Ryujinx/moltenvk
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix memory leaks when compiling shaders and pipelines without default OS autorelease pool.

Browse Source 
Wrap MVKMetalCompiler compiling execution in an @autoreleasepool.
This commit is contained in:
Bill Hollings 2019-10-07 18:52:24 -04:00
parent c6f5a09004
commit 1d628baca4
2 changed files with 2 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
Docs/Whats_New.md
View File

@ -24,6 +24,7 @@ Released TBD
- Add request for feedback from people who reject **MoltenVK** to `README.md` document.
- Allow `MVK_CONFIG_SYNCHRONOUS_QUEUE_SUBMITS` build setting to be overridden.
- Fix memory leaks of system classes during `VkInstance` and `VkQueue` creation.
- Fix memory leaks when compiling shaders and pipelines without default OS autorelease pool.

2
MoltenVK/MoltenVK/GPUObjects/MVKSync.mm
View File

@ -310,7 +310,7 @@ void MVKMetalCompiler::compile(unique_lock<mutex>& lock, dispatch_block_t block)
MVKDevice* mvkDev = _owner->getDevice();
_startTime = mvkDev->getPerformanceTimestamp();
dispatch_async(dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, 0), block);
dispatch_async(dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, 0), ^{ @autoreleasepool { block(); } });
// Limit timeout to avoid overflow since wait_for() uses wait_until()
chrono::nanoseconds nanoTimeout(min(mvkDev->_pMVKConfig->metalCompileTimeout, kMVKUndefinedLargeUInt64));