Skip to content

Commit

Permalink
Converted everything but tris
Browse files Browse the repository at this point in the history
  • Loading branch information
@sauraen
sauraen committed Aug 25, 2024
1 parent a28b463 commit a3c2702
Show file tree
Hide file tree
Showing 2 changed files with 49 additions and 38 deletions.
84 changes: 47 additions & 37 deletions f3dex3.s
View file
Original file line number Diff line number Diff line change
Expand Up @@ -620,7 +620,7 @@ clipTempVerts:
.org ((clipTempVerts + 0xF) & 0xFF0)
// Vertex addresses, to avoid a multiply-add for each vertex index lookup
vertexTable:
.skip (G_MAX_VERTS * 2) // halfword for each vertex; EX2 has an extra end addr but this is not needed
.skip ((G_MAX_VERTS + 8) * 2) // halfword for each vertex; need 1 extra end addr, easier to write 8 extra

.if . > yieldDataFooter
// Need to fit everything through vertex buffer in yield buffer, would like
Expand Down Expand Up @@ -995,8 +995,7 @@ tempPrevVtxGarbage equ 0x50 // Up to 2 * 0x26 = 0x4C used -> to 0x9C
.endmacro

.macro instantiate_branch_wz
jal vtx_addrs_from_cmd // byte 3 = vtx being tested; addr -> $10
nop
lhu $10, (vertexTable)(cmd_w0) // Vertex addr from byte 3
.if CFG_G_BRANCH_W // G_BRANCH_W/G_BRANCH_Z difference; this defines F3DZEX vs. F3DEX2
lh $10, VTX_W_INT($10) // read the w coordinate of the vertex (f3dzex)
.else
Expand Down Expand Up @@ -1125,7 +1124,8 @@ continue_from_os_task:
lw perfCounterB, mITMatrix + YDF_OFFSET_PERFCOUNTERB
lw perfCounterC, mITMatrix + YDF_OFFSET_PERFCOUNTERC
lw perfCounterD, mITMatrix + YDF_OFFSET_PERFCOUNTERD
lw taskDataPtr, OSTask + OSTask_data_ptr
jal fill_vertex_table
lw taskDataPtr, OSTask + OSTask_data_ptr
finish_setup:
.if CFG_PROFILING_C
mfc0 $11, DPC_CLOCK
Expand Down Expand Up @@ -1245,19 +1245,9 @@ call_ret_common:
G_LOAD_UCODE_handler:
j load_overlay_0_and_enter // Delay slot is harmless
G_MODIFYVTX_handler:
// Command byte 3 = vtx being modified; its addr -> $10
li $ra, do_moveword // Moveword adds cmd_w0 to $10 for final addr
lbu cmd_w0, (inputBufferEnd - 0x07)(inputBufferPos) // offset in vtx
vtx_addrs_from_cmd:
// Treat eight bytes of last command each as vertex indices << 1
// inputBufferEnd is close enough to the end of DMEM to fit in signed offset
lpv $v27[0], (inputBufferEndSgn - 8)(inputBufferPos)
// Also out elem 3 -> $10, elem 7 -> $3 because these are used more than once
vmudn $v29, vOne, $v30[0] // Address of vertex buffer
vmadl $v27, $v27, $v30[1] // Plus vtx indices times length
sb $zero, materialCullMode // This covers modify vtx, branchZ, cull
jr $ra
mfc2 $10, $v27[6]
lhu $10, (vertexTable)(cmd_w0) // Byte 3 = vtx being modified
j do_moveword // Moveword adds cmd_w0 to $10 for final addr
lbu cmd_w0, (inputBufferEnd - 0x07)(inputBufferPos) // offset in vtx, bit 15 clear

G_TRIFAN_handler:
li $1, 0x8000 // $ra negative = flag for G_TRIFAN
Expand Down Expand Up @@ -1332,27 +1322,25 @@ G_MEMSET_handler:
.endif

G_VTX_handler:
srl $2, cmd_w0, 11 // n << 1
sub $2, cmd_w0, $2 // = v0 << 1
vmudn $v29, vOne, $v30[0] // Address of vertex buffer
sb $2, (inputBufferEnd - 8)(inputBufferPos) // Store v0 << 1 as byte 0
lpv $v27[0], (inputBufferEndSgn - 8)(inputBufferPos) // (v0 + n) << 1 is byte 3
sb $zero, materialCullMode // This covers vtx
lhu $1, (inputBufferEnd - 0x07)(inputBufferPos) // $1 = size in bytes = vtx count * 0x10
lhu dmemAddr, (vertexTable)(cmd_w0) // (v0 + n) end address; up to 56 inclusive
jal segmented_to_physical // Convert address in cmd_w1_dram to physical
vmadl $v27, $v27, $v30[1] // Plus vtx indices times length
mfc2 $10, $v27[6]
lhu $1, (inputBufferEnd - 0x07)(inputBufferPos) // $1 = size in bytes = vtx count * 0x10
andi dmemAddr, dmemAddr, 0xFFF8 // Round down end addr to DMA word; one input vtx still fits in one internal vtx
sub dmemAddr, dmemAddr, $1 // Start addr = end addr - size
addi dmaLen, $1, -1 // DMA length is always offset by -1
j dma_read_write
li $ra, 0x8000 | vtx_after_dma // Negative = flag to not to return to clipping in vtx_setup_constants

vtx_after_dma:
lhu $5, geometryModeLabel + 1 // Load middle 2 bytes of geom mode
andi $10, $10, 0xFFF8 // Round down end addr to DMA word; one input vtx still fits in one internal vtx
jal dma_read_write
sub dmemAddr, $10, $1 // Start addr = end addr - size
mfc2 outputVtxPos, $v27[0] // Address of start
srl $2, cmd_w0, 11 // n << 1
sub $2, cmd_w0, $2 // = v0 << 1
lhu outputVtxPos, (vertexTable)($2) // Address of start
sb $zero, materialCullMode // This covers vtx
.if COUNTER_A_UPPER_VERTEX_COUNT
sll $11, $1, 12 // Vtx count * 0x10000
add perfCounterA, perfCounterA, $11 // Add to vertex count
.endif
li $ra, 0 // Flag to not return to clipping
vtx_setup_constants:
// Computes modified viewport scale and offset including fog info, and stores
// these to temp memory in the RDP buffer. This is only used during vertex write
Expand Down Expand Up @@ -1407,7 +1395,7 @@ vtx_setup_constants:
.endif
vmov sVPS[1], $v20[1] // Negate vscale[1] because RDP top = y=0
.if CFG_LEGACY_VTX_PIPE
bnez $ra, clip_after_constants // Return to clipping if from there
bgtz $ra, clip_after_constants // Return to clipping if from there
.endif
vmov sVPS[5], $v20[1] // Same for second half
vtx_matrix_load:
Expand Down Expand Up @@ -1451,7 +1439,7 @@ vtx_after_lt_setup:
sqv sVPO, (tempViewportOffset)(rdpCmdBufEndP1) // Store viewport offset
vmrg $v30, $v30, $v31[2] // 0; zero AO values
@@skipzeroao:
bnez $ra, clip_after_constants // Return to clipping if from there
bgtz $ra, clip_after_constants // Return to clipping if from there
sqv sVPS, (tempViewportScale)(rdpCmdBufEndP1) // Store viewport scale
lqv vM0I, (mMatrix + 0x00)($zero) // Load M matrix
lqv vM2I, (mMatrix + 0x10)($zero)
Expand Down Expand Up @@ -2342,9 +2330,31 @@ clip_draw_tris_loop:
bne clipPolyWrite, clipPolySelect, clip_draw_tris_loop
addi clipPolySelect, clipPolySelect, 2
clip_done:
lqv $v30, (v30Value)($zero) // Need this repeated here in case we exited early
lh $ra, tempTriRA

fill_vertex_table:
// Create bytes 00-07
li $1, 7
@@loop1:
sb $1, (vertexTable)($1)
bgtz $1, @@loop1
addi $1, $1, -1
// Load to vu and multiply by 2 to get vertex indexes. It would be more cycles
// to change the loop above to count by 2s than the stalls here.
li $2, vertexTable
lpv $v3[0], (0)($2)
li $3, vertexTable + ((G_MAX_VERTS + 8) * 2) // Need 0-56 inclusive, so do 0-63
vmudh $v3, $v3, $v31[3] // 2; now 0x0000, 0x0200, ..., 0x0E00
@@loop2:
vmudn $v29, vOne, $v30[0] // Address of vertex buffer
vmadl $v4, $v3, $v30[1] // Plus vtx indices times length
vadd $v3, $v3, $v30[2] // 0x1000; increment by 8 verts = 16
addi $2, $2, 0x10
bne $2, $3, @@loop2
sqv $v4[0], (-0x10)($2)
jr $ra
lqv $v30, (v30Value)($zero) // Need this repeated here in case we exited early
nop

ovl3_end:
.align 8
Expand Down Expand Up @@ -3115,7 +3125,8 @@ segmented_to_physical:
add cmd_w1_dram, cmd_w1_dram, $11 // Add the segment's address to the masked input address, resulting in the virtual address

G_CULLDL_handler:
jal vtx_addrs_from_cmd // Load start vtx addr in $10
lhu $10, (vertexTable)(cmd_w0) // Start vtx addr
lhu $3, (vertexTable)(cmd_w1_dram) // End vertex
/*
CLIP_OCCLUDED can't be included here because: Suppose the list consists of N-1
verts which are behind the occlusion plane, and 1 vert which is behind the camera
Expand All @@ -3125,8 +3136,7 @@ G_CULLDL_handler:
the occlusion plane if the vert is behind the camera, because this only matters for
G_CULLDL and not for tris.
*/
li $1, (CLIP_SCRN_NPXY | CLIP_CAMPLANE)
mfc2 $3, $v27[14] // End vertex
li $1, (CLIP_SCRN_NPXY | CLIP_CAMPLANE)
lhu $11, VTX_CLIP($10)
culldl_loop:
and $1, $1, $11
Expand Down
3 changes: 2 additions & 1 deletion gbi.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -2974,7 +2974,8 @@ _DW({ \
*
* Internally, a material is defined to start with any set image command, and
* end on any of the following: call, branch, return, vertex, all tri commands,
* modify vertex, branch Z/W, or cull. The physical address of the display list
* tex/fill rectangles, and successes on cull or branch w/z (which are usually
* preceded by vertex loads anyway). The physical address of the display list
* --not the address of the image--is stored when a material is started. If a
* material starts and its physical address is the same as the stored last start
* address, i.e. we're executing the same material display list as the last
Expand Down

0 comments on commit a3c2702

Please sign in to comment.