On Sat, Feb 06, 2016 at 01:48:19PM +0000, Peter Maydell wrote: > On 6 February 2016 at 12:17, Edgar E. Iglesias <edgar.igles...@gmail.com> > wrote: > > It seems to me like if EL3 is running in AArch32, then we shouldn't > > trap accesses from Secure EL1 but I can't find that logic. Am I missing > > something? > > If EL3 is running in AArch32 then there is no Secure EL1 -- all > of SVC, IRQ, and the other PL1 modes run at EL3 privilege, > and arm_current_el() will return 3 in this situation.
Yep, I keep forgetting that for AArch32... Cheers, Edgar