| Message ID | 20190710045310.12397-1-damien.lemoal@wdc.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [V2] block: Fix potential overflow in blk_report_zones() | expand |
On 7/9/19 10:53 PM, Damien Le Moal wrote: > For large values of the number of zones reported and/or large zone > sizes, the sector increment calculated with > > blk_queue_zone_sectors(q) * n > > in blk_report_zones() loop can overflow the unsigned int type used for > the calculation as both "n" and blk_queue_zone_sectors() value are > unsigned int. E.g. for a device with 256 MB zones (524288 sectors), > overflow happens with 8192 or more zones reported. > > Changing the return type of blk_queue_zone_sectors() to sector_t, fixes > this problem and avoids overflow problem for all other callers of this > helper too. The same change is also applied to the bdev_zone_sectors() > helper. Applied, thanks.
diff --git a/block/blk-zoned.c b/block/blk-zoned.c index 79ad269b545d..6c503824ba3f 100644 --- a/block/blk-zoned.c +++ b/block/blk-zoned.c @@ -73,7 +73,7 @@ EXPORT_SYMBOL_GPL(__blk_req_zone_write_unlock); static inline unsigned int __blkdev_nr_zones(struct request_queue *q, sector_t nr_sectors) { - unsigned long zone_sectors = blk_queue_zone_sectors(q); + sector_t zone_sectors = blk_queue_zone_sectors(q); return (nr_sectors + zone_sectors - 1) >> ilog2(zone_sectors); } diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 5ace0bb77213..760668ed555a 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -686,7 +686,7 @@ static inline bool blk_queue_is_zoned(struct request_queue *q) } } -static inline unsigned int blk_queue_zone_sectors(struct request_queue *q) +static inline sector_t blk_queue_zone_sectors(struct request_queue *q) { return blk_queue_is_zoned(q) ? q->limits.chunk_sectors : 0; } @@ -1434,7 +1434,7 @@ static inline bool bdev_is_zoned(struct block_device *bdev) return false; } -static inline unsigned int bdev_zone_sectors(struct block_device *bdev) +static inline sector_t bdev_zone_sectors(struct block_device *bdev) { struct request_queue *q = bdev_get_queue(bdev);
For large values of the number of zones reported and/or large zone sizes, the sector increment calculated with blk_queue_zone_sectors(q) * n in blk_report_zones() loop can overflow the unsigned int type used for the calculation as both "n" and blk_queue_zone_sectors() value are unsigned int. E.g. for a device with 256 MB zones (524288 sectors), overflow happens with 8192 or more zones reported. Changing the return type of blk_queue_zone_sectors() to sector_t, fixes this problem and avoids overflow problem for all other callers of this helper too. The same change is also applied to the bdev_zone_sectors() helper. Fixes: e76239a3748c ("block: add a report_zones method") Cc: stable@vger.kernel.org Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> --- block/blk-zoned.c | 2 +- include/linux/blkdev.h | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-)