From Openmoko
- P#797 splash screen / boot menu support for GTA02 missing
- Wifi driver on GTA02
- Samuel will focus on this driver.
- Samuel release driver for kernel 2.6.22.5/2.6.24
- Reported working but iperf chokes it, maybe firmware
- WPA / wpa_supplicant support
- Spinlock debug Panics, posted on kernel list
- Sameo examines it Jan 29th
- New patches Jan30th should fix
- Power management
- Confirmed working on WPA WLAN here 1st Feb (andy)
- Add V5 platform defines
- U-Boot (Werner)
- Kernel (Werner)
- Devirginator support for release partitioning scheme (Werner)
- Werner extends kernel to 8MB and adds 256KB "factory" part
- CPU speed only 211MHz. I#333
- This is high priority for test CPU module. Andy will focus on this problem. (Done)
- Change the default CPU speed to 400MHz.
- BBT(bad block test), low level foo validation for GTA02 I#335
- Because flash image time (4mins) is too long. (Done)
- Werner already has modified this and has committed to openmoko SVN.
- U-Boot initialization race condition: Hangs on boot P#887
- Werner has finished this bug.
- Git trees
- 2.6.24 tree available on git.openmoko.org - Andy (done)
- 2.6.22.5 tree - Andy (done)
- Documentation on internal Wiki - Andy (done)
- Assist Willie
- NOR flash device untested + no MTD support yet P#1072
- Andy focus on this issue. Patches have been sent to kernel mailing list.
- NOR is verified and MTD support working, waiting on devirginator integration
- Splash screen while system is booting. (Andy)
- Init Glamo (done)
- Init LCM (done)
- display bitmap (done)
- 3.5s - 4s deadtime before display
- Micro-SD card verification. I#320
- Andy will focus on this. (Conti.)
- 1-bit communication is verified and cards can be recognized and partition table scanned (done)
- 4-bit mode (done)
- Interrupt mode is working, no spinning for bulk (done)
- MCI driven SD Vdd voltage scaling (done)
- 2.6.24 makes a new problem with bulk read payloads showing missing words (~0.05 probability per word) in the Glamo RAM -- seems failed to be written by hardware yet problem not seen on 2.6.22.5
- requested assistance from S-Media FAE (worked around)
- achieve reliable partition detection / mount (done)
- test bulk write (done)
- low probability 16-bit wide corruption on write blocks... sometimes 2GB okay sometimes can't write a 1MB test file... corrupted AFTER written to Glamo
- requested assistance from S-Media FAE
- problem can be provoked by physical pressure on the 3362
- problem not found on second A4 device
- releasing to get wider testing (done)
- stress tests (done)
- refactor a bit (done)
- JFFS2 Mount time reduction
- Optimization attempt (Werner) done
- 20%/80% rootfs splitting, guaranteed 80% reduction in initial mount -- not needed
- 2.6.24 PBE (Milosch)
- Milosch has this working
- Bluetooth module not recognized -- examine (Willie)
- Willie changed PSU voltage to Bluetooth device
- Andy Confirmed recognized after
echo 1 >/sys/bus/platform/drivers/neo1973-pm-bt/neo1973-pm-bt.0/power_on
echo 1 >/sys/bus/platform/drivers/neo1973-pm-bt/neo1973-pm-bt.0/reset
echo 0 >/sys/bus/platform/drivers/neo1973-pm-bt/neo1973-pm-bt.0/reset
- No Wifi driver in there (done)
- Sameo ported it, included in kernel.git now with config patched (done)
- Test vibrator (Andy)
- Vibrator patch was missing
- needed adjusting for 2.6.24 (done)
- echo <n> >/sys/bus/platform/devices/neo1973-vibrator.0/leds/neo1973\:vibrator/brightness
- Battery charging function (500mA/1000mA) is implemented in kernel driver. (Matt)
- Motion sensors/SPI validation I#334
- last week, Andy and I found the problem, we changed SPI driver. Then we can communication with motion sersor. (Done)
- Motion sensors didn't work very well. It seems be broken.
- 2.6.24 S3C SPI is disabled by comment in platform, unsure if it was ever working
- Werner examines it
- Found the INT signals are stuck low
- Andy examines it
- Sensors are verified (done)
- GPIO SPI broken for multiple chip selects (fixed)
- bug with SPI load vs interrupts, to be solved by FIQ
- replace SPI with FIQ ISR FSM
- Timer 3 source used for FIQ and Vibrator implemented in ISR (done)
- Motion sensor FSM in FIQ ISR - implemented but problematic for Motion
- Changed to private SPI bitbang in ISR (done)
- Fixed /dev/input/event interface code (done)
- Confirmed by hexdump (done)
