We only need to set up the device links when software connection manager
path is used. The firmware connection manager does not need them and if
they are present they may even cause problems.
Reviewed-by: Yehezkel Bernat <YehezkelShB@gmail.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
We should not dereference ->dual_link_port if it is NULL and lane bonding
is requested. For this reason move lane bonding configuration happen
inside the block where ->dual_link_port != NULL.
Fixes: 54509f5005 ("thunderbolt: Add KUnit tests for path walking")
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Yehezkel Bernat <YehezkelShB@gmail.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
If the NVM authentication fails immediately, like if the firmware
detects that the image is not valid for some reason, better to read the
status once and if set to non-zero fail the operation accordingly.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The same way we support these two operations for USB4 routers we can
extend the retimer NVM operations to support retimers also.
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>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
It may be useful if the actual NVM authentication can be delayed to be
run later, for instance when the user logs out. For this reason add a
new NVM operation (AUHENTICATE_ONLY) that just triggers the authentication
procedure over whatever was written to the NVM storage.
This is not supported with Thunderbolt 1-3 devices, though.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently these write ops are used for updating router firmware images
only. Moving to tb.h helps the retimers also to use the same ops.
Also add tb_ prefix to the enum while there.
Signed-off-by: Rajmohan Mani <rajmohan.mani@intel.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
With help from platform firmware (ACPI) it is possible to power on
retimers even when there is no USB4 link (e.g nothing is connected to
the USB4 ports). This allows us to bring the USB4 sideband up so that we
can access retimers and upgrade their NVM firmware.
If the platform has support for this, we expose two additional
attributes under USB4 ports: offline and rescan. These can be used to
bring the port offline, rescan for the retimers and put the port online
again. The retimer NVM upgrade itself works the same way than with cable
connected.
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>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When accessing retimers when there is no cable connected we are going to
need additional USB4 port operations. First the port needs to be put
into offline mode, and then the sideband channel transactions must be
enabled on the SBTX line. This adds support for these operations.
Signed-off-by: Rajmohan Mani <rajmohan.mani@intel.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Typically retimers can be accessed only when the USB4 link is up (e.g
there is a cable connected). However, sometimes it is useful to be able
to access retimers even if there is nothing connected to the USB4 port.
For instance we may still want to be able to upgrade the retimer NVM
firmware even if the user does not have any USB4 devices. This is
something that USB4 spec leaves to implementers.
In case of ACPI based systems, we can support this by providing a
special _DSM method under each USB4 port. This _DSM can be used to turn
on power to on-board retimers (and cycle it through different modes so
that the sideband becomes usable).
This patch adds support for this _DSM and makes the functionality
available to the rest of the driver through tb_acpi_power_[on|off]_retimers().
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>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Create devices for each USB4 port. This is needed when we add retimer
access when there is no device connected but may be useful for other
purposes too following what USB subsystem does. This exports a single
attribute "link" that shows the type of the USB4 link (or "none" if
there is no cable connected).
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The upstream port can be connected to any previous generation
Thunderbolt port so logging as "TBT" is more accurate than "TBT3.
No functional changes.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Intel Goshen Ridge reports wrong DP main credits in NVM 27 and earlier,
so add a quirk that fixes it. We also need to expand the quirk table to
match on hardware vendor/device IDs too.
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>
Once lane bonding has been enabled (or disabled) both lane adapters may
update their total credits accordingly. For this reason re-read the port
credits after lane bonding has been enabled or disabled.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
USB4 routers must expose their preferred credit (buffer) allocation
information through router operation. This information tells the
connection manager how the router prefers its buffers to be allocated to
get the expected bandwidth for the supported protocols.
Read this information and store it as part of struct tb_switch for each
USB4 router.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
It may take some time until the two lanes enter bonded state so poll for
the link width to match what is expected before going forward. This ensures
the link is in expected state before we start establishing paths through
it.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
With the USB4 buffer allocation the number of credits (and non-flow
credits) may be different depending on the router buffer allocation
preferences. To allow this move the nfc_credits field to struct
tb_path_hop.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
The function does not modify the object in any way so make the parameter
const to reflect this.
No functional changes intended.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
The USB4 Configuration Manager guide suggests that the USB4 port wakes
are configured in a certain way, like that when the port is configured
the wake-on-connect should not be set and so forth, so align the driver
with this.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Latest USB4 spec added a new wake bit for DisplayPort so add this to the
driver when runtime suspending. This way wake up the domain when a new
monitor is plugged in to any of the device routers.
Also do the same for pre-USB4 devices through the link controller
registers as documented in chapter 13 of the USB4 spec.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Now that we have generic functionality available in nvm.c make the DMA
port code call it instead of duplicating the functionality.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
We do this for Thunderbolt 2/3 devices through DMA port, USB4 devices
and retimers pretty much the same way. Only the actual block read/write
is different. For this reason split out the NVM read/write functions
from usb4.c to nvm.c and make USB4 device code call these when needed.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Add new device known to support self-authenticate on disconnect.
Signed-off-by: Crag Wang <crag.wang@dell.com>
Reviewed-by: Mario Limonciello <mario.limonciello@outlook.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Up to 64 bytes of data can be read from NVM in one go.
Read address must be dword aligned. Data is read into a local buffer.
If caller asks to read data starting at an unaligned address then full
dword is anyway read from NVM into a local buffer. Data is then copied
from the local buffer starting at the unaligned offset to the caller
buffer.
In cases where asked data length + unaligned offset is over 64 bytes
we need to make sure we don't read past the 64 bytes in the local
buffer when copying to caller buffer, and make sure that we don't
skip copying unaligned offset bytes from local buffer anymore after
the first round of 64 byte NVM data read.
Fixes: b04079837b ("thunderbolt: Add initial support for USB4")
Cc: stable@vger.kernel.org
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Up to 64 bytes of data can be read from NVM in one go. Read address
must be dword aligned. Data is read into a local buffer.
If caller asks to read data starting at an unaligned address then full
dword is anyway read from NVM into a local buffer. Data is then copied
from the local buffer starting at the unaligned offset to the caller
buffer.
In cases where asked data length + unaligned offset is over 64 bytes
we need to make sure we don't read past the 64 bytes in the local
buffer when copying to caller buffer, and make sure that we don't
skip copying unaligned offset bytes from local buffer anymore after
the first round of 64 byte NVM data read.
Fixes: 3e13676862 ("thunderbolt: Add support for DMA configuration based mailbox")
Cc: stable@vger.kernel.org
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
This includes following Thunderbolt/USB4 changes for v5.13 merge window:
* Debugfs improvements
* Align the inter-domain (peer-to-peer) support with the USB4
inter-domain spec for better interoperability
* Add support for USB4 DROM and the new product descriptor
* More KUnit tests
* Detailed uevent for routers
* Few miscellaneous improvements
All these have been in linux-next without reported issues.
-----BEGIN PGP SIGNATURE-----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=RzJl
-----END PGP SIGNATURE-----
Merge tag 'thunderbolt-for-v5.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/westeri/thunderbolt into usb-next
Mika writes:
thunderbolt: Changes for v5.13 merge window
This includes following Thunderbolt/USB4 changes for v5.13 merge window:
* Debugfs improvements
* Align the inter-domain (peer-to-peer) support with the USB4
inter-domain spec for better interoperability
* Add support for USB4 DROM and the new product descriptor
* More KUnit tests
* Detailed uevent for routers
* Few miscellaneous improvements
All these have been in linux-next without reported issues.
* tag 'thunderbolt-for-v5.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/westeri/thunderbolt: (24 commits)
thunderbolt: Hide authorized attribute if router does not support PCIe tunnels
thunderbolt: Add details to router uevent
thunderbolt: Unlock on error path in tb_domain_add()
thunderbolt: Add support for USB4 DROM
thunderbolt: Check quirks in tb_switch_add()
thunderbolt: Add KUnit tests for DMA tunnels
thunderbolt: Add KUnit tests for XDomain properties
net: thunderbolt: Align the driver to the USB4 networking spec
thunderbolt: Allow multiple DMA tunnels over a single XDomain connection
thunderbolt: Drop unused tb_port_set_initial_credits()
thunderbolt: Use dedicated flow control for DMA tunnels
thunderbolt: Add support for maxhopid XDomain property
thunderbolt: Add tb_property_copy_dir()
thunderbolt: Align XDomain protocol timeouts with the spec
thunderbolt: Use pseudo-random number as initial property block generation
thunderbolt: Do not re-establish XDomain DMA paths automatically
thunderbolt: Add more logging to XDomain connections
Documentation / thunderbolt: Drop speed/lanes entries for XDomain
thunderbolt: Decrease control channel timeout for software connection manager
thunderbolt: Do not pass timeout for tb_cfg_reset()
...
With USB4 devices PCIe tunneling is optional so for device routers
without PCIe upstream adapter it does not make much sense to expose the
authorized attribute. For this reason hide it if PCIe tunneling is not
supported by the device router.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Expose two environment variables for routers as part of the initial
uevent:
USB4_VERSION=1.0
USB4_TYPE=host|device|hub
Userspace can use this information to expose more details about each
connected device. Only USB4 devices have USB4_VERSION but all devices
have USB4_TYPE.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This array uses 1-based indexing so it corrupts memory one element
beyond of the array. Fix it by making the array one element larger.
Fixes: dacb12877d ("thunderbolt: Add support for on-board retimers")
Cc: stable@vger.kernel.org
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
After the device_register() succeeds, then the correct way to clean up
is to call device_unregister(). The unregister calls both device_del()
and device_put(). Since this code was only device_del() it results in
a memory leak.
Fixes: dacb12877d ("thunderbolt: Add support for on-board retimers")
Cc: stable@vger.kernel.org
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
We accidentally deleted this unlock on the error path. Undelete it.
Fixes: 7f0a34d790 ("thunderbolt: Decrease control channel timeout for software connection manager")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
USB4 router DROM differs sligthly from Thunderbolt 1-3 DROM. For
instance it does not include UID and CRC8 in the header section, and it
has product descriptor genereric entry to describe the product IDs and
related information. If the "Version" field in the DROM header section
reads 3 it means the router only has USB4 DROM and if it reads 1 it
means the router supports TBT3 compatible DROM.
For this reason, update the DROM parsing code to support "pure" USB4
DROMs too.
While there drop the extra empty line at the end of tb_drom_read().
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>
The USB4 inter-domain service spec recommends using dedicated flow
control scheme so update the driver accordingly.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
USB4 inter-domain spec mandates that the compatible hosts expose a new
property "maxhopid" that tells the connection manager on the other side
what is the maximum supported input HopID over the connection. Since
this is depend on the lane adapter the cable is connected it needs to be
filled in dynamically.
For this reason we take a copy of the global properties and fill then
for each XDomain connection upon first connect, and then keep updating
it if the generation changes as services are being added/removed. We
also take advantage of this copy to fill in the hostname.
We also expose this maxhopid as an attribute under each XDomain device.
While there drop kernel-doc entry for property_lock which seems to be
left there when the structure was originally introduced.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
This function takes a deep copy of the properties. We need this in order
to support more dynamic properties per XDomain connection as required by
the USB4 inter-domain service spec.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
The USB4 inter-domain service spec has slightly different recommended
timeouts for the XDomain protocol so align the driver with those.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
As recommended by USB4 inter-domain service spec use pseudo-random value
instead of zero as initial XDomain property block generation value.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
This step is actually not needed. The service drivers themselves will
handle this once they have negotiated the service up and running again
with the remote side. Also dropping this makes it easier to add support
for multiple DMA tunnels over a single XDomain connection.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Currently the driver is pretty quiet when another host is connected
which makes debugging possible issues harder. For this reason add more
logging on debug level that can be turned on as needed.
While there log the host-to-host connection on info level analogous to
routers and retimers.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
When the firmware connection manager is not proxying between the
software and the hardware we can decrease the timeout for control
packets significantly. The USB4 spec recommends 10 ms +- 1 ms but we use
slightly larger value (100 ms) which is recommendation from Intel
Thunderbolt firmware folks. When firmware connection manager is running
then we keep using the existing 5000 ms.
To implement this we move the control channel allocation to
tb_domain_alloc(), and pass the timeout from that function to the
tb_ctl_alloc(). Then make both connection manager implementations pass
the timeout when they alloc the domain structure.
While there update kernel-doc of struct tb_ctl to match the reality.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
There is only one user for this function and it passes the default
timeout to it anyway, so remove the parameter completely. This is also
needed in the subsequent patch where we allow connection manager
implementations to use different timeout for non-raw control channel
messages.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
In most cases the response packet is lost because the router in question
was disconnected by the user. Resending the control packet in that case
just adds unnecessary delays, so disable that for intra-domain control
packets. For inter-domain (XDomain) packets we continue retrying.
This also aligns the driver better what the Intel connection manager
firmware is doing.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Currently, when first failure occurs while reading of the block,
we stop reading the block and jump to the next capability.
This doesn't cover the case of block with "holes" of inaccessible
dwords, followed by accessible dwords.
This patch address this problem.
In case of failure while reading the complete block in one transaction,
(because of one or more dwords is inaccessible), we read the remaining
dwords of the block dword-by-dword, one dword per transaction,
till the end of the block.
By doing this, we handle the case of block with "holes" of inaccessible
dwords, followed by accessible dwords. The accessible dwords are shown
with the fields: <offset> <relative_offset> <cap_id> <vs_cap_id> <value>
E.g.:
0x01eb 236 0x05 0x06 0x0000d166
While the inaccesible dwords are shown as: <offset> <not accessible>
E.g.:
0x01ed <not accessible>
Signed-off-by: Gil Fine <gil.fine@intel.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
If the driver is unbound and then bound back it goes over the topology
and figure out the existing tunnels. However, if it finds DP tunnel it
should make sure the domain does not runtime suspend as otherwise it
will tear down the DP tunnel unexpectedly.
Fixes: 6ac6faee5d ("thunderbolt: Add runtime PM for Software CM")
Cc: stable@vger.kernel.org
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
