- the timeout computation for the spurious IRQ test confused numerator
and denominator in a fraction
- the timeout offset between timer channels was hardcoded to 5000 from
when the timer was only tested with 1 MHz as frequency
- This resulted in slooooow test runs when running at slow
frequencies
- fix overflow handling in the spinning wait
- likely this would never overflow anyway assuming that
`timer_init()` resets the counter value, but let's not rely on
this and just fix the bug for good
Assertions are used instead of returning a NULL pointer to detect errors in the MTD definition and access in the case that the return value is not evaluated.
When using `make flash` or `make term` with `MOST_RECENT_PORT=1`, RIOT
should now only consider plausible TTYs and select the most recently
connected one. The filter is a bit broad, though, and will match
any board with a CH430 UART2USB bridge. But depending on the boards
attached, it might be good enough to reliably tell it apart from
other boards.
When using `make flash` or `make term` with `MOST_RECENT_PORT=1`, RIOT
should now only consider plausible TTYs and select the most recently
connected one. The filter is a bit broad, though, and will match
any board with a CH430 USB 2.0 UART2USB bridge. But depending on the
boards attached, it might be good enough to reliably tell it apart from
other boards.
When using `make flash` or `make term` with `MOST_RECENT_PORT=1`, RIOT
should now only consider plausible TTYs and select the most recently
connected one. The filter is a bit broad, though, and will match
any board with a CP2104 USB2UART bridge. But depending on the boards
attached, it might be good enough to reliably tell it apart from other
boards.
When using `make flash` or `make term` with `MOST_RECENT_PORT=1`, RIOT
should now only consider plausible TTYs and select the most recently
connected one. The filter is a bit broad, though, and will match
any board with a CP2102N USB2UART bridge. But depending on the boards
attached, it might be good enough to reliably tell it apart from other
boards.
When using `make flash` or `make term` with `MOST_RECENT_PORT=1`, RIOT
should now only consider plausible TTYs and select the most recently
connected one. The filter is a bit broad, though, and will match
any board with a CP2102 USB2UART bridge. But depending on the boards
attached, it might be good enough to reliably tell it apart from other
boards.
When using `make flash` or `make term` with `MOST_RECENT_PORT=1`, RIOT
should now select relatively robustly the most recently connected
ESP32 Ethernet Kit / ESP32 Wrover Kit board, even with other TTYs
present. This is because the FTDI Dual RS232-HS with two USB2UART
bridges in the chip would be an expensive overkill for most boards,
unless one is using the FTDI chip to big-bang JTAG as well - as done
on the Ethernet Kit and Wrover Kit boards.
If the feature periph_timer_query_freqs is provided, use the
corresponding function to iterate over all supported frequencies to
increase test coverage.
Allow accessing supported timer frequencies with a dedicated API.
This API needs to be implemented per platform and is available with
the feature periph_timer_query_freqs.
This test sporadically failed on multiple machines compiled with both
GNU and LLVM toolchains. To not waste more CI time by sporadically
failing unrelated CI runs, let's disable the test for now.
This fixes issues with `make compile-commands`, as this by default
generates clangd/LLVM compatible flags. Without `TARGET_ARCH_LLVM`
exported, this will fall back to `TARGET_ARCH`, which is different
for LLVM and GCC.
