On 27/02/2026 19:05, Benjamin Marzinski wrote:
On Wed, Feb 25, 2026 at 03:32:20PM +0000, John Garry wrote:
Add helpers for driver sysfs code for the following functionality:
- get/set iopolicy with mpath_iopolicy_store() and mpath_iopolicy_show()
- show device path per NUMA node
- "multipath" attribute group, equivalent to nvme_ns_mpath_attr_group
- device groups attribute array, similar to nvme_ns_attr_groups but not
containing NVMe members.
Note that mpath_iopolicy_store() has a update callback to allow same
functionality as nvme_subsys_iopolicy_update() be run for clearing paths.
Signed-off-by: John Garry <[email protected]>
diff --git a/lib/multipath.c b/lib/multipath.c
index 1ce57b9b14d2e..c05b4d25ca223 100644
--- a/lib/multipath.c
+++ b/lib/multipath.c
@@ -745,6 +745,116 @@ void mpath_device_set_live(struct mpath_disk *mpath_disk,
}
EXPORT_SYMBOL_GPL(mpath_device_set_live);
+static struct attribute dummy_attr = {
+ .name = "dummy",
+};
+
+static struct attribute *mpath_attrs[] = {
+ &dummy_attr,
+ NULL
+};
+
+static bool multipath_sysfs_group_visible(struct kobject *kobj)
+{
+ struct device *dev = container_of(kobj, struct device, kobj);
+ struct gendisk *disk = dev_to_disk(dev);
+
+ return is_mpath_head(disk);
+}
+
+static bool multipath_sysfs_attr_visible(struct kobject *kobj,
+ struct attribute *attr, int n)
+{
+ return false;
+}
+
+DEFINE_SYSFS_GROUP_VISIBLE(multipath_sysfs)
nitpick: this could use DEFINE_SIMPLE_SYSFS_GROUP_VISIBLE instead.
Yes, that seems reasonable. And, FWIW, I think that
multipath_sysfs_attr_visible() should return umode_t.
BTW, this is same as mainline NVMe code, so that could be updated first.
+
+const struct attribute_group mpath_attr_group = {
+ .name = "multipath",
+ .attrs = mpath_attrs,
+ .is_visible = SYSFS_GROUP_VISIBLE(multipath_sysfs),
+};
+EXPORT_SYMBOL_GPL(mpath_attr_group);
+
+const struct attribute_group *mpath_device_groups[] = {
+ &mpath_attr_group,
+ NULL
+};
+EXPORT_SYMBOL_GPL(mpath_device_groups);
+
+ssize_t mpath_iopolicy_show(struct mpath_iopolicy *mpath_iopolicy, char *buf)
+{
+ return sysfs_emit(buf, "%s\n",
+ mpath_iopolicy_names[mpath_read_iopolicy(mpath_iopolicy)]);
+}
+EXPORT_SYMBOL_GPL(mpath_iopolicy_show);
+
+static void mpath_iopolicy_update(struct mpath_iopolicy *mpath_iopolicy,
+ int iopolicy, void (*update)(void *), void *data)
+{
+ int old_iopolicy = READ_ONCE(mpath_iopolicy->iopolicy);
+
+ if (old_iopolicy == iopolicy)
+ return;
+
+ WRITE_ONCE(mpath_iopolicy->iopolicy, iopolicy);
+
+ /*
+ * iopolicy changes clear the mpath by design, which @update
+ * must do.
+ */
+ update(data);
+
+ pr_err("iopolicy changed from %s to %s\n",
+ mpath_iopolicy_names[old_iopolicy],
+ mpath_iopolicy_names[iopolicy]);
I not sure this warrants a pr_err().
Agreed, I can downgrade this to something like pr_info or notice.
Thanks
