Since some Nordic platforms don't use an external clock source, they
need to be built with the TARGET_NRF_LFCLK_RC flag. If they don't,
SystemInit will wait indefinitely for an external clock to start.
This patch adds a 1s timeout to the external oscillator initialisation,
and falls back to using the internal RC one if no LFCLKSTARTED event was
received in this interval.
For most applications, this won't matter. When building for the wrong
Nordic target, pin numbers will be mixed up as well and the application
will break later on.
But on targets that can only be updated using FOTA and don't have an
interface chip, this patch avoid ending up with a bricked device after a
slight mistake. Indeed, BLE DFU service is portable across all boards
and will run even with a mixed-up target setup, allowing the user to
re-install the right application.
An example scenario is transferring an application on a standalone
NRF51 chip, after spending some time prototyping on an mKIT, which tends
to happen a bit too often...
Signed-off-by: Jean-Philippe Brucker <jean-philippe.brucker@arm.com>
This patch clears the EVENT_COMPARE register even when we're not
expecting to get a COMPARE interrupt.
Here is a scenario where we get an unwelcome interrupt, and where it
currently results in a deadlock:
* A Ticker T1 runs at 500ms
* Its handler registers a Timeout T2 at 20µs. Since LF clock runs at
32.768kHz, this will be rounded up to a multiple of 30.5µs (actually
61µs, to ensure we're in the next tick window)
T1 T2 T1' T2'
-------|---|---------------------------------|---|--------->
: : :
<---> n * 30.5µs :
<--------------- 500ms --------------->
* When the ticker API handles T1, it calls us_ticker_set_interrupt a
first time for T1', and then for registering T2.
* Since T2 period is less than one RTC tick (30.5µs), there is a high
chance it will go past *while* we're still in this first handler.
* So ticker_irq_handler also handles T2 as well, and removes it from the
queue.
* us_ticker_set_interrupt is called for T1' before returning from
ticker_irq_handler
The problem resides in the fact that us_ticker_set_interrupt takes more
than 2 RTC ticks to execute: while inside the handler, T2 interrupt will
fire.
* Because of this pending interrupt, RTC1_IRQHandler will be called
right away, but since we removed T2 from the queue, the handler is
waiting for T1' instead, and will not clear EVENT_COMPARE until we
reach the T1' tick, in about 500ms.
* This results in a lock and main isn't executed anymore.
With this patch, we avoid being stuck in RTC1_IRQHandler while waiting
for T1'.
Note: even when T2 isn't handled in the same loop as T1, we may get a
spurious interrupt, because us_ticker_set_interrupt will be called twice
for T2... and will register T2+m the second time, while T2 interrupt is
fired in the background.
That case isn't as harmful, since RTC1_IRQHandler will be waiting for
T2+m, which is only one or two ticks further, and then it won't be
called again before T1'.
Signed-off-by: Jean-Philippe Brucker <jean-philippe.brucker@arm.com>
Happy Gecko's USB Low Energy Mode does not work properly with
isochronous endpoints.
Note that Happy Gecko errata USB_E112 states:
Do not use the SUSPEND mode of LEMOSCCTRL in USB_CTRL.
In rare cases with high data throughput, a transmission can
fail when this mode is enabled.
Use the GATED mode of LEMOSCCTRL for the best energy
efficiency. The NONE mode can be used to disable energy
savings
However, even using GATE mode causes problems for high-frequency
isochronous transfers. The primary effect for OUT endpoints is that
the first read will succeed, and the following one (targetting the
next frame) will fail with zero bytes incoming data, and PKTDRPSTS
is set in the related interrupt status register.
Disabling LEMOSCCTRL (set to NONE) solves the problem. Since this
will cause increased energy usage, do this only when an ISOC endpoint
is added.
USB device support for Wonder, Leopard and Happy Gecko. See
details on configuration etc on the previous commit.
Known issue: USBAudio does not work reliably on Happy Gecko.
** Overview
This commit provides USB device-mode drivers for Silicon Labs chips.
The code is based on the SiLabs USB driver that is part of the
Simplicity dev environment.
Because Mbed USBDevice class wants to control the USB transactions
itself, large parts of the higher level code originally present in the
driver have been stripped out. An attempt has been made to keep Mbed
code separate enough that patches can be applied with minor effort, if
needed.
** Use of dynamic memory
Due to requirements imposed by the USB hardware, the driver needs read
and write buffers for every enabled endpoint. Since information on
what EPs will be used is only available during runtime, dynamic memory
allocation is used to minimize memory usage.
If needed, dynamic memory use can be disabled via a flag in
usbconfig.h. In this case, please read the documentation to see what
limitations this imposes. Static memory usage can also be reduced by
limiting the maximum number of active endpoints - see configuration in
usbconfig.h and USBEndpoints_EFM32.h.
** Power management
Driver supports the new SiLabs functionality in Mbed sleep(). Normal
sleep (EM1) is always available, and deep sleep (EM2) is available
when USB is either disconnected or suspended. How and when different
sleep modes are made possible can be configured in usbconfig.h
** Supported hardware
Current release officially supports, and has only been tested on Giant
Gecko. Upcoming releases will add support for other SiLabs platforms
on Mbed.
Przemek Wirkus
Austin Blackstone
bcostm
Steven Cooreman
Devan Lai
tomoyuki yamanaka
Kevin Gillespie
0xc0170
Jean-Philippe Brucker
sarahmarshy
unknown
modtronix-com
Martin Simpson
dbestm
Jay Francis
Mikko Polojarvi
amveeq
hjjeon0608