@Oleksij, it appears tougher to locate than I thought. I'm back to kernel 3.11.0-15-generic. I tried to overheat by stressing the CPU, GPU and wireless device for 15 minutes by running stress --cpu 4, glxgears fullscreen (vsync on however) and setting txpower to 20 and pingflooding my gateway (sudo ping -f -s 1476 ). It did not succeed.
Only after 6-9 hours of moderately (ping -Di 0.2 -s 65507 192.168.1.1, you'll be the judge) pingflooding my gateway I got into this state where I could not ping my gateway anymore. Stopping the ping flood, waiting a while and trying a normal ping again did not appear to work. Only after suspending, waiting a while and resuming I can get the wifi adapter to function again. Resuming 1-2 seconds after suspending did not work for me. However, after this suspend-resume trick I can trigger the nonfunctional state quite easily after 10-15 minutes. So I am a bit confused about whether this is related to overheating, if it takes such a long while to trigger. Can I enable some kind of debugging to show this behaviour? -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1049383 Title: ath9k_usb: TL-WN821N v3 (AR7010+9287) Connection drops To manage notifications about this bug go to: https://bugs.launchpad.net/linux/+bug/1049383/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
