mirror/oot
1
0
Fork 0
mirror of https://github.com/zeldaret/oot.git synced 2025-07-04 15:04:31 +00:00
Code Issues Projects Releases Wiki Activity

Make func_800C213C match original ROM (#678)

Browse source 
* Make func_800C213C match original ROM

The C code for this function was very close to done, and was
functionally equivalent to the ROM.  This commit uses a few tricks
to convince the compiler to allocate the same registers as the original
ROM.  The two tricks used are:
1. Introduced a new temporary variable "yinc".
2. Added a "do while(0) {}" line in two places.

Both of these ideas came from the decomp-permuter tool.  I took the
do/while idea and guess as to what it might have been in the original
code.  I imagined that the "do while(0) {}" was probably left over from
a standard multi-line macro, which for some reason had an empty body.
I further suggest that maybe the empty body is because it was a feature
compiled out during RELEASE builds - perhaps a debug logging macro.

* Use simpler code for func_800C213C

Replaced dummy LOG macro with "do {} while(0);".  This version makes
less assumptions about what the original code looked like.
This commit is contained in:
Chris Teague 2021-02-12 14:26:42 -08:00 • committed by GitHub
parent b8c2ec1f60
commit bcf2694f93
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 8 additions and 265 deletions
Download patch file Download diff file Expand all files Collapse all files

17
src/code/PreRender.c
View file

@ -308,14 +308,13 @@ void func_800C2118(PreRenderContext* this, Gfx** gfxp) {
func_800C0F28(this, gfxp, this->zbufSave, this->zbuf);
}
#ifdef NON_MATCHING
// regalloc differences in gDPLoadMultiTile
void func_800C213C(PreRenderContext* this, Gfx** gfxp) {
Gfx* gfx;
s32 y;
s32 y2;
s32 add;
s32 uls;
s32 yinc; // vertical increment amount
s32 ult;
s32 lrx;
s32 lry;
@ -350,20 +349,23 @@ void func_800C213C(PreRenderContext* this, Gfx** gfxp) {
}
uls = 0;
yinc = 1;
ult = y2;
lry = (y2 + add - 1);
lry = (y2 + add - yinc);
gDPLoadMultiTile(gfx++, this->fbufSave, 0x0000, G_TX_RENDERTILE, G_IM_FMT_RGBA, G_IM_SIZ_16b, this->width,
this->height, uls, ult, lrx, lry, 0, G_TX_NOMIRROR | G_TX_WRAP, G_TX_NOMIRROR | G_TX_WRAP,
G_TX_NOMASK, G_TX_NOMASK, G_TX_NOLOD, G_TX_NOLOD);
gDPLoadMultiTile(gfx++, this->cvgSave, 0x0160, 1, G_IM_FMT_I, G_IM_SIZ_8b, this->width, this->height, uls, ult,
do {} while(0); // force register allocation behavior
gDPLoadMultiTile(gfx++, this->cvgSave, 0x0160, yinc, G_IM_FMT_I, G_IM_SIZ_8b, this->width, this->height, uls, ult,
lrx, lry, 0, G_TX_NOMIRROR | G_TX_WRAP, G_TX_NOMIRROR | G_TX_WRAP, G_TX_NOMASK, G_TX_NOMASK,
G_TX_NOLOD, G_TX_NOLOD);
gSPTextureRectangle(gfx++, uls << 2, ult << 2, (lrx + 1) << 2, (lry + 1) << 2, G_TX_RENDERTILE, uls << 5,
do {} while(0); // force register allocation behavior
gSPTextureRectangle(gfx++, uls << 2, ult << 2, (lrx + yinc) << 2, (lry + yinc) << 2, G_TX_RENDERTILE, uls << 5,
ult << 5, 1 << 10, 1 << 10);
if (1) {}
y -= add;
y2 += add;
}
@ -371,9 +373,6 @@ void func_800C213C(PreRenderContext* this, Gfx** gfxp) {
gDPPipeSync(gfx++);
*gfxp = gfx;
}
#else
#pragma GLOBAL_ASM("asm/non_matchings/code/PreRender/func_800C213C.s")
#endif
void func_800C24BC(PreRenderContext* this, Gfx** gfxp) {
func_800C0F28(this, gfxp, this->fbufSave, this->fbuf);