The USB device subsystem driver's set_interface() function calls
usb_dc_ep_configure() followed by usb_dc_ep_enable(). When
switching between alternate settings of a configuration's
interface, set_endpoint() can be followed by reset_endpoint() on
an endpoint. Some time later, set_endpoint() can be called again
on the same endpoint. This results in the HAL's
kUSB_DeviceControlEndpointInit function being called twice in a
row, which causes a memory allocation error. A simple solution is
to call the HAL's kUSB_DeviceControlEndpointDeinit function
before calling kUSB_DeviceControlEndpointInit. This overcomes the
memory allocation error.
Signed-off-by: Milind Paranjpe <mparanjpe@yahoo.com>
Change automated searching for files using "IRQ_CONNECT()" API not
including <zephyr/irq.h>.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
SAM V71 Xplained Ultra failed to re-enumerate on the bus after device
detach-attach sequence. The device was recognized by the host, but it
was not responding to control transfers, i.e. it was not acknowledging
SETUP data. Make device acknowledge SETUP data by reconfiguring EP0
when necessary.
Signed-off-by: Tomasz Moń <tomasz.mon@nordicsemi.no>
Device physically detaches when the DETACH bit is set. Fix the
assignment to actually set the bit instead of clearing it.
Signed-off-by: Tomasz Moń <tomasz.mon@nordicsemi.no>
Sometimes it is desired for the USB device to detach and reattach itself
to the host. USB device controller driver interface supports detach with
usb_dc_detach() function. Subsequently the device can be reattached to
the bus by calling usb_dc_attach().
USB device stack sets the control transfer callbacks in usb_enable().
Nordic usb_dc_detach() implementation inadvertedly overwrote registered
callbacks with zeroes by clearing endpoint configuration structure. This
lead to a NULL pointer dereference during enumeration after device has
reattached inself to the host.
Preserve the callbacks by removing the eps_ctx_uninit() function.
Endpoints runtime data is initialized in usb_dc_attach() so this should
pose no problems.
Signed-off-by: Tomasz Moń <tomasz.mon@nordicsemi.no>
The usb_dc_sam0 driver is leaking memory when the usb device is detached
using usb_disable().
When a device is detached, the driver does not reliably deallocate the
descriptor bank memory. Therefore, upon the next attach and configure,
another block of memory is allocated. After several connect /
disconnect sequences, the usb interface becomes unresponsive.
This commit adds a static function to release all memory allocated
to the ep descriptors when usb_dc_detach() is called. Since detach is
called from usb_disable(), this releases the memory when the
interface is disabled.
Avoid calling k_free() if buffer pointer is already
set to NULL.
Fixes#49177
Signed-off-by: Steven Slupsky <sslupsky@gmail.com>
Ensure that the FIFO of each endpoint is initialized during endpoint
configuration, to prevent a NULL FIFO from being encountered during the
first call to usb_dc_ep_write() for a given endpoint.
Signed-off-by: Nick Kraus <nick@nckraus.com>
1. Move the defines from usb_dc_mcux.h to usb_device_config.h
and fsl_os_abstraction.h. These headers are used by
the SDK USB driver. usb_dc_mcux.h header file is not longer
needed and hence deleted.
2. Delete the Zephyr implementation of the usb_device_struct
driver and use the one implemented inside the SDK USB
driver. This requires updating the references to
usb_device_struct inside the USB driver
3. Move defines and structures used by the driver
out of the header file that is included by the SDK and
into the MCUX USB driver.
4. Use end point defines provided by Zephyr instead of adding
them locally.
5. Add a Kconfig to set the thread stack size
6. Move code to enable interrupts back to usb_attach function.
Interrupts should be enabled after the init is successful,
else we see errors of the ISR getting called before the
init is complete causing Faults
6. Update west.yml to update the NXP HAL to get the updated
SDK USB driver.
Signed-off-by: Mahesh Mahadevan <mahesh.mahadevan@nxp.com>
So far thread was created as part of usb_dc_attach() by k_thread_create().
This means that if following function were executed:
* usb_enable()
* usb_disable()
* usb_enable()
then k_thread_create() was called second time. This results in undefined
behavior.
Fix above issue by moving k_thread_create() invocation to function called
during system initialization.
While at it, move IRQ_CONNECT() and irq_enable() invocations to init as
well.
Signed-off-by: Marcin Niestroj <m.niestroj@emb.dev>
So far thread was created as part of usb_dc_attach() by k_thread_create().
This means that if following function were executed:
* usb_enable()
* usb_disable()
* usb_enable()
then k_thread_create() was called second time. This results in undefined
behavior.
Fix above issue by moving k_thread_create() invocation to function called
during system initialization.
While at it, move IRQ_CONNECT() and irq_enable() invocations to init as
well.
Signed-off-by: Marcin Niestroj <m.niestroj@emb.dev>
So far thread was created as part of usb_dc_attach() by k_thread_create().
This means that if following function were executed:
* usb_enable()
* usb_disable()
* usb_enable()
then k_thread_create() was called second time. This results in undefined
behavior.
Fix above issue by moving k_thread_create() invocation to function called
during system initialization.
While at it, move IRQ_CONNECT() and irq_enable() invocations to init as
well.
Signed-off-by: Marcin Niestroj <m.niestroj@emb.dev>
Update usb device drivers to use DT_HAS_<compat>_ENABLED Kconfig symbol
to expose the driver and enable it by default based on devicetree.
We remove 'depend on' Kconfig for symbols that would be implied by
the devicetree node existing.
Signed-off-by: Kumar Gala <galak@kernel.org>
Many device pointers are initialized at compile and never changed. This
means that the device pointer can be constified (immutable).
Automated using:
```
perl -i -pe 's/const struct device \*(?!const)(.*)= DEVICE/const struct
device *const $1= DEVICE/g' **/*.c
```
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Soft resetting a STM32 device currently does not reset the USB
connection, which causes a few problems: The endpoints do not respond
anymore, and within zephyr any kind of status information like
CDC_SET_CONTROL_LINE_STATE is also missing as they are not
re-negotiated.
This is fixed by stopping the USB device from zephyr side, right after
initialising the USB device.
Signed-off-by: David Jablonski <dayjaby@gmail.com>
MISRA C:2012 Rule 14.4 (The controlling expression of an if statement
and the controlling expression of an iteration-statement shall have
essentially Boolean type.)
Use `do { ... } while (false)' instead of `do { ... } while (0)'.
Use comparisons with zero instead of implicitly testing integers.
The commit is a subset of the original auditable-branch commit:
5d02614e34a86b549c7707d3d9f0984bc3a5f22a
Signed-off-by: Simon Hein <SHein@baumer.com>
- If the HAL headers expose the USBPRE flag, then we're probably
dealing with a F103 - using the same fundamental logic as the
code for the OTG models, set-up the USB Prescaler correctly.
- Fixes#47146
Signed-off-by: Chris Collins <chris@realsimgear.com>
On SoCs with two USB controller, Zephyr selects the FS one, and in that
case the ULPI clock needs to be disabled in sleep mode for the
controller to work.
On SoCs with a single USB HS controller, this operation is not needed.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The ULPI clock uses slightly different names on STM32H7X as those
SoC can have more than one USB controller.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
On STM32H7X SoCs, both USB controllers are HS capable, however it is not
possible to connect an ULPI PHY to the USB2 one, which limit it to FS.
Therefore the ULPI clock in sleep mode has to be disabled for USB2 in
all cases. The ULPI clock for USB1 is already disabled when needed in
the usb_dc_stm32_clock_enable() function like for other SoCs.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Do not manually enable the USB clock on STM32H7, as it is already done
in usb_dc_stm32_clock_enable() using the values in the device tree.
This partially fixes the build for STM32H7 devices with a single USB
controller.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The ULPI clock is disabled by default in run mode, there is no need to
disable it again. It is however enabled by default in sleep/low power
mode, so it needs to be disabled.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Adds few missing zephyr/ prefixes to leftover #include statements that
either got added recently or were using double quote format.
Signed-off-by: Fabio Baltieri <fabiobaltieri@google.com>
For some reason, rather than testing for the presence of
disconnect-gpios in the devicetree, the STM32 USB Device driver was
relying on a hidden Kconfig flag to be set.
This patch removes the Kconfig option completely and simply tests for
the DT property - if it's set, you obviously know what you're doing and
obviously need the pull-up GPIO behaviour.
Signed-off-by: Chris Collins <chris@realsimgear.com>
Adding support for the GIC_V1 to the dc_dw USB driver
to be used by Cyclone V SoC FPGA Development Kit
Signed-off-by: Esteban Valverde <esteban.valverde.vega@intel.com>
In order to bring consistency in-tree, migrate all drivers to the new
prefix <zephyr/...>. Note that the conversion has been scripted, refer
to #45388 for more details.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
This add support to pinctrl at Atmel sam0 usb dc driver. It updates
all boards with new pinctrl groups format and drop pinmux entries.
Signed-off-by: Gerson Fernando Budke <nandojve@gmail.com>
This update Atmel sam usb drivers to use pinctrl driver and API. It
updates all boards with new pinctrl groups format.
Signed-off-by: Gerson Fernando Budke <nandojve@gmail.com>
Enable second usb EHCI if usb2 node has status="okay" Note that this
driver is still an single instance driver, this change simply enables
the driver to work with the usb2 peripheral if that one is enabled, and
usb1 is disabled.
Signed-off-by: Daniel DeGrasse <daniel.degrasse@nxp.com>
According to Kconfig guidelines, boolean prompts must not start with
"Enable...". The following command has been used to automate the changes
in this patch:
sed -i "s/bool \"[Ee]nables\? \(\w\)/bool \"\U\1/g" **/Kconfig*
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
MCUX usb ISR was making usb callbacks directly, which caused assertion
failures when a callback attempted to lock a mutex. Move USB callback
handler to separate thread, and make ISR notify thread via message
queue.
Fixes#40638
Signed-off-by: Daniel DeGrasse <daniel.degrasse@nxp.com>
The PINCTRL_DT_(INST_)DEFINE macros already defined the trailing ;,
making its usage inconsistent with other macros such as
DEVICE_DT_DEFINE.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
This flag may cause host to more ambitious power saving behavior,
not all Zephyr USB classes or applications may be ready for this
now. Allow to disable option USB_DEVICE_REMOTE_WAKEUP.
Signed-off-by: Johann Fischer <johann.fischer@nordicsemi.no>
Use the new pinctrl API to configure pins.
Additionally, rename usb_pinctrl to usb_pcfg to better fit
new pinctrl API.
Signed-off-by: Erwan Gouriou <erwan.gouriou@linaro.org>
Fill in received bytes in USBIP_RET_SUBMIT packet
Currently this field is unconditionally set to 0 and therefore not
filled in conveniently. This may mislead the USB host which is correctly
acknowledged that the transaction was sucessful but it cannot check the
actual received bytes.
Test application:
```python
import usb
data = (0xFF, 0xFF)
print("Opening loopback device")
device = usb.core.find(idVendor=0x2FE3, idProduct=0x0009)
print("Writing test data", data)
written = device.write(0x1, data)
print("Written", written, "bytes")
```
Before:
```
$ ./test_loopback.py
Opening loopback device
Writing test data (255, 255)
Written 0 bytes
```
After:
```
$ ./test_loopback.py
Opening loopback device
Writing test data (255, 255)
Written 2 bytes
```
Signed-off-by: Raúl Sánchez Siles <rsanchezs@k-lagan.com>
Use DT_INST_ENUM_IDX_OR and always default to full-speed
if CONFIG_USB_DC_HAS_HS_SUPPORT is not set or maximum-speed
property is not defined.
Remove low-speed setting since device stack does not support it.
Signed-off-by: Johann Fischer <johann.fischer@nordicsemi.no>
In the current USB device support, the sizes of bulk endpoint
are mostly configure through Kconfig and do not care if a device
is high-speed capable. The information if a USB device controller
supports high-speed comes from devicetree. Add a Kconfig option to
map this information and configure bulk endpoint sizes
accordingly.
Signed-off-by: Johann Fischer <johann.fischer@nordicsemi.no>
Check return value from recv() passed by usbip_recv().
Fixes: #39849Fixes: #39869
Coverity-CID: 240221
Coverity-CID: 240244
Signed-off-by: Johann Fischer <johann.fischer@nordicsemi.no>