[PATCH v2] erofs/zmap.c: Fix incorrect offset calculation

[Siddh Raman Pant]

Effective offset to add to length was being incorrectly calculated,
which resulted in iomap->length being set to 0, triggering a WARN_ON
in iomap_iter_done().

Fix that, and describe it in comments.

This was reported as a crash by syzbot under an issue about a warning
encountered in iomap_iter_done(), but unrelated to erofs.

C reproducer: https://syzkaller.appspot.com/text?tag=ReproC&x=1037a6b2880000
Kernel config: https://syzkaller.appspot.com/text?tag=KernelConfig&x=e2021a61197ebe02
Dashboard link: https://syzkaller.appspot.com/bug?extid=a8e049cd3abd342936b6

Reported-by: syzbot+a8e049cd3abd342936b6 at syzkaller.appspotmail.com
Suggested-by: Gao Xiang <hsiangkao at linux.alibaba.com>
Signed-off-by: Siddh Raman Pant <code at siddh.me>
---
Changes since v2:
- Fix the calculation instead of bailing out.

 fs/erofs/zmap.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c
index 0bb66927e3d0..a171e4caba3c 100644
--- a/fs/erofs/zmap.c
+++ b/fs/erofs/zmap.c
@@ -790,12 +790,16 @@ static int z_erofs_iomap_begin_report(struct inode *inode, loff_t offset,
 		iomap->type = IOMAP_HOLE;
 		iomap->addr = IOMAP_NULL_ADDR;
 		/*
-		 * No strict rule how to describe extents for post EOF, yet
-		 * we need do like below. Otherwise, iomap itself will get
+		 * No strict rule on how to describe extents for post EOF, yet
+		 * we need to do like below. Otherwise, iomap itself will get
 		 * into an endless loop on post EOF.
+		 *
+		 * Calculate the effective offset by subtracting extent start
+		 * (map.m_la) from the requested offset, and add it to length.
+		 * (NB: offset >= map.m_la always)
 		 */
 		if (iomap->offset >= inode->i_size)
-			iomap->length = length + map.m_la - offset;
+			iomap->length = length + offset - map.m_la;
 	}
 	iomap->flags = 0;
 	return 0;
-- 
2.35.1