Adds handling for arduino_default for specifying SPI device, and not CS #560
Conversation
|
And obviously it's failing here, because the pre-requisite change isn't in Portduino. |
|
if your PR targets a specific version of meshtastic native platform you should as well update the platformio_native.ini file to point to the relevant commit hash so that the CI job doesn't fail |
| @@ -67,14 +67,15 @@ namespace lgfx | |||
| }; | |||
| }; | |||
| int16_t pin_cs = -1; | |||
| bool is_SPI = false; | |||
There was a problem hiding this comment.
I don't think this change at the interface is needed. The interface already has a way to specify if it's spi or i2c by querying the value of cs.
Rather than changing this interface the application that uses this interface has to be changed.
There was a problem hiding this comment.
What about setting all four values (cs, mosi, miso, sclk) to 0 to indicate that this interface is managed by the Linux kernel?
There was a problem hiding this comment.
That might work. 0 is a valid pin number, but not all of them set to it. There is the problem of the union, which is in itself a really weird thing. So if we set the SPI elements to 0, it overwrites the I2C elements.
|
Closing this in favor of #561 |
This is the so-simple-its-dumb approach to the SPI touchscreen issue. The idea is that we'd add SPI1 and SPI2 to portduino as extern defines, so they're always available. We can tie them to a real hardware device in our init code, and here we just use the spi_host int to specify which one to use. This is an alternative solution to #559
Edit: the addition of is_SPI is because now that we don't necessarily need a CS, there's no other way to determine for sure that it's an SPI device.