On 11/6/19 4:14 PM, Christoph Hellwig wrote:
> A lot of the logic in invalidate_partitions and invalidate_partitions
... rescan_partitions and invalidate_partitions ...
> is shared. Merge the two functions to simplify things. There is
> a small behavior change in that we now send the keven change notice
> also if we were not invalidating but no partitions were found, which
> seems like the right thing to do.
>
> Signed-off-by: Christoph Hellwig <hch@xxxxxx>
> ---
> block/ioctl.c | 2 +-
> block/partition-generic.c | 38 ++++++++++++++------------------------
> fs/block_dev.c | 5 +----
> include/linux/genhd.h | 4 ++--
> 4 files changed, 18 insertions(+), 31 deletions(-)
>
> diff --git a/block/ioctl.c b/block/ioctl.c
> index 15a0eb80ada9..8a7e33ce2097 100644
> --- a/block/ioctl.c
> +++ b/block/ioctl.c
> @@ -171,7 +171,7 @@ int __blkdev_reread_part(struct block_device *bdev)
>
> lockdep_assert_held(&bdev->bd_mutex);
>
> - return rescan_partitions(disk, bdev);
> + return rescan_partitions(disk, bdev, false);
> }
> EXPORT_SYMBOL(__blkdev_reread_part);
>
> diff --git a/block/partition-generic.c b/block/partition-generic.c
> index f113be069b40..eae9daa8a523 100644
> --- a/block/partition-generic.c
> +++ b/block/partition-generic.c
> @@ -632,7 +632,8 @@ static int blk_add_partitions(struct gendisk *disk, struct block_device *bdev)
> return ret;
> }
>
> -int rescan_partitions(struct gendisk *disk, struct block_device *bdev)
> +int rescan_partitions(struct gendisk *disk, struct block_device *bdev,
> + bool invalidate)
> {
> int ret;
>
> @@ -641,13 +642,22 @@ int rescan_partitions(struct gendisk *disk, struct block_device *bdev)
> if (ret)
> return ret;
>
> - if (disk->fops->revalidate_disk)
> + if (invalidate)
> + set_capacity(disk, 0);
> + else if (disk->fops->revalidate_disk)
> disk->fops->revalidate_disk(disk);
> - check_disk_size_change(disk, bdev, true);
> +
> + check_disk_size_change(disk, bdev, !invalidate);
> bdev->bd_invalidated = 0;
>
> - if (!get_capacity(disk))
> + if (!get_capacity(disk)) {
> + /*
> + * Tell userspace that the media / partition table may have
> + * changed.
> + */
> + kobject_uevent(&disk_to_dev(disk)->kobj, KOBJ_CHANGE);
> return 0;
> + }
>
I wonder; wouldn't we miss a true size change here?
check_disk_size_change() will issue a kernel message (but no usevent) if
we have a real disk size change; if we fail to revalidate the disk an
uevent is issued, but no kernel message.
Can't we combine both by making check_disk_size_change() a bool function
and drop the call to 'get_capacity()', seeing that it's done in
check_disk_size_change() anyway?
Cheers,
Hannes
--
Dr. Hannes Reinecke Teamlead Storage & Networking
hare@xxxxxxx +49 911 74053 688
SUSE Software Solutions Germany GmbH, Maxfeldstr. 5, 90409 Nürnberg
HRB 247165 (AG München), GF: Felix Imendörffer
