Am 19.04.2017 um 18:31 schrieb Álvaro Fernández Rojas:
Hi Andreas,
El 19/04/2017 a las 18:25, Andreas Färber escribió:
Am 19.04.2017 um 18:22 schrieb Álvaro Fernández Rojas:
In case you decide to apply this version (I see you're yet deciding which is the best way to go), I tested it and it solves my problem with sysreset not bein probed correctly.
Can you point us to which call site that is in your case exactly?
Sure, see: https://github.com/Noltari/u-boot/commits/bmips-next-v3-fixes And: https://lists.denx.de/pipermail/u-boot/2017-April/288041.html
https://lists.denx.de/pipermail/u-boot/2017-April/288091.html
Hmm, to me that looks like the exact opposite symptom of what we're discussing right now. You seem to be getting into the for loop and adding checks there, whereas our issue is that we're not getting into the for loop due to NULL dev.
IIUC your code would be another candidate for _available_device() under discussion, as your patch is skipping devices that don't probe.
Regards, Andreas