mm/zswap.c: avoid unnecessary copy-in at map time

The buf mapped via zpool_map_handle() is only used to store compressed
page buffer and there is no information to extract from it. So we could
use ZPOOL_MM_WO instead to avoid unnecessary copy-in at map time.

Link: https://lkml.kernel.org/r/20210522092242.3233191-3-linmiaohe@huawei.com
Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
Cc: Colin Ian King <colin.king@canonical.com>
Cc: Dan Streetman <ddstreet@ieee.org>
Cc: Nathan Chancellor <nathan@kernel.org>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Seth Jennings <sjenning@redhat.com>
Cc: Tian Tao <tiantao6@hisilicon.com>
Cc: Vitaly Wool <vitaly.wool@konsulko.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

diff --git a/mm/zswap.c b/mm/zswap.c
index e934593..c27ce9f 100644 (file)
--- a/mm/zswap.c
+++ b/mm/zswap.c
@@ -1203,7 +1203,7 @@ static int zswap_frontswap_store(unsigned type, pgoff_t offset,
                zswap_reject_alloc_fail++;
                goto put_dstmem;
        }
-       buf = zpool_map_handle(entry->pool->zpool, handle, ZPOOL_MM_RW);
+       buf = zpool_map_handle(entry->pool->zpool, handle, ZPOOL_MM_WO);
        memcpy(buf, &zhdr, hlen);
        memcpy(buf + hlen, dst, dlen);
        zpool_unmap_handle(entry->pool->zpool, handle);