On 12 December 2013 12:14, C Fontana <claudio.fontana@xxxxxxxxxx> wrote: > I think that there is more than the missing return: > > we need to handle the case 0 as well, as it's a perfectly valid form > of a load/store pair: > it's the Load/Store no-allocate pair (offset) (LDNP, STNP). > > So in my view we need to add a case 0 where we handle the load/store > no-allocate pair, > and no default case, or a default case where we assert(0) for unreachable code, > since all possible index values (2 bits) should be handled by the > switch statement. Yep. I've added support for the non-temporal versions (pretty easy since qemu doesn't care about cache allocation hints) to my patchstack so we'll get this in the next version of the patchset. thanks -- PMM _______________________________________________ kvmarm mailing list kvmarm@xxxxxxxxxxxxxxxxxxxxx https://lists.cs.columbia.edu/cucslists/listinfo/kvmarm
