Age | Commit message (Collapse) | Author | Files | Lines |
|
The latest poky commit is requiring us to have all of
our template configs in a subdirectory instead of directly in
the `conf` directory. Without this we end up with errors during
setup like:
```
Error: TEMPLATECONF value (which is .../openbmc/meta-facebook/meta-bletchley/conf) must point to meta-some-layer/conf/templates/template-name
```
Fix this by moving all of our template files into the 'default'
template subdirectory (following the pattern of poky) and modifying
`setup` as necessary to follow.
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: Iecefde73d55acbb6bc63ae3d68c4311adaf327ae
|
|
- Pass column results to `expand` to eliminate tabs.
This enables easy copy/paste into markdown files for updating.
- Filter out QEMU machines.
We enable the QEMU machines from Yocto to be used for building but
they're not actively supported. Remove them to reduce the noise
in the tool output.
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I53c3f039cb36c705aa614627aeb40b20471c499c
|
|
The existing support for qemuarm wasn't working. Sometimes, when
submitting recipes to upstream, it is useful to test image builds
against various qemu* machine models. Enhance the setup script so that
we can build any of the qemu* machines that Poky supports and treat them
like a generic meta-phosphor machine.
Tested: Used `setup qemumips` and was able to `bitbake phosphor-image`
successfully.
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: Ide141c879dd674fe3771984f7fd41006a64ce54f
|
|
Change-Id: Id9d3d4ff314c0635f9f5ad6fc81d5f19977f65c1
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
|
|
If 'setup' is sourced with an invalid machine name, the error message is
now sent to stderr and the failure is propagated back up the stack with
a non-zero return value.
Change-Id: I50ad40b58c1479ce9ee80ea52536075eddd09b4c
Signed-off-by: Zev Weiss <zev@bewilderbeest.net>
|
|
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I4c09e61294506659500f9760993dd4b8b6f22f23
|
|
We already had logic to skip non-phosphor machines in the
'list' operation, but did not skip those machines when targeting
them. Those machines are almost certain to not operate as
expected and in some cases, such as EVB models, we have two
meta-layers associated with the machine. Skipping non-phosphor
machines ensures that the setup script picks the right
configuration and does not attempt to target a machine that won't
work.
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: If68c27e2a8675405672df497827dde6e86147108
|
|
If this script is sourced in a shell that has something like
`set -e`, then it fails if the column program does not exist
because the `which columm` fails. Make a few changes to
tolerate this and not fail:
- Use $(...) rather than `...`.
- Use `... || true` when appropriate to allow continuance.
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I9008778e8aee4747faa4decff74a2fb361a206ed
|
|
Modify setup script to support any level of machine conf depth.
Internally Google uses machines that sit yet another layer deep
in the repository. Instead of adding yet another level of depth
to the expansion, look for all configs using `find`, if it's
available. If it's not, ball back to the old wildcard expansion.
Signed-off-by: Maksym Sloyko <maxims@google.com>
Change-Id: I1066fe5f6ece4454e3555ca97f75bd91d6da1dba
|
|
The witherspoon config was moved directly into meta-ibm, which caused
the setup script to not be able to find it. Modify the setup script
to also find machine config files in single-deep meta directories.
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: Icb24649c2e90cd29f72e8b661085b85cde6fcad5
|
|
bitbake can struggle to cope with switching machine configurations in
the one build directory, so default to using machine-specific build
directories.
It's recommended that the following be added to your .profile
to provide a common location for bitbake data caches:
```
BB_ENV_EXTRAWHITE="SSTATE_DIR DL_DIR"
SSTATE_DIR="/var/cache/bitbake/sstate-cache"
DL_DIR="/var/cache/bitbake/downloads"
export BB_ENV_EXTRAWHITE SSTATE_DIR DL_DIR
```
Change-Id: I70abb6a6bc901c42266f92c9f8c230b793b7a0f9
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
|
|
Due to the import of meta-layers from out of openbmc, some of which have
their own machine configs, running `source setup` reveils a long list of
machines many of which are unsupported. Any supported machine should be
using the `meta-phosphor` layer in some way, so look for that in the
`bblayers.conf.sample` file.
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I3ee4aa33a643d5f7faf0c6e1bd6f3d9c84cddab8
|
|
zsh populates $0 even when sourcing a script. This caused sourcing
this script to think that it was being executed and it `exit 1`d
the running shell. Since no sane distribution would ever set /bin/sh
to zsh, the presence of $ZSH_NAME implies that the script is being
sourced.
Tested with both bash and zsh in both modes of operation.
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I2843d96a7fdaa46f9ee6edb9031ca9c16ca3bd98
|
|
The list of targets is getting longer and tends to not fit into
one screen when displayed in one column. Filter it through `column`
tool if available. If not available, then fall back to the old
representation.
Change-Id: I1e827d0835608d2e37ca58705e77a14b0667b170
Signed-off-by: Alexander Amelkin <a.amelkin@yadro.com>
|
|
Fixes: d70e7e81bfd44553f425f (setup: Use config name instead layer name)
Change-Id: Iee4a49f31bfbe137532f9be30277857d408cf7d5
Signed-off-by: Alexander Filippov <a.filippov@yadro.com>
|
|
Allows user to pass a config name instead of a layer name.
This can be extremely useful for layers, which contain several configs.
E.g. to build firmware for swift you can call:
```
. setup swift
bitbake obmc-phosphor-image
```
Change-Id: I397a53b0cc135f622e4b3dc712ff93f6a0c10b8a
Signed-off-by: Alexander Filippov <a.filippov@yadro.com>
|
|
This can be extremely useful if one is working on multiple
platforms simultaneously
Change-Id: I316ef106d972413221ab2691dfc727f8f9bcffa8
Signed-off-by: Amithash Prasad <amithash@fb.com>
Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>
|
|
The new subtree model brings the subtrees up from the openbmc-machines
layer.
Change-Id: I58a03ae1be374bc79ae1438e65e888375d12d0c0
Signed-off-by: Dave Cobbley <david.j.cobbley@linux.intel.com>
Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>
|
|
* Fix subshell spawning due to `sort`.
Subshell spawning effectively prevented the script from working.
Now `sort` is only invoked for `setup` without arguments;
* Add a license/copyright header
Change-Id: I643a91cae14e5ddb8e6efe154d325baaab3d0100
Signed-off-by: Alexander Amelkin <a.amelkin@yadro.com>
|
|
Before this commit it was required to specify a long
TEMPLATECONF string before running openbmc-env if one
wanted to build a non-qemu image.
This commit adds a simple setup script that can be
sourced like the following to setup the environment for
any machine (e.g., for palmetto):
. setup palmetto
Change-Id: Id3fa73dca9b8ce9347cf38e41fe8662c03ea84e3
Signed-off-by: Alexander Amelkin <a.amelkin@yadro.com>
|