On Tue, Jul 08, 2014 at 11:12:20AM -0700, Richard Henderson wrote: > On 07/08/2014 09:13 AM, Al Viro wrote: > > Frankly, I suspect that it's better to have qemu-system-alpha behave like > > the actual hardware does (including "FPCR.DNOD can't be set") and keep the > > linux-user behaviour as is, for somebody brave and masochistic enough to > > fight that one. And no, it's nowhere near "just let denorms ride through > > the normal softfloat code and play a bit with the flags it might raise". > > And then there's netbsd/alpha and openbsd/alpha, so in theory somebody might > > want to play with their software completion semantics (not identical to > > Linux > > one) for the sake of yet-to-be-written bsd-user alpha support... > > You're probably right there. > > I've pushed a couple more patches to the branch, split out from your patch > here. I believe I've got it all, and havn't mucked things up in the process. > I'll run some tests later today when I've got time.
Just one thing - 0x1fffffffffffff will make 32bit hosts whine about integer constant being too large. So will 0x1ffffffffffffful, unfortunately - it really ought to be ull.