Hi, On Fri, Jun 16, 2023 at 12:13:45PM +0200, Miquel Raynal wrote: > sre@xxxxxxxxxx wrote on Thu, 15 Jun 2023 01:27:24 +0200: > > On Sat, Jun 10, 2023 at 10:45:25PM +0200, Sam Ravnborg wrote: > > > On Fri, Jun 09, 2023 at 04:59:51PM +0200, Miquel Raynal wrote: > > > > A very basic debugging rule when a device is connected for the first > > > > time is to access a read-only register which contains known data in > > > > order to ensure the communication protocol is properly working. This > > > > driver lacked any read helper which is often a critical peace for > > > > fastening bring-ups. > > > > > > > > Add a read helper and use it to verify the communication with the panel > > > > is working as soon as possible in order to fail early if this is not the > > > > case. > > > > > > The read helper seems like a log of general boiler plate code. > > > I briefly looked into the use of regmap for the spi communication, > > > but it did not look like it supported the bit9 stuff. > > > > > > I did not stare enough to add a reviewd by, but the approach is fine > > > and it is good to detech issues early. > > > > The st7789v datasheet describes a setup where SPI is connected > > unidirectional (i.e. without MISO). In that case the ID check > > will fail. > > Right. I'll add a (spi->mode & SPI_NO_RX) check, as the default is to > have both lines, if there is no MISO line, I'd expect it to be > described in the DT, otherwise the description is broken. Checking for SPI_NO_RX sounds good to me. -- Sebastian
Attachment:
signature.asc
Description: PGP signature