[PATCH 2/2] f2fs: fix panic of IO alignment feature

From: Chao Yu
Date: Fri Jul 12 2019 - 04:56:13 EST


Since 07173c3ec276 ("block: enable multipage bvecs"), one bio vector
can store multi pages, so that we can not calculate max IO size of
bio as PAGE_SIZE * bio->bi_max_vecs. However IO alignment feature of
f2fs always has that assumption, so finally, it may cause panic during
IO submission as below stack.

kernel BUG at fs/f2fs/data.c:317!
RIP: 0010:__submit_merged_bio+0x8b0/0x8c0
Call Trace:
f2fs_submit_page_write+0x3cd/0xdd0
do_write_page+0x15d/0x360
f2fs_outplace_write_data+0xd7/0x210
f2fs_do_write_data_page+0x43b/0xf30
__write_data_page+0xcf6/0x1140
f2fs_write_cache_pages+0x3ba/0xb40
f2fs_write_data_pages+0x3dd/0x8b0
do_writepages+0xbb/0x1e0
__writeback_single_inode+0xb6/0x800
writeback_sb_inodes+0x441/0x910
wb_writeback+0x261/0x650
wb_workfn+0x1f9/0x7a0
process_one_work+0x503/0x970
worker_thread+0x7d/0x820
kthread+0x1ad/0x210
ret_from_fork+0x35/0x40

This patch adds one extra condition to check left space in bio while
trying merging page to bio, to avoid panic.

This bug was reported in bugzilla:

https://bugzilla.kernel.org/show_bug.cgi?id=204043

Signed-off-by: Chao Yu <yuchao0@xxxxxxxxxx>
---
v2: add IO type check.
fs/f2fs/data.c | 10 ++++++++++
fs/f2fs/super.c | 2 +-
include/linux/f2fs_fs.h | 1 +
3 files changed, 12 insertions(+), 1 deletion(-)

diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
index f1e401f9fc13..1971e59cdedb 100644
--- a/fs/f2fs/data.c
+++ b/fs/f2fs/data.c
@@ -504,6 +504,16 @@ static bool io_is_mergeable(struct f2fs_sb_info *sbi, struct bio *bio,
block_t last_blkaddr,
block_t cur_blkaddr)
{
+ if (F2FS_IO_ALIGNED(sbi) && (fio->type == DATA || fio->type == NODE)) {
+ unsigned int filled_blocks =
+ F2FS_BYTES_TO_BLK(bio->bi_iter.bi_size);
+ unsigned int io_size = F2FS_IO_SIZE(sbi);
+ unsigned int left_vecs = bio->bi_max_vecs - bio->bi_vcnt;
+
+ /* IOs in bio is aligned and left space of vectors is not enough */
+ if (!(filled_blocks % io_size) && left_vecs < io_size)
+ return false;
+ }
if (!page_is_mergeable(sbi, bio, last_blkaddr, cur_blkaddr))
return false;
return io_type_is_mergeable(io, fio);
diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index d95a681ef7c9..a98e3b93395d 100644
--- a/fs/f2fs/super.c
+++ b/fs/f2fs/super.c
@@ -3204,7 +3204,7 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
if (err)
goto free_bio_info;

- if (F2FS_IO_SIZE(sbi) > 1) {
+ if (F2FS_IO_ALIGNED(sbi)) {
sbi->write_io_dummy =
mempool_create_page_pool(2 * (F2FS_IO_SIZE(sbi) - 1), 0);
if (!sbi->write_io_dummy) {
diff --git a/include/linux/f2fs_fs.h b/include/linux/f2fs_fs.h
index 65559900d4d7..52af9ac164b4 100644
--- a/include/linux/f2fs_fs.h
+++ b/include/linux/f2fs_fs.h
@@ -41,6 +41,7 @@
#define F2FS_IO_SIZE_BYTES(sbi) (1 << (F2FS_OPTION(sbi).write_io_size_bits + 12)) /* B */
#define F2FS_IO_SIZE_BITS(sbi) (F2FS_OPTION(sbi).write_io_size_bits) /* power of 2 */
#define F2FS_IO_SIZE_MASK(sbi) (F2FS_IO_SIZE(sbi) - 1)
+#define F2FS_IO_ALIGNED(sbi) (F2FS_IO_SIZE(sbi) > 1)

/* This flag is used by node and meta inodes, and by recovery */
#define GFP_F2FS_ZERO (GFP_NOFS | __GFP_ZERO)
--
2.18.0.rc1