Using the USB port (as device)

Load the usbnet driver on the host:

Linux creates an Ethernet interface on the host side, also called usb0.

Give this new Ethernet interface an IP address in the same subnet as the target’s IP address. (The following command requires root permissions:)

Now the target can be accessed via the USB cable as if it was an Ethernet port. You can send a ping or open a telnet session from the host to the target, or vice versa.

Connect the USB cable from the target to an MS-Windows host computer. MS-Windows detects a new USB RNDIS/Ethernet Gadget, installs the appropriate driver, and creates a new network adapter on the system.

Configure the IP address of this network adapter by going to Control Panel > Network and Internet > Network Connections. Set, for example, an IP address of 192.168.44.1, subnet mask 255.255.255.0.

Now the target can be accessed via the USB cable as if it was an Ethernet port. You can send a ping or open a telnet session from the host to the target, or vice versa.

You must prepare the backing storage file before the gadget can use it. The backing storage is a regular file that must be created with its full desired size. This must be done before loading the gadget driver, but it only has to be done once.

The following example shows how to create a backing file of 32 MiB in the root directory of the target:

Creating the backing storage file is like having a raw disk; you still need to create partitions and install a file system before you can use it.

Use the fdisk program to partition the backing file and to specify the values of Cylinders, Heads, and Sectors of the backing file. These values are arbitrary, but since the storage gadget driver uses a sector size of 512 bytes, the formula is:

512 * Cylinders * Heads * Sectors = Total bytes of backing file storage

For our example of 32 MiB storage file, we can use values of:

Cylinders = 32

Heads = 16

Sectors = 128

So that 512*32*16*128 = 32 MiB

On the target we run fdisk with the CHS values over the backing file:

To create a primary partition that occupies the full size of the backing file:

The new partition is created by default as a Linux (ext4 file system) partition. To change the partition type to FAT32:

Press p to print the partition table details and verify everything is correct:

Press w to write out the partition table to the backing storage:

At this point, the partition has been created but it doesn’t hold a file system yet.

The easiest way to create a file system in the partition is to load the driver, connect the USB cable to the host PC, and create the file system on the recently created partition from the host computer.

A Linux host computer recognizes the target as a new SCSI device. Linux assigns a device node in the form sdx, where x is a letter index, and the partition appears as sdx1. To see exactly which device the host assigned, print out the system log messages with:

In the example, the host assigned sdb1 as the device node for the partition of our backing file.

On the host, create a VFAT file system into that partition (you might need root permissions):

When formatting a device on your Host PC, verify that the device node dev/sdX corresponds to the connected USB device and not to your PC hard disk. Running the mkfs.vfat command on the wrong device node might format your hard disk.

Now, mount the device in the host so the host has read/write access to it. If your host PC is prepared to automatically mount devices, you can simply disconnect the USB cable and connect it again. After a few seconds, the PC automatically detects and mounts the device as a USB storage drive, with the size of the partition we created. Initially, the disk is empty.

If the PC doesn’t auto mount the device you need to do it by hand (you might need root permissions):

In this example we’ve mounted the device into the folder /media/usbdisk. Now we can create files and folders within this folder.

The placeholder for the files and folders that we create in this folder is the 32 MiB file /backing_file that we created in the first steps.

As with other USB devices, before removing the USB cable you need to unmount this folder:

An MS-Windows host computer recognizes the target as a new USB mass storage device. A warning message appears to inform you that the disk is not yet formatted. Click Yes to enter the formatting tool.

Select FAT file system and an optional volume label (less than 11 characters). Then click Start.

Once the volume has been formatted with FAT file system you can use it as any other removable media and create folders, copy files, etc.

The placeholder for the files and folders that we create in the unit is the 32 MiB file /backing_file that we created in the first steps.

When using a block device for the file storage gadget, we recommended you first unmount the media on the target side. Otherwise, access from the host won’t be synchronized with accesses from the target, as the system would behave like having the same media mounted twice in different mount points.

Make sure any partitions of the block device are unmounted, for example:

Load the mass-storage gadget passing the block device node (in the example the SD card device node) as backing file:

Now, when you connect the USB cable to a host computer, the two partitions of the SD card will be accessible from the host.