BadAccess, BadShmSeg

[duart38]

Please describe the issue in as much detail as possible, including any errors and traces

I get this when i run the test command

X Error: BadAccess (attempt to access private resource denied) 10
Extension: 130 (MIT-SHM)
Minor opcode: 1 (X_ShmAttach)
Resource id: 0x2e00003
X Error: BadShmSeg (invalid shared segment parameter) 128
Extension: 130 (MIT-SHM)
Minor opcode: 5 (X_ShmCreatePixmap)
Resource id: 0x2e0000d

---

Linux distribution (if applicable): Ubuntu 18.04.1
Howdy version:

[boltgolt]

Could you try running it as sudo QT_X11_NO_MITSHM=1 howdy test?

To apply this fix permenatly add QT_X11_NO_MITSHM=1 to the bottom of /etc/environment.

[keckes]

Hi there, thanks for putting this together!

I had the same issue upon trying to run sudo howdy test and the viewer window was blank.

I tried running sudo QT_X11_NO_MITSHM=1 howdy test and it worked, though the image it returned was basically blank (see attached) and not dynamic (the image didn't really change even as I moved my face closer to the camera/emitters or waved my hand in front of them).

When I try using howdy normally it just times out and I have to enter the password. I assume it's because the image is not actually dynamic (i.e. not even recognizing something that resembles a face).

I'm running Ubuntu 18.04 LTS on a Thinkpad T480s.

[boltgolt]

This is a known decoding error in some cameras, could you try running sudo howdy config and setting force_mjpeg to true?

[keckes]

OK, did that and the test procedure gives an actual image -- great!

I just added two versions of my face (with and without glasses) to my profile then tried a few tests. Unfortunately I still get the timeout problem, so I still have to enter my password manually.

[boltgolt]

Try upping the certainty value, howdy might be recognizing you with too little match certainty.

[keckes]

I upped the certainty value to 9 and lowered the dark_threshold to 1 but still got the timeout error.

I noticed that when I run the test procedure, the frame rate and exposure consistency seems to depend heavily on the distance between my face and the emitters/camera. Still, even when I added a new face based on the distance and position I felt was most consistent in the test procedure, I still get the timeout errors.

Any other suggestions?

[boltgolt]

Following up on issues i forgot to reply to, are you still having this problem or did you fix it/uninstall Howdy?

[keckes]

Hi, thanks for checking in. I disabled (but didn't uninstall) Howdy because I could never get it to recognize me, no matter how much I tweaked the parameters. I had written earlier about how the output was just greenish static when I used the "sudo howdy test" command, so perhaps the authentication routine is getting the same data every time it turns on the camera and of course it in no way resembles a face.

…

On Thu, Oct 4, 2018 at 9:48 AM boltgolt ***@***.***> wrote: Following up on issues i forgot to reply to, are you still having this problem or did you fix it/uninstall Howdy? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#60 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AoT2S8yL8kX54v949QXRJge38pwuLLAZks5uhb1UgaJpZM4VvIJN> .

[boltgolt]

Greenish static is a new symptom for me, does this static also show up when you open Cheese (search for Cheese in the dash)? You might need to try out different webcams/video sizes in the Preferences menu.

[keckes]

Indeed it does show up initially in Cheese because the default camera is the one associated with the IR emitters. I always have to select the normal webcam in the Preferences menu to get a clear picture.

…

On Thu, Oct 4, 2018 at 8:41 PM boltgolt ***@***.***> wrote: Greenish static is a new symptom for me, does this static also show up when you open Cheese (search for Cheese in the dash)? You might need to try out different webcams/video sizes in the Preferences menu. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#60 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AoT2S8CoWbMBlMSLNypNGI91V6A6ly_6ks5uhlZQgaJpZM4VvIJN> .

[boltgolt]

So Cheese is also just getting the garbage data? Does switching resolutions in the preferences change anything?

[keckes]

Yep, the cheese default is the IR camera at 480x400 and it's the green garbage. Upon switching resolutions to 400x400 it looks as expected, however.

…

On Fri, Oct 5, 2018 at 11:53 AM boltgolt ***@***.***> wrote: So Cheese is also just getting the garbage data? Does switching resolutions in the preferences change anything? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#60 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AoT2S_tNxTGCH7SYr_2Y2S9wsiN37Toeks5uhywfgaJpZM4VvIJN> .

[boltgolt]

Great! Could you set the frame size in sudo howdy config to that format:

frame_width = 400
frame_height = 400

This should fix the problem.

[nilmerg]

Had the same problem, running it with sudo QT_X11_NO_MITSHM=1 howdy test finally worked.

[mbprado]

I had exactly the same issue.
I was able so see my image in test, but it never had been recognized. Also in test, using IR camera, it never shown the red circle around my face (testing with the normal camera, everything worked fine). I completely solved the issue with the steps bellow:

1 - To fix green image:
In howdy config, simply set ( no need to set force_mjpeg or anything else):
frame_width = 400
frame_height = 400

2 - To fix testing interface:
add the line to /etc/environment
QT_X11_NO_MITSHM=1

3 - Finally to get the correct detection, including the red circle in howdy test, I had to set bright and contrast the parameters in v4l:

v4l2-ctl -d /dev/video0 --set-ctrl=contrast=0
and:
v4l2-ctl -d /dev/video0 --set-ctrl=brightness=255
check with:

$ v4l2-ctl -d /dev/video0 --list-ctrls
brightness 0x00980900 (int) : min=0 max=255 step=1 default=128 value=255
contrast 0x00980901 (int) : min=0 max=255 step=1 default=32 value=0

Doing this, I had no more problem with detection using IR camera.