[U-Boot] [PATCH v2 1/4] ns16550: Add function to drain serial output

Tom Warren TWarren at nvidia.com
Fri Mar 9 22:23:31 CET 2012


Simon/Stephen,

> -----Original Message-----
> From: sjg at google.com [mailto:sjg at google.com] On Behalf Of Simon Glass
> Sent: Friday, March 09, 2012 2:09 PM
> To: Stephen Warren
> Cc: U-Boot Mailing List; Tom Warren
> Subject: Re: [PATCH v2 1/4] ns16550: Add function to drain serial output
> 
> Hi,
> 
> On Fri, Mar 9, 2012 at 1:00 PM, Stephen Warren <swarren at wwwdotorg.org>
> wrote:
> > On 03/09/2012 01:32 PM, Simon Glass wrote:
> >> Sometimes we want to be sure that the output FIFO has fully drained (e.g.
> >> because we are about to reset or the port or reset the board). Add a
> >> function for this.
> >>
> >> Signed-off-by: Simon Glass <sjg at chromium.org>
> >
> > The series,
> >
> > Tested-by: Stephen Warren <swarren at wwwdotorg.org>
> >
> > Tom,
> >
> > I'm OK with you adding these patches to u-boot-tegra/master and
> > sending a pull request for that. Just please put these patches before
> > any that actually enable CONFIG_OF_CONTROL for any boards, so there's
> > never a window of broken commits.
> >
> > I think it'd be nice at this stage to also drop the commit that turns
> > Ventana into a DT board, since re-using the Seaboard .dts file for
> > Ventana isn't correct. However, IIRC that commit was added to WAR some
> > build issue with Simon's patch series? If that's still a problem, then
> > I'm OK with converting Ventana if we have to, but we'll need to be
> > very careful that tegra-seaboard.dts doesn't add stuff that'll fail
> > on, or damage, Ventana.
> 
> A few points:
> 
> 1. Yes you can drop the Ventana commit and I agree it is a good idea.
> We should have a real Ventana .dts file I think. Ventana should built ok
> without a device tree now. The problem was in my clock.c patch.
> 
> 2. I have held off responding to Stephen's patch on the ML to see what other
> say. My view is that it is controversial since it changes the so-far
> accepted meaning of u-boot.bin and the behaviour of the U-Boot Makefile.
> Plus it is not really necessary as a means of informing the user since we
> put the pre-console putc() for exactly this problem. So I would rather leave
> Stephen's patch out at until people have time to decide that I am wrong
> about it. We already have CONFIG_OF_EMBED to build the fdt into u-boot.bin.
> Grant Likely had big reservations about this feature - let's not bring it in
> by stealth.
> 

I see now why it's so hard to find custodians for the U-Boot repos. :/

I've done as Stephen requested and re-ordered the commits so his Makefile fix comes before the config file changes to add CONFIG_OF_CONTROL to Seaboard and Ventana.

I've left the Ventana change that uses the Seaboard .dts file in, as I think it can easily be edited to use a Ventana-specific dts file just as soon as one of you finds one and submits a patch for it. For now, the Seaboard .dts just works.

I've pushed my latest changes to u-boot-tegra/master, and I'm going to let it percolate over the weekend. I'd have liked to get a pull request in to Albert today, before the ARM HEAD moves again, but I'm not going sweat to it since there doesn't seem to be a consensus on how to proceed here. IMO, I see no problem with putting these patches in as they stand, since I've tested them to work OK on Seaboard and Ventana, other (non-DT) Tegra2 boards build fine, and anything here that needs to change can always be edited with a future patch. But I'm going to head out early (it's my birthday today), and pick this up again next week.

Tom

-- 
nvpublic
> Regards,
> Simon


More information about the U-Boot mailing list