[ Upstream commit 8607edcd17 ]
A recent change to hide USB3 root hubs of USB2-only controllers broke
registration of USB2 root hubs - allow_single_roothub is set too late,
and by this time xhci_run() has already deferred root hub registration
until after the shared HCD is added, which will never happen.
This makes such controllers unusable, but testers didn't notice since
they were only bothered by warnings about empty USB3 root hubs. The bug
causes problems to other people who actually use such HCs and I was
able to confirm it on an ordinary HC by patching to ignore USB3 ports.
Setting allow_single_roothub during early setup fixes things.
Reported-by: Arisa Snowbell <arisa.snowbell@gmail.com>
Closes: https://lore.kernel.org/linux-usb/CABpa4MA9unucCoKtSdzJyOLjHNVy+Cwgz5AnAxPkKw6vuox1Nw@mail.gmail.com/
Reported-by: Michal Kubecek <mkubecek@suse.cz>
Closes: https://lore.kernel.org/linux-usb/lnb5bum7dnzkn3fc7gq6hwigslebo7o4ccflcvsc3lvdgnu7el@fvqpobbdoapl/
Fixes: 719de070f7 ("usb: xhci-pci: add support for hosts with zero USB3 ports")
Tested-by: Arisa Snowbell <arisa.snowbell@gmail.com>
Tested-by: Michal Kubecek <mkubecek@suse.cz>
Suggested-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Michal Pecio <michal.pecio@gmail.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 719de070f7 ]
Add xhci support for PCI hosts that have zero USB3 ports.
Avoid creating a shared Host Controller Driver (HCD) when there is only
one root hub. Additionally, all references to 'xhci->shared_hcd' are now
checked before use.
Only xhci-pci.c requires modification to accommodate this change, as the
xhci core already supports configurations with zero USB3 ports. This
capability was introduced when xHCI Platform and MediaTek added support
for zero USB3 ports.
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=220181
Tested-by: Nick Nielsen <nick.kainielsen@free.fr>
Tested-by: grm1 <grm1@mailbox.org>
Signed-off-by: Niklas Neronin <niklas.neronin@linux.intel.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250917210726.97100-4-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 41cf11946b ]
Allow autosuspend to be used by xhci plat device. For Qualcomm SoCs,
when in host mode, it is intended that the controller goes to suspend
state to save power and wait for interrupts from connected peripheral
to wake it up. This is particularly used in cases where a HID or Audio
device is connected. In such scenarios, the usb controller can enter
auto suspend and resume action after getting interrupts from the
connected device.
Signed-off-by: Krishna Kurapati <krishna.kurapati@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20250916120436.3617598-1-krishna.kurapati@oss.qualcomm.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit f3d12ec847 upstream.
DbC may add 1024 bogus bytes to the beginneing of the receiving endpoint
if DbC hw triggers a STALL event before any Transfer Blocks (TRBs) for
incoming data are queued, but driver handles the event after it queued
the TRBs.
This is possible as xHCI DbC hardware may trigger spurious STALL transfer
events even if endpoint is empty. The STALL event contains a pointer
to the stalled TRB, and "remaining" untransferred data length.
As there are no TRBs queued yet the STALL event will just point to first
TRB position of the empty ring, with '0' bytes remaining untransferred.
DbC driver is polling for events, and may not handle the STALL event
before /dev/ttyDBC0 is opened and incoming data TRBs are queued.
The DbC event handler will now assume the first queued TRB (length 1024)
has stalled with '0' bytes remaining untransferred, and copies the data
This race situation can be practically mitigated by making sure the event
handler handles all pending transfer events when DbC reaches configured
state, and only then create dev/ttyDbC0, and start queueing transfers.
The event handler can this way detect the STALL events on empty rings
and discard them before any transfers are queued.
This does in practice solve the issue, but still leaves a small possible
gap for the race to trigger.
We still need a way to distinguish spurious STALLs on empty rings with '0'
bytes remaing, from actual STALL events with all bytes transmitted.
Cc: stable <stable@kernel.org>
Fixes: dfba2174dc ("usb: xhci: Add DbC support in xHCI driver")
Tested-by: Łukasz Bartosik <ukaszb@chromium.org>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 2bbd38fcd2 upstream.
DbC is currently only enabled back if it's in configured state during
suspend.
If system is suspended after DbC is enabled, but before the device is
properly enumerated by the host, then DbC would not be enabled back in
resume.
Always enable DbC back in resume if it's suspended in enabled,
connected, or configured state
Cc: stable <stable@kernel.org>
Fixes: dfba2174dc ("usb: xhci: Add DbC support in xHCI driver")
Tested-by: Łukasz Bartosik <ukaszb@chromium.org>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 08fa726e66 ]
This reverts commit 28a76fcc4c.
No actual HW bugs are known where Endpoint Context shows Running state
but Stop Endpoint fails repeatedly with Context State Error and leaves
the endpoint state unchanged. Stop Endpoint retries on Running EPs have
been performed since early 2021 with no such issues reported so far.
Trying to handle this hypothetical case brings a more realistic danger:
if Stop Endpoint fails on an endpoint which hasn't yet started after a
doorbell ring and enough latency occurs before this completion event is
handled, the driver may time out and begin removing cancelled TDs from
a running endpoint, even though one more retry would stop it reliably.
Such high latency is rare but not impossible, and removing TDs from a
running endpoint can cause more damage than not giving back a cancelled
URB (which wasn't happening anyway). So err on the side of caution and
revert to the old policy of always retrying if the EP appears running.
[Remove stable tag as we are dealing with theoretical cases -Mathias]
Fixes: 28a76fcc4c ("usb: xhci: Avoid Stop Endpoint retry loop if the endpoint seems Running")
Signed-off-by: Michal Pecio <michal.pecio@gmail.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250917210726.97100-2-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 186e8f2bdb ]
The kthread_run() function returns error pointers so the
max3421_hcd->spi_thread pointer can be either error pointers or NULL.
Check for both before dereferencing it.
Fixes: 05dfa5c9bc ("usb: host: max3421-hcd: fix "spi_rd8" uses dynamic stack allocation warning")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Link: https://lore.kernel.org/r/aJTMVAPtRe5H6jug@stanley.mountain
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
Revert commit 9b28ef1e4c which is upstream
commit e1b0fa8639
It causes regression in 6.12.49 stable, no issues in upstream.
Commit 9b28ef1e4c ("usb: xhci: remove option to change a default
ring's TRB cycle bit") introduced a regression in 6.12.49 stable kernel.
The original commit was never intended for stable kernels, but was added
as a dependency for commit a5c98e8b13 ("xhci: dbc: Fix full DbC
transfer ring after several reconnects").
Since this commit is more of an optimization, revert it and solve the
dependecy by modifying one line in xhci_dbc_ring_init(). Specifically,
commit a5c98e8b13 ("xhci: dbc: Fix full DbC transfer ring after
several reconnects") moved function call xhci_initialize_ring_info()
into a separate function. To resolve the dependency, the arguments for
this function call are also reverted.
Closes: https://lore.kernel.org/stable/01b8c8de46251cfaad1329a46b7e3738@stwm.de/
Tested-by: Wolfgang Walter <linux@stwm.de>
Cc: stable@vger.kernel.org # v6.12.49
Signed-off-by: Niklas Neronin <niklas.neronin@linux.intel.com>
Acked-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit a5c98e8b13 ]
Pending requests will be flushed on disconnect, and the corresponding
TRBs will be turned into No-op TRBs, which are ignored by the xHC
controller once it starts processing the ring.
If the USB debug cable repeatedly disconnects before ring is started
then the ring will eventually be filled with No-op TRBs.
No new transfers can be queued when the ring is full, and driver will
print the following error message:
"xhci_hcd 0000:00:14.0: failed to queue trbs"
This is a normal case for 'in' transfers where TRBs are always enqueued
in advance, ready to take on incoming data. If no data arrives, and
device is disconnected, then ring dequeue will remain at beginning of
the ring while enqueue points to first free TRB after last cancelled
No-op TRB.
s
Solve this by reinitializing the rings when the debug cable disconnects
and DbC is leaving the configured state.
Clear the whole ring buffer and set enqueue and dequeue to the beginning
of ring, and set cycle bit to its initial state.
Cc: stable@vger.kernel.org
Fixes: dfba2174dc ("usb: xhci: Add DbC support in xHCI driver")
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250902105306.877476-3-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 220a0ffde0 ]
Decouple allocation of endpoint ring buffer from initialization
of the buffer, and initialization of endpoint context parts from
from the rest of the contexts.
It allows driver to clear up and reinitialize endpoint rings
after disconnect without reallocating everything.
This is a prerequisite for the next patch that prevents the transfer
ring from filling up with cancelled (no-op) TRBs if a debug cable is
reconnected several times without transferring anything.
Cc: stable@vger.kernel.org
Fixes: dfba2174dc ("usb: xhci: Add DbC support in xHCI driver")
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250902105306.877476-2-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Stable-dep-of: a5c98e8b13 ("xhci: dbc: Fix full DbC transfer ring after several reconnects")
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit e1b0fa8639 ]
The TRB cycle bit indicates TRB ownership by the Host Controller (HC) or
Host Controller Driver (HCD). New rings are initialized with 'cycle_state'
equal to one, and all its TRBs' cycle bits are set to zero. When handling
ring expansion, set the source ring cycle bits to the same value as the
destination ring.
Move the cycle bit setting from xhci_segment_alloc() to xhci_link_rings(),
and remove the 'cycle_state' argument from xhci_initialize_ring_info().
The xhci_segment_alloc() function uses kzalloc_node() to allocate segments,
ensuring that all TRB cycle bits are initialized to zero.
Signed-off-by: Niklas Neronin <niklas.neronin@linux.intel.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20241106101459.775897-12-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Stable-dep-of: a5c98e8b13 ("xhci: dbc: Fix full DbC transfer ring after several reconnects")
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 3f970bd06c ]
Add macro to streamline and standardize the iteration over ring
segment list.
xhci_for_each_ring_seg(): Iterates over the entire ring segment list.
The xhci_free_segments_for_ring() function's while loop has not been
updated to use the new macro. This function has some underlying issues,
and as a result, it will be handled separately in a future patch.
Suggested-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Niklas Neronin <niklas.neronin@linux.intel.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20241106101459.775897-11-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Stable-dep-of: a5c98e8b13 ("xhci: dbc: Fix full DbC transfer ring after several reconnects")
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit edcbe06453 upstream.
Suspend-resume cycle test revealed a memory leak in 6.17-rc3
Turns out the slot_id race fix changes accidentally ends up calling
xhci_free_virt_device() with an incorrect vdev parameter.
The vdev variable was reused for temporary purposes right before calling
xhci_free_virt_device().
Fix this by passing the correct vdev parameter.
The slot_id race fix that caused this regression was targeted for stable,
so this needs to be applied there as well.
Fixes: 2eb0337615 ("usb: xhci: Fix slot_id resource race conflict")
Reported-by: David Wang <00107082@163.com>
Closes: https://lore.kernel.org/linux-usb/20250829181354.4450-1-00107082@163.com
Suggested-by: Michal Pecio <michal.pecio@gmail.com>
Suggested-by: David Wang <00107082@163.com>
Cc: stable@vger.kernel.org
Tested-by: David Wang <00107082@163.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250902105306.877476-4-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 2eb0337615 upstream.
xHC controller may immediately reuse a slot_id after it's disabled,
giving it to a new enumerating device before the xhci driver freed
all resources related to the disabled device.
In such a scenario, device-A with slot_id equal to 1 is disconnecting
while device-B is enumerating, device-B will fail to enumerate in the
follow sequence.
1.[device-A] send disable slot command
2.[device-B] send enable slot command
3.[device-A] disable slot command completed and wakeup waiting thread
4.[device-B] enable slot command completed with slot_id equal to 1 and
wakeup waiting thread
5.[device-B] driver checks that slot_id is still in use (by device-A) in
xhci_alloc_virt_device, and fail to enumerate due to this
conflict
6.[device-A] xhci->devs[slot_id] set to NULL in xhci_free_virt_device
To fix driver's slot_id resources conflict, clear xhci->devs[slot_id] and
xhci->dcbba->dev_context_ptrs[slot_id] pointers in the interrupt context
when disable slot command completes successfully. Simultaneously, adjust
function xhci_free_virt_device to accurately handle device release.
[minor smatch warning and commit message fix -Mathias]
Cc: stable@vger.kernel.org
Fixes: 7faac1953e ("xhci: avoid race between disable slot command and host runtime suspend")
Signed-off-by: Weitao Wang <WeitaoWang-oc@zhaoxin.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250819125844.2042452-2-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit f9420f4757 upstream.
Increase the External ROM access timeouts to prevent failures during
programming of External SPI EEPROM chips. The current timeouts are
too short for some SPI EEPROMs used with uPD720201 controllers.
The current timeout for Chip Erase in renesas_rom_erase() is 100 ms ,
the current timeout for Sector Erase issued by the controller before
Page Program in renesas_fw_download_image() is also 100 ms. Neither
timeout is sufficient for e.g. the Macronix MX25L5121E or MX25V5126F.
MX25L5121E reference manual [1] page 35 section "ERASE AND PROGRAMMING
PERFORMANCE" and page 23 section "Table 8. AC CHARACTERISTICS (Temperature
= 0°C to 70°C for Commercial grade, VCC = 2.7V ~ 3.6V)" row "tCE" indicate
that the maximum time required for Chip Erase opcode to complete is 2 s,
and for Sector Erase it is 300 ms .
MX25V5126F reference manual [2] page 47 section "13. ERASE AND PROGRAMMING
PERFORMANCE (2.3V - 3.6V)" and page 42 section "Table 8. AC CHARACTERISTICS
(Temperature = -40°C to 85°C for Industrial grade, VCC = 2.3V - 3.6V)" row
"tCE" indicate that the maximum time required for Chip Erase opcode to
complete is 3.2 s, and for Sector Erase it is 400 ms .
Update the timeouts such, that Chip Erase timeout is set to 5 seconds,
and Sector Erase timeout is set to 500 ms. Such lengthy timeouts ought
to be sufficient for majority of SPI EEPROM chips.
[1] https://www.macronix.com/Lists/Datasheet/Attachments/8634/MX25L5121E,%203V,%20512Kb,%20v1.3.pdf
[2] https://www.macronix.com/Lists/Datasheet/Attachments/8750/MX25V5126F,%202.5V,%20512Kb,%20v1.1.pdf
Fixes: 2478be82de ("usb: renesas-xhci: Add ROM loader for uPD720201")
Cc: stable <stable@kernel.org>
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Link: https://lore.kernel.org/r/20250802225526.25431-1-marek.vasut+renesas@mailbox.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 4b9c60e440 ]
When a USB4 dock is unplugged from a system it won't respond to ring
events. The PCI core handles the surprise removal event and notifies
all PCI drivers. The XHCI PCI driver sets a flag that the device is
being removed as well.
When that flag is set don't show messages in the cleanup path for
marking the controller dead.
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Acked-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250717073107.488599-2-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit f72b9aa821 ]
There is a subtle contradiction between sections of the xHCI 1.2 spec
regarding the initialization of Input Endpoint Context fields. Section
4.8.2 ("Endpoint Context Initialization") states that all fields should
be initialized to 0. However, Section 6.2.3 ("Endpoint Context", p.453)
specifies that the Average TRB Length (avg_trb_len) field shall be
greater than 0, and explicitly notes (p.454): "Software shall set
Average TRB Length to '8' for control endpoints."
Strictly setting all fields to 0 during initialization conflicts with
the specific recommendation for control endpoints. In practice, setting
avg_trb_len = 0 is not meaningful for the hardware/firmware, as the
value is used for bandwidth calculation.
Motivation: Our company is developing a custom Virtual xHC hardware
platform that strictly follows the xHCI spec and its recommendations.
During validation, we observed that enumeration fails and a parameter
error (TRB Completion Code = 5) is reported if avg_trb_len for EP0 is
not set to 8 as recommended by Section 6.2.3. This demonstrates the
importance of assigning a meaningful, non-zero value to avg_trb_len,
even in virtualized or emulated environments.
This patch explicitly sets avg_trb_len to 8 for EP0 in
xhci_setup_addressable_virt_dev(), as recommended in Section 6.2.3, to
prevent potential issues with xHCI host controllers that enforce the
spec strictly.
Link: https://bugzilla.kernel.org/show_bug.cgi?id=220033
Signed-off-by: Jay Chen <shawn2000100@gmail.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250717073107.488599-4-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 65fc0fc137 ]
When a USB4 dock is unplugged from a system it won't respond to ring
events. The PCI core handles the surprise removal event and notifies
all PCI drivers. The XHCI PCI driver sets a flag that the device is
being removed, and when the device stops responding a flag is also
added to indicate it's dying.
When that flag is set don't bother to show warnings about a missing
controller.
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Acked-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250717073107.488599-3-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 7919407eca ]
When encounters some errors like these:
xhci_hcd 0000:4a:00.2: xHCI dying or halted, can't queue_command
xhci_hcd 0000:4a:00.2: FIXME: allocate a command ring segment
usb usb5-port6: couldn't allocate usb_device
It's hard to know whether xhc_state is dying or halted. So it's better
to print xhc_state's value which can help locate the resaon of the bug.
Signed-off-by: Su Hui <suhui@nfschina.com>
Link: https://lore.kernel.org/r/20250725060117.1773770-1-suhui@nfschina.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit d9e496a9fb ]
The variable `of_match` was incorrectly declared as a `bool`.
It is assigned the return value of of_match_device(), which is a pointer of
type `const struct of_device_id *`.
Fixes: 16b7e0cccb ("USB: xhci-plat: fix legacy PHY double init")
Signed-off-by: Seungjin Bae <eeodqql09@gmail.com>
Link: https://lore.kernel.org/r/20250619055746.176112-2-eeodqql09@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit efe3e3ae5a upstream.
Flush dbc requests when dbc is stopped and transfer rings are freed.
Failure to flush them lead to leaking memory and dbc completing odd
requests after resuming from suspend, leading to error messages such as:
[ 95.344392] xhci_hcd 0000:00:0d.0: no matched request
Cc: stable <stable@kernel.org>
Fixes: dfba2174dc ("usb: xhci: Add DbC support in xHCI driver")
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250627144127.3889714-5-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 2b857d69a5 upstream.
When /dev/ttyDBC0 device is created then by default ECHO flag
is set for the terminal device. However if data arrives from
a peer before application using /dev/ttyDBC0 applies its set
of terminal flags then the arriving data will be echoed which
might not be desired behavior.
Fixes: 4521f16139 ("xhci: dbctty: split dbc tty driver registration and unregistration functions.")
Cc: stable <stable@kernel.org>
Signed-off-by: Łukasz Bartosik <ukaszb@chromium.org>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/stable/20250610111802.18742-1-ukaszb%40chromium.org
Link: https://lore.kernel.org/r/20250627144127.3889714-4-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit cbc889ab01 upstream.
During the High-Speed Isochronous Audio transfers, xHCI
controller on certain AMD platforms experiences momentary data
loss. This results in Missed Service Errors (MSE) being
generated by the xHCI.
The root cause of the MSE is attributed to the ISOC OUT endpoint
being omitted from scheduling. This can happen when an IN
endpoint with a 64ms service interval either is pre-scheduled
prior to the ISOC OUT endpoint or the interval of the ISOC OUT
endpoint is shorter than that of the IN endpoint. Consequently,
the OUT service is neglected when an IN endpoint with a service
interval exceeding 32ms is scheduled concurrently (every 64ms in
this scenario).
This issue is particularly seen on certain older AMD platforms.
To mitigate this problem, it is recommended to adjust the service
interval of the IN endpoint to not exceed 32ms (interval 8). This
adjustment ensures that the OUT endpoint will not be bypassed,
even if a smaller interval value is utilized.
Cc: stable <stable@kernel.org>
Signed-off-by: Raju Rangoju <Raju.Rangoju@amd.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250627144127.3889714-2-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 7aed15379d upstream.
This reverts commit 6ccb83d6c4.
Commit 6ccb83d6c4 ("usb: xhci: Implement xhci_handshake_check_state()
helper") was introduced to workaround watchdog timeout issues on some
platforms, allowing xhci_reset() to bail out early without waiting
for the reset to complete.
Skipping the xhci handshake during a reset is a dangerous move. The
xhci specification explicitly states that certain registers cannot
be accessed during reset in section 5.4.1 USB Command Register (USBCMD),
Host Controller Reset (HCRST) field:
"This bit is cleared to '0' by the Host Controller when the reset
process is complete. Software cannot terminate the reset process
early by writinga '0' to this bit and shall not write any xHC
Operational or Runtime registers until while HCRST is '1'."
This behavior causes a regression on SNPS DWC3 USB controller with
dual-role capability. When the DWC3 controller exits host mode and
removes xhci while a reset is still in progress, and then tries to
configure its hardware for device mode, the ongoing reset leads to
register access issues; specifically, all register reads returns 0.
These issues extend beyond the xhci register space (which is expected
during a reset) and affect the entire DWC3 IP block, causing the DWC3
device mode to malfunction.
Cc: stable <stable@kernel.org>
Fixes: 6ccb83d6c4 ("usb: xhci: Implement xhci_handshake_check_state() helper")
Signed-off-by: Roy Luo <royluo@google.com>
Link: https://lore.kernel.org/r/20250522190912.457583-3-royluo@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 3eff494f6e upstream.
xhci_reset() currently returns -ENODEV if XHCI_STATE_REMOVING is
set, without completing the xhci handshake, unless the reset completes
exceptionally quickly. This behavior causes a regression on Synopsys
DWC3 USB controllers with dual-role capabilities.
Specifically, when a DWC3 controller exits host mode and removes xhci
while a reset is still in progress, and then attempts to configure its
hardware for device mode, the ongoing, incomplete reset leads to
critical register access issues. All register reads return zero, not
just within the xHCI register space (which might be expected during a
reset), but across the entire DWC3 IP block.
This patch addresses the issue by preventing xhci_reset() from being
called in xhci_resume() and bailing out early in the reinit flow when
XHCI_STATE_REMOVING is set.
Cc: stable <stable@kernel.org>
Fixes: 6ccb83d6c4 ("usb: xhci: Implement xhci_handshake_check_state() helper")
Suggested-by: Mathias Nyman <mathias.nyman@intel.com>
Signed-off-by: Roy Luo <royluo@google.com>
Link: https://lore.kernel.org/r/20250522190912.457583-2-royluo@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 68c1f16716 ]
The current xHCI driver does not validate whether a page size of 4096
bytes is supported. Address the issue by setting the page size to the
value supported by the xHCI controller, as read from the Page Size
register. In the event of an unexpected value; default to a 4K page size.
Additionally, this commit removes unnecessary debug messages and instead
prints the supported and used page size once.
The xHCI controller supports page sizes of (2^{(n+12)}) bytes, where 'n'
is the Page Size Bit. Only one page size is supported, with a maximum
page size of 128 KB.
Signed-off-by: Niklas Neronin <niklas.neronin@linux.intel.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250306144954.3507700-10-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit dfc88357b6 ]
When the device stalls an endpoint, current TD is assigned -EPIPE
status and Reset Endpoint is queued. If a Stop Endpoint is pending
at the time, it will run before Reset Endpoint and fail due to the
stall. Its handler will change TD's status to -EPROTO before Reset
Endpoint handler runs and initiates giveback.
Check if the stall has already been handled and don't try to do it
again. Since xhci_handle_halted_endpoint() performs this check too,
not overwriting td->status is the only difference.
I haven't seen this case yet, but I have seen a related one where
the xHC has already executed Reset Endpoint, EP Context state is
now Stopped and EP_HALTED is set. If the xHC took a bit longer to
execute Reset Endpoint, said case would become this one.
Signed-off-by: Michal Pecio <michal.pecio@gmail.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250311154551.4035726-3-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit cab63934c3 ]
Event polling delay is set to 0 if there are any pending requests in
either rx or tx requests lists. Checking for pending requests does
not work well for "IN" transfers as the tty driver always queues
requests to the list and TRBs to the ring, preparing to receive data
from the host.
This causes unnecessary busylooping and cpu hogging.
Only set the event polling delay to 0 if there are pending tx "write"
transfers, or if it was less than 10ms since last active data transfer
in any direction.
Cc: Łukasz Bartosik <ukaszb@chromium.org>
Fixes: fb18e5bb96 ("xhci: dbc: poll at different rate depending on data transfer activity")
Cc: stable@vger.kernel.org
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250505125630.561699-3-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 03e3d9c2bd ]
Queue event polling work with 0 delay in case there are pending transfers
queued up. This is part 2 of a 3 part series that roughly triples dbc
performace when using adb push and pull over dbc.
Max/min push rate after patches is 210/118 MB/s, pull rate 171/133 MB/s,
tested with large files (300MB-9GB) by Łukasz Bartosik
First performance improvement patch was commit 31128e7492
("xhci: dbc: add dbgtty request to end of list once it completes")
Cc: Łukasz Bartosik <ukaszb@chromium.org>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20241227120142.1035206-2-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Stable-dep-of: cab63934c3 ("xhci: dbc: Avoid event polling busyloop if pending rx transfers are inactive.")
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 732f35cf8b upstream.
When a USB device is connected to the OTG port, the tegra_xhci_id_work()
routine transitions the PHY to host mode and calls xhci_hub_control()
with the SetPortFeature command to enable port power.
In certain cases, the XHCI controller may be in a low-power state
when this operation occurs. If xhci_hub_control() is invoked while
the controller is suspended, the PORTSC register may return 0xFFFFFFFF,
indicating a read failure. This causes xhci_hc_died() to be triggered,
leading to host controller shutdown.
Example backtrace:
[ 105.445736] Workqueue: events tegra_xhci_id_work
[ 105.445747] dump_backtrace+0x0/0x1e8
[ 105.445759] xhci_hc_died.part.48+0x40/0x270
[ 105.445769] tegra_xhci_set_port_power+0xc0/0x240
[ 105.445774] tegra_xhci_id_work+0x130/0x240
To prevent this, ensure the controller is fully resumed before
interacting with hardware registers by calling pm_runtime_get_sync()
prior to the host mode transition and xhci_hub_control().
Fixes: f836e78430 ("usb: xhci-tegra: Add OTG support")
Cc: stable <stable@kernel.org>
Signed-off-by: Jim Lin <jilin@nvidia.com>
Signed-off-by: Wayne Chang <waynec@nvidia.com>
Link: https://lore.kernel.org/r/20250422114001.126367-1-waynec@nvidia.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit a5c7973539 upstream.
Device tree bindings state that the clock is optional for UHCI platform
controllers, and some existing device trees don't provide those - such
as those for VIA/WonderMedia devices.
The driver however fails to probe now if no clock is provided, because
devm_clk_get returns an error pointer in such case.
Switch to devm_clk_get_optional instead, so that it could probe again
on those platforms where no clocks are given.
Cc: stable <stable@kernel.org>
Fixes: 26c502701c ("usb: uhci: Add clk support to uhci-platform")
Signed-off-by: Alexey Charkov <alchark@gmail.com>
Link: https://lore.kernel.org/r/20250425-uhci-clock-optional-v1-1-a1d462592f29@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 9e3a28793d upstream.
A Short Packet event before the last TRB of a TD is followed by another
event on the final TRB on spec-compliant HCs, which is most of them.
A 'last_td_was_short' flag was added to know if a TD has just completed
as Short Packet and another event is to come. The flag was cleared after
seeing the event (unless no TDs are pending, but that's a separate bug)
or seeing a new TD complete as something other than Short Packet.
A rework replaced the flag with an 'old_trb_comp_code' variable. When
an event doesn't match the pending TD and the previous event was Short
Packet, the new event is silently ignored.
To preserve old behavior, 'old_trb_comp_code' should be cleared at this
point, but instead it is being set to current comp code, which is often
Short Packet again. This can cause more events to be silently ignored,
even though they are no longer connected with the old TD that completed
short and indicate a serious problem with the driver or the xHC.
Common device classes like UAC in async mode, UVC, serial or the UAS
status pipe complete as Short Packet routinely and could be affected.
Clear 'old_trb_comp_code' to zero, which is an invalid completion code
and the same value the variable starts with. This restores original
behavior on Short Packet and also works for illegal Etron events, which
the code has been extended to cover too.
Fixes: b331a3d809 ("xhci: Handle spurious events on Etron host isoc enpoints")
Signed-off-by: Michal Pecio <michal.pecio@gmail.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250410151828.2868740-4-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 64eb182d5f ]
Compatible "marvell,armada3700-xhci" match data uses the
struct xhci_plat_priv::init_quirk() function pointer to add
XHCI_RESET_ON_RESUME as quirk on XHCI.
Instead, use the struct xhci_plat_priv::quirks field.
Signed-off-by: Théo Lebrun <theo.lebrun@bootlin.com>
Link: https://lore.kernel.org/r/20250205-s2r-cdns-v7-1-13658a271c3c@bootlin.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 28a76fcc4c ]
Nothing prevents a broken HC from claiming that an endpoint is Running
and repeatedly rejecting Stop Endpoint with Context State Error.
Avoid infinite retries and give back cancelled TDs.
No such cases known so far, but HCs have bugs.
Signed-off-by: Michal Pecio <michal.pecio@gmail.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250311154551.4035726-4-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit b331a3d809 ]
Unplugging a USB3.0 webcam from Etron hosts while streaming results
in errors like this:
[ 2.646387] xhci_hcd 0000:03:00.0: ERROR Transfer event TRB DMA ptr not part of current TD ep_index 18 comp_code 13
[ 2.646446] xhci_hcd 0000:03:00.0: Looking for event-dma 000000002fdf8630 trb-start 000000002fdf8640 trb-end 000000002fdf8650
[ 2.646560] xhci_hcd 0000:03:00.0: ERROR Transfer event TRB DMA ptr not part of current TD ep_index 18 comp_code 13
[ 2.646568] xhci_hcd 0000:03:00.0: Looking for event-dma 000000002fdf8660 trb-start 000000002fdf8670 trb-end 000000002fdf8670
Etron xHC generates two transfer events for the TRB if an error is
detected while processing the last TRB of an isoc TD.
The first event can be any sort of error (like USB Transaction or
Babble Detected, etc), and the final event is Success.
The xHCI driver will handle the TD after the first event and remove it
from its internal list, and then print an "Transfer event TRB DMA ptr
not part of current TD" error message after the final event.
Commit 5372c65e13 ("xhci: process isoc TD properly when there was a
transaction error mid TD.") is designed to address isoc transaction
errors, but unfortunately it doesn't account for this scenario.
This issue is similar to the XHCI_SPURIOUS_SUCCESS case where a success
event follows a 'short transfer' event, but the TD the event points to
is already given back.
Expand the spurious success 'short transfer' event handling to cover
the spurious success after error on Etron hosts.
Kuangyi Chiang reported this issue and submitted a different solution
based on using error_mid_td. This commit message is mostly taken
from that patch.
Reported-by: Kuangyi Chiang <ki.chiang65@gmail.com>
Closes: https://lore.kernel.org/linux-usb/20241028025337.6372-6-ki.chiang65@gmail.com/
Tested-by: Kuangyi Chiang <ki.chiang65@gmail.com>
Tested-by: Michal Pecio <michal.pecio@gmail.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250306144954.3507700-16-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 906dec15b9 ]
The TRB pointer of these events points at enqueue at the time of error
occurrence on xHCI 1.1+ HCs or it's NULL on older ones. By the time we
are handling the event, a new TD may be queued at this ring position.
I can trigger this race by rising interrupt moderation to increase IRQ
handling delay. Similar delay may occur naturally due to system load.
If this ever happens after a Missed Service Error, missed TDs will be
skipped and the new TD processed as if it matched the event. It could
be given back prematurely, risking data loss or buffer UAF by the xHC.
Don't complete TDs on xrun events and don't warn if queued TDs don't
match the event's TRB pointer, which can be NULL or a link/no-op TRB.
Don't warn if there are no queued TDs at all.
Now that it's safe, also handle xrun events if the skip flag is clear.
This ensures completion of any TD stuck in 'error mid TD' state right
before the xrun event, which could happen if a driver submits a finite
number of URBs to a buggy HC and then an error occurs on the last TD.
Signed-off-by: Michal Pecio <michal.pecio@gmail.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250306144954.3507700-6-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit bfa8459942 ]
Missed Service Error after an error mid TD means that the failed TD has
already been passed by the xHC without acknowledgment of the final TRB,
a known hardware bug. So don't wait any more and give back the TD.
Reproduced on NEC uPD720200 under conditions of ludicrously bad USB link
quality, confirmed to behave as expected using dynamic debug.
Signed-off-by: Michal Pecio <michal.pecio@gmail.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250306144954.3507700-5-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 41d5e3806c ]
"maxim,max3421" DT compatible is missing its SPI device ID entry, not
allowing module autoloading and leading to the following message:
"SPI driver max3421-hcd has no spi_device_id for maxim,max3421"
Fix this by adding the spi_device_id table.
Signed-off-by: Alexander Stein <alexander.stein@mailbox.org>
Link: https://lore.kernel.org/r/20250128195114.56321-1-alexander.stein@mailbox.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit bcb60d4385 upstream.
The OHCI controller (rev 0x02) under LS7A PCI host has a hardware flaw.
MMIO register with offset 0x60/0x64 is treated as legacy PS2-compatible
keyboard/mouse interface, which confuse the OHCI controller. Since OHCI
only use a 4KB BAR resource indeed, the LS7A OHCI controller's 32KB BAR
is wrapped around (the second 4KB BAR space is the same as the first 4KB
internally). So we can add an 4KB offset (0x1000) to the OHCI registers
(from the PCI BAR resource) as a quirk.
Cc: stable <stable@kernel.org>
Suggested-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Alan Stern <stern@rowland.harvard.edu>
Tested-by: Mingcong Bai <baimingcong@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Link: https://lore.kernel.org/r/20250328040059.3672979-1-chenhuacai@loongson.cn
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 1ea050da55 upstream.
This check is performed before prepare_transfer() and prepare_ring(), so
enqueue can already point at the final link TRB of a segment. And indeed
it will, some 0.4% of times this code is called.
Then enqueue + 1 is an invalid pointer. It will crash the kernel right
away or load some junk which may look like a link TRB and cause the real
link TRB to be replaced with a NOOP. This wouldn't end well.
Use a functionally equivalent test which doesn't dereference the pointer
and always gives correct result.
Something has crashed my machine twice in recent days while playing with
an Etron HC, and a control transfer stress test ran for confirmation has
just crashed it again. The same test passes with this patch applied.
Fixes: 5e1c67abc9 ("xhci: Fix control transfer error on Etron xHCI host")
Cc: stable@vger.kernel.org
Signed-off-by: Michal Pecio <michal.pecio@gmail.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Reviewed-by: Kuangyi Chiang <ki.chiang65@gmail.com>
Link: https://lore.kernel.org/r/20250410151828.2868740-5-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit bea5892d0e upstream.
Current xhci bus resume implementation prevents xHC host from generating
interrupts during high-speed USB 2 and super-speed USB 3 bus resume.
Only reason to disable interrupts during bus resume would be to prevent
the interrupt handler from interfering with the resume process of USB 2
ports.
Host initiated resume of USB 2 ports is done in two stages.
The xhci driver first transitions the port from 'U3' to 'Resume' state,
then wait in Resume for 20ms, and finally moves port to U0 state.
xhci driver can't prevent interrupts by keeping the xhci spinlock
due to this 20ms sleep.
Limit interrupt disabling to the USB 2 port resume case only.
resuming USB 2 ports in bus resume is only done in special cases where
USB 2 ports had to be forced to suspend during bus suspend.
The current way of preventing interrupts by clearing the 'Interrupt
Enable' (INTE) bit in USBCMD register won't prevent the Interrupter
registers 'Interrupt Pending' (IP), 'Event Handler Busy' (EHB) and
USBSTS register Event Interrupt (EINT) bits from being set.
New interrupts can't be issued before those bits are properly clered.
Disable interrupts by clearing the interrupter register 'Interrupt
Enable' (IE) bit instead. This way IP, EHB and INTE won't be set
before IE is enabled again and a new interrupt is triggered.
Reported-by: Devyn Liu <liudingyuan@huawei.com>
Closes: https://lore.kernel.org/linux-usb/b1a9e2d51b4d4ff7a304f77c5be8164e@huawei.com/
Cc: stable@vger.kernel.org
Tested-by: Devyn Liu <liudingyuan@huawei.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250410151828.2868740-6-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 55741c7233 ]
The ffs() function returns the index of the first set bit, starting from 1.
If no bits are set, it returns zero. This behavior causes an off-by-one
page size in the debug message, as the page size calculation [1]
is zero-based, while ffs() is one-based.
Fix this by subtracting one from the result of ffs(). Note that since
variable 'val' is unsigned, subtracting one from zero will result in the
maximum unsigned integer value. Consequently, the condition 'if (val < 16)'
will still function correctly.
[1], Page size: (2^(n+12)), where 'n' is the set page size bit.
Fixes: 81720ec532 ("usb: host: xhci: use ffs() in xhci_mem_init()")
Signed-off-by: Niklas Neronin <niklas.neronin@linux.intel.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250306144954.3507700-9-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit bb0ba4cb10 upstream.
Two clearly different specimens of NEC uPD720200 (one with start/stop
bug, one without) were seen to cause IOMMU faults after some Missed
Service Errors. Faulting address is immediately after a transfer ring
segment and patched dynamic debug messages revealed that the MSE was
received when waiting for a TD near the end of that segment:
[ 1.041954] xhci_hcd: Miss service interval error for slot 1 ep 2 expected TD DMA ffa08fe0
[ 1.042120] xhci_hcd: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0005 address=0xffa09000 flags=0x0000]
[ 1.042146] xhci_hcd: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0005 address=0xffa09040 flags=0x0000]
It gets even funnier if the next page is a ring segment accessible to
the HC. Below, it reports MSE in segment at ff1e8000, plows through a
zero-filled page at ff1e9000 and starts reporting events for TRBs in
page at ff1ea000 every microframe, instead of jumping to seg ff1e6000.
[ 7.041671] xhci_hcd: Miss service interval error for slot 1 ep 2 expected TD DMA ff1e8fe0
[ 7.041999] xhci_hcd: Miss service interval error for slot 1 ep 2 expected TD DMA ff1e8fe0
[ 7.042011] xhci_hcd: WARN: buffer overrun event for slot 1 ep 2 on endpoint
[ 7.042028] xhci_hcd: All TDs skipped for slot 1 ep 2. Clear skip flag.
[ 7.042134] xhci_hcd: WARN: buffer overrun event for slot 1 ep 2 on endpoint
[ 7.042138] xhci_hcd: ERROR Transfer event TRB DMA ptr not part of current TD ep_index 2 comp_code 31
[ 7.042144] xhci_hcd: Looking for event-dma 00000000ff1ea040 trb-start 00000000ff1e6820 trb-end 00000000ff1e6820
[ 7.042259] xhci_hcd: WARN: buffer overrun event for slot 1 ep 2 on endpoint
[ 7.042262] xhci_hcd: ERROR Transfer event TRB DMA ptr not part of current TD ep_index 2 comp_code 31
[ 7.042266] xhci_hcd: Looking for event-dma 00000000ff1ea050 trb-start 00000000ff1e6820 trb-end 00000000ff1e6820
At some point completion events change from Isoch Buffer Overrun to
Short Packet and the HC finally finds cycle bit mismatch in ff1ec000.
[ 7.098130] xhci_hcd: ERROR Transfer event TRB DMA ptr not part of current TD ep_index 2 comp_code 13
[ 7.098132] xhci_hcd: Looking for event-dma 00000000ff1ecc50 trb-start 00000000ff1e6820 trb-end 00000000ff1e6820
[ 7.098254] xhci_hcd: ERROR Transfer event TRB DMA ptr not part of current TD ep_index 2 comp_code 13
[ 7.098256] xhci_hcd: Looking for event-dma 00000000ff1ecc60 trb-start 00000000ff1e6820 trb-end 00000000ff1e6820
[ 7.098379] xhci_hcd: Overrun event on slot 1 ep 2
It's possible that data from the isochronous device were written to
random buffers of pending TDs on other endpoints (either IN or OUT),
other devices or even other HCs in the same IOMMU domain.
Lastly, an error from a different USB device on another HC. Was it
caused by the above? I don't know, but it may have been. The disk
was working without any other issues and generated PCIe traffic to
starve the NEC of upstream BW and trigger those MSEs. The two HCs
shared one x1 slot by means of a commercial "PCIe splitter" board.
[ 7.162604] usb 10-2: reset SuperSpeed USB device number 3 using xhci_hcd
[ 7.178990] sd 9:0:0:0: [sdb] tag#0 UNKNOWN(0x2003) Result: hostbyte=0x07 driverbyte=DRIVER_OK cmd_age=0s
[ 7.179001] sd 9:0:0:0: [sdb] tag#0 CDB: opcode=0x28 28 00 04 02 ae 00 00 02 00 00
[ 7.179004] I/O error, dev sdb, sector 67284480 op 0x0:(READ) flags 0x80700 phys_seg 5 prio class 0
Fortunately, it appears that this ridiculous bug is avoided by setting
the chain bit of Link TRBs on isochronous rings. Other ancient HCs are
known which also expect the bit to be set and they ignore Link TRBs if
it's not. Reportedly, 0.95 spec guaranteed that the bit is set.
The bandwidth-starved NEC HC running a 32KB/uframe UVC endpoint reports
tens of MSEs per second and runs into the bug within seconds. Chaining
Link TRBs allows the same workload to run for many minutes, many times.
No negative side effects seen in UVC recording and UAC playback with a
few devices at full speed, high speed and SuperSpeed.
The problem doesn't reproduce on the newer Renesas uPD720201/uPD720202
and on old Etron EJ168 and VIA VL805 (but the VL805 has other bug).
[shorten line length of log snippets in commit messge -Mathias]
Signed-off-by: Michal Pecio <michal.pecio@gmail.com>
Cc: stable@vger.kernel.org
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250306144954.3507700-14-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 58d0a3fab5 upstream.
Up until commit d56b0b2ab1 ("usb: xhci: ensure skipped isoc TDs are
returned when isoc ring is stopped") in v6.11, the driver didn't skip
missed isochronous TDs when handling Stoppend and Stopped - Length
Invalid events. Instead, it erroneously cleared the skip flag, which
would cause the ring to get stuck, as future events won't match the
missed TD which is never removed from the queue until it's cancelled.
This buggy logic seems to have been in place substantially unchanged
since the 3.x series over 10 years ago, which probably speaks first
and foremost about relative rarity of this case in normal usage, but
by the spec I see no reason why it shouldn't be possible.
After d56b0b2ab1, TDs are immediately skipped when handling those
Stopped events. This poses a potential problem in case of Stopped -
Length Invalid, which occurs either on completed TDs (likely already
given back) or Link and No-Op TRBs. Such event won't be recognized
as matching any TD (unless it's the rare Link TRB inside a TD) and
will result in skipping all pending TDs, giving them back possibly
before they are done, risking isoc data loss and maybe UAF by HW.
As a compromise, don't skip and don't clear the skip flag on this
kind of event. Then the next event will skip missed TDs. A downside
of not handling Stopped - Length Invalid on a Link inside a TD is
that if the TD is cancelled, its actual length will not be updated
to account for TRBs (silently) completed before the TD was stopped.
I had no luck producing this sequence of completion events so there
is no compelling demonstration of any resulting disaster. It may be
a very rare, obscure condition. The sole motivation for this patch
is that if such unlikely event does occur, I'd rather risk reporting
a cancelled partially done isoc frame as empty than gamble with UAF.
This will be fixed more properly by looking at Stopped event's TRB
pointer when making skipping decisions, but such rework is unlikely
to be backported to v6.12, which will stay around for a few years.
Fixes: d56b0b2ab1 ("usb: xhci: ensure skipped isoc TDs are returned when isoc ring is stopped")
Cc: stable@vger.kernel.org
Signed-off-by: Michal Pecio <michal.pecio@gmail.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250306144954.3507700-4-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit c133ec0e57 upstream.
Raspberry Pi is a major user of those chips and they discovered a bug -
when the end of a transfer ring segment is reached, up to four TRBs can
be prefetched from the next page even if the segment ends with link TRB
and on page boundary (the chip claims to support standard 4KB pages).
It also appears that if the prefetched TRBs belong to a different ring
whose doorbell is later rung, they may be used without refreshing from
system RAM and the endpoint will stay idle if their cycle bit is stale.
Other users complain about IOMMU faults on x86 systems, unsurprisingly.
Deal with it by using existing quirk which allocates a dummy page after
each transfer ring segment. This was seen to resolve both problems. RPi
came up with a more efficient solution, shortening each segment by four
TRBs, but it complicated the driver and they ditched it for this quirk.
Also rename the quirk and add VL805 device ID macro.
Signed-off-by: Michal Pecio <michal.pecio@gmail.com>
Link: https://github.com/raspberrypi/linux/issues/4685
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=215906
CC: stable@vger.kernel.org
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250225095927.2512358-2-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 487cfd4a8e upstream.
When adding support for USB3-over-USB4 tunnelling detection, a check
for an Intel-specific capability was added. This capability, which
goes by ID 206, is used without any check that we are actually
dealing with an Intel host.
As it turns out, the Cadence XHCI controller *also* exposes an
extended capability numbered 206 (for unknown purposes), but of
course doesn't have the Intel-specific registers that the tunnelling
code is trying to access. Fun follows.
The core of the problems is that the tunnelling code blindly uses
vendor-specific capabilities without any check (the Intel-provided
documentation I have at hand indicates that 192-255 are indeed
vendor-specific).
Restrict the detection code to Intel HW for real, preventing any
further explosion on my (non-Intel) HW.
Cc: stable <stable@kernel.org>
Fixes: 948ce83fbb ("xhci: Add USB4 tunnel detection for USB3 devices on Intel hosts")
Signed-off-by: Marc Zyngier <maz@kernel.org>
Acked-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20250227194529.2288718-1-maz@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit c81d9fcd5b upstream.
Some Renesas HCs require firmware upload to work, this is handled by the
xhci_pci_renesas driver. Other variants of those chips load firmware from
a SPI flash and are ready to work with xhci_pci alone.
A refactor merged in v6.12 broke the latter configuration so that users
are finding their hardware ignored by the normal driver and are forced to
enable the firmware loader which isn't really necessary on their systems.
Let xhci_pci work with those chips as before when the firmware loader is
disabled by kernel configuration.
Fixes: 25f51b76f9 ("xhci-pci: Make xhci-pci-renesas a proper modular driver")
Cc: stable <stable@kernel.org>
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=219616
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=219726
Signed-off-by: Michal Pecio <michal.pecio@gmail.com>
Tested-by: Nicolai Buchwitz <nb@tipi-net.de>
Link: https://lore.kernel.org/r/20250128104529.58a79bfc@foxbook
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
