s6-linux-init/src/init, branch v1.0.6.1 Helpers to build an s6-based init program https://git.skarnet.org/cgit/s6-linux-init/atom/src/init?h=v1.0.6.1 2021-02-14T21:56:52Z bugfix: utmps default should be none 2021-02-14T21:56:52Z Laurent Bercot ska-skaware@skarnet.org 2021-02-14T21:56:52Z urn:sha1:6e0007d32054d0074d93bfa71cfb1f593ebddbf7 bugfix: actually log to catch-all in shutdownd. 2021-01-27T17:33:39Z Laurent Bercot ska-skaware@skarnet.org 2021-01-27T17:33:39Z urn:sha1:355a75e08bbc8af8af97576bad07471dd1b431d8 Also support more options in telinit. Make it work with the new s6. Add s6-linux-init-nuke. 2020-12-09T18:19:21Z Laurent Bercot ska-skaware@skarnet.org 2020-12-09T18:19:21Z urn:sha1:ca8d2c96ea09cb33ff6fef33c0314c24fc6a026a Adapt to skalibs-2.10.0.0 and s6-2.10.0.0 2020-11-30T18:27:41Z Laurent Bercot ska-skaware@skarnet.org 2020-11-30T18:27:41Z urn:sha1:2fa0acdd248c006270010196ad02e530d8ab09ea Accept 1 as cmdline runlevel; fix rc.shutdown comment for s6-rc 2020-11-11T12:39:44Z Laurent Bercot ska-skaware@skarnet.org 2020-11-11T12:39:44Z urn:sha1:c1171b979c236ec2c486bc12e4866112a12b9a84 Fix s6-l-i-m init $@ bug; prepare for 1.0.5.1 2020-11-10T23:03:01Z Laurent Bercot ska-skaware@skarnet.org 2020-11-10T23:03:01Z urn:sha1:86c99b5be2ff091520decdf1a792e9d06f259c30 Don't test for shutdown.allow in containers 2020-06-22T13:29:03Z Laurent Bercot ska-skaware@skarnet.org 2020-06-22T13:29:03Z urn:sha1:bc77187b8d2140eb46ec8b7557858557b26a44ee Properly prefix s6-log in s6-l-i-m 2020-06-09T17:18:00Z Laurent Bercot ska-skaware@skarnet.org 2020-06-09T17:18:00Z urn:sha1:abb464345e953e168bfb95b90a6f910b66b39cbf Use s6-sudod -d in runleveld 2020-03-16T13:17:27Z Laurent Bercot ska-skaware@skarnet.org 2020-03-16T13:17:27Z urn:sha1:3436ed5b69bc5488439458c08b7073f5894acb48 Simplify kbrequest management 2020-02-26T17:27:24Z Laurent Bercot ska-skaware@skarnet.org 2020-02-26T17:27:24Z urn:sha1:559a8893841d3126e5a1cf2490cee8f34cb65d7a - the ioctl never works on /dev/console, so trying to use it on stdin is useless. - /dev/tty0 always exists in a devtmpfs, so the late test is okay. It works on serial consoles too. If no devtmpfs, creating a correct /dev/tty0 is the responsibility of the integrator. - Also added SIGINT blocking before disablecad, to avoid potential early panic. For it to work on kernels without signalfd, it requires the skalibs commit that unblocks selfpipe-trapped signals when not using signalfd. - No such block is needed for kbrequest, because the default action for SIGWINCH is ignore, not terminate.