aboutsummaryrefslogtreecommitdiffstats
path: root/scripts
AgeCommit message (Collapse)Author
2017-04-05build-perf-git-import.py: handle Mexican timezones and localemarquiz/buildperf/git-importMarkus Lehtonen
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
2017-03-31build-perf-git-import.py: optimize buildstats by summing rusageMarkus Lehtonen
Store sum of rusage and child_rusage instead of separate values. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
2017-03-31build-perf-git-import.py: support conversion of xml reportsMarkus Lehtonen
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
2017-03-31build-perf-git-import.py: support importing arbitrarily named dirsMarkus Lehtonen
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
2017-03-31build-perf-git-import.py: implement --buildstats=c optionMarkus Lehtonen
Combine buildstats into one (optimized) JSON file. Commit this file under refs/notes/buildstats in order to save space. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
2017-03-31build-perf-git-import.py: implement --buildstats=o optionMarkus Lehtonen
Optimize buildstats for size. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
2017-03-31build-perf-git-import.py: implement --buildstats optionMarkus Lehtonen
Makes it possible to skip importing buildstats. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
2017-03-31build-perf-git-import.py: don't import bitbake log filesMarkus Lehtonen
In order to save space. Bitbake logs take a huge amount of space and are not generally interesting data. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
2017-03-31build-perf-git-import.py: support overriding metadataMarkus Lehtonen
Now the template always overrides original data. Also, mangling xml reports is now supported (even if conversion from xml to json is unimplemented). Change --metadata-template to --metadata-override which emphasizes that the values from the template now always replace values from the original results data. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
2017-03-31build-perf-git-import.py: implement --bareMarkus Lehtonen
For creating bare Git repositories. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
2017-03-31build-perf-git-import.py: improve commit messagesMarkus Lehtonen
Align with the (default) commit message used by oe-git-archive script (from yocto bug #10582) Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
2017-03-31build-perf-git-import.py: adapt JSON format to latest changesMarkus Lehtonen
Change JSON format to the match Yocto bugzilla #10590. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
2017-03-31build-perf-git-import.py: support conversion to xml formatMarkus Lehtonen
At the same time, split test metadata into a separate metadata JSON/XML file. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
2017-03-31build-perf-git-import.py: write a log fileMarkus Lehtonen
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
2017-03-31build-perf-git-import.py: hard code test descriptionsMarkus Lehtonen
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
2017-03-31build-perf-git-import.py: don't use globalres.logMarkus Lehtonen
Get rid of globalres.log and parse all data from the test log files. Formerly disk usage values were taken from globalres.log but now these are parsed from 'output.log'. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
2017-03-31Add scripts/contrib/build-perf-git-import.pyMarkus Lehtonen
Script for converting old archived build perf results in new JSON-based format and committing them in a Git repository. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
2017-03-30multilib_header: Update wrapper to handle arm 32/64 bitRichard Purdie
Having arm 32/64 bit headers coexisting turns out to be tricky. Unfortunately our wrapper works using wordsize.h and this differs on arm so we can't use it. Therefore replicate the logic here for arm. I did look into writing our own wordsize.h but we also need to remap kernel headers on arm and since wordsize.h comes from libc, that doesn't work for kernel headers. Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-30scripts/lib/compatlayer: detect_layers always use realpath'sAníbal Limón
If you are using relative paths and change to other folder for execution it will fail, so use realpaths always. [YOCTO #11164] Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-28sysroot-relativelinks: also consider links to dirs on the hostChristopher Larson
Dead symlinks, or symlinks to existing files will show up in 'files' of an os.walk, but symlinks to existing directories show up in 'dirs', so we need to consider both. As one example where this is an issue, the symlink from /usr/lib/ssl/certs was left pointing to /etc/ssl/certs rather than the relative path when the sdk was built on hosts where the latter exists. Signed-off-by: Christopher Larson <chris_larson@mentor.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-28update_gio_module_cache: fix host user contaminationAbdur Rehman
update_gio_module_cache intercept creates file: $D${libdir}/gio/modules/giomodule.cache Change ownership of this file to root:root to avoid user contamination by host. Signed-off-by: Abdur Rehman <abdur_rehman@mentor.com> Signed-off-by: Christopher Larson <chris_larson@mentor.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-28scripts/yocto-compat-layer.py: Add dependency validation in add_layerAníbal Limón
Some layers don't have dependencies so add a validation to avoid exception when trying to None.split(). Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-28devtool: open kconfig fragment in binary modeStephano Cetola
When devtool writes to the kconfig fragment, it writes the output of the diff command returned from pipe.communicate(). This function returns binary objects. We should open the kconfig fragment file in binary mode if we expect to write binary objects to it. [YOCTO #11171] Signed-off-by: Stephano Cetola <stephano.cetola@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-27wic: fix generation of partition UUIDEd Bartosh
Partition UUID for msdos partitioning is based on a 4 bytes long system identifier. Wic uses random number to generate system identifier. For the numbers starting with 0 partition uuid is shorter, which makes wic images non-bootable as kernel expects 8 charactest in PARTUUID parameter. Padded system identifier with '0' when generating partition UUID to make it always 8 characters long. This should fix the boot issue. Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-26scripts/wic: fix typoRobert Yang
roofs -> rootfs Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-26build-perf-test-wrapper.sh: store buildstats under git-notesMarkus Lehtonen
In order to dramatically reduce the data footprint of the result branches. This makes cloning/pulling result repositories a lot faster. Buildstats can still be accessed, when needed by doing git fetch refs/notes/buildstats/*:refs/notes/buildstats/* [YOCTO #10582] Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-26scripts/oe-git-archive: implement --notesMarkus Lehtonen
Option for adding git-notes to the commit. [YOCTO #10582] Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-26scripts/oe-git-archive: implement --excludeMarkus Lehtonen
May be used for excluding certain files from the commit. [YOCTO #10582] Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-26scripts/oe-git-archive: fix pushingMarkus Lehtonen
Git arguments were badly laid out. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-26oe-build-perf-test: pack all buildstat in one fileMarkus Lehtonen
Write out all buildstats into one big json file, instead of using multiple per-measurement files. Individual buildstats will be indexed using "<test_name>.<measurement_name>" as the key. Also, changes the per-testcase working directories into temporary directories that will be removed after test execution as there are no more per-testcase data files to store permanently. [YOCTO #10582] Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-24wic: use kernel_dir to find systemd-efi bootloaderEd Bartosh
Fixed test failure caused by using bootimg_dir path instead of kernel_dir to find systemd-efi bootloader: FAIL: test_systemd_bootdisk (oeqa.selftest.wic.Wic) ... FileNotFoundError: [Errno 2] No such file or directory: 'tmp/work/qemux86_64-poky-linux/core-image-minimal/1.0-r0/recipe-sysroot/usr/share' Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-23runqemu-gen-tapdevs: fix file path in exampleChen Qi
Fix the path in example so that '/usr/bin/tunctl' could be found under this path. Signed-off-by: Chen Qi <Qi.Chen@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
2017-03-22wic: improve getting syslinux pathEd Bartosh
Used wic-tools STAGING_DATADIR if syslinux can't be found in default bootimg_dir. Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-22wic: remove empty plugin methodsEd Bartosh
Removed do_install_disk and do_configure_partition methods from bootimage-partition, fsimage and rawcopy plugins as they're empy and not used. [YOCTO #10618] Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-22wic: use STAGING_DATADIR as a default for bootimg_dirEd Bartosh
Default value for bootimg_dir was not set in main wic script unlike the rest of artifacts: kernel_dir, roofts_dir and native_sysroot. Set it to the value of STAGING_DATADIR for consistency and to avoid confusion of wic plugin developers. Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-22wic: fix bug in sparse_copyEd Bartosh
sparse_copy creates output file that is smaller than input file when input file ends with unmapped blocks. Used truncate(<input file size>) when output file is created to ensure the size of output file will always be equal to the size of input file. Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-22wic: use kernel_dir instead of bootimg_dirEd Bartosh
bootimg_dir is usually set to the value of STAGING_DATADIR and kernel_dir - to the value of DEPLOY_DIR_IMAGE, so usage of kernel_dir is more logical in bootimg-efi, bootimg-partition and rawcopy plugins. Replaced usage of bootimg_dir to kernel_dir in 3 above mentioned plugins that use DEPLOY_DIR_IMAGE as a default artifact location. Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-22wic: don't set creator.bootimg_dir in pluginsEd Bartosh
Changing bootimg_dir in plugins is a dangerous hack as this value comes from wic command line and can be used by multiple plugins. Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-22wic: remove rootfs_pcbios_ext pluginEd Bartosh
Removed broken unused plugin. Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2017-03-21scripts/yocto-compat-layer.py: Handle layer dependencies when testAníbal Limón
If some layer depends on other tries to find layer dependency, if the layer dependency isn't found avoid to test the layer and notice the user. Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
2017-03-21scripts/yocto-compat-layer.py: Add option to disable layer autodiscoveryAníbal Limón
Sometimes there is a need to only analyze the layer specified by the command line, the new option -n will disable autodiscovery of layers and only will try to test specified layers. Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
2017-03-21scripts/lib/compatlayer: Remove require of meta- in layer dir nameAníbal Limón
The layers isn't required to have a dirctory name start with meta- so remove the validation. Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
2017-03-21scripts/compatlayer: Add exclude of meta-world-pkgdata on get_signaturesAníbal Limón
The meta-world-pkgdata recipe can be modified when a layer is added may be can add recipes to world target, so exclude by default. [YOCTO #11162] Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
2017-03-21scripts/yocto-compat-layer.py: Make output log argument optionalAníbal Limón
Only create a log file when --output-log option is specified, since logger is dumping to stdout by default is better to let the user decide if a log needs to be created. [YOCTO #11160] Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
2017-03-21scripts/yocto-compat-layer.py: Dump log to stdout instead of stderrAníbal Limón
The common unix tools uses stdout as standard for log output, by default python logging uses stderr if not stream is specified. [YOCTO #11160] Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
2017-03-21scriptutils: Add support for specify stream on logger_createAníbal Limón
It is a good idea to let the script to choose what stream wants to dump the logging output. [YOCTO #11160] Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
2017-03-21wic: partition.py: Add fsck to avoid corrupt EXT file systemsDaniel Schultz
This patch avoids the creation of a corrupt EXT file system. Since there are no checks if a EXT file system was successfully created, this should add to prevent possible system failures. Signed-off-by: Daniel Schultz <d.schultz@phytec.de> Signed-off-by: Ross Burton <ross.burton@intel.com>
2017-03-21gen-lockedsig-cache: catch os.link errorbrian avery
We do a hard link to speed up sdk creation but if your sstate-cache is across a file system boundary, this tries and fails. This patch catches that error and does a copy instead. Signed-off-by: brian avery <brian.avery@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
2017-03-21devtool: tidy up handling of parse failuresPaul Eggleton
Since the tinfoil2 refactoring, if an error occurred during parsing, we were showing a traceback and not correctly exiting (since we weren't calling shutdown()). Fix both of these issues. Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
2017-03-21wic/partionedfs: Avoid reserving space for non-existing ext. partitionKristian Amlie
We don't need the gap that the extended partition occupies if we already know that we have less than five partitions. Saves up to one full alignment of space. Signed-off-by: Kristian Amlie <kristian.amlie@mender.io> Signed-off-by: Ross Burton <ross.burton@intel.com>