Re: [PATCH v2] zsmalloc: prefer the the original page's node for compressed data

[Dan Williams <dan.j.williams@intel.com>]

Nhat Pham wrote:
> Currently, zsmalloc, zswap's and zram's backend memory allocator, does
> not enforce any policy for the allocation of memory for the compressed
> data, instead just adopting the memory policy of the task entering
> reclaim, or the default policy (prefer local node) if no such policy is
> specified. This can lead to several pathological behaviors in
> multi-node NUMA systems:
> 
> 1. Systems with CXL-based memory tiering can encounter the following
>    inversion with zswap/zram: the coldest pages demoted to the CXL tier
>    can return to the high tier when they are reclaimed to compressed
>    swap, creating memory pressure on the high tier.
> 
> 2. Consider a direct reclaimer scanning nodes in order of allocation
>    preference. If it ventures into remote nodes, the memory it
>    compresses there should stay there. Trying to shift those contents
>    over to the reclaiming thread's preferred node further *increases*
>    its local pressure, and provoking more spills. The remote node is
>    also the most likely to refault this data again. This undesirable
>    behavior was pointed out by Johannes Weiner in [1].
> 
> 3. For zswap writeback, the zswap entries are organized in
>    node-specific LRUs, based on the node placement of the original
>    pages, allowing for targeted zswap writeback for specific nodes.
> 
>    However, the compressed data of a zswap entry can be placed on a
>    different node from the LRU it is placed on. This means that reclaim
>    targeted at one node might not free up memory used for zswap entries
>    in that node, but instead reclaiming memory in a different node.
> 
> All of these issues will be resolved if the compressed data go to the
> same node as the original page. This patch encourages this behavior by
> having zswap and zram pass the node of the original page to zsmalloc,
> and have zsmalloc prefer the specified node if we need to allocate new
> (zs)pages for the compressed data.
> 
> Note that we are not strictly binding the allocation to the preferred
> node. We still allow the allocation to fall back to other nodes when
> the preferred node is full, or if we have zspages with slots available
> on a different node. This is OK, and still a strict improvement over
> the status quo:
> 
> 1. On a system with demotion enabled, we will generally prefer
>    demotions over compressed swapping, and only swap when pages have
>    already gone to the lowest tier. This patch should achieve the
>    desired effect for the most part.
> 
> 2. If the preferred node is out of memory, letting the compressed data
>    going to other nodes can be better than the alternative (OOMs,
>    keeping cold memory unreclaimed, disk swapping, etc.).
> 
> 3. If the allocation go to a separate node because we have a zspage
>    with slots available, at least we're not creating extra immediate
>    memory pressure (since the space is already allocated).
> 
> 3. While there can be mixings, we generally reclaim pages in
>    same-node batches, which encourage zspage grouping that is more
>    likely to go to the right node.
> 
> 4. A strict binding would require partitioning zsmalloc by node, which
>    is more complicated, and more prone to regression, since it reduces
>    the storage density of zsmalloc. We need to evaluate the tradeoff
>    and benchmark carefully before adopting such an involved solution.
> 
> [1]: https://lore.kernel.org/linux-mm/20250331165306.GC2110528@cmpxchg.org/
> 
> Suggested-by: Gregory Price <gourry@gourry.net>
> Signed-off-by: Nhat Pham <nphamcs@gmail.com>
[..]
> diff --git a/mm/zsmalloc.c b/mm/zsmalloc.c
> index 961b270f023c..8ba6cdf222ac 100644
> --- a/mm/zsmalloc.c
> +++ b/mm/zsmalloc.c
> @@ -243,9 +243,9 @@ static inline void zpdesc_dec_zone_page_state(struct zpdesc *zpdesc)
>  	dec_zone_page_state(zpdesc_page(zpdesc), NR_ZSPAGES);
>  }
>  
> -static inline struct zpdesc *alloc_zpdesc(gfp_t gfp)
> +static inline struct zpdesc *alloc_zpdesc(gfp_t gfp, const int nid)
>  {
> -	struct page *page = alloc_page(gfp);
> +	struct page *page = alloc_pages_node(nid, gfp, 0);

Why do the work to pass in @nid only to hardcode 0 to
alloc_pages_node()?