[U-Boot] [PATCH 0/7] Bootgraph.pl instrumentation support for UBoot

[Simon Glass]

Hi Andrew,

On Sat, Sep 10, 2011 at 5:40 AM, Andrew Murray <amurray at theiet.org> wrote:
> On 1 September 2011 00:53, Andrew Murray <amurray at theiet.org> wrote:
>>
>> >
>> > I will assume that we have a microsecond timer, update my patch and
>> > resubmit so you can take a look and see what you think. Hopefully we
>> > can unify this, your patch and the boot_progress stuff.
>>
>> Excellent! OK, well I will await the patch, read up on the original
>> intentions and we can go from there. I'll look forward to making UBoot
>> more boot time friendly. Good night.
>
> I've updated my patches based on your feedback. I guess the next step is to
> see how
> to integrate with the bootstage work.
> Andrew Murray

Thanks for the patch set. I have a few general comments.

- textbase should come from System.map also - it is hard coded at present
- Perl is ick, but it was good because it reminded me of why I am
happier in Python :-)
- Patch 4 didn't apply cleanly on master
- Should add to hush parser also
- Perl script could use some error checking. I had a bit of fun
debugging it as it silent failed
- The resulting image was very long and I had to zoom in a lot to see
any text. Perhaps increase font size a lot so you can see the timeline
and function names?
- You should use do_div for the 64-bit division I think
- I think the mention of CONFIG_BOOT_TIME should be CONFIG_BOOT_TRACE

This patch touches on Graeme's initcall patch. If board_init_r were
just a list of function pointers then your patch would be easier! Not
much we can do about that at this stage, and I think your solution is
reasonable.

Regards,
Simon