Upstream kinetis

Approved-by: Gregory Nutt <gnutt@nuttx.org>

   Initalize SPI1 on connector J6. No real use, as of yet.

testbuild:Added -x to fail build on errors for CI

Approved-by: Gregory Nutt <gnutt@nuttx.org>

   On CI we want to know ASAP of a failure. This adds the
   -x (exit on build failures) option to faclitate that
   behavior.

   Use ${MAKE} ${MAKE_FLAGS} for make invocation.
   When -x is provided change MAKE_FLAGS to
   --silent --no-print-directory and set -e

   Ignore exit status when using grep for checking
   for CONFIG_NXWM=y

mtd/w25: wait for BUSY flag to clear in w25_readid and w25_unprotect.  W25Q128 datasheet says that all instructions expect 'Read Status Register' and 'Erase/Program Suspend' are ignored when BUSY flag in status register is '1'. Therefore wait for busy flag to clear in w25_readid() and w25_unprotect().

stm32/stm32l4 PWM: While attempting to output a 70 MHz square wave from the timer output of a STM32 clocked at 140 MHz (which works fine in baremetal C), I stumbled on what I believe to be an error in arch/arm/src/stm32/stm32_pwm.c.  Line 1304 we are told that

reload = timclk / info->frequency;

which I belive to be incorrect, it should be

reload = timclk / info->frequency - 1;

since starting to count from 0, if I want to output half of the TIM clock, I must count to 1 and not to 2.

Surely enough, the original code did output 140/3=47 MHz, while this correction does allow the output up to 70 MHz.

I am not sure this affects most users generating slow PWM (e.g. PX4) but for frequencies
close to the PCLK, indeed the difference becomes significant.

stm32_hrtim: GPIOs configuration + EEV and FAULT strucutres

Approved-by: Gregory Nutt <gnutt@nuttx.org>

MTD M2PX: If we READ while a write/erase is pending, the command is ignored and the write/erase continues. If we dont catch this situation we will return garbage to the user because the flash will not execute the command.  So READ MUST wait for write completion, and before that, the bus must be locked since it's a precondition to calling waitwritecomplete().

Revert "MTD: Fix problems in SPI locking in mp25x.c driver.  Same problem probably in exists in several other drivers that derive from this this driver as well."

This reverts commit 85645284.

MTD: Fix problems in SPI locking in mp25x.c driver.  Same problem probably in exists in several other drivers that derive from this this driver as well.

Based on the last PR, review all serial driver vector attachment.  Found one additional error and updated all relevant drivers to current interrupt parameter passing.

i.MX6: Fix a wrong parameter passed when calling irq_attach() in imx_serial.c

Approved-by: Gregory Nutt <gnutt@nuttx.org>

stm32_hrtim: add character driver

Approved-by: Gregory Nutt <gnutt@nuttx.org>

configs:  Add readme for B-L465E-IOT01A Development kit.  This port is currently under consideration (but may not actually happen)

stm32f103-minimum: Fix a small BUG when reading from output pin.  We need a different read_ops to read from output pin. This patch fixes the issue:

nsh> gpio -o 0 /dev/gpout1
Driver: /dev/gpout1
  Output pin:    Value=1
  Writing:       Value=0
  Verify:        Value=1

Now:

nsh> gpio -o 0 /dev/gpout1
Driver: /dev/gpout1
  Output pin:    Value=1
  Writing:       Value=0
  Verify:        Value=0

Whoops -- that #include <cxxabi.h> wasn't supposed to sneak in there. If present, it may already define a __cxxabiv1::__guard that we could use, but that file comes from libstdc++, and I don't think the NuttX buildroot-based toolchain would have that, which is why we need libxx in the first place.

Fix C++ __guard implementation for ARM.  The standard C++ ABI that most platforms follow defines __guard to be 64 bits.  The existing implementation of libxx_cxa_guard.cxx follows this.  However, the 32-bit ARM C++ ABI defines it as 32 bits instead, and changes the meaning slightly so only the lowest bit is used. This matters because GCC creates guard symbols without regards to what libxx_cxa_guard.cxx says.  So on ARM, gcc allocates 4 bytes, but __cxa_guard_release writes 8 bytes, zeroing out another unlucky variable nearby. Fix it by special-casing 32-bit ARM in libxx_cxa_guard.