On Thu, Oct 14, 2010 at 09:15:41PM -0500, Mike Isely wrote: >On Thu, 14 Oct 2010, [email protected] wrote: > >> I encountered a CPU run with rmmod pvrusb2 and then several seconds later >> modprobing pvrusb2. >> >> From what top shows, pvrusb2-context either hung on modprobe or rmmod. >> >> >> # tail /var/log/everything/current -n 20 >> Oct 14 00:53:36 [kernel] tda18271 0-0060: creating new instance >> Oct 14 00:53:36 [kernel] TDA18271HD/C1 detected @ 0-0060 >> Oct 14 00:53:37 [kernel] tda829x 0-0042: type set to tda8295+18271 >> Oct 14 00:53:40 [kernel] usb 3-4: firmware: requesting v4l-cx2341x-enc.fw >> Oct 14 00:53:41 [kernel] pvrusb2: firmware2 upload transfer failure >> Oct 14 00:53:41 [kernel] pvrusb2: Clearing driver error statuss >> Oct 14 00:53:43 [kernel] cx25840 0-0044: 0x0000 is not a valid video input! >> Oct 14 00:53:43 [kernel] DVB: registering adapter 0 frontend 0 (Samsung >> S5H1411 >> QAM/8VSB Frontend)... >> Oct 14 00:53:43 [kernel] tda829x 0-0042: type set to tda8295 >> Oct 14 00:53:43 [kernel] tda18271 0-0060: attaching existing instance >> Oct 14 00:53:47 [kernel] usbcore: deregistering interface driver pvrusb2 >> Oct 14 00:53:47 [kernel] pvrusb2: Device being rendered inoperable > > >> Oct 14 00:53:47 [kernel] pvrusb2: unregistered device ■■■o0 [mpeg] > >This line is the key. Notice the garbage characters in place of the >device name. What's happened is that the printk which generated this >line referenced deallocated memory - because it had earlier been freed >as part of the driver tear-down. This is one of the bugs that I stomped >out last Spring (by reordering the tear-down steps slightly to avoid >this case).
Yup. After grabbing the dmesg log, I noticed the garbage chars and suspected something odd like that, but you're the expert -- and I today noticed your Changelog entry on this bug. > >I don't think you're running the latest driver. > > >> Oct 14 00:53:47 [kernel] pvrusb2: unregistering DVB devices >> Oct 14 00:54:02 [kernel] SysRq : Emergency Sync >> - Last output repeated 2 times - >> >> >> Top shows: >> PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND >> 26142 root 20 0 0 0 0 R 100 0.0 5:31.30 pvrusb2-context >> 4085 roger 39 19 2636 1188 904 S 1 0.1 4:12.39 top >> 26299 root 20 0 2656 1176 896 R 1 0.1 0:00.04 top >> >> (I've have to re-enable pvrusb2 mailing list to start sending mailing list >> emails.) > >Two things here: > >1. The mailing list software will disable a subscriber if his address >creates too many bounces. Note that once a month the list software >sends out a reminder that you're subscribed along with instructions (as >a reminder) for unsubscribing if you so choose. If that reminder gets >bounced, then it counts against you - and I've seen a lot of bounces >because people seem to have overly aggressive filters that treat the >reminder as junk. > >2. You posted from an address that I think is different than what you've >used in the past. There's no way the mailing list software can infer >two addresses are related so if you wanted to post from another address >you would also have to subscribe from it first. I purposely deactivated/suspended receiving mailing list emails. I wasn't sure how long it would be to re-enable, so I made a note here I might not see follow-ups just in case. -- Roger http://rogerx.freeshell.org/ _______________________________________________ pvrusb2 mailing list [email protected] http://www.isely.net/cgi-bin/mailman/listinfo/pvrusb2
