Permalink
Browse files

GE Debugger: Stall less liberally on GE dump exec.

This makes it easier to compare performance and flushing bugs.
  • Loading branch information...
unknownbrackets committed Jun 11, 2018
1 parent f7443aa commit 582bc2d60ae80e6e8523776d99751890d363ed45
Showing with 73 additions and 28 deletions.
  1. +73 −28 GPU/Debugger/Record.cpp
View
@@ -17,6 +17,7 @@
#include <algorithm>
#include <cstring>
#include <functional>
#include <vector>
#include <snappy-c.h>
#include "base/stringutil.h"
@@ -93,6 +94,7 @@ class DumpExecute {
bool Run();
private:
void SyncStall();
bool SubmitCmds(void *p, u32 sz);
void SubmitListEnd();
@@ -114,6 +116,7 @@ class DumpExecute {
u32 execListID = 0;
const int LIST_BUF_SIZE = 256 * 1024;
std::vector<u32> execListQueue;
u16 lastBufw_[8]{};
};
// This class maps pushbuffer (dump data) sections to PSP memory.
@@ -125,7 +128,7 @@ class DumpExecute {
class BufMapping {
public:
// Returns a pointer to contiguous memory for this access, or else 0 (failure).
u32 Map(u32 bufpos, u32 sz);
u32 Map(u32 bufpos, u32 sz, const std::function<void()> &flush);
// Clear and reset allocations made.
void Reset() {
@@ -140,8 +143,8 @@ class BufMapping {
}
protected:
u32 MapSlab(u32 bufpos);
u32 MapExtra(u32 bufpos, u32 sz);
u32 MapSlab(u32 bufpos, const std::function<void()> &flush);
u32 MapExtra(u32 bufpos, u32 sz, const std::function<void()> &flush);
enum {
// These numbers kept low because we only have 24 MB of user memory to map into.
@@ -212,20 +215,20 @@ class BufMapping {
static BufMapping execMapping;
u32 BufMapping::Map(u32 bufpos, u32 sz) {
u32 BufMapping::Map(u32 bufpos, u32 sz, const std::function<void()> &flush) {
int slab1 = bufpos / SLAB_SIZE;
int slab2 = (bufpos + sz - 1) / SLAB_SIZE;
if (slab1 == slab2) {
// Doesn't straddle, so we can just map to a slab.
return MapSlab(bufpos);
return MapSlab(bufpos, flush);
} else {
// We need contiguous, so we'll just allocate separately.
return MapExtra(bufpos, sz);
return MapExtra(bufpos, sz, flush);
}
}
u32 BufMapping::MapSlab(u32 bufpos) {
u32 BufMapping::MapSlab(u32 bufpos, const std::function<void()> &flush) {
u32 slab_pos = (bufpos / SLAB_SIZE) * SLAB_SIZE;
int best = 0;
@@ -239,21 +242,27 @@ u32 BufMapping::MapSlab(u32 bufpos) {
}
}
// Stall before mapping a new slab.
flush();
// Okay, we need to allocate.
if (!slabs_[best].Setup(slab_pos)) {
return 0;
}
return slabs_[best].Ptr(bufpos);
}
u32 BufMapping::MapExtra(u32 bufpos, u32 sz) {
u32 BufMapping::MapExtra(u32 bufpos, u32 sz, const std::function<void()> &flush) {
for (int i = 0; i < EXTRA_COUNT; ++i) {
// Might be likely to reuse larger buffers straddling slabs.
if (extra_[i].Matches(bufpos, sz)) {
return extra_[i].Ptr();
}
}
// Stall first, so we don't stomp existing RAM.
flush();
int i = extraOffset_;
extraOffset_ = (extraOffset_ + 1) % EXTRA_COUNT;
@@ -754,6 +763,17 @@ void NotifyFrame() {
}
}
void DumpExecute::SyncStall() {
gpu->UpdateStall(execListID, execListPos);
s64 listTicks = gpu->GetListTicks(execListID);
if (listTicks != -1) {
currentMIPS->downcount -= listTicks - CoreTiming::GetTicks();
}
// Make sure downcount doesn't overflow.
CoreTiming::ForceCheck();
}
bool DumpExecute::SubmitCmds(void *p, u32 sz) {
if (execListBuf == 0) {
u32 allocSize = LIST_BUF_SIZE;
@@ -784,22 +804,42 @@ bool DumpExecute::SubmitCmds(void *p, u32 sz) {
Memory::Write_U32((GE_CMD_JUMP << 24) | (execListBuf & 0x00FFFFFF), execListPos + 4);
execListPos = execListBuf;
// Don't continue until we've stalled.
SyncStall();
}
Memory::MemcpyUnchecked(execListPos, execListQueue.data(), pendingSize);
execListPos += pendingSize;
u32 writePos = execListPos;
Memory::MemcpyUnchecked(execListPos, p, sz);
execListPos += sz;
execListQueue.clear();
gpu->UpdateStall(execListID, execListPos);
s64 listTicks = gpu->GetListTicks(execListID);
if (listTicks != -1) {
currentMIPS->downcount -= listTicks - CoreTiming::GetTicks();
// TODO: Unfortunate. Maybe Texture commands should contain the bufw instead.
// The goal here is to realistically combine prims in dumps. Stalling for the bufw flushes.
u32_le *ops = (u32_le *)Memory::GetPointer(writePos);
for (u32 i = 0; i < sz / 4; ++i) {
u32 cmd = ops[i] >> 24;
if (cmd >= GE_CMD_TEXBUFWIDTH0 && cmd <= GE_CMD_TEXBUFWIDTH7) {
int level = cmd - GE_CMD_TEXBUFWIDTH0;
u16 bufw = ops[i] & 0xFFFF;
// NOP the address part of the command to avoid a flush too.
if (bufw == lastBufw_[level])
ops[i] = GE_CMD_NOP << 24;
else
ops[i] = (gstate.texbufwidth[level] & 0xFFFF0000) | bufw;
lastBufw_[level] = bufw;
}
// Since we're here anyway, also NOP out texture addresses.
// This makes Step Tex not hit phantom textures.
if (cmd >= GE_CMD_TEXADDR0 && cmd <= GE_CMD_TEXADDR7) {
ops[i] = GE_CMD_NOP << 24;
}
}
// Make sure downcount doesn't overflow.
CoreTiming::ForceCheck();
execListQueue.clear();
return true;
}
@@ -814,13 +854,8 @@ void DumpExecute::SubmitListEnd() {
Memory::Write_U32(GE_CMD_END << 24, execListPos + 4);
execListPos += 8;
gpu->UpdateStall(execListID, execListPos);
currentMIPS->downcount -= gpu->GetListTicks(execListID) - CoreTiming::GetTicks();
SyncStall();
gpu->ListSync(execListID, 0);
// Make sure downcount doesn't overflow.
CoreTiming::ForceCheck();
}
void DumpExecute::Init(u32 ptr, u32 sz) {
@@ -833,7 +868,7 @@ void DumpExecute::Registers(u32 ptr, u32 sz) {
}
void DumpExecute::Vertices(u32 ptr, u32 sz) {
u32 psp = execMapping.Map(ptr, sz);
u32 psp = execMapping.Map(ptr, sz, std::bind(&DumpExecute::SyncStall, this));
if (psp == 0) {
ERROR_LOG(SYSTEM, "Unable to allocate for vertices");
return;
@@ -844,7 +879,7 @@ void DumpExecute::Vertices(u32 ptr, u32 sz) {
}
void DumpExecute::Indices(u32 ptr, u32 sz) {
u32 psp = execMapping.Map(ptr, sz);
u32 psp = execMapping.Map(ptr, sz, std::bind(&DumpExecute::SyncStall, this));
if (psp == 0) {
ERROR_LOG(SYSTEM, "Unable to allocate for indices");
return;
@@ -855,7 +890,7 @@ void DumpExecute::Indices(u32 ptr, u32 sz) {
}
void DumpExecute::Clut(u32 ptr, u32 sz) {
u32 psp = execMapping.Map(ptr, sz);
u32 psp = execMapping.Map(ptr, sz, std::bind(&DumpExecute::SyncStall, this));
if (psp == 0) {
ERROR_LOG(SYSTEM, "Unable to allocate for clut");
return;
@@ -866,12 +901,15 @@ void DumpExecute::Clut(u32 ptr, u32 sz) {
}
void DumpExecute::TransferSrc(u32 ptr, u32 sz) {
u32 psp = execMapping.Map(ptr, sz);
u32 psp = execMapping.Map(ptr, sz, std::bind(&DumpExecute::SyncStall, this));
if (psp == 0) {
ERROR_LOG(SYSTEM, "Unable to allocate for transfer");
return;
}
// Need to sync in order to access gstate.transfersrcw.
SyncStall();
execListQueue.push_back((gstate.transfersrcw & 0xFF00FFFF) | ((psp >> 8) & 0x00FF0000));
execListQueue.push_back(((GE_CMD_TRANSFERSRC) << 24) | (psp & 0x00FFFFFF));
}
@@ -886,6 +924,7 @@ void DumpExecute::Memset(u32 ptr, u32 sz) {
const MemsetCommand *data = (const MemsetCommand *)(pushbuf.data() + ptr);
if (Memory::IsVRAMAddress(data->dest)) {
SyncStall();
gpu->PerformMemorySet(data->dest, (u8)data->value, data->sz);
}
}
@@ -896,20 +935,23 @@ void DumpExecute::MemcpyDest(u32 ptr, u32 sz) {
void DumpExecute::Memcpy(u32 ptr, u32 sz) {
if (Memory::IsVRAMAddress(execMemcpyDest)) {
SyncStall();
Memory::MemcpyUnchecked(execMemcpyDest, pushbuf.data() + ptr, sz);
gpu->PerformMemoryUpload(execMemcpyDest, sz);
}
}
void DumpExecute::Texture(int level, u32 ptr, u32 sz) {
u32 psp = execMapping.Map(ptr, sz);
u32 psp = execMapping.Map(ptr, sz, std::bind(&DumpExecute::SyncStall, this));
if (psp == 0) {
ERROR_LOG(SYSTEM, "Unable to allocate for texture");
return;
}
execListQueue.push_back((gstate.texbufwidth[level] & 0xFF00FFFF) | ((psp >> 8) & 0x00FF0000));
execListQueue.push_back(((GE_CMD_TEXADDR0 + level) << 24) | (psp & 0x00FFFFFF));
u32 bufwCmd = GE_CMD_TEXBUFWIDTH0 + level;
u32 addrCmd = GE_CMD_TEXADDR0 + level;
execListQueue.push_back((bufwCmd << 24) | ((psp >> 8) & 0x00FF0000) | lastBufw_[level]);
execListQueue.push_back((addrCmd << 24) | (psp & 0x00FFFFFF));
}
void DumpExecute::Display(u32 ptr, u32 sz) {
@@ -920,6 +962,9 @@ void DumpExecute::Display(u32 ptr, u32 sz) {
DisplayBufData *disp = (DisplayBufData *)(pushbuf.data() + ptr);
// Sync up drawing.
SyncStall();
__DisplaySetFramebuf(disp->topaddr.ptr, disp->linesize, disp->pixelFormat, 1);
__DisplaySetFramebuf(disp->topaddr.ptr, disp->linesize, disp->pixelFormat, 0);
}

0 comments on commit 582bc2d

Please sign in to comment.