GCS_MAVLink: Fix for the Gopro/Solo Gimbal issue

[mtbsteve]

@rmackay9 @davidbuzz @proficnc This PR replaces the PR #22692 which I accidentally closed last week.
It should work for CubeSolo, CubeGreen and fmuv3 boards when HAL_SOLO_GIMBAL_ENABLED is met and when a Gopro heartbeat is broadcasted. I tested it successfully with AC 4.4beta on a Solo with a CubeSolo board and a non-Solo copter with CubeBlack installed.

[rmackay9]

nitpick: we could change, "gopro_status_check == false" to "!gopro_status_check"

[tridge]

@mtbsteve we'd like a description of what the issue is so we can test the fix. We will test a solo here if you can describe what to look for

[rmackay9]

@peterbarker and @tridge decided to have a look at it over the weekend.

More generally we think it would be good to have a "one way private" option in SERIALx_OPTIONS. In this case we think the camera/gimbal needs to send its traffic to the GCS but all the other mavlink traffic cannot be sent to the camera/gimbal without it becoming unhappy. This would be scope creep and is not a blocker on this PR.

The blocking issue with this PR is the global variable.

[mtbsteve]

@mtbsteve we'd like a description of what the issue is so we can test the fix. We will test a solo here if you can describe what to look for

The issue is that all GOPRO mavlink messages (GOPRO_HEARTBEAT, GOPRO_GET_RESPONSE, GOPRP_SET_RESPONSE,..) are no longer transmitted between GCS, Solo Gimbal, and Controller since Arducopter 4.1.5.

[mtbsteve]

@peterbarker and @tridge decided to have a look at it over the weekend.

More generally we think it would be good to have a "one way private" option in SERIALx_OPTIONS. In this case we think the camera/gimbal needs to send its traffic to the GCS but all the other mavlink traffic cannot be sent to the camera/gimbal without it becoming unhappy. This would be scope creep and is not a blocker on this PR.

The blocking issue with this PR is the global variable.

Thanks for having a look at the issue.
I am not sure if a one-way communication would fix it since the communication of the GOPRO mavlink messages is bi-directional - between Solo-Gimbal and GCS as well as between Solo-Gimbal and the Controller, and between the Controller and the GCS.
I tested my fix extensively with AC 4.3, and also with the 4.4 dev version w/o problems so far.

[AndKe]

@rmackay9 - what's the verdict on this? I've been waiting for months on #23676
the process is too slow, can you please bring this up on a devcall?

[IamPete1]

Should be in the header file.

[rmackay9]

Let's correct the commit name to start with "GCS_MAVLink:"

[mtbsteve]

@rmackay9

Let's correct the commit name to start with "GCS_MAVLink:"

Done :-) Thanks a lot for merging!

[rmackay9]

Hi @mtbsteve,

We discussed this on the dev call and agreed we'd merge after these small changes are made:

1. the new variable in the .cpp is moved to the header file
2. commit message (not this PR description) is prefixed with "GCS_MAVLink"

I think we agreed that Tridge, PeterB or I will take care of making these small changes but you're welcome to as well. In any case, I've re-tagged for next week's dev call so it'll come up again if one of us doesn't get to it.

[mtbsteve]

Hi @mtbsteve,

We discussed this on the dev call and agreed we'd merge after these small changes are made:

1. the new variable in the .cpp is moved to the header file
2. commit message (not this PR description) is prefixed with "GCS_MAVLink"

I think we agreed that Tridge, PeterB or I will take care of making these small changes but you're welcome to as well. In any case, I've re-tagged for next week's dev call so it'll come up again if one of us doesn't get to it.

Hi @rmackay9
yes I would appreciate if one of you folks could take care of those changes since I am currently away from home.
Thanks for all your work!

[rmackay9]

We haven't gotten to this yet but we're leaving the dev call topic on it so that we don't forget next week!

[peterbarker]

@mtbsteve we discussed at DevCall and made the suggested changes; correcting commit message, variable scope, generally NFC unless we're compiling with the thing built in. Please test?

fixed