The index will never be negative and will never exceed 255. Change them
from i16 to u8 so SDCC will generate more efficient code.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Replace bare int types with stdint types. This was done with:
grep -rwl 'int' src/ | xargs sed -i 's/\<int\>/int16_t/g'
grep -rwl 'unsigned long' src/ | xargs sed -i 's/\<unsigned long\>/uint32_t/g'
grep -rwl 'unsigned char' src/ | xargs sed -i 's/\<unsigned char\>/uint8_t/g'
Then reverted for *main(), putchar(), and getchar().
The Arduino declarations for parallel_main() were also corrected to
match their definitions.
SDCC does *not* generate the same code in all instances, due to `int`
being treated different than `short int`.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Per the charger for the 3050 variants:
DC OUTPUT (输出/輸出): 19.5V⎓7.7A 150W
The 3060 charger does list 9.23A.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Use u8 instead of i16 as a delay can never be negative and should never
block for a long period of time.
Use a decrementing loop as SDCC generates more efficient code.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Most of the code is the same between IT8587E and IT5570E.
Identical files were detected with:
diff -rs src/ec/it8587e/ src/ec/it5570e/ | egrep '^Files .+ and .+ are identical$'
Then manually moved from one variant and deleted from the other.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Per the specs, a 16-byte signature is stored at 0x40. All proprietary EC
ROMs store a space-padded 16-byte vendor string at 0x50, but it isn't
required.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Logic for changing power limit is the same for all boards with GPUs. It
is still called from board_event() instead of in power_event() to
maintain the current behavior of checking every main cycle.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
The eSPI logic is the same for all TGL boards using it. None of it is
moved to IT5570E due to the call to power_cpu_reset().
Signed-off-by: Tim Crawford <tcrawford@system76.com>
* SMOOTH_FANS, SMOOTH_FANS_UP, SMOOTH_FANS_DOWN build flags to smooth fan speed changes.
Defaults 40, set SMOOTH_FANS=0 to disable and keep sharp fan speed changes
* fix for ACPI CPU temperature
* allow for a floor to be set for fan smoothing and specifically configure oryp6/7 to start smoothing at 25% to mitigate fan buzzing below 25%
* update default config values for fans
* update all devices to use defaults for heatup, cooldown, update galp5 fan curves
* Decrease default cooldown time from 20 to 10
Co-authored-by: Jacob Kauffmann <jacob@system76.com>
Use of shorter sentences should reduce cognitive load. Example shell
commands should improve understanding of the description.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Internal flashing can result in corrupted or lost OS data. Make this
more explicit for internal flashing, and make it a point for external
flashing.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Give the process from 0 to built. Due to the implications of doing
something wrong (unbootable machine), flashing is kept separate.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
- General copyright notice
- License info for firmware and tools
- Note about ITE datasheets being unavailable
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Explain why there are no tagged releases, and why the commit hash on
official releases does not match commits from this repo.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Use hard line breaks to keep the line under 80 columns. Break it into 2
sentences: one explaining the issue, and the other explaining the fix.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Fix ACPI timeout and delay on S3 resume when holding a key.
A timeout greater than 1ms results in the excessive delay (~10s) still
occurring on an oryp5 (IT8587).
People who are just reading the keyboard customization instructions might overread the dependencies installations sections.
Additionally, installing rustup for the first time requires sourcing cargo's env vars in case the firmware is flashed in the same shell session after installing rustup.
