Heiko,
Am 22.04.2016 um 12:20 schrieb Heiko Schocher:
Think of places where work is scheduled but the caller blocked the worker because the work has to be done later. Fastmap is one of these use cases but AFAIK it won't matter as no CPU scheduler is involved and will interrupt Fastmap.
Can you explain this a little bit?
As I said, when you are in a code region where parallel work must not happen as it will, for example, confuse your state you block the worker. But you are still allowed to schedule new work which will executed after you unblock it. For the fastmap example I gave it should be fine but I didn't check all code paths in UBI for u-boot single thread safety. :-)
What I wanted to say is that executing work directly at schedule time does not match 1:1 the POV of Linux UBI and is error prone. These are issues you won't notice by compile testing.
An alternative approach would be not executing work directly while scheduling it but in produce_free_peb(). UBI is designed to work with the worker being disabled. All UBI work will then happen synchronous and should also work in u-boot.
In the long run I suggest removing the whole Linux UBI implementation from u-boot and add a small (read only!) implementation which can also read UBIFS. Reading UBIFS is not a big deal. Also journal reply can be done in-memory.
Hmm.. I think read only is not for all boards an option, as we also create UBI Volumes and/or write to them in U-Boot ...
Depends. IMHO a bootloader has exactly one job, loading a kernel and booting it. And not being a poor man's general purpose operating system where you can also do management stuff like managing UBI volumes. ;-)
Thanks, //richard