fs/sync.c: make sync_file_range(2) use WB_SYNC_NONE writeback

sync_file_range(2) is documented to issue writeback only for pages that
are not currently being written.  After all the system call has been
created for userspace to be able to issue background writeout and so
waiting for in-flight IO is undesirable there.  However commit
ee53a891f474 ("mm: do_sync_mapping_range integrity fix") switched
do_sync_mapping_range() and thus sync_file_range() to issue writeback in
WB_SYNC_ALL mode since do_sync_mapping_range() was used by other code
relying on WB_SYNC_ALL semantics.

These days do_sync_mapping_range() went away and we can switch
sync_file_range(2) back to issuing WB_SYNC_NONE writeback.  That should
help PostgreSQL avoid large latency spikes when flushing data in the
background.

Andres measured a 20% increase in transactions per second on an SSD disk.

Signed-off-by: Jan Kara <jack@suse.com>
Reported-by: Andres Freund <andres@anarazel.de>
Tested-By: Andres Freund <andres@anarazel.de>
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

diff --git a/fs/sync.c b/fs/sync.c
index 4ec430a..dd5d171 100644 (file)
--- a/fs/sync.c
+++ b/fs/sync.c
@@ -348,7 +348,8 @@ SYSCALL_DEFINE4(sync_file_range, int, fd, loff_t, offset, loff_t, nbytes,
        }
 
        if (flags & SYNC_FILE_RANGE_WRITE) {
-               ret = filemap_fdatawrite_range(mapping, offset, endbyte);
+               ret = __filemap_fdatawrite_range(mapping, offset, endbyte,
+                                                WB_SYNC_NONE);
                if (ret < 0)
                        goto out_put;
        }