Skip to contentSkip to conversation

Fix wrong accouting of anon and file

PATCH In review Minchan Kim 2 participants 2 messages Feb 3, 2010

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

PATCH 1/1 Fix wrong accouting of anon and file

Unfortunately, Kame said he doesn't support this series. I am not sure we need this patch or revert patch.

Who need this?

David. Do you want to remain this patch in mmotm for your OOM patch in future?

If anyone doesn't reply my question, Do we have to make revert patch?

== CUT_HERE ==

mm-count-lowmem-rss.patch added lowmem accouting. But it changed file and rss accouting by mistake. Unfortunately my review also doesn't found it.

This patch fixes it.

---
 mm/memory.c |    8 ++++----
 1 files changed, 4 insertions(+), 4 deletions(-)
mm/memory.c +4 -4
index 3becdc3..ce8ff9d 100644
@@ -210,14 +210,14 @@ void sync_mm_rss(struct task_struct *task, struct mm_struct *mm)
 
 unsigned long get_file_rss(struct mm_struct *mm)
 {
-	return get_mm_counter(mm, MM_ANONPAGES)
-		+ get_mm_counter(mm, MM_ANON_LOWPAGES);
+	return get_mm_counter(mm, MM_FILEPAGES)
+		+ get_mm_counter(mm, MM_FILE_LOWPAGES);
 }
 
 unsigned long get_anon_rss(struct mm_struct *mm)
 {
-	return get_mm_counter(mm, MM_FILEPAGES)
-		+ get_mm_counter(mm, MM_FILE_LOWPAGES);
+	return get_mm_counter(mm, MM_ANONPAGES)
+		+ get_mm_counter(mm, MM_ANON_LOWPAGES);
 }
 
 unsigned long get_low_rss(struct mm_struct *mm)
Signature · 14 lines
1.6.5



-- 
Kind regards,
Minchan Kim


--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
#1 lore raw
Quoted 3 lines · Minchan Kim

Unfortunately, Kame said he doesn't support this series. I am not sure we need this patch or revert patch.

Nobody is currently using it and it adds an overhead just by doing the extra branches, so I'd be inclined to drop mm-count-lowmem-rss.patch from -mm and then reintroduce it later if something needs it down the line.

Quoted 5 lines

Who need this?

David. Do you want to remain this patch in mmotm for your OOM patch in future?

We'll need to do something for lowmem ooms so that we aren't needlessly killing taks that don't consume it. At this point, I think it's better to just fail GFP_DMA allocations where direct reclaim (and, later, memory compaction) has failed unless it is __GFP_NOFAIL, which none of them are. So this would be a change to the page allocator to defer the oom killer and return NULL on GFP_DMA instead of needlessly killing tasks.

Quoted 2 lines

If anyone doesn't reply my question, Do we have to make revert patch?

We won't need a revert patch, Andrew will be able to simply drop mm-count-lowmem-rss.patch from -mm.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>