btrfs: validate queue limits

Call blk_validate_limits on the queue limits used for zone append
splitting so that calculated values get filled in and any stacking
conflicts get cought.

Without this there isn't a max_zone_append_sectors limits as of commit
559218d ("block: pre-calculate max_zone_append_sectors").

Fixes: 559218d ("block: pre-calculate max_zone_append_sectors")
Reported-by: Yi Zhang <yi.zhang@redhat.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Link: https://lore.kernel.org/r/20241113084541.34315-3-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>

fs/btrfs/zoned.c

@@ -711,6 +711,12 @@ int btrfs_check_zoned_mode(struct btrfs_fs_info *fs_info)
 			blk_stack_limits(lim, bdev_limits(device->bdev), 0);
 	}
 
+	ret = blk_validate_limits(lim);
+	if (ret) {
+		btrfs_err(fs_info, "zoned: failed to validate queue limits");
+		return ret;
+	}
+
 	/*
 	 * stripe_size is always aligned to BTRFS_STRIPE_LEN in
 	 * btrfs_create_chunk(). Since we want stripe_len == zone_size,