Re: [Paparazzi-devel] Lisa/L not working - bug on last commit ?

[Felix Ruess]

Hi,

> Lisa/L + Aspirin 1.5 + ppm radio + ublox GPS + MKK actuators , on a rotorcraft

Can you please post your airframe file, e.g. as a gist or on pastebin.com

> The v3.9 branch was working last week. I installed maybe two week ago. In
> fact, I did a "git pull" for installation, and realized after that I got the
> v3.9. I do a "git pull" last week, before to notice that I have a different
> branch. So the problem is the moment I get the v3.9 and last week. 

Can you please provide the exact versions of v3.9 it worked with and which one it doesn't work with?

Did you check that you are using the unmodified version from the Paparazzi repo, so now changes you made yourself?

Easiest to check is with "gitk --all" which will show you the branches graphically... and best post the SHA1 of the commit you tested it on.

Also make sure to run "make" after pulling.

> By the way, it's working in "master" branch, so it's something different between
> the two branch.

in master nearly all the low level drivers are ported to libopencm3 and also the new I2C driver is used.

In v3.9 the old I2C driver is still the default. If you want to test it with the new driver, add 

<configure name="USE_NEW_I2C_DRIVER" value="1"/>

to your firmware section.

>> Why do you think there was a bug introduced here? Sounds rather
>> unlikely, since you can see on
>> https://github.com/paparazzi/paparazzi/commits/v3.9 there were no
>> significant changes regarding hw.
>
> I know. But maybe AHRS or guidance change introduce an endless loop ? Or an
> exception fault ? Even without hw link, it can be a bad hw call or tasks
> conflict.

Possible, but I very much doubt it... it was just some renames/cleanup and some sim fixes...

More likely that it is the same old I2C issue that happens now and then, with the old driver it sometimes gets stuck...

Cheers, Felix