On Thu, Jul 19, 2018 at 3:52 PM, Tom Rini trini@konsulko.com wrote:
On Wed, Jul 18, 2018 at 04:09:33PM +0200, Wolfgang Denk wrote:
Dear Tom,
In message 20180718125351.GE4609@bill-the-cat you wrote:
Loading Environment from FAT... --> MMC: no card present --> ** Bad device mmc 0 ** --> Failed (-5) Loading Environment from MMC... --> OKinstead of:
Loading Environment from FAT... MMC: no card present ** Bad device mmc 0 ** Failed (-5) Loading Environment from MMC... OKSo, I think maybe (and given Wolfgang's comments) we should think about how the output might want to look, and how to get there without GD changes. Perhaps: Attempting to load Environment from FAT (do we have more easily available info at this point?): MMC: no card present ** Bad device mmc 0 ** Failed (-5) Loading Environment from MMC... Attempting to load Environment from MMC: Succeeded
Just my 0.02€:
In the non-error case, the output should be a single (ideally short) line.
Rationale: to many lines of ourput clutter your screen and make you miss context faster; to many/long lines take time to print so they make booting slower.
In the error case, the user should be able to understand what the problem was and decide if it was critical or can be ignored (like here when intentionally booting without SDCard).
I understand, but I don't know if we can get there still. The problem is we don't know if we've succeeded until we've done the relevant probing and that in turn is what's breaking the single line, and got us to where we are now.
Actually we can, please see my new RFC patch [1]. It's a bit hacky, but the only other way to do so is to rework drivers (MMC, etc).
Also, I figured how to do prefixing (to display MMC errors as nested w.r.t. "Loading environment), without adding new field to gd. We can just add some new GD_LG_ and print prefix when it's installed. I'm gonna send new RFC soon. Please let me know what you think about [1].
[1] https://lists.denx.de/pipermail/u-boot/2018-July/335223.html
-- Tom