linux/poky.git/meta/recipes-devtools/run-postinsts, branch zeus-next

run-postinsts: Fix full execution of scripts at first boot

2019-05-12T08:04:26+00:00

run-postinsts runs a given set of scripts during the first boot of the device, when one of these scripts prints something to stdout (isnt daemonized correctly), since stdout is not available at that time, the script execution immediately returns with an error (exit_group()), this error causes the script to terminate all threads within the process, causing undesired behavior since the script might still had to execute some other code. Replace eval built-in with (), since () executes in a subshell, even if one of the scripts exits, all threads of that process will only be within that session, this ensures other scripts meant to be run are still run afterwards. [YOCTO #13266] (From OE-Core rev: 706410c847ac9c89317d098de5d5c580736edbbb) Signed-off-by: Alejandro Enedino Hernandez Samaniego Signed-off-by: Richard Purdie

run-postinsts: for dpkg/opkg, do not rely on /etc/*-postinsts

2018-06-18T10:07:58+00:00

Start opkg/dpkg as soon as the respective package managers status file is present, no matter whether /etc/$pm-postinsts exists. This decouples the implicit link between postinsts scripts in /etc and the package manager: Currently the package manager is only started if those scripts are present, although the package manager does not use those scripts at all! Package managers install their own set of postinst scripts. The behavior when using rpm packages stays the same. Note that using the package managers capability to execute postinst scripts is preferred for good reasons: It makes sure that the package managers database reflects that the packages have been completely installed and configured. This change allows to drop installation of the postinsts scripts when package management is present. This will be done in a separate change. Note: Before commit 5aae19959a44 ("rootfs.py: Change logic to unistall packages") rootfs.py did not install /etc/$pm-postinsts when package management is installed! The change caused YOCTO #8235 which lead to the behavior change of run-postinsts in first place. (From OE-Core rev: 1af421fcf05457223bf271f3951599b6c150cdff) Signed-off-by: Stefan Agner Signed-off-by: Richard Purdie

run-postinsts: Replace pi_dir variable test

2018-04-18T17:57:06+00:00

Since commit 5159ddcb62682e1b7e63a20a9218ea96e3fe10a2 string length test performed against pi_dir has effectively never been able to succeed. Change this to rather test if pi_dir is not an existing directory. By doing we remove the chance of seeing the following console error message during first boot to a pristine rootfs: 'ls: /etc/ipk-postinsts: No such file or directory' (From OE-Core rev: b1600ec8991dfbd3b73d209b9c620a171c5b13c4) Signed-off-by: Niko Mauno Signed-off-by: Ross Burton Signed-off-by: Richard Purdie

run-postinsts: do not reload daemon configuration

2017-08-15T23:03:14+00:00

In case a systemd service disables itself while init is still in its boot sequence the reloading of the service files can be problematic. In that case: It seems that systemd looses the state of .device units, and some services depend on such units (namely serial consoles such as serial-getty@ttymxc0.service). As a result no getty is spawned on the affected serial tty. After a power-cycle the second boot (which does not disable services) succeeds. The following sequence shows this problem: | Jan 09 16:36:28 apalis-t30 systemctl[162]: Removed /etc/systemd/system/sysinit.target.wants/run-postinsts.service. | Jan 09 16:36:28 apalis-t30 systemd[1]: Reloading. | ... | And then the failing one: | Feb 22 15:33:15 apalis-t30 systemd[1]: dev-ttyS0.device: Job dev-ttyS0.device/start timed out. | Feb 22 15:33:15 apalis-t30 systemd[1]: Timed out waiting for device dev-ttyS0.device. | Feb 22 15:33:15 apalis-t30 systemd[1]: Dependency failed for Serial Getty on ttyS0. | Feb 22 15:33:15 apalis-t30 systemd[1]: serial-getty@ttyS0.service: Job serial-getty@ttyS0.service/start failed with result 'dependency'. | Feb 22 15:33:15 apalis-t30 systemd[1]: dev-ttyS0.device: Job dev-ttyS0.device/start failed with result 'timeout'. | Feb 22 15:33:15 apalis-t30 systemd[1]: Reached target Login Prompts. (the time has been updated between this two events, but that does not influence the issue) Using --no-reload in the service file avoids the "Reloading." message above and seems to not cause such issues anymore. Reported-by: Stefan Agner (From OE-Core rev: 23b359b6e26d0b17037bf955bd15a16a3fd9ab8f) Signed-off-by: Max Krummenacher (cherry picked from commit 4a425d05bac5dcba023b67aa3d726f7e7869404f) Signed-off-by: Richard Purdie

run-postinsts: Print message before running deferred postinst scripts

2016-12-08T10:31:30+00:00

Opkg can defer running postinst scripts to first boot, which can take a while on some systems. The output of `opkg configure` (or whatever pm is used) is redirected to a file when logging is enabled (I.e. $POSTINST_LOGGING == 1), making the machine appear hung during this process. This change simply prints a wait message on the console to inform the user of this potentially long and silent operation so that they do not mistakenly reboot their machine. Why not simply `tee` the output instead? Tee might be provided by BusyBox in some distros, which may need to run update-alternatives in the very postinst scripts being executed by this process. It's therefore not safe to assume Tee (or any other packaged util) is available until the configure process finishes. (From OE-Core rev: 3f8f1d8e4b129e5dfdd63d450887320aa0e33228) Signed-off-by: Haris Okanovic Signed-off-by: Ross Burton Signed-off-by: Richard Purdie

run-postinsts: Use opkg/dpkg to configure when possible

2016-12-07T10:38:00+00:00

Currently run-postinsts script has code to run postinst scripts via opkg/dpkg configure but that code is never used. The advantage of using package managers instead of just executing the scripts is to keep the package manager DB updated. Fix the script so that the package managers are used when appropriate. Also use $localstatedir for the opkg runtime file location. Fixes [YOCTO #10478]. (From OE-Core rev: b645919f173512f9e75aeb26348d60b63dcdc53c) Signed-off-by: Jussi Kukkonen Signed-off-by: Ross Burton Signed-off-by: Richard Purdie

Remove $COREBASE/LICENSE from LIC_FILES_CHKSUM

2016-10-28T15:15:18+00:00

Several recipes reference the LICENSE file in their LIC_FILES_CHKSUM variable as ${COREBASE}/LICENSE. This forces distribution providers to keep this file verbatim or to overload the affected recipes. The section "Moving to the Yocto Project 1.6 Release" in the Yocto manual suggests removing the LICENSE file where possible. Remove LICENSE in cases where COPYING.MIT is also given and replace LICENSE with COPYING.MIT if the former was the only entry. All modified recipes specify LICENSE = "MIT" and none of the in-tree files specify a different license either. As the packages do not change (the license files are not contained in them), do not increase PR. (From OE-Core rev: 0059e0661826c857a07c862bcb46162671e0e330) Signed-off-by: Olaf Mandel Signed-off-by: Ross Burton Signed-off-by: Richard Purdie

run-postinsts: Correct comment misspelling, "rpm-posinsts".

2016-07-08T08:57:24+00:00

(From OE-Core rev: 45048442913687f2817ef9a52352fe3214f5f023) Signed-off-by: Robert P. J. Day Signed-off-by: Richard Purdie

opkg: upgrade to v0.3.0

2015-09-03T11:43:14+00:00

Changes required: - Rename opkg-cl to opkg - Add libarchive dependency - Drop backport patches - Drop obsolete directory options - Add patch to handle empty index files Based on initial work by Paul Barker. (From OE-Core rev: 1dd2a9ea54f5a5497e23814f144f35ff15430d71) Signed-off-by: Alejandro del Castillo CC: Paul Barker Signed-off-by: Richard Purdie

run-postinsts: use ${localstatedir} instead of hardcoding /var/lib

2015-07-31T09:32:38+00:00

(From OE-Core rev: 48d735a8938ec7c49754a57019182ebcd9a2d8b6) Signed-off-by: Ross Burton Signed-off-by: Richard Purdie