On Mon, 15 Jun 2026 at 11:56, Alex Bennée <[email protected]> wrote: > > Peter Maydell <[email protected]> writes: > > > On Fri, 29 May 2026 at 09:29, Alex Bennée <[email protected]> wrote: > >> > >> Now WFE can actually suspend on A-profile we also need to handle when > >> its trapped. To do this we need to pass the instruction size so we can > >> deal with the is_16bit syndrome encoding. > >> > >> Reviewed-by: Richard Henderson <[email protected]> > >> Signed-off-by: Alex Bennée <[email protected]> > > > > Doesn't this need to go earlier in the patchset? Otherwise there's > > a point where WFE can go to sleep but doesn't honour the trap bit. > > Hmm - the way I read the description was we would only trap if we were > going to sleep. Pre 6/8 we never sleep so we can't trap.
Yes, that's the point. The guarantee we have to uphold with the trap bits is "if the WFE would actually go to sleep, then we need to take the trap". - before this patchset: we never sleep, so never need to trap - after patch 6: we can sleep, but we incorrectly won't trap - after this patch: we can sleep and will trap So there's a window in this patchseries (after 6 but before 7) where we don't do the right thing. -- PMM
