Age | Commit message (Collapse) | Author |
|
|
|
Should work, but doesn't. At least it doesn't hardlock.
|
|
The MAX-based function can produce values that are non-monotonic for a span
which causes glitches in texture filtering. The sqrt-based one avoids that.
This is perhaps slightly slower than before, but the difference
probably isn't noticable given we're doing software mipmap filtering.
Issue reported by Nir Radian <nirr@horizonsemi.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
interfaces.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Also a fair amount of cleanup.
|
|
This keeps texture emit from invalidating CS.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
COS, SIN, and CSC are not simple.
|
|
|
|
We're cookin' now.
|
|
|
|
|
|
|
|
Both the windows and fallback versions are void already.
|
|
|
|
|
|
Signed-off-by: Thomas Hellstrom <thellstrom-at-vmware-dot-com>
|
|
Signed-off-by: Thomas Hellstrom <thellstrom-at-vmware-dot-com>
|
|
|
|
|
|
|
|
This makes mad.txt draw correctly. Yay!
|
|
|