Date: Mon, 9 May 2022 12:29:39 +0200 From: Milan Obuch <freebsd-ports@dino.sk> To: freebsd-ports@freebsd.org Subject: Re: Recent u-boot port change Message-ID: <20220509122939.4e8a9d48@mp2.dino.sk> In-Reply-To: <83B78F74-9C4D-4727-AC71-E21ED5E7E1AA@chrisbowman.com> References: <83B78F74-9C4D-4727-AC71-E21ED5E7E1AA@chrisbowman.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, 8 May 2022 15:29:08 -0700
Christopher Bowman <crb@chrisbowman.com> wrote:
> First, I truly do appreciate all the effort every one puts into
> FreeBSD. Apologies if I=E2=80=99m not quite following protocol here:
>
> I see that u-boot master port was recently upgraded to pull u-boot
> 2022.04. I have a private port that I=E2=80=99ve been using but looks me =
to
> an earlier version of the port since the recent one uses a version of
> u-boot I can=E2=80=99t get to work. I attempted to update my private port
> based on the new version to see if I can get it working and committed
> However, I=E2=80=99m confused how the master port ever built.
Hi,
I am building some of the u-boot ports in order to verify the process
and test them locally. I think master port is not built directly, it
serves as common place for u-boot related patches and similar things.
You should build your slave port.
> When I try to build with my patches applied I get the following
> errors:
>=20
> tools/mkeficapsule.c:18:10: fatal error: 'uuid/uuid.h' file not found
> #include <uuid/uuid.h>
> ^~~~~~~~~~~~~
> 1 error generated.
>=20
> There is a uuid.h in /usr/include so I patched this.
>=20
> Then I come across:
>=20
> tools/mkeficapsule.c:21:10: fatal error: 'gnutls/gnutls.h' file not
> found #include <gnutls/gnutls.h>
> ^~~~~~~~~~~~~~~~~
> 1 error generated.
>=20
> Ok so I install gnutls and patch the make file to also search
> /usr/local/include where that gets installed. Next I get:
>=20
> tools/mkeficapsule.c:631:8: warning: implicit declaration of function
> 'uuid_parse' is invalid in C99 [-Wimplicit-function-declaration] if
> (uuid_parse(optarg, uuid_buf)) { ^
> 1 warning generated.
>=20
> uuid_parse doesn=E2=80=99t exist in /usr/include/uudi.h but
> /usr/include/sys/uuid.h which is included by /usr/include/uuid.h does
> have a parse_uuid (note: not uuid_parse) which looks right so I patch
> again but now I get:
>=20
> tools/mkeficapsule.c:631:8: warning: implicit declaration of function
> 'parse_uuid' is invalid in C99 [-Wimplicit-function-declaration] if
> (parse_uuid(optarg, uuid_buf)) { ^
> 1 warning generated.
>=20
> at which point I give up and start to wonder how any of the other
> u-boot ports build? I mean this doesn=E2=80=99t look very specific to my
> port which is a Xilinx Zynq/arm architecture. I didn=E2=80=99t see any ot=
her
> Zynq ports in the tree so I couldn=E2=80=99t look at them for reference.
> This all happens on a 13.0 system with a recent git copy of ports.
>
I have some local ports for Zynq based boards. I just did not verified
them yet with new 2022.04 u-boot, I used old FreeBSD box for tests
which needs full upgrade (in process). But testing u-boot ports locally
is just before trying local ports...
One hint - there is freebsd-uboot mailing list, designed specifically
for u-boot related issues. While it is not much activity there, I think
it is more appropriate for this discussion...
Regards,
Milan
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20220509122939.4e8a9d48>