From 9d5b897600a8f9405ec37a141b9417f34a11c557 Mon Sep 17 00:00:00 2001 From: Nat Goodspeed Date: Mon, 2 Dec 2019 14:39:24 -0500 Subject: DRTVWR-494: Defend LLInstanceTracker against multi-thread usage. The previous implementation went to some effort to crash if anyone attempted to create or destroy an LLInstanceTracker subclass instance during traversal. That restriction is manageable within a single thread, but becomes unworkable if it's possible that a given subclass might be used on more than one thread. Remove LLInstanceTracker::instance_iter, beginInstances(), endInstances(), also key_iter, beginKeys() and endKeys(). Instead, introduce key_snapshot() and instance_snapshot(), the only means of iterating over LLInstanceTracker instances. (These are intended to resemble functions, but in fact the current implementation simply presents the classes.) Iterating over a captured snapshot defends against container modifications during traversal. The term 'snapshot' reminds the coder that a new instance created during traversal will not be considered. To defend against instance deletion during traversal, a snapshot stores std::weak_ptrs which it lazily dereferences, skipping on the fly any that have expired. Dereferencing instance_snapshot::iterator gets you a reference rather than a pointer. Because some use cases want to delete all existing instances, add an instance_snapshot::deleteAll() method that extracts the pointer. Those cases used to require explicitly copying instance pointers into a separate container; instance_snapshot() now takes care of that. It remains the caller's responsibility to ensure that all instances of that LLInstanceTracker subclass were allocated on the heap. Replace unkeyed static LLInstanceTracker::getInstance(T*) -- which returned nullptr if that instance had been destroyed -- with new getWeak() method returning std::weak_ptr. Caller must detect expiration of that weak_ptr. Adjust tests accordingly. Use of std::weak_ptr to detect expired instances requires engaging std::shared_ptr in the constructor. We now store shared_ptrs in the static containers (std::map for keyed, std::set for unkeyed). Make LLInstanceTrackerBase a template parameterized on the type of the static data it manages. For that reason, hoist static data class declarations out of the class definitions to an LLInstanceTrackerStuff namespace. Remove the static atomic sIterationNestDepth and its methods incrementDepth(), decrementDepth() and getDepth(), since they were used only to forbid creation and destruction during traversal. Add a std::mutex to static data. Introduce an internal LockStatic class that locks the mutex while providing a pointer to static data, making that the only way to access the static data. The LLINSTANCETRACKER_DTOR_NOEXCEPT macro goes away because we no longer expect ~LLInstanceTracker() to throw an exception in test programs. That affects LLTrace::StatBase as well as LLInstanceTracker itself. Adapt consumers to the new LLInstanceTracker API. --- indra/llrender/llgl.cpp | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'indra/llrender') diff --git a/indra/llrender/llgl.cpp b/indra/llrender/llgl.cpp index c0f0cec80b..a24de45fde 100644 --- a/indra/llrender/llgl.cpp +++ b/indra/llrender/llgl.cpp @@ -2376,9 +2376,8 @@ void LLGLNamePool::release(GLuint name) //static void LLGLNamePool::upkeepPools() { - for (tracker_t::instance_iter iter = beginInstances(); iter != endInstances(); ++iter) + for (auto& pool : instance_snapshot()) { - LLGLNamePool & pool = *iter; pool.upkeep(); } } @@ -2386,9 +2385,8 @@ void LLGLNamePool::upkeepPools() //static void LLGLNamePool::cleanupPools() { - for (tracker_t::instance_iter iter = beginInstances(); iter != endInstances(); ++iter) + for (auto& pool : instance_snapshot()) { - LLGLNamePool & pool = *iter; pool.cleanup(); } } -- cgit v1.2.3 From 30fa24966463a2c10c620b782ac5c8d1b8303ceb Mon Sep 17 00:00:00 2001 From: Nat Goodspeed Date: Thu, 25 Oct 2018 11:17:21 -0400 Subject: DRTVWR-476: Fix glVertexAttrib{IPointer,PointerARB}() OpenGL calls. VS 2017 complains about the same thing that clang does: casting S32 to GLvoid* can't possibly produce a valid pointer value because S32 can't fit a whole 64-bit pointer. To appease it, not only must we use reinterpret_cast, but we must first cast S32 to intptr_t and then reinterpret_cast THAT. --- indra/llrender/llvertexbuffer.cpp | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) (limited to 'indra/llrender') diff --git a/indra/llrender/llvertexbuffer.cpp b/indra/llrender/llvertexbuffer.cpp index 1312f6afda..f89522ea57 100644 --- a/indra/llrender/llvertexbuffer.cpp +++ b/indra/llrender/llvertexbuffer.cpp @@ -1471,7 +1471,12 @@ void LLVertexBuffer::setupVertexArray() //glVertexattribIPointer requires GLSL 1.30 or later if (gGLManager.mGLSLVersionMajor > 1 || gGLManager.mGLSLVersionMinor >= 30) { - glVertexAttribIPointer(i, attrib_size[i], attrib_type[i], sTypeSize[i], (const GLvoid*) mOffsets[i]); + // nat 2018-10-24: VS 2017 also notices the issue + // described below, and warns even with reinterpret_cast. + // Cast via intptr_t to make it painfully obvious to the + // compiler that we're doing this intentionally. + glVertexAttribIPointer(i, attrib_size[i], attrib_type[i], sTypeSize[i], + reinterpret_cast(intptr_t(mOffsets[i]))); } #endif } @@ -1486,7 +1491,7 @@ void LLVertexBuffer::setupVertexArray() // rather than as an actual pointer, so it's okay. glVertexAttribPointerARB(i, attrib_size[i], attrib_type[i], attrib_normalized[i], sTypeSize[i], - reinterpret_cast(mOffsets[i])); + reinterpret_cast(intptr_t(mOffsets[i]))); } } else -- cgit v1.2.3