Skip to content

Conversation

@ev-mp
Copy link
Contributor

@ev-mp ev-mp commented Apr 3, 2019

  • Promote T265 Firmware tag to 0.0.18.5448
  • Update log level for T265-specific cases
  • SDK's Cmake private files list - remove duplicates, add missing headers
  • GCC [wpedantic] warning - Remove unnecessary semicolon

"${CMAKE_CURRENT_LIST_DIR}/types.h"
"${CMAKE_CURRENT_LIST_DIR}/backend.h"
"${CMAKE_CURRENT_LIST_DIR}/device.h"
"${CMAKE_CURRENT_LIST_DIR}/api.h"
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@ev-mp because you sorted this list and changed its contents at the same time, it's hard to see what changed here. I see at least ivcam/ivcam-private.h is now gone, which is unrelated to T265 as far as I know.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

These changes rectify inconsistencies discovered during the validation of the new TM2 FW. And they are separate from the FW update commit

_color_device_idx = add_sensor(create_color_device(ctx, color_devs_info));
}
};
}
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This seems unrelated to T265, so probably shouldn't be in this PR

else
{
LOG_WARNING("Dropped frame. alloc_frame(...) returned nullptr");
LOG_INFO("Dropped frame. alloc_frame(...) returned nullptr");
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

In what scenario would alloc_frame return a nullptr that doesn't justify a warning?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The warning occurs whenever frames acquired by users are not released in a timely manner.
We observe it some OpenGl-enabled demos where the default frames queue allocated in the SDK (size=16) is probably not enough to compensate for 3D scene render.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The main problem is that this is a negative feedback loop - frame drops cause console output, slowing the app and causing more frame drops.
A better solution that we discussed for a while (I think @aangerma tried implementing at some point) is for processing blocks to implement notification extension (like the sensor does) and manage some basic internal statistics. This way, we can avoid spamming the output with such warnings...

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@ev-mp Should we make the frame queue deeper for imu data which is small but high rate?

@ev-mp ev-mp changed the title Promote T265 Firmware selection to 0.0.18.5441 Promote T265 FW Tag; Minor SDK fixes Apr 3, 2019
Evgeni Raikhel added 2 commits April 4, 2019 12:54
Change-Id: I44cf76237e005f175523274717570e22a39eaf43
(cherry picked from commit d72b3ab0b9e34bb997ba5c775d77c1e1f4e25487)
…ing header

Reduce log report level in TM2-specific scenarios.
Handle remove trailing whitespaces and GCC -wPedantic warning
Fix Librelasense cmake private files list

Change-Id: If3c1c873aefec7c5a876892dac399f75199e6499
(cherry picked from commit 7df5e5bcc7ef6650df03a705398375d26ebc5a87)
Signed-off-by: Evgeni Raikhel <[email protected]>
@ev-mp ev-mp merged commit 8791358 into realsenseai:development Apr 4, 2019
@ev-mp ev-mp deleted the tm2_5441 branch April 18, 2019 08:41
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants