Age | Commit message (Collapse) | Author | Files |
|
Used truncate instead of dd to create wic images for the
following reasons:
- dd doesn't preserve sparseness
- truncate syntax is much more clear
- dd requires additional calculations of the image size
in blocks
- the way dd was used in the code is not always correct.
In some cases it was writing one block to the file which makes
it not 100% sparse.
[YOCTO #9099]
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Added explanation of --system-id option to the output of
wic help kickstart.
[YOCTO #9096]
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Used sfdisk to set partition system id if --system-id parameter
is used for a partition in wks file.
[YOCTO #9096]
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Added new argument to add_partition call to pass partition
system id down the stack.
[YOCTO #9096]
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Added Partition.system_id attribute and initialized it
from parse result of wks option --system-id. It will be
used by the wic code below the call stack to set partition
system id.
[YOCTO #9096]
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Added sfdisk -> util-linux pair to the dictionary
executable -> recipe as sfdisk is going to be used by wic
to set partition system id.
[YOCTO #9096]
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Added new option --system-id to wks parser. The option
will be used to set partition system id.
[YOCTO #9096]
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
os.path.join discards the cr_workdir var contents if the path of the
second arguments is absolute.
Signed-off-by: Ioan-Adrian Ratiu <adrian.ratiu@ni.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
The latest wic kickstart refactoring introduced a bootloader option
"--configfile" which lets wks' specify a custom grub.cfg for use
while booting. This is very useful for creating stuff like boot menus.
This change lets isoimage-isohybrid use --configfile; if this option is
not specified in a wks, it generates a default cfg as before.
Signed-off-by: Ioan-Adrian Ratiu <adrian.ratiu@ni.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
If UNINATIVE_CHECKSUM changes over an SDK update, bitbake within the
extensible SDK will be broken because it will see that the matching
uninative tarball doesn't exist and if there is a default value of
UNINATIVE_URL it will attempt to download the file and will then fail
because the checksums don't match up; alternatively if no UNINATIVE_URL
is set then it'll also fail with an error about misconfiguration. To fix
this, add some logic to devtool sdk-update to download the matching
uninative tarball(s) for the checksum(s) in the newly fetched SDK
configuration.
Fixes [YOCTO #9301].
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Replace git pull with fetch and reset to avoid the merge logic in the
event that the layers repo in the published SDK we're updating to isn't
fast-forward merge from the local repo.
Also add gitignore and committer info during publish to avoid errors and
to be sure that the first commit has a dummy user in it.
[ YOCTO #9368 ]
Signed-off-by: Stephano Cetola <stephano.cetola@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
It is unusual but not impossible to find recipes whose first entry is
not the main source URL but instead some patch or other local file, for
example python-cryptography in meta-python (which sets SRC_URI before
inheriting pypi). There's nothing inherently wrong with this, and we
shouldn't assume that the first entry is the main source URL, so just
take the first non-local entry instead.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
For example, support both:
$ runqemu qemux86-64 ramfs
$ runqemu qemux86-64 cpio.gz (new)
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Partitions specifying --fsoptions were silently skipped by wic
due to the old bug introduced when removing code related to
subvolume handling:
- if mountpoint == "/" or not fsopts or fsopts.find("subvol=") == -1
+ if mountpoint == "/" or not fsopts:
[YOCTO #9396]
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
It turns out that dd's conv=sparse doesn't look at the file extents, but simply
checks if a "block" is all zero. If the block of zero was meaningful it gets
lost and if the image is subsequently written to media using a sparse-aware
writer then the block of zeros won't be written at all.
This reverts commit 5fd592fbae2e046bcb8c3a6c3ef4993fe0400676.
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The individual partitions created by wic are sparse but without
this change the assembled image is written as one (potentially
very) large file.
Preserve sparseness in the assembled image by passing the sparse
conversion symbol.
[YOCTO #9099]
Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
When a recipe is added to the workspace, the signatures for the tasks
will change. This means that bitbake must be told to allow the
signatures to be different if they are in locked-sigs.inc.
This is done by creating an unlocked-sigs.inc file which contains all
the recipes in the workspace each time devtool reads the workspace.
So not only will necessary things get added, previously added items will
be removed by virtue of them no longer being in the workspace.
This also makes sure that the extensible sdk picks up unlocked-sigs.inc
as part of the configuration.
[YOCTO #9195]
Signed-off-by: Randy Witt <randy.e.witt@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
NetworkManager can clobber tap devices if left alone, this gives a
note about how to set tap* as unmanaged
[YOCTO #8587]
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Fixed when git > 2.1.0:
$ ./scripts/create-pull-request -r HEAD^ -u contrib -b rbt/git
fatal: Not a valid revision: rbt/git
ERROR: git request-pull reported an error
This is because newer git requires both local and remote branch named as
rbt/git, but usually, we only named the remote branch as rbt/foo, and
foo for local branch.
Add a option '-l' to fix the problem, default is HEAD.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
installed
Coverage subprocessing file is being created even when coverage
is not installed, which causes errors of "module not found" to
be send to the oe-selftest output.
This patch adds indent to the block of code creating this coverage
file, so it can only be executed when coverage is actually
installed.
[Yocto #9334]
Signed-off-by: Humberto Ibarra <humberto.ibarra.lopez@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Coverage temporal file for sub-processing is being removed only
when the --coverage option was found. This is wrong since the
file is created when any coverage option (source, include or omit)
is given, even if --coverage is not one of them.
This patch makes sure to remove the file if any coverage option was
given.
Signed-off-by: Humberto Ibarra <humberto.ibarra.lopez@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
There is a message that warns the user about enabling subprocessing
for coverage to work. After the fix for Yocto #8930,
this task is done automatically, so the warning is not needed
anymore.
Signed-off-by: Humberto Ibarra <humberto.ibarra.lopez@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The commit OE-Core:ada4639 (oe-buildenv-internal: simplify derivation
of BB_ENV_EXTRAWHITE) changed the format of BB_ENV_EXTRAWHITE
variable to:
,----
| BB_ENV_EXTRAWHITE='ALL_PROXY
| BB_NO_NETWORK
| BB_NUMBER_THREADS
| ...
| '
`----
Instead of:
,----
| BB_ENV_EXTRAWHITE='ALL_PROXY BB_NO_NETWORK BB_NUMBER_THREADS...'
`----
The old format allow for external script parsing easier and there is
no need to change the format as it has no benefit from usage
perspective.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Invoking oe-pkgdata-util in turn for every package in the list was slow
with a large image. Modify oe-pkgdata-util's read-value command to take
an option to read the list of packages from a file, as well as prefix
the value with the package name; we can then use this to get all of the
package sizes at once. This reduces the time to gather this information
from minutes to just a second or two.
Fixes [YOCTO #7339].
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
By default cpio preserves the uid&guid's of the original user which
leads to host contamination and boot failures because commands like
mount from initramfs expect to be run by root and the original host
user might not even exist on the target.
Signed-off-by: Ioan-Adrian Ratiu <adrian.ratiu@ni.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Some filenames can omit 'initramfs', or use other names. This makes
detection more flexible by using only the image name, machine arch and
image type in a glob wildcard.
Signed-off-by: Ioan-Adrian Ratiu <adrian.ratiu@ni.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
When installing SDK in a non-default location and the path length
of the SDK install location is longer than the path length of the
default SDK location, relocation of .ldsochache section will overwrite
file location outside of the .ldsocache section size.
In addition, additional checks were added to ensure that any
path in sections .gccrelocprefix and .ldsochache will not exceed
the space allocated for it within the file, which would also result
in file corruption.
[YOCTO #9268]
Signed-off-by: Juro Bystricky <juro.bystricky@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Strictly speaking we ought to explicitly shut down a tinfoil instance
when we're done with it. This doesn't affect modify's operation but is
important if you want to be able to call into modify() from another
plugin (though anyone doing so should be advised that the function is
by no means a stable API and is subject to change in future releases).
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
If SDK_EXT_TYPE is set to "minimal" then the SDK won't contain many
sstate artifacts, and you're required to set up an sstate mirror in this
case anyway so there's no point publishing the "stub" sstate-cache
directory from within the SDK since it won't be useful for update
purposes and may be confused with the real sstate-cache.
There is however a possibility that people might publish the real
sstate-cache directory under the same output directory provided to
oe-publish-sdk, thus deleting it after extracting (as we were doing with
other files we wanted to clean up at the end) would be problematic,
besides which extracting it and then deleting it is wasteful. Thus,
introduce a "-p" command line option to the SDK installer that we can
use to tell tar not to extract the items we don't want when publishing.
This has the added benefit of mostly keeping references to these in the
place they belong i.e. in populate_sdk_ext.bbclass.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The SDK argument is expected to be an installer .sh file; if a directory
is specified we can get an ugly failure later on; best to check up
front.
Fixes [YOCTO #9065].
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
There appears to have been a lot of copy and pasting of the code
which prints tracebacks upon failure and limits the stack trace to
5 entries. This obscures the real error and is very confusing to the user
it look me an age to work out why some tracebacks weren't useful.
This patch removes the limit, making tracebacks much more useful for
debugging.
[YOCTO #9230]
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
It should be the similar type as hddimg, rather than ext234 or btrfs.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This fixes the problem by changing the name to the coverage data file,
using the timestamp as an identifier.
The name for the coverage data file is constructed based on the tests
ran; this has created a couple of issues so far, affecting coverage report.
If --run-tests-by option is given, the data file name won't have any
identifier, causing following runs to overwrite themselves. On the
other hand, if too many tests are given, the file name exceeds linux
limits and fails to store the coverage data all together.
[Yocto #9253]
Signed-off-by: Humberto Ibarra <humberto.ibarra.lopez@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
When fixing paths for .gccrelocprefix section, it will corrupt the next
entry during updating the current one if "new_prefix" length is more
than "DEFAULT_INSTALL_DIR", this problem is obvious on the code, but it's
only found when install sdk onto a net file system.
Signed-off-by: Wenlin Kang <wenlin.kang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The new class uses gio-querymodules tool on postinst and postrm.
This regenerates the module cache which is useful to avoid loading
modules that are not needed at runtime: If a Gio module is not
listed in the cache file it will always get loaded.
* Add a postinst-intercept 'gio-module-cache': it runs
gio-querymodules using qemuwrapper. This is required because the tool
actually loads the modules to generate the cache.
* Add a gio-module-cache class that adds postinstall and postrm
scripts. In the sysroot population case use the new intercept.
* Inherit the new class in glib-2.0, glib-networking and gconf.
Fixes [YOCTO #9241].
Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
We also make sure to correctly support usernames that contain spaces.
For simplicity sed + regex has been replaced with shell parameter expansion,
which works in both, bash and dash.
Signed-off-by: André Draszik <adraszik@tycoint.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: André Draszik <adraszik@tycoint.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Type of --no-table option was incorrectly set in new wks parser.
It causes parser to require argument for this option, which makes
wic to fail with wks files that use --no-table:
Error: argument --no-table: expected one argument
Changed action parameter to 'store_true' to fix the issue.
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Use 'sort --unique' to combine the two lists and remove duplicates.
Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Use config_<machine> to split the code into separate functions, so that
different machines won't affect each other, and they will have a better
structure.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
|
|
* remove akita and spitz related code
They are not supported by runqemu anymore:
$ runqemu spitz
Error: unable to classify arg [spitz]
So remove related code.
* Remove checking of 256M for qemuarm, qemu can check it, for example:
$ runqemu qemuarm qemuparams="-m 1024"
[snip]
qemu: Too much memory for this machine: 1024 MB, maximum 256 MB
[snip]
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
|
|
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
|
|
[YOCTO #9168]
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
|
|
They don't work, and the script can check the type correctly.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
|
|
* So that we can add more image support easliy.
* I think that wic should be vm images.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
|
|
* Make it can boot scsi and virtio block drive such as root=/dev/sdX and
/dev/vdX.
* Drop VM from help info, id doesn't work, and the script can check
whether it is a vm disk or not.
* Make it can be run by:
$ runqemu tmp/deploy/images/qemux86-64/core-image-minimal-qemux86-64.vmdk
or:
$ runqemu qemux86-64 vmdk
[YOCTO #9170]
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
|
|
A typo in the command prevented the last part of the publish from
succeeding.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|