On Fri, Oct 13, 2017 at 12:09 PM, Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx> wrote: > On Fri, Oct 13, 2017 at 6:56 AM, Andrey Ryabinin > <aryabinin@xxxxxxxxxxxxx> wrote: >> >> This could be fixed by s/vmovdqa/vmovdqu change like bellow, but maybe the right fix >> would be to align the data properly? > > I suspect anything that has the SHA extensions should also do > unaligned loads efficiently. The whole "aligned only" model is broken. > It's just doing two loads from the state pointer, there's likely no > point in trying to align it. > > So your patch looks fine, but maybe somebody could add the required > alignment to the sha256 context allocation (which I don't know where > it is). IIRC if we try the latter, then we'll risk hitting the #*!&@% gcc bug that mostly prevents 16-byte alignment from working on GCC before 4.8 or so. That way lies debugging disasters.
