paparazzi-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Paparazzi-devel] Lisa debugging


From: antoine drouin
Subject: Re: [Paparazzi-devel] Lisa debugging
Date: Wed, 1 Sep 2010 20:01:22 +0200

Hi Marko

I know the I2C code needs quite a bit of rework. I2C1 is something
close from the libstm32 example - I2C2 uses different code. When I2C2
is stable I'll make I2C1 use the same code. About the missing files,
they are from libstm32, which is distributed with the paparazzi-stm32
package. You can look at the script used to build the package in the
toolchain directory on svn. What I find strange is that I don't get
the crashes you have here, but I'm using a slightly different setup
with asctec esc. Another difference is that I'm using a different
version of libstm32 where I've turned off the assert_param. In the
version of the package they are function call to an empty function (
in  sw/airborne/stm32/stm32_vector_table.c ). libstm32 make a generous
use of it, slowing down every interrupt handler by about a factor of
two.
I'll release a new version of the toolchain with the assert_param disabled asap.

Poine



On Wed, Sep 1, 2010 at 7:38 PM, Marko Thaler <address@hidden> wrote:
> Hello,
>
> Eric and Poine, thank you for your suggestions and help.
>
> I have set up the debug environment in Eclipse which seems to work fine.
>
> The only problem now is that if I randomly suspend (halt) the code
> execution, I sometimes get an error:
>
> Can't find a source file at "some_file.c"
>
> Until now some_file.c were:
> - stm32_usart.c
> - stm32_i2c.c
>
> I can not find this files anywhere in the paparazzi3 repository. The
> interesting part is that during compilation there are no errors and those
> files are not needed!?
>
> I have not managed to make a proper debug session with regards to Lisa/MKK
> ESCs freeze. However, in a quick test debug I have observed two different
> error behaviours:
>
> a) MKK ESCs block the motors (red ESC LED goes on) while Lisa is running
> normally (I am testing motors without props). To regain a normal motor
> operation I only have to power down and power up the Lisa. During this
> period ESCs have an uninterrupted power supply and the red ESC LEDs turn off
> as soon as the Lisa is reset (powered back up).
>
> b) Lisa can freeze in flight (Lisa's LEDs stop blinking and stay in their
> last state) and ESCs turn the motors off (red ESC LEDs go on). In this case
> I can regain normal operation by suspending (halting) Lisa in the debug
> session and resuming the session. As soon as I hit the resume button in
> Eclipse motors and Lisa start to operate normally.
>
> For the b) example I could not take a look at where in the program the error
> occurred since I got "Can't find a source file" message from above and
> <symbol is not available> error in the debug window.
>
> I'll have a go at a proper debug tomorrow morning.
>
> Regards,
>
> Marko
>
>
> On Tue, Aug 31, 2010 at 8:19 PM, antoine drouin <address@hidden> wrote:
>>
>> looking at the I2C issue
>>
>> for JTAG:
>> xterm -e "/opt/paparazzi/stm32/bin/openocd -f interface/lisa-l.cfg -f
>> board/lisa-l.cfg"&
>> xterm -e "telnet localhost 4444" &
>>
>> in that window : halt
>>
>> /opt/paparazzi/stm32/bin/arm-none-eabi-gdbtui --eval-command="target
>> remote localhost:3333" var/BOOZ2_A7/ap/ap.elf
>>
>> Regards
>>
>> Poine
>>
>>
>> On Tue, Aug 31, 2010 at 6:40 PM, Marko Thaler <address@hidden>
>> wrote:
>> > Hello everybody,
>> >
>> > I am testing the new Lisa on a quadrotor and have a few questions:
>> > - I use Lisa with MKK ESCs. Has anybody encountered a freeze of ESCs
>> > while
>> > in flight? The red LEDs on all the MKK ESCs turn on and all the motors
>> > shut
>> > down. This happens while Lisa is running normally. I have tested all the
>> > I2C
>> > cabling but haven't observed anything unusual. Sometimes red LEDs turn
>> > on at
>> > power up, but after a reset everything is normal and functioning
>> > properly.
>> > - I am new to gdb (been previously using CrossWorks from Rowley). Is
>> > there a
>> > quick and dirty procedure write up on how to connect to Lisa? I had a
>> > couple
>> > of tries to attach gdb (gdbserver) to a running Lisa but with no luck.
>> > - When I do not have the MKK ESCs connected to the I2C port and the
>> > autopilot code is running, I can freeze the code with holding the
>> > throttle
>> > stick in the motor on/off position (down and right) for a couple of
>> > seconds
>> > (anything from 3 to 20 seconds). This condition is not always
>> > reproducible
>> > so sometimes it needs a couple of retries (go to center position and
>> > than
>> > back to down and right). I could practically always reproduce the freeze
>> > in
>> > the first three tries when ap_mode is in ATTITUDE_Z_HOLD. When in
>> > ATTITUDE_RC_CLIMB and ATTITUDE_DIRECT the freeze is rarely reproducible.
>> > The
>> > freeze never happened when ESCs were connected to Lisa.
>> >
>> > I know you should not abuse Lisa in this way, I am just writing up my
>> > observations if anybody encounters something similar.
>> >
>> > I will examine all the unusual behaviour in the code myself as soon as I
>> > get
>> > more comfortable with gdb (any help is deeply appreciated :).
>> >
>> > Thanks for your input and kind regards,
>> >
>> > Marko
>> >
>> >
>> > _______________________________________________
>> > Paparazzi-devel mailing list
>> > address@hidden
>> > http://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>> >
>> >
>>
>> _______________________________________________
>> Paparazzi-devel mailing list
>> address@hidden
>> http://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>
>
> _______________________________________________
> Paparazzi-devel mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>
>



reply via email to

[Prev in Thread] Current Thread [Next in Thread]