Skip to contentSkip to conversation

revert-mm-dont-reclaim-inodes-with-many-attached-pages.patch removed from -mm tree

merged [email protected] 1 participant 1 message Feb 13, 2019 Stable branch

lore
Share & tools
lore

[email protected]

author tree notice Applied
#0 lore raw

The patch titled
Subject: Revert "mm: don't reclaim inodes with many attached pages"
has been removed from the -mm tree. Its filename was
revert-mm-dont-reclaim-inodes-with-many-attached-pages.patch

This patch was dropped because it was merged into mainline or a subsystem tree

------------------------------------------------------
From: Dave Chinner <[email protected]>
Subject: Revert "mm: don't reclaim inodes with many attached pages"

This reverts a76cf1a474d7d ("mm: don't reclaim inodes with many attached pages").

This change causes serious changes to page cache and inode cache behaviour and balance, resulting in major performance regressions when combining worklaods such as large file copies and kernel compiles.

https://bugzilla.kernel.org/show_bug.cgi?id=202441

This change is a hack to work around the problems introduced by changing how agressive shrinkers are on small caches in commit 172b06c32b94 ("mm:
slowly shrink slabs with a relatively small number of objects"). It creates more problems than it solves, wasn't adequately reviewed or tested, so it needs to be reverted.

Fixes a76cf1a474d7d ("mm: don't reclaim inodes with many attached pages")
Cc Wolfgang Walter <[email protected]>
Cc Roman Gushchin <[email protected]>
Cc Rik van Riel <[email protected]>
Cc Michal Hocko <[email protected]>

---

 fs/inode.c |    7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)
fs/inode.c renamed from fs/inode.c~revert-mm-dont-reclaim-inodes-with-many-attached-pages +2 -5
@@ -730,11 +730,8 @@ static enum lru_status inode_lru_isolate
 		return LRU_REMOVED;
 	}
 
-	/*
-	 * Recently referenced inodes and inodes with many attached pages
-	 * get one more pass.
-	 */
-	if (inode->i_state & I_REFERENCED || inode->i_data.nrpages > 1) {
+	/* recently referenced inodes get one more pass */
+	if (inode->i_state & I_REFERENCED) {
 		inode->i_state &= ~I_REFERENCED;
 		spin_unlock(&inode->i_lock);
 		return LRU_ROTATE;

_

Patches currently in -mm which might be from [email protected] are