stm32: serial Allow configuring Rx DMA buffer size

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

Upstream kinetis

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

Upstream kinetis i2c pin def

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

   The output structure of the GPIO for I2C needs to be open
   drain.  When left at the default, one can observe on a scope
   the slave contending with the push-pull during the ACK

   The output structure of the GPIO for I2C needs to be open
   drain.  When left at the default, one can observe on a scope
   the slave contending with the push-pull during the ACK

   The output structure of the GPIO for I2C needs to be open
   drain.  When left at the default, one can observe on a scope
   the slave contending with the push-pull during the ACK

   The output structure of the GPIO for I2C needs to be open
   drain.  When left at the default, one can observe on a scope
   the slave contending with the push-pull during the ACK

   The output structure of the GPIO for I2C needs to be open
   drain.  When left at the default, one can observe on a scope
   the slave contending with the push-pull during the ACK

Fix errors introduced into Kinetis serial when I unsuccessfully tried to correct coding standard violations.  Folks, things will be better for everyone if you just follow that standard.

STM32L4: add support for many new MCUs from STM32L4X3XX product line and Nucleo-L452 board

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

configs/nucleo-l496zg: Kconfig was copied from nucleo-144

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

Removed as most options have not been tested.

(This also removes DPFPU/DTCM/ITCM features again, fixing a
recent git history hickup.)

Not tested for any product family, but now it at least compiles.
L496 devices can have one bit wider Volatile Data Segment.

STM32L4: stm32l4_i2c: change wrong macro to CONFIG_I2C_POLLED

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

GPIO and UART seem similar across STMicro product matrix,
so renamed files accordingly. RCC is cloned just in case,
while conflicting differences there seem to be very minor.

This is modeled after STM32F7. Idea is to declare each chip in Kconfig
but allow for flash size override. Commit adds many STM32L4_HAVE_XXX
feature test macros.