On 2018-10-01 21:33, Ajay Gupta wrote: > Hi Heikki and Peter, > >>>>> Can we get the working set in while the issues is being debugged? >>>> >>>> I'm not the one making the decision, and I don't even know if this >>>> is going through the i2c or the usb tree? If it's going through the >>>> i2c tree you need a tag from the usb people (Greg?) on patch 2/2, >>>> and if it's going through the usb tree, you need a tag from Wolfram >>>> on patch 1/2. As I said, I'm not involved with that part, I'm just >>>> reviewing these patches because I felt like it. >>>> >>>> The remaining issue that bothers me is the looping reads, and your >>>> email address reveals that you should be in a very good position to >>>> work out why they do not work, and fix it or let us know why they >>>> don't. >> I am working with different teams to debug this and I think it may take some >> time to know the root cause. > We got confirmation from HW team about 4 byte read limitation. There has to > be a STOP after every single read cycle. One read cycle supports maximum of > 4 byte burst. I will update the patches with a comment on this. Thanks for digging into this! And if the HW team says it's not possible, then of course my objection falls flat. However, you only mention "read cycle", and based on your defines (I2C_MST_CNTL_CYCLE_TRIGGER) that seems to be terminology from the spec. Yet, you apparently do writes without triggering a cycle. Do the HW team have anything to say about doing reads without triggering a "cycle"? Cheers, Peter
