The TIM2 timer is expected to run at 1 MHz, but it's clocked by the
CPU clock. The current code expects the CPU runs with 48 MHz.
Introduce TIM2_CLOCK_SPEED and calculate a proper divider.
This allows a much lighter implementation of led_indicate_trx(), which
is called for every frame !
Test with 'cangen -L0 -g0 -p10 can0'
After : 13174 frames/sec
Not necessary to spam it every mainloop. Currently throughputs of 17-18k
frames per second is realistic, which means function calls from mainloop
are very frequent and must be "short" compared to the processing time
for each frame.
Convert from the queue implementation to a Linux compatible list
implementation. Get rid of This way we can avoid dynamic memory
allocation altogether.
As the struct gs_host_frame is not placed into the data segment,
requirements for the static memory grows. On the low end processors
the heap reserved in the linker file is too big, resulting in a linker
error. On STM32F042 and STM32F072 set a HEAP size of 0 bytes.
It turned out, that the locked functions added in
61baf8ce9039 ("list.h: add locked variants") are not needed. Remove
these and replace them by locked variants of functions needed in the
upcoming patches.
Place the struct gs_host_frame msgbuf directly into
USBD_GS_CAN_HandleTypeDef. Get rid of another calloc().
This also requires a tweak to heap size in order to fit on the smaller
targets.
USBD_GS_CAN_PrepareReceive() is only used in usbd_gs_can.c, remove
function prototype from usbd_gs_can.h, mark as static, and move in
front to avoid forward declaration.
Marking structures as packed as the following 2 results:
- the compiler doesn't insert any holes in the structures
e.g. if variables are not naturally aligned
- the compiler doesn't assume the base address of the struct to be
naturally aligned
The 2nd effect results in the compiler (depending on the target
architecture) to generate bad code, as unaligned access is not
possible on all archs.
Mark the structs as aligned to 4 bytes, using the not-so-beautiful
__ALIGN_BEGIN and __ALIGN_END macros.
This saves some bytes and generates probably faster code.
| add/remove: 0/0 grow/shrink: 0/2 up/down: 0/-468 (-468)
| Function old new delta
| USBD_GS_CAN_EP0_RxReady 684 568 -116
| main 2712 2360 -352
| Total: Before=17138, After=16670, chg -2.73%
There is a pair of USB requests ( GS_USB_BREQ_GET_USER_ID and
_SET_USER_ID) that is not supported by the linux driver, and not used by
any known software :
https://github.com/candle-usb/candleLight_fw/issues/91
Since the firmware returns a unique serial #, there doesn't appear to be
any use case for assigning a user ID. Using a whole flash page for that
purpose can be wasteful on bigger targets .
This removes all the relevant code and adjusts the linker scripts
accordingly.
Looking at the queue code it turns out, a lot of the functions are not
used. Remove the unused queue functionality. Not sure why the compiler
hasn't removed the malloc() and free() functions entirely, but this
saves some bytes:
| add/remove: 0/3 grow/shrink: 0/0 up/down: 0/-232 (-232)
| Function old new delta
| malloc 20 - -20
| free 20 - -20
| _free_r 192 - -192
| Total: Before=17370, After=17138, chg -1.34%
Move CAN_INTERFACE and CAN_CLOCK_SPEED to the config file as the CAN
clock can vary based on HW implementation.
Better to add it here than to keep adding defines into the
usbd_gs_can.h file.
CAN_INTERFACE can vary based on the number of channels you are using
so also makes sense to add here even though most chips use the same
interface ID.
In commit
40e1997d45
[1] GS_CAN_FEATURE_GET_STATE has been added to the Linux driver.
Document this bit here accordingly.
The new feature ("GS_CAN_FEATURE_GET_STATE") indicates that the state
of the CAN controller can be queried with the new
GS_USB_BREQ_GET_STATE control message.
[1] https://git.kernel.org/torvalds/c/40e1997d4551e6053fc2f61959628da474775dc5
In commit
1f1835264d
[1] Bus Error Reporting has been added to the Linux driver. Document
this bit here accordingly.
The new feature ("GS_CAN_FEATURE_BERR_REPORTING") indicates that the
bus error reporting in the CAN controller can switched on and off with
the GS_CAN_MODE_BERR_REPORTING mode bit in the GS_USB_BREQ_MODE
control message.
[1] https://git.kernel.org/torvalds/c/1f1835264d81da03fcd05646542a622678b379d6
Since all bit errors already trigger sending a error msg,
remove the code triggering on counters going upwards, since they
are send anyway. See [1] "This if statement doesn't normally do
anything", get rid of it.
[1] https://github.com/candle-usb/candleLight_fw/pull/61
Add a USB message which controls a GPIO for a specified CAN bus.
The GPIO activates a SSR to connect a 120Ohm terminating resistor
to the specified CAN bus.
Signed-off-by: Daniel Trevitz <daniel.trevitz@wika.com>
with help from include-what-you-use (iwyu-tool). It's not quite clever
enough to understand our intent of including "hal_include.h" but still
offers good suggestions.
