oe4t / meta-tegra Goto Github PK

Hello madisongh,

first of all thank you for this BSP. We have the following problem:

We basically followed issue #12 and were
successful with the core-image-sato, but with core-image-minimal the boot process failed.

The flashing looks good but when U-Boot scans mmc 0:1, it doesn't recognize anything to boot from. Indeed ls mmc 0:1 boot reveals, that /boot is empty.

We tried this with branch morty-l4t-r27.1 as well as with commit 1436d4c from master.

Do you have any clues for us what to try next?

Thanks in advance,

selbstereg

ERROR: cuda-toolkit-7.0-76-r0 do_package_qa: QA Issue: -dev package contains non-symlink .so: cuda-toolkit-dev path 'work/aarch64-poky-linux/cuda-toolkit/7.0-76-r0/packages-split/cuda-toolkit-dev/usr/local/cuda-7.0/lib/stubs/libnppc.so'
-dev package contains non-symlink .so: cuda-toolkit-dev path 'work/aarch64-poky-linux/cuda-toolkit/7.0-76-r0/packages-split/cuda-toolkit-dev/usr/local/cuda-7.0/lib/stubs/libcufft.so'
-dev package contains non-symlink .so: cuda-toolkit-dev path 'work/aarch64-poky-linux/cuda-toolkit/7.0-76-r0/packages-split/cuda-toolkit-dev/usr/local/cuda-7.0/lib/stubs/libcufftw.so'
-dev package contains non-symlink .so: cuda-toolkit-dev path 'work/aarch64-poky-linux/cuda-toolkit/7.0-76-r0/packages-split/cuda-toolkit-dev/usr/local/cuda-7.0/lib/stubs/libcusparse.so'
-dev package contains non-symlink .so: cuda-toolkit-dev path 'work/aarch64-poky-linux/cuda-toolkit/7.0-76-r0/packages-split/cuda-toolkit-dev/usr/local/cuda-7.0/lib/stubs/libnpps.so'
-dev package contains non-symlink .so: cuda-toolkit-dev path 'work/aarch64-poky-linux/cuda-toolkit/7.0-76-r0/packages-split/cuda-toolkit-dev/usr/local/cuda-7.0/lib/stubs/libcusolver.so'
-dev package contains non-symlink .so: cuda-toolkit-dev path 'work/aarch64-poky-linux/cuda-toolkit/7.0-76-r0/packages-split/cuda-toolkit-dev/usr/local/cuda-7.0/lib/stubs/libcublas.so'
-dev package contains non-symlink .so: cuda-toolkit-dev path 'work/aarch64-poky-linux/cuda-toolkit/7.0-76-r0/packages-split/cuda-toolkit-dev/usr/local/cuda-7.0/lib/stubs/libcurand.so'
-dev package contains non-symlink .so: cuda-toolkit-dev path 'work/aarch64-poky-linux/cuda-toolkit/7.0-76-r0/packages-split/cuda-toolkit-dev/usr/local/cuda-7.0/lib/stubs/libnppi.so' [dev-elf]
ERROR: cuda-toolkit-7.0-76-r0 do_package_qa: QA Issue: cuda-toolkit rdepends on cuda-toolkit-dev [dev-deps]
ERROR: cuda-toolkit-7.0-76-r0 do_package_qa: QA run found fatal errors. Please consider fixing them.
ERROR: cuda-toolkit-7.0-76-r0 do_package_qa: Function failed: do_package_qa
ERROR: Logfile of failure stored in: /home/ubuntu/Build/embedded_distribution/poky/build/tmp/work/aarch64-poky-linux/cuda-toolkit/7.0-76-r0/temp/log.do_package_qa.23811
ERROR: Task 208 (/home/ubuntu/Build/embedded_distribution/poky/meta-tegra/recipes-devtools/cuda/cuda-toolkit_7.0-76.bb, do_package_qa) failed with exit code '1'

Add inherit cuda to a recipe that sets BBCLASSEXTEND = "native". Now try building the -native variant with a MACHINE that does not set "cuda" as a MACHINE_FEATURE. The _appends mess up the resulting variable definitions.

Hi,

I successfully built core-image-minimal and flashed it to the TX2. My Build Configurations:
BB_VERSION = "1.34.0"
BUILD_SYS = "x86_64-linux"
NATIVELSBSTRING = "universal-4.8"
TARGET_SYS = "aarch64-poky-linux"
MACHINE = "jetson-tx2"
DISTRO = "poky"
DISTRO_VERSION = "2.3.3"
TUNE_FEATURES = "aarch64"
TARGET_FPU = ""
meta
meta-poky
meta-yocto-bsp = "pyro:e31d17d2e1a7a011f5ba0f14f392763846bd80fd"
meta-tegra = "pyro-l4t-r28.1:c19e9bc1f5b058fbae3ebd118e5379aee296fe86"

Added the meta-tegra layer to bblayers.conf and following to local.conf:
MACHINE ?= "jetson-tx2"
IMAGE_CLASSES += "image_types_tegra"
IMAGE_FSTYPES = "tegraflash"
GCCVERSION = "5.%" # Cuda support

Now I want to add the cuda toolkit to the core-image-minimal. Therefore I created a new layer:
./poky/scripts/yocto-layer create tx2-cuda
mkdir -p ./meta-tx2-cuda/recipes-core/images
Added the following to new recipe meta-tx2-cuda/recipes-core/images/core-image-tx2-cuda.bb:
DESCRIPTION = "Minimal image with cuda installation."
LICENSE = "MIT"
require recipes-core/images/core-image-minimal.bb
IMAGE_INSTALL += "cuda-toolkit"

The exception thrown is:
ERROR: Task do_populate_sdk in /home/user/work/meta-tx2-cuda/recipes-core/images/core-image-tx2-cuda.bb rdepends upon non-existent task do_package_write_rpm in /home/user/work/meta-tegra/recipes-devtools/cuda/cuda-toolkit_8.0.84-1.bb
ERROR: Command execution failed: 1

Furthermore I was a little bit confused about the cuda-toolkit_8.0.84-1.bb "dummy recipe". Did it work in the past? I did the same procedure on the TX1 (morty branch). Everthing worked fine there. Any help is greatly appreciated.

@madisongh I'm not sure this is really an issue or not. I'm trying to get access to the cuda tools, such as cuda-gdbserver.

I've added "cuda-binaries" to my IMAGE_INSTALL packages but I get this

ERROR: Task do_populate_sdk in /opt/oe/krogoth/meta-tempmine/recipes-core/image/core-image-mimbal.bb rdepends upon non-existent task do_populate_sysroot in /opt/oe/krogoth/meta-tegra/recipes-devtools/cuda/cuda-binaries_8.0.34-1.bb
ERROR: Command execution failed: 1

I see a lot of cuda-* packages. Maybe I'm including the wrong one? what's a good set to include?

@madisongh Any thoughts on the error below.

When I add

IMAGE_CLASSES += "image_types_tegra"
IMAGE_FSTYPES = "tegraflash"

to the end of my local.conf and try to bitbake core-image-sato I get

NOTE: Resolving any missing task queue dependencies
ERROR: Nothing PROVIDES 'u-boot'
ERROR: u-boot-tegra PROVIDES u-boot but was skipped: incompatible with machine jetson-tx2 (not in COMPATIBLE_MACHINE)
ERROR: u-boot was skipped: PREFERRED_PROVIDER_virtual/bootloader set to u-boot-tegra, not u-boot
ERROR: u-boot-tegra PROVIDES u-boot but was skipped: incompatible with machine jetson-tx2 (not in COMPATIBLE_MACHINE)
ERROR: Required build target 'core-image-sato' has no buildable providers.
Missing or unbuildable dependency chain was: ['core-image-sato', 'u-boot']

without adding the IMAGE_FSTYPES it builds fine.

I've got an issue that I'm wondering might have been encountered by others before.
After flashing the TX-1 using the doflash.sh script, something strange seems to have happened with the partition table.

Running lsblk, we get some pretty normal-looking results:

NAME         MAJ:MIN RM  SIZE RO TYPE MOUNTPOINT
mmcblk0rpmb  179:48   0    4M  0 disk 
mmcblk0boot0 179:16   0    4M  1 disk 
mmcblk0boot1 179:32   0    4M  0 disk 
mmcblk0      179:0    0 14.7G  0 disk 
|-mmcblk0p1  179:1    0   12G  0 part /
|-mmcblk0p2  179:2    0    2M  0 part 
|-mmcblk0p3  179:3    0    4M  0 part 
|-mmcblk0p4  179:4    0    2M  0 part 
|-mmcblk0p5  179:5    0    6M  0 part 
|-mmcblk0p6  179:6    0    4M  0 part 
|-mmcblk0p7  179:7    0    6M  0 part 
|-mmcblk0p8  179:8    0    2M  0 part 
|-mmcblk0p9  179:9    0    2M  0 part 
|-mmcblk0p10 179:10   0   20M  0 part 
|-mmcblk0p11 179:11   0   64M  0 part 
|-mmcblk0p12 179:12   0   64M  0 part 
|-mmcblk0p13 179:13   0    4M  0 part 
|-mmcblk0p14 179:14   0    2M  0 part 
|-mmcblk0p15 179:15   0    6M  0 part 
|-mmcblk0p16 259:0    0    6M  0 part 
|-mmcblk0p17 259:1    0    2M  0 part 
`-mmcblk0p18 259:2    0  2.5G  0 part

But taking a look at our actually useable space with df -ah, something is up:

Filesystem      Size  Used Avail Use% Mounted on
/dev/mmcblk0p1  667M  566M   65M  90% /

A partition that should have 12 GB of space has been reduced to only 667 MB!!!

We can further confirm that we're definitely only using about half a gig with du:

root@TX1:/# du -sbh /
612M    /

This has rendered the system unusable for my purposes. Anyone have any idea as to what might be going on or how to fix this? Thanks!

When loading a kernel from the eMMC device, the stock u-boot is able to copy data at a rate of ~36.2 MiB/s while the yocto u-boot only transfers at ~314.5 KiB/s.

Stock u-boot:

Tegra186 (P2771-0000-500) # version
U-Boot 2016.07-g0ce7ca2 (Jul 20 2017 - 00:45:25 -0700)
aarch64-unknown-linux-gnu-gcc (GCC) 4.8.5
GNU ld (GNU Binutils) 2.24

Tegra186 (P2771-0000-500) # mmc info
Device: Tegra SD/MMC
Manufacturer ID: 11
OEM: 100
Name: 032G3 
Tran Speed: 52000000
Rd Block Len: 512
MMC version 5.1
High Capacity: Yes
Capacity: 29.1 GiB
Bus Width: 8-bit
Erase Group Size: 512 KiB
HC WP Group Size: 4 MiB
User Capacity: 29.1 GiB WRREL
Boot Capacity: 4 MiB ENH
RPMB Capacity: 4 MiB ENH

Retrieving file: /boot/Image
20280368 bytes read in 534 ms (36.2 MiB/s)

Built u-boot:

Tegra186 (P2771-0000-500) # version
U-Boot 2016.07 (Aug 16 2017 - 13:55:00 +0000)
aarch64-poky-linux-gcc (GCC) 6.3.0
GNU ld (GNU Binutils) 2.28.0.20170307

Retrieving file: /boot/Image-4.4.38-l4t-r28.1+gebda89f
10563776 bytes read in 32753 ms (314.5 KiB/s)

Tegra186 (P2771-0000-500) # mmc info
Device: Tegra SD/MMC
Manufacturer ID: 11
OEM: 100
Name: 032G3 
Tran Speed: 52000000
Rd Block Len: 512
MMC version 5.1
High Capacity: Yes
Capacity: 29.1 GiB
Bus Width: 8-bit
Erase Group Size: 512 KiB
HC WP Group Size: 4 MiB
User Capacity: 29.1 GiB WRREL
Boot Capacity: 4 MiB ENH
RPMB Capacity: 4 MiB ENH

Can you please add some kind of "getting started" to the README file with instructions about building a basic system?

Thanks
Avishay

Hi,
i am trying to build an image for tx2. I have changed the machine name to jetson-tx2 and added IMAGE_CLASSES += "image_types_tegra" in the local.conf file
I am getting the following error when i execute the command bitbake core-image-sato

ERROR: linux-tegra-4.4+gitAUTOINC+79e4600e81-r0 do_compile: oe_runmake failed
ERROR: linux-tegra-4.4+gitAUTOINC+79e4600e81-r0 do_compile: Function failed: do_compile (log file is located at /home/madhumitha/yocto/tx2/tmp/work/jetson_tx2-poky-linux/linux-tegra/4.4+gitAUTOINC+79e4600e81-r0/temp/log.do_compile.10904)
ERROR: Logfile of failure stored in: /home/madhumitha/yocto/tx2/tmp/work/jetson_tx2-poky-linux/linux-tegra/4.4+gitAUTOINC+79e4600e81-r0/temp/log.do_compile.10904

Builds fail when using the pyro-l4t-r28.1 branch during the compilation of pango, librsvg, and gstreamer1.0-plugins-nveglgles if you have gobject-introspection-data in your DISTRO_FEATURES (which is on by default).

The scanner runs under qemu-aarch64, and that is exiting with a signal 11. Disabling gobject-introspection-data by setting that in DISTRO_FEATURES_BACKFILL_CONSIDERED, or setting GI_DATA_ENABLED to "False" for those recipes, works around the issue.

This appears to be similar to #20.

I have this in my IMAGE_INSTALL +=

    gstreamer1.0 \
    gstreamer1.0-meta-base \
    gstreamer1.0-omx-tegra \
    gstreamer1.0-plugins-bad \
    gstreamer1.0-plugins-base \
    gstreamer1.0-plugins-good \
    gstreamer1.0-plugins-nveglgles \
    gstreamer1.0-plugins-tegra \
    gstreamer1.0-vaapi \

But, I can't see any omx plugins

root@jetson-tx2:/mnt/gstpipes# gst-inspect-1.0 | grep omx
root@jetson-tx2:/mnt/gstpipes# 

I did try this from #20

DISTRO_FEATURES_append = " x11"

which I feel like I successfully used on TX1 but can't remember.

I'm trying to build this layer with rocko for the Jetson TX2. The wiki mentions that another layer is required to provide GCC version 5. What options are there for this? What's been tested to work? I tried using meta-linaro, but had problems just building the toolchain. That may be something I've done wrong though, so it'd be good to know what works before I spend more time on it.

I can see a few initial messages from the driver at boot:

[ 2.530680] e1000e: Intel(R) PRO/1000 Network Driver - 3.2.6-k
[ 2.530683] e1000e: Copyright(c) 1999 - 2015 Intel Corporation.
[ 2.530720] igb: Intel(R) Gigabit Ethernet Network Driver - version 5.3.0-k
[ 2.530722] igb: Copyright (c) 2007-2014 Intel Corporation.

But it never reaches the point where it binds to eth1.

Is there anything special I need to add to my image contents? As far as I can tell there is no firmware file required for the e1000e.

Is anyone else seeing this issue?

I'm hitting an issue compiling gnutls with what I assume is the tx1 kernel. Are you having the sames issues? You must be using an oe-core with gnutls 3.5.3 or greater. The gnutls feature that broke it I believe to be:

** libgnutls: Use getrandom where available via the syscall interface.
This works around an issue of not-using getrandom even if it exists
since glibc doesn't declare such function.

As far as I understand, the nvidia 3.10 kernel doesn't implement it, and the fallback is also failing?

When bitbake 'image' -c populate_sdk with cuda-misc-headers-dev in " RDEPENDS_${PN} += " path of my .bbappend file for toolchain-target I see bitbake error with:
.....
nothing provides cuda-misc-headers = 8.0.84+1-r0 needed by cuda-misc-headers-dev-8.0.84+1-r0.aarch64_tegra186
.....

So, adding
RDEPENDS_${PN}-dev = ""
in the end of recipes-devtools/cuda/cuda-misc-headers_8.0.84-1.bb solves the problem.

I'm using pyro-l4t-r28.1 , compiling for Jetson-tx2.

patchelf is introducing some corruption into the binary, so we need to figure out a different way of dealing with libdrm that will let us drop in the NVIDIA version but still use the built version for gobject introspection purposes.

Hi i'm new in this project, so i don't understand all..
First, before to put our yocto layer, we must to download Jetson TX1 64-bit Driver Package or all is integrated in this meta-tegra ?
Second, how we put this in the nvidia jetson tx1 ?
And third, how we choice our target for the bitbake (core-image-minimal, core-image-sato, ...) ?

Thank you for your attention.

cuda-toolkit combines a bunch of the cuda libraries that are packaged as separate .deb files. While that's convenient for building, it's not great from a packaging perspective -- a dependency on any one of the shareables in that package will bring them all in, driving up disk space use on the target. This came up with a build that uses tensorrt - that package needs libcublas, and so draws in the entire set of cuda-toolkit libraries.

So we need to package the libraries separately, just as NV does.

Hello,

I'm working with krogoth-l4t-r24.2.1 branch and I had an error about ext3 rootfs alignment during do_image_tegraflash :

DEBUG: Executing shell function do_image_tegraflash
4.0K	/home/jc/oe-core/build/tmp-glibc/deploy/images/jetson-tx1/core-image-minimal-jetson-tx1.ext3
size of /home/jc/oe-core/build/tmp-glibc/deploy/images/jetson-tx1/core-image-minimal-jetson-tx1.ext3 is not mulple of 4096.

I found that IMAGE_ROOTFS_ALIGNMENT ?= "4096" line, in meta-tegra/classes/image_types_tegra.bbclass, fails to override the value assigned in meta/classes/image_types.bbclass.

As this variable is in KiB, i beleive the correct value would be 4.

Would you accept a pull request on this matter?

I am trying to build an example image, core-image-sato.

When trying to build, I get this.

pknopf@ubuntu:~/git/tegra-test2/build$ bitbake core-image-sato
Parsing recipes: 100% |######################################################################################################################################################################| Time: 00:00:47
Parsing of 902 .bb files complete (0 cached, 902 parsed). 2242 targets, 230 skipped, 0 masked, 0 errors.
NOTE: There are 3 recipes to be removed from sysroot jetson-tx1, removing...
NOTE: Resolving any missing task queue dependencies
ERROR: Nothing PROVIDES 'virtual/libgl' (but /home/pknopf/git/tegra-test2/poky/meta/recipes-graphics/libsdl/libsdl_1.2.15.bb DEPENDS on or otherwise requires it)
ERROR: tegra-libraries PROVIDES virtual/libgl but was skipped: incompatible with host aarch64-poky-linux (not in COMPATIBLE_HOST)
ERROR: lib32-mesa-gl PROVIDES virtual/libgl but was skipped: incompatible with machine jetson-tx1 (not in COMPATIBLE_MACHINE)
ERROR: mesa-gl PROVIDES virtual/libgl but was skipped: incompatible with machine jetson-tx1 (not in COMPATIBLE_MACHINE)
NOTE: Runtime target 'libsdl' is unbuildable, removing...
Missing or unbuildable dependency chain was: ['libsdl', 'virtual/libgl']
NOTE: Runtime target 'packagegroup-core-x11-sato-games' is unbuildable, removing...
Missing or unbuildable dependency chain was: ['packagegroup-core-x11-sato-games', 'libsdl', 'virtual/libgl']
ERROR: Required build target 'core-image-sato' has no buildable providers.
Missing or unbuildable dependency chain was: ['core-image-sato', 'packagegroup-core-x11-sato-games', 'libsdl', 'virtual/libgl']

It seems that the NVidia libraries provided are only 32bit. Is this correct? What does this layer even support multi-lib then? Why not just compile everything (even then kernel) at 32bit?

Hi Matt,
I stumbled upon another error while compiling with CUDA nvcc:
standard library <new> and <cmath> are not found, below is just the first error about new: file not found. If I fix manually I get next error about cmath.

How do you set up gcc when compiling with nvcc for targets?
Is this a bug?

Errors below

| /home/ubuntu/Build/embedded_distribution/poky/build/tmp/sysroots/x86_64-linux/usr/local/cuda-7.0/bin/nvcc -M -D__CUDACC_ /home/ubuntu/
Build/embedded_distribution/poky/build/tmp/work/aarch64-poky-linux/sensorpro/0.0.1-r0/git/core/CudaWrapper.cu -o /home/ubuntu/Build/embed
ded_distribution/poky/build/tmp/work/aarch64-poky-linux/sensorpro/0.0.1-r0/build/core/CMakeFiles/sensorprocore.dir//sensorprocore_generat
ed_CudaWrapper.cu.o.NVCC-depend -ccbin /home/ubuntu/Build/embedded_distribution/poky/build/tmp/sysroots/x86_64-linux/usr/bin/aarch64-poky
-linux/aarch64-poky-linux-gcc -m64 -arch=sm_20 -std=c++11 -O2 -DVERBOSE -g -DNVCC -I/home/ubuntu/Build/embedded_distribution/poky/build/t
mp/sysroots/jetson-tx1/usr/local/cuda-7.0/include -I/home/ubuntu/Build/embedded_distribution/poky/build/tmp/sysroots/jetson-tx1/usr/local
/cuda-7.0/samples/common/inc -I/home/ubuntu/Build/embedded_distribution/poky/build/tmp/sysroots/jetson-tx1/usr/include -I/home/ubuntu/Bui
ld/embedded_distribution/poky/build/tmp/work/aarch64-poky-linux/sensorpro/0.0.1-r0/build -I/home/ubuntu/Build/embedded_distribution/poky/
build/tmp/work/aarch64-poky-linux/sensorpro/0.0.1-r0/git -I/home/ubuntu/Build/embedded_distribution/poky/build/tmp/sysroots/jetson-tx1/us
r/local/cuda-7.0/include
| In file included from /home/ubuntu/Build/embedded_distribution/poky/build/tmp/sysroots/jetson-tx1/usr/local/cuda-7.0/include/cuda_runti
me.h:105:0,
| from :0:
| /home/ubuntu/Build/embedded_distribution/poky/build/tmp/sysroots/jetson-tx1/usr/local/cuda-7.0/include/common_functions.h:87:15: fatal
error: new: No such file or directory
| compilation terminated.
| CMake Error at sensorprocore_generated_CudaWrapper.cu.o.cmake:203 (message):
| Error generating
| /home/ubuntu/Build/embedded_distribution/poky/build/tmp/work/aarch64-poky-linux/sensorpro/0.0.1-r0/build/core/CMakeFiles/sensorprocor
e.dir//./sensorprocore_generated_CudaWrapper.cu.o
|
|_

I would like to contribute to this but I'm not sure where to start. Is there an IRC channel on freenode that anyone here frequents where we can chat? I have a TX1 I'm planning on using for a project but would like to help out.

Hi,
when I try to run the X server on the TX1, I get the message that the X server video driver ABI version 23.0 is not supported by the nvidia driver. Is it possible to solve this somehow?

Last time I checked, you added a patch to supposedly fix Xorg, but it didn't work. Have you been able to successfully get it up and running yet? Maybe the qt5 cinematic experience demo?

If not, if you are interested in some contract work, email me at [email protected].

Hi,

I am new to Yocto but have been bringing myself up-to-speed. I am attempting both core-image-minimal and core-image-sato builds for the TX2. My setup steps:

• Checkout pyro branch of Yocto
• Checkout pyro-l4t-r28.1 branch of meta-tegra
• oe-init-build-env as usual
• Added the meta-tegra layer to bblayers.conf
• Added the following to local.conf:
  MACHINE ?= "jetson-tx2"
  IMAGE_CLASSES += "image_types_tegra"
  IMAGE_FSTYPES = "tegraflash"
  GCCVERSION = "5.%" # Cuda support

The build seems to progress nearly to completion but fails in the core-image-sato recipe do_image_complete() function. It gives a python traceback showing that it is trying to write the license manifest. The exception thrown is:

FileNotFound: /tmp/deploy/licenses/u-boot-tegra/recipeinfo

I am running this inside a preconfigured Docker image of Ubuntu 14.04 with all Yocto dependencies present. (I can build the basic default image for qemu/x86.) Any help is greatly appreciated.

Crash at boot time on TX1 systems if the kernel is compiled with gcc 7.3.0 from YP 2.4/OE-core rocko:

[    4.653481] Unexpected kernel BRK exception at EL1
[    4.653486] Unhandled debug exception: ptrace BRK handler (0xf20003e8) at 0x0000000000000000
[    4.653492] Internal error: : 0 [#1] PREEMPT SMP
[    4.653497] Modules linked in:
[    4.653505] CPU: 1 PID: 1 Comm: swapper/0 Not tainted 4.4.38-l4t-r28.2+g1a8a1c0ea08a #1
[    4.653507] Hardware name: jetson_tx1 (DT)
[    4.653512] task: ffffffc0fb2f0000 ti: ffffffc0fb2a4000 task.ti: ffffffc0fb2a4000
[    4.653524] PC is at update_clock_tree_delay.isra.0+0x908/0xc70
[    4.653530] LR is at update_clock_tree_delay.isra.0+0x908/0xc70
[    4.653534] pc : [<ffffffc000926cf0>] lr : [<ffffffc000926cf0>] pstate: 800000c5
[    4.653536] sp : ffffffc0fb2a7770
[    4.653542] x29: ffffffc0fb2a7770 x28: 0000000000040000 
[    4.653548] x27: 0000000000000004 x26: 000000000000000a 
[    4.653553] x25: 0000000000000271 x24: ffffffc0015bf000 
[    4.653558] x23: 0000000003200000 x22: 00000000201a51d8 
[    4.653562] x21: ffffffc0f9dc6118 x20: 0000000000000001 
[    4.653567] x19: 00000000000a2800 x18: 0000000000000001 
[    4.653572] x17: 0000000000000000 x16: 0000000000000000 
[    4.653577] x15: ffffffc00153dfb8 x14: 0000000000020000 
[    4.653581] x13: 00000000fffffff0 x12: ffffffc00153dbda 
[    4.653586] x11: ffffffc00135a760 x10: ffffffc00153c000 
[    4.653591] x9 : 0000000000000000 x8 : ffffffc00155abe3 
[    4.653596] x7 : 0000000000000000 x6 : 000000000004e8a4 
[    4.653600] x5 : 000000000000000b x4 : 0000000000000001 
[    4.653605] x3 : 0000000000000002 x2 : ffffffc0f9dcad58 
[    4.653610] x1 : 00000000000002a0 x0 : 00000000000002a0 
[    4.653611] 
[    4.653614] Process swapper/0 (pid: 1, stack limit = 0xffffffc0fb2a4020)
[    4.653616] Call trace:
[    4.653624] [<ffffffc000926cf0>] update_clock_tree_delay.isra.0+0x908/0xc70
[    4.653631] [<ffffffc0009277fc>] emc_set_clock_r21021+0x53c/0x1cf0
[    4.653637] [<ffffffc0009205bc>] emc_set_clock.constprop.1+0x34/0xb0
[    4.653643] [<ffffffc000920b28>] tegra210_emc_set_rate+0x4f0/0x550
[    4.653651] [<ffffffc000915464>] clk_emc_set_rate+0xa4/0x130
[    4.653659] [<ffffffc0008fd290>] clk_change_rate+0x2e0/0x550
[    4.653665] [<ffffffc0008fd598>] clk_core_set_rate_nolock+0x98/0xd0
[    4.653670] [<ffffffc0008fd63c>] clk_set_rate+0x6c/0x90
[    4.653676] [<ffffffc00090fb54>] _simple_shared_update+0x2c/0x38
[    4.653681] [<ffffffc00090e040>] clk_shared_bus_update+0x30/0x48
[    4.653686] [<ffffffc00090ea1c>] clk_shared_set_rate+0x74/0x90
[    4.653691] [<ffffffc0008fd290>] clk_change_rate+0x2e0/0x550
[    4.653697] [<ffffffc0008fd598>] clk_core_set_rate_nolock+0x98/0xd0
[    4.653702] [<ffffffc0008fd63c>] clk_set_rate+0x6c/0x90
[    4.653708] [<ffffffc0008eb1e8>] tegra_bwmgr_set_emc+0x3c0/0x440
[    4.653716] [<ffffffc0008f2e14>] actmon_dev_platform_init_t21x+0x9c/0x1c0
[    4.653722] [<ffffffc0008f2324>] tegra_actmon_register+0x624/0xc50
[    4.653728] [<ffffffc00123273c>] tegra21x_actmon_probe+0xa0/0xb4
[    4.653738] [<ffffffc000628be8>] platform_drv_probe+0x60/0xc0
[    4.653744] [<ffffffc0006266e4>] driver_probe_device+0x17c/0x3e8
[    4.653749] [<ffffffc0006269b8>] __driver_attach+0x68/0xa0
[    4.653756] [<ffffffc00062459c>] bus_for_each_dev+0x4c/0xa0
[    4.653762] [<ffffffc0006260e8>] driver_attach+0x20/0x28
[    4.653768] [<ffffffc000625a78>] bus_add_driver+0x140/0x278
[    4.653773] [<ffffffc000627838>] driver_register+0x98/0xe0
[    4.653779] [<ffffffc000628b30>] __platform_driver_register+0x48/0x50
[    4.653785] [<ffffffc001232694>] tegra21x_actmon_driver_init+0x18/0x20
[    4.653792] [<ffffffc0000814a4>] do_one_initcall+0x12c/0x1c0
[    4.653801] [<ffffffc0011edb58>] kernel_init_freeable+0x1dc/0x28c
[    4.653810] [<ffffffc000b9eaa0>] kernel_init+0x10/0xe0
[    4.653815] [<ffffffc0000847a0>] ret_from_fork+0x10/0x30
[    4.654162] ---[ end trace 1d8a0a9e4064d500 ]---

Looks like incorrect code being generated by the compiler.
Since rocko has GCC 6 available, and since CUDA 9 doesn't support GCC 7 anyway, for now just note that you have to use GCC 6 when compiling for the TX1.

I'm having 2 problems with the image generated with tegraflash:

1. u-boot-jetson-tx1.bin flashes properly but will not boot. I've been running morty, but have also tried pyro-l4t-r28.1 with the same results. I get the following output on the serial console from morty:

<LOTS OF OUTPUT DURING FLASH - NO ERRORS NOTED>
...
[0068.504] partition DTB write successful.
[0068.515] Writing BCT partition
[0068.518] bct_init bctinit
[0068.602] partition BCT write successful.
[0068.610] Change state to cold boot mode
[0068.614] Flushing data to stroage
[0000.183] [TegraBoot] (version 24.00.2015.42-mobile-ec3b827e)
[0000.189] Processing in cold boot mode Bootloader 2
[0000.193] A02 Bootrom Patch rev = 63
[0000.196] Power-up reason: software reset
[0000.200] No Battery Present
[0000.203] RamCode = 0
[0000.205] Platform has Ddr4 type ram
[0000.209] max77620 disabling SD1 Remote Sense
[0000.213] Setting Ddr voltage to 1125mv
[0000.217] Serial Number of Pmic Max77663: 0x811ad
[0000.224] Entering ramdump check
[0000.227] Get RamDumpCarveOut = 0x0
[0000.231] RamDumpCarveOut=0x0, RamDumperFlag=0xe59ff3f8
[0000.236] Last reboot was clean, booting normally!
[0000.240] Sdram initialization is successful
[0000.244] SecureOs Carveout Base=0xff800000 Size=0x00800000
[0000.250] GSC1 Carveout Base=0xff700000 Size=0x00100000
[0000.255] GSC2 Carveout Base=0xff600000 Size=0x00100000
[0000.260] GSC3 Carveout Base=0xff500000 Size=0x00100000
[0000.265] GSC4 Carveout Base=0xff400000 Size=0x00100000
[0000.270] GSC5 Carveout Base=0xff300000 Size=0x00100000
[0000.276] BpmpFw Carveout Base=0xff2c0000 Size=0x00040000
[0000.281] Lp0 Carveout Base=0xff2bf000 Size=0x00001000
[0000.296] RamDump Carveout Base=0xff23f000 Size=0x00080000
[0000.302] Platform-DebugCarveout: 0
[0000.305] Nck Carveout Base=0xff03f000 Size=0x00200000
[0000.310] Non secure mode. Disable rollback prevention
[0000.315] AOTAG Init Done
[0000.352] Using GPT Primary to query partitions
[0000.358] Loading Tboot-CPU binary
[0000.407] Verifying bootloader in OdmNonSecureSBK mode
[0000.417] Bootloader load address is 0xa0000000, entry address is 0xa0000258
[0000.424] Bootloader downloaded successfully.
[0000.428] Downloaded Tboot-CPU binary to 0xa0000258
[0000.433] MAX77620_GPIO1 Configured.
[0000.436] MAX77620_GPIO5 Configured.
[0000.440] CPU power rail is up
[0000.442] CPU clock enabled
[0000.446] Performing RAM repair
[0000.449] Updating A64 Warmreset Address to 0xa00002e9
[0000.466] Bootloader DTB Load Address: 0x83000000
[0000.483] Kernel DTB Load Address: 0x83080000
[0000.488] Bootloader is not valid
[0000.491] Error in NvTbootLoadBinary: 0x14 !
[0000.495] GPT: Partition NOT found !
[0000.498] Find Partition via GPT Failed
[0000.502] Find Partition via PT Failed
[0000.506] function NvTbootGetBinaryOffsets: 0x1 error
[0000.510] Error in NvTbootLoadBinary: 0x1 !
[0000.514] Error is 1

And the following from pyro:

[0032.845] partition DTB write successful.
[0032.856] Writing BCT partition
[0032.860] bct_init bctinit
[0032.942] partition BCT write successful.
[0032.950] Change state to cold boot mode
[0032.954] Flushing data to stroage
[0000.162] [TegraBoot] (version 00.00.2014.50-mobile-d44d4bf0)
[0000.168] Processing in cold boot mode Bootloader 2
[0000.172] A02 Bootrom Patch rev = 63
[0000.176] Power-up reason: software reset
[0000.179] No Battery Present
[0000.182] RamCode = 0
[0000.184] Platform has Ddr4 type ram
[0000.188] max77620 disabling SD1 Remote Sense
[0000.192] Setting Ddr voltage to 1125mv
[0000.196] Serial Number of Pmic Max77663: 0x811ad
[0000.203] Entering ramdump check
[0000.206] Get RamDumpCarveOut = 0x0
[0000.210] RamDumpCarveOut=0x0, RamDumperFlag=0xe59ff3f8
[0000.215] Last reboot was clean, booting normally!
[0000.219] Sdram initialization is successful
[0000.223] SecureOs Carveout Base=0xff800000 Size=0x00800000
[0000.229] GSC1 Carveout Base=0xff700000 Size=0x00100000
[0000.234] GSC2 Carveout Base=0xff600000 Size=0x00100000
[0000.239] GSC3 Carveout Base=0xff500000 Size=0x00100000
[0000.244] GSC4 Carveout Base=0xff400000 Size=0x00100000
[0000.249] GSC5 Carveout Base=0xff300000 Size=0x00100000
[0000.255] BpmpFw Carveout Base=0xff2c0000 Size=0x00040000
[0000.260] Lp0 Carveout Base=0xff2bf000 Size=0x00001000
[0000.275] RamDump Carveout Base=0xff23f000 Size=0x00080000
[0000.281] Platform-DebugCarveout: 0
[0000.284] Nck Carveout Base=0xff03f000 Size=0x00200000
[0000.289] Non secure mode, and RB not enabled.
[0000.319] Using GPT Primary to query partitions
[0000.324] Loading Tboot-CPU binary
[0000.373] Verifying bootloader in OdmNonSecureSBK mode
[0000.383] Bootloader load address is 0xa0000000, entry address is 0xa0000258
[0000.390] Bootloader downloaded successfully.
[0000.394] Downloaded Tboot-CPU binary to 0xa0000258
[0000.399] MAX77620_GPIO1 Configured.
[0000.403] MAX77620_GPIO5 Configured.
[0000.406] CPU power rail is up
[0000.409] CPU clock enabled
[0000.413] Performing RAM repair
[0000.416] Updating A64 Warmreset Address to 0xa00002e9
[0000.444] Bootloader DTB Load Address: 0x83000000
[0000.472] Kernel DTB Load Address: 0x83100000
[0000.477] Loading cboot binary
[0000.570] Verifying bootloader in OdmNonSecureSBK mode
[0000.609] Bootloader load address is 0x92c00000, entry address is 0x92c00258
[0000.616] Bootloader downloaded successfully.
[0000.620] GPT: Partition NOT found !
[0000.623] Find Partition via GPT Failed
[0000.627] Find Partition via PT Failed
[0000.631] function NvTbootGetBinaryOffsets: 0x1 error
[0000.635] Error in NvTbootLoadBinary: 0x1 !
[0000.639] Next binary entry address: 0x92c00258
[0000.644] BoardId: 2180
[0000.649] dram memory type is 3
[0000.653] WB0 init successful
[0000.679] Bpmp FW successfully loaded
[0000.682] Set NvDecSticky Bits
[0000.685] GSC1 address : ff700000
[0000.689] GSC2 address ff63fffc value c0edbbcc
[0000.694] GSC2 address : ff600000
[0000.697] GSC3 address : ff500000
[0000.701] GSC4 address : ff400000
[0000.705] GSC5 address : ff300000
[0000.708] GSC MC Settings done
[0000.711] TOS plaintext Image length 28752
[0000.715] Monitor size 28752
[0000.718] OS size 0
[0000.721] *** Secure OS image signature not verified ***
[0000.726] Loading and Validation of Secure OS Successful
[0000.732] NvTbootPackSdramParams: start.
[0000.737] NvTbootPackSdramParams: done.
[0000.740] Tegraboot started after 141606 us
[0000.744] Basic modules init took 289640 us
[0000.748] Sec Bootdevice Read Time = 215 ms, Read Size = 9476 KB
[0000.754] Sec Bootdevice Write Time = -1940251271 ms, Write Size = -68719477 KB
[0000.761] Next stage binary read took 23772 us
[0000.766] Carveout took 207062 us
[0000.769] CPU initialization took 96466 us
[0000.773] Total time taken by TegraBoot 616940 us

[0000.777] Starting CPU & Halting co-processor

64NOTICE: BL31: v1.2(release):cc5fd7c
NOTICE: BL31: Built : 00:37:02, Jul 20 2017
NOTICE: Trusty image missing.
ERROR: Error initializing runtime service trusty_fast
[0000.918] RamCode = 0
[0000.939] LPDDR4 Training: Read DT: Number of tables = 10
[0000.944] EMC Training (SRC-freq: 204000; DST-freq: 40800)
[0000.949] EMC Training Skipped
[0000.952] EMC Training (SRC-freq: 204000; DST-freq: 68000)
[0000.957] EMC Training Skipped
[0000.960] EMC Training (SRC-freq: 204000; DST-freq: 102000)
[0000.965] EMC Training Skipped
[0000.968] EMC Training (SRC-freq: 204000; DST-freq: 204000)
[0000.973] EMC Training Skipped
[0000.976] EMC Training (SRC-freq: 204000; DST-freq: 408000)
[0000.982] EMC Training Successful
[0000.985] EMC Training (SRC-freq: 204000; DST-freq: 665600)
[0000.991] EMC Training Successful
[0000.994] EMC Training (SRC-freq: 204000; DST-freq: 800000)
[0001.006] EMC Training Successful
[0001.009] EMC Training (SRC-freq: 204000; DST-freq: 1065600)
[0001.032] EMC Training Successful
[0001.035] EMC Training (SRC-freq: 204000; DST-freq: 1331200)
[0001.056] EMC Training Successful
[0001.059] EMC Training (SRC-freq: 204000; DST-freq: 1600000)
[0001.079] EMC Training Successful
[0001.082] Switching to 800000 KHz Success
[0001.092] RamCode = 0
[0001.096] DT Write: emc-table@40800 succeeded
[0001.102] DT Write: emc-table@68000 succeeded
[0001.108] DT Write: emc-table@102000 succeeded
[0001.114] DT Write: emc-table@204000 succeeded
[0001.120] DT Write: emc-table@408000 succeeded
[0001.126] DT Write: emc-table@665600 succeeded
[0001.132] DT Write: emc-table@800000 succeeded
[0001.138] DT Write: emc-table@1065600 succeeded
[0001.144] DT Write: emc-table@1331200 succeeded
[0001.150] DT Write: emc-table@1600000 succeeded
[0001.155] LPDDR4 Training: Write DT: Number of tables = 10
[0001.189]
[0001.190] Debug Init done
[0001.193] Marked DTB cacheable
[0001.196] Bootloader DTB loaded at 0x83000000
[0001.201] Marked DTB cacheable
[0001.204] Kernel DTB loaded at 0x83100000
[0001.208] DeviceTree Init done
[0001.227] Pinmux applied successfully
[0001.233] gicd_base: 0x50041000
[0001.237] gicc_base: 0x50042000
[0001.240] Interrupts Init done
[0001.246] Using base:0x60005008 & irq:33 for tick-timer
[0001.251] Using base:0x60005000 for delay-timer
[0001.255] platform_init_timer: DONE
[0001.259] Timer(tick) Init done
[0001.264] osc freq = 38400 khz
[0001.269]
[0001.271] welcome to cboot
[0001.273]
[0001.274] Cboot Version: 00.00.2014.50-t210-fadd1be5
[0001.279] calling constructors
[0001.282] initializing heap
[0001.285] initializing threads
[0001.288] initializing timers
[0001.291] creating bootstrap completion thread
[0001.295] top of bootstrap2()
[0001.298] CPU: ARM Cortex A57
[0001.301] CPU: MIDR: 0x411FD071, MPIDR: 0x80000000
[0001.306] initializing platform
[0001.332] config for ddr50 mode completed
[0001.336] sdmmc bdev is already initialized
[0001.340] Enable APE clock
[0001.342] Un-powergate APE partition
[0001.346] of_register: registering tegra_udc to of_hal
[0001.351] of_register: registering inv20628-driver to of_hal
[0001.357] of_register: registering ads1015-driver to of_hal
[0001.362] of_register: registering lp8557-bl-driver to of_hal
[0001.368] of_register: registering bq2419x_charger to of_hal
[0001.373] of_register: registering cpc to of_hal
[0001.378] of_register: registering bq27441_fuel_gauge to of_hal
[0001.398] gpio framework initialized
[0001.401] of_register: registering tca9539_gpio to of_hal
[0001.406] of_register: registering tca9539_gpio to of_hal
[0001.412] of_register: registering i2c_bus_driver to of_hal
[0001.417] of_register: registering i2c_bus_driver to of_hal
[0001.423] of_register: registering i2c_bus_driver to of_hal
[0001.428] pmic framework initialized
[0001.432] of_register: registering max77620_pmic to of_hal
[0001.437] regulator framework initialized
[0001.441] of_register: registering tegra_xhci to of_hal
[0001.446] initializing target
[0001.455] gpio_driver_register: register 'tegra_gpio_driver' driver
[0001.467] fixed regulator driver initialized
[0001.538] initializing OF layer
[0001.541] NCK carveout not present
[0001.545] Skipping dts_overrides
[0001.549] of_children_init: Ops found for compatible string nvidia,tegra210-xhci
[0001.557] of_children_init: Ops found for compatible string nvidia,tegra210-i2c
[0001.588] I2C Bus Init done
[0001.591] of_children_init: Ops found for compatible string nvidia,tegra210-i2c
[0001.606] I2C Bus Init done
[0001.608] of_children_init: Ops found for compatible string ti,tca9539
[0001.622] tca9539_init: i2c bus: 2, slave addr: 0xe8
[0001.627] I2C slave not started
[0001.630] I2C write failed
[0001.633] i2c write failed
[0001.635] Slave: 0xe8, Register 0x4 Instance 2, Buffer Size: 0
[0001.641] tca9539_device_init: failed to write polar reg
[0001.647] tca9539_init: failed to init device!
[0001.651] of_children_init: Ops found for compatible string ti,tca9539
[0001.665] tca9539_init: i2c bus: 2, slave addr: 0xee
[0001.670] I2C slave not started
[0001.673] I2C write failed
[0001.676] i2c write failed
[0001.678] Slave: 0xee, Register 0x4 Instance 2, Buffer Size: 0
[0001.684] tca9539_device_init: failed to write polar reg
[0001.689] tca9539_init: failed to init device!
[0001.694] of_children_init: Ops found for compatible string nvidia,tegra210-i2c
[0001.709] I2C Bus Init done
[0001.711] of_children_init: Ops found for compatible string nvidia,tegra210-i2c
[0001.726] I2C Bus Init done
[0001.729] of_children_init: Ops found for compatible string nvidia,tegra210-i2c
[0001.744] I2C Bus Init done
[0001.747] of_children_init: Ops found for compatible string maxim,max77620
[0001.762] max77620_init using irq 118
[0001.767] register 'maxim,max77620' pmic
[0001.772] gpio_driver_register: register 'max77620-gpio' driver
[0001.778] of_children_init: Ops found for compatible string nvidia,tegra210-i2c
[0001.793] I2C Bus Init done
[0001.798] NCK carveout not present
[0001.801] shim_invoke: No NCT, Calling dts updates
[0001.822] Find /i2c@7000c000's alias i2c0
[0001.826] get eeprom at 1-a2, size 256, type 0
[0001.830] get eeprom at 1-ae, size 256, type 0
[0001.846] Find /i2c@7000c400's alias i2c1
[0001.849] get eeprom at 2-a0, size 256, type 0
[0001.865] Find /i2c@7000c500's alias i2c2
[0001.869] get eeprom at 3-a0, size 256, type 0
[0001.873] get eeprom at 3-ae, size 256, type 0
[0001.885] Find /host1x/i2c@546c0000's alias i2c6
[0001.889] get eeprom at 7-a8, size 256, type 0
[0001.894] pm_ids_update: Updating 1,a2, size 256, type 0
[0001.899] I2C slave not started
[0001.902] I2C write failed
[0001.905] Writing offset failed
[0001.908] eeprom_init: EEPROM read failed
[0001.912] pm_ids_update: eeprom init failed
[0001.916] pm_ids_update: Updating 1,ae, size 256, type 0
[0001.921] I2C slave not started
[0001.924] I2C write failed
[0001.927] Writing offset failed
[0001.930] eeprom_init: EEPROM read failed
[0001.934] pm_ids_update: eeprom init failed
[0001.938] pm_ids_update: Updating 2,a0, size 256, type 0
[0001.943] I2C slave not started
[0001.946] I2C write failed
[0001.949] Writing offset failed
[0001.952] eeprom_init: EEPROM read failed
[0001.956] pm_ids_update: eeprom init failed
[0001.960] pm_ids_update: Updating 3,a0, size 256, type 0
[0001.990] pm_ids_update: The pm board id is 2180-1000-400
[0001.999] Adding plugin-manager/ids/2180-1000-400=/i2c@7000c500:module@0x50
[0002.009] pm_ids_update: pm id update successful
[0002.014] pm_ids_update: Updating 3,ae, size 256, type 0
[0004.020] I2C Rx fifo filled timeout
[0004.023] I2c Read failed
[0004.025] Slave: 0xae, Register 0x0 Instance 3, Buffer Size: 0
[0004.031] eeprom_init: EEPROM read failed
[0004.035] pm_ids_update: eeprom init failed
[0004.039] pm_ids_update: Updating 7,a8, size 256, type 0
[0004.046] GPIO driver for phandle 0xd6 could not be found
[0006.052] I2C Rx fifo filled timeout
[0006.056] I2c Read failed
[0006.058] Slave: 0xa8, Register 0x0 Instance 3, Buffer Size: 0
[0006.064] eeprom_init: EEPROM read failed
[0006.068] pm_ids_update: eeprom init failed
[0006.074] GPIO driver for phandle 0xd6 could not be found
[0006.107] updating /chosen/nvidia,wifi-mac node 00:04:4b:65:c3:bf
[0006.116] updating /chosen/nvidia,bluetooth-mac node 00:04:4b:65:c3:c0
[0006.125] updating /chosen/nvidia,ethernet-mac node 00:04:4b:65:c3:c1
[0006.132] Plugin Manager: Parse ODM data 0x00084000
[0006.168] Find /i2c@7000c000's alias i2c0
[0006.172] get eeprom at 1-a2, size 256, type 0
[0006.176] get eeprom at 1-ae, size 256, type 0
[0006.192] Find /i2c@7000c400's alias i2c1
[0006.196] get eeprom at 2-a0, size 256, type 0
[0006.211] Find /i2c@7000c500's alias i2c2
[0006.215] get eeprom at 3-a0, size 256, type 0
[0006.219] get eeprom at 3-ae, size 256, type 0
[0006.231] Find /host1x/i2c@546c0000's alias i2c6
[0006.235] get eeprom at 7-a8, size 256, type 0
[0006.240] pm_ids_update: Updating 1,a2, size 256, type 0
[0006.245] I2C slave not started
[0006.248] I2C write failed
[0006.251] Writing offset failed
[0006.254] eeprom_init: EEPROM read failed
[0006.258] pm_ids_update: eeprom init failed
[0006.262] pm_ids_update: Updating 1,ae, size 256, type 0
[0006.267] I2C slave not started
[0006.270] I2C write failed
[0006.273] Writing offset failed
[0006.276] eeprom_init: EEPROM read failed
[0006.280] pm_ids_update: eeprom init failed
[0006.284] pm_ids_update: Updating 2,a0, size 256, type 0
[0006.289] I2C slave not started
[0006.292] I2C write failed
[0006.295] Writing offset failed
[0006.298] eeprom_init: EEPROM read failed
[0006.302] pm_ids_update: eeprom init failed
[0006.306] pm_ids_update: Updating 3,a0, size 256, type 0
[0006.336] pm_ids_update: The pm board id is 2180-1000-400
[0006.344] Adding plugin-manager/ids/2180-1000-400=/i2c@7000c500:module@0x50
[0006.351] pm_ids_update: pm id update successful
[0006.355] pm_ids_update: Updating 3,ae, size 256, type 0
[0008.361] I2C Rx fifo filled timeout
[0008.365] I2c Read failed
[0008.367] Slave: 0xae, Register 0x0 Instance 3, Buffer Size: 0
[0008.373] eeprom_init: EEPROM read failed
[0008.377] pm_ids_update: eeprom init failed
[0008.381] pm_ids_update: Updating 7,a8, size 256, type 0
[0008.388] GPIO driver for phandle 0xd6 could not be found
[0010.394] I2C Rx fifo filled timeout
[0010.397] I2c Read failed
[0010.400] Slave: 0xa8, Register 0x0 Instance 3, Buffer Size: 0
[0010.406] eeprom_init: EEPROM read failed
[0010.409] pm_ids_update: eeprom init failed
[0010.416] GPIO driver for phandle 0xd6 could not be found
[0010.448] Chip UID is 00000001641225830800000007fb02c0
[0010.456] shim_cmdline_install: /chosen/bootargs: androidboot.modem=none androidboot.serialno=03225160366050c002fb androidboot.security=non-secure
[0010.469] Add serial number:03225160366050c002fb as DT property
[0010.477] calling apps_init()
[0010.482] Found 18 GPT partitions in "sdmmc3_user"
[0010.487] Proceeding to Cold Boot
[0010.490] starting app android_boot_app
[0010.494] Device state: unlocked
[0010.497] display console init
[0010.502] skip-display-init in bootloader
[0010.506] set_instance_and_out_type: invalid out_type
[0010.511] display_console_init: No display init
[0010.522] Gpio keyboard init success
[0010.525] bct_init bctinit
[0010.528] device_query_partition_size: failed to open partition sdmmc3_user:BMP !
[0010.535] BMP partition read failed!
[0010.539] blob_init: blob-partitDT entry for leds-pwm not found
i[0010.553] on BMP header read failed
[0010.556] load_bmp_blob: BMP blob initialization failed
[0010.561] Could not load/initialize BMP blob...ignoring
[0010.566] bl_battery_charging: connected to external power supply
[0010.578] display_console_ioctl: No display init
[0010.582] switch_backlight failed
[0010.585] bct_init bctinit
[0010.588] device_query_partition_size: failed to open partition sdmmc3_user:MSC !
[0010.596] MSC Partition not found
[0010.599] bct_init bctinit
[0010.601] bct_init bctinit
[0010.605] blob_init: USP partition does not have valid Blob
[0010.610] Loading android kernel
[0010.614] Android boot image magic number invalid
[0010.619] Failed while loading kernel and ramdisk images
[0010.624] Failed to boot Android
[0010.627] starting fastboot mode
[0010.633] fastboot cmd_init done.
[0010.636] platform does not support off-mode-charge
[0010.641] usbdcd_init Initialize USBF driver
[0010.645] usbdcd_phy_open oscfreq = 5
[0010.654] usbdcd_start Start the initialized controller
[0010.[0010.662] -- suspend --
662] display_console_clear: No display init
[0010.669] Error in fastboot menu
[0010.672] Continue booting ...
[0010.675]
[0010.676] display_console_clear: No display init
[0010.680] Loading android kernel
[0010.684] Android boot image magic number invalid
[0010.689] Failed while loading kernel and ramdisk images
[0010.694] Failed to boot Android
[0010.697] starting fastboot mode
[0010.732] fastboot cmd_init done.
[0010.735] platform does not support off-mode-charge
[0010.740] usbdcd_idisplay_console_clear: No display init
[0010.748] Error in fastboot menu
[0010.751] Continue booting ...
[0010.754]
[0010.755] display_console_clear: No display init
[0010.759]
[0010.761] -----------------------------------------------
[0010.766] Synchronous Exception: DATA ABORT (FAR: 8)
[0010.771] -----------------------------------------------
[0010.776] PAR_ELX: 0x80b
[0010.779]
[0010.780] ESR 0x96000005: ec 0x25, il 0x1, iss 0x5
[0010.785] -----------------------------------------------
[0010.790] [Stack Trace]
[0010.792]
[0010.794] => pc:0x92C39C94, sp:0x92C9A6F0
[0010.797] => pc:0x92C3893C, sp:0x92C9A920
[0010.801] => pc:0x92C38854, sp:0x92C9A940
[0010.805] => pc:0x92C03DEC, sp:0x92C9A950
[0010.809] => pc:0x92C38948, sp:0x92C9A9D0
[0010.813] => pc:0x92C38854, sp:0x92C9AB30
[0010.817] => pc:0x92C03DEC, sp:0x92C9AB40
[0010.821] => pc:0x92C0464C, sp:0x92C9ABC0
[0010.825] => pc:0x92C036C0, sp:0x92C9AD20
[0010.829] => pc:0x92C02CF8, sp:0x92C9ADA0
[0010.833] => pc:0x92C02CC0, sp:0x92C9ADB0
[0010.837] -----------------------------------------------
[0010.842] iframe 0x92c9a600:
[0010.845] x0 0x 92c64000 x1 0x 0 x2 0x 260 x3 0x 2
[0010.854] x4 0x a x5 0x 92c9a6e0 x6 0x 20 x7 0x 6d365b1f
[0010.864] x8 0x a x9 0x 7 x10 0x 92c9260d x11 0x 401
[0010.873] x12 0x 10000000 x13 0x 40 x14 0x 1 x15 0x 92cb6958
[0010.882] x16 0x ffff0 x17 0x 10000 x18 0x 0 x19 0x 92c90000
[0010.892] x20 0x 92c64518 x21 0x 83028a30 x22 0x 92c64000 x23 0x 92c90000
[0010.901] x24 0x 400 x25 0x 24 x26 0x 92c9a9c0 x27 0x 0
[0010.910] x28 0x 0 x29 0x 92c9a920 lr 0x 92c38940 sp 0x 92c9a6f0
[0010.919] elr 0x 92c39c94
[0010.923] spsr 0x 60000309
[0010.926] -----------------------------------------------
[0010.931] panic (caller 0x92c01238): die
[0010.935] HALT: spinning forever...

2. I can workaround the above problem by changing the u-boot image created to the NVIDIA provided one in flash.xml (changing u-boot-jetson-tx1.bin to u-boot-dtb.bin). However, when I flash this way, the DTB doesn't seem to be included in the main image. I checked the ext3 file and it doesn't have the DTB in /boot and listing the contents of /boot on the eMMC in u-boot shows that it is missing there as well.

How can I resolve these issues? Both morty and pyro exhibit the same behaviors (though I really like not relying on the NVIDIA JetPack install for the flashing tools). Is there a specific hash I need to be on with poky and meta-openembedded?

What am I doing wrong?

And on a related note I'd like to know the best way of specifying a different DTB for my hardware. I'm utilizing carrier boards from ConnectTech and they provide DTB files defining their hardware. How can I make those override the default DTB?

Thanks!

The deps are in place for the application that is being built:
RDEPENDS_${PN} = "cuda-toolkit cuda-cudart-dev"
DEPENDS = "cuda-cudart"
| CMake Error at /home/ubuntu/poky/build/tmp/sysroots/x86_64-linux/usr/share/cmake-3.4/Modules/FindCUDA.cmake:617 (message):
| Specify CUDA_TOOLKIT_ROOT_DIR
| Call Stack (most recent call first):
| CMakeLists.txt:33 (find_package)

not really an issue, but any plans to support the TX2? This layer has been very useful for the TX1.

With L4T R28.1, NVIDIA provides experimental support for Wayland. Need to investigate that and see if it can be integrated.

Hi,

I compiled core-image-sato-sdk system, it boots and everything works except that when I try to compile in the Jetson board using nvcc it throws an error:

sh: cicc: command not found

After reading in the blog of NVIDIA, I understood that the error is produced because nvcc is not finding nvvm (cicc). I reviewed the nvcc.profile file an effectively, the route of the file does not exists in the root file system. The weirdest part is that in the Packages.gz file of cuda-repo-l4t-8-0-local_8.0.34-1_arm64.deb, cicc is not mentioned. In equivalent files, such cuda-repo-ubuntu1504-7-5-local_7.5-18_amd64.deb, the Packages.gz files indicates that cicc is part of the cuda-core-.deb package. I checked that core, command-line-tools, and other packages appear in the list of cuda-toolkit_8.0.34-1.bb file. Also, I added IMAGE_INSTALL+= " cudnn cuda-toolkit" to local.conf file. What do you think can be the problem?

Thank you very much, you did an amazing job developing meta-tegra.

Sergio

Hi Matt,

What are you thoughts on ways to integrate this layer with an external layer that provides CUDA support for the desktop? I have a hacked up layer locally that (somewhat) emulates the CUDA packages available in this layer but for x86_64 desktop systems, so I use the same naming convention as here to keep compatibility with the cuda bbclass.

I had envisioned something like a meta-nvidia which provides x86 cuda support, and maybe x86 driver support (my current use case is in containers with the drivers passed through). However I don't know how to share the cuda bbclass without either layer depending on the other.

One idea would be to move the CUDA support out of meta-tegra into meta-nvidia and have cuda support in there, maybe have a tree like

meta-nvidia -> meta-cuda
-> meta-nvidia-driver

then meta-cuda would be a dependancy of meta-tegra if you wanted the cuda support and the bbclass and all recipe names would be in sync.

I understand that this would be a pain, so I'm open to suggestions a good way to handle it.

Hi, I want to build Linux which should support Elroy Carrier. What would be the best way of doing it? connect tech has provided there BSP to support TX2, can I use the .dtb given by connect tech.
Please help

They are currently extracted in binary form, but sources are available.

Tied a couple different approaches to running deviceQuery on Jetson TX2. Segfaults every time. I'm fairly certain it has to do with mismatches in library versions etc.

Is there a way to build it as part of generating the image? Can I generate an SDK that will allow me to build it and match the versions?

Building for "core-image-minimal" or "core-image-base" I don't seem to end up with anything in tmp-glibc/deploy/images/jetson-tk1 that would obviously be dd'ed to an sdcard:

core-image-base-jetson-tk1-20170530054448.rootfs.manifest
core-image-base-jetson-tk1-20170530054448.rootfs.tar.gz
core-image-base-jetson-tk1-20170530054448.testdata.json
core-image-base-jetson-tk1.manifest -> core-image-base-jetson-tk1-20170530054448.rootfs.manifest
core-image-base-jetson-tk1.tar.gz -> core-image-base-jetson-tk1-20170530054448.rootfs.tar.gz
core-image-base-jetson-tk1.testdata.json -> core-image-base-jetson-tk1-20170530054448.testdata.json
...

There is the *rootfs.tar.gz, but wouldn't that assume I've already partitioned and formatted the card?

Hi! Once I built my img with core-image-sato, I want to add ROS in jetson-tx1.
So for that, I created an initial build directory with in my local.conf meta-ros and his dependencies.
Build Configuration:
BB_VERSION = "1.30.0"
BUILD_SYS = "x86_64-linux"
NATIVELSBSTRING = "Ubuntu-16.04"
TARGET_SYS = "aarch64-poky-linux"
MACHINE = "jetson-tx1"
DISTRO = "poky"
DISTRO_VERSION = "2.1.1"
TUNE_FEATURES = "aarch64"
TARGET_FPU = ""
meta
meta-poky
meta-yocto-bsp = "krogoth:204b2bae4a5958735fdd19c63a69f4ed3780bba7"
meta-oe
meta-python
meta-multimedia = "krogoth:851a064b53dca3b14dd33eaaaca9573b1a36bf0e"
meta-tegra = "krogoth-l4t-r24.2:c5494deee41d2ec1165e2490339117b7f771a731"
meta-ros = "master:35cb79a0a1cd4ec6e92a197ec1d4f5d26ea8ae29"

Then I run with core-image-ros-roscore.
But the building completed, I have a file core-image-ros-roscore-jetson-tx1.tar.gz and inside there is several directory :
bin dev home media opt run sys usr
boot etc lib mnt proc sbin tmp var

After I don't know how to add this in my project with core-image-sato img in order to flash all in jetson-tx1.

Thank you.

As mentioned in #12, the xusb firmware needs to be available during early kernel initialization, particularly if the rootfs is on a USB device.

Possible approaches to resolve this:

1. Build the firmware into the kernel image via CONFIG_EXTRA_FIRMWARE. This might not be great from a licensing perspective.
2. For non-USB boots, use systemd-udevd instead of eudev. systemd-udevd includes the firmware loading helper, which was removed from eudev. Patches for eudev to restore that functionality could also be added.
3. Use an initrd that at least has the needed firmware. This is the approach taken by L4T.

The CUDA libraries are installed in the image, but are located at a non-standard location (/usr/local/cuda-X.Y/lib) and therefore cannot be found at runtime. What's the best way to solve it? Add a line to ld.so.conf? Create symbolic links for /lib? How can I implement those solutions?

Hi

Now I am build AGL with meta-tegra layer.
So, I get AGL image for jetson-tx1 EVK successfully Thanks.

But When I boot AGL on jetson-tx1 EVK, I can't see weston-launcher screen.
So, I excuted weston-laucher manually and get follow errors.

root@jetson-tx1:~# /usr/bin/weston-launch --tty /dev/tty1 --user root -- --idle-time=4294967
Date: 2017-06-02 UTC
[03:34:31.978] weston 1.11.0
               http://wayland.freedesktop.org
               Bug reports to: https://bugs.freedesktop.org/enter_bug.cgi?product=Wayland
component=weston&version=1.11.0
               Build: 1.10.93-2-g2d825ed configure.ac: bump to version 1.11.0 for the official release (2016-05-31 17:10:40 -0700)
[03:34:31.978] OS: Linux, 3.10.96+, #14 SMP PREEMPT Mon Jun 5 18:42:52 KST 2017, aarch64
[03:34:31.979] Using config file '/etc/xdg/weston/weston.ini'
[03:34:31.979] Output repaint window is 7 ms maximum.
[03:34:31.979] Loading module '/usr/lib/weston/drm-backend.so'
[03:34:31.982] initializing drm backend
[03:34:31.982] logind: failed to get session seat
[03:34:31.982] logind: cannot setup systemd-logind helper (-61), using legacy fallback
[03:34:31.983] no drm device found
[03:34:31.983] fatal: failed to create compositor backend
failed to restore keyboard mode: Bad file descriptor
failed to set KD_TEXT mode on tty: Bad file descriptor

I think that I need DRM for executing weston-launcher.
So, I added TEGRA_DRM using 'make menuconfig'.
But Still I got same error as well. I think DRM is not working very well.

Can I get some information about that?
Thanks,

I'm trying to have this command to work:

gst-launch-1.0 fakesrc \
    ! omxh265enc ! 'video/x-h265, stream-format=(string)byte-stream' \
    ! filesink location=test.mkv

Everything is OK running this on Nvidia's Ubuntu system, but on the one I built with oe, it gives:

WARNING: erroneous pipeline: no element "omxh265enc"

gst-inspect-1.0 output doesn't show all the omx* pipelines :

root@jetson-tx1:~# gst-inspect-1.0 |grep omx
omx:  omxmpeg4videodec: OpenMAX MPEG4 Video Decoder
omx:  omxh264dec: OpenMAX H.264 Video Decoder
omx:  omxmpeg4videoenc: OpenMAX MPEG4 Video Encoder
omx:  omxaacenc: OpenMAX AAC Audio Encoder

Wheras all of these are present in nvidia the system:

ubuntu@tegra-ubuntu:~$ gst-inspect-1.0  | grep omx
omx:  omxmpeg4videodec: OpenMAX MPEG4 Video Decoder
omx:  omxh264dec: OpenMAX H.264 Video Decoder
omx:  omxh265dec: OpenMAX H.265 Video Decoder
omx:  omxvp8dec: OpenMAX VP8 Video Decoder
omx:  omxvp9dec: OpenMAX VP9 Video Decoder
omx:  omxmpeg2videodec: OpenMAX MPEG2 Video Decoder
omx:  omxwmvdec: OpenMAX WMV Video Decoder
omx:  omxaacdec: OpenMax AAC decoder
omx:  omxmpegaudiodec: OpenMax MP3/MP2 decoder
omx:  omxh264enc: OpenMAX H.264 Video Encoder
omx:  omxh265enc: OpenMAX H.265 Video Encoder
omx:  omxvp8enc: OpenMAX VP8 Video Encoder
omx:  nvhdmioverlaysink: OpenMax HDMI Video Sink
omx:  nvoverlaysink: OpenMax Video Sinkk

Some omxh265enc seem to be compiled into libgstomx.so though.

root@jetson-tx1:~# grep -c omxh265enc /usr/lib/gstreamer-1.0/libgstomx.so
3

Am I missing some configuration options or depency ?

Here is my OE local.conf :

MACHINE ?= "jetson-tx1"
DISTRO_FEATURES = "opengl x11"
IMAGE_INSTALL_append = " \
kernel \
kernel-devicetree \
kernel-image \
kernel-modules \
kernel-vmlinux \
libdrm \
gstreamer1.0-libav \
gstreamer1.0-plugins-bad \
gstreamer1.0-plugins-base \
gstreamer1.0-plugins-tegra \
gstreamer1.0-plugins-good \
gstreamer1.0-omx-tegra \
gstreamer1.0-plugins-nveglgles"

And the configuration information :

Build Configuration:
BB_VERSION        = "1.30.0"
BUILD_SYS         = "x86_64-linux"
NATIVELSBSTRING   = "Ubuntu-16.04"
TARGET_SYS        = "aarch64-oe-linux"
MACHINE           = "jetson-tx1"
DISTRO            = "nodistro"
DISTRO_VERSION    = "nodistro.0"
TUNE_FEATURES     = "aarch64"
TARGET_FPU        = ""
meta              = "krogoth:28da89a20b70f2bf0c85da6e8af5d94a3b7d76c9"
meta-tegra        = "krogoth-l4t-r24.2.1:119e0fd8c2f49b33fbd5a5849f3926a8e3c58eb2"

NVIDIA released L4T R21.6 in October 2017. Do you plan to support it?

Working through #12

and having

IMAGE_CLASSES += "image_types_tegra"
IMAGE_FSTYPES = "tegraflash"

at the end of my local.conf.

I get this when I try to bitbake core-image-sato (core-image-minimal works fine)

| DEBUG: Executing python function set_image_size
| DEBUG: Python function set_image_size finished
| DEBUG: Executing shell function do_image_tegraflash
| size of /opt/Yocto/krogoth/tmp/deploy/images/jetson-tx1/core-image-sato-jetson-tx1.ext3 is not mulple of 4096.
| WARNING: exit code 255 from a shell command.
| ERROR: Function failed: do_image_tegraflash

I tried turning on debugging and looking at the log file but there's not much information beyond the above. be curious if you have any thoughts on things I may be doing wrong.

Building against Yocto Rocko (2.4), after commit 7c89171, the Kernel recipe fails at the packaging stage:

ERROR: linux-tegra-4.4+gitAUTOINC+79e4600e81-r0 do_package: QA Issue: linux-tegra: Files/directories were installed but not shipped in any package:
  /boot/initrd
  /boot/extlinux
  /boot/extlinux/extlinux.conf

The initscripts (or systemd services and related scripts) and udev rules have been updated in R28.1. They need revewing and either direct importing or translating.

I am using this layer (pyro) together with meta-ros. When I try to compile the image-geometry recipe I get the following error:

Loading cache...done.
Loaded 3145 entries from dependency cache.
Parsing recipes...done.
Parsing of 2379 .bb files complete (2375 cached, 4 parsed). 3149 targets, 197 skipped, 0 masked, 0 errors.
NOTE: Resolving any missing task queue dependencies

Build Configuration:
BB_VERSION        = "1.34.0"
BUILD_SYS         = "x86_64-linux"
NATIVELSBSTRING   = "universal"
TARGET_SYS        = "aarch64-poky-linux"
MACHINE           = "jetson-tx1"
DISTRO            = "poky"
DISTRO_VERSION    = "2.3"
TUNE_FEATURES     = "aarch64"
TARGET_FPU        = ""
meta              
meta-poky         = "HEAD:e5aa1ac23669a45491de9ecffac491c01e70203f"
meta-oe           
meta-multimedia   
meta-python       
meta-networking   = "HEAD:5e82995148a2844c6f483ae5ddd1438d87ea9fb7"
meta-tegra        = "HEAD:dd2912a420796eb4fb3742dca670721146941930"
meta-ros          = "HEAD:aebd5b3195b35a5a15c8e21397e357dae25b7fa2"

Initialising tasks...done.
NOTE: Executing SetScene Tasks
NOTE: Executing RunQueue Tasks
NOTE: Running task 2122 of 2122 (/data/test/jetson-ros-bsp/sources/meta-ros/recipes-ros/vision-opencv/image-geometry_1.11.15.bb:do_configure)
NOTE: recipe image-geometry-1.11.15-r0: task do_configure: Started
Log data follows:
| DEBUG: Executing shell function do_configure
| NOTE: cmake.bbclass no longer uses OECMAKE_BUILDPATH.  The default behaviour is now out-of-tree builds with B=WORKDIR/build.
| -- The C compiler identification is GNU 5.4.0
| -- The CXX compiler identification is GNU 5.4.0
| -- Check for working C compiler: /data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/recipe-sysroot-native/usr/bin/aarch64-poky-linux/aarch64-poky-linux-gcc
| -- Check for working C compiler: /data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/recipe-sysroot-native/usr/bin/aarch64-poky-linux/aarch64-poky-linux-gcc -- works
| -- Detecting C compiler ABI info
| -- Detecting C compiler ABI info - done
| -- Detecting C compile features
| -- Detecting C compile features - done
| -- Check for working CXX compiler: /data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/recipe-sysroot-native/usr/bin/aarch64-poky-linux/aarch64-poky-linux-g++
| -- Check for working CXX compiler: /data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/recipe-sysroot-native/usr/bin/aarch64-poky-linux/aarch64-poky-linux-g++ -- works
| -- Detecting CXX compiler ABI info
| -- Detecting CXX compiler ABI info - done
| -- Detecting CXX compile features
| -- Detecting CXX compile features - done
| -- Using CATKIN_DEVEL_PREFIX: /data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/build/devel
| -- Using CMAKE_PREFIX_PATH: /data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/recipe-sysroot/opt/ros/indigo;/data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/recipe-sysroot/usr;/data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/recipe-sysroot-native/opt/ros/indigo;/data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/recipe-sysroot-native/usr
| -- This workspace overlays: /data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/recipe-sysroot/opt/ros/indigo;/data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/recipe-sysroot/usr;/data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/recipe-sysroot-native/opt/ros/indigo;/data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/recipe-sysroot-native/usr
| -- Found PythonInterp: /data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/recipe-sysroot-native/usr/bin/python-native/python (found version "2.7.13")
| -- Using PYTHON_EXECUTABLE: /data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/recipe-sysroot-native/usr/bin/python-native/python
| -- Using default Python package layout
| -- Found PY_em: /data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/recipe-sysroot-native/usr/lib/python2.7/site-packages/em.pyc
| -- Using empy: /data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/recipe-sysroot-native/usr/lib/python2.7/site-packages/em.pyc
| -- Using CATKIN_ENABLE_TESTING: 0
| -- catkin 0.6.19
| CMake Error at /data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/recipe-sysroot-native/usr/share/cmake-3.7/Modules/FindPackageHandleStandardArgs.cmake:138 (message):
|   Could NOT find CUDA (missing: CUDA_INCLUDE_DIRS CUDA_CUDART_LIBRARY) (found
|   suitable exact version "8.0")
| Call Stack (most recent call first):
|   /data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/recipe-sysroot-native/usr/share/cmake-3.7/Modules/FindPackageHandleStandardArgs.cmake:378 (_FPHSA_FAILURE_MESSAGE)
|   /data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/recipe-sysroot-native/usr/share/cmake-3.7/Modules/FindCUDA.cmake:1013 (find_package_handle_standard_args)
|   /data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/toolchain.cmake:42 (find_package)
|   /data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/recipe-sysroot/usr/share/OpenCV/OpenCVConfig.cmake:105 (find_host_package)
|   CMakeLists.txt:5 (find_package)
| 
| 
| -- Configuring incomplete, errors occurred!
| See also "/data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/build/CMakeFiles/CMakeOutput.log".
| WARNING: exit code 1 from a shell command.
| ERROR: Function failed: do_configure (log file is located at /data/test/jetson-ros-bsp/build/tmp/work/jetson_tx1-poky-linux/image-geometry/1.11.15-r0/temp/log.do_configure.15756)
NOTE: recipe image-geometry-1.11.15-r0: task do_configure: Failed
NOTE: Tasks Summary: Attempted 2122 tasks of which 2121 didn't need to be rerun and 1 failed.

Summary: 1 task failed:
  /data/test/jetson-ros-bsp/sources/meta-ros/recipes-ros/vision-opencv/image-geometry_1.11.15.bb:do_configure
Summary: There was 1 ERROR message shown, returning a non-zero exit code.

I also tried to inherit the cuda.bbclass in the image-geometry_1.11.15.bb recipe

diff --git a/recipes-ros/vision-opencv/image-geometry_1.11.15.bb b/recipes-ros/vision-opencv/image-geometry_1.11.15.bb
index 3ed5a84..f2cbc9a 100644
--- a/recipes-ros/vision-opencv/image-geometry_1.11.15.bb
+++ b/recipes-ros/vision-opencv/image-geometry_1.11.15.bb
@@ -6,3 +6,5 @@ LIC_FILES_CHKSUM = "file://package.xml;beginline=12;endline=12;md5=d566ef916e9de
 DEPENDS = "opencv sensor-msgs"
 
 require vision-opencv.inc
+
+inherit cuda

@madisongh i'm following the instructions at the end of this issue #12. The one difference being I'm using core-image-minimal

it seems to go ok but it doesn't appear to flash a kernel or rootfs. After it reboots I just get a u-boot prompt. the doflash.sh happens very quickly but does seem to end ok

[   3.2199 ] Flashing completed

[   3.2199 ] Coldbooting the device
[   3.2206 ] tegradevflash --reboot coldboot
[   3.2211 ] Cboot version 00.01.0000
[   3.2231 ] 

Is there a next step I need to do?

See here.

Do you plan to update this? Is this an issue at run-time? Care for a pull-request?

When building an image, the device tree doesn't get populated in the final image, although it is being referenced by extlinux.conf.
Adding kernel-devicetree to IMAGE_INSTALL solves the problem, but to my understanding it should have happened automatically.