On Tue, Oct 11, 2011 at 18:38, Dan McGee <dpmcgee@xxxxxxxxx> wrote: > In trying to answer the question "What is the proper way to restart > udevd on a running system?", the following questions came up: > > 1. I was informed something such as `killall udevd && udevd --daemon` > was not safe, is this true? You might not be able to start a new udevd while the old one is still running. > 2. Is doing `udevadm control --exit && udevd --daemon` any better? Unlike signals, --exit is sync and blocks properly. You can only lose events between the stop and start that way - but that has and can always happen in simple setups that do not support socket activation. > 3. Now that the decision [1] has been made to move udevd out of the > system path, how is one supposed to even locate the binary to restart? > 4. Should there be a `udevadm control --restart` command? As mentioned in the NEWS of udev, for non-systemd systems create a symlink in /sbin where udevd was located . On systemd systems we can update and restart udevd without knowing any details, calling parameters/sequence, install locations, and without losing any event from the kernel or a tool talking to udevd during the restart. For older systems with the symlink in /sbin nothing should have changed. Kay -- To unsubscribe from this list: send the line "unsubscribe linux-hotplug" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
