Wolfgang Denk wrote: [ snip ]
Ok, well I've successfully compiled with CONFIG_HARD_I2C and have communicated with my clock, so that is good. Next up is transferring the date in u-boot to linux, but I'm sure that will be documented somewhere.
You don't "transfer" the daye. Linux will pick it up itself (which requires an RTC driver and eventually some modifications).
Let me check that I understand: As far as I can see the u-boot realtime clock drivers just allow you to set and read the time from the u-boot command line; a useful utility. If linux is to read the time from the realtime clock it must have a realtime clock driver of its own; there's no way that u-boot can pass the time to linux at boot time? (Of course, this would be fraught with difficulties, because depending on how fast your cpu boots linux, a different amount of time would elapse before linux could actually do something with the passed date stamp).
So, next up for me is to write a linux driver for the ds1307. I was actually in the middle of that before getting sidetracked by the u-boot rtc code. However, the u-boot code has enabled me to hardware debug my realtime clock circuit, so that was very useful. Has anyone written a linux driver for the ds1307 or similar? (I'm cross posting to the linuxppc-embedded list to ask this question).
As for CONFIG_SOFT_I2C; Attached (really this time) is a diff between the original u-boot-0.4.0/include/configs/TQM823L.h and the one that I have altered. The compile of soft_i2c.c fails with the following error messages:
That's because you did not define the required macros (see for example the configuration in lwmon.h).
Right, so it's not the case that you just #define CONFIG_SOFT_I2C; you have to correctly define the macros. I shall do some work on the readme and post a diff later on. Thanks for the pointers.
best,
Seb