Re: [PATCH v3 3/4] mm/page_owner: add NUMA node filter with nodelist support
From: zhen.ni
Date: Wed Apr 29 2026 - 06:21:32 EST
在 2026/4/29 09:28, SeongJae Park 写道:
On Tue, 28 Apr 2026 15:11:11 +0800 Zhen Ni <zhen.ni@xxxxxxxxxxxx> wrote:Good suggestion.
Add NUMA node filtering functionality to page_owner to allow
filtering pages by specific NUMA node(s) using nodelist format.
The filter allows users to focus on pages from specific NUMA nodes,
which is useful for NUMA-aware memory allocation analysis and debugging.
Supported input formats:
- Single node: echo "2" > nid
- Multiple nodes: echo "0,2,3" > nid
- Node range: echo "0-3" > nid
- Mixed format: echo "0,2-4,7" > nid
- Disable filter: echo "-1" > nid
Link: https://lore.kernel.org/linux-mm/20260417154638.22370-4-zhen.ni@xxxxxxxxxxxx/
Link: https://lore.kernel.org/linux-mm/20260419155540.376847-4-zhen.ni@xxxxxxxxxxxx/
Seems the above two links are for v1 and v2 of this patch. I think putting
those with the context at commentary area [1] could be useful.
The reason is that `owner_filter.nid_mask` is a nodemask_t, which is aSuggested-by: Zi Yan <ziy@xxxxxxxxxx>[...]
Signed-off-by: Zhen Ni <zhen.ni@xxxxxxxxxxxx>
---
diff --git a/mm/page_owner.c b/mm/page_owner.c
index 6d87b6948cfa..e674a374669a 100644
--- a/mm/page_owner.c
+++ b/mm/page_owner.c
@@ -685,6 +685,7 @@ read_page_owner(struct file *file, char __user *buf, size_t count, loff_t *ppos)
struct page_ext *page_ext;
struct page_owner *page_owner;
depot_stack_handle_t handle;
+ nodemask_t mask;
if (!static_branch_unlikely(&page_owner_inited))
return -EINVAL;
@@ -698,6 +699,8 @@ read_page_owner(struct file *file, char __user *buf, size_t count, loff_t *ppos)
while (!pfn_valid(pfn) && (pfn & (MAX_ORDER_NR_PAGES - 1)) != 0)
pfn++;
+ mask = owner_filter.nid_mask;
+
READ_ONCE() was used for owner_filter.print_mode. Should nid_mask also read
using READ_ONCE()?
128-byte structure. READ_ONCE() only supports types up to 8 bytes and will trigger a compile-time assertion failure for larger structures.
This was actually an issue in v2 - the AI review tool (sashiko.dev) and
Andrew both caught the compilation error with READ_ONCE/WRITE_ONCE on
nodemask_t, so v3 removed them.
Yes, empty input (echo > nid) works because nodelist_parse() handles it/* Find an allocated page */
for (; pfn < max_pfn; pfn++) {
/*
@@ -730,6 +733,14 @@ read_page_owner(struct file *file, char __user *buf, size_t count, loff_t *ppos)
if (unlikely(!page_ext))
continue;
+ /* NUMA node filter using bitmask */
+ if (!nodes_empty(mask)) {
+ int nid = page_to_nid(page);
+
+ if (!node_isset(nid, mask))
+ goto ext_put_continue;
+ }
+
/*
* Some pages could be missed by concurrent allocation or free,
* because we don't hold the zone lock.
@@ -1009,6 +1020,75 @@ DEFINE_SIMPLE_ATTRIBUTE(page_owner_print_mode_fops,
&page_owner_print_mode_get,
&page_owner_print_mode_set, "%lld");
+static ssize_t nid_filter_write(struct file *file,
+ const char __user *buf,
+ size_t count, loff_t *ppos)
+{
+ char *kbuf;
+ nodemask_t mask;
+ int ret;
+ int val;
+
+ /*
+ * Limit input size to handle worst-case nodelist (all nodes).
+ * Worst case per node: ",NNNNN" (comma + 5-digit node number) = 6 bytes.
+ * Formula: 100 bytes overhead + 6 * MAX_NUMNODES
+ */
+ if (count > (100 + 6 * MAX_NUMNODES))
+ return -EINVAL;
+
+ kbuf = kmalloc(count + 1, GFP_KERNEL);
+ if (!kbuf)
+ return -ENOMEM;
+
+ if (copy_from_user(kbuf, buf, count)) {
+ ret = -EFAULT;
+ goto out_free;
+ }
+ kbuf[count] = '\0';
+
+ /* Support: "-1" to clear, or nodelist format like "0", "0,2", "0-3" */
+ if (kstrtoint(kbuf, 10, &val) == 0 && val == -1)
+ nodes_clear(mask);
+ else if (nodelist_parse(kbuf, mask)) {
+ ret = -EINVAL;
+ goto out_free;
+ }
Doesn't empty string input to nodelist_parse() clears the mask? Can't it be
reused?
correctly. However, nodelist_parse() - which is implemented via
bitmap_parselist() - cannot handle "-1" as it's not a valid range format
and would return an error. The explicit "-1" check is necessary to support `echo "-1" > nid` without returning an error.
So the "-1" check handles a case that nodelist_parse() cannot handle.
For consistency with the other file_operations+
+ owner_filter.nid_mask = mask;
+ ret = count;
+
+out_free:
+ kfree(kbuf);
+ return ret;
+}
+
+static int nid_filter_show(struct seq_file *m, void *v)
+{
+ nodemask_t mask = owner_filter.nid_mask;
+
+ if (nodes_empty(mask))
+ seq_puts(m, "-1\n");
+ else
+ seq_printf(m, "%*pbl\n", nodemask_pr_args(&mask));
+
+ return 0;
+}
+
+static int nid_filter_open(struct inode *inode, struct file *file)
+{
+ return single_open(file, nid_filter_show, NULL);
+}
+
+static const struct file_operations nid_filter_fops = {
+ .owner = THIS_MODULE,
+ .open = nid_filter_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+ .write = nid_filter_write,
+ .release = single_release,
+};
+
static int __init pageowner_init(void)
{
@@ -1024,6 +1104,8 @@ static int __init pageowner_init(void)
filter_dir = debugfs_create_dir("page_owner_filter", NULL);
debugfs_create_file("print_mode", 0600, filter_dir, NULL,
&page_owner_print_mode_fops);
+ debugfs_create_file("nid", 0600, filter_dir, NULL,
+ &nid_filter_fops);
Why don't you use 'page_owner_' prefix like other fops, for consistency?
in this module (page_owner_fops, page_owner_threshold_fops,
page_owner_print_mode_fops), I'll rename nid_filter_fops to
page_owner_nid_filter_fops.
I'll incorporate these improvements in the next version.
Thanks for the detailed review!
Best regards,dir = debugfs_create_dir("page_owner_stacks", NULL);
debugfs_create_file("show_stacks", 0400, dir,
--
2.20.1
[1] https://docs.kernel.org/process/submitting-patches.html#commentary
Thanks,
SJ
Zhen Ni