Arch Linux User Repository

The 100% cpu load on a single core still persists with kernel 4.0.7-2-ARCH. The dmesg | tail gives this when one of the cores runs at 100%:
[ 337.987781] SwitchChannel#14
[ 338.104140] BW=40 EXTCHA_BELOW CentralChannel=9
[ 338.104144] ==>rt2800_config_channel_rf7630 ch9
[ 338.104502] SwitchChannel#9
[ 338.111896] 0x400000a1 (status)
[ 338.111912] rt2800pci_txdone status = 0x400000a1
[ 338.111915] rt2800pci_txdone qid = 0x0
[ 338.112643] 0xa1 (status)
[ 338.112656] rt2800pci_txdone status = 0xa1
[ 338.112659] rt2800pci_txdone qid = 0x0

The messages in the kernel log has also been fixed by a developer, see comment #245 from https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1220146

Maybe you can do the same?

I hereby confirm that the 100% cpu core load is gone with the update. My battery
usage has been reduced by 50%. Thanks Benjarobin!

The amount of kernel messages is less too. On my system it is reduced to a long list with these type of messages.

...
[ 71.092849] ==>rt2800_config_channel_rf7630 ch12
[ 71.209642] ==>rt2800_config_channel_rf7630 ch1
[ 71.416487] ==>rt2800_config_channel_rf7630 ch13
[ 71.533344] ==>rt2800_config_channel_rf7630 ch1
[ 71.740245] ==>rt2800_config_channel_rf7630 ch14
[ 71.857026] ==>rt2800_config_channel_rf7630 ch1
...

I have the exact same issue.

After that it is even hard to start my laptop again, since the bios has a hard time to initialize.

So, confirmed.

even with the new update (18-6) I still most times when I turn off wifi my system get frozen
does anyone have the same problem? (I use Gnome)

Cool! I will test the updated driver a bit later and report back.

Did you also had time to make all the prinkt statements only appear when the kernel/system is in debugging level?

As reported before, for me the statement below did not had the effect I was hoping for.

printk(KERN_DEBUG "example debug message\n");

Coming from:

https://wiki.ubuntu.com/Kernel/KernelDebuggingTricks

Maybe DKMS or modules have a different mechanism / level?

I did update the source code using https://github.com/kuba-moo/mt7630e/
I back ported the fix that produce high CPU load.
If you were talking about another deposit please name it. Thank you :-)

@benjarobin

In the post below I noticed a lot of people reporting success of using a fixed driver for the mt7630. Maybe this can be of use?

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1220146

If the printk("xxx") statements are replaced by printk(KERN_DEBUG "xxx"), for some reason the messages still end up in the logs, while have kernel log level at 4 (= KERN_WARNING).

When I removed all the printk statements from rt2800pci.c, rt2800lib.c and rt2x00dev.c, the output is reduced a lot not, but not stopped.

Unfortunately, the removal of the printk statments does not have an effect on the performance issue.

I am going to fix it properly, I am kind of busy now (sorry). The update will be in aur4 since I think I may not have the time tomorrow (Sunday 7 June)