v5.11 changes the blkdev lookup mechanism completely since commit
22ae8ce8b892 ("block: simplify bdev/disk lookup in blkdev_get"),
and small part of the change is to unhash part bdev inode when
deleting partition. Turns out this kind of change does fix one
nasty issue in case of BLOCK_EXT_MAJOR:
1) when one partition is deleted & closed, disk_put_part() is always
called before bdput(bdev), see blkdev_put(); so the part's devt can
be freed & re-used before the inode is dropped
2) then new partition with same devt can be created just before the
inode in 1) is dropped, then the old inode/bdev structurein 1) is
re-used for this new partition, this way causes use-after-free and
kernel panic.
It isn't possible to backport the whole fbig patchset of "merge struct
block_device and struct hd_struct v4" for addressing this issue.
https://lore.kernel.org/linux-block/20201128161510.347752-1-hch@lst.de/
So fixes it by unhashing part bdev in delete_partition(), and this way
is actually aligned with v5.11+'s behavior.
Reported-by: cuishw(a)inspur.com
Tested-by: cuishw(a)inspur.com
Cc: Christoph Hellwig <hch(a)lst.de>
Cc: Jan Kara <jack(a)suse.cz>
Signed-off-by: Ming Lei <ming.lei(a)redhat.com>
---
block/partitions/core.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/block/partitions/core.c b/block/partitions/core.c
index a02e22411594..e3d61ec4a5a6 100644
--- a/block/partitions/core.c
+++ b/block/partitions/core.c
@@ -329,6 +329,7 @@ void delete_partition(struct hd_struct *part)
struct gendisk *disk = part_to_disk(part);
struct disk_part_tbl *ptbl =
rcu_dereference_protected(disk->part_tbl, 1);
+ struct block_device *bdev;
/*
* ->part_tbl is referenced in this part's release handler, so
@@ -346,6 +347,12 @@ void delete_partition(struct hd_struct *part)
* "in-use" until we really free the gendisk.
*/
blk_invalidate_devt(part_devt(part));
+
+ bdev = bdget_part(part);
+ if (bdev) {
+ remove_inode_hash(bdev->bd_inode);
+ bdput(bdev);
+ }
percpu_ref_kill(&part->ref);
}
--
2.37.3
This bug is marked as fixed by commit:
net: core: netlink: add helper refcount dec and lock function
net: sched: add helper function to take reference to Qdisc
net: sched: extend Qdisc with rcu
net: sched: rename qdisc_destroy() to qdisc_put()
net: sched: use Qdisc rcu API instead of relying on rtnl lock
But I can't find it in any tested tree for more than 90 days.
Is it a correct commit? Please update it by replying:
#syz fix: exact-commit-title
Until then the bug is still considered open and
new crashes with the same signature are ignored.