Re: [PATCH 3/3] mm/mempolicy: introduce MPOL_WEIGHTED_INTERLEAVE for weighted interleaving

On Thu, Jan 18, 2024 at 11:05:52AM +0800, Huang, Ying wrote:

Gregory Price [off-list ref] writes:

quoted

+static unsigned long alloc_pages_bulk_array_weighted_interleave(gfp_t gfp,
+		struct mempolicy *pol, unsigned long nr_pages,
+		struct page **page_array)
+{
+	struct task_struct *me = current;
+	unsigned long total_allocated = 0;
+	unsigned long nr_allocated;
+	unsigned long rounds;
+	unsigned long node_pages, delta;
+	u8 weight;
+	struct iw_table __rcu *table;
+	u8 *weights;
+	unsigned int weight_total = 0;
+	unsigned long rem_pages = nr_pages;
+	nodemask_t nodes;
+	int nnodes, node, weight_nodes;
+	int prev_node = NUMA_NO_NODE;
+	int i;
+
+	nnodes = read_once_policy_nodemask(pol, &nodes);
+	if (!nnodes)
+		return 0;
+
+	/* Continue allocating from most recent node and adjust the nr_pages */
+	if (pol->wil.cur_weight) {
+		node = next_node_in(me->il_prev, nodes);
+		node_pages = pol->wil.cur_weight;
+		if (node_pages > rem_pages)
+			node_pages = rem_pages;
+		nr_allocated = __alloc_pages_bulk(gfp, node, NULL, node_pages,
+						  NULL, page_array);
+		page_array += nr_allocated;
+		total_allocated += nr_allocated;
+		/* if that's all the pages, no need to interleave */
+		if (rem_pages <= pol->wil.cur_weight) {
+			pol->wil.cur_weight -= rem_pages;
+			return total_allocated;
+		}
+		/* Otherwise we adjust nr_pages down, and continue from there */
+		rem_pages -= pol->wil.cur_weight;
+		pol->wil.cur_weight = 0;
+		prev_node = node;
+	}
+
+	/* fetch the weights for this operation and calculate total weight */
+	weights = kmalloc(nnodes, GFP_KERNEL);
+	if (!weights)
+		return total_allocated;
+
+	rcu_read_lock();
+	table = rcu_dereference(iw_table);
+	weight_nodes = 0;
+	for_each_node_mask(node, nodes) {
+		weights[weight_nodes++] = table->weights[node];
+		weight_total += table->weights[node];
+	}
+	rcu_read_unlock();
+
+	if (!weight_total) {
+		kfree(weights);
+		return total_allocated;
+	}
+
+	/* Now we can continue allocating as if from 0 instead of an offset */
+	rounds = rem_pages / weight_total;
+	delta = rem_pages % weight_total;
+	for (i = 0; i < nnodes; i++) {
+		node = next_node_in(prev_node, nodes);
+		weight = weights[i];
+		node_pages = weight * rounds;
+		if (delta) {
+			if (delta > weight) {
+				node_pages += weight;
+				delta -= weight;
+			} else {
+				node_pages += delta;
+				delta = 0;
+			}
+		}
+		nr_allocated = __alloc_pages_bulk(gfp, node, NULL, node_pages,
+						  NULL, page_array);
+		page_array += nr_allocated;
+		total_allocated += nr_allocated;
+		if (total_allocated == nr_pages)
+			break;
+		prev_node = node;
+	}
+
+	/*
+	 * Finally, we need to update me->il_prev and pol->wil.cur_weight
+	 * if there were overflow pages, but not equivalent to the node
+	 * weight, set the cur_weight to node_weight - delta and the
+	 * me->il_prev to the previous node. Otherwise if it was perfect
+	 * we can simply set il_prev to node and cur_weight to 0
+	 */
+	if (node_pages) {
+		me->il_prev = prev_node;
+		node_pages %= weight;
+		pol->wil.cur_weight = weight - node_pages;
+	} else {
+		me->il_prev = node;
+		pol->wil.cur_weight = 0;
+	}

It appears that we should set me->il_prev and pol->wil.cur_weight when
delta becomes 0?  That is, following allocation should start from there?

So the observation is that when delta reaches 0, we know what the prior
node should be.  The only corner case being that delta is 0 when we
enter the loop (in which case current prev_node is the correct
prev_node).

Eyeballing it, this seems correct, but I'll do some additional
validation tomorrow. That should clean up the last block a bit.

Thanks!
~Gregory