Attention is currently required from: Nico Huber, Thomas Heijligen, Edward O'Callaghan, Angel Pons.
4 comments:
Commit Message:
Patch Set #1, Line 14: adding another argument to register_spi_master.
The new version seems better! If the shutdown function can be […]
I am so happy to hear this!
Patchset:
> I wasn't sure which one of the structs is the right place for the shutdown function, and even more […]
thank you, this is actually very useful! (as usual!)
If master is a right place, than the whole change can be done in 3 chunks spi/par/opaque (like we did with the data). Then moving data and shutdown into registered_master can be next step? But that would be internal to the core, callers won't even notice (or shouldn't notice to be precise).
>>> Only in the same file for the "few" masters that you have been looking at
so far ;)
I had commit 4f5169df5ff2b63a40d654ab133b8a3108b6a599 in mind when I said "in the same file", looks like lots of masters, but maybe I misunderstand something?
File linux_spi.c:
if (register_spi_master(&spi_master_linux, spi_data))
return 1; /* core does cleanup */
return 0;
Or just […]
Done
File programmer.h:
Patch Set #1, Line 359: shutdown
I think we'll need shutdown functions for both eventually. Starting with […]
>>>for this to be ideally called `deinit` and logically one level up in `struct programmer_entry`
There is no data one level up, but shutdown function needs data. I started thinking about programmer_entry, but then returned back to master struct, because it has data.
Maybe deinit is a different thing to what is currently called "shutdown function"?
To view, visit change 55932. To unsubscribe, or for help writing mail filters, visit settings.