[PATCH] z3fold: fix reclaim lock-ups

From: Vitaly Wool
Date: Mon Apr 30 2018 - 06:58:12 EST


Do not try to optimize in-page object layout while the page is
under reclaim. This fixes lock-ups on reclaim and improves reclaim
performance at the same time.

Reported-by: Guenter Roeck <linux@xxxxxxxxxxxx>
Signed-off-by: Vitaly Wool <vitaly.vul@xxxxxxxx>
---
mm/z3fold.c | 42 ++++++++++++++++++++++++++++++------------
1 file changed, 30 insertions(+), 12 deletions(-)

diff --git a/mm/z3fold.c b/mm/z3fold.c
index c0bca6153b95..901c0b07cbda 100644
--- a/mm/z3fold.c
+++ b/mm/z3fold.c
@@ -144,7 +144,8 @@ enum z3fold_page_flags {
PAGE_HEADLESS = 0,
MIDDLE_CHUNK_MAPPED,
NEEDS_COMPACTING,
- PAGE_STALE
+ PAGE_STALE,
+ UNDER_RECLAIM
};

/*****************
@@ -173,6 +174,7 @@ static struct z3fold_header *init_z3fold_page(struct page *page,
clear_bit(MIDDLE_CHUNK_MAPPED, &page->private);
clear_bit(NEEDS_COMPACTING, &page->private);
clear_bit(PAGE_STALE, &page->private);
+ clear_bit(UNDER_RECLAIM, &page->private);

spin_lock_init(&zhdr->page_lock);
kref_init(&zhdr->refcount);
@@ -756,6 +758,10 @@ static void z3fold_free(struct z3fold_pool *pool, unsigned long handle)
atomic64_dec(&pool->pages_nr);
return;
}
+ if (test_bit(UNDER_RECLAIM, &page->private)) {
+ z3fold_page_unlock(zhdr);
+ return;
+ }
if (test_and_set_bit(NEEDS_COMPACTING, &page->private)) {
z3fold_page_unlock(zhdr);
return;
@@ -840,6 +846,8 @@ static int z3fold_reclaim_page(struct z3fold_pool *pool, unsigned int retries)
kref_get(&zhdr->refcount);
list_del_init(&zhdr->buddy);
zhdr->cpu = -1;
+ set_bit(UNDER_RECLAIM, &page->private);
+ break;
}

list_del_init(&page->lru);
@@ -887,25 +895,35 @@ static int z3fold_reclaim_page(struct z3fold_pool *pool, unsigned int retries)
goto next;
}
next:
- spin_lock(&pool->lock);
if (test_bit(PAGE_HEADLESS, &page->private)) {
if (ret == 0) {
- spin_unlock(&pool->lock);
free_z3fold_page(page);
return 0;
}
- } else if (kref_put(&zhdr->refcount, release_z3fold_page)) {
- atomic64_dec(&pool->pages_nr);
+ spin_lock(&pool->lock);
+ list_add(&page->lru, &pool->lru);
+ spin_unlock(&pool->lock);
+ } else {
+ z3fold_page_lock(zhdr);
+ clear_bit(UNDER_RECLAIM, &page->private);
+ if (kref_put(&zhdr->refcount,
+ release_z3fold_page_locked)) {
+ atomic64_dec(&pool->pages_nr);
+ return 0;
+ }
+ /*
+ * if we are here, the page is still not completely
+ * free. Take the global pool lock then to be able
+ * to add it back to the lru list
+ */
+ spin_lock(&pool->lock);
+ list_add(&page->lru, &pool->lru);
spin_unlock(&pool->lock);
- return 0;
+ z3fold_page_unlock(zhdr);
}

- /*
- * Add to the beginning of LRU.
- * Pool lock has to be kept here to ensure the page has
- * not already been released
- */
- list_add(&page->lru, &pool->lru);
+ /* We started off locked to we need to lock the pool back */
+ spin_lock(&pool->lock);
}
spin_unlock(&pool->lock);
return -EAGAIN;
--
2.15.1