EC 4.0 Firmware
Ver: 7 0.5.01
- Add support for new Simplo LiPoly battery
- Remove support for NiMH battery (which was never used in mass production, but was made in small quantities for testing; SKU295 through to SKU298), see further change detail, and workaround for old prototypes.
- Change default compiler to SDCC
- Requires new kernel, see kernel patch
Ver: 7 0.4.12
- emit a new event when the power button is released, so a kernel can time the press.
- increase the hard powerdown button press timeout from 2 to 5 seconds to give user-level more flexibility when looking for long presses.
Ver: 7 0.4.11
- reinit AC detect timer when leaving forced discharge mode, to avoid erroneous low-battery power-downs. this mode is only during testing (runin, and battery-capacity tests).
- this release didn't make it into the next release of Open Firmware, Q7B37. --Quozl 01:35, 26 February 2014 (UTC)
Ver: 7 0.4.10
- Fix for #12704, which caused the power LED to behave incorrectly after short resumes.
Ver: 7 0.4.09
- Reduce post-reprogramming power draw. After an EC reflash, an extra flash of the LEDs will be seen the first time the laptop is powered off while on battery only.
Ver: 7 0.4.08
- Change runin discharge disable so its ahead of poweroff
- Fix stupid mistake in commit b7e7fd817c86
Ver: 7 0.4.07
- NiMH charging in very cold temperature tweaks
- Disable the 2 stop mode problem checks
- Reduce stop mode power draw
Ver: 7 0.4.06
- Allow monitor DEBUG_CMDS to enable support functions
- Drop max LiFe charge rate to 2.5A
- Fix runin discharge XO-4 C2 crash
- ec_irq: add support for suppressing the fallback pulse, for debug
- ec_irq: move the remaining EC timing support from monitor.c to ec_irq.c
Ver: 7 0.4.05
- debug support for wakeup issues
Ver: 7 0.4.04
- further changes for managing keyboard and touchpad flow control, fixing yet another lockup-on-resume scenario.
Ver: 7 0.4.03
- fallout from the changes in 0.4.02 caused keyboard and touchpad flow control to not be asserted in some cases. this in turn caused SoC lockups on resume from suspend.
Ver: 7 0.4.02
- fix EC/SoC communication problems (#12525).
Ver: 7 0.4.01
- further lengthen EC_IRQ# timings for 1Ghz machines
Ver: 7 0.4.00
Ver: 7 0.3.13
- revert wakeup timings to 3.0.11 values
Ver: 7 0.3.12
- ec_irq: add fallback to bail us out of incomplete suspends
- ec_irq: implement timing tuning commands in monitor
Ver: 7 0.3.11
- Allow battery debugging to be enabled via serial port
- Fix incorrect battery critical LED state
- Change W to mW in battery debug output to make the scale clearer
- Add debounce for the thermal alarm signal and enable a new LED state when triggered. (Alternating power LED and red battery LED)
- Blink code for high battery current prior to stop mode (Alternating Green/Yellow)
- Blink code for exiting stop mode while power is off (Alternating Green/Red)
Ver: 7 0.3.10
- Fix NiMH pre-maturely marked full problem
Ver: 7 0.3.09
- Fix logic bug in ext power detect after runin discharge. Machines turning off runin discharge could fail to sense external power.
- Print ACR value and SOC when going into stop mode
- Refactor critical and low flag handling to fix corner cases. There were several corner cases where you could get the previous code to display the wrong LED state.
- Add support for GP LiFePO4 batteries. There are few around the Twine offices.
- Turn off battery state debugs in LiFePO4 code
- kbdwatch: wait longer after seeing SOC_SLEEP to release flow-control. Fixes a lockup on 1Ghz MMP3.
Ver: 7 0.3.08
- Make sure full flag is clear when SOC is < 95%
- Up bad ACR read difference threshold to 20 counts
- Come out of stop mode once in a while, to do housekeeping
- adjust the timings for the EC_IRQ assertions
Ver: 7 0.3.07
- Check battery EEPROM bank 0 values and fix them if wrong
- Detect bad ACR register reads from the battery due to ESD and discard them
- Generate 2 wake up interrupts to the SoC on resume. One to make it wake up and then a 2nd to actually make it start executing. Works around resume problems.
Ver: 7 0.3.06
- Change battery over temperature indicator to be a pulse of dark (both LEDs off) rather than a pulse of red. Red is now exclusively for errors.
- Fix typo in IO map causing LV_SET code to mess with EC_IRQ.