Keith Packard [Tue, 17 Feb 2015 06:22:37 +0000 (22:22 -0800)]
altosdroid: Ignore automatic tab changing while activity is saved
When the activity state is saved (after onSaveInstanceState()), we
can't update the UI until the activity is restarted or restored; that
means any UI changes we make, like switching tabs, must deal with this
by allowing those changes to be ignored, using commitAllowingStateLoss
instead of commit.
Keith Packard [Tue, 17 Feb 2015 05:19:09 +0000 (21:19 -0800)]
altosdroid: Explicitly disconnect BT on termination or 'disconnect'
This adds an explicit message to the telemetry service telling it when
to stop trying to talk to the bluetooth device. Until this message is
received, the service will reconnect to the specified BT device.
That message is sent when you 'quit' the application, or when you 'disconnect'.
Keith Packard [Sun, 15 Feb 2015 07:18:38 +0000 (23:18 -0800)]
altos/cc1200: Adjust bit-sync configuration
The default bit timing adjustment mechanism allows for only a 0.2%
deviation from the programmed bit timing. I found one TeleMini device
which is beyond that tolerance as it was built with an older crystal
with more error.
Switch to the more expensive synchronization mechanism which allows up
to 2% timing error, but requires a multi-byte preamble (which we
have). This fixes packet mode nicely.
Keith Packard [Sat, 14 Feb 2015 09:13:21 +0000 (01:13 -0800)]
altosuilib: Detect pair programming by product name, not USB id
With TeleDongle, TeleBT and TeleMetrum coming in both pair- and self-
programmable versions, we can't use the USB id to tell them
apart. Instead, fetch the device name and use that instead.
Keith Packard [Sat, 14 Feb 2015 09:11:30 +0000 (01:11 -0800)]
altosui: Run all igniter status requests from non-GUI thread
Anything run from the UI thread blocks the UI entirely; the Fire
Igniters startup code to collect the number of pyro channels when
building the UI was doing that from the UI thread. Switch that around
so that the UI doesn't get built until that reply comes back, allowing
the user to see the 'connecting' dialog, and cancel it if required.
Keith Packard [Sat, 14 Feb 2015 07:51:10 +0000 (23:51 -0800)]
altos: Replace ao_alarm/ao_clear_alarm with ao_sleep_for
Having arbitrary alarms firing in the middle of complicated device
logic makes no sense at all. Therefore only correct use of ao_alarm
and ao_clear_alarm was around a specific ao_sleep call, with correct
recovery in case the alarm fires.
This patch replaces all uses of ao_alarm/ao_sleep/ao_clear_alarm with
ao_sleep_for, a new function which takes the alarm timeout directly.
A few cases which weren't simply calling ao_sleep have been reworked
to pass the timeout value down to the place where sleep *is* being
called, and having that code deal with the return correctly.
Keith Packard [Sun, 8 Feb 2015 04:36:10 +0000 (20:36 -0800)]
altosuilib: Don't offer to graph some GPS details that TM doesn't log
When using a TM eeprom file, various minor GPS details are logged
(course, ground speed, climb rate, etc). Make sure these aren't
offered up for graphing.
Keith Packard [Sun, 8 Feb 2015 04:22:19 +0000 (20:22 -0800)]
altosuilib: Get the Eeprom download progress bar working again
The eeprom download code wasn't computing the start of each state
transition correctly, so the progress bar was snapping to the end of
the chunk for each state.
Invalid state values would snap the bar to the right side.
Landed state wasn't ever seen, so the bar would not ever fill.
Keith Packard [Sun, 8 Feb 2015 01:09:40 +0000 (17:09 -0800)]
altosuilib: Reset telemetry state after processing it
This avoids re-using stale state after switching telemetry reception
parameters around, which otherwise generates spurious entries for the
wrong frequencies in the scan results list.
Keith Packard [Sun, 8 Feb 2015 01:08:03 +0000 (17:08 -0800)]
altoslib: Drop telemetry packets processed while monitoring is disabled
A lag between the thread queuing telemetry packets and one pulling
them out can result in stale telemetry data being returned to the
reader. Fix this by dropping telemetry read while monitoring is disabled.
Keith Packard [Sat, 7 Feb 2015 22:40:17 +0000 (14:40 -0800)]
altosui: Remove the dregs of AltosDroid load-old-telem code
AltosDroid used to scan the old .telem file to return to the previous
flight state on restart. Now it just loads the old state object
instead, a vast improvement in performance.
To do that, there were some changes in the altoslib code
required. This patch just removes those, fixing replay bugs in TeleGPS
along the way.
Keith Packard [Sat, 7 Feb 2015 21:09:18 +0000 (13:09 -0800)]
altos: ao_packet_getchar needs to return 'int' so AO_READ_AGAIN works
With ao_packet_getchar returning 'char', AO_READ_AGAIN would get
converted to 0xff instead of -1 on return, making ao_packet_echo
either produce a bogus character on output, breaking command
responses, or wedge and never exit when the link was dropped.
Keith Packard [Sat, 7 Feb 2015 09:05:57 +0000 (01:05 -0800)]
altos/lpc: Give up on interrupt-driven SPI driver
There are just too many limitations in when interrupts are delivered
to make them useful. Instead, just drive the SPI directly with the
CPU. At higher spi speeds (6Mhz or more), it's probably faster to do
it this way anyways.
Keith Packard [Sat, 7 Feb 2015 04:40:05 +0000 (20:40 -0800)]
altosuilib: Don't show 'Sats in view' for TeleGPS eeprom graphing
We don't log the visible sats from TeleGPS, only the basic GPS
info. Have AltosFlightStats track whether sat info is present, then
use that to elide the Sats In View graph entry as needed.
Keith Packard [Sat, 7 Feb 2015 04:36:44 +0000 (20:36 -0800)]
altosuilib: EEProm download from TeleGPS doesn't have states
As TeleGPS doesn't have flight states, the EEProm download progress
bar shouldn't show 'state invalid', and it should use a different
scale for each block.
Fix this by having AltosConfigData know which devices have state based
on their log format, then mark the progress bar state limits as
'invalid/invalid' instead of 'boost/landed'. Then have the progress
bar use that to set a more reasonable scale for each block.
Keith Packard [Fri, 6 Feb 2015 12:45:17 +0000 (04:45 -0800)]
altoslib: Store MS5607 data in AltosConfigData for use by AltosMs5607
When doing 'Monitor Idle', we fetch new config data each iteration and
pass that to each of the readers, including ms5607. Instead of
re-fetching the config data there, just store the ms5607 parameters
when we fetch it the first time and copy it over.
Keith Packard [Fri, 6 Feb 2015 11:32:27 +0000 (03:32 -0800)]
micropeak: Support both FTDI and Altus Metrum USB IDs
The windows driver only returns FTDI devices when
altos_ftdi_list_start is invoked, so we need to call both that and the
regular altos_list_start to get all of the devices.
Keith Packard [Fri, 6 Feb 2015 09:29:56 +0000 (01:29 -0800)]
libaltos: Use more SetupDi API to get Windows 7 listing devices
My Windows 7 box doesn't include 'SymbolicName' in the registry, but
SetupDiGetDeviceInstanceId *does* return something that includes
vid/pid/serial, so use that in preference.
Keith Packard [Fri, 6 Feb 2015 00:08:44 +0000 (16:08 -0800)]
altos/lpc: Clean up USB endpoint access functions
The USB device endpoints can have two pointers in them, but we use
only the first. Fix the access functions to take an index as to which
we want so that we can disable the other address registers with them
instead of requiring open-coded access.
Keith Packard [Thu, 5 Feb 2015 04:30:15 +0000 (20:30 -0800)]
altos/telebt-v3.0: Set initial radio cal to a sensible value
CC1200 uses a different xtal than CC1120, so the default calibration
needs to change. This value hit 434.550 on SN 2407, so it seems like a
reasonable place to start.
Keith Packard [Sun, 1 Feb 2015 15:44:52 +0000 (16:44 +0100)]
altos/stmf0: Support PA11/PA12 remapping
Small pin-count versions of the STMF0 can remap PA11/PA12 on the same
pins as PA9/PA10. These are used by USB, so have the USB driver deal
with remapping them.
Keith Packard [Sun, 1 Feb 2015 15:36:51 +0000 (16:36 +0100)]
altos/stm: Add ability to delay STDIO usage for serial ports
Bluetooth needs to delay adding the serial port to stdio until the
link is up and running. The cc1111 serial driver had
DELAY_SERIAL_*_STDIN bits which have been added to the STM serial driver.
Keith Packard [Tue, 27 Jan 2015 06:16:18 +0000 (22:16 -0800)]
altos/usbtrng-v2.0: Add random output command
Read chunks of random ADC data, do some CRC16 computations to de-bias
and merge 4 bytes of input data into 2 bytes of output data, then dump
that out over USB.
Keith Packard [Tue, 27 Jan 2015 06:12:40 +0000 (22:12 -0800)]
altos/stmf0: Add faster USB path
This adds a way to allocate private USB buffers for sending data
without needing to copy it again. It requires ensuring that all
accesses are 16 bits aligned to 16 bit boundaries.
Keith Packard [Fri, 23 Jan 2015 05:31:45 +0000 (21:31 -0800)]
altos/lpc: Fix double-buffered USB changes
These got merged when we were down in Auckland, but before they'd been
finished. Transmitting worked fine, but receiving was mis-configuring
the OUT buffer size in the hardware.
Keith Packard [Tue, 23 Dec 2014 01:11:51 +0000 (17:11 -0800)]
altos: Adjust CC1200 RSSI reporting value
On the test setup, we were about 13dB off, so tweak the settings to
suit. This result wasn't tested; it's just a rough offset until we get
a production run of TeleDongle boards back to do more systematic
measurements.
Keith Packard [Tue, 23 Dec 2014 01:10:33 +0000 (17:10 -0800)]
altos: Document cc115l power measured at RF setting 0x03 as -31.75dBm
We built a custom TeleGPS load for radio sensitivity measurements with
the lowest documented power level setting (0x03), and measured the
power out at that value of -31.75dBm on the test board.
Keith Packard [Mon, 24 Nov 2014 02:56:40 +0000 (18:56 -0800)]
altos/cc1200: With PQT wide open, we can't use PQT_REACHED for start
Because we're allowing even signals only weakly correlated with the
preamble through to sync detection, we can't use the PQT_REACHED
symbol to tell when a packet header has been seen. Instead, just look
for SYNC_FOUND.
Keith Packard [Mon, 24 Nov 2014 02:31:15 +0000 (18:31 -0800)]
altos/cc1200: Wait for packet to be placed in FIFO during receive
This changes the receive code to use MCU_STATUS, waiting for
MARC_STATUS1 to indicate that the packet is in the fifo before reading
it out.
It also fixes the receive timeout code to keep receiving if the
preamble or sync have been seen when the timeout fires. This makes
TeleLCO able to use short timeouts during scanning while still
successfully receiving packets.