Fixed a race condition with incremental defragmentation. The issue here is that VmaBlockVector::DefragmentationEnd() relied on the mutex being previously locked to safely mutate its data, but with incremental defrag this isn't guaranteed to be the case anymore.
diff --git a/src/vk_mem_alloc.h b/src/vk_mem_alloc.h
index 7b31463..a466e1f 100644
--- a/src/vk_mem_alloc.h
+++ b/src/vk_mem_alloc.h
@@ -6523,6 +6523,7 @@
VkCommandBuffer commandBuffer);
void DefragmentationEnd(
class VmaBlockVectorDefragmentationContext* pCtx,
+ uint32_t flags,
VmaDefragmentationStats* pStats);
uint32_t ProcessDefragmentations(
@@ -13180,22 +13181,36 @@
void VmaBlockVector::DefragmentationEnd(
class VmaBlockVectorDefragmentationContext* pCtx,
+ uint32_t flags,
VmaDefragmentationStats* pStats)
{
- // Destroy buffers.
- for(size_t blockIndex = pCtx->blockContexts.size(); blockIndex--; )
+ if(flags & VMA_DEFRAGMENTATION_FLAG_INCREMENTAL && m_hAllocator->m_UseMutex)
{
- VmaBlockDefragmentationContext& blockCtx = pCtx->blockContexts[blockIndex];
- if(blockCtx.hBuffer)
- {
- (*m_hAllocator->GetVulkanFunctions().vkDestroyBuffer)(
- m_hAllocator->m_hDevice, blockCtx.hBuffer, m_hAllocator->GetAllocationCallbacks());
- }
+ VMA_ASSERT(pCtx->mutexLocked == false);
+
+ // Incremental defragmentation doesn't hold the lock, so when we enter here we don't actually have any
+ // lock protecting us. Since we mutate state here, we have to take the lock out now
+ m_Mutex.LockWrite();
+ pCtx->mutexLocked = true;
}
- if(pCtx->res >= VK_SUCCESS)
+ // If the mutex isn't locked we didn't do any work and there is nothing to delete.
+ if(pCtx->mutexLocked || !m_hAllocator->m_UseMutex)
{
- FreeEmptyBlocks(pStats);
+ // Destroy buffers.
+ for(size_t blockIndex = pCtx->blockContexts.size(); blockIndex--;)
+ {
+ VmaBlockDefragmentationContext &blockCtx = pCtx->blockContexts[blockIndex];
+ if(blockCtx.hBuffer)
+ {
+ (*m_hAllocator->GetVulkanFunctions().vkDestroyBuffer)(m_hAllocator->m_hDevice, blockCtx.hBuffer, m_hAllocator->GetAllocationCallbacks());
+ }
+ }
+
+ if(pCtx->res >= VK_SUCCESS)
+ {
+ FreeEmptyBlocks(pStats);
+ }
}
if(pCtx->mutexLocked)
@@ -14117,7 +14132,7 @@
for(size_t i = m_CustomPoolContexts.size(); i--; )
{
VmaBlockVectorDefragmentationContext* pBlockVectorCtx = m_CustomPoolContexts[i];
- pBlockVectorCtx->GetBlockVector()->DefragmentationEnd(pBlockVectorCtx, m_pStats);
+ pBlockVectorCtx->GetBlockVector()->DefragmentationEnd(pBlockVectorCtx, m_Flags, m_pStats);
vma_delete(m_hAllocator, pBlockVectorCtx);
}
for(size_t i = m_hAllocator->m_MemProps.memoryTypeCount; i--; )
@@ -14125,7 +14140,7 @@
VmaBlockVectorDefragmentationContext* pBlockVectorCtx = m_DefaultPoolContexts[i];
if(pBlockVectorCtx)
{
- pBlockVectorCtx->GetBlockVector()->DefragmentationEnd(pBlockVectorCtx, m_pStats);
+ pBlockVectorCtx->GetBlockVector()->DefragmentationEnd(pBlockVectorCtx, m_Flags, m_pStats);
vma_delete(m_hAllocator, pBlockVectorCtx);
}
}
