David Gibson wrote:

On Tue, Feb 19, 2008 at 08:18:19PM -0500, Jerry Van Baren wrote:

quoted

Jon Loeliger wrote:

quoted

So, like, the other day David Gibson mumbled:

quoted

In light of the recently discovered bug with NOP handling, this adds
some more testcases for NOP handling.  Specifically, it adds a helper
program which will add a NOP tag after every existing tag in a dtb,
and runs the standard battery of tests over trees mangled in this way.

For now, this does not add a NOP at the very beginning of the
structure block.  This causes problems for libfdt at present, because
we assume in many places that the root node's BEGIN_NODE tag is at
offset 0.  I'm still contemplating what to do about this (with one
option being simply to declare such dtbs invalid).

Signed-off-by: David Gibson <redacted>

Applied.

BTW, declaring DTBs with BEGIN_NODES not at offset 0
as invalid seems like a fine choice to me.

jdl

FWIIW, I vote ditto on declaring DTBs with BEGIN_NODES not at offset 0 
as invalid.  The root being at offset 0 assumption is pretty well 
entrenched and I cannot think of any reason to change it that would be 
worth the effort.

Well, it's actually not that hard to deal with.  I've already been
planning to add a helper function/macro which validates a node offset
(something currently open-coded in a whole bunch of places).  It would
be fairly easy to make it skip over nops as well.

But, likewise I can think of no reason that NOPs before the root node
would be useful or likely to occur in practice.

Hi David,

Originally, finding the root node by searching the path "/" would return 
an error so I specifically caught that case and used 0 for the offset. 
I looked over the current u-boot and libfdt code and it looks like that 
works now (u-boot no longer traps "/") so the offset 0 == root node 
assumption is no longer built into u-boot.  That is good.

OTOH, now I got some comments in u-boot I need to fix. :-/  Oh well, it 
is a net win. :-)

Best regards,
gvb