The eglplatform.h header is updated to account for deprecation of
MESA_EGL_NO_X11_HEADERS macro. Update srcrev to pick this change.
Signed-off-by: Gowtham Tammana <g-tammana@ti.com>
The additional patch for ti-sgx-ddk-km remaps the arm-oe-linux-gnueabi prefix
to the arm-linux-gnueabihf preferred target. Same should be done for the
arm-poky-linux-gnueabi prefix.
Signed-off-by: Zan Dobersek <zan@falconsigh.net>
Reviewed-by: Gowtham Tammana <g-tammana@ti.com>
It appears common-csl-ip-rtos does not require ti-sysbios dependency to build,
but ti-pdk.bbclass adds it by default. While most of PDK-based components do
require ti-sysbios to be built from sources, it is safe to remove it from CSL,
reducing the dependency tree for special cases like PRU FW based on CSL, but
not needing ti-sysbios.
This only affects firmwares built from sources against PDK and CSL, not when
RTOS firmwares are delivered in binary form for newer K3 platforms.
Build-tested on am335x, am57xx, k2g and am65xx.
Signed-off-by: Denys Dmytriyenko <denis@denix.org>
Fixes:
ERROR: cuia-1.0.0.13-r0 do_package: QA Issue: cuia: Files/directories were installed but not shipped in any package:
/usr/lib/gnu.targets.arm.rtsv5T.av5T
/usr/lib/syscalls.am3g
/usr/lib/gnu.targets.codesourcery.rtsv5T.a470MV
/usr/lib/syscalls.am4g
/usr/lib/gnu.targets.codesourcery.rtsv6.av6
/usr/lib/gnu.targets.arm.rtsv7A.av7A
/usr/lib/syscalls.am4fg
/usr/lib/gnu.targets.arm.rtsv7M.am4fg
/usr/lib/gnu.targets.arm.rtsv7M.am4g
/usr/lib/boot.am3g
/usr/lib/gnu.targets.arm.rtsv7M.am3g
/usr/lib/gnu.targets.arm.rtsv6.av6
/usr/lib/boot.am4fg
/usr/lib/boot.am4g
Please set FILES such that these items are packaged. Alternatively if they are unneeded, avoid installing them or delete them within do_install.
cuia: 14 installed and not shipped files. [installed-vs-shipped]
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
It is a pre-built binary, hence no control over passing correct LDFLAGS.
This used to be a warning and recently became an error. Fixes:
ERROR: vis-01.50.07.15-r1 do_package_qa: QA Issue: File /usr/bin/RadioApp in package vis doesn't have GNU_HASH (didn't pass LDFLAGS?) [ldflags]
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Workaround for these:
.../ti/drv/cppi/test/k2h/armv7/linux/fw_main.c:50: multiple definition of `coreNum';
./obj/k2h/cppi/test/cppi_test.o:.../ti/drv/cppi/test/src/cppi_test.c:102: first defined here
.../ti/drv/cppi/test/k2h/armv7/linux/fw_main.c:50: multiple definition of `errorCount';
./obj/k2h/cppi/test/cppi_test.o:.../ti/drv/cppi/test/src/cppi_test.c:102: first defined here
.../ti/drv/cppi/test/k2h/armv7/linux/fw_test.h:85: multiple definition of `fw_passCfgVaddr';
./obj/k2h/cppi/test/cppi_test.o:.../ti/drv/cppi/test/k2h/armv7/linux/fw_test.h:85: first defined here
.../ti/drv/cppi/test/k2h/armv7/linux/fw_test.h:84: multiple definition of `fw_srioCfgVaddr';
./obj/k2h/cppi/test/cppi_test.o:.../ti/drv/cppi/test/k2h/armv7/linux/fw_test.h:84: first defined here
.../ti/drv/cppi/test/k2h/armv7/linux/fw_test.h:83: multiple definition of `fw_qmssDataVaddr';
./obj/k2h/cppi/test/cppi_test.o:.../ti/drv/cppi/test/k2h/armv7/linux/fw_test.h:83: first defined here
.../ti/drv/cppi/test/k2h/armv7/linux/fw_test.h:82: multiple definition of `fw_qmssCfgVaddr';
./obj/k2h/cppi/test/cppi_test.o:.../ti/drv/cppi/test/k2h/armv7/linux/fw_test.h:82: first defined here
Needs to be fixed properly later:
https://gcc.gnu.org/gcc-10/porting_to.html
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Workaround for these:
qmss/mpm_transport_qmss_interface.so:(.bss+0x0): multiple definition of `fw_qmssCfgVaddr'; mpm_transport.so:(.bss+0x0): first defined here
qmss/mpm_transport_qmss_interface.so:(.bss+0x4): multiple definition of `fw_qmssDataVaddr'; mpm_transport.so:(.bss+0x4): first defined here
qmss/mpm_transport_qmss_interface.so:(.bss+0x8): multiple definition of `fw_srioCfgVaddr'; mpm_transport.so:(.bss+0x8): first defined here
qmss/mpm_transport_qmss_interface.so:(.bss+0xc): multiple definition of `fw_srioSerdesCfgVaddr'; mpm_transport.so:(.bss+0xc): first defined here
qmss/mpm_transport_qmss.so:(.bss+0x0): multiple definition of `fw_qmssCfgVaddr'; mpm_transport.so:(.bss+0x0): first defined here
qmss/mpm_transport_qmss.so:(.bss+0x4): multiple definition of `fw_qmssDataVaddr'; mpm_transport.so:(.bss+0x4): first defined here
qmss/mpm_transport_qmss.so:(.bss+0x8): multiple definition of `fw_srioCfgVaddr'; mpm_transport.so:(.bss+0x8): first defined here
qmss/mpm_transport_qmss.so:(.bss+0xc): multiple definition of `fw_srioSerdesCfgVaddr'; mpm_transport.so:(.bss+0xc): first defined here
qmss/mpm_transport_qmss_osal.so:(.bss+0x0): multiple definition of `fw_qmssCfgVaddr'; mpm_transport.so:(.bss+0x0): first defined here
qmss/mpm_transport_qmss_osal.so:(.bss+0x4): multiple definition of `fw_qmssDataVaddr'; mpm_transport.so:(.bss+0x4): first defined here
qmss/mpm_transport_qmss_osal.so:(.bss+0x8): multiple definition of `fw_srioCfgVaddr'; mpm_transport.so:(.bss+0x8): first defined here
qmss/mpm_transport_qmss_osal.so:(.bss+0xc): multiple definition of `fw_srioSerdesCfgVaddr'; mpm_transport.so:(.bss+0xc): first defined here
Needs to be fixed properly later:
https://gcc.gnu.org/gcc-10/porting_to.html
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Workaround for the following gcc10 errors:
.../gdbc6x/gdb/varobj.h:221: multiple definition of `ada_varobj_ops'; mi-cmd-var.o:.../gdbc6x/gdb/./varobj.h:221: first defined here
.../gdbc6x/gdb/varobj.h:220: multiple definition of `java_varobj_ops'; mi-cmd-var.o:.../gdbc6x/gdb/./varobj.h:220: first defined here
.../gdbc6x/gdb/varobj.h:219: multiple definition of `cplus_varobj_ops'; mi-cmd-var.o:.../gdbc6x/gdb/./varobj.h:219: first defined here
.../gdbc6x/gdb/varobj.h:218: multiple definition of `c_varobj_ops'; mi-cmd-var.o:.../gdbc6x/gdb/./varobj.h:218: first defined here
Needs to be resolved properly in the code later:
https://gcc.gnu.org/gcc-10/porting_to.html
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Workaround for this gcc10 error:
.../linux/src/api/gates/GateHWSpinlock.c:96: multiple definition of `_GateHWSpinlock_cfgParams';
.libs/Ipc.o:.../linux/src/api/Ipc.c:97: first defined here
Needs to be resolved properly in the code:
https://gcc.gnu.org/gcc-10/porting_to.html
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Cc: Suman Anna <s-anna@ti.com>
It was giving this warning for years, finally gone:
| warning: `sys_errlist' is deprecated; use `strerror' or `strerror_r' instead
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Cc: Suman Anna <s-anna@ti.com>
update am64x rtos firmware to keep pinmux and control MMRs
remain unlocked now since linux needs this to be unlock.
this fix the USB and performance degradation issues.
am64xx: 07.03.00.19
Signed-off-by: Shiju Sivasankaran <shiju@ti.com>
Signed-off-by: Praneeth Bajjuri <praneeth@ti.com>
Switch to using all of multi-certificate boot images. This allows us
to create bootable images - in this case, we are missing tiboot3.bin
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Denys Dmytriyenko <denys@konsulko.com>
IMAGE type and dependencies can easily be centralized to SoC include
file. This is necessary to make sure that we don't do the duplication
and having multiple places to fix up.
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Denys Dmytriyenko <denys@konsulko.com>
Switch to using all of multi-certificate boot images. This allows us
to create bootable images - in this case, we are missing tiboot3.bin
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Denys Dmytriyenko <denys@konsulko.com>
wic and tar dependency on trusted firmware is wrong. It is u-boot build
that depends on trusted firmware deployment and is correctly handled by
u-boot recipe. No point in duplicating that dependency, so just clean it
all up.
Reported-by: Denys Dmytriyenko <denys@konsulko.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Denys Dmytriyenko <denys@konsulko.com>
We use the default IMAGE_BOOT_FILES configuration which matches up with
legacy boot configuration. Document the same to allow folks to
understand by just looking at the SoC include file.
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Denys Dmytriyenko <denys@konsulko.com>
We use the default IMAGE_BOOT_FILES configuration which matches up with
legacy boot configuration. Document the same to allow folks to
understand by just looking at the SoC include file.
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Denys Dmytriyenko <denys@konsulko.com>
Lets keep things consistent by providing two possibilities for platforms
to pick from - legacy boot and multi-certificate boot.
In legacy boot, the base bootloader and system firmware are
maintained as separate binaries (tiboot3.bin and sysfw.itb).
In multi-certificate boot that newer K3 devices support, ROM is smarter
and can handle multiple x509 certificate based images: so we can strip
out the sysfw.itb and integrate it as part of tiboot3.bin itself. This
improves authentication and overall system boot times since we are now
able to boot both the system controller and the boot processor in
parallel.
We do have a scheme currently to identify the images necessary for boot
etc, but things are handled on a platform conf file basis. We can
improve that by introducing the pattern at the top level include and use
the relevant pattern in platforms as needed.
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Denys Dmytriyenko <denys@konsulko.com>
Remove duplicate machine-specific redefines of UBOOT_BINARY, UBOOT_IMAGE and
UBOOT_SYMLINK variables, as they are already defined with same values in the
common k3r5.inc file.
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Reviewed-by: Nishanth Menon <nm@ti.com>
Fixes:
.../scripts/secure-binary-image.sh: line 69: openssl: command not found
.../scripts/secure-binary-image.sh: line 78: openssl: command not found
Signed-off-by: Denys Dmytriyenko <denys@konsulko.com>
Signed-off-by: Praneeth Bajjuri <praneeth@ti.com>
ti-upstream-tools[1] repository carries the overlays and defconfig
fragments that we don't wish to overload upstream kernel community with.
The reason for that being these are constantly on churn as our test
coverage increases etc. Currently we reuse the "BRANCH" variable to also
matchup with kernel and ti-upstream-tools repositories as well.
This can tend to be a little painful if we decide to start testing
various kernel configurations - RT etc.. So, lets just split out the
TOOLS_BRANCH separated from the branch we use for kernel. Even though
the branch remains the exact same at this point.
[1] https://git.ti.com/cgit/ti-linux-kernel/ti-upstream-tools/
Signed-off-by: Dan Murphy <dmurphy@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Add GNU Multiple Precision Arithmetic Library as a dependency for
mainline linux builds. These are necesary for kernel to handle gcc
plugins for builtins.h
Signed-off-by: Dan Murphy <dmurphy@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Denys Dmytriyenko <denys@konsulko.com>
The J7200/AM64 ti-sci-fw needs to be installed in the deploy
directory so that it can be packaged in Processor SDK bundle
This is required because the ti-sci-fw is needed in the
installer prebuilt-images in order to build the combined boot
image for J7200/AM64
Signed-off-by: Lucas Weaver <l-weaver@ti.com>
update ti-rtos-metadata and ti-rtos-firmware recipes of meta-ti to
pick AM64x rtos firmware from mcu-plus-sdk instead of PDK.
Other devices such as AM65x, J7 continue to use fw from PDK.
Built and tested both IPC echo test applications - "rpmsg_client_sample”
& “rpmsg_char_simple”. Both tests passed for all remote R5F cores
on AM64x EVM
Signed-off-by: Shiju Sivasankaran <shiju@ti.com>
[praneeth@ti.com: rebase to dunfell/next]
Signed-off-by: Praneeth Bajjuri <praneeth@ti.com>
Praneeth Bajjuri
Gowtham Tammana
Zan Dobersek
Denys Dmytriyenko
Denys Dmytriyenko
Shiju Sivasankaran
Nishanth Menon
Dan Murphy
Lucas Weaver