[Chore] (Config): triage backend-scoped env toggles under MOBILEGL_ESPRYT_ and MOBILEGL_MAGMA_ prefixes

This commit is contained in:
2026-08-28 00:38:17 -04:00
parent 7fd4550968
commit 5d51372c44
32 changed files with 150 additions and 150 deletions
+30 -30
View File
@@ -69,34 +69,34 @@ namespace MobileGL::MG_Config {
struct FeaturesTable {
// MOBILEGL_DISABLE_TIMERQUERY: do not advertise or use GPU timer queries.
Bool DisableTimerQuery = false;
// MOBILEGL_ENABLE_GLES_TEXTURE_VIEW: advertise GL_ARB_texture_view on DirectGLES when
// MOBILEGL_ESPRYT_ENABLE_TEXTURE_VIEW: advertise GL_ARB_texture_view on DirectGLES when
// the host ES driver has EXT/OES_texture_view. Off by default: the host extension is
// present on Adreno 830 and the functional half of KHR-GL4{2,3}.texture_view still fails
// there, because the view's ES internalformat is normalized independently of the storage
// it aliases (see BackendObject_DirectGLES::BuildAdvertisedExtensions). The flag exists
// so that work can be done without editing the gate.
Bool EnableGlesTextureView = false;
Bool EsprytEnableTextureView = false;
// MOBILEGL_ENABLE_SPIRV_VALIDATION: validate generated and transformed SPIR-V.
// Disabled by default because validation is a diagnostics-only cost.
Bool EnableSpirvValidation = false;
// MOBILEGL_USE_ANGLE: load ANGLE EGL/GLES libraries.
Bool UseAngle = false;
// MOBILEGL_ESPRYT_USE_ANGLE: load ANGLE EGL/GLES libraries.
Bool EsprytUseAngle = false;
#if defined(MOBILEGL_TRACE_ANGLE_VARIANTS)
// MOBILEGL_TRACE_ANGLE_VARIANT: signed trace-APK ANGLE build short hash.
String TraceAngleVariant;
#endif
// MOBILEGL_DISABLE_SUBGROUP: force-disable Vulkan shader subgroup support,
// MOBILEGL_MAGMA_DISABLE_SUBGROUP: force-disable Vulkan shader subgroup support,
// including the opt-in emulated compute path below.
Bool DisableSubgroup = false;
Bool MagmaDisableSubgroup = false;
// MOBILEGL_MAGMA_EMULATE_SUBGROUP: implement GL_KHR_shader_subgroup's compute
// stage on a 32-lane VIRTUAL subgroup lowered to workgroup-shared memory
// (ShaderTranspiler::EmulateSubgroupsPass). Strictly a last resort: it only ever
// engages when this flag is set AND the device has no native subgroup support at
// all - a device with real subgroup operations always uses them natively,
// whatever their width (the known iterationRP defect is patched by
// FixIterationRPSubgroupScratch below instead). Off by default.
// MagmaFixIterationRPSubgroupScratch below instead). Off by default.
Bool MagmaEmulateSubgroup = false;
// MOBILEGL_FIX_ITERATIONRP_SUBGROUP_SCRATCH: patch iterationRP's own bug - the
// MOBILEGL_MAGMA_FIX_ITERATIONRP_SUBGROUP_SCRATCH: patch iterationRP's own bug - the
// pack declares `shared vec2 prefixSumCache[32]` for a 512-invocation exposure
// reduction and indexes it by gl_SubgroupID, so any device with sub-16-lane
// subgroups (8-lane lavapipe -> 64 subgroups) writes shared memory out of
@@ -106,12 +106,12 @@ namespace MobileGL::MG_Config {
// so every other shader passes through byte-identical - as does iterationRP
// itself on >= 16-lane devices. Auto is ON; ForceOff replays the pack's bug
// verbatim.
QuirkOverride FixIterationRPSubgroupScratch = QuirkOverride::Auto;
// MOBILEGL_ITERATIONRP_FIX_BARRIER: repair Program 203's missing workgroup
QuirkOverride MagmaFixIterationRPSubgroupScratch = QuirkOverride::Auto;
// MOBILEGL_MAGMA_ITERATIONRP_FIX_BARRIER: repair Program 203's missing workgroup
// rendezvous between its two reductions over prefixSumCache. Off by default and
// fingerprint-gated by FixIterationRPBarrierPass when enabled.
Bool IterationRPFixBarrier = false;
// MOBILEGL_DERIVE_NUM_SUBGROUPS: replace compute gl_NumSubgroups loads with
Bool MagmaIterationRPFixBarrier = false;
// MOBILEGL_MAGMA_DERIVE_NUM_SUBGROUPS: replace compute gl_NumSubgroups loads with
// ceil(workgroup invocations / gl_SubgroupSize) on the NATIVE subgroup path
// (ShaderTranspiler::DeriveNumSubgroupsPass). Auto is ON: GL requires
// gl_SubgroupID < gl_NumSubgroups, Adreno's builtin reports 1 while the same
@@ -119,7 +119,7 @@ namespace MobileGL::MG_Config {
// whenever the pipeline can request REQUIRE_FULL_SUBGROUPS (which the renderer
// does whenever local_size_x is a multiple of the native width). ForceOff returns
// to the raw driver builtin.
QuirkOverride DeriveNumSubgroups = QuirkOverride::Auto;
QuirkOverride MagmaDeriveNumSubgroups = QuirkOverride::Auto;
// MOBILEGL_ADVERTISE_FP64: add GL_ARB_gpu_shader_fp64 to the advertised extension
// string. `double` in a shader always WORKS - it is narrowed to 32 bits before any
// module reaches a backend (ShaderTranspiler::DemoteFloat64Pass) - but the extension
@@ -132,16 +132,16 @@ namespace MobileGL::MG_Config {
Bool MagmaR11G11B10FFallback = false;
// MOBILEGL_MAGMA_FRAMESINFLIGHT: requested Magma frames in flight, defaulting to 3.
Uint32 MagmaFramesInFlight = 3;
// MOBILEGL_AVOID_SAMPLER_MIPMAP_MIN_FILTER: avoid mipmap min filters in samplers,
// MOBILEGL_ESPRYT_AVOID_SAMPLER_MIPMAP_MIN_FILTER: avoid mipmap min filters in samplers,
// resolves certain rendering bugs on ANGLE + llvmpipe.
Bool AvoidSamplerMipmapMinFilter = false;
// MOBILEGL_AVOID_EXPLICIT_LOD_BIAS: leave an already-explicit LOD argument alone when
Bool EsprytAvoidSamplerMipmapMinFilter = false;
// MOBILEGL_ESPRYT_AVOID_EXPLICIT_LOD_BIAS: leave an already-explicit LOD argument alone when
// emulating GL_TEXTURE_LOD_BIAS, instead of adding the bias uniform to it. Injecting
// the uniform turns a compile-time-constant LOD into a runtime expression, which
// sends ANGLE + llvmpipe down a mip-selection path that dereferences a NULL
// descriptor and kills the process. Deviates from spec (Vulkan adds the bias to
// OpImageSampleExplicitLod), so it is an avoidance for that stack only.
Bool AvoidExplicitLodBias = false;
Bool EsprytAvoidExplicitLodBias = false;
// MOBILEGL_ESPRYT_UNLOCATED_IO_BLOCKS: emit a tessellation/geometry program's
// inter-stage interface blocks WITHOUT their layout(location=) qualifier, letting ES
// match them by block name and member sequence instead. The Mali ES driver delivers
@@ -161,29 +161,29 @@ namespace MobileGL::MG_Config {
Bool CoherentAsFlush = false;
// MOBILEGL_TRACE_SKIP_AUTODESTROY: skip teardown in the ELF destructor (Init.cpp).
Bool TraceSkipAutodestroy = false;
// MOBILEGL_DISABLE_UBO_RING: force the DirectGLES global-UBO upload back to the
// MOBILEGL_ESPRYT_DISABLE_UBO_RING: force the DirectGLES global-UBO upload back to the
// per-draw glBufferSubData path instead of the persistent-mapped ring allocator
// (negative control / driver-bug escape hatch).
Bool DisableUboRing = false;
// MOBILEGL_DISABLE_UNPACK_RING: force DirectGLES texture uploads back to
Bool EsprytDisableUboRing = false;
// MOBILEGL_ESPRYT_DISABLE_UNPACK_RING: force DirectGLES texture uploads back to
// glTexSubImage from the client pointer instead of staging them through the
// persistent-mapped unpack-PBO ring (negative control / driver-bug escape
// hatch).
Bool DisableUnpackRing = false;
// MOBILEGL_DISABLE_UPLOAD_RING: force DirectGLES app buffer updates
Bool EsprytDisableUnpackRing = false;
// MOBILEGL_ESPRYT_DISABLE_UPLOAD_RING: force DirectGLES app buffer updates
// (glBufferSubData / map flushes) back to the immediate driver upload instead
// of queueing them for the staged-copy flush through the persistent-mapped
// upload ring (negative control / driver-bug escape hatch; the immediate
// upload stalls on drivers that resolve the WAR hazard on the CPU, e.g. Mali).
Bool DisableUploadRing = false;
// MOBILEGL_DISABLE_INVALIDATE_FLUSH: skip the glMapBufferRange(WRITE |
Bool EsprytDisableUploadRing = false;
// MOBILEGL_ESPRYT_DISABLE_INVALIDATE_FLUSH: skip the glMapBufferRange(WRITE |
// INVALIDATE_RANGE) tier of the DirectGLES pending-range flush and go straight
// to the upload ring's staged glCopyBufferSubData (negative control / escape
// hatch for a driver whose range-invalidating map misbehaves). The map tier is
// what keeps a partial write into a large in-flight buffer priced by the RANGE:
// on Mali both the immediate glBufferSubData and a staged copy into a busy
// mutable store ghost the whole destination on the CPU.
Bool DisableInvalidateFlush = false;
Bool EsprytDisableInvalidateFlush = false;
// MOBILEGL_ESPRYT_FORCE_DS_READBACK_EMULATION: make DirectGLES skip the native ES
// depth/stencil reads and always go through the shader-sampling emulation. Core GL
// ES has no depth or stencil readback, but some drivers accept it anyway (Mesa does,
@@ -204,10 +204,10 @@ namespace MobileGL::MG_Config {
// gl_FragDepth writers, and fully color-masked attachments are exempt (see
// PipelineFactory::ShouldSuppressDepthWrite). Auto detects Qualcomm.
QuirkOverride MagmaDisableBlendedDepthWriteQuirk = QuirkOverride::Auto;
// MOBILEGL_DISABLE_ROBUST_BUFFER_ACCESS: leave the Vulkan robustBufferAccess device
// MOBILEGL_MAGMA_DISABLE_ROBUST_BUFFER_ACCESS: leave the Vulkan robustBufferAccess device
// feature off. It is enabled by default to match GL's defined out-of-range fetch
// behavior; this escape hatch exists to measure or dodge its GPU cost on a device.
Bool DisableRobustBufferAccess = false;
Bool MagmaDisableRobustBufferAccess = false;
// MOBILEGL_MAGMA_MULTIDRAW_MODE: preferred DirectVulkan multi-draw dispatch tier
// ("ext" | "indirect" | "unroll", see MultiDrawMode). Clamped to device support;
// unset picks the best supported tier.
@@ -248,7 +248,7 @@ namespace MobileGL::MG_Config {
// miscompiled shader: if a device ever renders differently with the cache
// on, one run with this falsy says so.
QuirkOverride ShaderTranslationCache = QuirkOverride::Auto;
// MOBILEGL_FORCE_VIEWPORT_ARRAY_EMULATION: DirectGLES' gl_ViewportIndex routing
// MOBILEGL_ESPRYT_FORCE_VIEWPORT_ARRAY_EMULATION: DirectGLES' gl_ViewportIndex routing
// emulation - the builtin becomes a flat varying, the fragment stage gets a
// per-pass gate, and a routed draw is REPLAYED once per distinct viewport state
// with the real glViewport/glScissor/glDepthRangef set for it. Auto is ON, and
@@ -260,8 +260,8 @@ namespace MobileGL::MG_Config {
// the pre-emulation path, extension passthrough where it exists and
// LowerViewportIndexPass' demote-to-a-plain-global where it does not - and is
// the negative control the emulation is measured against.
QuirkOverride ViewportArrayEmulation = QuirkOverride::Auto;
// MOBILEGL_WIDEN_PACKED16_STORAGE: DirectGLES stores GL_RGB565/GL_RGB5(A1)/GL_RGBA4
QuirkOverride EsprytViewportArrayEmulation = QuirkOverride::Auto;
// MOBILEGL_ESPRYT_WIDEN_PACKED16_STORAGE: DirectGLES stores GL_RGB565/GL_RGB5(A1)/GL_RGBA4
// images as 8-bit-per-channel ES storage (GL_RGB8/GL_RGBA8) instead of the driver's
// native 16-bit packed formats. Auto defers to a POST driver-bug probe
// (SelfTest::CopyImageMirrorsPacked16FieldOrder): some Mali drivers store SOME
+18 -18
View File
@@ -162,37 +162,37 @@ namespace MobileGL::MG_ConfigLoader {
inline void InitFeatures() {
auto& features = MG_Config::Features;
features.DisableTimerQuery = QueryEnvFlag("MOBILEGL_DISABLE_TIMERQUERY");
features.EnableGlesTextureView = QueryEnvFlag("MOBILEGL_ENABLE_GLES_TEXTURE_VIEW");
features.EsprytEnableTextureView = QueryEnvFlag("MOBILEGL_ESPRYT_ENABLE_TEXTURE_VIEW");
features.EnableSpirvValidation = QueryEnvFlag("MOBILEGL_ENABLE_SPIRV_VALIDATION");
features.UseAngle = QueryEnvFlag("MOBILEGL_USE_ANGLE");
features.EsprytUseAngle = QueryEnvFlag("MOBILEGL_ESPRYT_USE_ANGLE");
#if defined(MOBILEGL_TRACE_ANGLE_VARIANTS)
QueryEnvVariable("MOBILEGL_TRACE_ANGLE_VARIANT", features.TraceAngleVariant, "");
#endif
features.DisableSubgroup = QueryEnvFlag("MOBILEGL_DISABLE_SUBGROUP");
features.MagmaDisableSubgroup = QueryEnvFlag("MOBILEGL_MAGMA_DISABLE_SUBGROUP");
features.MagmaEmulateSubgroup = QueryEnvFlag("MOBILEGL_MAGMA_EMULATE_SUBGROUP");
features.FixIterationRPSubgroupScratch =
QueryEnvQuirkOverride("MOBILEGL_FIX_ITERATIONRP_SUBGROUP_SCRATCH");
features.IterationRPFixBarrier = QueryEnvFlag("MOBILEGL_ITERATIONRP_FIX_BARRIER");
features.DeriveNumSubgroups = QueryEnvQuirkOverride("MOBILEGL_DERIVE_NUM_SUBGROUPS");
features.MagmaFixIterationRPSubgroupScratch =
QueryEnvQuirkOverride("MOBILEGL_MAGMA_FIX_ITERATIONRP_SUBGROUP_SCRATCH");
features.MagmaIterationRPFixBarrier = QueryEnvFlag("MOBILEGL_MAGMA_ITERATIONRP_FIX_BARRIER");
features.MagmaDeriveNumSubgroups = QueryEnvQuirkOverride("MOBILEGL_MAGMA_DERIVE_NUM_SUBGROUPS");
features.AdvertiseFp64 = QueryEnvFlag("MOBILEGL_ADVERTISE_FP64");
features.MagmaR11G11B10FFallback = QueryEnvFlag("MOBILEGL_MAGMA_R11G11B10F_FALLBACK");
features.MagmaFramesInFlight = QueryEnvUint32("MOBILEGL_MAGMA_FRAMESINFLIGHT", 3, 1, 64);
features.AvoidSamplerMipmapMinFilter =
QueryEnvFlag("MOBILEGL_AVOID_SAMPLER_MIPMAP_MIN_FILTER");
features.AvoidExplicitLodBias = QueryEnvFlag("MOBILEGL_AVOID_EXPLICIT_LOD_BIAS");
features.EsprytAvoidSamplerMipmapMinFilter =
QueryEnvFlag("MOBILEGL_ESPRYT_AVOID_SAMPLER_MIPMAP_MIN_FILTER");
features.EsprytAvoidExplicitLodBias = QueryEnvFlag("MOBILEGL_ESPRYT_AVOID_EXPLICIT_LOD_BIAS");
features.EsprytUnlocatedIoBlocks = QueryEnvQuirkOverride("MOBILEGL_ESPRYT_UNLOCATED_IO_BLOCKS");
features.CoherentAsFlush = QueryEnvFlag("MOBILEGL_COHERENT_AS_FLUSH");
features.TraceSkipAutodestroy = QueryEnvFlag("MOBILEGL_TRACE_SKIP_AUTODESTROY");
features.DisableUboRing = QueryEnvFlag("MOBILEGL_DISABLE_UBO_RING");
features.DisableUnpackRing = QueryEnvFlag("MOBILEGL_DISABLE_UNPACK_RING");
features.DisableUploadRing = QueryEnvFlag("MOBILEGL_DISABLE_UPLOAD_RING");
features.DisableInvalidateFlush = QueryEnvFlag("MOBILEGL_DISABLE_INVALIDATE_FLUSH");
features.EsprytDisableUboRing = QueryEnvFlag("MOBILEGL_ESPRYT_DISABLE_UBO_RING");
features.EsprytDisableUnpackRing = QueryEnvFlag("MOBILEGL_ESPRYT_DISABLE_UNPACK_RING");
features.EsprytDisableUploadRing = QueryEnvFlag("MOBILEGL_ESPRYT_DISABLE_UPLOAD_RING");
features.EsprytDisableInvalidateFlush = QueryEnvFlag("MOBILEGL_ESPRYT_DISABLE_INVALIDATE_FLUSH");
features.EsprytForceDepthStencilReadbackEmulation =
QueryEnvFlag("MOBILEGL_ESPRYT_FORCE_DS_READBACK_EMULATION");
features.RelaxedSemantics = QueryEnvFlag("MOBILEGL_RELAXED_SEMANTICS");
features.MagmaDisableBlendedDepthWriteQuirk =
QueryEnvQuirkOverride("MOBILEGL_MAGMA_DISABLE_BLENDED_DEPTH_WRITE");
features.DisableRobustBufferAccess = QueryEnvFlag("MOBILEGL_DISABLE_ROBUST_BUFFER_ACCESS");
features.MagmaDisableRobustBufferAccess = QueryEnvFlag("MOBILEGL_MAGMA_DISABLE_ROBUST_BUFFER_ACCESS");
features.MagmaMultiDrawMode = QueryEnvMultiDrawMode("MOBILEGL_MAGMA_MULTIDRAW_MODE");
features.EsprytMultiDrawMode = QueryEnvGLESMultiDrawMode("MOBILEGL_ESPRYT_MULTIDRAW_MODE");
features.AsyncShaderCompile = QueryEnvQuirkOverride("MOBILEGL_ASYNC_SHADER_COMPILE");
@@ -200,10 +200,10 @@ namespace MobileGL::MG_ConfigLoader {
features.AsyncOptimisticShaderStatus =
QueryEnvQuirkOverride("MOBILEGL_ASYNC_OPTIMISTIC_SHADER_STATUS");
features.ShaderTranslationCache = QueryEnvQuirkOverride("MOBILEGL_SHADER_CACHE");
features.ViewportArrayEmulation =
QueryEnvQuirkOverride("MOBILEGL_FORCE_VIEWPORT_ARRAY_EMULATION");
features.EsprytViewportArrayEmulation =
QueryEnvQuirkOverride("MOBILEGL_ESPRYT_FORCE_VIEWPORT_ARRAY_EMULATION");
features.EsprytWidenPacked16Storage =
QueryEnvQuirkOverride("MOBILEGL_WIDEN_PACKED16_STORAGE");
QueryEnvQuirkOverride("MOBILEGL_ESPRYT_WIDEN_PACKED16_STORAGE");
}
inline void InitBackendType() {
@@ -1198,8 +1198,8 @@ namespace MobileGL::MG_Backend::DirectGLES {
//
// Until that reconciliation exists, advertising here would be the same lie the comment
// above refuses to tell, just with an extra prerequisite met. Set
// MOBILEGL_ENABLE_GLES_TEXTURE_VIEW=1 to re-enable it for that work.
if (textureViewSupported && MG_Config::Features.EnableGlesTextureView) {
// MOBILEGL_ESPRYT_ENABLE_TEXTURE_VIEW=1 to re-enable it for that work.
if (textureViewSupported && MG_Config::Features.EsprytEnableTextureView) {
extensions.push_back(E_GL_ARB_texture_view);
}
// Only advertised when the host ES driver actually filters anisotropically: the sampler
+11 -11
View File
@@ -54,7 +54,7 @@ namespace MobileGL::MG_Backend::DirectGLES {
// the driver fifteen more rectangles to rasterize against, and nothing in MobileGL has ever
// programmed the indexed state it would need.
Bool ViewportArrayEmulationEnabled() {
return MG_Config::Features.ViewportArrayEmulation != MG_Config::QuirkOverride::ForceOff;
return MG_Config::Features.EsprytViewportArrayEmulation != MG_Config::QuirkOverride::ForceOff;
}
Bool g_anyProgramRoutesViewportIndex = false;
@@ -75,13 +75,13 @@ namespace MobileGL::MG_Backend::DirectGLES {
static Bool ShouldAvoidSamplerMipmapMinFilterOnAngleLlvmpipe() {
// IsAngleLlvmpipeRenderer combined with the
// MOBILEGL_AVOID_SAMPLER_MIPMAP_MIN_FILTER feature toggle,
// MOBILEGL_ESPRYT_AVOID_SAMPLER_MIPMAP_MIN_FILTER feature toggle,
// both resolved in FillInGLESCapabilities.
return g_GLESCapabilities.AvoidSamplerMipmapMinFilter;
}
static Bool ShouldAvoidExplicitLodBiasOnAngleLlvmpipe() {
// IsAngleLlvmpipeRenderer combined with the MOBILEGL_AVOID_EXPLICIT_LOD_BIAS
// IsAngleLlvmpipeRenderer combined with the MOBILEGL_ESPRYT_AVOID_EXPLICIT_LOD_BIAS
// feature toggle, both resolved in FillInGLESCapabilities.
return g_GLESCapabilities.AvoidExplicitLodBias;
}
@@ -823,7 +823,7 @@ namespace MobileGL::MG_Backend::DirectGLES {
// now: kill switch off, the ES copy entry point resolved, and the ring's
// own availability gate (EXT_buffer_storage + fences + live context) up.
Bool UploadRingUsableNow() {
if (MG_Config::Features.DisableUploadRing) return false;
if (MG_Config::Features.EsprytDisableUploadRing) return false;
if (!g_GLESFuncs.glCopyBufferSubData) return false;
return RingAvailable(g_uploadRing);
}
@@ -846,7 +846,7 @@ namespace MobileGL::MG_Backend::DirectGLES {
// stores also kills the ghost, but eagerly commits every arena's
// full extent - +hundreds of MB - which LMK'd the whole device.)
// 2. The staging ring + glCopyBufferSubData: the copy is ordered on
// the GPU timeline, no CPU wait (MOBILEGL_DISABLE_INVALIDATE_FLUSH
// the GPU timeline, no CPU wait (MOBILEGL_ESPRYT_DISABLE_INVALIDATE_FLUSH
// forces this tier as the map path's negative control).
// 3. Direct glBufferSubData (potentially stalling) when neither the
// map entry points nor the ring exist.
@@ -871,7 +871,7 @@ namespace MobileGL::MG_Backend::DirectGLES {
// frontend size and the backend store disagree (a pending respecify
// resolves that later; bytes past either end have nowhere to land).
const SizeT limit = std::min(bufferObject.GetSize(), resource.storageSize);
const Bool mapUsable = !MG_Config::Features.DisableInvalidateFlush &&
const Bool mapUsable = !MG_Config::Features.EsprytDisableInvalidateFlush &&
g_GLESFuncs.glMapBufferRange && g_GLESFuncs.glUnmapBuffer;
const Bool ringUsable = UploadRingUsableNow();
for (const auto& range : ranges) {
@@ -1065,7 +1065,7 @@ namespace MobileGL::MG_Backend::DirectGLES {
// keeps the legacy immediate upload: draw-time sync never flushes
// ranges for it, and its mapping publishes writes by itself.
if ((resource->persistentMapped && resource->persistentPtr) ||
MG_Config::Features.DisableUploadRing) {
MG_Config::Features.EsprytDisableUploadRing) {
UploadRangeNow(*resource, bufferObject, offset, offset + size);
resource->syncedChangeSerial = bufferObject.GetChangeSerial();
return;
@@ -1093,7 +1093,7 @@ namespace MobileGL::MG_Backend::DirectGLES {
// flush at draw-time sync; only the zero-copy persistent store and the
// negative-control kill switch keep the immediate paths below.
if (!(resource->persistentMapped && resource->persistentPtr) &&
!MG_Config::Features.DisableUploadRing) {
!MG_Config::Features.EsprytDisableUploadRing) {
const std::lock_guard<std::mutex> lock(resource->pendingMutex);
resource->pendingRanges.Add(range);
return;
@@ -1901,7 +1901,7 @@ namespace MobileGL::MG_Backend::DirectGLES {
} // namespace
Bool UboRingAvailable() {
if (MG_Config::Features.DisableUboRing) return false;
if (MG_Config::Features.EsprytDisableUboRing) return false;
return RingAvailable(g_uboRing);
}
@@ -1914,7 +1914,7 @@ namespace MobileGL::MG_Backend::DirectGLES {
void UboRingOnPresent() { RingOnPresent(g_uboRing); }
Bool UnpackRingAvailable() {
if (MG_Config::Features.DisableUnpackRing) return false;
if (MG_Config::Features.EsprytDisableUnpackRing) return false;
return RingAvailable(g_unpackRing);
}
@@ -6489,7 +6489,7 @@ namespace MobileGL::MG_Backend::DirectGLES {
const Int advertisedMaxSamples =
std::max(g_GLESCapabilities.MaxSamples, kFrontendMaxSamples);
// Armed by the EMULATION as well as by the missing extension, and the emulation is on
// by default (MOBILEGL_FORCE_VIEWPORT_ARRAY_EMULATION). Having the extension is not a
// by default (MOBILEGL_ESPRYT_FORCE_VIEWPORT_ARRAY_EMULATION). Having the extension is not a
// reason to keep the builtin: it only ever gave the SHADER a compilable name, while
// the driver's INDEXED viewport state was never programmed by anything in MobileGL
// (SyncRenderState pushes index 0 and stops), so an extension-capable driver
+3 -3
View File
@@ -230,7 +230,7 @@ namespace MobileGL::MG_Backend::DirectGLES {
// still holding what glViewport/glScissor/glDepthRange broadcast to all sixteen - collapses
// to a single pass with an all-ones gate mask, i.e. one draw and no behaviour change at all.
//
// Whether emulation runs. Off only under MOBILEGL_FORCE_VIEWPORT_ARRAY_EMULATION falsy, which
// Whether emulation runs. Off only under MOBILEGL_ESPRYT_FORCE_VIEWPORT_ARRAY_EMULATION falsy, which
// restores the pre-emulation path as a negative control.
Bool ViewportArrayEmulationEnabled();
// Whether ANY program built in this process has come out with a viewport gate. Sticky once
@@ -627,7 +627,7 @@ namespace MobileGL::MG_Backend::DirectGLES {
// which is what to watch if this ring ever shows up in an RSS regression: it
// grows on demand from 4 MiB and is capped, not unbounded.
//
// False when the feature is disabled (MOBILEGL_DISABLE_UNPACK_RING),
// False when the feature is disabled (MOBILEGL_ESPRYT_DISABLE_UNPACK_RING),
// EXT_buffer_storage / fences are missing, the ES context is not current, or
// ring creation already failed under this context. Callers then upload from
// the client pointer exactly as before.
@@ -656,7 +656,7 @@ namespace MobileGL::MG_Backend::DirectGLES {
// draw-time sync drains them: bytes staged into this ring, then one
// glCopyBufferSubData per merged range - the copy is ordered on the GPU
// timeline, so the hazard costs no CPU wait. Reclamation contract identical
// to the other two rings. MOBILEGL_DISABLE_UPLOAD_RING restores the
// to the other two rings. MOBILEGL_ESPRYT_DISABLE_UPLOAD_RING restores the
// historical immediate-upload path (negative control / escape hatch).
void UploadRingOnPresent();
} // namespace BufferImpl
+2 -2
View File
@@ -50,7 +50,7 @@ namespace MobileGL::MG_Backend::DirectGLES {
// driver stores some packed16 allocations with a mirrored field order
// (PixelFormatNormalizeOptionBit::WidenPacked16Norm). True only for
// GL_RGB565/GL_RGB5(_A1)/GL_RGBA4, and only where the POST probe measured the
// divergence (or MOBILEGL_WIDEN_PACKED16_STORAGE forces it). The transfer paths
// divergence (or MOBILEGL_ESPRYT_WIDEN_PACKED16_STORAGE forces it). The transfer paths
// consult it too: the packed-norm re-upload leg must stand down when the ES storage
// is no longer 16-bit packed.
Bool UsesWidenedPacked16NormStorage(TextureInternalFormat internalFormat);
@@ -532,7 +532,7 @@ namespace MobileGL::MG_Backend::DirectGLES {
// avoidExplicitLodBias leaves lookups that already carry an explicit LOD untouched,
// so their constant level stays constant; only the implicit-LOD forms take the bias.
// Off by default and only ever set on ANGLE + llvmpipe, where injecting the uniform
// into a constant LOD crashes the driver (MOBILEGL_AVOID_EXPLICIT_LOD_BIAS).
// into a constant LOD crashes the driver (MOBILEGL_ESPRYT_AVOID_EXPLICIT_LOD_BIAS).
String EmulateTextureLodBias(const String& glslCode, Bool avoidExplicitLodBias = false);
} // namespace PrgramImpl
@@ -624,7 +624,7 @@ namespace MobileGL::MG_Backend::DirectVulkan {
if (nonZeroIndirectBaseInstanceSupported) {
extensions.push_back(E_GL_ARB_base_instance);
}
if (shaderSubgroupSupported && !MG_Config::Features.DisableSubgroup) {
if (shaderSubgroupSupported && !MG_Config::Features.MagmaDisableSubgroup) {
extensions.push_back(E_GL_KHR_shader_subgroup);
}
// GL_KHR_parallel_shader_compile is MobileGL's own capability, not the Vulkan
@@ -70,7 +70,7 @@ namespace MobileGL::MG_Backend::DirectVulkan {
const RendererInfo& GetRendererIdentity();
// The full OpenGL extension list Magma advertises (glGetString(GL_EXTENSIONS)) for
// a device with the given raw capabilities. The MOBILEGL_DISABLE_SUBGROUP and
// a device with the given raw capabilities. The MOBILEGL_MAGMA_DISABLE_SUBGROUP and
// MOBILEGL_DISABLE_TIMERQUERY escape hatches are applied inside, so callers pass
// the detected device support (passing an already-gated value is harmless).
Vector<GLExtension> BuildAdvertisedExtensions(Bool shaderSubgroupSupported, Bool timerQueriesSupported,
@@ -13214,8 +13214,8 @@ void main() {
// Match GL's robust buffer-fetch behavior where the Vulkan device supports it. This covers
// out-of-range fetches; arbitrary GL vertex strides/offsets still need the explicit tight
// repack in VertexInputStateFactory when they violate Vulkan's address-alignment rules.
// MOBILEGL_DISABLE_ROBUST_BUFFER_ACCESS leaves it off to measure or dodge its GPU cost.
deviceFeatures.robustBufferAccess = MG_Config::Features.DisableRobustBufferAccess
// MOBILEGL_MAGMA_DISABLE_ROBUST_BUFFER_ACCESS leaves it off to measure or dodge its GPU cost.
deviceFeatures.robustBufferAccess = MG_Config::Features.MagmaDisableRobustBufferAccess
? VK_FALSE
: supportedDeviceFeatures.robustBufferAccess;
deviceFeatures.geometryShader = supportedDeviceFeatures.geometryShader;
@@ -13590,13 +13590,13 @@ void main() {
subgroupPropertyQuery.pNext = &subgroupProperties;
getPhysicalDeviceProperties2(m_physicalDevice.handle, &subgroupPropertyQuery);
// Mirrors the loader's HasUsableShaderSubgroupSupport gate, including the
// MOBILEGL_DISABLE_SUBGROUP escape hatch, so the module lowerings can never
// MOBILEGL_MAGMA_DISABLE_SUBGROUP escape hatch, so the module lowerings can never
// disagree with the advertised capabilities.
const Bool usableSubgroups =
subgroupProperties.subgroupSize > 0 &&
(subgroupProperties.supportedStages & VK_SHADER_STAGE_COMPUTE_BIT) != 0 &&
(subgroupProperties.supportedOperations & VK_SUBGROUP_FEATURE_BASIC_BIT) != 0;
if (usableSubgroups && !MG_Config::Features.DisableSubgroup) {
if (usableSubgroups && !MG_Config::Features.MagmaDisableSubgroup) {
m_nativeSubgroupSize = subgroupProperties.subgroupSize;
m_nativeSubgroupSupported = true;
}
@@ -563,7 +563,7 @@ namespace MobileGL::MG_Backend::DirectVulkan {
// Native subgroup topology, queried at device creation for the compute-module
// subgroup repairs (SubgroupSupportPolicy.h) and the REQUIRE_FULL_SUBGROUPS
// stage flag; 0 / false when the device has no usable compute subgroups or
// MOBILEGL_DISABLE_SUBGROUP forced them off.
// MOBILEGL_MAGMA_DISABLE_SUBGROUP forced them off.
Uint32 m_nativeSubgroupSize = 0;
Bool m_nativeSubgroupSupported = false;
Bool m_computeFullSubgroupsFeatureEnabled = false;
@@ -39,18 +39,18 @@ namespace MobileGL::MG_Backend::DirectVulkan {
inline Bool ShouldEmulateSubgroups(const Bool nativeSubgroupSupported) {
return MG_Config::Features.MagmaEmulateSubgroup && !nativeSubgroupSupported &&
!MG_Config::Features.DisableSubgroup;
!MG_Config::Features.MagmaDisableSubgroup;
}
inline Bool ShouldFixIterationRPSubgroupScratch() {
// Auto is ON: the patch is fingerprint-gated to iterationRP's reduction and
// grows one under-declared array; every other module passes through untouched.
return MG_Config::Features.FixIterationRPSubgroupScratch !=
return MG_Config::Features.MagmaFixIterationRPSubgroupScratch !=
MG_Config::QuirkOverride::ForceOff;
}
inline Bool ShouldFixIterationRPBarrier() {
return MG_Config::Features.IterationRPFixBarrier;
return MG_Config::Features.MagmaIterationRPFixBarrier;
}
inline Bool ShouldDeriveNumSubgroups() {
@@ -58,6 +58,6 @@ namespace MobileGL::MG_Backend::DirectVulkan {
// contract to hold, and the derived ceil() value is the one the renderer can pin
// with REQUIRE_FULL_SUBGROUPS - the driver builtin is the value with no
// cross-driver guarantee (Adreno returns 1 for an 8-subgroup dispatch).
return MG_Config::Features.DeriveNumSubgroups != MG_Config::QuirkOverride::ForceOff;
return MG_Config::Features.MagmaDeriveNumSubgroups != MG_Config::QuirkOverride::ForceOff;
}
} // namespace MobileGL::MG_Backend::DirectVulkan
+5 -5
View File
@@ -288,9 +288,9 @@ if (MOBILEGL_ITEST_VK_ICD)
if (MOBILEGL_ITEST_VK_ICD MATCHES "lvp_icd|lavapipe")
message(STATUS "Integration tests: lavapipe ICD - forcing the iterationRP repairs on")
list(APPEND MGL_ITEST_VULKAN_ENV
"MOBILEGL_FIX_ITERATIONRP_SUBGROUP_SCRATCH=1"
"MOBILEGL_DERIVE_NUM_SUBGROUPS=1"
"MOBILEGL_ITERATIONRP_FIX_BARRIER=1")
"MOBILEGL_MAGMA_FIX_ITERATIONRP_SUBGROUP_SCRATCH=1"
"MOBILEGL_MAGMA_DERIVE_NUM_SUBGROUPS=1"
"MOBILEGL_MAGMA_ITERATIONRP_FIX_BARRIER=1")
endif()
endif()
@@ -353,7 +353,7 @@ mgl_itest_join_environment(MGL_ITEST_VULKAN_OPTIMISTIC_ENVIRONMENT
"MOBILEGL_BACKEND_TYPE=DirectVulkan" "MOBILEGL_ASYNC_SHADER_COMPILE=1"
"MOBILEGL_ASYNC_OPTIMISTIC_SHADER_STATUS=1" ${MGL_ITEST_VULKAN_ENV})
mgl_itest_join_environment(MGL_ITEST_GLES_NO_VIEWPORT_EMULATION_ENVIRONMENT
"MOBILEGL_BACKEND_TYPE=DirectGLES" "MOBILEGL_FORCE_VIEWPORT_ARRAY_EMULATION=0" ${MGL_ITEST_COMMON_ENV})
"MOBILEGL_BACKEND_TYPE=DirectGLES" "MOBILEGL_ESPRYT_FORCE_VIEWPORT_ARRAY_EMULATION=0" ${MGL_ITEST_COMMON_ENV})
# MOBILEGL_LOG_FILE_PATH alongside the pin, because the arming assertion needs somewhere to
# read the library's own report from. The strip's arming signal is a latched MGLOG_I and there
# is no other way for a test process to learn that it fired - MG_Config is not reachable from
@@ -364,7 +364,7 @@ mgl_itest_join_environment(MGL_ITEST_GLES_UNLOCATED_IO_BLOCKS_ENVIRONMENT
"MOBILEGL_LOG_FILE_PATH=${CMAKE_CURRENT_BINARY_DIR}/unlocated-io-blocks.log"
${MGL_ITEST_COMMON_ENV})
mgl_itest_join_environment(MGL_ITEST_GLES_WIDENED_PACKED16_ENVIRONMENT
"MOBILEGL_BACKEND_TYPE=DirectGLES" "MOBILEGL_WIDEN_PACKED16_STORAGE=1" ${MGL_ITEST_COMMON_ENV})
"MOBILEGL_BACKEND_TYPE=DirectGLES" "MOBILEGL_ESPRYT_WIDEN_PACKED16_STORAGE=1" ${MGL_ITEST_COMMON_ENV})
# TIMEOUT on every entry: a GPU test that wedges must fail the run, not hang it.
set(MGL_ITEST_TIMEOUT 120)
@@ -31,7 +31,7 @@
// The repair under test is the packed16 storage widening
// (PixelFormatNormalizeOptionBit::WidenPacked16Norm): where the POST probe
// (SelfTest::CopyImageMirrorsPacked16FieldOrder) measures the mirror - or
// MOBILEGL_WIDEN_PACKED16_STORAGE forces it - the three formats are stored as
// MOBILEGL_ESPRYT_WIDEN_PACKED16_STORAGE forces it - the three formats are stored as
// GL_RGB8/GL_RGBA8, leaving no 16-bit packed image for a copy to disagree about. The client
// word still round-trips exactly: the canonical shadow is already UNorm8, and an n-bit field
// encodes to UNorm8 and back losslessly for every n <= 8.
@@ -40,7 +40,7 @@
// * the ambient registrations take the narrow path on a clean driver (llvmpipe has no
// mirror, so Auto keeps the native 16-bit storage - the pre-existing behaviour stays
// covered);
// * the DirectGLES.WidenedPacked16. registration pins MOBILEGL_WIDEN_PACKED16_STORAGE=1,
// * the DirectGLES.WidenedPacked16. registration pins MOBILEGL_ESPRYT_WIDEN_PACKED16_STORAGE=1,
// which is the storage every affected device will actually run - without it the repair
// is unfalsifiable off-device, because no CI driver has the bug that arms it.
// The Mali mirror itself CANNOT be reproduced here; only the on-device CTS run can show the
@@ -525,7 +525,7 @@ void main() { fragColor = vec4(float(gsIndex) * 16.0 / 255.0, 0.0, 0.0, 1.0); }
//
// Everything above is a claim about pixels, and a claim about pixels cannot tell an
// emulation that works from a backend that was going to be right anyway. This case builds
// the SAME program in a process started with MOBILEGL_FORCE_VIEWPORT_ARRAY_EMULATION=0
// the SAME program in a process started with MOBILEGL_ESPRYT_FORCE_VIEWPORT_ARRAY_EMULATION=0
// (the NoViewportArrayEmulation. ctest entry) and requires case 1's
// result to COLLAPSE: with no routing, every geometry invocation rasterizes against
// viewport 0's rectangle, so the last invocation paints the whole surface and every cell
@@ -551,10 +551,10 @@ void main() { fragColor = vec4(float(gsIndex) * 16.0 / 255.0, 0.0, 0.0, 1.0); }
// entry for it, so the control still runs in every ctest run; anywhere else - the
// ambient ctest entries, or the binary run straight from a device shell - the
// emulation is on and this case skips.
if (AmbientQuirkFromEnvironment("MOBILEGL_FORCE_VIEWPORT_ARRAY_EMULATION") != AmbientQuirk::Off) {
if (AmbientQuirkFromEnvironment("MOBILEGL_ESPRYT_FORCE_VIEWPORT_ARRAY_EMULATION") != AmbientQuirk::Off) {
GTEST_SKIP() << "this is the negative control for the emulation and needs it off for the "
"whole process; the NoViewportArrayEmulation. ctest entry runs it with "
"MOBILEGL_FORCE_VIEWPORT_ARRAY_EMULATION=0";
"MOBILEGL_ESPRYT_FORCE_VIEWPORT_ARRAY_EMULATION=0";
}
IntTarget target = MakeIntTarget(kSurfaceSide, kSurfaceSide);
@@ -22,7 +22,7 @@
namespace MobileGL::MG_Util::BackendLoader {
static Bool UseAngle() {
return MG_Config::Features.UseAngle;
return MG_Config::Features.EsprytUseAngle;
}
#if defined(MOBILEGL_TRACE_ANGLE_VARIANTS) && defined(__ANDROID__)
@@ -1762,9 +1762,9 @@ namespace MobileGL::MG_Util::BackendLoader {
caps.IsAngleLlvmpipeRenderer =
caps.IsAngleRenderer && caps.GLESRendererString.find("llvmpipe") != String::npos;
caps.AvoidSamplerMipmapMinFilter =
caps.IsAngleLlvmpipeRenderer && MG_Config::Features.AvoidSamplerMipmapMinFilter;
caps.IsAngleLlvmpipeRenderer && MG_Config::Features.EsprytAvoidSamplerMipmapMinFilter;
caps.AvoidExplicitLodBias =
caps.IsAngleLlvmpipeRenderer && MG_Config::Features.AvoidExplicitLodBias;
caps.IsAngleLlvmpipeRenderer && MG_Config::Features.EsprytAvoidExplicitLodBias;
MGLOG_I(" GL_EXT_disjoint_timer_query supported: %s",
caps.SupportsDisjointTimerQuery ? "true" : "false");
MGLOG_I(" GL_KHR_parallel_shader_compile supported: %s",
@@ -1235,10 +1235,10 @@ namespace MobileGL {
// GL_RENDERER contains both "ANGLE" and "llvmpipe".
Bool IsAngleLlvmpipeRenderer = false;
// IsAngleLlvmpipeRenderer combined with the
// MOBILEGL_AVOID_SAMPLER_MIPMAP_MIN_FILTER feature toggle:
// MOBILEGL_ESPRYT_AVOID_SAMPLER_MIPMAP_MIN_FILTER feature toggle:
// sampler min filters should drop their mipmap component.
Bool AvoidSamplerMipmapMinFilter = false;
// IsAngleLlvmpipeRenderer combined with the MOBILEGL_AVOID_EXPLICIT_LOD_BIAS
// IsAngleLlvmpipeRenderer combined with the MOBILEGL_ESPRYT_AVOID_EXPLICIT_LOD_BIAS
// feature toggle: LOD-bias emulation should not touch explicit-LOD lookups.
Bool AvoidExplicitLodBias = false;
// True when indirect draws leak the command's baseInstance word ("reserved,
@@ -92,7 +92,7 @@ namespace MobileGL::MG_Util::BackendLoader {
}
Bool IsShaderSubgroupForcedDisabled() {
return MG_Config::Features.DisableSubgroup;
return MG_Config::Features.MagmaDisableSubgroup;
}
} // namespace
@@ -272,7 +272,7 @@ namespace MobileGL::MG_Util::BackendLoader {
supportsShaderSubgroup ? "true" : "false", caps.SupportsShaderSubgroup ? "true" : "false",
subgroupProps.subgroupSize, subgroupProps.supportedStages, subgroupProps.supportedOperations);
if (supportsShaderSubgroup && forceDisableShaderSubgroup) {
MGLOG_W("Vulkan shader subgroup support forced off by MOBILEGL_DISABLE_SUBGROUP");
MGLOG_W("Vulkan shader subgroup support forced off by MOBILEGL_MAGMA_DISABLE_SUBGROUP");
}
return true;
@@ -2407,7 +2407,7 @@ namespace MobileGL::MG_Util::SelfTest {
Optional<DriverBugFinding> ProbeCopyImagePacked16FieldOrderBug(const GLESFunctionsTable& gl) {
if (!CopyImageMirrorsPacked16FieldOrder(gl)) return std::nullopt;
// The mitigation is a knob (MOBILEGL_WIDEN_PACKED16_STORAGE), so the row consults
// The mitigation is a knob (MOBILEGL_ESPRYT_WIDEN_PACKED16_STORAGE), so the row consults
// it: under ForceOff - the documented negative control - the corruption is
// replayed verbatim, and a hardcoded "Fixed" would be exactly the kind of
// reassurance this file exists to refuse. Auto and ForceOn both widen once this
@@ -2433,12 +2433,12 @@ namespace MobileGL::MG_Util::SelfTest {
"already holds and the client word round-trips through exactly), so no "
"16-bit packed image is left for a copy to disagree about, at twice the "
"memory for images of those formats; override with "
"MOBILEGL_WIDEN_PACKED16_STORAGE";
"MOBILEGL_ESPRYT_WIDEN_PACKED16_STORAGE";
return DriverBugFinding{
"glCopyImageSubData mirrors 16-bit packed texels between differently-laid-out images",
DriverBugVerdict::Fixed, detail};
}
detail += "MOBILEGL_WIDEN_PACKED16_STORAGE=0 keeps the native narrow storage, so such "
detail += "MOBILEGL_ESPRYT_WIDEN_PACKED16_STORAGE=0 keeps the native narrow storage, so such "
"copies are left exactly as the driver delivers them, mirrored words included";
return DriverBugFinding{
"glCopyImageSubData mirrors 16-bit packed texels between differently-laid-out images",
+1 -1
View File
@@ -1557,7 +1557,7 @@ namespace MobileGL::MG_Util::SelfTest {
// Native iterationRP compute witness. This deliberately uses a separate
// throwaway Vulkan device rather than the real renderer's queues, and it
// treats MOBILEGL_DISABLE_SUBGROUP as irrelevant: the row reports what the
// treats MOBILEGL_MAGMA_DISABLE_SUBGROUP as irrelevant: the row reports what the
// driver does, not what MobileGL elects to advertise to applications.
void ProbeVulkanIterationRPWitness(ReportBuilder& builder, PFN_vkGetInstanceProcAddr getInstanceProcAddr,
VkInstance instance, VkPhysicalDevice physicalDevice,