Andrew Stubbs wrote in "[PATCH v7 0/2] libgomp: OpenMP pinned memory improvements":
Compared to the v6 patchset I sent last week, this series has had the gomp_debug noise and some vestigial cruft removed, and the custom mutexes have been replaced with standard gomp_mutex ones. The "simple" allocator has also been moved into the "gomp_" namespace.
Andrew Stubbs wrote in this thread:
This patch introduces a new custom memory allocator for use with pinned memory (in the case where the Cuda allocator isn't available). In future, this allocator will also be used for Managed Memory. Both memories are incompatible with the system malloc because allocated memory cannot share a page with memory allocated for other purposes. This means that small allocations will no longer consume an entire page of pinned memory. Unfortunately, it also means that pinned memory pages will never be unmapped (although they may be reused). This isn't a technical limitation; the "free" algorithm could be extended in future, if needed.
* * *
OK for mainline?
LGTM. Thanks for the patch! Tobias
