On Fri, Apr 06, 2018 at 06:30:50PM +0100, James Morse wrote: > Hi Mark, > > On 06/04/18 18:22, Mark Rutland wrote: > > Digging a bit, I also thing that our ct_user_exit and ct_user_enter > > usage is on dodgy ground today. > > [...] > > > I think similar applies to SDEI; we don't negotiate with RCU prior to > > invoking handlers, which might need RCU. > > The arch code's __sdei_handler() calls nmi_enter() -> rcu_nmi_enter(), which > does a rcu_dynticks_eqs_exit().
Ah, sorry. I missed that! Thanks, Mark.