This the first step in moving the collection of layer 2 netstats from
the low level driver to a central location, ie. gnrc_netif, to avoid
code duplication.
Add command mode that will be used like this:
mode <dev> <data bits> <parity> <stop bits>
This command must be called after init otherwise
the UART won't be fully initialized.
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Add support for specifying data bits, stop bits and parity at
runtime.
Introduce feature periph_uart_modecfg for uart_mode() till all
other CPUs implement it.
STM32 L1, F1, F2, F4 supports following modes:
* 7E1, 7E2
* 7O1, 7O2
* 8N1, 8N2
* 8E1, 8E2
* 8O1, 8O2
STM32 L0, L4, F0, F3, F7 supports following modes:
* 6E1, 6E2
* 6O1, 6O2
* 7E1, 7E2
* 7O1, 7O2
* 7N1, 7N2
* 8N1, 8N2
* 8E1, 8E2
* 8O1, 8O2
Use USART_CR1_M1 macro to detect 7-bit support because
even inside one family there could be devices that don't
support 7-bit mode. So just using a family macro is not
enough.
As stated in the datasheets for L0, L4, F0, F3, F7 devices,
data bits can only be changed when UART is disabled (UE=0).
Introduce uart_stop() routine to satisfy this requirement.
STM32 UART adds parity to the MSB of a byte to send. The same
also applies to the received bytes. As a result this bit must
be masked in order to get the pure data.
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
As suggested in PR#5899 add a routine uart_mode() that will
setup data bits, stop bits and parity at runtime.
uart.h provides a set of enums defining these settings and each
platform will override them to specify values corresponding to
its configuration registers.
The idea behind the enums is to specify default settings i.e. 8N1
through the 0 value item. Invoking uart_mode(uart, 0, 0, 0) will
setup 8N1 mode.
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
If netdev_driver_t::recv() is called and the provided buffer is smaller than
the frame then `-ENOBUFS` should be returned, the frame should be dropped, and
no data of the frame should be returned.
Addresses: https://github.com/RIOT-OS/RIOT/issues/10413
Checking by the send function that at least two maximum size Ethernet frames fit in the remaining heap before the LwIP packet buffer is allocated seems to increase stability. This can be caused by the fact that WLAN hardware interrupts allocate additional memory when receiving a frame during the send attempt.
Error case:
1. thread_yield_higher() stores the thread's ucontext
2. creates an "isr ucontext" for isr_thread_yield, switches to it
Case 1: no signals are pending, continues in isr_thread_yield()
3a. sched_run is called
4a. return to sched_active_thread ucontext
Case 2: signals pending (the crashing scenario), continues in native_irq_handler()
3b. handles signals
4b. if sched_context_switch_request is set, call sched_run
5b. return to sched_active_thread ucontext
4b misses the call to sched_run(), leading to a possible return into a
non-ready thread.
The msp430 toolchain is missing an `fputs()` implementation. This commit makes
them use the `printf("%s", str);` instead of `fputs(str, stdout);`, which is
semantically equivalent (but has more overhead).
