A user of the meta-ti layer may choose to use a combination
of the machines herein with another kernel, not necessarily
the one from the machine definition.
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org>
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
The AM654 SoC is a lead device of the K3 Multicore SoC architecture
platform, targeted for broad market and industrial control with aim to
meet the complex processing needs of modern embedded products.
See AM65x Technical Reference Manual (SPRUID7, April 2018)
for further details: http://www.ti.com/lit/pdf/spruid7
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
opencl-monitor does not distinguish between dra7xx and am57xx devices,
treating both as the same and attempting to build for am57xx regardless.
However, this is dependent on common-csl-ip-rtos, a component of PDK,
which in general does handle these devices differently. The definition
of am57xx SOCs in the PDK list from omap-15.conf gets inherited by
dra7xx-evm.conf, thereby creating a rule that all dra7xx PDK recipes
must build am57xx libraries. This results in a dra7xx image which is
much larger than necessary, and more severely, creates an artifical
dependency link between these two devices and introduces build failures
for various PDK recipes.
Instead, we now override the PDK SOC definition in dra7xx-evm.conf, and
only in the CSL RTOS recipe do we append the am57xx SOCS for dra7xx-evm
machine build. This allows for opencl-monitor to build successfully for
dra7xx while preventing any other PDK recipes from attempting to build
libraries for am57xx devices.
This is a workaround that must stay in place until opencl build can
properly comprehend am57xx and dra7xx independently, and/or omap-a15
is not implied to be synonymous with am57xx in various recipes.
Signed-off-by: Stephen Molfetta <sjmolfetta@ti.com>
Signed-off-by: Jacob Stiffler <j-stiffler@ti.com>
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
* The omapl137 family does not have linux support and is only
supported by baremetal and TI RTOS.
Signed-off-by: Jacob Stiffler <j-stiffler@ti.com>
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
* The c66x SOCs (e.g. c665x-evm and c667x-evm) are DSP-only devices,
and therefore do not have a supported kernel nor an integrated
toolchain.
* Add these machines to allow the creation of baremetal and TI RTOS
packages.
Signed-off-by: Jacob Stiffler <j-stiffler@ti.com>
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
* Set the ti-pdk class variables which sets the list of devices and
boards supported.
Signed-off-by: Jacob Stiffler <j-stiffler@ti.com>
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
The bb.data.getVar() API has been deprecated & removed. This change
follows that change and allows MACHINE=beaglebone to work again.
Signed-off-by: Gary Thomas <gary@mlbassoc.com>
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
AM43x was missing the hdmi dtb file in the
KERNEL_DEVICETREE variable. Adding this in allows for
a copy to deploy directory.
Signed-off-by: Dan Murphy <dmurphy@ti.com>
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
Set DEFAULTTUNE optimization in BSP from specific Cortex level to more generic
armv7a, but keep thumb, neon and hard-float options enabled due to numerous BSP
components requiring them. While enabling additional Cortex optimization might
gain a tiny performance improvement, it makes binary feeds incompatible between
A8, A9 and A15 machines, leads to unnecessary rebuilds and hides potential
packaging or recipe issues. Hence the decision to enable Cortex optimization
should be left to the Distro.
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
Correct the name of the variable used to define the size of the logo:
GUI_MACHINE_CLASS renamed MACHINE_GUI_CLASS.
Signed-off-by: Thomas Perrot <thomas.perrot@tupi.fr>
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
Daniel D?az
Denys Dmytriyenko
e-ruei1@ti.com
Stephen Molfetta
Mahesh Radhakrishnan
Hongmei Gou
Jacob Stiffler
Gary Thomas
Dan Murphy
Thomas Perrot