swung0x48/MobileGL
1
0
Fork 0
mirror of https://github.com/MobileGL-Dev/MobileGL synced 2026-09-08 20:28:32 +09:00
Code Issues Packages Projects Releases Wiki Activity

[Test] (Espryt): run the sanity binary on the twin arm it was compiled for, and pin that it does

Browse Source 
- SanityTest never calls MG_ConfigLoader::Init, so MG_Config::Features.PipePush kept its
  static default of 0 - the value a PULL build ships - and every case in the binary took
  StateBackendObjectRegistry's legacy UnorderedMap arm. In build-linux, build-push AND
  build-verify alike, the whole D13 "must not break" list (the scratch-FBO scrub, the three
  context-generation guards on the texture/framebuffer/renderbuffer twins, the sampled-set
  staleness walk and the whole-registry ScopedDirectGLESTextureBindings fixture) was
  therefore evidence about code this package did not change; a gdb breakpoint on
  MGPipeSlotAllocator::Acquire was the only way to see it.
- A gtest global Environment now seeds Features.PipePush with ConfigLoader's own push-build
  default, so the binary runs the arm that SHIPS in the build it was compiled for: legacy in
  build-linux (where the handle arm is not compiled and every DirectGLESSlotTable case skips
  visibly), handles in build-push and build-verify. MOBILEGL_PIPE_PUSH in the environment
  overrides it with ConfigLoader's decimal/0x contract, so the legacy-arm run of the same
  binary is one env var. It is an Environment and not a static initializer because
  MG_Config::Features has a String member and is dynamically initialised.
- TheTwinRegistryCasesInThisBinaryRunOnTheHandleArm is that gdb probe made falsifiable:
  delete the environment and it goes red naming the arm rather than a symptom.
This commit is contained in:
swung0x48
2026-09-07 23:18:09 -04:00
parent 5a3c0616b7
commit e10f5d6750
1 changed files with 69 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
MobileGL/MG_Test/SanityTest.cpp
+69
View File
@@ -37,9 +37,54 @@
#include <MG_Util/ShaderTranspiler/ShaderSourceProcessor.h>
#include <MG_Util/Debug/Log.h>
#include <MG_Util/Types.h>
#include <Config.h>
#include <MG_Pipe/MGPipe.h>
#include <limits>
#include <set>
#if MOBILEGL_PIPE_PUSH
namespace {
// Which arm of the twin table this binary runs on.
//
// SanityTest never calls MG_ConfigLoader::Init, so MG_Config::Features keeps its static
// defaults - and Features.PipePush's static default is 0 ("pull everything", Config.h),
// which is the value a PULL build ships. Without this the D13 "must not break" cases
// (the scratch-FBO scrub, the three context-generation guards on the texture /
// framebuffer / renderbuffer twins, the sampled-set staleness walk and the whole-registry
// ScopedDirectGLESTextureBindings fixture) exercised the legacy UnorderedMap arm in EVERY
// build directory - so a push build's 82 sanity cases said nothing about the code this
// package actually changed.
//
// The default here is therefore ConfigLoader's own push-build default
// (kMGPipeSubsystemsMigratedAtP2, ConfigLoader.cpp), i.e. this binary runs the arm that
// SHIPS in the build it was compiled for: legacy in build-linux (where the handle arm is
// not compiled at all and every DirectGLESSlotTable case skips), handles in build-push and
// build-verify. MOBILEGL_PIPE_PUSH in the environment overrides it with the same
// decimal/0x contract ConfigLoader.cpp:172-192 gives it, so `MOBILEGL_PIPE_PUSH=0
// ctest -R Sanity` is the legacy-arm run of the same binary and the A/B is one env var.
//
// It runs as a gtest Environment rather than a static initializer on purpose:
// MG_Config::Features has a String member, so it is dynamically initialised, and writing
// to it from another TU's static initializer would be an initialisation-order race.
// SetUp() runs inside RUN_ALL_TESTS, long after every static initializer, and before the
// first test - hence before anything can latch EsprytSlotTablesEnabled().
class EsprytSlotArmEnvironment final : public ::testing::Environment {
public:
void SetUp() override {
MobileGL::Uint64 bits = MobileGL::MG_Pipe::kMGPipeSubsystemsMigratedAtP2;
const char* knob = std::getenv("MOBILEGL_PIPE_PUSH");
if (knob != nullptr && *knob != '\0') {
bits = std::strtoull(knob, nullptr, 0);
}
MobileGL::MG_Config::Features.PipePush = bits;
}
};
const ::testing::Environment* g_esprytSlotArmEnvironment =
::testing::AddGlobalTestEnvironment(new EsprytSlotArmEnvironment());
} // namespace
#endif // MOBILEGL_PIPE_PUSH
namespace {
class DynamicParameterBackend final : public MobileGL::MG_Backend::BackendObject {
public:
@@ -3351,6 +3396,26 @@ TEST(DirectGLESSlotTable, GetOrCreateToleratesANullStateObject) {
EXPECT_EQ(table.Find(nullptr), nullptr);
EXPECT_TRUE(MG_Pipe::MGPipeHandleIsNull(table.HandleOf(nullptr)));
}
// The gate on MAJOR 1 of the round-2 review: this binary's OTHER 82 cases - among them every
// D13 "must not break" item - are only evidence about this package if they run on the arm this
// package wrote. Before EsprytSlotArmEnvironment existed they did not, in any build directory
// the P2 brief defines, and nothing said so; a gdb breakpoint on MGPipeSlotAllocator::Acquire
// was the only way to find out. This case is that breakpoint, made falsifiable: delete the
// environment and it goes red, and it goes red naming the arm rather than the symptom.
TEST(DirectGLESSlotTable, TheTwinRegistryCasesInThisBinaryRunOnTheHandleArm) {
const char* knob = std::getenv("MOBILEGL_PIPE_PUSH");
if (knob != nullptr && *knob != '\0') {
GTEST_SKIP() << "the operator pinned the arm with MOBILEGL_PIPE_PUSH=" << knob;
}
EXPECT_TRUE(MobileGL::MG_Backend::DirectGLES::EsprytSlotTablesEnabled())
<< "a push build of SanityTest resolved the LEGACY twin registry, so every case in this "
"binary that builds a twin - the scratch-FBO scrub, the three context-generation "
"guards, the sampled-set staleness walk, ScopedDirectGLESTextureBindings - is "
"exercising code this package did not change";
EXPECT_NE(MobileGL::MG_Config::Features.PipePush & MobileGL::MG_Pipe::kMGPipeSubsystemEsprytSlots,
0ull);
}
#else
// G2 wants the pull and the push build to list the SAME ctest entries. The twin table only
// exists under MOBILEGL_PIPE_PUSH, so in the pull build each case above keeps its name and
@@ -3382,4 +3447,8 @@ TEST(DirectGLESSlotTable, ObjectChurnAloneDrivesTheSweep) {
TEST(DirectGLESSlotTable, GetOrCreateToleratesANullStateObject) {
GTEST_SKIP() << "the {slot, gen} twin table is compiled only under MOBILEGL_PIPE_PUSH";
}
TEST(DirectGLESSlotTable, TheTwinRegistryCasesInThisBinaryRunOnTheHandleArm) {
GTEST_SKIP() << "the {slot, gen} twin table is compiled only under MOBILEGL_PIPE_PUSH";
}
#endif // MOBILEGL_PIPE_PUSH