xref: /linux/Documentation/admin-guide/thunderbolt.rst (revision ae99fb8baafc881b35aa0b79d7ac0178a7c40c89)
1.. SPDX-License-Identifier: GPL-2.0
2
3======================
4 USB4 and Thunderbolt
5======================
6USB4 is the public specification based on Thunderbolt 3 protocol with
7some differences at the register level among other things. Connection
8manager is an entity running on the host router (host controller)
9responsible for enumerating routers and establishing tunnels. A
10connection manager can be implemented either in firmware or software.
11Typically PCs come with a firmware connection manager for Thunderbolt 3
12and early USB4 capable systems. Apple systems on the other hand use
13software connection manager and the later USB4 compliant devices follow
14the suit.
15
16The Linux Thunderbolt driver supports both and can detect at runtime which
17connection manager implementation is to be used. To be on the safe side the
18software connection manager in Linux also advertises security level
19``user`` which means PCIe tunneling is disabled by default. The
20documentation below applies to both implementations with the exception that
21the software connection manager only supports ``user`` security level and
22is expected to be accompanied with an IOMMU based DMA protection.
23
24Security levels and how to use them
25-----------------------------------
26The interface presented here is not meant for end users. Instead there
27should be a userspace tool that handles all the low-level details, keeps
28a database of the authorized devices and prompts users for new connections.
29
30More details about the sysfs interface for Thunderbolt devices can be
31found in ``Documentation/ABI/testing/sysfs-bus-thunderbolt``.
32
33Those users who just want to connect any device without any sort of
34manual work can add following line to
35``/etc/udev/rules.d/99-local.rules``::
36
37  ACTION=="add", SUBSYSTEM=="thunderbolt", ATTR{authorized}=="0", ATTR{authorized}="1"
38
39This will authorize all devices automatically when they appear. However,
40keep in mind that this bypasses the security levels and makes the system
41vulnerable to DMA attacks.
42
43Starting with Intel Falcon Ridge Thunderbolt controller there are 4
44security levels available. Intel Titan Ridge added one more security level
45(usbonly). The reason for these is the fact that the connected devices can
46be DMA masters and thus read contents of the host memory without CPU and OS
47knowing about it. There are ways to prevent this by setting up an IOMMU but
48it is not always available for various reasons.
49
50The security levels are as follows:
51
52  none
53    All devices are automatically connected by the firmware. No user
54    approval is needed. In BIOS settings this is typically called
55    *Legacy mode*.
56
57  user
58    User is asked whether the device is allowed to be connected.
59    Based on the device identification information available through
60    ``/sys/bus/thunderbolt/devices``, the user then can make the decision.
61    In BIOS settings this is typically called *Unique ID*.
62
63  secure
64    User is asked whether the device is allowed to be connected. In
65    addition to UUID the device (if it supports secure connect) is sent
66    a challenge that should match the expected one based on a random key
67    written to the ``key`` sysfs attribute. In BIOS settings this is
68    typically called *One time saved key*.
69
70  dponly
71    The firmware automatically creates tunnels for Display Port and
72    USB. No PCIe tunneling is done. In BIOS settings this is
73    typically called *Display Port Only*.
74
75  usbonly
76    The firmware automatically creates tunnels for the USB controller and
77    Display Port in a dock. All PCIe links downstream of the dock are
78    removed.
79
80The current security level can be read from
81``/sys/bus/thunderbolt/devices/domainX/security`` where ``domainX`` is
82the Thunderbolt domain the host controller manages. There is typically
83one domain per Thunderbolt host controller.
84
85If the security level reads as ``user`` or ``secure`` the connected
86device must be authorized by the user before PCIe tunnels are created
87(e.g the PCIe device appears).
88
89Each Thunderbolt device plugged in will appear in sysfs under
90``/sys/bus/thunderbolt/devices``. The device directory carries
91information that can be used to identify the particular device,
92including its name and UUID.
93
94Authorizing devices when security level is ``user`` or ``secure``
95-----------------------------------------------------------------
96When a device is plugged in it will appear in sysfs as follows::
97
98  /sys/bus/thunderbolt/devices/0-1/authorized	- 0
99  /sys/bus/thunderbolt/devices/0-1/device	- 0x8004
100  /sys/bus/thunderbolt/devices/0-1/device_name	- Thunderbolt to FireWire Adapter
101  /sys/bus/thunderbolt/devices/0-1/vendor	- 0x1
102  /sys/bus/thunderbolt/devices/0-1/vendor_name	- Apple, Inc.
103  /sys/bus/thunderbolt/devices/0-1/unique_id	- e0376f00-0300-0100-ffff-ffffffffffff
104
105The ``authorized`` attribute reads 0 which means no PCIe tunnels are
106created yet. The user can authorize the device by simply entering::
107
108  # echo 1 > /sys/bus/thunderbolt/devices/0-1/authorized
109
110This will create the PCIe tunnels and the device is now connected.
111
112If the device supports secure connect, and the domain security level is
113set to ``secure``, it has an additional attribute ``key`` which can hold
114a random 32-byte value used for authorization and challenging the device in
115future connects::
116
117  /sys/bus/thunderbolt/devices/0-3/authorized	- 0
118  /sys/bus/thunderbolt/devices/0-3/device	- 0x305
119  /sys/bus/thunderbolt/devices/0-3/device_name	- AKiTiO Thunder3 PCIe Box
120  /sys/bus/thunderbolt/devices/0-3/key		-
121  /sys/bus/thunderbolt/devices/0-3/vendor	- 0x41
122  /sys/bus/thunderbolt/devices/0-3/vendor_name	- inXtron
123  /sys/bus/thunderbolt/devices/0-3/unique_id	- dc010000-0000-8508-a22d-32ca6421cb16
124
125Notice the key is empty by default.
126
127If the user does not want to use secure connect they can just ``echo 1``
128to the ``authorized`` attribute and the PCIe tunnels will be created in
129the same way as in the ``user`` security level.
130
131If the user wants to use secure connect, the first time the device is
132plugged a key needs to be created and sent to the device::
133
134  # key=$(openssl rand -hex 32)
135  # echo $key > /sys/bus/thunderbolt/devices/0-3/key
136  # echo 1 > /sys/bus/thunderbolt/devices/0-3/authorized
137
138Now the device is connected (PCIe tunnels are created) and in addition
139the key is stored on the device NVM.
140
141Next time the device is plugged in the user can verify (challenge) the
142device using the same key::
143
144  # echo $key > /sys/bus/thunderbolt/devices/0-3/key
145  # echo 2 > /sys/bus/thunderbolt/devices/0-3/authorized
146
147If the challenge the device returns back matches the one we expect based
148on the key, the device is connected and the PCIe tunnels are created.
149However, if the challenge fails no tunnels are created and error is
150returned to the user.
151
152If the user still wants to connect the device they can either approve
153the device without a key or write a new key and write 1 to the
154``authorized`` file to get the new key stored on the device NVM.
155
156DMA protection utilizing IOMMU
157------------------------------
158Recent systems from 2018 and forward with Thunderbolt ports may natively
159support IOMMU. This means that Thunderbolt security is handled by an IOMMU
160so connected devices cannot access memory regions outside of what is
161allocated for them by drivers. When Linux is running on such system it
162automatically enables IOMMU if not enabled by the user already. These
163systems can be identified by reading ``1`` from
164``/sys/bus/thunderbolt/devices/domainX/iommu_dma_protection`` attribute.
165
166The driver does not do anything special in this case but because DMA
167protection is handled by the IOMMU, security levels (if set) are
168redundant. For this reason some systems ship with security level set to
169``none``. Other systems have security level set to ``user`` in order to
170support downgrade to older OS, so users who want to automatically
171authorize devices when IOMMU DMA protection is enabled can use the
172following ``udev`` rule::
173
174  ACTION=="add", SUBSYSTEM=="thunderbolt", ATTRS{iommu_dma_protection}=="1", ATTR{authorized}=="0", ATTR{authorized}="1"
175
176Upgrading NVM on Thunderbolt device or host
177-------------------------------------------
178Since most of the functionality is handled in firmware running on a
179host controller or a device, it is important that the firmware can be
180upgraded to the latest where possible bugs in it have been fixed.
181Typically OEMs provide this firmware from their support site.
182
183There is also a central site which has links where to download firmware
184for some machines:
185
186  `Thunderbolt Updates <https://thunderbolttechnology.net/updates>`_
187
188Before you upgrade firmware on a device or host, please make sure it is a
189suitable upgrade. Failing to do that may render the device (or host) in a
190state where it cannot be used properly anymore without special tools!
191
192Host NVM upgrade on Apple Macs is not supported.
193
194Once the NVM image has been downloaded, you need to plug in a
195Thunderbolt device so that the host controller appears. It does not
196matter which device is connected (unless you are upgrading NVM on a
197device - then you need to connect that particular device).
198
199Note an OEM-specific method to power the controller up ("force power") may
200be available for your system in which case there is no need to plug in a
201Thunderbolt device.
202
203After that we can write the firmware to the non-active parts of the NVM
204of the host or device. As an example here is how Intel NUC6i7KYK (Skull
205Canyon) Thunderbolt controller NVM is upgraded::
206
207  # dd if=KYK_TBT_FW_0018.bin of=/sys/bus/thunderbolt/devices/0-0/nvm_non_active0/nvmem
208
209Once the operation completes we can trigger NVM authentication and
210upgrade process as follows::
211
212  # echo 1 > /sys/bus/thunderbolt/devices/0-0/nvm_authenticate
213
214If no errors are returned, the host controller shortly disappears. Once
215it comes back the driver notices it and initiates a full power cycle.
216After a while the host controller appears again and this time it should
217be fully functional.
218
219We can verify that the new NVM firmware is active by running the following
220commands::
221
222  # cat /sys/bus/thunderbolt/devices/0-0/nvm_authenticate
223  0x0
224  # cat /sys/bus/thunderbolt/devices/0-0/nvm_version
225  18.0
226
227If ``nvm_authenticate`` contains anything other than 0x0 it is the error
228code from the last authentication cycle, which means the authentication
229of the NVM image failed.
230
231Note names of the NVMem devices ``nvm_activeN`` and ``nvm_non_activeN``
232depend on the order they are registered in the NVMem subsystem. N in
233the name is the identifier added by the NVMem subsystem.
234
235Upgrading NVM when host controller is in safe mode
236--------------------------------------------------
237If the existing NVM is not properly authenticated (or is missing) the
238host controller goes into safe mode which means that the only available
239functionality is flashing a new NVM image. When in this mode, reading
240``nvm_version`` fails with ``ENODATA`` and the device identification
241information is missing.
242
243To recover from this mode, one needs to flash a valid NVM image to the
244host controller in the same way it is done in the previous chapter.
245
246Networking over Thunderbolt cable
247---------------------------------
248Thunderbolt technology allows software communication between two hosts
249connected by a Thunderbolt cable.
250
251It is possible to tunnel any kind of traffic over a Thunderbolt link but
252currently we only support Apple ThunderboltIP protocol.
253
254If the other host is running Windows or macOS, the only thing you need to
255do is to connect a Thunderbolt cable between the two hosts; the
256``thunderbolt-net`` driver is loaded automatically. If the other host is
257also Linux you should load ``thunderbolt-net`` manually on one host (it
258does not matter which one)::
259
260  # modprobe thunderbolt-net
261
262This triggers module load on the other host automatically. If the driver
263is built-in to the kernel image, there is no need to do anything.
264
265The driver will create one virtual ethernet interface per Thunderbolt
266port which are named like ``thunderbolt0`` and so on. From this point
267you can either use standard userspace tools like ``ifconfig`` to
268configure the interface or let your GUI handle it automatically.
269
270Forcing power
271-------------
272Many OEMs include a method that can be used to force the power of a
273Thunderbolt controller to an "On" state even if nothing is connected.
274If supported by your machine this will be exposed by the WMI bus with
275a sysfs attribute called "force_power".
276
277For example the intel-wmi-thunderbolt driver exposes this attribute in:
278  /sys/bus/wmi/devices/86CCFD48-205E-4A77-9C48-2021CBEDE341/force_power
279
280  To force the power to on, write 1 to this attribute file.
281  To disable force power, write 0 to this attribute file.
282
283Note: it's currently not possible to query the force power state of a platform.
284