On 10.03.21 17:14, Minchan Kim wrote:
ffer_head LRU caches will be pinned and thus cannot be migrated. This can prevent CMA allocations from succeeding, which are often used on platforms with co-processors (such as a DSP) that can only use physically contiguous memory. It can also prevent memory hot-unplugging from succeeding, which involves migrating at least MIN_MEMORY_BLOCK_SIZE bytes of memory, which ranges from 8 MiB to 1 GiB based on the architecture in use.
Actually, it's memory_block_size_bytes(), which can be even bigger (IIRC, 128MiB..2 GiB on x86-64) that fails to get offlined. But that will prevent bigger granularity (e.g., a whole DIMM) from getting unplugged.
Correspondingly, invalidate the BH LRU caches before a migration starts and stop any buffer_head from being cached in the LRU caches, until migration has finished.
Sounds sane to me. -- Thanks, David / dhildenb
