On Thu, Feb 06, 2025 at 06:50:45PM +0000, Frank van der Linden wrote:
> Commit b5389086ad7b ("hugetlbfs: extend the definition of hugepages parameter to support node allocation")
> changed the NUMA_NO_NODE round-robin allocation behavior in case of a
> failure to allocate from one NUMA node. The code originally moved on to
> the next node to try again, but now it immediately breaks out of the loop.
>
> Restore the original behavior.
Did you stumble upon this?
AFAICS, memblock_alloc_range_nid() will call memblock_find_in_range_node() with
NUMA_NO_NODE for exact_nide = false, which would be our case.
Meaning that if memblock_alloc_try_nid_raw() returns false here, it is because
we could not allocate the page in any node, right?
--
Oscar Salvador
SUSE Labs
