si5351: add support for si5351 #810
Conversation
si5351/si5351.go
Outdated
| "math" | ||
|
|
||
| "tinygo.org/x/drivers" | ||
| "tinygo.org/x/drivers/internal/legacy" |
There was a problem hiding this comment.
This is known to cause heap allocations, shall we try and avoid this from start?
There was a problem hiding this comment.
Good point @ysoldak I just added another commit to use the regmap package instead.
Adds support for the si5351 I2C programmable clock generator using code from @chiefMarlin which used code from @conotto which somehow never got merged. Thank you everyone! Signed-off-by: deadprogram <[email protected]>
deadprogram
force-pushed
the
si5351-redux
branch
from
99907b5 to
2c55886
Compare
soypat
left a comment
soypat
left a comment
There was a problem hiding this comment.
Very nice! Love to see regmap in use! FYI got a couple suggestions on how to avoid most if not all allocations
Signed-off-by: deadprogram <[email protected]>
deadprogram
force-pushed
the
si5351-redux
branch
from
2c55886 to
a48aaf0
Compare
|
Appears that there are no further heap allocations now. Should be ready to go. |
|
|
||
| // Power down all output drivers | ||
| buf := []byte{CLK0_CONTROL, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80} | ||
| d.bus.Tx(uint16(d.Address), buf, nil) |
There was a problem hiding this comment.
First real use of regmap in the wild!
I was trying to wrap my head around it and how could I use it in cases like #800 (as suggested).
I basically hit this same problem -- can't just use regmap functions everywhere, consistently, have to fall back to raw d.bus.Tx here and there => have to keep both rw and bus. And that sort of defies the purpose, to me at least...
We may be better off implementing helper methods for reading and writing variable size slices, like this, but in internal package somewhere to be reusable?
https://github.com/tinygo-org/drivers/blob/dev/lsm9ds1/lsm9ds1.go#L216
There was a problem hiding this comment.
I think that @soypat is working on something like that. In the meantime, we can add in the future if we find it is needed.
|
Any other feedback or can this get merged now? |
|
Good stuff! Curious how the compiler is managing to not escape an array that is being passed into an interface. Very curious indeed |
Adds support for the si5351 I2C programmable clock generator using code from @chiefMarlin which used code from @conotto which somehow never got merged.
Thank you everyone!