Age | Commit message (Collapse) | Author |
|
Pulls in multi CPU support (no more limitation to dual-core max).
Signed-off-by: Aniket Limaye <a-limaye@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
Updated the value(s) for:
ltp-ddt: SRCREV
Signed-off-by: LCPD Automation Script <lcpdbld@list.ti.com>
|
|
Updated the value(s) for:
ltp-ddt: SRCREV
Signed-off-by: LCPD Automation Script <lcpdbld@list.ti.com>
|
|
j721e and j784s4 support UFS but the package ufs-utils was missing for j784s4.
Adding the same.
Signed-off-by: Aniket Limaye <a-limaye@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
Updated the value(s) for:
ltp-ddt: SRCREV
Signed-off-by: LCPD Automation Script <lcpdbld@list.ti.com>
|
|
Updated the value(s) for:
ltp-ddt: SRCREV
Signed-off-by: LCPD Automation Script <lcpdbld@list.ti.com>
|
|
Image recipe for jailhouse does not have a summary and description. Due to
which oe-layersetup also shows No Summary available for tisdk-jailhouse-image
as compared to other targets for which we have proper summary.
So to ensure good practice and help other users understand the purpose and
functionality of recipe, add a summary and description for jailhouse image.
Signed-off-by: Paresh Bhagat <p-bhagat@ti.com>
Acked-by: Chirag Shilwant <c-shilwant@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
Updated the value(s) for:
ltp-ddt: SRCREV
Signed-off-by: LCPD Automation Script <lcpdbld@list.ti.com>
|
|
Add a new jailhouse-inmate image which will be used as initramfs
for jailhouse linux demo. This image will contain some additional
applications and packages need for OOB demo.
Signed-off-by: Paresh Bhagat <p-bhagat@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
Add Upstream-Status to a patch we've been carrying for a bit. This has been
accepted upstream and will be present in the version of Weston used in the next
Yocto LTS.
Signed-off-by: Randolph Sapp <rs@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
Backport some sanity patches related to plane ordering and positioning.
These are all mandatory for display subsystems that allow primary and
overlay mutable zpos.
They have all already been accepted upstream already and are only
necessary for Kirkstone's version of Weston.
Signed-off-by: Randolph Sapp <rs@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
Backport some sanity patches related to plane ordering and positioning.
These are all mandatory for display subsystems that allow primary and
overlay mutable zpos.
Otherwise Weston will attempt to place sprite/overlay planes above
whatever active primary plane it chooses for scanout, running into
issues if the primary plane is above zpos 0 and there isn't enough
headroom.
They have all already been accepted upstream already and are only
necessary for Kirkstone's version of Weston.
Signed-off-by: Randolph Sapp <rs@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
Updated the value(s) for:
ltp-ddt: SRCREV
Signed-off-by: LCPD Automation Script <lcpdbld@list.ti.com>
|
|
The init.d scripts for sysrepo and netopeer2-server requires certain
other init.d scripts that are now removed. And the systemd services generated
systemd-sysv-generator is not sufficient.
So introduce new unit files for sysrepo and netopeer2 and update the
respective bitbake recipes.
Signed-off-by: Ravi Gunasekaran <r-gunasekaran@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
Use Yocto variable names instead of path value
Signed-off-by: Ravi Gunasekaran <r-gunasekaran@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
The new dynamic meta-chromium-browser layer will do this same thing
and add chromium-ozone-wayland into the image if the layer is detected.
This is a hangover from dunfell and doesn't serve any purpose in
kirkstone so removing to uncomplicate things.
Signed-off-by: Darren Etheridge <detheridge@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
If the chromium-browser-layer is detected, then add
chromium-ozone-wayland into the default image.
If the layer is detected then append chromium-ozone-wayland to the
IMAGE_INSTALL variable. This increases the build time and image size
which is why we make this conditional on the browser layer being
present.
Signed-off-by: Darren Etheridge <detheridge@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
This applies a patch for meta-browser/chromium-ozone-wayland to make the
sandboxing work with the Imagination GPU components without the need for
the --no-sandbox flag.
GPU acceleration in Chromium is dependant on IMG DDK 23.3. It works
across AXE/BXS/8XE GPU's. No acceleration is expected for SGX.
Also add an upstream patch to stop Chromium from segfaulting when it is
run with no input devices connected to the board.
Patch was retrieved from:
https://chromium.googlesource.com/chromium/src/+/323077958301bc321d840a2c2b983ab469934753
Signed-off-by: Darren Etheridge <detheridge@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
Cleaning up left over chromium stuff that I suspect won't even build
with kirkstone. The meta-browser layer has moved on considerably and my
next patches will replace this with GPU support on IMG Rogue based GPU
devices within Chromium v111.
Signed-off-by: Darren Etheridge <detheridge@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
When multiple CAN's are present, then names that are getting assigned
change after every boot even after providing alias in the device
tree.
So add udev rules to specify the interface name to use for
each mcan device, to make it persistent across reboots.
Signed-off-by: Bhavya Kapoor <b-kapoor@ti.com>
Signed-off-by: Aniket Limaye <a-limaye@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
GLES3 headers are missing in the linux-devkit/sysroots since 09.00.
These are needed for cross compiling QT, OpenGLES and other graphics applications.
Signed-off-by: Sai Sree Kartheek Adivi <s-adivi@ti.com>
Acked-by: Chirag Shilwant <c-shilwant@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
Updated the value(s) for:
ltp-ddt: SRCREV
Signed-off-by: LCPD Automation Script <lcpdbld@list.ti.com>
|
|
We don't support x11, and while it can currently be enabled in tandem
with our graphics stack right now without explicitly causing issues on
it's own, Yocto still treats the presence of x11 as a distro feature to
mean that desktop GL is present.
This assumption leads to a lot of incorrectly configured packages in our
case so to prevent future incompatibility with layers adjusting their
PACKAGECONFIG values with this logic we should explicitly remove x11
again.
This decision will likely flip again if oe-core introduces a standard
"opengles" feature instead of just the blanket "opengl" one they
currently use.
Signed-off-by: Randolph Sapp <rs@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
- In past, there have been issues in do_fetch from infradead.org over git protocol in proxy based network [1].
- As per the latest release announcement from mtd-utils, infraroot.at is the official mirror for infradead.org [2].
- Since infradead.org has fetch failures over git, start using infraroot.at over https in SRC_URI.
Reference
---------
[1] https://e2e.ti.com/support/processors-group/processors/f/processors-forum/962170/processor-sdk-am65x-how-to-resolve-the-fetch-error-that-occurs-when-executing-bitbake
[2] https://lists.infradead.org/pipermail/linux-mtd/2023-August/100922.html
Signed-off-by: Chirag Shilwant <c-shilwant@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
Updated the value(s) for:
ltp-ddt: SRCREV
Signed-off-by: LCPD Automation Script <lcpdbld@list.ti.com>
|
|
Updated the value(s) for:
ltp-ddt: SRCREV
Signed-off-by: LCPD Automation Script <lcpdbld@list.ti.com>
|
|
The use of the variable SYSTEMD_DEFAULT_TARGET is as a filename target.
But the extra whitespace in the string caused by the link breaks and
indents added spaces between the directory and the filename which
resulted in the following error:
run.set_systemd_default_target.23991: line 147: [: too many arguments
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
This is required to avoid the frame drops and increase the tolrence
for CSI capture using gsteamer
Signed-off-by: Rahul T R <r-ravikumar@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
This will make it easier to maintain and it will be easier if we decide
to have TI fork for gstreamer.
Signed-off-by: Rahul T R <r-ravikumar@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
Fix tisdk-default-image do_rootfs errors after [1]
cherry-pick commit from master branch: 922420df0f445c8f7c830965f88a5fc566eb6ff0
Upstream strongswan recipe in meta-openembedded/meta-networking sets
runtime dependency for MGF1 plugin, but it is not enabled by default
and hence no such package gets produced.
[1]: https://git.openembedded.org/meta-openembedded/commit/meta-networking/recipes-support/strongswan?h=kirkstone&id=dd23a9930360f540c919ddfd35fd1f6088174ef3
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Signed-off-by: Aniket Limaye <a-limaye@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
Updated the value(s) for:
ltp-ddt: SRCREV
Signed-off-by: LCPD Automation Script <lcpdbld@list.ti.com>
|
|
Add patches to prefer contiguous video format and supporting
dmabuf-import mode for video decoder.
This is to support use-cases involving video codec with other components
viz. GPU, Camera e.t.c.
For 0002-v4l2-Give-preference-to-contiguous-format-if-support.patch :
There are certain blocks such as camera, GPU which only seem
to support well the contigous formats (sometimes at the S/W level)
which doesn't go well with gstreamer v4l2 which prefers non-contigous
by default and we fail at negotiation.
The crux of the issue is that unlike v4l2 gstreamer has only single GST
video format for both NV12 and NV12M (non-contigous) and it prefers to
advertise latter during caps negotiation.
There was a proposal made long back to have separate GST video
formats for NV12 and NV12M but that didn't converge, perhaps they
wanted to support this with a VideoMeta based scheme as per
logs [0].
For 0001-v4l2-Changes-for-DMA-Buf-import-j721s2.patch :
This is a local patch which complements a patch already made
available in driver to support dmabuf import for the decoder.
Here the gist of the issue is that v4l2 framework expects userspace
to make sure it uses same v4l2 buffer index each time with same backing
memory as kernel framework relies on these indices to identify frame
ordering [2].
There again is discussion around this to support this in upstream but
the upstream patch is not yet merged [1].
[0]:
https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/issues/414
[1]:
https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4550/diffs?commit_id=b12eaf65f489bab50917e22242758e000d69df58
[2]:
https://www.kernel.org/doc/html/latest/userspace-api/media/v4l/dev-decoder.html
Signed-off-by: Devarsh Thakkar <devarsht@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
All READMEs were still mentioning master and old release version, update that.
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
This layer does not depend on meta-qt5.
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
Updated the value(s) for:
ltp-ddt: SRCREV
Signed-off-by: LCPD Automation Script <lcpdbld@list.ti.com>
|
|
AM335x has some display/graphics hardware limitations and moreover
BeagleBone Black only has external HDMI output, hence add a custom
weston.ini that sets core/gbm-format=rgb565 and output/name=HDMI-A-1
to work out of the box. The rest of the .ini file is unmodified from
the generic one.
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Acked-by: Andrew Davis <afd@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
This was using a SysVinit initscript that was also called from a systemd
unit file when systemd-compat-units were enabled. With systemd and
SysVinit now fully separated in Arago, adjust initscripts accordingly to
work for both init managers.
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
This was using a SysVinit initscript that was also called from a systemd
unit file when systemd-compat-units were enabled. With systemd and
SysVinit now fully separated in Arago, adjust initscripts accordingly to
work for both init managers.
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
Updated the value(s) for:
ltp-ddt: SRCREV
Signed-off-by: LCPD Automation Script <lcpdbld@list.ti.com>
|
|
Add am62pxx as compatible machine for jailhouse image. Also since
this image is primarily for jailhouse remove the machine in
IMAGE_INSTALL:append.
Signed-off-by: Paresh Bhagat <p-bhagat@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
Updated the value(s) for:
SRCREV
Signed-off-by: LCPD Automation Script <lcpdbld@list.ti.com>
|
|
- Install systemd service file or init script based on the distro_features
systemd/sysvinit.
- systemd service also needs to depend on lighttpd service so that matrix
won't start before the webserver.
Signed-off-by: Sai Sree Kartheek Adivi <s-adivi@ti.com>
Acked-by: Sinthu Raja <sinthu.raja@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
Updated the value(s) for:
SRCREV
Signed-off-by: LCPD Automation Script <lcpdbld@list.ti.com>
|
|
A bit arbitrary, but it seems like a good point in time to bump Arago Distro
version, since things are relatively stable currently and to coincide with
the Yocto Project Nanbield release.
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Acked-by: Ryan Eatmon <reatmon@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
Change the inmate ramfs which is used for jailhouse linux demo
to use a smaller and dedicated tisdk-tiny-initramfs instead of
tisdk-tiny-image.
Signed-off-by: Paresh Bhagat <p-bhagat@ti.com>
Reviewed-by: Chirag Shilwant <c-shilwant@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
Updated the value(s) for:
SRCREV
Signed-off-by: LCPD Automation Script <lcpdbld@list.ti.com>
|
|
Updated the value(s) for:
SRCREV
Signed-off-by: LCPD Automation Script <lcpdbld@list.ti.com>
|
|
Signed-off-by: Sai Sree Kartheek Adivi <s-adivi@ti.com>
Acked-by: Aniket Limaye <a-limaye@ti.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|
|
As tisdk-bundle was based on the image class long time ago, it inherited
some of its old attributes and artifacts used back in the days. There
were several updates and reworks done to this class through the years to
keep it up to date and in sync with upstream image class, but some pieces
fell through the cracks. In recent and not so recent years upstream image
class have deprecated some things, changed and extended others, such as
consolidated umask handling, improved sstate handling, expanded list of
cleaned directories and changed stamping and locking policy.
Since tisdk-bundle inherits image class, corresponding attributes for
do_rootfs, do_image and do_image_complete don't need to be modified here
and can be left to their default values, as well as the order of those
tasks.
This change addresses the issue with sstate/deploy file collisions we've
been seeing at times lately resulting in pseudo abort errors. And it
improves long term maintainability of tisdk-bundle class as well.
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Signed-off-by: Ryan Eatmon <reatmon@ti.com>
|