SPI peripheral/CS naming issues

catfact commented 8 years ago

encountered this discrepancy when refactoring aleph for building against libavr32

in aleph board conf header (https://github.com/tehn/aleph/blob/dev/avr32_lib/conf/aleph_board.h#L82) we stick to the ASF conventions when defining the address of an SPI peripheral:

#define ADC_SPI (&AVR32_SPI1)

and defining a chip select index:

#define ADC_SPI_NPCS 0

in skeleton/system/libavr32, there is only one SPIperipheral definition, and ADC_SPIetc are defined as the chip select numbers and not as peripheral addresses. like so: https://github.com/monome/libavr32/blob/master/conf/teletype/conf_board.h#L138

i can't see the reason for this change, which breaks all the drivers and muddies a perfectly clear naming convention.

i would like to:

[x] change these names back. this is simple, but requires a simultaneous change to libavr32 and, alas, all the module repos (because there is no DAC driver, just pasted direct calls to SPI functions.) there might be other similar changes things so i'll come back to it at a later stage in aleph refactor, leaving this issue as a reminder (and in case someone wants to tell me why i shouldn't do this.)

and while i'm at it:

[ ] factor direct calls to DAC SPI out of module main code and into the lib, so changing details of board configuration doesn't impact modules in the future.

samdoshi commented 8 years ago

Both of these sound good to me.

Add a CHANGELOG.md file when you make the changes so that they can be tracked easily.

catfact commented 7 years ago

we kind of stalled out on the aleph refactor (it's basically done on a branch, but hasn't been properly tested)

but we did change the pin defines with PR #18 and related. i didn't bother refactoring TT code. closing this anyway since it's been over a year

monome / libavr32