
up 1 day, 7:46.
Here is my statistic but I don't know if it's good or bad. The ERR counter seems to be quite high in my case...
CPU0 CPU1 CPU2 CPU3
8: 20652222 20652209 20652205 20652195 MIPS GIC Local 1 timer
9: 24380 0 0 0 MIPS GIC 63 IPI call
10: 0 12277125 0 0 MIPS GIC 64 IPI call
11: 0 0 5090606 0 MIPS GIC 65 IPI call
12: 0 0 0 16608850 MIPS GIC 66 IPI call
13: 495613 0 0 0 MIPS GIC 67 IPI resched
14: 0 416821 0 0 MIPS GIC 68 IPI resched
15: 0 0 568266 0 MIPS GIC 69 IPI resched
16: 0 0 0 765120 MIPS GIC 70 IPI resched
19: 12 0 0 0 MIPS GIC 33 ttyS0
20: 0 0 0 0 MIPS GIC 29 xhci-hcd:usb1
21: 44215558 0 0 0 MIPS GIC 10 1e100000.ethernet
22: 32 0 0 0 MIPS GIC 30 gsw
23: 439 0 76432716 0 MIPS GIC 11 mt76x2e
24: 6368129 0 0 0 MIPS GIC 31 mt76x2e
26: 0 0 0 0 GPIO 7 keys
27: 0 0 0 0 GPIO 18 keys
ERR: 31936
07:14:31 up 2 days, 9:23, load average: 0.04, 0.03, 0.00
I don't see any Kernel Errors in my log. I'm not sure if this is related to my setup...
ERR
is incremented in the case of errors in the IO-APIC bus (the bus that connects the CPUs in a SMP system. This means that an error has been detected, the IO-APIC automatically retry the transmission, so it should not be a big problem, but you should read the SMP-FAQ.
Source: https://github.com/torvalds/linux/blob/master/Documentation/filesystems/proc.txt
Never figured out what the most optimum tweaked smp_affinity values are
Just let you guys aware that there’s a hw offloading fix in the pull request, at here.
The next dream is to get a stable switch driver.
We need this patch for the DIR860L...