[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20110928231112.GE2838@ponder.secretlab.ca>
Date: Wed, 28 Sep 2011 18:11:12 -0500
From: Grant Likely <grant.likely@...retlab.ca>
To: Arnd Bergmann <arnd@...db.de>
Cc: Mark Salter <msalter@...hat.com>, linux-kernel@...r.kernel.org,
linux-arch@...r.kernel.org, devicetree-discuss@...ts.ozlabs.org
Subject: Re: [PATCH v3 07/24] C6X: devicetree support
On Wed, Sep 28, 2011 at 03:31:36PM +0200, Arnd Bergmann wrote:
> On Tuesday 27 September 2011, Mark Salter wrote:
> > diff --git a/arch/c6x/boot/install-dtb.c b/arch/c6x/boot/install-dtb.c
> > new file mode 100644
> > index 0000000..fa979ce
> > --- /dev/null
> > +++ b/arch/c6x/boot/install-dtb.c
> > @@ -0,0 +1,333 @@
> > +/*
> > + * Program to hack in a DTB to an ELF file having a placeholder
> > + * section named __fst_blob.
> > + *
> > + * This allows for building multiple images with builtin DTBs
> > + * using a single vmlinux image. This is only necessary until
> > + * bootloader support exists.
> > + *
> > + * Copyright 2011 Texas Instruments Incorporated
> > + *
> > + * This program is free software; you can redistribute it and/or
> > + * modify it under the terms of the GNU General Public License
> > + * as published by the Free Software Foundation; either version
> > + * 2 of the License, or (at your option) any later version.
> > + *
> > + * Usage: install-dtb <image.dtb> <kernel.elf>
> > + */
>
> Other architectures allow simply appending the device tree to the
> kernel image file, which is more convenient for users and does
> not require much kernel support. Have you tried this?
>
> > diff --git a/arch/c6x/platforms/platform.c b/arch/c6x/platforms/platform.c
> > new file mode 100644
> > index 0000000..ad4651c
> > --- /dev/null
> > +++ b/arch/c6x/platforms/platform.c
> > @@ -0,0 +1,17 @@
> > +
> > +static int __init c6x_device_probe(void)
> > +{
> > + of_platform_bus_probe(NULL, of_default_bus_match_table, NULL);
> > + return 0;
> > +}
> > +device_initcall(c6x_device_probe);
>
> device_initcall is run rather late. I think it would be more useful
> to run this as early as possible, in order to make it possible to
> have more of the core services probed through the device tree.
>
> I also think that for probing the device tree root, you should be
> using of_platform_populate instead of of_platform_bus_probe
> nowadays. Maybe Grant can provide more information on this.
Yes. Use of_platform_populate(). It's better.
g.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists