Daniel Egger
Daniel Egger
@robbie01 👋 As you've noticed there's currently no code to implement the slave part. My SPI slave knowledge is unfortunately rather limited so I'm not sure that we already have...
Not to spoil your fun but a bitbang implementation doesn't really have any relevance here. Any reason why you're not using the internal pull-ups?
I've just pulled up a F072 Nucleo and some I2C hardware... Let's see whether I can find something. Which version did you test? Master?
I can confirm the behaviour. On an empty write a NACK to address 0 is sometimes not detected/reported by the hardware. Checking with the RM I don't see any way...
> Can you check whether you can reproduce with other addresses, too? 0 is general call address so might be special. I haven't tried. I have tried a better version...
> i2cdetect man page does not recommend scanning either with just write-byte or with just quick-write: https://linux.die.net/man/8/i2cdetect It does suggest not to use SMBus methods to scan the bus. I...
Why would that be an SMBus method? 🤔 And what is the point of using I2C if you're not going to transmit anything at all? Also I have not see...
What isn't a hack? 😅 Maybe we can improve the documentation to set the expectations right?
Adding an example seems fine to me, though this is unnecessarily complex and convoluted: - The macros (just to deal with the global variables) seem a bit over the top...