| Commit message (Collapse) | Author | Age |
|---|
| |
|
|
|
|
|
|
|
|
| |
This gpio subsystem keeps track of the GPIO pins which
have been reserved and takes care of the housekeeping
with keeping them running.
This gpio subsystem also knows which alternate functions
belong to which pins, so it can automatically configure
the pins for the alternate functions.
|
| |
|
|
|
|
| |
Before this fix, hfree would neglect to set the prev pointer in the next
used block and such was leaving the prev pointers invalid after
coalescing frees.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
What was in core/ is now moved to arch/stm34l4xxx/peripherals. This new
directory is *supposed to* to contain raw header files defining just the
pertinent register structures for the various peripherals. Peripheral
management belongs somewhere in the new `kern/..` directories. This is
not completely the case at the moment, so more refactoring needs to be
done.
What was sitting in the root has now been moved into the kern/
directory. The kern/ directory is to contain everything else
other than raw device register definitions. The root of the kern/
tree is reserved for standard library-esque headers.
The kern/<peripheral> directory contains management systems for that
peripheral. (At the moment DMA is the only peripheral with a decent
management system.) Preferably these peripheral systems should only
include their correlating header in arch/stm34l4xxx/peripherals, and
use other management systems for handling other peripherals rather
than manipulating their raw registers directly. (Though this ideal
will require much more critical mass of management systems.)
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
| |
The new halloc() call allocates memory on the
STM32l's SRAM2 starting right above the DATA
section.
The implementation uses a very-dense, albeit slower, linked-list
allocation as opposed to fancy B-trees or something. However, the
overhead is just 1 32-bit word per allocation and thus allows for
reasonably dense memory-packing on the small 16K memory chip.
|
| |
|
|
|
| |
This abstraction makes it much more intuitive to use the
DMA features on the STM32L4 boards.
|
| | |
|
| |
|
|
| |
and the rest of the project.
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
| |
Apparently people really don't like bitfields in practice,
so I have devised a compromise to retain some readability while
using bitmasks instead by writing a regset() macro.
|
| |
|
|
| |
subdirectory.
|
| |
|
|
|
|
|
|
|
|
|
| |
- The init() function renamed to on_reset()
- on_reset() now responsible for tight-looping at the end
- on_reset() now set the VTable offset to the base of the FLASH
- included exhaustive list of irqs in isrs.i
- interrupt routines by default flash a code indicating their
isr number.
- interrupt routines are weak-linked allowing the programmer
to override them at-will.
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
testing_harness with fake environment to allow testing on x86
development machines.
|
| |
|
