The USB4 spec defines an optional feature that allows the connection
manager to negotiate with the graphics through DPCD registers changes in
the bandwidth allocation dynamically. This is referred as "bandwidth
allocation mode" in the spec. The connection manager uses DP IN adapters
registers to communicate with the graphics, and also gets notifications
from these adapters when the graphics wants to change the bandwidth
allocation. Both the connection manager and the graphics driver needs to
support this.
We check if the DP IN adapter supports this and if it does enable it
before establishing a DP tunnel. Then we react on DP_BW notifications
coming from the DP IN adapter and update the bandwidth allocation
accordingly (within the maximum common capabilities the DP IN/OUT
support).
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Brad reported that on Apple hardware with Light Ridge or Falcon Ridge
controller, plugging in a chain of Thunderbolt displays (Light Ridge
based controllers) causes all kinds of tearing and flickering. The
reason for this is that on Thunderbolt 1 hardware there is no lane
bonding so we have two independent 10 Gb/s lanes, and currently Linux
tunnels both displays through the lane 1. This makes the displays to
share the 10 Gb/s bandwidth which may not be enough for higher
resolutions.
For this reason make the second tunnel go through the lane 0 instead.
This seems to match what the macOS connection manager is also doing.
Reported-by: Brad Campbell <lists2009@fnarfbargle.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Tested-by: Brad Campbell <lists2009@fnarfbargle.com>
If the boot firmware implements connection manager of its own it may not
create the paths in the same way or order we do. For example it may
create first PCIe tunnel and then USB3 tunnel. When we restore our
tunnels (first de-activating them) we may be doing that over completely
different tunnels and that leaves them possibly non-functional. For this
reason we re-use the tunnel discovery functionality and find out all the
existing tunnels, and tear them down. Once that is done we can restore
our tunnels.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
The USB4 Connection Manager guide provides detailed information how the
USB4 router buffer (credit) allocation information should be used by the
connection manager when it allocates buffers for different paths. This
patch implements it for Linux. For USB 3.x and DisplayPort we use
directly the router preferences. The rest of the buffer space is then
used for PCIe and DMA (peer-to-peer, XDomain) traffic. DMA tunnels
require at least one buffer and PCIe six, so if there is not enough
buffers we fail the tunnel creation.
For the legacy Thunderbolt 1-3 devices we use the existing hard-coded
scheme except for DMA where we use the values suggested by the USB4 spec
chapter 13.
Co-developed-by: Gil Fine <gil.fine@intel.com>
Signed-off-by: Gil Fine <gil.fine@intel.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Currently we have had an artificial limitation of a single DMA tunnel
per XDomain connection. However, hardware wise there is no such limit
and software based connection manager can take advantage of all the DMA
rings available on the host to establish tunnels.
For this reason make the tb_xdomain_[enable|disable]_paths() to take the
DMA ring and HopID as parameter instead of storing them in the struct
tb_xdomain. We also add API functions to allocate input and output
HopIDs of the XDomain connection that the service drivers can use
instead of hard-coding.
Also convert the two existing service drivers over to this API.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
USB3 supports both isochronous and non-isochronous traffic. The former
requires guaranteed bandwidth and can take up to 90% of the total
bandwidth. With USB4 USB3 is tunneled over USB4 fabric which means that
we need to make sure there is enough bandwidth allocated for the USB3
tunnels in addition to DisplayPort tunnels.
Whereas DisplayPort bandwidth management is static and done before the
DP tunnel is established, the USB3 bandwidth management is dynamic and
allows increasing and decreasing the allocated bandwidth according to
what is currently consumed. This is done through host router USB3
downstream adapter registers.
This adds USB3 bandwidth management to the software connection manager
so that we always try to allocate maximum bandwidth for DP tunnels and
what is left is allocated for USB3.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Whereas DisplayPort bandwidth is consumed only in one direction (from DP
IN adapter to DP OUT adapter), USB3 adds separate bandwidth for both
upstream and downstream directions.
For this reason extend the tunnel consumed bandwidth routines to support
both directions and implement this for DP.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
USB4 added a capability to tunnel USB 3.x protocol over the USB4
fabric. USB4 device routers may include integrated SuperSpeed HUB or a
function or both. USB tunneling follows PCIe so that the tunnel is
created between the parent and the child router from USB3 downstream
adapter port to USB3 upstream adapter port over a single USB4 link.
This adds support for USB 3.x tunneling and also capability to discover
existing USB 3.x tunnels (for example created by connection manager in
boot firmware).
Signed-off-by: Rajmohan Mani <rajmohan.mani@intel.com>
Co-developed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Link: https://lore.kernel.org/r/20191217123345.31850-9-mika.westerberg@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Titan Ridge supports Display Port 1.4 which adds HBR3 (High Bit Rate)
rates that may be up to 8.1 Gb/s over 4 lanes. This translates to
effective data bandwidth of 25.92 Gb/s (as 8/10 encoding is removed by
the DP adapters when going over Thunderbolt fabric). If another high
rate monitor is connected we may need to reduce the bandwidth it
consumes so that it fits into the total 40 Gb/s available on the
Thunderbolt fabric.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
In addition to PCIe and Display Port tunnels it is also possible to
create tunnels that forward DMA traffic from the host interface adapter
(NHI) to a NULL port that is connected to another domain through a
Thunderbolt cable. These tunnels can be used to carry software messages
such as networking packets.
To support this we introduce another tunnel type (TB_TUNNEL_DMA) that
supports paths from NHI to NULL port and back.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Display Port tunnels are somewhat more complex than PCIe tunnels as it
requires 3 tunnels (AUX Rx/Tx and Video). In addition we are not
supposed to create the tunnels immediately when a DP OUT is enumerated.
Instead we need to wait until we get hotplug event to that adapter port
or check if the port has HPD set before tunnels can be established. This
adds Display Port tunneling support to the software connection manager.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
In Apple Macs the boot firmware (EFI) connects all devices automatically
when the system is started, before it hands over to the OS. Instead of
ignoring we discover all those PCIe tunnels and record them using our
internal structures, just like we do when a device is connected after
the OS is already up.
By doing this we can properly tear down tunnels when devices are
disconnected. Also this allows us to resume the existing tunnels after
system suspend/resume cycle.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
To be able to tunnel non-PCIe traffic, separate tunnel functionality
into generic and PCIe specific parts. Rename struct tb_pci_tunnel to
tb_tunnel, and make it hold an array of paths instead of just two.
Update all the tunneling functions to take this structure as parameter.
We also move tb_pci_port_active() to switch.c (and rename it) where we
will be keeping all port and switch related functions.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
In order to tunnel non-PCIe traffic as well rename tunnel_pci.[ch] to
tunnel.[ch] to reflect this fact. No functional changes.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
