Skip to content

Commit

Permalink
Changing psyqo's DMA modes
Browse files Browse the repository at this point in the history 
- Chained DMA needs to use GPU DMA mode 2.
- Normal DMA can't go beyond 8 words per block.
  • Loading branch information
@nicolasnoble
nicolasnoble committed Oct 20, 2024
1 parent 06d7890 commit 0964370
Show file tree
Hide file tree
Showing 3 changed files with 324 additions and 2 deletions.
6 changes: 4 additions & 2 deletions src/mips/psyqo/src/gpu.cpp
View file
Original file line number Diff line number Diff line change
Expand Up @@ -176,7 +176,7 @@ void psyqo::GPU::initialize(const psyqo::GPU::Configuration &config) {
}
} break;
}
// GPU back in Fifo polling mode, in case we were uploading to VRAM
// GPU back in Fifo polling mode
Hardware::GPU::Ctrl = 0x04000001;
if (m_flushCacheAfterDMA) {
Prim::FlushCache fc;
Expand Down Expand Up @@ -350,7 +350,7 @@ void psyqo::GPU::uploadToVRAM(const uint16_t *data, Rect region, eastl::function
bcr >>= 1;

unsigned bs = 1;
while (((bcr & 1) == 0) && (bs < 16)) {
while (((bcr & 1) == 0) && (bs < 8)) {
bs <<= 1;
bcr >>= 1;
}
Expand Down Expand Up @@ -474,6 +474,8 @@ void psyqo::GPU::sendChain(eastl::function<void()> &&callback, DMA::DmaCallback
void psyqo::GPU::scheduleChainedDMA(uintptr_t head) {
Kernel::assert((DMA_CTRL[DMA_GPU].CHCR & 0x01000000) == 0, "GPU DMA busy");
while ((Hardware::GPU::Ctrl & uint32_t(0x10000000)) == 0);
// Using block command mode, probably?
Hardware::GPU::Ctrl = 0x04000002;
DMA_CTRL[DMA_GPU].MADR = head;
eastl::atomic_signal_fence(eastl::memory_order_release);
DMA_CTRL[DMA_GPU].CHCR = 0x01000401;
Expand Down
11 changes: 11 additions & 0 deletions src/mips/tests/gpu/Makefile
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,11 @@
TARGET = gpu
TYPE = ps-exe

SRCS += \
../../common/syscalls/printf.s \
../../common/crt0/memory-s.s \
../../common/crt0/memory-c.c \
../../common/crt0/crt0.s \
gpu.c \

include ../../common.mk
309 changes: 309 additions & 0 deletions src/mips/tests/gpu/gpu.c
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,309 @@
/*

Check warning on line 1 in src/mips/tests/gpu/gpu.c

View check run for this annotation

CodeScene Delta Analysis / CodeScene Cloud Delta Analysis (main)

❌ New issue: Overall Code Complexity 

This module has a mean cyclomatic complexity of 5.57 across 7 functions. The mean complexity threshold is 4. This file has many conditional statements (e.g. if, for, while) across its implementation, leading to lower code health. Avoid adding more conditionals.
MIT License
Copyright (c) 2024 PCSX-Redux authors
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
*/

#include "common/hardware/gpu.h"

#include <stdint.h>

#include "common/hardware/dma.h"
#include "common/hardware/hwregs.h"
#include "common/hardware/irq.h"
#include "common/syscalls/syscalls.h"

static const uint8_t polydata1[] = {
0xff, 0x80, 0x33, 0x38, 0x01, 0x00, 0xff, 0xff, 0xff, 0x80, 0x33, 0x00, 0x30, 0x00, 0x03, 0x00,
0xff, 0x80, 0x33, 0x00, 0x05, 0x00, 0x29, 0x00, 0xff, 0x80, 0x33, 0x00, 0x30, 0x00, 0x2e, 0x00,
};
static const uint8_t polydata2[] = {
0x80, 0xff, 0x33, 0x38, 0x3b, 0x00, 0x0a, 0x00, 0x80, 0xff, 0x33, 0x00, 0x71, 0x00, 0x02, 0x00,
0x80, 0xff, 0x33, 0x00, 0x44, 0x00, 0x31, 0x00, 0x80, 0xff, 0x33, 0x00, 0x73, 0x00, 0x23, 0x00,
};
static const uint8_t polydata3[] = {
0x33, 0xff, 0x80, 0x38, 0x89, 0x00, 0x0b, 0x00, 0x33, 0xff, 0x80, 0x00, 0xa6, 0x00, 0x0a, 0x00,
0x33, 0xff, 0x80, 0x00, 0x7f, 0x00, 0x37, 0x00, 0x33, 0xff, 0x80, 0x00, 0xa9, 0x00, 0x31, 0x00,
};
static const uint8_t polydata4[] = {
0x33, 0x80, 0xff, 0x38, 0xb6, 0x00, 0x0f, 0x00, 0x33, 0x80, 0xff, 0x00, 0xda, 0x00, 0x0d, 0x00,
0x33, 0x80, 0xff, 0x00, 0xbb, 0x00, 0x3e, 0x00, 0x33, 0x80, 0xff, 0x00, 0xdf, 0x00, 0x3e, 0x00,
};
static const uint8_t polydata5[] = {
0xff, 0x00, 0x00, 0x38, 0x01, 0x00, 0x4e, 0x00, 0xff, 0x00, 0x00, 0x00, 0x2d, 0x00, 0x55, 0x00,
0xff, 0x00, 0x00, 0x00, 0x08, 0x00, 0x77, 0x00, 0xff, 0x00, 0x00, 0x00, 0x30, 0x00, 0x7e, 0x00,
};
static const uint8_t polydata6[] = {
0x00, 0xff, 0x00, 0x38, 0x3e, 0x00, 0x57, 0x00, 0x00, 0xff, 0x00, 0x00, 0x6f, 0x00, 0x56, 0x00,
0x00, 0xff, 0x00, 0x00, 0x43, 0x00, 0x7e, 0x00, 0x00, 0xff, 0x00, 0x00, 0x72, 0x00, 0x75, 0x00,
};
static const uint8_t polydata7[] = {
0x00, 0x00, 0xff, 0x38, 0x87, 0x00, 0x5e, 0x00, 0x00, 0x00, 0xff, 0x00, 0xa9, 0x00, 0x57, 0x00,
0x00, 0x00, 0xff, 0x00, 0x80, 0x00, 0x86, 0x00, 0x00, 0x00, 0xff, 0x00, 0xa8, 0x00, 0x81, 0x00,
};
static const uint8_t polydata8[] = {
0xff, 0xff, 0x00, 0x38, 0xb5, 0x00, 0x5e, 0x00, 0xff, 0xff, 0x00, 0x00, 0xdb, 0x00, 0x5d, 0x00,
0xff, 0xff, 0x00, 0x00, 0xbc, 0x00, 0x8f, 0x00, 0xff, 0xff, 0x00, 0x00, 0xde, 0x00, 0x8e, 0x00,
};
static const uint8_t polydata9[] = {
0xff, 0x00, 0xff, 0x38, 0xff, 0xff, 0xa2, 0x00, 0xff, 0x00, 0xff, 0x00, 0x2f, 0x00, 0xa3, 0x00,
0xff, 0x00, 0xff, 0x00, 0x08, 0x00, 0xc6, 0x00, 0xff, 0x00, 0xff, 0x00, 0x30, 0x00, 0xcd, 0x00,
};
static const uint8_t polydata10[] = {
0x00, 0xff, 0xff, 0x38, 0x3e, 0x00, 0xa7, 0x00, 0x00, 0xff, 0xff, 0x00, 0x6e, 0x00, 0xa5, 0x00,
0x00, 0xff, 0xff, 0x00, 0x43, 0x00, 0xce, 0x00, 0x00, 0xff, 0xff, 0x00, 0x73, 0x00, 0xc6, 0x00,
};
static const uint8_t polydata11[] = {
0xff, 0xff, 0xff, 0x38, 0x86, 0x00, 0xad, 0x00, 0xff, 0xff, 0xff, 0x00, 0xaa, 0x00, 0xa7, 0x00,
0xff, 0xff, 0xff, 0x00, 0x7f, 0x00, 0xd6, 0x00, 0xff, 0xff, 0xff, 0x00, 0xa9, 0x00, 0xd2, 0x00,
};
static const uint8_t polydata12[] = {
0x80, 0x80, 0x80, 0x38, 0xb5, 0x00, 0xae, 0x00, 0x80, 0x80, 0x80, 0x00, 0xdb, 0x00, 0xae, 0x00,
0x80, 0x80, 0x80, 0x00, 0xbc, 0x00, 0xdf, 0x00, 0x80, 0x80, 0x80, 0x00, 0xde, 0x00, 0xdd, 0x00,
};

static void reset() {
IMASK = 0;
IREG = 0;
for (unsigned i = 0; i < 7; i++) {
DMA_CTRL[i].CHCR = 0;
DMA_CTRL[i].BCR = 0;
DMA_CTRL[i].MADR = 0;
}
DPCR = 0x800;
uint32_t dicr = DICR;
DICR = dicr;
DICR = 0;
GPU_STATUS = 0x00000000;
struct DisplayModeConfig config = {
.hResolution = HR_320,
.vResolution = VR_240,
.videoMode = VM_NTSC,
.colorDepth = CD_15BITS,
.videoInterlace = VI_OFF,
.hResolutionExtended = HRE_NORMAL,
};
setDisplayMode(&config);
setHorizontalRange(0, 0xa00);
setVerticalRange(16, 255);
setDisplayArea(0, 0);
setDrawingArea(0, 0, 320, 240);
}

static int s_frame = 0;
static void setRelativeDrawingOffset(int16_t x, int16_t y) { setDrawingOffset(s_frame ? x : x + 320, y); }

static void flip() {
uint32_t imask = IMASK;
IMASK = imask | IRQ_VBLANK;
while ((IREG & IRQ_VBLANK) == 0);
IREG &= ~IRQ_VBLANK;
IMASK = imask;
s_frame ^= 1;
setDisplayArea(s_frame ? 320 : 0, 0);
setDrawingArea(s_frame ? 0 : 320, 0, s_frame ? 320 : 640, 240);
setRelativeDrawingOffset(0, 0);
struct FastFill ff = {
.c = {{.r = 0x68, .g = 0xb0, .b = 0xd8}},
.x = s_frame ? 0 : 320,
.y = 0,
.w = 320,
.h = 240,
};
fastFill(&ff);
}

static void sendOnePolygon(const uint8_t* data_) {
const uint32_t* data = (const uint32_t*)data_;
waitGPU();
for (unsigned i = 0; i < 8; i++) {
GPU_DATA = data[i];
}
}

static void sendOnePolygonNoWait(const uint8_t* data_) {
const uint32_t* data = (const uint32_t*)data_;
for (unsigned i = 0; i < 8; i++) {
GPU_DATA = data[i];
}
}

static int s_frameCount = 0;
static int s_sequenceId = 0;

static uint32_t s_singleBlock[8 * 12];
static uint32_t s_DMAChain[9 * 12];

static void sendSingleBlock(unsigned blockSize) {
uint32_t bcr = (sizeof(s_singleBlock) >> 2) / blockSize;
DMA_CTRL[DMA_GPU].MADR = (uintptr_t)s_singleBlock;
DMA_CTRL[DMA_GPU].BCR = (bcr << 16) | blockSize;
DMA_CTRL[DMA_GPU].CHCR = 0x01000201;
}

int main() {
__builtin_memcpy(s_singleBlock, polydata1, sizeof(polydata1));
__builtin_memcpy(s_singleBlock + 8, polydata2, sizeof(polydata2));
__builtin_memcpy(s_singleBlock + 16, polydata3, sizeof(polydata3));
__builtin_memcpy(s_singleBlock + 24, polydata4, sizeof(polydata4));
__builtin_memcpy(s_singleBlock + 32, polydata5, sizeof(polydata5));
__builtin_memcpy(s_singleBlock + 40, polydata6, sizeof(polydata6));
__builtin_memcpy(s_singleBlock + 48, polydata7, sizeof(polydata7));
__builtin_memcpy(s_singleBlock + 56, polydata8, sizeof(polydata8));
__builtin_memcpy(s_singleBlock + 64, polydata9, sizeof(polydata9));
__builtin_memcpy(s_singleBlock + 72, polydata10, sizeof(polydata10));
__builtin_memcpy(s_singleBlock + 80, polydata11, sizeof(polydata11));
__builtin_memcpy(s_singleBlock + 88, polydata12, sizeof(polydata12));

__builtin_memcpy(s_DMAChain + 1, polydata1, sizeof(polydata1));
__builtin_memcpy(s_DMAChain + 10, polydata2, sizeof(polydata2));
__builtin_memcpy(s_DMAChain + 19, polydata3, sizeof(polydata3));
__builtin_memcpy(s_DMAChain + 28, polydata4, sizeof(polydata4));
__builtin_memcpy(s_DMAChain + 37, polydata5, sizeof(polydata5));
__builtin_memcpy(s_DMAChain + 46, polydata6, sizeof(polydata6));
__builtin_memcpy(s_DMAChain + 55, polydata7, sizeof(polydata7));
__builtin_memcpy(s_DMAChain + 64, polydata8, sizeof(polydata8));
__builtin_memcpy(s_DMAChain + 73, polydata9, sizeof(polydata9));
__builtin_memcpy(s_DMAChain + 82, polydata10, sizeof(polydata10));
__builtin_memcpy(s_DMAChain + 91, polydata11, sizeof(polydata11));
__builtin_memcpy(s_DMAChain + 100, polydata12, sizeof(polydata12));

for (unsigned i = 0; i < 11; i++) {
s_DMAChain[i * 9] = 0x08000000 | ((uintptr_t)(&s_DMAChain[i * 9 + 9]) & 0xffffff);
}
s_DMAChain[99] = 0x08ffffff;

reset();
flip();
flip();
enableDisplay();
while (1) {
if (s_frameCount == 0) {
*(volatile uint8_t*)(s_sequenceId) = s_sequenceId;
}
setRelativeDrawingOffset(0, 0);
switch (s_sequenceId) {
case 0:
sendOnePolygon(polydata1);
sendOnePolygon(polydata2);
sendOnePolygon(polydata3);
sendOnePolygon(polydata4);
sendOnePolygon(polydata5);
sendOnePolygon(polydata6);
sendOnePolygon(polydata7);
sendOnePolygon(polydata8);
sendOnePolygon(polydata9);
sendOnePolygon(polydata10);
sendOnePolygon(polydata11);
sendOnePolygon(polydata12);
break;
case 1: // derps
sendOnePolygonNoWait(polydata1);
sendOnePolygonNoWait(polydata2);
sendOnePolygonNoWait(polydata3);
sendOnePolygonNoWait(polydata4);
sendOnePolygonNoWait(polydata5);
sendOnePolygonNoWait(polydata6);
sendOnePolygonNoWait(polydata7);
sendOnePolygonNoWait(polydata8);
sendOnePolygonNoWait(polydata9);
sendOnePolygonNoWait(polydata10);
sendOnePolygonNoWait(polydata11);
sendOnePolygonNoWait(polydata12);
break;
case 2:
sendGPUStatus(0x04000001);
sendSingleBlock(1);
break;
case 3: // freezes
// sendGPUStatus(0x04000002);
// sendSingleBlock(1);
break;
case 4:
sendGPUStatus(0x04000001);
sendSingleBlock(2);
break;
case 5: // freezes
// sendGPUStatus(0x04000002);
// sendSingleBlock(2);
break;
case 6:
sendGPUStatus(0x04000001);
sendSingleBlock(4);
break;
case 7: // freezes
// sendGPUStatus(0x04000002);
// sendSingleBlock(4);
break;
case 8:
sendGPUStatus(0x04000001);
sendSingleBlock(8);
break;
case 9:
sendGPUStatus(0x04000002);
sendSingleBlock(8);
break;
case 10: // derps
sendGPUStatus(0x04000001);
sendSingleBlock(12);
break;
case 11: // freezes
// sendGPUStatus(0x04000002);
// sendSingleBlock(12);
break;
case 12: // derps
sendGPUStatus(0x04000001);
sendSingleBlock(16);
break;
case 13:
sendGPUStatus(0x04000002);
sendSingleBlock(16);
break;
case 14:
sendGPUStatus(0x04000001);
DMA_CTRL[DMA_GPU].MADR = (uintptr_t)s_DMAChain;
DMA_CTRL[DMA_GPU].CHCR = 0x01000401;
break;
case 15:
sendGPUStatus(0x04000002);
DMA_CTRL[DMA_GPU].MADR = (uintptr_t)s_DMAChain;
DMA_CTRL[DMA_GPU].CHCR = 0x01000401;
break;
}
while ((DMA_CTRL[DMA_GPU].CHCR & 0x01000000) != 0);
for (unsigned i = 0; i < s_sequenceId; i++) {
setRelativeDrawingOffset(i * 10 + 10, 220);
sendGPUData(0x70ffffff);
GPU_DATA = 0x00000000;
}
flip();
if (s_frameCount++ == 90) {
s_sequenceId++;
s_frameCount = 0;
if (s_sequenceId == 16) {
s_sequenceId = 0;
}
}
}
return 0;
}

Check warning on line 309 in src/mips/tests/gpu/gpu.c

View check run for this annotation

CodeScene Delta Analysis / CodeScene Cloud Delta Analysis (main)

❌ New issue: Complex Method 

main has a cyclomatic complexity of 24, threshold = 9. This function has many conditional statements (e.g. if, for, while), leading to lower code health. Avoid adding more conditionals and code to it without refactoring.

Check warning on line 309 in src/mips/tests/gpu/gpu.c

View check run for this annotation

CodeScene Delta Analysis / CodeScene Cloud Delta Analysis (main)

❌ New issue: Bumpy Road Ahead 

main has 2 blocks with nested conditional logic. Any nesting of 2 or deeper is considered. Threshold is one single, nested block per function. The Bumpy Road code smell is a function that contains multiple chunks of nested conditional logic. The deeper the nesting and the more bumps, the lower the code health.

0 comments on commit 0964370

Please sign in to comment.