24 Jun
2021
24 Jun
'21
3:54 p.m.
On 6/24/21 4:18 AM, Paul Gofman wrote:
On 6/24/21 08:16, Zebediah Figura wrote:
ok(!xs->Mask, "Got unexpected Mask %s.\n", wine_dbgstr_longlong(xs->Mask));/* The save area has garbage in this case, the state should be restored to INIT_STATE* without using these data. */
ok(xs->Mask == (xsaveopt_enabled ? 0 : 4), "Got unexpected Mask %#I64x.\n", xs->Mask);/* The save area has garbage if xsaveopt is available.** Clear the mask; the state should be restored to INIT_STATE without* using this data. */xs->Mask = 0; memcpy(context_ymm_data, test_extended_context_spoil_data1 + 4, sizeof(M128A)); }Since you probably have to resend anyway due to test failures, may I suggest to change a comment a bit, like:
/* The save area has garbage if xs->Mask is zero. * * Clear the mask: the state should be restored to INIT_STATE without * using context data. */Or somehow I had to read these a few times before understood how this makes sense. It was not immediately obvious that the second part ("clear the mask: ...") does not concern the garbage mentioned in the first part but quite the opposite: mask clear is need in the test when there was actually not a garbage.
I guess I was trying to separate the comment, because it's a little confusing as-is. I'll do a better job of that...