From mboxrd@z Thu Jan  1 00:00:00 1970
From: jgunthorpe@obsidianresearch.com (Jason Gunthorpe)
Date: Thu, 4 Oct 2012 11:59:07 -0600
Subject: [PATCH] [ARM] Use AT() in the linker script to create correct
 program headers
In-Reply-To: <20121004113637.GA2117@linaro.org>
References: <20120930232116.GC30637@obsidianresearch.com>
 <20121001153952.GB2100@linaro.org>
 <20121001160639.GA31620@obsidianresearch.com>
 <20121001175647.GD2100@linaro.org>
 <20121001183543.GC22342@obsidianresearch.com>
 <20121002102346.GB2108@linaro.org>
 <20121002174759.GC23733@obsidianresearch.com>
 <20121003104335.GA2254@linaro.org>
 <20121003184437.GB12231@obsidianresearch.com>
 <20121004113637.GA2117@linaro.org>
Message-ID: <20121004175907.GB2994@obsidianresearch.com>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

On Thu, Oct 04, 2012 at 12:36:37PM +0100, Dave Martin wrote:

> OK, so it is supported, but not for ARM, yet.  I'm not sure that
> such a patch would be rejected, since building in a DTB is not
> really that different from building in a command-line.

Maybe I will be able to look at this in a few months..

> The other solution to this problem is to distribute the dtb
> alongside the kernel as a supplementary image (similar to
> initramfs), with the bootloader doing the bare minimum of
> modifications to it.

Yes, but we still need rely on complex code like I2C/MTD to create a
correct DTB, which again puts us back to patching the kernel for that
functionality.
 
> Naturally, if zero modifications are needed then the bootloader does not
> need libfdt (or equivalent code) at all.

If only :)

> One other option open to you would be to provide an ELF image loadable
> by your bootloader via a simple wrapper.  This would probably be a more
> robust approach than carrying patches against head.S, since it removes
> any intimate dependency on the kernel code itself.

This is what I ment by having a 2nd stage loader, but this example is
not complete because at this point you also have to patch into the DTB
the initrd address (passed in r0/r1), and the various MAC addresses
(getting them is the hard part..).

The nice thing about the head.S patch is that it lives right after
_entry, so there actually are no dependencies on the kernel code, it
executes in the same environment as the example you presented. 

Anyhow, exploring a CONFIG_ARM_BUILT_IN_DTB.. I think I'd add
something like this to head.S:

#ifdef CONFIG_ARM_BUILT_IN_DTB
       adr     r3,bootloader_r0
       stmia   r3,{r0,r1,r2}
       mov     r2, #0
       mov     r1, #0
       mov     r0, #0
#else
       bl      __vet_atags
#endif

And something like this to early_dt_fixup:

#ifdef CONFIG_ARM_BUILT_IN_DTB
        for_each_dt_provider(dtp) {
	      devtree = dtp->get_dtb();
	      if (devtree)
	      	   break;
        }
#else
        devtree = phys_to_virt(dt_phys);
#endif

Where the 'dev tree provider' would use the stored bootloader
registers and any other information to return the proper DTB. 

Jason