libcamera
libcamera copied to clipboard
raspberrypi
compiler warnings introduced recently.
Suppressing with: -Wno-sign-compare
> [136/219] Compiling C++ object src/libcamera/libcamera.so.0.3.2.p/pipeline_rpi_pisp_pisp.cpp.o
> [137/219] Compiling C++ object src/ipa/rpi/cam_helper/librpi_ipa_cam_helper.a.p/cam_helper_imx708.cpp.o
> [138/219] Compiling C++ object src/ipa/rpi/controller/librpi_ipa_controller.a.p/algorithm.cpp.o
> [139/219] Compiling C++ object src/libcamera/libcamera.so.0.3.2.p/v4l2_subdevice.cpp.o
> [140/219] Compiling C++ object src/ipa/rpi/controller/librpi_ipa_controller.a.p/controller.cpp.o
> [141/219] Compiling C++ object src/ipa/rpi/controller/librpi_ipa_controller.a.p/rpi_black_level.cpp.o
> [142/219] Compiling C++ object src/ipa/rpi/controller/librpi_ipa_controller.a.p/rpi_cac.cpp.o
> [143/219] Compiling C++ object src/ipa/rpi/controller/librpi_ipa_controller.a.p/rpi_af.cpp.o
> [144/219] Compiling C++ object src/ipa/rpi/controller/librpi_ipa_controller.a.p/rpi_agc.cpp.o
> [145/219] Compiling C++ object src/ipa/rpi/controller/librpi_ipa_controller.a.p/rpi_alsc.cpp.o
> [146/219] Compiling C++ object src/ipa/rpi/cam_helper/librpi_ipa_cam_helper.a.p/imx500_tensor_parser_imx500_tensor_parser.cpp.o
> ../src/ipa/rpi/cam_helper/imx500_tensor_parser/imx500_tensor_parser.cpp: In function 'int {anonymous}::parseInputTensorBody(RPiController::IMX500InputTensorInfo&, const uint8_t*, const InputTensorApParams&, const DnnHeader&)':
> ../src/ipa/rpi/cam_helper/imx500_tensor_parser/imx500_tensor_parser.cpp:589:42: warning: comparison of integer expressions of different signedness: 'unsigned int' and 'int' [-Wsign-compare]
> 589 | if (((pixelIndex == inputApParams.width * inputApParams.height) ||
> | ~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> ../src/ipa/rpi/cam_helper/imx500_tensor_parser/imx500_tensor_parser.cpp:590:42: warning: comparison of integer expressions of different signedness: 'unsigned int' and 'int' [-Wsign-compare]
> 590 | (pixelIndex == inputApParams.width * inputApParams.height * 2) ||
> | ~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> ../src/ipa/rpi/cam_helper/imx500_tensor_parser/imx500_tensor_parser.cpp:591:42: warning: comparison of integer expressions of different signedness: 'unsigned int' and 'int' [-Wsign-compare]
> 591 | (pixelIndex == inputApParams.width * inputApParams.height * 3))) {
> | ~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> [147/219] Compiling C++ object src/ipa/rpi/controller/librpi_ipa_controller.a.p/rpi_awb.cpp.o
> [148/219] Compiling C++ object src/ipa/rpi/controller/librpi_ipa_controller.a.p/rpi_agc_channel.cpp.o
> [149/219] Compiling C++ object src/ipa/rpi/common/librpi_ipa_common.a.p/ipa_base.cpp.o
> ninja: build stopped: subcommand failed.
Suppressing with -Wno-stringop-truncation
> [133/219] Compiling C++ object src/ipa/rpi/controller/librpi_ipa_controller.a.p/device_status.cpp.o
> [134/219] Compiling C++ object src/ipa/rpi/cam_helper/librpi_ipa_cam_helper.a.p/cam_helper_imx500.cpp.o
> FAILED: src/ipa/rpi/cam_helper/librpi_ipa_cam_helper.a.p/cam_helper_imx500.cpp.o
> g++ -Isrc/ipa/rpi/cam_helper/librpi_ipa_cam_helper.a.p -Isrc/ipa/rpi/cam_helper -I../src/ipa/rpi/cam_helper -Isrc/ipa/rpi -I../src/ipa/rpi -Iinclude -I../include -Iinclude/libcamera/ipa -Iinclude/libcamera -fdiagnostics-color=always -D_GLIBCXX_ASSERTIONS=1 -D_FILE_OFFSET_BITS=64 -Wall -Winvalid-pch -Wextra -Werror -std=c++17 -O3 -Wno-redundant-move -Wmissing-declarations -Wshadow -include /build/source/build/config.h -Wno-sign-compare -fPIC -DLIBCAMERA_BASE_PRIVATE -MD -MQ src/ipa/rpi/cam_helper/librpi_ipa_cam_helper.a.p/cam_helper_imx500.cpp.o -MF src/ipa/rpi/cam_helper/librpi_ipa_cam_helper.a.p/cam_helper_imx500.cpp.o.d -o src/ipa/rpi/cam_helper/librpi_ipa_cam_helper.a.p/cam_helper_imx500.cpp.o -c ../src/ipa/rpi/cam_helper/cam_helper_imx500.cpp
> ../src/ipa/rpi/cam_helper/cam_helper_imx500.cpp: In member function 'void CamHelperImx500::parseInferenceData(libcamera::Span<const unsigned char>, RPiController::Metadata&)':
> ../src/ipa/rpi/cam_helper/cam_helper_imx500.cpp:259:40: error: 'char* __builtin_strncpy(char*, const char*, long unsigned int)' specified bound 64 equals destination size [-Werror=stringop-truncation]
> 259 | strncpy(exported.networkName, inputTensorInfo.networkName.c_str(),
> | ^
> cc1plus: all warnings being treated as errors
> [135/219] Compiling C++ object src/ipa/rpi/cam_helper/librpi_ipa_cam_helper.a.p/cam_helper_imx708.cpp.o
> [136/219] Compiling C++ object src/libcamera/libcamera.so.0.3.2.p/pipeline_rpi_pisp_pisp.cpp.o
> [137/219] Compiling C++ object src/ipa/rpi/controller/librpi_ipa_controller.a.p/algorithm.cpp.o
> [138/219] Compiling C++ object src/libcamera/libcamera.so.0.3.2.p/v4l2_subdevice.cpp.o
> [139/219] Compiling C++ object src/ipa/rpi/controller/librpi_ipa_controller.a.p/controller.cpp.o
> [140/219] Compiling C++ object src/ipa/rpi/controller/librpi_ipa_controller.a.p/rpi_black_level.cpp.o
> [141/219] Compiling C++ object src/ipa/rpi/controller/librpi_ipa_controller.a.p/rpi_cac.cpp.o
> [142/219] Compiling C++ object src/ipa/rpi/controller/librpi_ipa_controller.a.p/rpi_af.cpp.o
> [143/219] Compiling C++ object src/ipa/rpi/controller/librpi_ipa_controller.a.p/rpi_ccm.cpp.o
> [144/219] Compiling C++ object src/ipa/rpi/controller/librpi_ipa_controller.a.p/rpi_agc.cpp.o
> [145/219] Compiling C++ object src/ipa/rpi/cam_helper/librpi_ipa_cam_helper.a.p/imx500_tensor_parser_imx500_tensor_parser.cpp.o
> [146/219] Compiling C++ object src/ipa/rpi/controller/librpi_ipa_controller.a.p/rpi_alsc.cpp.o
> [147/219] Compiling C++ object src/ipa/rpi/controller/librpi_ipa_controller.a.p/rpi_awb.cpp.o
> [148/219] Compiling C++ object src/ipa/rpi/controller/librpi_ipa_controller.a.p/rpi_agc_channel.cpp.o
> [149/219] Compiling C++ object src/ipa/rpi/common/librpi_ipa_common.a.p/ipa_base.cpp.o
> ninja: build stopped: subcommand failed.
Rather than suppressing the warnings lets fix the actual potential issues.
As far as I know we don't suppress any warnings in the libcamera build. I also don't see these warnings in my local gcc/clang compilers. Can you tell me what compiler you are using?
gcc (GCC) 14.2.1 20241116
when I set: CXXFLAGS="-Wno-sign-compare -Wno-stringop-truncation"
Everything seems to be ok. strange that you don't see the warnings at least.
I can confirm I am seeing the problem on archlinuxarm (armv7, well a systemd-nspawn container if you want to know), with gcc 14.2.1 20250207:
[167/535] Compiling C++ object src/ipa/rpi/cam_helper/librpi_ipa_cam_helper.a.p/cam_helper_imx500.cpp.o
FAILED: src/ipa/rpi/cam_helper/librpi_ipa_cam_helper.a.p/cam_helper_imx500.cpp.o
c++ -Isrc/ipa/rpi/cam_helper/librpi_ipa_cam_helper.a.p -Isrc/ipa/rpi/cam_helper -I../libcamera/src/ipa/rpi/cam_helper -Isrc/ipa/rpi -I../libcamera/src/ipa/rpi -Iinclude -I../libcamera/include -Iinclude/libcamera/ipa -Iinclude/libcamera -fdiagnostics-color=always -D_GLIBCXX_ASSERTIONS=1 -D_FILE_OFFSET_BITS=64 -Wall -Winvalid-pch -Wextra -Werror -std=c++17 -Wno-redundant-move -Wno-psabi -Wmissing-declarations -Wshadow -include /home/alarm/pkg/PKGBUILDs/extra/libcamera/src/build/config.h -march=armv7-a -mfloat-abi=hard -mfpu=neon -O2 -pipe -fstack-protector-strong -fno-plt -fexceptions -Wp,-D_FORTIFY_SOURCE=3 -Wformat -Werror=format-security -fstack-clash-protection -fno-omit-frame-pointer -Wp,-D_GLIBCXX_ASSERTIONS -fPIC -DLIBCAMERA_BASE_PRIVATE -MD -MQ src/ipa/rpi/cam_helper/librpi_ipa_cam_helper.a.p/cam_helper_imx500.cpp.o -MF src/ipa/rpi/cam_helper/librpi_ipa_cam_helper.a.p/cam_helper_imx500.cpp.o.d -o src/ipa/rpi/cam_helper/librpi_ipa_cam_helper.a.p/cam_helper_imx500.cpp.o -c ../libcamera/src/ipa/rpi/cam_helper/cam_helper_imx500.cpp
distcc[161274] (dcc_build_somewhere) Warning: failed to distribute, running locally instead
../libcamera/src/ipa/rpi/cam_helper/cam_helper_imx500.cpp: Dans la fonction membre « void CamHelperImx500::parseInferenceData(libcamera::Span<const unsigned char>, RPiController::Metadata&) »:
../libcamera/src/ipa/rpi/cam_helper/cam_helper_imx500.cpp:248:40: erreur: « char* __builtin_strncpy(char*, const char*, unsigned int) » la limite spécifiée 64 est égale à la taille de la destination [-Werror=stringop-truncation]
248 | strncpy(exported.networkName, inputTensorInfo.networkName.c_str(),
| ^
cc1plus : tous les avertissements sont traités comme des erreurs
distcc[161274] ERROR: compile ../libcamera/src/ipa/rpi/cam_helper/cam_helper_imx500.cpp on localhost failed
[168/535] Compiling C++ object src/ipa/rpi/cam_helper/librpi_ipa_cam_helper.a.p/cam_helper_imx519.cpp.o
I had to use the CXXFLAGS given above
The following patch should fix the -Wno-stringop-truncation error
diff --git a/src/ipa/rpi/cam_helper/cam_helper_imx500.cpp b/src/ipa/rpi/cam_helper/cam_helper_imx500.cpp
index 9a5623f14ac5..6b12f3021afd 100644
--- a/src/ipa/rpi/cam_helper/cam_helper_imx500.cpp
+++ b/src/ipa/rpi/cam_helper/cam_helper_imx500.cpp
@@ -246,7 +246,7 @@ void CamHelperImx500::parseInferenceData(libcamera::Span<const uint8_t> buffer,
exported.height = inputTensorInfo.height;
exported.numChannels = inputTensorInfo.channels;
strncpy(exported.networkName, inputTensorInfo.networkName.c_str(),
- sizeof(exported.networkName));
+ sizeof(exported.networkName) - 1);
exported.networkName[sizeof(exported.networkName) - 1] = '\0';
metadata.set("cnn.input_tensor_info", exported);
metadata.set("cnn.input_tensor", std::move(inputTensorInfo.data));
@@ -279,7 +279,7 @@ void CamHelperImx500::parseInferenceData(libcamera::Span<const uint8_t> buffer,
<< "IMX500 output tensor info export failed, numTensors > MaxNumTensors";
}
strncpy(exported.networkName, outputTensorInfo.networkName.c_str(),
- sizeof(exported.networkName));
+ sizeof(exported.networkName) - 1);
exported.networkName[sizeof(exported.networkName) - 1] = '\0';
metadata.set("cnn.output_tensor_info", exported);
metadata.set("cnn.output_tensor", std::move(outputTensorInfo.data));
However, I'm not sure about what's going on with the -Wno-sign-compare error. As far as I can tell, both LHS and RHS are unsigned values, so everything should be ok. Perhaps the code has changed since the error was first reported?
@solsticedhiver can you test the above patch and report it it fixes your compiler warning?
Change has now been merged (https://github.com/raspberrypi/libcamera/commit/2993e7953e2f3ce374fd1d5a5adde15ce6918692) and will make its way into the next release.