[PATCH v2 5/9] bootstd: Correct creating of bootdev sibling

[Simon Glass]

Hi Bin,

On Sun, 30 Jul 2023 at 21:37, Bin Meng <bmeng.cn at gmail.com> wrote:
>
> Hi Simon,
>
> On Mon, Jul 31, 2023 at 10:50 AM Simon Glass <sjg at chromium.org> wrote:
> >
> > Hi Bin,
> >
> > On Sun, 30 Jul 2023 at 19:59, Bin Meng <bmeng.cn at gmail.com> wrote:
> > >
> > > Hi Simon,
> > >
> > > On Mon, Jul 31, 2023 at 1:15 AM Simon Glass <sjg at chromium.org> wrote:
> > > >
> >
> > >
> > > >         }
> > > >
> > > >         return 0;
> > > > --
> > >
> > > Regards,
> > > Bin
> > > > Use the correct function here, since there may be multiple IDE devices
> > > > available.
> > > >
> > > > Signed-off-by: Simon Glass <sjg at chromium.org>
> > > > Reviewed-by: Mattijs Korpershoek <mkorpershoek at baylibre.com>
> > > > ---
> > > >
> > > > (no changes since v1)
> > > >
> > > >  drivers/block/ide.c | 4 ++--
> > > >  1 file changed, 2 insertions(+), 2 deletions(-)
> > > >
> > > > diff --git a/drivers/block/ide.c b/drivers/block/ide.c
> > > > index 89201dd4d229..c698f9cbd558 100644
> > > > --- a/drivers/block/ide.c
> > > > +++ b/drivers/block/ide.c
> > > > @@ -1059,9 +1059,9 @@ static int ide_probe(struct udevice *udev)
> > > >                 desc->lba48 = pdesc.lba48;
> > > >                 desc->type = pdesc.type;
> > > >
> > > > -               ret = bootdev_setup_for_dev(udev, "ide_bootdev");
> > > > +               ret = bootdev_setup_for_sibling_blk(blk, "ide_bootdev");
> > > >                 if (ret)
> > > > -                       return log_msg_ret("bootdev", ret);
> > > > +                       return log_msg_ret("bd", ret);
> > >
> > > Why changes from bootdev to bd here? To me, bd is unclear, and sounds
> > > like "boardinfo" in U-Boot.
> >
> > I don't need to...the goal is to have the strings be 4 bytes or less,
> > so they don't use up too much space in the image...but also have them
> > unique enough that you can look at the function and see which call
> > site failed.
>
> I see the needs of saving size in TPL, or maybe SPL, but not everywhere, right?

The thing is, we only have one knob to control this - CONFIG_LOG_ERROR_RETURN

When you enable it, all of the log_msg_ret() calls in U-Boot produce a
string. So there is quite a big impact.

For example, on coral, the text size increases by about 35KB. So it is
about a 6% increase. If it gets to much then it might not fit, or it
might cause other prolbems. For example, we have fixed-size regions in
binman for U-Boot.

Regards,
Simon


>
> >
> > For this one, I don't mind either way.

Regards,
Simon