This is a cleanup of the gameini database. Specifically:
It removes the "Issues=..." lines and their respective comments since
they don't show up in the gui, are very old and wrong in most cases.
They contain probably more than 4 year old comments (i don't have a clue
who or when wrote them).
Also remove the "EFBCopyCacheEnable = True" lines from the database.
These were put at a time that efb to Ram had resolution issues without
cache being enabled, safe texture cache could be disabled and speed was
detrimental. Now that efb to ram doesn't have the same res issues as
back then, safe texture cache can't be disabled with speed gains being
non existent (you get even a speed hit if texture cache is put to safe)
i think it should be removed from the database.
Missed a define in x64MemTools for when the thought process was Android == ARM
Also changes the variable we use for choosing which folders to copy to and from our jni file.
This has changed since the x86_64 build target uses the library folder x86-64, which is stupid and annoying.
It didn't behave correctly with an input of zero, resulting in some games
breaking (at the least, Fight Night 2). This should be fixed now.
Also clean it up, add a few comments, and fix some variants of the instruction
that are so rare that they probably never got tested.
Google has gotten their act together and fixes a few of the signal handling headers.
Change over to a header that works on both r10 32bit and r10 64bit.
32bit has the old "broken" headers as in some didn't even exist.
64bit has the "fixed" headers that one would expect on any regular unix system.
There were some fixes back on March 13th, 2014 for fixing compiling on MIPS64.
Also some fixes on June 25th, 2014 for SPARC64 fixes.
Probably more things, but those are what I care about.
To avoid FPRs being pushed unnecessarily, I checked the uses: DSPEmitter
doesn't use FPRs, and VertexLoader doesn't use anything but RAX, so I
specified the register list accordingly. The regular JIT, however, does
use FPRs, and as far as I can tell, it was incorrect not to save them in
the outer routine. Since the dispatcher loop is only exited when
pausing or stopping, this should have no noticeable performance impact.
- Factor common work into a helper function.
- Replace confusingly named "noProlog" with "rsp_alignment". Now that
x86 is not supported, we can just specify it explicitly as 8 for
clarity.
- Add the option to include more frame size, which I'll need later.
- Revert a change by magumagu in March which replaced MOVAPD with MOVUPD
on account of 32-bit Windows, since it's no longer supported. True,
apparently recent processors don't execute the former any faster if the
pointer is, in fact, aligned, but there's no point using MOVUPD for
something that's guaranteed to be aligned...
(I discovered that GenFrsqrte and GenFres were incorrectly passing false
to noProlog - they were, in fact, functions without prologs, the
original meaning of the parameter - which caused the previous change to
break. This is now fixed.)
Linktothepast