From: "Artur Wiebe" <artur@4wiebe.de>
To: ptxdist@pengutronix.de
Subject: Re: [ptxdist] ?==?utf-8?q? ?==?utf-8?q? [PATCH] machine-id: make using rc-once optional
Date: Fri, 27 Nov 2020 10:52:05 +0100 [thread overview]
Message-ID: <234bd5-5fc0cc80-9-426f2080@104488028> (raw)
In-Reply-To: <20201127083841.GC1592430@pengutronix.de>
Hi Michael,
On Friday, November 27, 2020 09:38 CET, Michael Olbrich <m.olbrich@pengutronix.de> wrote:
> On Wed, Nov 25, 2020 at 05:56:49PM +0100, Artur Wiebe wrote:
> > If not set only an empty /etc/machine-id will be installed.
> >
> > Signed-off-by: Artur Wiebe <artur@4wiebe.de>
> > ---
> > rules/machine-id.in | 22 ++++++++++++++++++----
> > rules/machine-id.make | 4 ++++
> > rules/machine-id.postinst | 5 +++--
> > 3 files changed, 25 insertions(+), 6 deletions(-)
> >
> > diff --git a/rules/machine-id.in b/rules/machine-id.in
> > index f18fec554..2a3901f41 100644
> > --- a/rules/machine-id.in
> > +++ b/rules/machine-id.in
> > @@ -1,11 +1,25 @@
> > ## SECTION=initmethod
> >
> > -config MACHINE_ID
> > +menuconfig MACHINE_ID
> > tristate
> > prompt "create /etc/machine-id"
> > - select RC_ONCE if RUNTIME
> > - select UTIL_LINUX_NG if !INITMETHOD_SYSTEMD
> > - select UTIL_LINUX_NG_UUIDGEN if !INITMETHOD_SYSTEMD
> > + select RC_ONCE if MACHINE_ID_RC_ONCE && RUNTIME
> > + select UTIL_LINUX_NG if MACHINE_ID_RC_ONCE && !INITMETHOD_SYSTEMD
> > + select UTIL_LINUX_NG_UUIDGEN if MACHINE_ID_RC_ONCE && !INITMETHOD_SYSTEMD
> > help
> > Enable this to make sure there is always a unique /etc/machine-id
> > available in the system.
> > +
> > +if MACHINE_ID
> > +
> > +config MACHINE_ID_RC_ONCE
> > + bool "generate (bbinit) or save (systemd) machine-id using rc-once at first boot"
> > + default y
>
> Hmmm, maybe should have no prompt and 'depends on RC_ONCE' (instead of
> selecting it above). I don't think there is a use-case for not persisting
> the machine-id if rc-once is available.
>
> This is different from the ssh server key case: There are other ways to
> create the keys.
Why should one select rc-once, if it is not needed by any other package? Maybe make it even invisible?
In my projects I use an initramfs where I mount an overlayfs over /etc. When systemd starts /etc/mancine-id is writable while / is never mounted rw.
So I want just to install an empty /etc/machine-id and let systemd generate and write the new one. rc-once just jeopardize this.
>
> > + help
> > + If not set only an empty /etc/machine-id will be installed.
> > + Systemd:
> > + If /etc/machine-id exists and is empty, systemd will bind-mount
> > + a file containing the new machine-id over it and later try
> > + to commit it to disk (if /etc/ is writable).
> > +
> > +endif
> > diff --git a/rules/machine-id.make b/rules/machine-id.make
> > index 3883d64cb..e1cbdcc42 100644
> > --- a/rules/machine-id.make
> > +++ b/rules/machine-id.make
> > @@ -27,7 +27,11 @@ $(STATEDIR)/machine-id.targetinstall:
> > @$(call install_fixup,machine-id,AUTHOR,"Michael Olbrich <m.olbrich@pengutronix.de>")
> > @$(call install_fixup,machine-id,DESCRIPTION,missing)
> >
> > + @$(call install_copy, machine-id, 0, 0, 0644, /dev/null, /etc/machine-id)
>
> Maybe install_alternative and an empty on in projectroot/. Sometimes it's
> better to share the same machine-id across devices than having a new one
> after rebooting.
Good idea. I'll change this.
>
> Also, this should only happen for !PTXCONF_MACHINE_ID_RC_ONCE. It's not a
> good idea to modify files that are installed by a package.
This is the case for almost all files in /etc if they are modified on the target. I avoid it with the overlayfs. In this way defaults are installed by the packages and changes are stored on the target.
>
> > +
> > +ifdef MACHINE_ID_RC_ONCE
>
> ifdef PTXCONF_MACHINE_ID_RC_ONCE
>
> > @$(call install_alternative, machine-id, 0, 0, 0755, /etc/rc.once.d/machine-id)
> > +endif
> >
> > @$(call install_finish,machine-id)
> >
> > diff --git a/rules/machine-id.postinst b/rules/machine-id.postinst
> > index 2060129d1..f1648226b 100644
> > --- a/rules/machine-id.postinst
> > +++ b/rules/machine-id.postinst
> > @@ -1,3 +1,4 @@
> > #!/bin/sh
> > -touch "$DESTDIR/etc/machine-id"
>
> So this is still needed inside the 'if'.
I moved it into the package because otherwise an empty package is not generated.
ERROR: no files installed
With this patch machine-id would be reset on reinstall, but why should someone reinstall the machine-id package?
Artur
>
> Michael
>
> > -$DESTDIR/usr/sbin/enable-rc-once machine-id
> > +if [ -f $DESTDIR/etc/rc.once.d/machine-id ]; then
> > + $DESTDIR/usr/sbin/enable-rc-once machine-id
> > +fi
> > --
> > 2.29.2
> >
> >
> > _______________________________________________
> > ptxdist mailing list
> > ptxdist@pengutronix.de
> > To unsubscribe, send a mail with subject "unsubscribe" to ptxdist-request@pengutronix.de
> >
>
> _______________________________________________
> ptxdist mailing list
> ptxdist@pengutronix.de
> To unsubscribe, send a mail with subject "unsubscribe" to ptxdist-request@pengutronix.de
_______________________________________________
ptxdist mailing list
ptxdist@pengutronix.de
To unsubscribe, send a mail with subject "unsubscribe" to ptxdist-request@pengutronix.de
next prev parent reply other threads:[~2020-11-27 9:52 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-11-25 16:56 [ptxdist] " Artur Wiebe
2020-11-27 8:38 ` Michael Olbrich
2020-11-27 9:52 ` Artur Wiebe [this message]
2020-11-27 11:02 ` Ladislav Michl
2020-12-04 6:32 Michael Olbrich
2020-12-04 9:33 ` [ptxdist] ?==?utf-8?q? ?==?utf-8?q? " Artur Wiebe
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=234bd5-5fc0cc80-9-426f2080@104488028 \
--to=artur@4wiebe.de \
--cc=ptxdist@pengutronix.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