09:48 <cyndis> digetx: I don't understand your question about how many syncpoints a client(?) will have. Shared syncpoints don't need fixed IDs, they can be communicated between VMs through IVC or networking or other mechanisms

10:11 <digetx> cyndis: all sync points are readable and only some are writable?

10:11 <cyndis> correct

16:57 <DavidHeidelberg> Hello did anyone tried ever play with KernelCI, TuxSuite or TuxMake?

17:29 <digetx> never tried, how many kernels are you going to build and test? if it's one-two, then perhaps a real CI could be an overkill

17:30 <DavidHeidelberg> well, many users porting their devices to mainline like kernels, having some webpage with recent builds, some multiple compilers assembled kernels and these stuff could help a bit

17:31 <DavidHeidelberg> also thinking that having option just git push, wait a bit and then take a binary or boot.img and flash it would be nice for people with lowpower laptops

17:32 <DavidHeidelberg> + some stats, warnings check collector, maybe tests

17:32 <DavidHeidelberg> (plus)

17:33 <digetx> it may be not easy to set it up and maintain

17:34 <DavidHeidelberg> well, I'm thinking about reusing most of already done stuff (looking mostly to tuxsuite and tuxmake)

17:42 <digetx> I only briefly touched kernelci failure reports, can't suggest which is better

22:45 <Ultrasauce> can anyone slip me a copy of the max77620 datasheet? a register bitfield I need documented is not in the driver and maxim basically told me to get bent