This includes the following changes:
75985aa Add Kconfig option for new systemd files
bdc1d3c systemd generic startup
16790d1 doc: fix broken link for talks
7c8926d ssbl_handler: fix age comparison
f130541 travis: set libwebsockets version
9409e9a Prepare 2019.11-rc1
d8d9fd4 handler: add a copy handler
21594da Add SSBL Handler
34c55f1 handlers: sort the list of modules in Makefile
ede66e7 doc: add scenarios how to use SWUpdate
a5e82b9 doc: add links to single and double copy scenarios
c744dfe Cosmetic: rename entry points in raw handler
864dedd raw handler: return error number instead of -1
e38ad73 handler: push progress status from bootloader handler
cfaec8c u-boot handler: fix removing variable
27c40c4 doc: ELC video published on YouTube
07811c7 swupdate: add pre-update command hook
da974cd doc: update list of talks with ELC 2019
Signed-off-by: Adrian Freihofer <adrian....@siemens.com>
Hi Adrian,
This includes the following changes: 75985aa Add Kconfig option for new systemd files bdc1d3c systemd generic startup 16790d1 doc: fix broken link for talks 7c8926d ssbl_handler: fix age comparison f130541 travis: set libwebsockets version 9409e9a Prepare 2019.11-rc1 d8d9fd4 handler: add a copy handler 21594da Add SSBL Handler 34c55f1 handlers: sort the list of modules in Makefile ede66e7 doc: add scenarios how to use SWUpdate a5e82b9 doc: add links to single and double copy scenarios c744dfe Cosmetic: rename entry points in raw handler 864dedd raw handler: return error number instead of -1 e38ad73 handler: push progress status from bootloader handler cfaec8c u-boot handler: fix removing variable 27c40c4 doc: ELC video published on YouTube 07811c7 swupdate: add pre-update command hook da974cd doc: update list of talks with ELC 2019
Your changelog it's wrong, it's missing:
045a618 raw_handler: handle ro block devices
5700346 Makefile: rename tools binaries
48f22c6 add .editorconfig file
e5e7e02 doc: describe REST api
Signed-off-by: Adrian Freihofer <adrian.f...@siemens.com>
--- recipes-support/swupdate/swupdate_git.bb | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/recipes-support/swupdate/swupdate_git.bb b/recipes-support/swupdate/swupdate_git.bb index 3f69b87..9fea83a 100644 --- a/recipes-support/swupdate/swupdate_git.bb +++ b/recipes-support/swupdate/swupdate_git.bb @@ -3,5 +3,5 @@ require swupdate_tools.inc DEFAULT_PREFERENCE = "-1" -SRCREV ?= "a8d45381253838b34e4e035c065a1cdf429b1f09" +SRCREV ?= "045a618a725d0a2fce64161f10101c0004ac5d85" PV = "2019.04+git${SRCPV}"
Best Regards,
Joris
> Signed-off-by: Adrian Freihofer <adrian....@siemens.com>
Hi Stefan,There is an option in Kconfig but this option is not useful for a Yocto build system. Therefore the default value of Kbuild allows to provide the value like done in the line you are referring.systemd_system_unitdir is a system specific variable provided by Yocto. Every other value as the one set by this line, would break almost everything here. What's the advantage of a configuration flag where exactly one value is valid? Or what is the other value you would like to configure?
--
You received this message because you are subscribed to the Google Groups "swupdate" group.
To unsubscribe from this group and stop receiving emails from it, send an email to swupdate+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/swupdate/962bbe3f-6185-4c7f-b539-51af4546fbeb%40googlegroups.com.
Hi Adrian,
Am 23.10.19 um 23:11 schrieb Adrian Freihofer:
> In case of singed and/or encrypted images the corresponding keys and
> certificates need to be installed into the image.
>
> If the variables SWUPDATE_CMS_CERT and SWUPDATE_AES_FILE are set for
> the image (not only for the image-update) as well, the required
> certificate and key files get installed and the -k and the -K paramter
> are added to the swupdate configuration.
>
> Signed-off-by: Adrian Freihofer <adrian....@siemens.com>
Am 23.10.19 um 23:10 schrieb Adrian Freihofer:
> Support the following in a sw-description:
> Â Â filename = "@@ROOTFS_IMAGE@@";
> Â Â sha256 = "@@@ROOTFS_IMAGE@@";
>
> First the @@ROOTFS_IMAGE@@ gets replaced e.g. by @my-image. Second
> @my-image gets replaced by the sha256 hash.
>
Please don't use a single @ and use an obviously solution which is easy
readable. Maybe we should use ${} to support nested variables:
$${ROOTFS_IMAGE}_SHA256}
Alternative a sw-description.bbclass could create the sw-description
on-the-fly.
Regards
  Stefan
Hi Adrian,
Am 24.10.19 um 12:43 schrieb adrian....@gmail.com:
> Hi Stefan,
>
> There is an option in Kconfig but this option is not useful for a Yocto
> build system.
Why?
Couldn't you append the value to the .config?
> Therefore the default value of Kbuild allows to provide
> the value like done in the line you are referring.
> systemd_system_unitdir is a system specific variable provided by Yocto.
> Every other value as the one set by this line, would break almost
> everything here. What's the advantage of a configuration flag where
> exactly one value is valid? Or what is the other value you would like to
> configure?
The problem is the break of the common design. You have three step:
configure, build, install. You break this concept because you set a
configuration option during install. All other build tools (autotools,
cmake, meson) use the configuration step to configure the installation step.
Regards
  Stefan
Hi Adrian,
On 23/10/19 23:10, Adrian Freihofer wrote:
> swupdate-www requires the swupdate-progress service from swupdate-tools
> package.
>
Why ? swupdate-progress is not mandatory. I can have swupdate-www
(Website) without swupdate-progress.
Best regards,
Stefano
> Signed-off-by: Adrian Freihofer <adrian....@siemens.com>
Best regards,
Stefano Babic
>
> Signed-off-by: Adrian Freihofer <adrian....@siemens.com>
> Signed-off-by: Adrian Freihofer <adrian....@siemens.com>
> + Â Â Â Â grep -v 'CONFIG_SYSTEMD' ${WORKDIR}/defconfig > ${S}/.config
> + Â Â Â Â echo "# Global settings from swupdate recipe" >> ${S}/.config
> + Â Â Â Â echo "CONFIG_SYSTEMD=y" >> ${S}/.config
> + Â Â Â Â echo "CONFIG_SYSTEMD_SYSTEM_UNITDIR=\"${systemd_system_unitdir}\"" >> ${S}/.config
I understand the logic, but I have an understanding problem. In fact,
who is the master ?
In current implementation, defconfig is the master. If a user run
"bitbake -c menuconfig swupdate" or add fragments to the configuration,
this becomes the master. In fact, the current recipes do not try to
change defconfig, but set OE to reflect the values in defconfig. This is
the reason for the anonymous python function in swupdate.inc. This adds
DEPENDS to the list, quite an exception in OE. Because the list of
depends can be short or long, on depends of the configuration.
However, the single case above seems "quite" ok to me. CONFIG_SYSTEMD
comes from defconfig. But should not be responsibility of the user to
set CONFIG_SYSTEMD_SYSTEM_UNITDIR instead of forcing it ?
Hi Stefan, Adrian,
On 03/11/19 09:48, Stefan Herbrechtsmeier wrote:
> Hi Adrian and Stefano,
>
> Am 02.11.19 um 22:48 schrieb adrian....@gmail.com: