Should fix the flicker in Castlevania.
Fixes#17517
The heuristic worked for Rainbow Six but broke Castlevania, so I'd rather
use a compat flag instead of breaking a different game until we can find a
more reliable heuristic for Rainbow Six.
Fixes#17451, while also keeping the Dante performance fix from #17032.
Of course, it's possible that something else could slow down now... But
hopefully not. This could also fix other problems.
The framebuffer does contain the top of the video frame at the present
time though, so things have partially gone well. However the rest of the
frame lies after it in VRAM. What we probably technically should do when
we see this is to assemble an image from both the framebuffer we find
and VRAM contents exceeding the range of that image, but seems kinda
impractical.
This works, but there seems to be some other issue with how the text is
rendered or copied. Could be a color precision issue confusing the game
I suppose.
Fixes#17030 , or at least improves on it - for optimal performance that
big framebuffer used for bloom should be split like in Killzone, but it's not trivial.
The regression in 1.14 is fixed with this, at least.
I tried it with a few other games with no issues - it seems games are
using TexFlush when needed. But let's see if it really is safe to rely
on that...
There might also be other places we should call DiscardFramebufferCopy
in.
This shouldn't normally happen as conforming drivers are required to
support that texture format, but the software driver that we
accidentally choose on Poco C40 (see issue #16391) doesn't.
That we choose that driver will be fixed separately. This fix on its own
at least lets the user comfortably navigate to settings and switch to
OpenGL.
* Wrap the display globals in a struct
Makes it easier to search/replace these, for future refactorings.
* Some renaming
* Qt buildfix, also fix the Qt build on Mac (got broken with battery changes)
* Attempt at buildfixing ios
* UWP buildfix
And thus change of depth buffer scale/offset.
Previously, old depth buffers with values that now are out of range
could stick around, causing #16941. This clears them to the expected 0
value, which helps Outrun. Ideally we should convert depth buffers to
the new format, but if we can get away without that, that's also nice.
This is enough for #16941.
