Re: [Buildroot] [PATCH 1/1] package/libmnl: fix build failure

["Yann E. MORIN" <yann.morin.1998@free.fr>]

Dario, All,

On 2023-05-21 20:36 +0200, Dario Binacchi spake thusly:
> On Sun, May 21, 2023 at 5:16 PM Yann E. MORIN <yann.morin.1998@free.fr> wrote:
> > On 2023-05-21 17:10 +0200, Yann E. MORIN spake thusly:
> > > On 2023-05-20 23:01 +0200, Dario Binacchi spake thusly:
> > > > The added patch is an adaptation of the one submitted to the libmnl
> > > > project.
> > > >
> > > > Fixes:
> > > > - http://autobuild.buildroot.net/results/36c1619c8b02efeaf9caebae8bb5c3eb8a26238b/
> > > > - http://autobuild.buildroot.net/results/15f6dc103eec7bdeb5020acbc0ae99dc97c00e4e/
> > > Those failures are due to improper dependency on kernel headers:
> > Still, the proper way is to add a dependency on kernel headers >= 4.11
> > for the libmnl examples.
> In the libmnl repository, there are commits adding kernel headers to avoid
> dependency issues. The CAN header was missing because the added example
> was the first one related to CAN. It's worth noting that even the
> can-utils package
> synchronizes its internal headers with the kernel headers whenever they change.
> In fact, in the can-utils package in buildroot, there are no
> dependencies on kernel
> headers. If you agree, I would wait to verify if the patch I submitted
> to libmnl is
> accepted. If it is, we can apply only that patch.
> 
> What do you think?

I think that I should not have applied the CAN example patch, as it was
clearly a feature patch; I only applied it because the patch was already
applied upstream, so it was a patch we would not have to carry forever,
*and* examples are just that, examples, and are not needed for a
functional package, *and* it looked prety self-contained.

However, it b reaks the build, so this is not nice.

So, my position is that we should really add the dependency on kernel
headers in master, and if/when upstream applies the headers include
patch, then we can see to relax the restriction.

A slightly less restrictive solution would be to only include the CAN
example when the toolchain headers are fitting, e.g. something like
(with full reorganisation of examples which I refrained from doing when
applying your previous patch 271745c37a04, but now we'd need it to be
even cleaner...):

     LIBMNL_EXAMPLES_genl = genl-family-get genl-group-events

     LIBMNL_EXAMPLES_kobject = kobject-event

     LIBMNL_EXAMPLES_netfilter = \
         nfct-create-batch \
         nfct-daemon \
         nfct-dump \
         nfct-event \
         nf-log \
         nf-queue

     LIBMNL_EXAMPLES_rtnl = \
         rtnl-addr-add \
         rtnl-addr-dump \
         rtnl-link-dump \
         rtnl-link-dump2 \
         rtnl-link-dump3 \
         rtnl-link-event \
         rtnl-link-set \
         rtnl-neigh-dump \
         rtnl-route-add \
         rtnl-route-dump \
         rtnl-route-event

     ifeq ($(BR2_TOOLCHAIN_HEADERS_AT_LEAST_4_11),y)
     LIBMNL_EXAMPLES_rtnl += rtnl-link-can
     endif

     LIBMNL_EXAMPLES_INSTALL_TARGETS = \
         $(foreach t, \
             genl kobject netfilter rtnl, \
             $(addprefix examples/$(t)/,$(LIBMNL_EXAMPLES_$(t))))

Regards,
Yann E. MORIN.

> Thanks and regards,
> Dario
> >
> > Regards,
> > Yann E. MORIN.
> >
> > --
> > .-----------------.--------------------.------------------.--------------------.
> > |  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.  |
> > '------------------------------^-------^------------------^--------------------'
> 
> 
> 
> -- 
> 
> Dario Binacchi
> 
> Senior Embedded Linux Developer
> 
> dario.binacchi@amarulasolutions.com
> 
> __________________________________
> 
> 
> Amarula Solutions SRL
> 
> Via Le Canevare 30, 31100 Treviso, Veneto, IT
> 
> T. +39 042 243 5310
> info@amarulasolutions.com
> 
> www.amarulasolutions.com

-- 
.-----------------.--------------------.------------------.--------------------.
|  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.  |
'------------------------------^-------^------------------^--------------------'
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot