* tmk_core/common: Fixing TIMER_DIFF macro to calculate difference correctly after the timer wraps.
Let's go through an example, using the following macro:
If the first timer read is 0xe4 and the second one is 0x32, the timer wrapped.
If the timer would have had more bits, it's new value would have been 0x132,
and the correct difference in time is 0x132 - 0xe4 = 0x4e
old code TIMER_DIFF_8(0x32, 0xe4) = 0xff - 0xe4 + 0x32 = 0x4d, which is wrong.
new code TIMER_DIFF_8(0x32, 0xe4) = 0xff + 1 - 0xe4 + 0x32 = 0x4e, which is correct.
This also gives a chance for a smart compiler to optimize the code using normal
integer overflow.
For example on AVR, the following C code:
uint8_t __attribute__ ((noinline)) test(uint8_t current_timer, uint8_t start_timer)
{
return TIMER_DIFF_8(current_timer, start_timer);
}
With the original code, it gets translated to the following list of instructions:
00004c6e <test>:
4c6e: 98 2f mov r25, r24
4c70: 86 1b sub r24, r22
4c72: 96 17 cp r25, r22
4c74: 08 f4 brcc .+2 ; 0x4c78 <test+0xa>
4c76: 81 50 subi r24, 0x01 ; 1
4c78: 08 95 ret
But with this commit, it gets translated to a single instruction:
00004c40 <test>:
4c40: 86 1b sub r24, r22
4c42: 08 95 ret
This unfortunately doesn't always work so nicely, for example the following C code:
int __attribute__ ((noinline)) test(uint8_t current_timer, uint8_t start_timer)
{
return TIMER_DIFF_8(current_timer, start_timer);
}
(Note: return type changed to int)
With the original code it gets translated to:
00004c6e <test>:
4c6e: 28 2f mov r18, r24
4c70: 30 e0 ldi r19, 0x00 ; 0
4c72: 46 2f mov r20, r22
4c74: 50 e0 ldi r21, 0x00 ; 0
4c76: 86 17 cp r24, r22
4c78: 20 f0 brcs .+8 ; 0x4c82 <test+0x14>
4c7a: c9 01 movw r24, r18
4c7c: 84 1b sub r24, r20
4c7e: 95 0b sbc r25, r21
4c80: 08 95 ret
4c82: c9 01 movw r24, r18
4c84: 84 1b sub r24, r20
4c86: 95 0b sbc r25, r21
4c88: 81 50 subi r24, 0x01 ; 1
4c8a: 9f 4f sbci r25, 0xFF ; 255
4c8c: 08 95 ret
Wth this commit it gets translated to:
00004c40 <test>:
4c40: 28 2f mov r18, r24
4c42: 30 e0 ldi r19, 0x00 ; 0
4c44: 46 2f mov r20, r22
4c46: 50 e0 ldi r21, 0x00 ; 0
4c48: 86 17 cp r24, r22
4c4a: 20 f0 brcs .+8 ; 0x4c54 <test+0x14>
4c4c: c9 01 movw r24, r18
4c4e: 84 1b sub r24, r20
4c50: 95 0b sbc r25, r21
4c52: 08 95 ret
4c54: c9 01 movw r24, r18
4c56: 84 1b sub r24, r20
4c58: 95 0b sbc r25, r21
4c5a: 93 95 inc r25
4c5c: 08 95 ret
There is not much performance improvement in this case, however at least with this
commit it functions correctly.
Note: The following commit will improve compiler output for the latter example.
* tmk_core/common: Improve code generation for TIMER_DIFF* macros
Because of integer promotion the compiler is having a hard time generating
efficient code to calculate TIMER_DIFF* macros in some situations.
In the below example, the return value is "int", and this is causing the
trouble.
Example C code:
int __attribute__ ((noinline)) test(uint8_t current_timer, uint8_t start_timer)
{
return TIMER_DIFF_8(current_timer, start_timer);
}
BEFORE: (with -Os)
00004c40 <test>:
4c40: 28 2f mov r18, r24
4c42: 30 e0 ldi r19, 0x00 ; 0
4c44: 46 2f mov r20, r22
4c46: 50 e0 ldi r21, 0x00 ; 0
4c48: 86 17 cp r24, r22
4c4a: 20 f0 brcs .+8 ; 0x4c54 <test+0x14>
4c4c: c9 01 movw r24, r18
4c4e: 84 1b sub r24, r20
4c50: 95 0b sbc r25, r21
4c52: 08 95 ret
4c54: c9 01 movw r24, r18
4c56: 84 1b sub r24, r20
4c58: 95 0b sbc r25, r21
4c5a: 93 95 inc r25
4c5c: 08 95 ret
AFTER: (with -Os)
00004c40 <test>:
4c40: 86 1b sub r24, r22
4c42: 90 e0 ldi r25, 0x00 ; 0
4c44: 08 95 ret
Note: the example is showing -Os but improvements can be seen at all optimization levels,
including -O0. We never use -O0, but I tested it to make sure that no extra code is
generated in that case.OA
* quantum/debounce: Fix custom wrapping timers in eager_pr and eager_pk debounce algorithms
Please see the below simulated sequence of events:
Column A is the 16-bit value returned by read_timer();
Column B is the value returned by custom_wrap_timer_read();
Column C is the original code: (timer_read() % MAX_DEBOUNCE)
A, B, C
65530, 19, 30
65531, 20, 31
65532, 21, 32
65533, 22, 33
65534, 23, 34
65535, 24, 35
0 25, 0
1, 26, 1
2, 27, 2
3, 28, 3
4, 29, 4
5, 30, 5
read_timer() wraps about every 1.09 seconds, and so debouncing might
fail at these times without this commit.
* quantum/debounce/eager_pr and eager_pk: modifications for code readability according to code review.
* quantum/debounce/eager_pr and eager_pk: modifications for code readability according to code review. (2)
* Define NO_ACTION_MACRO/FUNCTION in header instead of makefile when LTO is enabled
Currently, boards and keymaps that define NO_ACTION_MACRO/FUNCTION unconditionally
will not compile with LTO_ENABLE (#8604). This fixes the issue by moving the
definitions from common.mk to action.h, which enables us to check for previous
definitions of those macros (this cannot be done in a makefile).
* Remove LTO checks in templates
Since now NO_ACTION_MACRO/FUNCTION are defined as needed in action.h (which is
included by quantum.h), checking for LTO in keyboard and user code is no
longer required.
* Update LTO_ENABLE docs
* Update tmk_core/common/progmem.h
Co-Authored-By: Ryan <fauxpark@gmail.com>
* Update quantum/rgblight.c
Co-Authored-By: Ryan <fauxpark@gmail.com>
* fixed problem with implicit declaration in quantum/rgblight.c (#8381)
Co-authored-by: Ryan <fauxpark@gmail.com>
* Port over some AVR backlight logic to SLEEP_LED
* Port over some AVR backlight logic to SLEEP_LED - add timer 3
* Port over some AVR backlight logic to SLEEP_LED - clang format
* Enable SLEEP_LED within vusb protocol
* Improve process_record system
Code based on @colinta's
* Rename and better handle functions
* Fix incorrect function call to process_record_user
* Add documentation for post_process_record
* Add both get_event_keycode and get_record_keycode functions
And add some comments about these functions
* Update code format
* Cleanup merge artifacts
* Add support for Bootmagic lite when using SPLIT_HAND_PIN
* Deduplicate bootmagic_lite logic from within via
* Revert location of defaults so that user overrides still work for now
* Tidy up code slightly
* Fix extra keyboard report during test_fixture teardown
* Add tests for pressing two keys with only different modifers
* Fix#1708
When two keys that use the same keycode, but different modifiers were
pressed at the same time, the second keypress wasn't registered. This is
fixed by forcing a key release when we detect a new press for the same
keycode.
* Fix the NKRO version of is_key_pressed
* Fix uninitalized loop variable
Co-authored-by: Jack Humbert <jack.humb@gmail.com>
* Moved contents of rgblight_reconfig.h to rgblight_post_config.h.
In #3582, rgblight_reconfig.h had to be newly created. Now, the build system of qmk_firmware has a post_cofig feature, so that what was done in rgblight_reconfig.h can now be realized in rgblight_post_config.h.
**This commit does not change the build result.**
Testing script
```shell
# build on master
git checkout master
echo master > /tmp/master_md5.txt
# RGBLIGHT_ENABLE = no
make HELIX=verbose helix/rev2:default:clean
make HELIX=verbose helix/rev2:default
md5 helix_rev2_default.hex >> /tmp/master_md5.txt
# RGBLIGHT_ENABLE = yes, with animations
make HELIX=verbose helix/rev2/back:default:clean
make HELIX=verbose helix/rev2/back:default
md5 helix_rev2_back_default.hex >> /tmp/master_md5.txt
# RGBLIGHT_ENABLE = yes, without animations
make HELIX=verbose,no_ani helix/rev2/back:default:clean
make HELIX=verbose,no_ani helix/rev2/back:default
md5 helix_rev2_back_default.hex >> /tmp/master_md5.txt
# build on refactor_rgblight_reconfig.h
git checkout refactor_rgblight_reconfig.h
echo refactor_rgblight_reconfig.h > /tmp/branch_md5.txt
# RGBLIGHT_ENABLE = no
make HELIX=verbose helix/rev2:default:clean
make HELIX=verbose helix/rev2:default
md5 helix_rev2_default.hex >> /tmp/branch_md5.txt
# RGBLIGHT_ENABLE = yes, with animations
make HELIX=verbose helix/rev2/back:default:clean
make HELIX=verbose helix/rev2/back:default
md5 helix_rev2_back_default.hex >> /tmp/branch_md5.txt
# RGBLIGHT_ENABLE = yes, without animations
make HELIX=verbose,no_ani helix/rev2/back:default:clean
make HELIX=verbose,no_ani helix/rev2/back:default
md5 helix_rev2_back_default.hex >> /tmp/branch_md5.txt
diff -u /tmp/master_md5.txt /tmp/branch_md5.txt
```
Test result:
```
--- /tmp/master_md5.txt 2020-01-03 15:42:22.000000000 +0900
+++ /tmp/branch_md5.txt 2020-01-03 15:42:42.000000000 +0900
@@ -1,4 +1,4 @@
-master
+refactor_rgblight_reconfig.h
MD5 (helix_rev2_default.hex) = f360032edd522448366d471d8f4f8181
MD5 (helix_rev2_back_default.hex) = 0c663acc6cccc44476b3b969ad22a48f
MD5 (helix_rev2_back_default.hex) = e66b1195ff6d38e6e22c975b8ae42fd3
```
* Expressions that are too long are difficult to read, so wrap them.
* Edit the expression again
* remove `defined(RGBLIGHT_ANIMATIONS)` in `tmk_core/common/*/suspend.c`, `tmk_core/protocol/*/main.c`
move contents of rgblight_reconfig.h to rgblight.h.
The following changes were made to rgblight.h.
```diff
+#ifdef RGBLIGHT_USE_TIMER
void rgblight_task(void);
void rgblight_timer_init(void);
void rgblight_timer_enable(void);
void rgblight_timer_disable(void);
void rgblight_timer_toggle(void);
+#else
+#define rgblight_task()
+#define rgblight_timer_init()
+#define rgblight_timer_enable()
+#define rgblight_timer_disable()
+#define rgblight_timer_toggle()
+#endif
```
The following changes were made to tmk_core/common/avr/suspend.c, tmk_core/common/chibios/suspend.c, tmk_core/protocol/chibios/main.c, tmk_core/protocol/lufa/lufa.c, tmk_core/protocol/vusb/main.c.
```diff
-# ifdef RGBLIGHT_ANIMATIONS
rgblight_timer_enable();
-# endif
```
```diff
-#if defined(RGBLIGHT_ANIMATIONS) && defined(RGBLIGHT_ENABLE)
+#if defined(RGBLIGHT_ENABLE)
rgblight_task();
#endif
```
* remove 'defined(RGBLIGHT_ANIMATIONS)' in tmk_core/common/keyboard.c
Co-authored-by: Joel Challis <git@zvecr.com>
* uart.c fix from TMK
Backport from tmk/tmk_keyboard@c41e48a0ab0712d2667feb6b5dd8a4d5491cfcc5
* Avoid deadlock when uart.c is usind in ISR
Backport from tmk/tmk_keyboard@55443fabb731459e21b45781c6d951edac5d75f4
* Compile error if ONESHOT_TIMEOUT defined but oneshot disabled
When ONESHOT_TIMEOUT and NO_ACTION_ONESHOT are both defined, this code
fails to compile.
Wrap the one usage of ONESHOT_TIMEOUT that is inconsistent with the
rest so all usages are properly wrapped by a check of NO_ACTION_ONESHOT.
* Run file through clang-format
Co-authored-by: Ted M Lin <tedmlin@gmail.com>
* Option to use function for KEYCODE2 routines.
Convert the KEYCODE2SYSTEM and KEYCODE2CONSUMER macros to functions,
defaulting to using the macros. The function form allows the compiler
to optimize the switch statement itself, over the macro nested
ternaries.
To enable this feature, #define USE_KEYCODE2_FUNCTION.
Testing against a random selection of avr-based keyboards, this
increased available flash by ~500 bytes. For arm-based keyboards,
the available flash increased by ~400 bytes.
* Replace macro with function entirely.
As zvecr states, go bold and just commit to using the function instead
of the macro.
* Reformat whitespace now that functional review is done
Verified against clang-format output.
* Dedupe extrakey report struct, and send functions in V-USB & LUFA
* Doc comment for consistency
* Wrap it in ifdef to prevent unused function error
* Do the same for ATSAM
- uprintf -> dprintf
- Fix atsam "vendor" eeprom.
- Bump Kinetis K20x to 64 bytes, too.
- Rollback Kinetis to 32 bytes as partitioning can only be done once. Add warning about changing the value.
- Change RAM-backed "fake" EEPROM implementations to match eeconfig's current usage.
- Add 24LC128 by request.
* Run clang-format manually to fix recently changed files
* Run clang-format manually to fix recently changed files - revert template files
* Run clang-format manually to fix recently changed files - format off for ascii_to_keycode_lut
Joel Challis
Joshua Moses Diamond
Chris Ko
Konstantin Đorđević
Drashna Jaelre
QMK Bot
Purdea Andrei
Ryan
Casper Weiss Bang
fredizzimo
francislan
Takeshi ISHII
James Young
ridingqwerty
Ted Lin
Ted M Lin
Pavel Župa
Nick Brassel