XO 1.5 EC dev changelog
Revision as of 05:47, 9 August 2009 by Wad
1.5 EC development change log
- Catch soft off. Soft off is dictated by SUSC and MAIN_ON (SUSB) when the signals go into a specific state it means that the host has turned itself off and the EC should shut down the voltage regulators.
- Fix the polarity of the SUSC high low reporting. It was inverted.
- Optimize the way the SCI timing reporting was happening the numbers reported by tMO now match the timestamps happening from MO L and MO H.
- Disable sequence warnings fro srec_cat when building the .bin image
- Update the make dev rule so that it copies the test code to 'ec_test.img' in the EC firmware dir.
- Remove code that deals with a register we don't have. Leftover from another code base.
- Compile option to disable timestamp printing. (if you need diagnostic stuff that has to go really fast)
- Rename all the assert/deassert to enable/disable. This makes it so you can search for specific IO ops. Searching for assert_<bla> would also get you deassert_<BLA>.
- General purpose EC timer host wakeups. EC cmd 0x36 takes a signed long value that is the number of ms to wait and then it will issue a SCI 0x01. Allows wakeups with around 8 - 12 ms resolution for about 26 days in the future. (ie 2^31 milliseconds).
- Autowak is synced with MAIN_ON going low to help duplicate a resume crash.
- LiFePO4 debug dumps can now enabled or disabled via OFW.
Ver: 1.9.3 (Unreleased)
- Fix a typo bug that was turing off CC0 rather than CHG
- Clear the 1 wire guage cache values for voltage and current when there is not a respose to a presence pulse.
- Tweak the power up sequecing state machine to remove the PWR_GOOD state since it was unused.
- Make the POWER_BUT_IN# strobe pulse the same width for both power up from off and for the strobe when power is on.
- Add support for setting the value on the CHGCTL pin so we can test this feature.
- Refactor keyboard and touchpad wakeups so that you can tell the difference between the 2 and don't wakeup on keyboard break codes.
- reorder the scancode generation table to match the new ordering from the rewritten GetMatrixKeys.
- Modify the polarity of some signals so they don't create leakage in S5
- Reduce the VR_ON to RSMRST# delay to approx 20ms as requested by Beckham Chen @ quanta.
- Fix no low voltage host shut off bug.
- Battery charge taper current tweak 100mA -> 150mA.
- Refactor 1-w states to break up some of the larger complex states.
- 1 wire state machines are now run every 100ms.
- Battery debugging output was left enabled if you connect up a battery you will get a dump of various variables every second.