Skip to contentSkip to conversation

mm/memory_hotplug: return zero from do_migrate_range() for only success

PATCH In review SeongJae Park 2 participants 5 messages Feb 14, 2023 – Feb 15, 2023

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 mm/memory_hotplug: return zero from do_migrate_range() for only success

do_migrate_range() returns migrate_pages() return value, which zero means perfect success, in usual cases. If all pages are failed to be isolated, however, it returns isolate_{lru,movalbe}_page() return values, or zero if all pfn were invalid, were hugetlb or hwpoisoned. So do_migrate_range() returning zero means either perfect success, or special cases of isolation total failure.

Actually, the return value is not checked by any caller, so it might be better to simply make it a void function. However, there is a TODO for checking the return value.

Make it easier to understand what it means.

---
 mm/memory_hotplug.c | 7 +++++++
 1 file changed, 7 insertions(+)
mm/memory_hotplug.c +7 -0
index a1e8c3e9ab08..db2c02d502a2 100644
@@ -1620,6 +1620,12 @@ static int scan_movable_pages(unsigned long start, unsigned long end,
 	return 0;
 }
 
+/*
+ * migrate pages in the given pfn range.
+ *
+ * Returns the number of {normal folio, large folio, hugetlb} that were not
+ * migrated, or an error code.
+ */
 static int
 do_migrate_range(unsigned long start_pfn, unsigned long end_pfn)
 {
@@ -1685,6 +1691,7 @@ do_migrate_range(unsigned long start_pfn, unsigned long end_pfn)
 		}
 		put_page(page);
 	}
+	ret = -ENOENT;
 	if (!list_empty(&source)) {
 		nodemask_t nmask = node_states[N_MEMORY];
 		struct migration_target_control mtc = {
Signature · 1 line
2.25.1
#1 lore raw
Quoted 10 lines · SeongJae Park

do_migrate_range() returns migrate_pages() return value, which zero means perfect success, in usual cases. If all pages are failed to be isolated, however, it returns isolate_{lru,movalbe}_page() return values, or zero if all pfn were invalid, were hugetlb or hwpoisoned. So do_migrate_range() returning zero means either perfect success, or special cases of isolation total failure.

Actually, the return value is not checked by any caller, so it might be better to simply make it a void function. However, there is a TODO for checking the return value.

I'd prefer to not add more dead code ;) Let's not return an error instead.

It's still unclear which kind of fatal migration issues we actually care about and how to really detect them.

Signature · 3 lines
Thanks,

David / dhildenb
#2 lore raw
Quoted 8 lines · On Wed, 15 Feb 2023 14:16:05 +0100 David Hildenbrand <[email protected]> wrote:
Quoted 7 lines · SeongJae Park

do_migrate_range() returns migrate_pages() return value, which zero means perfect success, in usual cases. If all pages are failed to be isolated, however, it returns isolate_{lru,movalbe}_page() return values, or zero if all pfn were invalid, were hugetlb or hwpoisoned. So do_migrate_range() returning zero means either perfect success, or special cases of isolation total failure.

Quoted 5 lines
Quoted 3 lines

Actually, the return value is not checked by any caller, so it might be better to simply make it a void function. However, there is a TODO for checking the return value.

I'd prefer to not add more dead code ;) Let's not return an error instead.

Makes sense, I will send next spin soon.

Quoted 3 lines

It's still unclear which kind of fatal migration issues we actually care about and how to really detect them.

What do you think about treating the isolation/migration rate limit (migrate_rs) hit in do_migrate_range() as fatal? It warns for the event already, so definitely a bad sign.

If that's not that bad enough to be treated as fatal, I think we could have yet another rate limit to be considered fatal.

Thanks,
SJ

Quoted 5 lines
Signature · 4 lines
-- 
Thanks,

David / dhildenb
#3 lore raw
Quoted 22 lines · SeongJae Park
Quoted 8 lines · On Wed, 15 Feb 2023 14:16:05 +0100 David Hildenbrand <[email protected]> wrote:
Quoted 7 lines · SeongJae Park

do_migrate_range() returns migrate_pages() return value, which zero means perfect success, in usual cases. If all pages are failed to be isolated, however, it returns isolate_{lru,movalbe}_page() return values, or zero if all pfn were invalid, were hugetlb or hwpoisoned. So do_migrate_range() returning zero means either perfect success, or special cases of isolation total failure.

Quoted 5 lines
Quoted 3 lines

Actually, the return value is not checked by any caller, so it might be better to simply make it a void function. However, there is a TODO for checking the return value.

I'd prefer to not add more dead code ;) Let's not return an error instead.

Makes sense, I will send next spin soon.

Quoted 3 lines

It's still unclear which kind of fatal migration issues we actually care about and how to really detect them.

Quoted 6 lines

What do you think about treating the isolation/migration rate limit (migrate_rs) hit in do_migrate_range() as fatal? It warns for the event already, so definitely a bad sign.

If that's not that bad enough to be treated as fatal, I think we could have yet another rate limit to be considered fatal.

IIRC, there are some setups where offlining might take several minutes (e.g., heavy O_DIRECT load) and that's to be expected.

So the existing code warns for better debugging, but keeps trying. So the ratelimit is rather to not produce too much debug output, not to really indicate that something is fatal.

Signature · 3 lines
Thanks,

David / dhildenb
#4 lore raw
Quoted 30 lines · On Wed, 15 Feb 2023 21:00:50 +0100 David Hildenbrand <[email protected]> wrote:
Quoted 22 lines · SeongJae Park
Quoted 8 lines · On Wed, 15 Feb 2023 14:16:05 +0100 David Hildenbrand <[email protected]> wrote:
Quoted 7 lines · SeongJae Park

do_migrate_range() returns migrate_pages() return value, which zero means perfect success, in usual cases. If all pages are failed to be isolated, however, it returns isolate_{lru,movalbe}_page() return values, or zero if all pfn were invalid, were hugetlb or hwpoisoned. So do_migrate_range() returning zero means either perfect success, or special cases of isolation total failure.

Quoted 5 lines
Quoted 3 lines

Actually, the return value is not checked by any caller, so it might be better to simply make it a void function. However, there is a TODO for checking the return value.

I'd prefer to not add more dead code ;) Let's not return an error instead.

Makes sense, I will send next spin soon.

Quoted 3 lines

It's still unclear which kind of fatal migration issues we actually care about and how to really detect them.

Quoted 6 lines

What do you think about treating the isolation/migration rate limit (migrate_rs) hit in do_migrate_range() as fatal? It warns for the event already, so definitely a bad sign.

If that's not that bad enough to be treated as fatal, I think we could have yet another rate limit to be considered fatal.

Quoted 6 lines

IIRC, there are some setups where offlining might take several minutes (e.g., heavy O_DIRECT load) and that's to be expected.

So the existing code warns for better debugging, but keeps trying. So the ratelimit is rather to not produce too much debug output, not to really indicate that something is fatal.

Thank you for clarification, David!

Thanks,
SJ

Quoted 5 lines
Signature · 4 lines
-- 
Thanks,

David / dhildenb