Addy, On Thu, Nov 6, 2014 at 12:11 AM, Addy Ke <addy.ke at rock-chips.com> wrote: > high_ns calculated from the low division of CLKDIV register is the sum of > actual measured high_ns and rise_ns. The rise time which related to > external pull-up resistor can be up to the maximum rise time in I2C spec. > > In my test, if external pull-up resistor is 4.7K, rise_ns is about 700ns. > So the actual measured high_ns is about 3900ns, which is less than 4000ns > (the minimum high_ns in I2C spec). It's a little unfortunate to have to make the assumption that the rise time for a board is going to be the maximum the spec allows. I think this is something that's better to let a board specify in the device tree. Allowing us to specify it also gives us a little extra slop and makes it easier to specify a "fall time" without affecting the resulting frequency. Right now your code assumes that the fall time is 0. I've prototyped what things could look like if the rise and fall times were specified in the device tree. You can see my patch (atop yours) at: https://chromium-review.googlesource.com/#/c/232774/ ...perhaps you could squash that into your patch and post up v2? -Doug
