swung0x48/MobileGL
1
0
Fork 0
mirror of https://github.com/MobileGL-Dev/MobileGL synced 2026-09-12 06:08:30 +09:00
Code Issues Packages Projects Releases Wiki Activity

[Fix] (MG_Backend/DirectGLES): always use native indirect draws with a bound buffer

Browse Source 
Adreno (830) exposes no GL_EXT_base_instance, and gating the native path
on it sent Flywheel's whole MDI call to the CPU loop, which reads the
stale shadow instanceCount (0) and draws nothing. A non-zero reserved
word is benign on mobile drivers, instanced arrays were never
baseInstance-offset in the emulation anyway, and the CPU loop can never
see GPU-written commands - native is strictly better.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
swung0x48
2026-07-08 13:47:45 +00:00
co-authored by Claude Fable 5
parent 41b15955b0
commit 2395a6ded2
1 changed files with 9 additions and 32 deletions
Download Patch File Download Diff File Expand all files Collapse all files
MobileGL/MG_Backend/DirectGLES/DirectGLES.cpp
+9 -32
View File
@@ -1042,26 +1042,16 @@ namespace MobileGL::MG_Backend::DirectGLES {
// execute natively on the GPU so commands written by compute shaders (e.g. Flywheel's // execute natively on the GPU so commands written by compute shaders (e.g. Flywheel's
// culling pipeline updating instanceCount) are honored; the CPU shadow is still consulted // culling pipeline updating instanceCount) are honored; the CPU shadow is still consulted
// for the per-command baseInstance, which is CPU-authored, to feed the mg_BaseInstance // for the per-command baseInstance, which is CPU-authored, to feed the mg_BaseInstance
// shader emulation. Falls back to the CPU per-command loop for client-memory commands or // shader emulation. Without GL_EXT_base_instance a non-zero baseInstance in the command is
// when the driver cannot consume the command's baseInstance field (no GL_EXT_base_instance). // technically undefined in ES; mobile drivers ignore the reserved word, instanced-array
// fetches were never baseInstance-offset here anyway, and the CPU fallback cannot see
// GPU-written command fields at all - so native is never worse. Only client-memory
// commands take the CPU per-command loop.
static void ExecuteIndexedIndirectCommands(GLenum mode, GLenum type, SizeT indexSize, const Uint8* commandBytes, static void ExecuteIndexedIndirectCommands(GLenum mode, GLenum type, SizeT indexSize, const Uint8* commandBytes,
SizeT commandOffset, Bool hasIndirectBuffer, GLsizei drawcount, SizeT commandOffset, Bool hasIndirectBuffer, GLsizei drawcount,
GLsizei stride, const char* label) { GLsizei stride, const char* label) {
Bool useNative = hasIndirectBuffer && SupportsNativeIndirectDraws(); (void)label;
if (useNative && !g_GLESCapabilities.SupportsBaseInstance) { const Bool useNative = hasIndirectBuffer && SupportsNativeIndirectDraws();
for (GLsizei i = 0; i < drawcount; ++i) {
DrawElementsIndirectCommand cmd{};
std::memcpy(&cmd, commandBytes + static_cast<SizeT>(i) * stride, sizeof(cmd));
if (cmd.baseInstance != 0) {
useNative = false;
MGLOG_W("%s: non-zero baseInstance without GL_EXT_base_instance, falling back to CPU "
"emulation (GPU-written command fields will not be honored)",
label);
break;
}
}
}
if (useNative) { if (useNative) {
for (GLsizei i = 0; i < drawcount; ++i) { for (GLsizei i = 0; i < drawcount; ++i) {
DrawElementsIndirectCommand cmd{}; DrawElementsIndirectCommand cmd{};
@@ -1093,21 +1083,8 @@ namespace MobileGL::MG_Backend::DirectGLES {
static void ExecuteArraysIndirectCommands(GLenum mode, const Uint8* commandBytes, SizeT commandOffset, static void ExecuteArraysIndirectCommands(GLenum mode, const Uint8* commandBytes, SizeT commandOffset,
Bool hasIndirectBuffer, GLsizei drawcount, GLsizei stride, Bool hasIndirectBuffer, GLsizei drawcount, GLsizei stride,
const char* label) { const char* label) {
Bool useNative = hasIndirectBuffer && SupportsNativeIndirectDraws(); (void)label;
if (useNative && !g_GLESCapabilities.SupportsBaseInstance) { const Bool useNative = hasIndirectBuffer && SupportsNativeIndirectDraws();
for (GLsizei i = 0; i < drawcount; ++i) {
DrawArraysIndirectCommand cmd{};
std::memcpy(&cmd, commandBytes + static_cast<SizeT>(i) * stride, sizeof(cmd));
if (cmd.baseInstance != 0) {
useNative = false;
MGLOG_W("%s: non-zero baseInstance without GL_EXT_base_instance, falling back to CPU "
"emulation (GPU-written command fields will not be honored)",
label);
break;
}
}
}
if (useNative) { if (useNative) {
for (GLsizei i = 0; i < drawcount; ++i) { for (GLsizei i = 0; i < drawcount; ++i) {
DrawArraysIndirectCommand cmd{}; DrawArraysIndirectCommand cmd{};