This patch series resolves runtime failures in GStreamer pipelines
that use TensorFlow Lite for object detection. The issue
was related to as undefined Abseil logging symbols when loading the
tensorflow2-lite
NNStreamer plugin.
The root cause was incomplete linking of Abseil logging components in the
TensorFlow Lite shared library build. While the library loaded successfully,
specific logging symbols (_ZN4absl12lts_2025012712log_internal9kCharNullE,
ZN4absl12lts_2025012712log_internal10LogMessageC1EPKciNS2_8ErrorTagE) were
missing at runtime.
The fix adds the missing Abseil logging targets (log_internal_nullguard,
log_internal_message, log_severity, raw_logging_internal) to both the CMake
module discovery and the tensorflow-lite target linking, ensuring all required
symbols are embedded in the shared library.
Pratham Deshmukh (1):
tensorflow-lite: fix undefined Abseil logging symbols in shared
library
...nsorFlow-Lite-pipeline-failures-due-.patch | 62 +++++++++++++++++++
.../tensorflow-lite/tensorflow-lite_2.20.0.bb | 1 +
2 files changed, 63 insertions(+)
create mode 100644
meta-arago-extras/recipes-framework/tensorflow-lite/tensorflow-lite/0003-Fix-GStreamer-TensorFlow-Lite-pipeline-failures-due-.patch
--
2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#17320):
https://lists.yoctoproject.org/g/meta-arago/message/17320
Mute This Topic: https://lists.yoctoproject.org/mt/118279632/21656
Group Owner: [email protected]
Unsubscribe: https://lists.yoctoproject.org/g/meta-arago/unsub
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-
