Debian ARM board: 'INZI' is not natively supported ... please upgrade to kernel 4.12 or later.

[KeCh96]

---

Required Info
Camera Model	SR300
Firmware Version
Operating System & Version	Debian
Kernel Version (Linux Only)	4.11.0
Platform	ARM board
SDK Version	SDK2

I want to get depth frames and process by ARM board with Debian system. I have followed this guide and build pyrealsense2.so in my ARM board. Then I run the following code in python:

import pyrealsense2 as rs
pipeline = rs.pipeline()
pipeline.start()

I run into an error:

Traceback (most recent call last):
File "test_realsense.py", line 5, in
pipeline.start()
RuntimeError: The requested pixel format 'INZI' is not natively supported by the Linux kernel and likely requires a patch!
Alternatively please upgrade to kernel 4.12 or later. Last Error: Inappropriate ioctl for device

So I tried to update my linux kernel. I use aptitude search linux-image | awk '{print $2}' to see all support kernels. I get the following result:

linux-image-4.11.0-qcomlt
linux-image-4.11.0-qcomlt-arm:armhf
linux-image-4.11.0-qcomlt-arm64
linux-image-4.11.0-qcomlt-arm64-dbg
linux-image-4.11.0-qcomlt-dbg
linux-image-4.14.0-qcomlt-arm:armhf
linux-image-4.14.0-qcomlt-arm64
linux-image-4.14.0-qcomlt-arm64-dbg
linux-image-4.9.0-qcomlt
linux-image-4.9.0-qcomlt-dbg

I download 4.14 kernel and use apt-get purge linux-image-4.11.0-qcomlt to remove the 4.11 kernel. Then I type dpkg --get-selections |grep linux-image and only see 4.14. But I type uname -a and find the kernel is still 4.11
I try the python code pipeline.start() in this environment, and get "No device" error.

I also tried 4.9.0, and the result is exactly the same with 4.14.

I wonder if the camera will work in 4.10.0 or 4.4.0 kernel, but as the above information shows, my ARM board does not seem to support 4.10.0 or 4.4.0 kernel. Thus this solution seems not applicable for me.

My main goal to get depth frame, (and not some of more advanced controls / metadata support), I notice this guide, but I still confused what I should do in my condition. So I open this new issue. Could you please help me? @dorodnic @zivsha

[dorodnic]

Hi @KeCh96
The same answer from #1186 still holds: you can either manually patch the kernel or run cmake with -DFORCE_LIBUVC=true.

[KeCh96]

Thank you!! I run cmake ../ -DFORCE_LIBUVC=true -DBUILD_PYTHON_BINDINGS=true. Now I get depth frame successfully! @dorodnic

After several seconds, I get 391 frames, but an Segmentation fault error occurs:

372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
Segmentation fault

I run my code again and get Segmentation fault immediately (I get no frame this time).
Then I plug out SR300 and plug it in, then run my code. After about 300 frames, the Segmentation fault error occurs.

[KeCh96]

According to this guide, this error may caused by multiple instances of the same device on Linux. But I only use one SR300. @sgorsten

According to this issue, we can use the following code to avoid this bug:

            if (
                    !ir_frame || ir_frame.get_data() == NULL ||
                    !color_frame || color_frame.get_data() == NULL ||
                    !depth_frame || depth_frame.get_data() == NULL) {

                --i;
                continue;
            }

So I write my code as following:

pipeline = rs.pipeline()
pipeline.start()

i=0
while True:
    try:
        frames = pipeline.wait_for_frames()
        depth = frames.get_depth_frame()
        depth_data = depth.as_frame().get_data()
        i=i+1
        print i
    except:
        continue

But the Segmentation fault error still occurs.

Could someone provide some guidance? Thanks in advance.

[KeCh96]

I found the following annotation in /third_party/sqlite/sqlite3.h. This information may reveal the cause of this bug. But I still don't know how to solve my problem.

**
** ^Calling sqlite3_free() with a pointer previously returned
** by sqlite3_malloc() or sqlite3_realloc() releases that memory so
** that it might be reused.  ^The sqlite3_free() routine is
** a no-op if is called with a NULL pointer.  Passing a NULL pointer
** to sqlite3_free() is harmless.  After being freed, memory
** should neither be read nor written.  Even reading previously freed
** memory might result in a segmentation fault or other severe error.
** Memory corruption, a segmentation fault, or other severe error
** might result if sqlite3_free() is called with a non-NULL pointer that
** was not obtained from sqlite3_malloc() or sqlite3_realloc().
**

[KeCh96]

I open a new issue on this problem. To analyze the reason of the error, I show my 'cmake' and 'make' output. Could someone provide some help?

[RealSense-Customer-Engineering]

[Realsense Customer Engineering Team Comment]
Hi @KeCh96,

Since you resolve this issue, still need any support for this topic?

[RealSense-Customer-Engineering]

[Realsense Customer Engineering Team Comment]
If no other further c