From: "'Heinisch, Alexander' via isar-users" <isar-users@googlegroups.com>
To: "Prusty, Badrikesh" <badrikesh.prusty@siemens.com>,
"isar-users@googlegroups.com" <isar-users@googlegroups.com>,
"wzh@ilbers.de" <wzh@ilbers.de>
Cc: "Kiszka, Jan" <jan.kiszka@siemens.com>,
"MOESSBAUER, Felix" <felix.moessbauer@siemens.com>,
"Cetin, Gokhan" <gokhan.cetin@siemens.com>
Subject: Re: [PATCH v7] image-postproc: gate systemd preset-all on masked unit state
Date: Tue, 14 Jul 2026 05:08:57 +0000 [thread overview]
Message-ID: <523277bfe9f93b4cdfb2a63b4c8f23607a7371f0.camel@siemens.com> (raw)
In-Reply-To: <SG2PR06MB51077C6DF7E9182EE88C71C991F92@SG2PR06MB5107.apcprd06.prod.outlook.com>
On Tue, 2026-07-14 at 04:05 +0000, Prusty, Badrikesh (FT FDS CES LX PBU
2) wrote:
> Hi,
>
> > I'd propose to revert that "fix" and fix the preset shipped with
> > your distribution accordingly:
> >
> > e.g.:
> > ```
> > echo "disable reboot.target" > /usr/lib/systemd/system-preset/00-
> > reboot-target-mask.preset
> > ```
>
> Thanks for the suggestion!
>
> Unfortunately, shipping a preset file with a "disable" rule won't
> work
> in this case, because populate_systemd_preset applies presets
> with:
>
> systemctl preset-all --preset-mode="enable-only"
>
> This means "disable" rules in preset files are silently ignored.
>
The preset-all with enable-only won't take care about the disabling nor
about masking, you just have to resolve the conflict between masking
and the preset for said unit by unsetting the units/targets preset
enable policy. (you could also just remove the line that states preset
enable for said unit - but imo the drop in config version is nicer)
The proposed handling below is not a replacement for masking the unit,
but a fix for the error you experienced during preset-all.
Based on the example you gave, that would result in something like:
```
cat meta-isar/recipes-example/no-ctrl-alt-del/files/postinst
#!/bin/sh
systemctl mask ctrl-alt-del.target
echo "disable reboot.target" > /usr/lib/systemd/system-preset/00-
reboot-target-mask.preset
```
> To make disable rules effective, the --preset-mode flag would need to
> be dropped from populate_systemd_preset.
>
>
> Many thanks,
> Badrikesh
>
> -----Original Message-----
> From: Heinisch, Alexander (FT RPD CED SES-AT)
> <alexander.heinisch@siemens.com>
> Sent: 01 July 2026 15:12
> To: Prusty, Badrikesh (FT FDS CES LX PBU 2)
> <badrikesh.prusty@siemens.com>; isar-users@googlegroups.com;
> wzh@ilbers.de
> Cc: Kiszka, Jan (FT RPD CED) <jan.kiszka@siemens.com>; Moessbauer,
> Felix (FT RPD CED OES-DE) <felix.moessbauer@siemens.com>; Çetin,
> Gökhan (FT D EU TR C&E) <gokhan.cetin@siemens.com>
> Subject: Re: [PATCH v7] image-postproc: gate systemd preset-all on
> masked unit state
>
> Thx for clarifying steps to reproduce.
>
> I followed your steps below with a minor modification to build an
> existing image (isar-image-base)
>
> ```
> ./kas/kas-container build kas/isar.yaml:kas/machine/x86-
> pc.yaml:kas/distro/debian-
> bookworm.yaml:kas/package/pkg_hello.yaml:kas/image/isar-image-
> base.yaml
> ```
>
> I could reproduce described log output in bookworm.
>
> But, the error message is caused by a conflict you introduced by
> masking /usr/lib/systemd/system/ctrl-alt-del.target!
Note: My proposed solution is about resolving described conflict, not
about a replacement for masking the unit.
BR Alexander
>
> ctrl-alt-del.target is a link to reboot.target which is explicitly
> mentioned enabled in /usr/lib/systemd/system-preset/90-systemd.preset
>
> Therefore, systemd (in bookworm) emits an error.
> Later versions of systemd (e.g. the one shipped with trixie) handle
> that more gracefully [1], as mentioned previously.
>
>
> TL;DR:
>
> I'd propose to revert that "fix" and fix the preset shipped with your
> distribution accordingly:
>
> e.g.:
> ```
> echo "disable reboot.target" > /usr/lib/systemd/system-preset/00-
> reboot-target-mask.preset
> ```
>
> BR Alexander
>
>
> [1]:
> https://github.com/systemd/systemd/commit/d41f08bd2ac4cfa58191b64d31ca9e6f3dec7552
>
> On Sat, 2026-06-20 at 04:35 +0000, Prusty, Badrikesh (FT FDS CES LX
> PBU
> 2) wrote:
> > Hi,
> >
> > Here are the simple steps to reproduce the issue:
> >
> > 1. Check out Isar without this patch applied.
> >
> > 2. Create a recipe to mask a unit. Here I am using `ctrl-alt-
> > del.target` ```
> > > cat meta-isar/recipes-example/no-ctrl-alt-del/no-ctrl-alt-del.bb
> > inherit dpkg-raw
> >
> > SRC_URI = "file://postinst"
> >
> > > cat meta-isar/recipes-example/no-ctrl-alt-del/files/postinst
> > #!/bin/sh
> >
> > systemctl mask ctrl-alt-del.target
> > ```
> >
> > 3. Install it to the rootfs:
> > ```
> > > git diff
> > diff --git a/kas/package/pkg_hello.yaml
> > b/kas/package/pkg_hello.yaml
> > index ea2bce8e..382acd29 100644
> > --- a/kas/package/pkg_hello.yaml
> > +++ b/kas/package/pkg_hello.yaml
> > @@ -6,4 +6,4 @@ header:
> >
> > local_conf_header:
> > package-hello: |
> > - IMAGE_INSTALL:append = " hello"
> > + IMAGE_INSTALL:append = " hello no-ctrl-alt-del"
> > ```
> >
> > 4. Trigger a build with distro set to debian-bookworm:
> > ```
> > > ./kas/kas-container build kas/isar.yaml:kas/machine/x86-
> > > pc.yaml:kas/distro/debian-
> > > bookworm.yaml:kas/package/pkg_hello.yaml
> > ```
> >
> > Error log:
> > ```
> > > DEBUG: Executing shell function
> > > image_postprocess_populate_systemd_preset
> > > Failed to preset unit, unit /etc/systemd/system/systemd-
> > > firstboot.service is masked.
> > > Failed to preset unit, unit /etc/systemd/system/ctrl-alt-
> > > del.target
> > > is masked.
> > > Failed to preset unit, unit /lib/systemd/system/x11-
> > > common.service
> > > is masked.
> > > Failed to preset unit, unit /lib/systemd/system/rc.service is
> > > masked.
> > > Failed to preset unit, unit
> > > /lib/systemd/system/cryptdisks.service
> > > is masked.
> > > Failed to preset unit, unit /lib/systemd/system/cryptdisks-
> > > early.service is masked.
> > > Failed to preset unit, unit /lib/systemd/system/rcS.service is
> > > masked.
> > > Failed to preset unit, unit /lib/systemd/system/hwclock.service
> > > is
> > > masked.
> > > Unit /etc/systemd/system/systemd-firstboot.service is masked,
> > > ignoring.
> > > Unit /etc/systemd/system/ctrl-alt-del.target is masked, ignoring.
> > > Created symlink /etc/systemd/system/multi-
> > > user.target.wants/remote-
> > > cryptsetup.target → /lib/systemd/system/remote-cryptsetup.target.
> > > Created symlink /etc/systemd/system/sysinit.target.wants/systemd-
> > > sysext.service → /lib/systemd/system/systemd-sysext.service.
> > > Unit /lib/systemd/system/x11-common.service is masked, ignoring.
> > > Unit /lib/systemd/system/rc.service is masked, ignoring.
> > > Failed to preset unit, file "/etc/systemd/system/ctrl-alt-
> > > del.target" already exists and is a symlink to "/dev/null".
> > ```
> >
> > Many thanks,
> > Badrikesh
> >
> > -----Original Message-----
> > From: Heinisch, Alexander (FT RPD CED SES-AT)
> > <alexander.heinisch@siemens.com>
> > Sent: 20 June 2026 02:07
> > To: Prusty, Badrikesh (FT FDS CES LX PBU 2)
> > <badrikesh.prusty@siemens.com>; isar-users@googlegroups.com;
> > wzh@ilbers.de
> > Cc: Kiszka, Jan (FT RPD CED) <jan.kiszka@siemens.com>; Moessbauer,
> > Felix (FT RPD CED OES-DE) <felix.moessbauer@siemens.com>; Çetin,
> > Gökhan (FT D EU TR C&E) <gokhan.cetin@siemens.com>
> > Subject: Re: [PATCH v7] image-postproc: gate systemd preset-all on
> > masked unit state
> >
> > On Mon, 2026-06-15 at 11:02 +0200, Zhihang Wei wrote:
> > > Applied to next, thanks.
> > >
> > > Zhihang
> > >
> > > On 4/29/26 15:02, 'Badrikesh Prusty' via isar-users wrote:
> > > > Skip systemd preset-all during image postprocessing if masked
> > > > unit
> > > > files are detected, to avoid non-fatal failures when presetting
> > > > units that are already masked.
> > > >
> > > > Move execution fully inside a single chroot invocation for
> > > > consistency and reduced overhead.
> > > >
> > > > Fixes: 72b88a12 ("classes/image-postproc: Enable systemd units
> > > > based on systemd presets")
> > Sorry for the (too)late reply.
> >
> > Can you elaborate what this fix is about?
> >
> > At least on my systems (and according to documentation) preset-all
> > --
> > preset-mode="enable-only" handles masked services gracefully.
> >
> > The command emits:
> > "Unit [redacted].service is masked, ignoring." correctly.
> > but exits with exit code 0.
> >
> > Further, the previous behavior is exactly what systemd does on
> > first
> > boot anyways.
> >
> > BR Alexander
> >
> > > >
> > > > Signed-off-by: Badrikesh Prusty <badrikesh.prusty@siemens.com>
> > > > ---
> > > > meta/classes-recipe/rootfs.bbclass | 13 +++++++------
> > > > 1 file changed, 7 insertions(+), 6 deletions(-)
> > > >
> > > > diff --git a/meta/classes-recipe/rootfs.bbclass b/meta/classes-
> > > > recipe/rootfs.bbclass index 8b502a50..d250690f 100644
> > > > --- a/meta/classes-recipe/rootfs.bbclass
> > > > +++ b/meta/classes-recipe/rootfs.bbclass
> > > > @@ -569,13 +569,14 @@ EOSUDO
> > > >
> > > > ROOTFS_POSTPROCESS_COMMAND +=
> > > > "${@bb.utils.contains('ROOTFS_FEATURES', 'populate-systemd-
> > > > preset', 'image_postprocess_populate_systemd_preset', '', d)}"
> > > > image_postprocess_populate_systemd_preset() {
> > > > - SYSTEMD_INSTALLED=$(sudo chroot '${ROOTFSDIR}' dpkg-query
> > > > \
> > > > - --showformat='${db:Status-Status}' \
> > > > - --show systemd || echo "" )
> > > > + sudo chroot "${ROOTFSDIR}" /bin/sh <<'EOSH'
> > > > + SYSTEMD_INSTALLED=$(dpkg-query --
> > > > showformat='${db:Status-
> > > > Status}' --show systemd 2>/dev/null)
> > > >
> > > > - if (test "$SYSTEMD_INSTALLED" = "installed"); then
> > > > - sudo chroot '${ROOTFSDIR}' systemctl preset-all --
> > > > preset-
> > > > mode="enable-only"
> > > > - fi
> > > > + if [ "${SYSTEMD_INSTALLED}" = "installed" ]; then
> > > > + systemctl list-unit-files --state=masked --no-
> > > > legend
> > > > >
> > > > grep -q '^' \
> > > > + || systemctl preset-all --preset-mode="enable-
> > > > only"
> > > > + fi
> > > > +EOSH
> > > > }
> > > >
> > > > do_rootfs_postprocess[vardeps] =
> > > > "${ROOTFS_POSTPROCESS_COMMAND}"
> > > > --
> > > > 2.47.3
> > > >
> >
> > --
> > Alexander Heinisch
> > Siemens AG
> > http://www.siemens.com/
>
> --
> Alexander Heinisch
> Siemens AG
> http://www.siemens.com/
--
Alexander Heinisch
Siemens AG
http://www.siemens.com/
--
You received this message because you are subscribed to the Google Groups "isar-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to isar-users+unsubscribe@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/isar-users/523277bfe9f93b4cdfb2a63b4c8f23607a7371f0.camel%40siemens.com.
prev parent reply other threads:[~2026-07-14 5:09 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-04-29 13:02 'Badrikesh Prusty' via isar-users
2026-06-15 9:02 ` Zhihang Wei
2026-06-19 20:37 ` 'Heinisch, Alexander' via isar-users
2026-06-20 4:35 ` 'Prusty, Badrikesh' via isar-users
2026-07-01 9:41 ` 'Heinisch, Alexander' via isar-users
2026-07-14 4:05 ` 'Prusty, Badrikesh' via isar-users
2026-07-14 5:08 ` 'Heinisch, Alexander' via isar-users [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=523277bfe9f93b4cdfb2a63b4c8f23607a7371f0.camel@siemens.com \
--to=isar-users@googlegroups.com \
--cc=alexander.heinisch@siemens.com \
--cc=badrikesh.prusty@siemens.com \
--cc=felix.moessbauer@siemens.com \
--cc=gokhan.cetin@siemens.com \
--cc=jan.kiszka@siemens.com \
--cc=wzh@ilbers.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox