mjb/altos/src/stm, branch altosdroid

altos: Increase default STM stack to 648 bytes

2012-06-28T06:09:16+00:00

512 seems cozy given the printf implementation we're using and the extensive interrupts. Signed-off-by: Keith Packard

altos: Clean up STM I2C clock computations

2012-06-28T00:18:57+00:00

Fix both clock time and rise time. Signed-off-by: Keith Packard

altos: get MPU6000 I2C link working reliably

2012-06-27T21:38:35+00:00

This slows the i2c bus to 100kHz (yuck), sets the rise time to spec (it was way off) and adds more delays during bus setup. I've run this for hours now without trouble. Will try to adjust things back to fast mode and see if I can make that work as 100kHz isn't fast enough to reliably get data at 100 samples/sec. Signed-off-by: Keith Packard

altos: reorder stm USB state stores to avoid races

2012-06-27T21:35:56+00:00

Must set ao_usb_in_pending before telling USB about new data or an interrupt could arrive at the wrong time to clear it. Same for ao_usb_in_flushed. Without these changes, I've seen the USB bus lock up on occasion, waiting for an IN packet to consume data, but with no IN data pending in the hardware. Signed-off-by: Keith Packard

altos: include targe SPI speed in get request

2012-06-27T21:34:53+00:00

STM needs it to be provided when enabling the SPI device, so just fix AVR and cc1111 to do the same. Signed-off-by: Keith Packard

altos: Wait for i2c START condition before setting interrupt bits

2012-06-27T09:46:27+00:00

This seems better than the random loop that it replaces, but I still have no idea why this is required; it doesn't coorespond to the docs at all... Signed-off-by: Keith Packard

altos: stm i2c debug code was calling flush() even when disabled

2012-06-27T08:19:05+00:00

The fancy stm i2c debugging code had calls to flush() that were invoked outside of the conditionals leading to all kinds of fun -- flush() may re-enable interrupts, yield or do all kinds of wacky stuff, none of which is appropriate from the middle of a device driver Signed-off-by: Keith Packard

altos: mpu6000 requires a delay during start

2012-06-27T08:17:51+00:00

I have no idea why this is required, but the mpu6000 will not come up and run if this isn't present. Signed-off-by: Keith Packard

altos: Create a 32-bit 1MHz timer for use in profiling execution

2012-06-27T06:18:44+00:00

This provides a simple method for getting high-resolution timer data to use in performance tuning code. It's not used by default anywhere. Signed-off-by: Keith Packard

altos: Optimize FEC encode and decode

2012-06-27T06:11:10+00:00

Integrate interleaving, CRC and padding within the decode/encode functions. Provide for ISR priorities so that the 1120 RX interrupt takes precedence over the other interrupts or we risk losing bits. Optimize the viterbi decoder a bit (goes from 10ms per packet to 7ms per packet). Signed-off-by: Keith Packard