Andrew Geissler | f034379 | 2020-11-18 10:42:21 -0600 | [diff] [blame] | 1 | .. SPDX-License-Identifier: CC-BY-SA-2.0-UK |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 2 | |
| 3 | ***************************** |
| 4 | QA Error and Warning Messages |
| 5 | ***************************** |
| 6 | |
| 7 | .. _qa-introduction: |
| 8 | |
| 9 | Introduction |
| 10 | ============ |
| 11 | |
| 12 | When building a recipe, the OpenEmbedded build system performs various |
| 13 | QA checks on the output to ensure that common issues are detected and |
| 14 | reported. Sometimes when you create a new recipe to build new software, |
| 15 | it will build with no problems. When this is not the case, or when you |
| 16 | have QA issues building any software, it could take a little time to |
| 17 | resolve them. |
| 18 | |
| 19 | While it is tempting to ignore a QA message or even to disable QA |
| 20 | checks, it is best to try and resolve any reported QA issues. This |
| 21 | chapter provides a list of the QA messages and brief explanations of the |
| 22 | issues you could encounter so that you can properly resolve problems. |
| 23 | |
| 24 | The next section provides a list of all QA error and warning messages |
| 25 | based on a default configuration. Each entry provides the message or |
| 26 | error form along with an explanation. |
| 27 | |
| 28 | .. note:: |
| 29 | |
| 30 | - At the end of each message, the name of the associated QA test (as |
| 31 | listed in the ":ref:`insane.bbclass <ref-classes-insane>`" |
| 32 | section) appears within square brackets. |
| 33 | |
| 34 | - As mentioned, this list of error and warning messages is for QA |
| 35 | checks only. The list does not cover all possible build errors or |
| 36 | warnings you could encounter. |
| 37 | |
| 38 | - Because some QA checks are disabled by default, this list does not |
| 39 | include all possible QA check errors and warnings. |
| 40 | |
| 41 | .. _qa-errors-and-warnings: |
| 42 | |
| 43 | Errors and Warnings |
| 44 | =================== |
| 45 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 46 | .. _qa-check-libexec: |
| 47 | |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 48 | - ``<packagename>: <path> is using libexec please relocate to <libexecdir> [libexec]`` |
| 49 | |
| 50 | The specified package contains files in ``/usr/libexec`` when the |
| 51 | distro configuration uses a different path for ``<libexecdir>`` By |
| 52 | default, ``<libexecdir>`` is ``$prefix/libexec``. However, this |
| 53 | default can be changed (e.g. ``${libdir}``). |
| 54 | |
| 55 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 56 | .. _qa-check-rpaths: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 57 | |
| 58 | - ``package <packagename> contains bad RPATH <rpath> in file <file> [rpaths]`` |
| 59 | |
| 60 | The specified binary produced by the recipe contains dynamic library |
| 61 | load paths (rpaths) that contain build system paths such as |
| 62 | :term:`TMPDIR`, which are incorrect for the target and |
| 63 | could potentially be a security issue. Check for bad ``-rpath`` |
| 64 | options being passed to the linker in your |
| 65 | :ref:`ref-tasks-compile` log. Depending on the build |
| 66 | system used by the software being built, there might be a configure |
| 67 | option to disable rpath usage completely within the build of the |
| 68 | software. |
| 69 | |
| 70 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 71 | .. _qa-check-useless-rpaths: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 72 | |
| 73 | - ``<packagename>: <file> contains probably-redundant RPATH <rpath> [useless-rpaths]`` |
| 74 | |
| 75 | The specified binary produced by the recipe contains dynamic library |
| 76 | load paths (rpaths) that on a standard system are searched by default |
| 77 | by the linker (e.g. ``/lib`` and ``/usr/lib``). While these paths |
| 78 | will not cause any breakage, they do waste space and are unnecessary. |
| 79 | Depending on the build system used by the software being built, there |
| 80 | might be a configure option to disable rpath usage completely within |
| 81 | the build of the software. |
| 82 | |
| 83 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 84 | .. _qa-check-file-rdeps: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 85 | |
| 86 | - ``<packagename> requires <files>, but no providers in its RDEPENDS [file-rdeps]`` |
| 87 | |
| 88 | A file-level dependency has been identified from the specified |
| 89 | package on the specified files, but there is no explicit |
| 90 | corresponding entry in :term:`RDEPENDS`. If |
| 91 | particular files are required at runtime then ``RDEPENDS`` should be |
| 92 | declared in the recipe to ensure the packages providing them are |
| 93 | built. |
| 94 | |
| 95 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 96 | .. _qa-check-build-deps: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 97 | |
| 98 | - ``<packagename1> rdepends on <packagename2>, but it isn't a build dependency? [build-deps]`` |
| 99 | |
| 100 | A runtime dependency exists between the two specified packages, but |
| 101 | there is nothing explicit within the recipe to enable the |
| 102 | OpenEmbedded build system to ensure that dependency is satisfied. |
| 103 | This condition is usually triggered by an |
| 104 | :term:`RDEPENDS` value being added at the packaging |
| 105 | stage rather than up front, which is usually automatic based on the |
| 106 | contents of the package. In most cases, you should change the recipe |
| 107 | to add an explicit ``RDEPENDS`` for the dependency. |
| 108 | |
| 109 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 110 | .. _qa-check-dev-so: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 111 | |
| 112 | - ``non -dev/-dbg/nativesdk- package contains symlink .so: <packagename> path '<path>' [dev-so]`` |
| 113 | |
| 114 | Symlink ``.so`` files are for development only, and should therefore |
| 115 | go into the ``-dev`` package. This situation might occur if you add |
| 116 | ``*.so*`` rather than ``*.so.*`` to a non-dev package. Change |
| 117 | :term:`FILES` (and possibly |
| 118 | :term:`PACKAGES`) such that the specified ``.so`` |
| 119 | file goes into an appropriate ``-dev`` package. |
| 120 | |
| 121 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 122 | .. _qa-check-staticdev: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 123 | |
| 124 | - ``non -staticdev package contains static .a library: <packagename> path '<path>' [staticdev]`` |
| 125 | |
| 126 | Static ``.a`` library files should go into a ``-staticdev`` package. |
| 127 | Change :term:`FILES` (and possibly |
| 128 | :term:`PACKAGES`) such that the specified ``.a`` file |
| 129 | goes into an appropriate ``-staticdev`` package. |
| 130 | |
| 131 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 132 | .. _qa-check-libdir: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 133 | |
| 134 | - ``<packagename>: found library in wrong location [libdir]`` |
| 135 | |
| 136 | The specified file may have been installed into an incorrect |
| 137 | (possibly hardcoded) installation path. For example, this test will |
| 138 | catch recipes that install ``/lib/bar.so`` when ``${base_libdir}`` is |
| 139 | "lib32". Another example is when recipes install |
| 140 | ``/usr/lib64/foo.so`` when ``${libdir}`` is "/usr/lib". False |
| 141 | positives occasionally exist. For these cases add "libdir" to |
| 142 | :term:`INSANE_SKIP` for the package. |
| 143 | |
| 144 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 145 | .. _qa-check-debug-files: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 146 | |
| 147 | - ``non debug package contains .debug directory: <packagename> path <path> [debug-files]`` |
| 148 | |
| 149 | The specified package contains a ``.debug`` directory, which should |
| 150 | not appear in anything but the ``-dbg`` package. This situation might |
| 151 | occur if you add a path which contains a ``.debug`` directory and do |
| 152 | not explicitly add the ``.debug`` directory to the ``-dbg`` package. |
| 153 | If this is the case, add the ``.debug`` directory explicitly to |
| 154 | ``FILES_${PN}-dbg``. See :term:`FILES` for additional |
| 155 | information on ``FILES``. |
| 156 | |
| 157 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 158 | .. _qa-check-arch: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 159 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 160 | - ``Architecture did not match (<file_arch>, expected <machine_arch>) in <file> [arch]`` |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 161 | |
| 162 | By default, the OpenEmbedded build system checks the Executable and |
| 163 | Linkable Format (ELF) type, bit size, and endianness of any binaries |
| 164 | to ensure they match the target architecture. This test fails if any |
| 165 | binaries do not match the type since there would be an |
| 166 | incompatibility. The test could indicate that the wrong compiler or |
| 167 | compiler options have been used. Sometimes software, like |
| 168 | bootloaders, might need to bypass this check. If the file you receive |
| 169 | the error for is firmware that is not intended to be executed within |
| 170 | the target operating system or is intended to run on a separate |
| 171 | processor within the device, you can add "arch" to |
| 172 | :term:`INSANE_SKIP` for the package. Another |
| 173 | option is to check the :ref:`ref-tasks-compile` log |
| 174 | and verify that the compiler options being used are correct. |
| 175 | |
| 176 | |
| 177 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 178 | - ``Bit size did not match (<file_bits>, expected <machine_bits>) in <file> [arch]`` |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 179 | |
| 180 | By default, the OpenEmbedded build system checks the Executable and |
| 181 | Linkable Format (ELF) type, bit size, and endianness of any binaries |
| 182 | to ensure they match the target architecture. This test fails if any |
| 183 | binaries do not match the type since there would be an |
| 184 | incompatibility. The test could indicate that the wrong compiler or |
| 185 | compiler options have been used. Sometimes software, like |
| 186 | bootloaders, might need to bypass this check. If the file you receive |
| 187 | the error for is firmware that is not intended to be executed within |
| 188 | the target operating system or is intended to run on a separate |
| 189 | processor within the device, you can add "arch" to |
| 190 | :term:`INSANE_SKIP` for the package. Another |
| 191 | option is to check the :ref:`ref-tasks-compile` log |
| 192 | and verify that the compiler options being used are correct. |
| 193 | |
| 194 | |
| 195 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 196 | - ``Endianness did not match (<file_endianness>, expected <machine_endianness>) in <file> [arch]`` |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 197 | |
| 198 | By default, the OpenEmbedded build system checks the Executable and |
| 199 | Linkable Format (ELF) type, bit size, and endianness of any binaries |
| 200 | to ensure they match the target architecture. This test fails if any |
| 201 | binaries do not match the type since there would be an |
| 202 | incompatibility. The test could indicate that the wrong compiler or |
| 203 | compiler options have been used. Sometimes software, like |
| 204 | bootloaders, might need to bypass this check. If the file you receive |
| 205 | the error for is firmware that is not intended to be executed within |
| 206 | the target operating system or is intended to run on a separate |
| 207 | processor within the device, you can add "arch" to |
| 208 | :term:`INSANE_SKIP` for the package. Another |
| 209 | option is to check the :ref:`ref-tasks-compile` log |
| 210 | and verify that the compiler options being used are correct. |
| 211 | |
| 212 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 213 | .. _qa-check-textrel: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 214 | |
| 215 | - ``ELF binary '<file>' has relocations in .text [textrel]`` |
| 216 | |
| 217 | The specified ELF binary contains relocations in its ``.text`` |
| 218 | sections. This situation can result in a performance impact at |
| 219 | runtime. |
| 220 | |
| 221 | Typically, the way to solve this performance issue is to add "-fPIC" |
| 222 | or "-fpic" to the compiler command-line options. For example, given |
| 223 | software that reads :term:`CFLAGS` when you build it, |
| 224 | you could add the following to your recipe: |
| 225 | :: |
| 226 | |
| 227 | CFLAGS_append = " -fPIC " |
| 228 | |
| 229 | For more information on text relocations at runtime, see |
Andrew Geissler | d1e8949 | 2021-02-12 15:35:20 -0600 | [diff] [blame] | 230 | https://www.akkadia.org/drepper/textrelocs.html. |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 231 | |
| 232 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 233 | .. _qa-check-ldflags: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 234 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 235 | - ``File '<file>' in package '<package>' doesn't have GNU_HASH (didn't pass LDFLAGS?) [ldflags]`` |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 236 | |
| 237 | This indicates that binaries produced when building the recipe have |
| 238 | not been linked with the :term:`LDFLAGS` options |
| 239 | provided by the build system. Check to be sure that the ``LDFLAGS`` |
| 240 | variable is being passed to the linker command. A common workaround |
| 241 | for this situation is to pass in ``LDFLAGS`` using |
| 242 | :term:`TARGET_CC_ARCH` within the recipe as |
| 243 | follows: |
| 244 | :: |
| 245 | |
| 246 | TARGET_CC_ARCH += "${LDFLAGS}" |
| 247 | |
| 248 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 249 | .. _qa-check-xorg-driver-abi: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 250 | |
| 251 | - ``Package <packagename> contains Xorg driver (<driver>) but no xorg-abi- dependencies [xorg-driver-abi]`` |
| 252 | |
| 253 | The specified package contains an Xorg driver, but does not have a |
| 254 | corresponding ABI package dependency. The xserver-xorg recipe |
| 255 | provides driver ABI names. All drivers should depend on the ABI |
| 256 | versions that they have been built against. Driver recipes that |
| 257 | include ``xorg-driver-input.inc`` or ``xorg-driver-video.inc`` will |
| 258 | automatically get these versions. Consequently, you should only need |
| 259 | to explicitly add dependencies to binary driver recipes. |
| 260 | |
| 261 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 262 | .. _qa-check-infodir: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 263 | |
| 264 | - ``The /usr/share/info/dir file is not meant to be shipped in a particular package. [infodir]`` |
| 265 | |
| 266 | The ``/usr/share/info/dir`` should not be packaged. Add the following |
| 267 | line to your :ref:`ref-tasks-install` task or to your |
| 268 | ``do_install_append`` within the recipe as follows: |
| 269 | :: |
| 270 | |
| 271 | rm ${D}${infodir}/dir |
| 272 | |
| 273 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 274 | .. _qa-check-symlink-to-sysroot: |
| 275 | |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 276 | - ``Symlink <path> in <packagename> points to TMPDIR [symlink-to-sysroot]`` |
| 277 | |
| 278 | The specified symlink points into :term:`TMPDIR` on the |
| 279 | host. Such symlinks will work on the host. However, they are clearly |
| 280 | invalid when running on the target. You should either correct the |
| 281 | symlink to use a relative path or remove the symlink. |
| 282 | |
| 283 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 284 | .. _qa-check-la: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 285 | |
| 286 | - ``<file> failed sanity test (workdir) in path <path> [la]`` |
| 287 | |
| 288 | The specified ``.la`` file contains :term:`TMPDIR` |
| 289 | paths. Any ``.la`` file containing these paths is incorrect since |
| 290 | ``libtool`` adds the correct sysroot prefix when using the files |
| 291 | automatically itself. |
| 292 | |
| 293 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 294 | .. _qa-check-pkgconfig: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 295 | |
| 296 | - ``<file> failed sanity test (tmpdir) in path <path> [pkgconfig]`` |
| 297 | |
| 298 | The specified ``.pc`` file contains |
| 299 | :term:`TMPDIR`\ ``/``\ :term:`WORKDIR` |
| 300 | paths. Any ``.pc`` file containing these paths is incorrect since |
| 301 | ``pkg-config`` itself adds the correct sysroot prefix when the files |
| 302 | are accessed. |
| 303 | |
| 304 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 305 | .. _qa-check-debug-deps: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 306 | |
| 307 | - ``<packagename> rdepends on <debug_packagename> [debug-deps]`` |
| 308 | |
| 309 | A dependency exists between the specified non-dbg package (i.e. a |
| 310 | package whose name does not end in ``-dbg``) and a package that is a |
| 311 | ``dbg`` package. The ``dbg`` packages contain debug symbols and are |
| 312 | brought in using several different methods: |
| 313 | |
| 314 | - Using the ``dbg-pkgs`` |
| 315 | :term:`IMAGE_FEATURES` value. |
| 316 | |
| 317 | - Using :term:`IMAGE_INSTALL`. |
| 318 | |
| 319 | - As a dependency of another ``dbg`` package that was brought in |
| 320 | using one of the above methods. |
| 321 | |
| 322 | The dependency might have been automatically added because the |
| 323 | ``dbg`` package erroneously contains files that it should not contain |
| 324 | (e.g. a non-symlink ``.so`` file) or it might have been added |
| 325 | manually (e.g. by adding to :term:`RDEPENDS`). |
| 326 | |
| 327 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 328 | .. _qa-check-dev-deps: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 329 | |
| 330 | - ``<packagename> rdepends on <dev_packagename> [dev-deps]`` |
| 331 | |
| 332 | A dependency exists between the specified non-dev package (a package |
| 333 | whose name does not end in ``-dev``) and a package that is a ``dev`` |
| 334 | package. The ``dev`` packages contain development headers and are |
| 335 | usually brought in using several different methods: |
| 336 | |
| 337 | - Using the ``dev-pkgs`` |
| 338 | :term:`IMAGE_FEATURES` value. |
| 339 | |
| 340 | - Using :term:`IMAGE_INSTALL`. |
| 341 | |
| 342 | - As a dependency of another ``dev`` package that was brought in |
| 343 | using one of the above methods. |
| 344 | |
| 345 | The dependency might have been automatically added (because the |
| 346 | ``dev`` package erroneously contains files that it should not have |
| 347 | (e.g. a non-symlink ``.so`` file) or it might have been added |
| 348 | manually (e.g. by adding to :term:`RDEPENDS`). |
| 349 | |
| 350 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 351 | .. _qa-check-dep-cmp: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 352 | |
| 353 | - ``<var>_<packagename> is invalid: <comparison> (<value>) only comparisons <, =, >, <=, and >= are allowed [dep-cmp]`` |
| 354 | |
| 355 | If you are adding a versioned dependency relationship to one of the |
| 356 | dependency variables (:term:`RDEPENDS`, |
| 357 | :term:`RRECOMMENDS`, |
| 358 | :term:`RSUGGESTS`, |
| 359 | :term:`RPROVIDES`, |
| 360 | :term:`RREPLACES`, or |
| 361 | :term:`RCONFLICTS`), you must only use the named |
| 362 | comparison operators. Change the versioned dependency values you are |
| 363 | adding to match those listed in the message. |
| 364 | |
| 365 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 366 | .. _qa-check-compile-host-path: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 367 | |
| 368 | - ``<recipename>: The compile log indicates that host include and/or library paths were used. Please check the log '<logfile>' for more information. [compile-host-path]`` |
| 369 | |
| 370 | The log for the :ref:`ref-tasks-compile` task |
| 371 | indicates that paths on the host were searched for files, which is |
| 372 | not appropriate when cross-compiling. Look for "is unsafe for |
| 373 | cross-compilation" or "CROSS COMPILE Badness" in the specified log |
| 374 | file. |
| 375 | |
| 376 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 377 | .. _qa-check-install-host-path: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 378 | |
| 379 | - ``<recipename>: The install log indicates that host include and/or library paths were used. Please check the log '<logfile>' for more information. [install-host-path]`` |
| 380 | |
| 381 | The log for the :ref:`ref-tasks-install` task |
| 382 | indicates that paths on the host were searched for files, which is |
| 383 | not appropriate when cross-compiling. Look for "is unsafe for |
| 384 | cross-compilation" or "CROSS COMPILE Badness" in the specified log |
| 385 | file. |
| 386 | |
| 387 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 388 | .. _qa-check-configure-unsafe: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 389 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 390 | - ``This autoconf log indicates errors, it looked at host include and/or library paths while determining system capabilities. Rerun configure task after fixing this. [configure-unsafe]`` |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 391 | |
| 392 | The log for the :ref:`ref-tasks-configure` task |
| 393 | indicates that paths on the host were searched for files, which is |
| 394 | not appropriate when cross-compiling. Look for "is unsafe for |
| 395 | cross-compilation" or "CROSS COMPILE Badness" in the specified log |
| 396 | file. |
| 397 | |
| 398 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 399 | .. _qa-check-pkgname: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 400 | |
| 401 | - ``<packagename> doesn't match the [a-z0-9.+-]+ regex [pkgname]`` |
| 402 | |
| 403 | The convention within the OpenEmbedded build system (sometimes |
| 404 | enforced by the package manager itself) is to require that package |
| 405 | names are all lower case and to allow a restricted set of characters. |
| 406 | If your recipe name does not match this, or you add packages to |
| 407 | :term:`PACKAGES` that do not conform to the |
| 408 | convention, then you will receive this error. Rename your recipe. Or, |
| 409 | if you have added a non-conforming package name to ``PACKAGES``, |
| 410 | change the package name appropriately. |
| 411 | |
| 412 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 413 | .. _qa-check-unknown-configure-option: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 414 | |
| 415 | - ``<recipe>: configure was passed unrecognized options: <options> [unknown-configure-option]`` |
| 416 | |
| 417 | The configure script is reporting that the specified options are |
| 418 | unrecognized. This situation could be because the options were |
| 419 | previously valid but have been removed from the configure script. Or, |
| 420 | there was a mistake when the options were added and there is another |
| 421 | option that should be used instead. If you are unsure, consult the |
| 422 | upstream build documentation, the ``./configure --help`` output, and |
| 423 | the upstream change log or release notes. Once you have worked out |
| 424 | what the appropriate change is, you can update |
| 425 | :term:`EXTRA_OECONF`, |
| 426 | :term:`PACKAGECONFIG_CONFARGS`, or the |
| 427 | individual :term:`PACKAGECONFIG` option values |
| 428 | accordingly. |
| 429 | |
| 430 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 431 | .. _qa-check-pn-overrides: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 432 | |
| 433 | - ``Recipe <recipefile> has PN of "<recipename>" which is in OVERRIDES, this can result in unexpected behavior. [pn-overrides]`` |
| 434 | |
| 435 | The specified recipe has a name (:term:`PN`) value that |
| 436 | appears in :term:`OVERRIDES`. If a recipe is named |
| 437 | such that its ``PN`` value matches something already in ``OVERRIDES`` |
| 438 | (e.g. ``PN`` happens to be the same as :term:`MACHINE` |
| 439 | or :term:`DISTRO`), it can have unexpected |
| 440 | consequences. For example, assignments such as |
| 441 | ``FILES_${PN} = "xyz"`` effectively turn into ``FILES = "xyz"``. |
| 442 | Rename your recipe (or if ``PN`` is being set explicitly, change the |
| 443 | ``PN`` value) so that the conflict does not occur. See |
| 444 | :term:`FILES` for additional information. |
| 445 | |
| 446 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 447 | .. _qa-check-pkgvarcheck: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 448 | |
| 449 | - ``<recipefile>: Variable <variable> is set as not being package specific, please fix this. [pkgvarcheck]`` |
| 450 | |
| 451 | Certain variables (:term:`RDEPENDS`, |
| 452 | :term:`RRECOMMENDS`, |
| 453 | :term:`RSUGGESTS`, |
| 454 | :term:`RCONFLICTS`, |
| 455 | :term:`RPROVIDES`, |
| 456 | :term:`RREPLACES`, :term:`FILES`, |
| 457 | ``pkg_preinst``, ``pkg_postinst``, ``pkg_prerm``, ``pkg_postrm``, and |
| 458 | :term:`ALLOW_EMPTY`) should always be set specific |
| 459 | to a package (i.e. they should be set with a package name override |
| 460 | such as ``RDEPENDS_${PN} = "value"`` rather than |
| 461 | ``RDEPENDS = "value"``). If you receive this error, correct any |
| 462 | assignments to these variables within your recipe. |
| 463 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 464 | |
| 465 | - ``recipe uses DEPENDS_${PN}, should use DEPENDS [pkgvarcheck]`` |
| 466 | |
| 467 | This check looks for instances of setting ``DEPENDS_${PN}`` |
| 468 | which is erroneous (:term:`DEPENDS` is a recipe-wide variable and thus |
| 469 | it is not correct to specify it for a particular package, nor will such |
| 470 | an assignment actually work.) Set ``DEPENDS`` instead. |
| 471 | |
| 472 | |
| 473 | .. _qa-check-already-stripped: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 474 | |
| 475 | - ``File '<file>' from <recipename> was already stripped, this will prevent future debugging! [already-stripped]`` |
| 476 | |
| 477 | Produced binaries have already been stripped prior to the build |
| 478 | system extracting debug symbols. It is common for upstream software |
| 479 | projects to default to stripping debug symbols for output binaries. |
| 480 | In order for debugging to work on the target using ``-dbg`` packages, |
| 481 | this stripping must be disabled. |
| 482 | |
| 483 | Depending on the build system used by the software being built, |
| 484 | disabling this stripping could be as easy as specifying an additional |
| 485 | configure option. If not, disabling stripping might involve patching |
| 486 | the build scripts. In the latter case, look for references to "strip" |
| 487 | or "STRIP", or the "-s" or "-S" command-line options being specified |
| 488 | on the linker command line (possibly through the compiler command |
| 489 | line if preceded with "-Wl,"). |
| 490 | |
| 491 | .. note:: |
| 492 | |
| 493 | Disabling stripping here does not mean that the final packaged |
| 494 | binaries will be unstripped. Once the OpenEmbedded build system |
Andrew Geissler | 4c19ea1 | 2020-10-27 13:52:24 -0500 | [diff] [blame] | 495 | splits out debug symbols to the ``-dbg`` package, it will then |
| 496 | strip the symbols from the binaries. |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 497 | |
| 498 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 499 | .. _qa-check-packages-list: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 500 | |
| 501 | - ``<packagename> is listed in PACKAGES multiple times, this leads to packaging errors. [packages-list]`` |
| 502 | |
| 503 | Package names must appear only once in the |
| 504 | :term:`PACKAGES` variable. You might receive this |
| 505 | error if you are attempting to add a package to ``PACKAGES`` that is |
| 506 | already in the variable's value. |
| 507 | |
| 508 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 509 | .. _qa-check-files-invalid: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 510 | |
| 511 | - ``FILES variable for package <packagename> contains '//' which is invalid. Attempting to fix this but you should correct the metadata. [files-invalid]`` |
| 512 | |
| 513 | The string "//" is invalid in a Unix path. Correct all occurrences |
| 514 | where this string appears in a :term:`FILES` variable so |
| 515 | that there is only a single "/". |
| 516 | |
| 517 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 518 | .. _qa-check-installed-vs-shipped: |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 519 | |
| 520 | - ``<recipename>: Files/directories were installed but not shipped in any package [installed-vs-shipped]`` |
| 521 | |
| 522 | Files have been installed within the |
| 523 | :ref:`ref-tasks-install` task but have not been |
| 524 | included in any package by way of the :term:`FILES` |
| 525 | variable. Files that do not appear in any package cannot be present |
| 526 | in an image later on in the build process. You need to do one of the |
| 527 | following: |
| 528 | |
| 529 | - Add the files to ``FILES`` for the package you want them to appear |
| 530 | in (e.g. ``FILES_${``\ :term:`PN`\ ``}`` for the main |
| 531 | package). |
| 532 | |
| 533 | - Delete the files at the end of the ``do_install`` task if the |
| 534 | files are not needed in any package. |
| 535 | |
| 536 | |
| 537 | |
| 538 | - ``<oldpackage>-<oldpkgversion> was registered as shlib provider for <library>, changing it to <newpackage>-<newpkgversion> because it was built later`` |
| 539 | |
| 540 | This message means that both ``<oldpackage>`` and ``<newpackage>`` |
| 541 | provide the specified shared library. You can expect this message |
| 542 | when a recipe has been renamed. However, if that is not the case, the |
| 543 | message might indicate that a private version of a library is being |
| 544 | erroneously picked up as the provider for a common library. If that |
| 545 | is the case, you should add the library's ``.so`` file name to |
| 546 | :term:`PRIVATE_LIBS` in the recipe that provides |
| 547 | the private version of the library. |
| 548 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 549 | |
| 550 | .. _qa-check-unlisted-pkg-lics: |
| 551 | |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 552 | - ``LICENSE_<packagename> includes licenses (<licenses>) that are not listed in LICENSE [unlisted-pkg-lics]`` |
| 553 | |
| 554 | The :term:`LICENSE` of the recipe should be a superset |
| 555 | of all the licenses of all packages produced by this recipe. In other |
| 556 | words, any license in ``LICENSE_*`` should also appear in |
| 557 | :term:`LICENSE`. |
| 558 | |
Andrew Geissler | 6ce62a2 | 2020-11-30 19:58:47 -0600 | [diff] [blame] | 559 | |
| 560 | .. _qa-check-configure-gettext: |
| 561 | |
| 562 | - ``AM_GNU_GETTEXT used but no inherit gettext [configure-gettext]`` |
| 563 | |
| 564 | If a recipe is building something that uses automake and the automake |
| 565 | files contain an ``AM_GNU_GETTEXT`` directive then this check will fail |
| 566 | if there is no ``inherit gettext`` statement in the recipe to ensure |
| 567 | that gettext is available during the build. Add ``inherit gettext`` to |
| 568 | remove the warning. |
| 569 | |
| 570 | |
| 571 | .. _qa-check-mime: |
| 572 | |
| 573 | - ``package contains mime types but does not inherit mime: <packagename> path '<file>' [mime]`` |
| 574 | |
| 575 | The specified package contains mime type files (``.xml`` files in |
| 576 | ``${datadir}/mime/packages``) and yet does not inherit the mime |
| 577 | class which will ensure that these get properly installed. Either |
| 578 | add ``inherit mime`` to the recipe or remove the files at the |
| 579 | ``do_install`` step if they are not needed. |
| 580 | |
| 581 | |
| 582 | .. _qa-check-mime-xdg: |
| 583 | |
| 584 | - ``package contains desktop file with key 'MimeType' but does not inhert mime-xdg: <packagename> path '<file>' [mime-xdg]`` |
| 585 | |
| 586 | The specified package contains a .desktop file with a 'MimeType' key |
| 587 | present, but does not inherit the mime-xdg class that is required in |
| 588 | order for that to be activated. Either add ``inherit mime`` to the |
| 589 | recipe or remove the files at the ``do_install`` step if they are not |
| 590 | needed. |
| 591 | |
| 592 | |
| 593 | .. _qa-check-src-uri-bad: |
| 594 | |
| 595 | - ``<recipename>: SRC_URI uses unstable GitHub archives [src-uri-bad]`` |
| 596 | |
| 597 | GitHub provides "archive" tarballs, however these can be re-generated |
| 598 | on the fly and thus the file's signature will not necessarily match that |
| 599 | in the SRC_URI checksums in future leading to build failures. It is |
| 600 | recommended that you use an official release tarball or switch to |
| 601 | pulling the corresponding revision in the actual git repository instead. |
| 602 | |
| 603 | |
| 604 | - ``SRC_URI uses PN not BPN [src-uri-bad]`` |
| 605 | |
| 606 | If some part of :term:`SRC_URI` needs to reference the recipe name, it should do |
| 607 | so using ${:term:`BPN`} rather than ${:term:`PN`} as the latter will change |
| 608 | for different variants of the same recipe e.g. when :term:`BBCLASSEXTEND` |
| 609 | or multilib are being used. This check will fail if a reference to ``${PN}`` |
| 610 | is found within the ``SRC_URI`` value - change it to ``${BPN}`` instead. |
| 611 | |
| 612 | |
| 613 | .. _qa-check-unhandled-features-check: |
| 614 | |
| 615 | - ``<recipename>: recipe doesn't inherit features_check [unhandled-features-check]`` |
| 616 | |
| 617 | This check ensures that if one of the variables that the :ref:`features_check <ref-classes-features_check>` |
| 618 | class supports (e.g. :term:`REQUIRED_DISTRO_FEATURES`) is used, then the recipe |
| 619 | inherits ``features_check`` in order for the requirement to actually work. If |
| 620 | you are seeing this message, either add ``inherit features_check`` to your recipe |
| 621 | or remove the reference to the variable if it is not needed. |
| 622 | |
| 623 | |
| 624 | .. _qa-check-missing-update-alternatives: |
| 625 | |
| 626 | - ``<recipename>: recipe defines ALTERNATIVE_<packagename> but doesn't inherit update-alternatives. This might fail during do_rootfs later! [missing-update-alternatives]`` |
| 627 | |
| 628 | This check ensures that if a recipe sets the :term:`ALTERNATIVE` variable that the |
| 629 | recipe also inherits :ref:`update-alternatives <ref-classes-update-alternatives>` such |
| 630 | that the alternative will be correctly set up. If you are seeing this message, either |
| 631 | add ``inherit update-alternatives`` to your recipe or remove the reference to the variable |
| 632 | if it is not needed. |
| 633 | |
| 634 | |
| 635 | .. _qa-check-shebang-size: |
| 636 | |
| 637 | - ``<packagename>: <file> maximum shebang size exceeded, the maximum size is 128. [shebang-size]`` |
| 638 | |
| 639 | This check ensures that the shebang line (``#!`` in the first line) for a script |
| 640 | is not longer than 128 characters, which can cause an error at runtime depending |
| 641 | on the operating system. If you are seeing this message then the specified script |
| 642 | may need to be patched to have a shorter in order to avoid runtime problems. |
| 643 | |
| 644 | |
| 645 | .. _qa-check-perllocalpod: |
| 646 | |
| 647 | - ``<packagename> contains perllocal.pod (<files>), should not be installed [perllocalpod]`` |
| 648 | |
| 649 | ``perllocal.pod`` is an index file of locally installed modules and so shouldn't be |
| 650 | installed by any distribution packages. The :ref:`cpan <ref-classes-cpan>` class |
| 651 | already sets ``NO_PERLLOCAL`` to stop this file being generated by most Perl recipes, |
| 652 | but if a recipe is using ``MakeMaker`` directly then they might not be doing this |
| 653 | correctly. This check ensures that perllocal.pod is not in any package in order to |
| 654 | avoid multiple packages shipping this file and thus their packages conflicting |
| 655 | if installed together. |
| 656 | |
| 657 | |
| 658 | .. _qa-check-usrmerge: |
| 659 | |
| 660 | - ``<packagename> package is not obeying usrmerge distro feature. /<path> should be relocated to /usr. [usrmerge]`` |
| 661 | |
| 662 | If ``usrmerge`` is in :term:`DISTRO_FEATURES`, this check will ensure that no package |
| 663 | installs files to root (``/bin``, ``/sbin``, ``/lib``, ``/lib64``) directories. If you are seeing this |
| 664 | message, it indicates that the ``do_install`` step (or perhaps the build process that |
| 665 | ``do_install`` is calling into, e.g. ``make install`` is using hardcoded paths instead |
| 666 | of the variables set up for this (``bindir``, ``sbindir``, etc.), and should be |
| 667 | changed so that it does. |
| 668 | |
| 669 | |
| 670 | .. _qa-check-patch-fuzz: |
| 671 | |
| 672 | - ``Fuzz detected: <patch output> [patch-fuzz]`` |
| 673 | |
| 674 | This check looks for evidence of "fuzz" when applying patches within the ``do_patch`` |
| 675 | task. Patch fuzz is a situation when the ``patch`` tool ignores some of the context |
| 676 | lines in order to apply the patch. Consider this example: |
| 677 | |
| 678 | Patch to be applied: :: |
| 679 | |
| 680 | --- filename |
| 681 | +++ filename |
| 682 | context line 1 |
| 683 | context line 2 |
| 684 | context line 3 |
| 685 | +newly added line |
| 686 | context line 4 |
| 687 | context line 5 |
| 688 | context line 6 |
| 689 | |
| 690 | Original source code: :: |
| 691 | |
| 692 | different context line 1 |
| 693 | different context line 2 |
| 694 | context line 3 |
| 695 | context line 4 |
| 696 | different context line 5 |
| 697 | different context line 6 |
| 698 | |
| 699 | Outcome (after applying patch with fuzz): :: |
| 700 | |
| 701 | different context line 1 |
| 702 | different context line 2 |
| 703 | context line 3 |
| 704 | newly added line |
| 705 | context line 4 |
| 706 | different context line 5 |
| 707 | different context line 6 |
| 708 | |
| 709 | Chances are, the newly added line was actually added in a completely |
| 710 | wrong location, or it was already in the original source and was added |
| 711 | for the second time. This is especially possible if the context line 3 |
| 712 | and 4 are blank or have only generic things in them, such as ``#endif`` or ``}``. |
| 713 | Depending on the patched code, it is entirely possible for an incorrectly |
| 714 | patched file to still compile without errors. |
| 715 | |
| 716 | *How to eliminate patch fuzz warnings* |
| 717 | |
| 718 | Use the ``devtool`` command as explained by the warning. First, unpack the |
| 719 | source into devtool workspace: :: |
| 720 | |
| 721 | devtool modify <recipe> |
| 722 | |
| 723 | This will apply all of the patches, and create new commits out of them in |
| 724 | the workspace - with the patch context updated. |
| 725 | |
| 726 | Then, replace the patches in the recipe layer: :: |
| 727 | |
| 728 | devtool finish --force-patch-refresh <recipe> <layer_path> |
| 729 | |
| 730 | The patch updates then need be reviewed (preferably with a side-by-side diff |
| 731 | tool) to ensure they are indeed doing the right thing i.e.: |
| 732 | |
| 733 | #. they are applied in the correct location within the file; |
| 734 | #. they do not introduce duplicate lines, or otherwise do things that |
| 735 | are no longer necessary. |
| 736 | |
| 737 | To confirm these things, you can also review the patched source code in |
| 738 | devtool's workspace, typically in ``<build_dir>/workspace/sources/<recipe>/`` |
| 739 | |
| 740 | Once the review is done, you can create and publish a layer commit with |
| 741 | the patch updates that modify the context. Devtool may also refresh |
| 742 | other things in the patches, those can be discarded. |
| 743 | |
| 744 | |
Andrew Geissler | c9f7865 | 2020-09-18 14:11:35 -0500 | [diff] [blame] | 745 | |
| 746 | Configuring and Disabling QA Checks |
| 747 | =================================== |
| 748 | |
| 749 | You can configure the QA checks globally so that specific check failures |
| 750 | either raise a warning or an error message, using the |
| 751 | :term:`WARN_QA` and :term:`ERROR_QA` |
| 752 | variables, respectively. You can also disable checks within a particular |
| 753 | recipe using :term:`INSANE_SKIP`. For information on |
| 754 | how to work with the QA checks, see the |
| 755 | ":ref:`insane.bbclass <ref-classes-insane>`" section. |
| 756 | |
| 757 | .. note:: |
| 758 | |
| 759 | Please keep in mind that the QA checks exist in order to detect real |
| 760 | or potential problems in the packaged output. So exercise caution |
| 761 | when disabling these checks. |