Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 20 Nov 2014 20:24:53 -0800
From:      John-Mark Gurney <jmg@funkthat.com>
To:        sbruno@freebsd.org
Cc:        freebsd-arm <freebsd-arm@freebsd.org>
Subject:   Re: ARMv6 -- pkg repo build
Message-ID:  <20141121042452.GD99957@funkthat.com>
In-Reply-To: <1416537224.7423.49.camel@bruno>
References:  <1416531121.7423.47.camel@bruno> <20141121011905.GC99957@funkthat.com> <1416537224.7423.49.camel@bruno>

next in thread | previous in thread | raw e-mail | index | archive | help
Sean Bruno wrote this message on Thu, Nov 20, 2014 at 18:33 -0800:
> On Thu, 2014-11-20 at 17:19 -0800, John-Mark Gurney wrote:
> > Sean Bruno wrote this message on Thu, Nov 20, 2014 at 16:52 -0800:
> > > http://chips.ysv.freebsd.org/build.html?mastername=11armv6-11armv6&build=2014-11-21_00h06m29s
> > > 
> > > I'm making an attempt at ARMv6 packages this week.  If you see
> > > ports that you know how to fix fail (e.g. bmake or deforas-libsystem)
> > > file a bugzilla ticket with your proposed fixes and I'll drive them into
> > > the tree.
> > > 
> > > Once this build is done, an ARMv6 package repo will be available on
> > > this server for alpha testing while we figure out some mechanical bits
> > > to the qemu-emulation I'm using.
> > 
> > re: cuse4bsd failing:
> > 
> > a) not sure why cuse4bsd is being built as a port, since it's been
> > integrated into HEAD as cuse..
> > 
> 
> Ah, that is odd.  I wonder if we've resolved all the dependencies to
> make ports depend on the base system component?

It's only been in head for 5 months, so it's possible that the porters
don't know about it yet..

Imported:
https://svnweb.freebsd.org/changeset/base/r266581

Enabled w/ version bump (1100023 has cuse):
https://svnweb.freebsd.org/changeset/base/r267440

After this I looked at the skipped ports, and there are a handful
(30+) that were skipped because of cuse being missing...

> > b) this looks like a mismatch w/ headers, trying to use machine local
> > headers instead of arm's...
> 
> Odd, this is being done in a jail, even with all my hackery, this
> shouldn't be possible.  However, if the build of the arm jail is copying
> the wrong headers into it in the first place, that's different.

Ahh, it's an issue w/ sys/conf/kmod.mk when it builds the machine
link...  In kmod.mk it does:
${.OBJDIR}/${_link}:
        @case ${.TARGET:T} in \
        machine) \
                path=${SYSDIR}/${MACHINE}/include ;; \

So, looks like MACHINE isn't set correctly, or MACHINE isn't what's
suppose to be here...  Not sure which is correct...

-- 
  John-Mark Gurney				Voice: +1 415 225 5579

     "All that I will do, has been done, All that I have, has not."



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20141121042452.GD99957>