On Fri, 17 Apr 2020 15:15:43 +0800 Leo Yan leo.yan@linaro.org wrote:
Good analysis, Masami :)
Agreed. Nice work Masami!
Seems to me, Arm64 should study Arm platform to change using trace_ipi_raise_rcuidle() to work around the restriction so can allow the event to work in idle context?
Also agree. Looks to need the rcuidle() variant.
-- Steve