summaryrefslogtreecommitdiffstats
path: root/openpower/scripts
Commit message (Collapse)AuthorAgeFilesLines
* Disable the following features to reduce initramfs sizeRaptor Engineering Development Team2020-02-191-0/+4
| | | | | | | | | * Nouveau driver (mostly fails on POWER w/ 64k pages anyway) * Non-English locales * Discrete HWDB files Once the -Os build failure has been resolved upstream, these can likely be reenabled.
* Do some extra cleanup in the skiroot.Brandon Bergren2019-04-241-0/+19
| | | | | | | | * Throw out elf libraries for nonbootable architectures * Throw out the c++ standard library (it's needed during build time for other pieces of the stack, but not the skiroot.) Signed-off-by: Brandon Bergren <git@bdragon.rtk0.net>
* Delete firmware whitelistRaptor Engineering Development Team2019-04-171-8/+1
| | | | No binary-only firmware components should be included in any Talos II firmware builds
* op-build v2.3-rc2 release notesStewart Smith2019-04-051-0/+1
| | | | Signed-off-by: Stewart Smith <stewart@linux.ibm.com>
* Add driver and firmware for QLogic devicesArtem Senichev2018-12-141-0/+1
| | | | | | Support for QLogic Corp. FastLinQ QL41000 Series 10/25/40/50GbE Controller. Signed-off-by: Artem Senichev <a.senichev@yadro.com>
* scripts/release-notes: Ignore witherspoon_dev for XMLSamuel Mendoza-Jonas2018-10-101-0/+1
| | | | | | | The XML for witherspoon_dev is the same as for the normal witherspoon platform so ignore it when cloning platform-xml repositories. Signed-off-by: Samuel Mendoza-Jonas <sam@mendozajonas.com>
* scripts/release-notes: Use markdown-flavoured line breaksSamuel Mendoza-Jonas2018-10-101-24/+24
| | | | | | | This way we can easily pass RELEASE.md through markdown without everything getting squashed into the same line. Signed-off-by: Samuel Mendoza-Jonas <sam@mendozajonas.com>
* firmware-whitelist: Drop e100 from whitelistJoel Stanley2018-08-221-1/+0
| | | | | | | | | We disabled the e100 driver in the kernel configuration back in op-build v1.22 with 3d17c64a1522 ("linux: config: Remove Intel E1000e and E100"). Remove the firmware from the whitelist as there's no chance anyone is using it. Signed-off-by: Joel Stanley <joel@jms.id.au>
* Drop untested graphic support from the buildJoel Stanley2018-08-221-6/+1
| | | | | | | | | | | | | | Back in the day Tuletta (P8) machines would use Radeon GPU drivers. No one has tested this in many moons. To save all systems from paying for the kernel driver, disable it in the kernel configuration and drop the firmware from the firenze, pseries and zz configurations. A similar story applies for Matrox cards. If someone comes hunting for support in the future they can revert this commit. Signed-off-by: Joel Stanley <joel@jms.id.au>
* release-notes: add hcode repositoryStewart Smith2018-05-091-0/+2
| | | | Signed-off-by: Stewart Smith <stewart@linux.ibm.com>
* openpower/scripts: Check (with travis) that defconfigs are correctStewart Smith2018-04-101-0/+18
| | | | Signed-off-by: Stewart Smith <stewart@linux.ibm.com>
* buildroot: Update to 2018.02Joel Stanley2018-04-101-1/+2
| | | | | | | | | | | | | | | | | | | | | | | | | Release notes and announcement: https://git.buildroot.net/buildroot/plain/CHANGES?id=2018.02 http://lists.busybox.net/pipermail/buildroot/2018-March/215002.html Updated packages: rsync: 3.1.2 -> 3.1.3 glibc: glibc-2.26-73 -> glibc-2.26-146 linux-firmware: a3a26af24e29 -> 65b1c68c63f9 util-linux: 2.31 -> 2.31.1 eudev: 3.2.4 -> 3.2.5 Updated host packages: libxml2: 2.9.5 -> 2.9.7 binutils: 2.28.1 -> 2.29.1 Signed-off-by: Joel Stanley <joel@jms.id.au> [stewart: remove old config option from pseries_defconfig, fix firmware-whitelist] Signed-off-by: Stewart Smith <stewart@linux.ibm.com>
* Merge pull request #1920 from sammj/release-notesStewart Smith2018-02-281-1/+1
|\ | | | | release-notes: Update env path
| * release-notes: Update env pathSamuel Mendoza-Jonas2018-02-281-1/+1
| | | | | | | | Signed-off-by: Samuel Mendoza-Jonas <sam@mendozajonas.com>
* | doc: initial sphinx-based documentation supportStewart Smith2018-02-261-9/+19
|/ | | | | | | | | | | | | | The release-notes script now produces Markdown compatible with pandoc to help convert it into RST for sphinx. Actual documentation is fairly limited currently. We use intersphinx to link to various sub-projects' docs. A .travis.yml file is added to build+publish documentation when new code is pushed to github. Signed-off-by: Stewart Smith <stewart@linux.vnet.ibm.com>
* Enable ql5000 firmware for BostonSamuel Mendoza-Jonas2017-09-271-0/+1
| | | | Signed-off-by: Samuel Mendoza-Jonas <sam@mendozajonas.com>
* release-notes: Update to account for single witherspoon platformSamuel Mendoza-Jonas2017-09-211-0/+17
| | | | | | | | | The release-notes script gets confused by the existence of both witherspoon_sequoia_defconfig and witherspoon_defconfig. In this scenario witherspoon_sequoia_defconfig only exists to placate the CI process, so treat it as if it does not exist. Signed-off-by: Samuel Mendoza-Jonas <sam@mendozajonas.com>
* release-notes: Clone project repos via HTTPSSamuel Mendoza-Jonas2017-08-171-8/+8
| | | | | | | | | | | The release-notes script currently specifies HTTP URLs for each project repository. These are all Github URLs and recently there's been a 50/50 chance of Github either a) redirecting to HTTPS or b) timing out. Update all URLs to use HTTPS so that the release-notes script can actually clone them. Signed-off-by: Samuel Mendoza-Jonas <sam@mendozajonas.com>
* release-notes: Don't try to clone witherspoon-foo-xmlSamuel Mendoza-Jonas2017-08-041-0/+9
| | | | | | | | | | | The proliferation of witherspoon flavours has confused the release-notes script again, causing it to attempt to clone the nonexistent witherspoon-{redbud,sequoia}-xml repo. Handle these extra variants to make sure we only clone the actual witherspoon-xml repository. Signed-off-by: Samuel Mendoza-Jonas <sam@mendozajonas.com>
* scripts/release-notes: half-treat sequoia as witherspoon for package changes.Stewart Smith2017-07-141-0/+10
| | | | | | | | | URGH. We kind of treat sequoia as witherspoon so we can show updated packages in release notes. Signed-off-by: Stewart Smith <stewart@linux.vnet.ibm.com>
* scripts/release-notes: quoting is funStewart Smith2017-07-141-2/+5
| | | | | | | On a similar level to .,' - table name quoting is odd and on the surface makes no sense. Signed-off-by: Stewart Smith <stewart@linux.vnet.ibm.com>
* scripts/release-notes: fixes around witherspoon-(sequoia|redbud)Stewart Smith2017-07-141-13/+25
| | | | | | | | Include some added/removed platform bug fixes, as well as fixes to do with a temporary bug in the op-build repo that was introduced and fixed 2 commits either side of the v1.17 tag Signed-off-by: Stewart Smith <stewart@linux.vnet.ibm.com>
* scripts/release-notes: Don't try to clone redbud-xmlSamuel Mendoza-Jonas2017-07-121-0/+1
| | | | | | | | The redbud platform uses the witherspoon-xml repository. There is no redbud-xml repository so ignore it when creating a list of -xml repositories. Signed-off-by: Samuel Mendoza-Jonas <sam@mendozajonas.com>
* scripts/release-notes: Update for machine-xml changes, diff package listStewart Smith2017-05-251-92/+220
| | | | | | | | | | | | | | | A bit of a decent rework to the release-notes tool, coping with the machine-xml rework (to be one package with different repos/versions per platform, see cd986e4cf84a526ee9653f1282e7889a9312f658) We also do a legal-info build for each platform, as the resulting manifest.csv allows us to work out what packages have been updated/added/removed for each platform. Also add the '--platform' option, to produce release notes for only one platform. Signed-off-by: Stewart Smith <stewart@linux.vnet.ibm.com>
* Add romulus to release-notes scriptStewart Smith2017-04-271-0/+3
| | | | Signed-off-by: Stewart Smith <stewart@linux.vnet.ibm.com>
* Update SBE code for P9Sachin Gupta2017-04-071-0/+3
|
* release-notes: Add support for Witherspoon and ZaiusPatrick Williams2017-01-161-0/+6
| | | | Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
* release-notes: Support P9 branch variable namesPatrick Williams2017-01-161-0/+11
| | | | Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
* Release Linux 4.9.2-openpower1Joel Stanley2017-01-151-1/+1
| | | | | | | | | | | | | | | | | | | This moves from the 4.4 based kernel series to the newer long term support kernel, Linux 4.9. To make the move we regenerate the kernel defconfig for the new kernel. We still have three out of tree patches. This commit makes the following changes: * Drops 5 of our out of tree patches * Two remain for XHCI kexec issues * One remains for VGA handover We also update bnx2x firmware version for newer kernel driver, which means modify the firmware whitelist. Signed-off-by: Joel Stanley <joel@jms.id.au>
* firmware-whitelist: Make whitelist accurateJoel Stanley2016-11-141-6/+7
| | | | | | | The whitelist was matching bnx2* as ok, so everything in bnx2x was being left in the image. Signed-off-by: Joel Stanley <joel@jms.id.au>
* openpower: Run firmware whitelist for all machinesJoel Stanley2016-11-011-0/+0
| | | | | | | | | | | | Firenze has a script that ensures only the correct firmware is present in the lib/firmware directory of skiroot. This is useful functionality for all machines, so rename it and add it to the ROOTFS_POST_BUILD_SCRIPT config option for all machines. Note that this means adding new drivers that have loadable firwmare in the future will require updating this script. Signed-off-by: Joel Stanley <joel@jms.id.au>
* firmware-whitelist: Whitelist explict bnx2x firmwareJoel Stanley2016-10-281-1/+1
| | | | | | | This is the one we need. Save a few hundred KB of space by not including them all. Signed-off-by: Joel Stanley <joel@jms.id.au>
* scripts: Drop support for CAICOS cardsSamuel Mendoza-Jonas2016-09-141-6/+1
| | | | | | | | The CAICOS series of radeon GPUs work on POWER but do not appear to survive kexec. Remove support for them from Skiroot so that the GPU won't be properly initialised until the host OS boots. Signed-off-by: Samuel Mendoza-Jonas <sam@mendozajonas.com>
* scripts: Include BTC binary for radeon cardsSamuel Mendoza-Jonas2016-08-101-1/+2
| | | | | | | | Some radeon chips require an additional _rlc.bin binary that doesn't match the name of its other required binary files. Include the BTC_rlc.bin binary as well for CAICOS chips. Signed-off-by: Samuel Mendoza-Jonas <sam@mendozajonas.com>
* Release notes tool.Patrick Williams2016-08-011-0/+180
| | | | Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
* platform-xml: DEPENDENCIES is a list of packages, not targetsJeremy Kerr2016-07-251-1/+0
| | | | | | | | | | | | | | | | | | | | | | | The current $target-xml packages use target names in their dependencies lists, where they should be using package names. This breaks certain targets, like 'legal-info', which create dependencies by adding other suffixes to the package names. From the buildroot manual: These dependencies are listed in terms of lower-case package names, which can be packages for the target (without the host- prefix) or packages for the host (with the host-) prefix). Buildroot will ensure that all these packages are built and installed before the current package starts its configuration. This change fixed these DEPENDENCIES defintions to use package names instead. Consequently, we don't need to filter out the -install-images target suffix in the op-target-dependencies script. Fixes #547. Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
* scripts: Add firenze-firmware-whitelist scriptSamuel Mendoza-Jonas2016-07-201-0/+46
| | | | | | | | | Firenze builds require several radeon binaries from linux-firmware, however the buildroot package installs support for many more varieties than we need. Add a post-build script that compares installed firmware binaries against a whitelist and removes any unnecessary binaries. Signed-off-by: Samuel Mendoza-Jonas <sam@mendozajonas.com>
* Remove dependency on realpath in post-build rootfs setup.Douglas Miller2015-10-051-4/+3
| | | | Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
* target: Don't symlink outside of /varJeremy Kerr2015-09-181-0/+12
| | | | | | | | | | We will want to share /var with petitboot plugins, but most of the directories in /var are actually symlinks to outside of var (to /tmp). This change replaces the symlinks with actual directories. Since we're running in an initramfs, we don't need to put everything in /tmp. Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
* Add tool to display targets and dependencies.Patrick Williams2015-03-011-0/+87
OpenPOWER on IntegriCloud