public inbox for buildroot@busybox.net
 help / color / mirror / Atom feed
From: Yann E. MORIN <yann.morin.1998@free.fr>
To: buildroot@busybox.net
Subject: [Buildroot] [PATCH v3 2/2] package/firewalld: new package
Date: Sat, 25 Apr 2020 00:14:48 +0200	[thread overview]
Message-ID: <20200424221448.GM5035@scaer> (raw)
In-Reply-To: <CAFSsvmqFEN7TM5JQOaeMrTd2KaxT6K=7xAku21YvzuLjW9-psg@mail.gmail.com>

Adam, All,

On 2020-04-24 12:14 -0700, Adam Duskett spake thusly:
> On Thu, Apr 23, 2020 at 2:19 PM Yann E. MORIN <yann.morin.1998@free.fr> wrote:
> >
> > On 2020-04-23 10:33 -0700, aduskett at gmail.com spake thusly:
> > > From: Adam Duskett <Aduskett@gmail.com>
> > >
> > > Firewalld provides a dynamically managed firewall with
> > > support for network or firewall zones to define the trust level of network
> > > connections or interfaces.
> > >
> > > Items of note:
> > >
> > >  - Iptables is set as a requirement because many applications such as Docker
> > >    require direct-passthrough rules which firewalld does not support.
> > >
> > >  - Setting FIREWALLD_AUTORECONF=YES results in the following error:
> > >    "error: cannot find input file: `po/Makefile.in.in'" As such, a manual call
> > >    to autogen.sh is necessary with PATH=$(BR_PATH) set to ensure the script
> > >    uses the host intltoolize binary.
> >
> > I think we can do better. ;-)
> >
> >     # Prpare the tree as autogen.sh would; allows using _AUTORECONF=YES
> >     define FIREWALLD_INTLTOOLIZE
> >         ls -1 $(@D)/po/*.po |sed -r -e 's:.*/::; s:.po$$::' >$(@D)/po/LINGUAS
> >         cd $(@D) && $(HOST_DIR)/bin/intltoolize --force --automake
> >     endef
> >     FIREWALLD_PRE_CONFIGURE_HOOKS += FIREWALLD_INTLTOOLIZE
> >
> I heavily disagree:
> 1) The existing autogen.sh script may change in the future which would
> make this macro
>     more unmaintanable.
> 2) package/sdl/sdl.mk has PATH=$(BR_PATH) ./autogen.sh, which means
> there is already
>     a precident for this kind of logic.

Few things:

 1- the autogen.sh script calls autoreconf with a different set of
    arguments as our infra does, notably it does not pass the paths to
    the other tools, and as such does not pass the paths to the autoconf
    local include directories;

 2- the SDL autoreconf does not work with our AUTORECONF=YES infra,
    while firewalld should work nicely after the intltoolize preparation
    step.

 3- indeed it puts a little bit more burden on our shoulders, but in
    fact all this discussion is moot is you switch to using the released
    tarball, rather than a git cnashot: be careful that the github
    helper macro will point to a github-generated tarball. However,
    firewalld maintainers has uploaded a released tarball:

    https://github.com/firewalld/firewalld/releases/download/v0.8.2/firewalld-0.8.2.tar.gz

So, if you switch to that, we no longer need to autoreconf (unless
you're going to patch one of the autotools files...)

> > >  - Setting PYTHON="/usr/bin/env python$(PYTHON3_VERSION_MAJOR)" prevents
> > >    Firewalld from setting the shebang in the installed python files to the full
> > >    path to the python interpreter used when building.
> > >
> > >  - The bundled provided SYSV init file has several bashisms and requires
> > >    /etc/init.d/functions which buildroot doesn't provide. So instead, we
> > >    provide a more simple init.d file in the package directory, which does not
> > >    require bash.
> > >
> > > Signed-off-by: Adam Duskett <Aduskett@gmail.com>
> > > ---
> > [--SNIP--]
> > > diff --git a/package/firewalld/firewalld.hash b/package/firewalld/firewalld.hash
> > > new file mode 100644
> > > index 0000000000..93cf67ccea
> > > --- /dev/null
> > > +++ b/package/firewalld/firewalld.hash
> > > @@ -0,0 +1,3 @@
> > > +# Locally computed
> > > +sha256  769c57d5c49218821ad30639a3f802831f9618899e403a811802c56e30f93c79  firewalld-v0.8.2.tar.gz
> > > +sha256  8177f97513213526df2cf6184d8ff986c675afb514d4e68a404010521b880643  COPYING
> > > diff --git a/package/firewalld/firewalld.mk b/package/firewalld/firewalld.mk
> > > new file mode 100644
> > > index 0000000000..aedc250be1
> > > --- /dev/null
> > > +++ b/package/firewalld/firewalld.mk
> > > @@ -0,0 +1,84 @@
> > > +# iptables, ip6tables, ebtables, and ipset *should* be unnecessary
> > > +# when the nftables backend is available, because nftables supersedes all of
> > > +# them. However we still need to build and install iptables and ip6tables
> > > +# because application relying on direct passthrough rules (IE docker) will
> > > +# break.
> >
> > So it should nt be the responsibility of firewalld to select iptables or
> > ebtables, but the individual applications, like docker, which should
> > either unconditionally select them, or conditionally select them, like
> > for example:
> >
> Possibly? This change would have the potential to make things far more
> complicated,
> as docker is just one example. I can't find a distribution (including
> OE) that does not always
> compile firewalld with iptables.

OK, so this comment of mine is moot: I read in diagonal, and too fast.
Ignore it and keep your code+comment as-is, sorry...

Yet, there is still a little nit, below...

> >     seelct BR2_PACKAGE_IPTABLES if BR2_PACKAGE_FIREWALLD
> >
> > > +# /etc/sysconfig/firewalld is a Red Hat-ism, only referenced by
> > > +# the Red Hat-specific init script which isn't used, so we set
> > > +# --disable-sysconfig.
> > > +FIREWALLD_CONF_OPTS += \
> > > +     --disable-rpmmacros \
> > > +     --disable-sysconfig \
> > > +     --with-ip6tables-restore=/usr/sbin/ip6tables-restore \
> > > +     --with-ip6tables=/usr/sbin/ip6tables \
> > > +     --with-iptables-restore=/usr/sbin/iptables-restore \
> > > +     --with-iptables=/usr/sbin/iptables \
> > > +     --with-nft=/usr/sbin/nft \
> > > +     --without-ebtables \
> > > +     --without-ebtables-restore \
> > > +     --without-ipset \

Why are you selecting BR2_PACKAGE_EBTABLES and BR2_PACKAGE_IPSET in
Config.in if you explicitly disable support for them here?

Regards,
Yann E. MORIN.

> > > +     --without-xml-catalog
> > > +
> > > +ifeq ($(BR2_PACKAGE_SYSTEMD),y)
> > > +FIREWALLD_DEPENDENCIES += systemd
> > > +FIREWALLD_CONF_OPTS += --with-systemd-unitdir=/usr/lib/systemd/system
> >
> > We also want an explicit --enable-systemd
> >
> > This was just a quick review, I may come back to it later in the WE...
> >
> > Regards,
> > Yann E. MORIN.
> >
> > > +else
> > > +FIREWALLD_CONF_OPTS += --disable-systemd
> > > +endif
> > > +
> > > +define FIREWALLD_INSTALL_INIT_SYSTEMD
> > > +     $(INSTALL) -D -m 0644 $(@D)/config/firewalld.service \
> > > +             $(TARGET_DIR)/usr/lib/systemd/system/firewalld.service
> > > +endef
> > > +
> > > +# The bundled sysvinit file requires /etc/init.d/functions which is not
> > > +# provided by buildroot. As such, we provide our own firewalld init file.
> > > +define FIREWALLD_INSTALL_INIT_SYSV
> > > +     $(INSTALL) -D -m 0755 $(FIREWALLD_PKGDIR)/S41firewalld \
> > > +             $(TARGET_DIR)/etc/init.d/S41firewalld
> > > +endef
> > > +
> > > +$(eval $(autotools-package))
> > > --
> > > 2.25.3
> > >
> > > _______________________________________________
> > > buildroot mailing list
> > > buildroot at busybox.net
> > > http://lists.busybox.net/mailman/listinfo/buildroot
> >
> > --
> > .-----------------.--------------------.------------------.--------------------.
> > |  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
> > | +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
> > | +33 561 099 427 `------------.-------:  X  AGAINST      |  \e/  There is no  |
> > | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
> > '------------------------------^-------^------------------^--------------------'

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 561 099 427 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'

  reply	other threads:[~2020-04-24 22:14 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-04-23 17:33 [Buildroot] [PATCH v3 1/2] package/python-slip-dbus: new package aduskett at gmail.com
2020-04-23 17:33 ` [Buildroot] [PATCH v3 2/2] package/firewalld: " aduskett at gmail.com
2020-04-23 21:19   ` Yann E. MORIN
2020-04-24 19:14     ` Adam Duskett
2020-04-24 22:14       ` Yann E. MORIN [this message]
2020-04-23 20:58 ` [Buildroot] [PATCH v3 1/2] package/python-slip-dbus: " Yann E. MORIN

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=20200424221448.GM5035@scaer \
    --to=yann.morin.1998@free.fr \
    --cc=buildroot@busybox.net \
    /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