<html> <head> <base href="https://bugs.freedesktop.org/" /> </head> <body> <div> <a class="bz_bug_link bz_status_ASSIGNED " title="ASSIGNED - Illegal instruction _mesa_x86_64_transform_points4_general" href="https://bugs.freedesktop.org/show_bug.cgi?id=27512#c19">Comment # 19</a> on <a class="bz_bug_link bz_status_ASSIGNED " title="ASSIGNED - Illegal instruction _mesa_x86_64_transform_points4_general" href="https://bugs.freedesktop.org/show_bug.cgi?id=27512">bug 27512</a> from <a class="email" href="mailto:t_arceri@yahoo.com.au" title="Timothy Arceri <t_arceri@yahoo.com.au>"> Timothy Arceri</a> <pre>(In reply to Roland Scheidegger from <a href="show_bug.cgi?id=27512#c18">comment #18</a>) > > That doesn't make sense to me. The offset is just part of the memory > operand. Unless the assembler encodes it wrong I can't see why that wouldn't > work (which I would think to be unlikely, but the locality hints are also > encoded into the mod r/m byte - what's the encoding of the instruction?) > I suppose a solution would just be to ditch prefetch - as was pointed out > it's not really far ahead enough in any case, even k8 and p4 had primitive > hw prefetchers which should make such a simple software prefetch completely > unnecessary. I shouldn't play with asm passed my bedtime ... I was having trouble with the patch applying so recreated it myself. Seems the problem was I missed one instruction. All works well once fixed so I've sent that patch to the mailing list.</pre> </div> <hr> You are receiving this mail because: <ul> <li>You are the assignee for the bug.</li> </ul> </body> </html>