load-timestamp.sh --save should only save the current systemtime and not
evaluate /etc/timestamp and change the systemtime.
If the systemtime has been changed to something in the past of /etc/timestamp
then that change should be accepted without further checks.
While at it correct the comment. The RTC/hwclock is not touched by the
timestamp service. This is a copy paste error from the init.d implementation
in openembedded-core/meta/recipes-core/initscripts/initscripts-1.0/bootmisc.sh
Signed-off-by: Max Krummenacher <max.oss.09@gmail.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
8228b2d56e
changes the time format written to /etc/timestamp
When the system time is set from the content of /etc/timestamp,
one has to use the new format also.
While at it
- remove the now unneeded cut and concatenation as the string is already
correctly ordered.
- consistently use absolute pathnames to programs
- don't print warnings when /etc/timestamp does not exist
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
(cherry picked from commit 7776d287c3a6666d3789ac0fc706a12b079c1dc7)
Signed-off-by: Max Krummenacher <max.oss.09@gmail.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Without this multimachine builds blow up due to packages having the same
license between architectures. Shocking and unexpected!
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
error caused in do_rootfs:
| NOTE: > Executing update_pixbuf_cache intercept ...
| FATAL: kernel too old
| WARNING: The postinstall intercept hook 'update_pixbuf_cache' failed (exit code: 127)! See log for details!
and many other 'kernel too old' messages.
So let's trust in OLDEST_KERNEL set by oe-core's bitbake.conf.
Signed-off-by: Andreas Müller <schnitzeltony@googlemail.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
This services generated from volatile-binds.bb recipe file only have
effect in a read-only filesystem. So if the rootfs is read-write, the
related service are not started.
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
When building out a systemd based image, the udev-hwdb postinstall
script always fails at rootfs time. This is because that the 'udevadm'
command used in udev-hwdb postinstall script is now in the udev
package and udev recommends udev-hwdb.
Instead of letting udev recommends udev-hwdb, we let systemd do it
and make udev-hwdb rdepend on udev.
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
In daisy update-alternatives has been moved from opkg to the opkg-utils recipe.
opkg-utils now creates an additional package update-alternatives-opkg.
Without this patch the opkg-utils recipe is built and the opkg-utils package is
installed for VIRTUAL-RUNTIME_update-alternatives.
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Remove unused pinnings, use linaro toolchains for arm
including gdb
use OE-Core toolchain for providing native and crossdk
components of toolchain. this helps to choose linaro or
OE-Core toolchain in a single tmpdir and per architecture
still being able to share common components. We really
dont care who provides native and crossdk components anyway
the arm code improvements that come from linaro gcc is
paramount for crossgcc and gcc on target and candian-gcc in
case of SDK where we will use it anyway
Update comments to match current state of affairs
beautify the files a bit
Signed-off-by: Khem Raj <raj.khem@gmail.com>
