Evan: d/patches/close-libusb-on-failure.patch looks fair enough, but what is debian/patches/test? It's a patch of a patch, not referenced/explained in the changelog, what does that do? Please either explain in the changelog or if it is just noise please remove and reupload. Thanks!
-- You received this bug notification because you are a member of Desktop Packages, which is subscribed to pcsc-lite in Ubuntu. https://bugs.launchpad.net/bugs/1551897 Title: Excessive CPU utilization Status in ccid package in Ubuntu: Fix Released Status in pcsc-lite package in Ubuntu: Invalid Status in ccid source package in Xenial: In Progress Status in pcsc-lite source package in Xenial: Invalid Status in pcsc-lite package in Debian: Fix Released Bug description: [Impact] When hotplugging a composite USB device that expose multiple interfaces (such as a Yubikey NEO or Yubikey 4), libccid does not correctly de-initialize the USB library when it skips over the non- CCID interface. This subsequently results in pcscd (or, presumably, anything that loads libccid, though I don't think anything else does) either using 100% CPU or segfaulting when the device is unplugged. This seems worthy of an SRU as it is an easily reproducible bug with a simple, targeted fix. [Test Case] 1. Make sure pcscd is running (sudo systemctl start pcscd; possibly sudo systemctl restart pcscd if a potentially fixed package has just been installed) 2. Plug and then unplug a composite USB device that includes a CCID interface (such as a Yubikey NEO) 3. Observe that pcscd has crashed with a segfault (sudo systemctl status pcscd) [Regression Potential] The patch is quite narrow, and thus the potential for regression should be limited. The new code adds calls to close_libusb_if_needed, which is already written to be fairly conservative. The worst case here is likely an unexpected call to libusb_exit, which could cause libusb to get into an inconsistent state. However, in practice the call seems quite safe, and the likelihood of regression low. ====== Original bug description: In xenial, pcscd CPU utilization occasionally goes haywire: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 27404 root 20 0 384264 4992 2228 S 100.3 0.0 75:39.03 /usr/sbin/pcscd --foreground --auto-exit I get this about once a day since installing xenial. I can't quite connect it to any specific event on the system, it just seems to happen sporadically -- fan turns on, I run top, and there it is again. Never had this issue on same hardware incl. smartcard with trusty. I wonder if it might be this? https://bugs.debian.org/cgi- bin/bugreport.cgi?bug=718473 To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/ccid/+bug/1551897/+subscriptions -- Mailing list: https://launchpad.net/~desktop-packages Post to : desktop-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~desktop-packages More help : https://help.launchpad.net/ListHelp
