Make sure clDNN lib is installed properly. Fixes:
| ERROR: dldt-inference-engine-2020r1-r0 do_package: QA Issue: dldt-inference-engine: Files/directories were installed but not shipped in any package:
| /usr/lib
| /usr/lib/libclDNN64.a
| Please set FILES such that these items are packaged. Alternatively if they are unneeded, avoid installing them or delete them within do_install.
| dldt-inference-engine: 2 installed and not shipped files. [installed-vs-shipped]
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Backport patches to fix compile problems like:
| /build/build/tmp/work/corei7-64-poky-linux/intel-graphics-compiler/1.0.10395-r0/git/visa/LocalScheduler/SWSB_G4IR.cpp:1459:49: error: expected primary-expression before 'int'
| 1459 | int maxTokenDelay = std::numeric_limits<int>::min(); //The delay may cause if reuse
| | ^~~
| /build/build/tmp/work/corei7-64-poky-linux/intel-graphics-compiler/1.0.10395-r0/git/visa/LocalScheduler/SWSB_G4IR.cpp:1460:37: error: 'numeric_limits' is not a member of 'std'
| 1460 | int minTokenDistance = std::numeric_limits<int>::max(); //The distance from the reused node
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
(cherry picked from commit 311b432394)
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Backport a patch to fix errors like:
12:14:05 | /poky/build/tmp/work/x86_64-linux/intel-compute-runtime-native/20.15.16524-r0/git/shared/source/helpers/basic_math.h:166:21: error: ânumeric_limitsâ is not a member of âstdâ
12:14:05 | 166 | return std::numeric_limits<T>::max();
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Fixes errors when building on debian-11 or Ubuntu 22.04:
10:59:41 | /build/tmp/work/x86_64-linux/intel-graphics-compiler-native/1.0.3771-r0/git/visa/iga/IGALibrary/IR/SWSBSetter.cpp:713:39: error: ânumeric_limitsâ is not a member of âstdâ
10:59:41 | 713 | const InstIDs math_id = {std::numeric_limits<uint32_t>::max(), 0};
10:59:41 | | ^~~~~~~~~~~~~~
22:06:45 | /build/tmp/work/x86_64-linux/intel-graphics-compiler-native/1.0.3771-r0/git/visa/iga/IGAExe/../IGALibrary/api/igax.hpp:501:30: error: argument to âallocaâ may be too large [-Werror=alloca-larger-than=]
22:06:45 | 501 | strPtr = (char *)alloca(strCap); \
22:06:45 | | ^~~~~~
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
We're checking for REQUIRED_DISTRO_FEATURES so make sure that it
actually is used and inherit features_check. Fixes:
| lms-2022.0.0.0-r0 do_package_qa: QA Issue: lms: recipe doesn't inherit features_check [unhandled-features-check]
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
(cherry picked from commit 34f55b4c75)
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
The project has been moved so update the SRC_URI and the original 2020.1
release commit is now only available on master branch.
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
as meta-intel drop support for 4.19 kernel, this bbappend are no longer
required.
reverts commit 108c6938a9.
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
The name of project has changed to oneDNN. Enable GPU engine to be built
by default which would need clang layer to be included.
Also include two minor fixes to install logic that have been submitted
for review upstream.
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Error log:
clang (LLVM option parsing): Unknown command line argument '-pgo-warn-misexpect'.
Try: 'clang (LLVM option parsing) --help'
Force opencl-clang to link with shared clang-cpp for target builds and
explicitly link with spirv library because of a change in behaviour in llvm 10.
Use the patch to link with all static libs only for native cases where meta-clang
default to building static libs only by default.
Signed-off-by: Naveen Saini <naveen.kumar.saini@intel.com>
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
We don't want to install sample sources. So delete them and install the
binaries instead. Also make sure that we're installing the python
samples and modules at correct location. Remove the tweaks from local
patch and copy in the recipe itself.
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Use a submitted patch instead to point to correct llvm-tblgen and include
a fix to really link with all libs when LLVM libs are built statically.
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Remove patches that are present in this version of llvm and backport a
fix to prevent linking errors.
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
In order to include visa/ headers from igc, set
IGC_DIR path
Signed-off-by: Naveen Saini <naveen.kumar.saini@intel.com>
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Dropped the patch which is already upstream.
Signed-off-by: Naveen Saini <naveen.kumar.saini@intel.com>
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Fix the dependency to be added when ptests are not enabled. Fixes:
| ERROR: Nothing PROVIDES '0' (but /meta-intel/dynamic-layers/openembedded-layer/recipes-support/opencv/dldt-inference-engine_2019r3.1.bb DEPENDS on or otherwise requires it)
| ERROR: Required build target 'dldt-inference-engine' has no buildable providers.
| Missing or unbuildable dependency chain was: ['dldt-inference-engine', '0']
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
LIBCPLUSPLUS is now set to use GNU libstdc++ by default. A new variable
RUNTIME can be used to change this behaviour.
6895c79e05
Remove this bbappend as it is not required any more.
Signed-off-by: Naveen Saini <naveen.kumar.saini@intel.com>
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Install the OpenCL kernels and cldnn_global_custom_kernels.xml to allow
specification of OpenCL kernels for custom layers.
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
* Make sure that USB udev rules for Intel(R) Movidius(TM) Neural Compute
Stick and Intel(R) Neural Compute Stick 2 are packaged.
* Package vpu firmware only when it is enabled.
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
* There is interesting issue in do_patch, I was debugging strange
behavior with .bbappend where I've added another small patch.
And it started failing to configure completely.
bitbake -e shows that all .patch files are in SRC_URI and
log.do_patch shows that all were applied, but git diff (as well as
patches/series) shows only the last one added from the bbappend
to be applied.
This was caused by 8 existing patches in .bb file using ;patchdir=../
and my patch in .bbappend using ;patchdir=.. without slash at the end,
it should be fixed in quilt (or how do_patch is using it), but for
now just drop the trailing slash, because 99.9% recipes use ;patchdir=..
without the slash.
It's easily reproducible by removing the slash from the last patch
(without any bbappend).
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
