[Buildroot] [PATCH 14/32] package/systemd: add optional support for tmpfiles

[Maxime Hadjinlian]

Hi Arnout, Thomas, Gabe, all

On Tue, Mar 1, 2016 at 11:00 PM, Arnout Vandecappelle <arnout at mind.be> wrote:
> On 02/29/16 22:56, Thomas Petazzoni wrote:
>> Dear Gabe Evans,
>>
>> On Thu, 25 Feb 2016 21:55:23 +0000, Gabe Evans wrote:
>>> Signed-off-by: Gabe Evans <gabe at hashrabbit.co>
>>> ---
>>>  package/systemd/Config.in  | 9 +++++++++
>>>  package/systemd/systemd.mk | 6 ++++++
>>>  2 files changed, 15 insertions(+)
>>>
>>> diff --git a/package/systemd/Config.in b/package/systemd/Config.in
>>> index 2ebd825..2f89b0e 100644
>>> --- a/package/systemd/Config.in
>>> +++ b/package/systemd/Config.in
>>> @@ -118,6 +118,15 @@ config BR2_PACKAGE_SYSTEMD_QUOTACHECK
>>>
>>>        http://www.freedesktop.org/software/systemd/man/systemd-quotacheck.service.html
>>>
>>> +config BR2_PACKAGE_SYSTEMD_TMPFILES
>>> +    bool "enable tmpfiles support"
>>> +    help
>>> +      systemd-tmpfiles creates, deletes, and cleans up volatile and temporary
>>> +      files and directories, based on the configuration file format and location
>>> +      specified in tmpfiles.d(5).
>>> +
>>> +      http://www.freedesktop.org/software/systemd/man/systemd-tmpfiles.html
>>
>> Can this really be made optional? We have other packages that rely on
>> the tmpfiles.d feature:
>
>  Well, this is a bit similar to installing init scripts in /etc/init.d: that
> assumes that it's a system with our /etc/init.d/rcS and an inittab that runs it.
Indeed, note that this will causes program to fail to start and/or
behave in-appropriately (especially with a RO system).
>
>  However, it is essential that all (or many) of these systemd tools default y,
> to mimic the current situation. I think for the ones that you committed up to
> now there is no harm in defaulting to no, but this one, random-seed, logind,
> hostnamed, and hwdb should definitely default y IMHO.
random-seed could default to n without a problem I think. I agree with
you for the other.
>
>  Regards,
>  Arnout
>
>>
>> package/audit/audit.mk: mkdir -p $(TARGET_DIR)/usr/lib/tmpfiles.d
>> package/audit/audit.mk: $(INSTALL) -D -m 644 package/audit/audit_tmpfiles.conf \
>> package/audit/audit.mk:         $(TARGET_DIR)/usr/lib/tmpfiles.d/audit.conf
>> package/avahi/avahi.mk: $(INSTALL) -D -m 644 package/avahi/avahi_tmpfiles.conf \
>> package/avahi/avahi.mk:         $(TARGET_DIR)/usr/lib/tmpfiles.d/avahi.conf
>> package/dhcp/dhcp.mk:           $(TARGET_DIR)/usr/lib/tmpfiles.d/dhcpd.conf
>> package/dhcp/dhcp.mk:           $(TARGET_DIR)/usr/lib/tmpfiles.d/dhcpd.conf
>> package/nfs-utils/nfs-utils.mk: $(INSTALL) -D -m 0644 package/nfs-utils/nfs-utils_tmpfiles.conf \
>> package/nfs-utils/nfs-utils.mk:         $(TARGET_DIR)/usr/lib/tmpfiles.d/nfs-utils.conf
>>
>> Maxime, what do you think?
Also, Gabe, if you want there's many more flags that you could
enable/disable, I was planning to make a proper patch instead of the
ugly one I have in my tree, but I currently add theses flags to
systemd, that are a bit more than your selection, some of them are
maybe redundant or even useless, it's a list of flags that follows me
for a long time. But it drastically helped (along with the one you
introduced) in reducing the bloat of systemd.

I will give your series a spin as it's really interesting for me to
replace my dirty patch :)

--disable-acl
--disable-apparmor
--disable-audit
--disable-chkconfig
--disable-compat-libs
--disable-coverage
--disable-dbus
--disable-efi
--disable-elfutils
--disable-gcrypt
--disable-gnutls
--disable-gtk-doc
--disable-gtk-doc-html
--disable-gtk-doc-pdf
--disable-gudev
--disable-ima
--disable-kdbus
--disable-ldconfig
--disable-libcryptsetup
--disable-libcurl
--disable-libidn
--disable-manpages
--disable-multi-seat-x
--disable-myhostname
--disable-networkd
--disable-nls
--disable-pam
--disable-python-devel
--disable-qrencode
--disable-resolved
--disable-seccomp
--disable-selinux
--disable-smack
--disable-terminal
--disable-tests
--disable-utmp
--disable-xattr
--disable-xkbcommon
--enable-blkid


>>
>> Thanks!
>>
>> Thomas
>>
>
>
> --
> Arnout Vandecappelle                          arnout at mind be
> Senior Embedded Software Architect            +32-16-286500
> Essensium/Mind                                http://www.mind.be
> G.Geenslaan 9, 3001 Leuven, Belgium           BE 872 984 063 RPR Leuven
> LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
> GPG fingerprint:  7493 020B C7E3 8618 8DEC 222C 82EB F404 F9AC 0DDF
> _______________________________________________
> buildroot mailing list
> buildroot at busybox.net
> http://lists.busybox.net/mailman/listinfo/buildroot