Date: Thu, 17 Apr 2014 13:07:08 -0600 From: Ian Lepore <ian@FreeBSD.org> To: Mark R V Murray <mark@grondar.org> Cc: Tim Kientzle <tim@kientzle.com>, freebsd-arm <freebsd-arm@FreeBSD.org>, ticso@cicely.de Subject: Re: Building an ARM/RPI-B release (hacked) on CURRENT/AMD64. Message-ID: <1397761628.1124.245.camel@revolution.hippie.lan> In-Reply-To: <ED243420-AC33-4D90-A23D-0266C142F3C1@grondar.org> References: <9FDD6F0E-B2A9-48D9-A3E4-181868995FDA@grondar.org> <EC41E53F-96EF-4652-9A02-D49448D104BE@kientzle.com> <20140417103117.GE44138@cicely7.cicely.de> <1397738961.1124.157.camel@revolution.hippie.lan> <ED243420-AC33-4D90-A23D-0266C142F3C1@grondar.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 2014-04-17 at 19:01 +0100, Mark R V Murray wrote:
> On 17 Apr 2014, at 13:49, Ian Lepore <ian@FreeBSD.org> wrote:
>=20
> > U-boot requires that a global register be set aside by the compiler a=
nd
> > it's used to access all global vars. As I vaguely understand it, u-b=
oot
> > used to want r8 for this, and clang didn't used to support the concep=
t
> > at all. Now clang supports it, but only for r9, and apparently more
> > recent u-boot expects r9 rather than r8. So the fix is probably to u=
se
> > more recent u-boot sources (I've been using 2014.01 for imx6 stuff), =
and
> > probably to add the new -ffixed-r9 flag for a clang build.
>=20
> Correct.
>=20
> The pig in trying to build u-boot 2004.04 with Clang/XDEV is the use of
>=20
> #define DECLARE_GLOBAL_DATA_PTR register volatile gd_t *=
gd asm ("r9=94)
>=20
> which means =93gd is an alias for the r9 register and is a pointer to t=
ype =85=94
>=20
> =85 I think. :-)
>=20
> Clang doesn=92t like this one bit. First objection is to =93global regi=
ster variables=94, so if I experimentally knock out the =93register=94, I=
simply get the second objection - to "multiple instances of the r9 globa=
l variable=94.
>=20
> Googling a bit suggests that Clang just plain can=92t do this. :-(
>=20
> M
Hmmm. After a bit of poking around in the llvm code, it looks like the
full extent of the support for -ffixed-r9 is that it doesn't consider
that register available for use by the code generator; that's only part
of what u-boot needs. =20
Some online notes I found for clang 3.5 claim that global register
variables aren't supported, and aren't likely to be any time soon.
-- Ian
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1397761628.1124.245.camel>