The current code uses commit fac323471d ("usb: udc: allow adding
and removing the same gadget device") as the workaround to let
the gadget device is re-used, but it is not allowed from driver
core point. In this commit, we allocate gadget structure dynamically,
and free it at its release function. Since the gadget device's
driver_data has already occupied by usb_composite_dev structure, we have
to use gadget device's platform data to store dwc3 structure.
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Alan Stern <stern@rowland.harvard.edu>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Peter Chen <peter.chen@nxp.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
According the programming guide (for all DWC3 IPs), when the driver
handles ClearFeature(halt) request, it should issue CLEAR_STALL command
_after_ the END_TRANSFER command completes. The END_TRANSFER command may
take some time to complete. So, delay the ClearFeature(halt) request
control status stage and wait for END_TRANSFER command completion
interrupt. Only after END_TRANSFER command completes that the driver
may issue CLEAR_STALL command.
Cc: stable@vger.kernel.org
Fixes: cb11ea56f3 ("usb: dwc3: gadget: Properly handle ClearFeature(halt)")
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
The function driver may queue new requests right after halting the
endpoint (i.e. queue new requests while the endpoint is stalled).
There's no restriction preventing it from doing so. However, dwc3
currently drops those requests after CLEAR_STALL. The driver should only
drop started requests. Keep the pending requests in the pending list to
resume and process them after the host issues ClearFeature(Halt) to the
endpoint.
Cc: stable@vger.kernel.org
Fixes: cb11ea56f3 ("usb: dwc3: gadget: Properly handle ClearFeature(halt)")
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
Currently dwc3 doesn't handle usb_request->zero for SG requests. This
change checks and prepares extra TRBs for the ZLP for SG requests.
Cc: <stable@vger.kernel.org> # v4.5+
Fixes: 04c03d10e5 ("usb: dwc3: gadget: handle request->zero")
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
The usb_request->zero doesn't apply for isoc. Also, if we prepare a
0-length (ZLP) TRB for the OUT direction, we need to prepare an extra
TRB to pad up to the MPS alignment. Use the same bounce buffer for the
ZLP TRB and the extra pad TRB.
Cc: <stable@vger.kernel.org> # v4.5+
Fixes: d6e5a549cc ("usb: dwc3: simplify ZLP handling")
Fixes: 04c03d10e5 ("usb: dwc3: gadget: handle request->zero")
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
The SG list may be set up with entry size more than the requested
length. Check the usb_request->length and make sure that we don't setup
the TRBs to send/receive more than requested. This case may occur when
the SG entry is allocated up to a certain minimum size, but the request
length is less than that. It can also occur when the request is reused
for a different request length.
Cc: <stable@vger.kernel.org> # v4.18+
Fixes: a31e63b608 ("usb: dwc3: gadget: Correct handling of scattergather lists")
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
When we have nothing left to be queued after handling the last trb
we have to stop the current transfer. This way we can ensure that
the next request will be queued with a new and valid timestamp
and will not directly run into an missed xfer.
Reviewed-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
Currently __dwc3_gadget_start_isoc must be called very shortly after
XferNotReady. Otherwise the frame number is outdated and start transfer
will fail, even with several retries.
DSTS provides the lower 14 bit of the frame number. Use it in combination
with the frame number provided by XferNotReady to guess the current frame
number. This will succeed unless more than one 14 rollover has happened
since XferNotReady.
Start transfer might still fail if the frame number is increased
immediately after DSTS is read. So retries are still needed.
Don't drop the current request if this happens. This way it is not lost and
can be used immediately to try again with the next frame number.
With this change, __dwc3_gadget_start_isoc is still not successfully in all
cases bit it increases the acceptable delay after XferNotReady
significantly.
Reviewed-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
This patch adds a define DWC3_FRNUMBER_MASK for the commonly used
0x3fff mask and uses it.
Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
Rationale:
Reduces attack surface on kernel devs opening the links for MITM
as HTTPS traffic is much harder to manipulate.
Deterministic algorithm:
For each file:
If not .svg:
For each line:
If doesn't contain `\bxmlns\b`:
For each link, `\bhttp://[^# \t\r\n]*(?:\w|/)`:
If neither `\bgnu\.org/license`, nor `\bmozilla\.org/MPL\b`:
If both the HTTP and HTTPS versions
return 200 OK and serve the same content:
Replace HTTP with HTTPS.
Signed-off-by: Alexander A. Klimov <grandmaster@al2klimov.de>
Link: https://lore.kernel.org/r/20200711135804.19735-1-grandmaster@al2klimov.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
For some reason, the TEST_ defines in the usb/ch9.h files did not have
the USB_ prefix on it, making it a bit confusing when reading the file,
as well as not the nicest thing to do in a uapi file.
So fix that up and add the USB_ prefix on to them, and fix up all
in-kernel usages. This included deleting the duplicate copy in the
net2272.h file.
Cc: Felipe Balbi <balbi@kernel.org>
Cc: Michal Simek <michal.simek@xilinx.com>
Cc: Mathias Nyman <mathias.nyman@intel.com>
Cc: Pawel Laszczak <pawell@cadence.com>
Cc: YueHaibing <yuehaibing@huawei.com>
Cc: Nathan Chancellor <natechancellor@gmail.com>
Cc: Jason Yan <yanaijie@huawei.com>
Cc: Jia-Ju Bai <baijiaju1990@gmail.com>
Cc: Stephen Boyd <swboyd@chromium.org>
Cc: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Jules Irenge <jbi.octave@gmail.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Thinh Nguyen <Thinh.Nguyen@synopsys.com>
Cc: Rob Gill <rrobgill@protonmail.com>
Cc: Macpaul Lin <macpaul.lin@mediatek.com>
Acked-by: Minas Harutyunyan <hminas@synopsys.com>
Acked-by: Bin Liu <b-liu@ti.com>
Acked-by: Chunfeng Yun <chunfeng.yun@mediatek.com>
Acked-by: Peter Chen <peter.chen@nxp.com>
Link: https://lore.kernel.org/r/20200618144206.2655890-1-gregkh@linuxfoundation.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Rather busy cycle. We have a total 99 non-merge commits going into v5.8
merge window. The majority of the changes are in dwc3 this around (31.7%
of all changes). It's composed mostly Thinh's recent updates to get dwc3
to behave correctly with stream transfers. We have also have Roger's for
Keystone platforms and Neil's updates for the meson glue layer.
Apart from those, we have the usual set of non-critical fixes, new
device IDs, spelling fixes all over the place.
Signed-off-by: Felipe Balbi <balbi@kernel.org>
-----BEGIN PGP SIGNATURE-----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=OiS0
-----END PGP SIGNATURE-----
Merge tag 'usb-for-v5.8' of git://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb into usb-next
Felipe writes:
USB: changes for v5.8 merge window
Rather busy cycle. We have a total 99 non-merge commits going into v5.8
merge window. The majority of the changes are in dwc3 this around (31.7%
of all changes). It's composed mostly Thinh's recent updates to get dwc3
to behave correctly with stream transfers. We have also have Roger's for
Keystone platforms and Neil's updates for the meson glue layer.
Apart from those, we have the usual set of non-critical fixes, new
device IDs, spelling fixes all over the place.
Signed-off-by: Felipe Balbi <balbi@kernel.org>
* tag 'usb-for-v5.8' of git://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb: (99 commits)
usb: dwc3: keystone: Turn on USB3 PHY before controller
dt-bindings: usb: ti,keystone-dwc3.yaml: Add USB3.0 PHY property
dt-bindings: usb: convert keystone-usb.txt to YAML
usb: dwc3: gadget: Check for prepared TRBs
usb: gadget: Fix issue with config_ep_by_speed function
usb: cdns3: ep0: delete the redundant status stage
usb: dwc2: Update Core Reset programming flow.
usb: gadget: fsl: Fix a wrong judgment in fsl_udc_probe()
usb: gadget: fix potential double-free in m66592_probe.
usb: cdns3: Fix runtime PM imbalance on error
usb: gadget: lpc32xx_udc: don't dereference ep pointer before null check
usb: dwc3: Increase timeout for CmdAct cleared by device controller
USB: dummy-hcd: use configurable endpoint naming scheme
usb: cdns3: gadget: assign interrupt number to USB gadget structure
usb: gadget: core: sync interrupt before unbind the udc
arm64: dts: qcom: sc7180: Add interconnect properties for USB
arm64: dts: qcom: sdm845: Add interconnect properties for USB
dt-bindings: usb: qcom,dwc3: Introduce interconnect properties for Qualcomm DWC3 driver
ARM: dts: at91: Remove the USB EP child node
dt-bindings: usb: atmel: Mark EP child node as deprecated
...
There are cases where the endpoint needs to be restarted. For example,
it may need to restart for NoStream rejection to reinitiate stream. If
so, check and make sure we don't prepare beyond the current transfer
when we restart the endpoint.
DWC_usb32 internal burst transfer feature will look into TRBs beyond a
transfer. Other controllers will stop on the last TRB, but not
DWC_usb32. This may cause the controller to incorrectly process TRBs of
a different transfer. Make sure to explicitly prevent preparing TRBs of
a different transfer.
This should only affect DWC_usb32 releases prior to v1.00a since it
doesn't use SET_ENDPOINT_PRIME to reinitiate stream. However, it's
better to be cautious in case users don't want to use SET_ENDPOINT_PRIME
command. Also, it's possible other controller IPs may share the same
features as DWC_usb32 in new releases.
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
If the SS PHY is in P3, there is no pipe_clk, HW may use suspend_clk
for function, as suspend_clk is slow so EP command need more time to
complete, e.g, imx8M suspend_clk is 32K, set ep configuration will
take about 380us per below trace time stamp(44.286278 - 44.285897
= 0.000381):
configfs_acm.sh-822 [000] d..1 44.285896: dwc3_writel: addr
000000006d59aae1 value 00000401
configfs_acm.sh-822 [000] d..1 44.285897: dwc3_readl: addr
000000006d59aae1 value 00000401
... ...
configfs_acm.sh-822 [000] d..1 44.286278: dwc3_readl: addr
000000006d59aae1 value 00000001
configfs_acm.sh-822 [000] d..1 44.286279: dwc3_gadget_ep_cmd:
ep0out: cmd 'Set Endpoint Configuration' [401] params 00001000
00000500 00000000 --> status: Successful
This was originally found on Hisilicon Kirin Soc that need more time
for the device controller to clear the CmdAct of DEPCMD.
Signed-off-by: Yu Chen <chenyu56@huawei.com>
Signed-off-by: John Stultz <john.stultz@linaro.org>
Signed-off-by: Li Jun <jun.li@nxp.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
DWC_usb32 v1.00a and later can use SET_EP_PRIME command to reinitiate a
stream. Use the command to handle NoStream rejection instead of ending
and restarting the endpoint.
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
Overview of stream transfer requirement:
* A transfer will have a set of TRBs of the same stream ID.
* A transfer is started with a stream ID in START_TRANSFER command.
* A new stream will only start when the previous completes.
Overview of stream events:
* A "prime" from host indicates that its endpoints are active
(buffers prepared and ready to receive/transmit data). The controller
automatically initiates stream if it sees this.
* A "NoStream" rejection event indicates that the host isn't ready.
Host will put the endpoint back to idle state. Device may need to
reinitiate the stream to start transfer again.
* A Stream Found event means host accepted device initiated stream.
Nothing needs to be done from driver.
To initiate a stream, the driver will issue START_TRANSFER command with
a stream ID. To reinitiate the stream, the driver must issue
END_TRANSFER and restart the transfer with START_TRANSFER command with
the same stream ID.
This implementation handles device-initated streams (e.g. UASP driver).
It also handles some hosts' quirky behavior where they only prime each
endpoint once.
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
Don't prepare TRBs beyond a transfer. In DWC_usb32, its transfer burst
capability may try to read and use TRBs beyond the active transfer. For
other controllers, they don't process the next transfer TRBs until the
current transfer is completed. Explicitly prevent preparing TRBs ahead
for all controllers.
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
If a transfer is in-progress, any new request should not kick off
another transfer. The driver needs to wait for the current transfer to
complete before starting off the next transfer. Introduce a new flag
DWC3_EP_WAIT_TRANSFER_COMPLETE for this.
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
In DWC3, to prepare TRBs for streams, all the TRBs of a transfer will
use the same stream ID. To start a new stream, the driver needs to wait
for the current transfer to complete or ended (by END_TRANFER command).
As a result, inform the controller of the last TRB of a transfer so that
it knows when a transfer completes and start a new transfer of a new
stream.
Even though the transfer completion handling can be applied for other
non-isoc endpoints, only do it for streams due to its requirement.
It's better to keep the controller's TRB cache full than waiting for
transfer completion and starting a new transfer.
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
To switch from one stream to another, this requires the driver to start
a new transfer with a specific stream ID. For a transfer to complete,
the driver needs to indicate the last TRB of a transfer, and it needs to
enable XferComplete event to handle completed TRBs of a transfer. Let's
enable this event only for stream capable endpoints.
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
To prepare for handling of XferComplete event, let's refactor and split
up dwc3_gadget_endpoint_transfer_in_progress() to handle TRBs completion
for different events. The handling of TRBs completion will be the same,
but the status of XferComplete event is different than XferInProgress.
No functional change in this commit.
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
While handling TRBs completion, if a END_TRANSFER command isn't
completed, don't kick new transfer or issue END_TRANSFER command.
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
DWC_usb32 supports MDWIDTH value larger than 255 and up to 1023. The
field HWPARAMS6[9:8] stores the upper 2-bit values of the DWC_usb32's
MDWIDTH. Check that parameter and properly get the MDWIDTH for
DWC_usb32.
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
Synopsys introduces a new controller DWC_usb32. It supports dual-lane
and speed up to 20 Gbps, and the DWC3 driver will drive this controller.
Currently the driver uses a single field dwc->revision to ID both
DWC_usb3 and DWC_usb31 and their version number. This was sufficient for
two IPs, but this method doesn't work with additional IPs. As a result,
let's separate the dwc->revision field to 2 separate fields: ip and
revision. The ip field now stores the ID of the controller's IP while
the revision field stores the controller's version number.
This new scheme enforces DWC3 to compare the revision within the same IP
only. As a result, we must update all the revision check of the
controller to check its corresponding IP.
To help with this enforcement, we create a few macros to help with
the common version checks:
DWC3_IP_IS(IP)
DWC3_VER_IS(IP, VERSION)
DWC3_VER_IS_PRIOR(IP, VERSION)
DWC3_VER_IS_WITHIN(IP, LOWER_VERSION, UPPER_VERSION)
DWC3_VER_TYPE_IS_WITHIN(IP, VERSION,
LOWER_VERSION_TYPE,
UPPER_VERSION_TYPE)
The DWC_usb32 controller operates using the same programming model and
with very similar configurations as its previous controllers. Please
note that the various IP and revision checks in this patch match the
current checks for DWC_usb31 version 1.90a. Additional configurations
that are unique to DWC_usb32 are applied separately.
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
If there are still pending requests because no TRB was available,
prepare more when started requests are completed.
Introduce dwc3_gadget_ep_should_continue() to check for incomplete and
pending requests to resume updating new TRBs to the controller's TRB
cache.
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
The check for the HWO flag in dwc3_gadget_ep_reclaim_trb_sg()
causes us to break out of the loop before we call
dwc3_gadget_ep_reclaim_completed_trb(), which is what likely
should be clearing the HWO flag.
This can cause odd behavior where we never reclaim all the trbs
in the sg list, so we never call giveback on a usb req, and that
will causes transfer stalls.
This effectively resovles the adb stalls seen on HiKey960
after userland changes started only using AIO in adbd.
Cc: YongQin Liu <yongqin.liu@linaro.org>
Cc: Anurag Kumar Vulisha <anurag.kumar.vulisha@xilinx.com>
Cc: Yang Fei <fei.yang@intel.com>
Cc: Thinh Nguyen <thinhn@synopsys.com>
Cc: Tejas Joglekar <tejas.joglekar@synopsys.com>
Cc: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Cc: Jack Pham <jackp@codeaurora.org>
Cc: Josh Gao <jmgao@google.com>
Cc: Todd Kjos <tkjos@google.com>
Cc: Felipe Balbi <balbi@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: linux-usb@vger.kernel.org
Cc: stable@vger.kernel.org #4.20+
Signed-off-by: John Stultz <john.stultz@linaro.org>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
If the driver issued START_TRANSFER and received a no-resource status,
then generally there are a few reasons for this:
1) The driver did not allocate resource for the endpoint during
power-on-reset initialization.
2) The transfer resource was reset. At this moment, we don't do this in
the driver, but it occurs when the driver issues START_CONFIG cmd to ep0
with resource index=2.
3) The driver issues the START_TRANSFER command to an already started
endpoint. Usually, this is because the END_TRANSFER command hasn't
completed yet.
Print out a warning to help debug this issue in the driver.
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
After a number of unsuccessful start isoc attempts due to bus-expiry
status, issue END_TRANSFER command and retry on the next XferNotReady
event.
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
As long as the START_TRANSFER command completes, it provides the
resource index of the endpoint. Use this when we need to issue
END_TRANSFER command to an isoc endpoint to retry with a new
XferNotReady event.
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
If dwc3 fails to issue START_TRANSFER/UPDATE_TRANSFER command, then we
should properly end an active transfer and give back all the started
requests. However if it's for an isoc endpoint, the failure maybe due to
bus-expiry status. In this case, don't give back the requests and wait
for the next retry.
Fixes: 72246da40f ("usb: Introduce DesignWare USB3 DRD Driver")
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
The flow from function dwc3_gadget_ep_dequeue() is not easy to follow.
Refactor it for easier read. No functional change in this commit.
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
Remove 2 unnecessary checks:
1) A request in the started_list must have its trb field set. So
checking for req->trb is unnecessary.
2) An endpoint must have started (and have not ended) for the request to
still be in the started_list. There's no point to check if the endpoint
is started. We had this check because previously the driver didn't
handle the endpoint's started/ended flags for END_TRANSFER command
properly. See commit 9f45581f5e ("usb: dwc3: gadget: early giveback
if End Transfer already completed").
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
If a request is dequeued, the transfer is cancelled. Give back all
the started requests.
In most scenarios, the function driver dequeues all requests of a
transfer when there's a failure. If the function driver follows this,
then it's fine. If not, then we'd be skipping TRBs at different points
within the dequeue and enqueue pointers, making dequeue/enqueue pointers
useless. To enforce and make sure that we're properly skipping TRBs,
cancel all the started requests and give back all the cancelled requests
to the function drivers.
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
DWC3 must not issue CLEAR_STALL command to control endpoints. The
controller automatically clears the STALL when it receives the SETUP
token. Also, when the driver receives ClearFeature(halt_ep), DWC3 must
stop any active transfer from the endpoint and give back all the
requests to the function drivers.
Fixes: 72246da40f ("usb: Introduce DesignWare USB3 DRD Driver")
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
A request may not be completed because not all the TRBs are prepared for
it. This happens when we run out of available TRBs. When some TRBs are
completed, the driver needs to prepare the rest of the TRBs for the
request. The check dwc3_gadget_ep_request_completed() shouldn't be
checking the amount of data received but rather the number of pending
TRBs. Revise this request completion check.
Cc: stable@vger.kernel.org
Fixes: e0c42ce590 ("usb: dwc3: gadget: simplify IOC handling")
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
The controller always supports link recovery for device in SS and SSP.
Remove the speed limit check. Also, when the device is in RESUME or
RESET state, it means the controller received the resume/reset request.
The driver must send the link recovery to acknowledge the request. They
are valid states for the driver to send link recovery.
Fixes: 72246da40f ("usb: Introduce DesignWare USB3 DRD Driver")
Fixes: ee5cd41c91 ("usb: dwc3: Update speed checks for SuperSpeedPlus")
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
Currently the calculation of max packet size limit for IN endpoints is
too restrictive. This prevents a matching of a capable hardware endpoint
during configuration. Below is the minimum recommended HW configuration
to support a particular endpoint setup from the databook:
For OUT endpoints, the databook recommended the minimum RxFIFO size to
be at least 3x MaxPacketSize + 3x setup packets size (8 bytes each) +
clock crossing margin (16 bytes).
For IN endpoints, the databook recommended the minimum TxFIFO size to be
at least 3x MaxPacketSize for endpoints that support burst. If the
endpoint doesn't support burst or when the device is operating in USB
2.0 mode, a minimum TxFIFO size of 2x MaxPacketSize is recommended.
Base on these recommendations, we can calculate the MaxPacketSize limit
of each endpoint. This patch revises the IN endpoint MaxPacketSize limit
and also sets the MaxPacketSize limit for OUT endpoints.
Reference: Databook 3.30a section 3.2.2 and 3.2.3
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
Change the macro name DWC3_GTXFIFOSIZ_TXFDEF to DWC3_GTXFIFOSIZ_TXFDEP
to match with the register name GTXFIFOSIZ.TXFDEP.
Fixes: 457e84b662 ("usb: dwc3: gadget: dynamically re-size TxFifos")
Fixes: 0cab8d26d6 ("usb: dwc3: Update DWC_usb31 GTXFIFOSIZ reg fields")
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
When skipping TRBs, we need to account for wrapping around the ring
buffer and not modifying some invalid TRBs. Without this fix, dwc3 won't
be able to check for available TRBs.
Cc: stable <stable@vger.kernel.org>
Fixes: 7746a8dfb3 ("usb: dwc3: gadget: extract dwc3_gadget_ep_skip_trbs()")
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
We track END_TRANSFER command completion. Don't clear transfer
started/ended flag prematurely. Otherwise, we'd run into the problem
with restarting transfer before END_TRANSFER command finishes.
Fixes: 6d8a019614 ("usb: dwc3: gadget: check for Missed Isoc from event status")
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
If scatter-gather operation is allowed, a large USB request is split
into multiple TRBs. For preparing TRBs for sg list, driver iterates
over the list and creates TRB for each sg and mark the chain bit to
false for the last sg. The current IOMMU driver is clubbing the list
of sgs which shares a page boundary into one and giving it to USB driver.
With this the number of sgs mapped it not equal to the the number of sgs
passed. Because of this USB driver is not marking the chain bit to false
since it couldn't iterate to the last sg. This patch addresses this issue
by marking the chain bit to false if it is the last mapped sg.
At a practical level, this patch resolves USB transfer stalls
seen with adb on dwc3 based db845c, pixel3 and other qcom
hardware after functionfs gadget added scatter-gather support
around v4.20.
Credit also to Anurag Kumar Vulisha <anurag.kumar.vulisha@xilinx.com>
who implemented a very similar fix to this issue.
Cc: Felipe Balbi <balbi@kernel.org>
Cc: Yang Fei <fei.yang@intel.com>
Cc: Thinh Nguyen <thinhn@synopsys.com>
Cc: Tejas Joglekar <tejas.joglekar@synopsys.com>
Cc: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Cc: Jack Pham <jackp@codeaurora.org>
Cc: Todd Kjos <tkjos@google.com>
Cc: Greg KH <gregkh@linuxfoundation.org>
Cc: Linux USB List <linux-usb@vger.kernel.org>
Cc: stable <stable@vger.kernel.org> #4.20+
Signed-off-by: Pratham Pratap <prathampratap@codeaurora.org>
[jstultz: Slight tweak to remove sg_is_last() usage, reworked
commit message, minor comment tweak]
Signed-off-by: John Stultz <john.stultz@linaro.org>
Link: https://lore.kernel.org/r/20200302214443.55783-1-john.stultz@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The current code in dwc3_gadget_ep_reclaim_completed_trb() will
check for IOC/LST bit in the event->status and returns if
IOC/LST bit is set. This logic doesn't work if multiple TRBs
are queued per request and the IOC/LST bit is set on the last
TRB of that request.
Consider an example where a queued request has multiple queued
TRBs and IOC/LST bit is set only for the last TRB. In this case,
the core generates XferComplete/XferInProgress events only for
the last TRB (since IOC/LST are set only for the last TRB). As
per the logic in dwc3_gadget_ep_reclaim_completed_trb()
event->status is checked for IOC/LST bit and returns on the
first TRB. This leaves the remaining TRBs left unhandled.
Similarly, if the gadget function enqueues an unaligned request
with sglist already in it, it should fail the same way, since we
will append another TRB to something that already uses more than
one TRB.
To aviod this, this patch changes the code to check for IOC/LST
bits in TRB->ctrl instead.
At a practical level, this patch resolves USB transfer stalls seen
with adb on dwc3 based HiKey960 after functionfs gadget added
scatter-gather support around v4.20.
Cc: Felipe Balbi <balbi@kernel.org>
Cc: Yang Fei <fei.yang@intel.com>
Cc: Thinh Nguyen <thinhn@synopsys.com>
Cc: Tejas Joglekar <tejas.joglekar@synopsys.com>
Cc: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Cc: Jack Pham <jackp@codeaurora.org>
Cc: Todd Kjos <tkjos@google.com>
Cc: Greg KH <gregkh@linuxfoundation.org>
Cc: Linux USB List <linux-usb@vger.kernel.org>
Cc: stable <stable@vger.kernel.org>
Tested-by: Tejas Joglekar <tejas.joglekar@synopsys.com>
Reviewed-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Anurag Kumar Vulisha <anurag.kumar.vulisha@xilinx.com>
[jstultz: forward ported to mainline, reworded commit log, reworked
to only check trb->ctrl as suggested by Felipe]
Signed-off-by: John Stultz <john.stultz@linaro.org>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
We had a 100us delay to synchronize the END_TRANSFER command completion
before giving back requests to the function drivers. Now, the controller
driver can handle cancelled TRBs with the requests' cancelled_list and
it can also wait until the END_TRANSFER completion before starting new
transfers. Synchronization can simply base on the controller's command
completion interrupt. The 100us delay is no longer needed. Remove this
arbitrary delay.
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If the END_TRANSFER command hasn't completed yet, then don't send the
START_TRANSFER command. The controller may not be able to start if
that's the case. Some controller revisions depend on this. See
commit 76a638f8ac ("usb: dwc3: gadget: wait for End Transfer to
complete"). Let's only send START_TRANSFER command after the
END_TRANSFER command had completed.
Fixes: 3aec99154d ("usb: dwc3: gadget: remove DWC3_EP_END_TRANSFER_PENDING")
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
While the END_TRANSFER command is sent but not completed, any request
dequeue during this time will cause the driver to issue the END_TRANSFER
command. The driver needs to submit the command only once to stop the
controller from processing further. The controller may take more time to
process the same command multiple times unnecessarily. Let's add a flag
DWC3_EP_END_TRANSFER_PENDING to check for this condition.
Fixes: 3aec99154d ("usb: dwc3: gadget: remove DWC3_EP_END_TRANSFER_PENDING")
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Send a no-action link state change request before the actual request
so DWC3 can send the same request whenever we call
dwc3_gadget_set_link_state().
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When DWC3 receives disconnect event, it needs to set the link state to
RX_Detect.
DWC_usb3 3.30a programming guide 4.1.7
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
DCTL.ULSTCHNGREQ is a write-only field. When doing a read-modify-write
to DCTL, the driver must make sure that there's no unintended link state
change request from whatever is read from DCTL.ULSTCHNGREQ. Set link
state change to no-action when the driver writes to DCTL.
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Peter Chen