On Fri, 13 Apr 2018 09:19:41 -0700 Martin Kelly <mkelly@xxxxxxxx> wrote: > On 04/13/2018 02:25 AM, Jean-Baptiste Maneyrol wrote: > > Hello, > > > > I am now able to reproduce the issue by generating kernel irq (just > > plug/unplug mouse or keyboard is sufficient). > > > > My last modification doesn't solve anything, since event the hard irq > > handler is disabled when processing another interrupt. Having a latched > > interrupt and triggering by level is a workaround, but clearly not perfect. > > > > > I'm glad you can reproduce the issue now! What board are you using? My > issues have been with the nanopi neo air (based on the Allwinner H3 SoC). > > > I think we need to work out a new solution for timestamping correctly > > the data. > > > > JB > > > > > I think we should try to solve the missing interrupts issue. Without > solving it, the best we can do is interpolate, as I do in the other patch. > > That said, I think supporting more interrupt types and interpolating are > good ideas regardless. Supporting more interrupt types is useful for > integrating with more types of hardware, and interpolating is useful for > being more robust against systems having issues, which can unexpectedly > happen even if we fix the immediate issue we see here. Agreed on the interrupt types. Interpolating is till rather 'nasty' so the case is a little less clear, but I'm being convinced I think... Anyhow, looking forward to v6 Will want JB reviewed-by or acked-by on this one! Jonathan -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
