This patchset (hopefully) fixes a problem with loading some Windows-generated UDF 2.5 filesystems, and two other problems I noticed along the way. This has only been tested on a single NetBSD-generated image (works before, works now) and a few Windows-7-generated images (triggered KASAN and failed to mount before, works fine now), but I don't have any other UDF 2.5 images to test it on. As I was getting ready to post this, I noticed on lkml that Laura Abbott has just posted a similar fix (https://lkml.org/lkml/2016/5/18/489) for the IS_ERR issue. However, I _think_ that fix is inadequate as iput() called from udf_free_partition() will still dereference the error pointer in s_mirror_fe on unmount. If I'm wrong and that patch is fine, then let me know if you want me to rework this to account for it. Also, should MAINTAINERS list linux-fsdevel under UDF? Alden Tondettar (3): udf: Don't BUG on missing metadata partition descriptor udf: Use IS_ERR when loading metadata mirror file entry udf: Use correct partition reference number for metadata fs/udf/partition.c | 13 +++++++++---- fs/udf/super.c | 22 ++++++++++++---------- fs/udf/udf_sb.h | 1 + 3 files changed, 22 insertions(+), 14 deletions(-) -- 2.1.4 -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
