diff mbox series

[5/7] mm/sparse.c: update code comment about section activate/deactivate

Message ID 20200209104826.3385-6-bhe@redhat.com (mailing list archive)
State New, archived
Headers show
Series mm/hotplug: Only use subsection in VMEMMAP case and fix hot add/remove failure in SPARSEMEM|!VMEMMAP case | expand

Commit Message

Baoquan He Feb. 9, 2020, 10:48 a.m. UTC
It's helpful to note that sub-section is only supported in
SPARSEMEM_VMEMMAP case, but not in SPARSEMEM|!VMEMMAP case. Add
sentences into the code comment above sparse_add_section.

And move the code comments inside section_deactivate() to be above
it, this makes code cleaner.

Signed-off-by: Baoquan He <bhe@redhat.com>
---
 mm/sparse.c | 40 +++++++++++++++++++++++-----------------
 1 file changed, 23 insertions(+), 17 deletions(-)
diff mbox series

Patch

diff --git a/mm/sparse.c b/mm/sparse.c
index 36e6565ec67e..a7e78bfe0dce 100644
--- a/mm/sparse.c
+++ b/mm/sparse.c
@@ -809,6 +809,23 @@  static void free_map_bootmem(struct page *memmap)
 }
 #endif /* CONFIG_SPARSEMEM_VMEMMAP */
 
+/*
+ * To deactivate a memory region, there are 3 cases across two
+ * two configurations (SPARSEMEM_VMEMMAP={y,n}):
+ *
+ * 1/ deactivation of a partial hot-added section (only possible
+ * in the SPARSEMEM_VMEMMAP=y case).
+ *    a/ section was present at memory init
+ *    b/ section was hot-added post memory init
+ * 2/ deactivation of a complete hot-added section
+ * 3/ deactivation of a complete section from memory init
+ *
+ * For 1/, when subsection_map does not empty we will not be
+ * freeing the usage map, but still need to free the vmemmap
+ * range.
+ *
+ * For 2/ and 3/ the SPARSEMEM_VMEMMAP={y,n} cases are unified
+ */
 static void section_deactivate(unsigned long pfn, unsigned long nr_pages,
 		struct vmem_altmap *altmap)
 {
@@ -821,23 +838,7 @@  static void section_deactivate(unsigned long pfn, unsigned long nr_pages,
 	rc = clear_subsection_map(pfn, nr_pages);
 	if(IS_ERR_VALUE((unsigned long)rc))
 		return;
-	/*
-	 * There are 3 cases to handle across two configurations
-	 * (SPARSEMEM_VMEMMAP={y,n}):
-	 *
-	 * 1/ deactivation of a partial hot-added section (only possible
-	 * in the SPARSEMEM_VMEMMAP=y case).
-	 *    a/ section was present at memory init
-	 *    b/ section was hot-added post memory init
-	 * 2/ deactivation of a complete hot-added section
-	 * 3/ deactivation of a complete section from memory init
-	 *
-	 * For 1/, when subsection_map does not empty we will not be
-	 * freeing the usage map, but still need to free the vmemmap
-	 * range.
-	 *
-	 * For 2/ and 3/ the SPARSEMEM_VMEMMAP={y,n} cases are unified
-	 */
+
 	if (!rc) {
 		unsigned long section_nr = pfn_to_section_nr(pfn);
 
@@ -913,6 +914,11 @@  static struct page * __meminit section_activate(int nid, unsigned long pfn,
  *
  * This is only intended for hotplug.
  *
+ * Note that the added memory region is either section aligned, or
+ * sub-section aligned. The sub-section aligned region can only be
+ * hot added in SPARSEMEM_VMEMMAP case, please refer to ZONE_DEVICE
+ * part of memory-model.rst for more details.
+ *
  * Return:
  * * 0		- On success.
  * * -EEXIST	- Section has been present.