summaryrefslogtreecommitdiffstats
path: root/meta-xilinx-standalone/conf
Commit message (Collapse)AuthorAgeFilesLines
* meta-xilinx-standalone: Rework embeddedsw integrationHEADmasterMark Hatle2024-11-241-1/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Introduce a new embeddedsw-source recipe that is a single point for unpack, and patch of the sources. This will allow someone to universally patch (via a bbappend or similar) the esw sources. As part of this work, we transfered the various patches for fsbl, pmu, plm, and psm to the universe source recipe. Most of this is transparent to any recipe using xlnx-embeddedsw. The class automaticaly detects if the recipe is a user our the special embeddedsw-source recipe and enables the correct behavior. Since the esw sources expect the S and B to be the same, the cmake objects get written into S, we are not able to follow the gcc-source example. Instead recipes that use the xlnx-embeddedsw class will hard link or copy the source files to their working directory. This avoid the penalty for fetch, unpack and patch. Note these recipes may also add their own SRC_URI which WILL be processed as normal. This is often used to introduce special tcl scripts. See the do_copy_shared_src task injected by the xlnx-embeddedsw.bbclass for technical details. Additionally, correct some of the existing patch Upstream-Status entries to conform to current YP guidelines. For 2024.2 introduce a special Xil_Assert fix for the current toolchain which is more strict on type conversion.. See the patch itself for more details. Signed-off-by: Mark Hatle <mark.hatle@amd.com>
* layer.conf: Move from scarthgap to styhead walnascar, bring in old recipes ↵Mark Hatle2024-11-241-1/+1
| | | | | | | | | for parse binutils-2.42 and gstreamer-1.22.12 were pulled in from the latest poky scarthgap, as of commit dce4163d42f7036ea216b52b9135968d51bec4c1. Signed-off-by: Mark Hatle <mark.hatle@amd.com>
* distro/xilinx-standalone: Add inclusion to petalinux-mirrors.confMark Hatle2024-10-111-0/+9
| | | | | | | | | | | | | | | | | By default PetaLinux uses the own-mirror class to access cached version of the software. While this distribution configuration is not PetaLinux, nor should it require PetaLinux to operating properly, including the same mirror configuration will unify the Linux and non-Linux (baremetal) configurations. This resolves an issue with BB_NO_NETWORK set to 1, and a local disk mirror is enabled. Uses 'include' instead of 'require' so no error is generated if meta-petalinux (which it not required) is not present in the user's configuration. In this case behavior reverts to default and any mirror configurations also return to default behavior. Signed-off-by: Mark Hatle <mark.hatle@amd.com>
* xilinx-standalone: Add workaround for warning about embedded buildpathsMark Hatle2024-10-051-0/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Normally the presence of buildpaths indicates a problem with a library or executable. This is resolved by the system passing in prefix mapping values to ensure all paths are target oriented. For some reason, only in baremetal configurations, part of the debug paths are NOT being remapped. These paths are not actually used anywhere: The Directory Table (offset 0x22, lines 6, columns 1): Entry Name 0 (indirect line string, offset: 0x74): /scratch/mhatle/git/2024.2/build/tmp-zynqmp-zcu104-sdt-full-cortexa53-fsbl/work/cortexa53-xilinx-elf/csudma/2024.2+git/build 1 (indirect line string, offset: 0x102): /usr/include/machine 2 (indirect line string, offset: 0x117): /usr/include/sys 3 (indirect line string, offset: 0x128): /usr/lib/aarch64-xilinx-elf/gcc/aarch64-xilinx-elf/13.3.0/include 4 (indirect line string, offset: 0x16a): /usr/include 5 (indirect line string, offset: 0x177): /usr/src/debug/csudma/2024.2+git/XilinxProcessorIPLib/drivers/csudma/src The File Name Table (offset 0x40, lines 12, columns 2): Entry Dir Name 0 5 (indirect line string, offset: 0x1c0): xcsudma.c 1 1 (indirect line string, offset: 0x1ca): _default_types.h 2 2 (indirect line string, offset: 0x1db): _stdint.h 3 3 (indirect line string, offset: 0x1e5): stddef.h 4 4 (indirect line string, offset: 0x1ee): xil_types.h 5 5 (indirect line string, offset: 0x1fa): xcsudma.h 6 4 (indirect line string, offset: 0x204): xil_util.h 7 4 (indirect line string, offset: 0x20f): string.h 8 4 (indirect line string, offset: 0x218): xil_cache.h 9 4 (indirect line string, offset: 0x224): xil_assert.h 10 5 (indirect line string, offset: 0x231): xcsudma.c 11 4 (indirect line string, offset: 0x23b): xil_io.h In the above example, entry '0' in the directory table is tiggering the the buildpath QA warning. However in the file name table the path is never referred to. Also I have verified that all of the path remapping arguments are being passed in correctly during compilation and linking. Based on this, it appears to be a valid workaround to simply disable the test while we determine why the invalid enty is present. Signed-off-by: Mark Hatle <mark.hatle@amd.com>
* layer.conf: Remove the default XILINX_WITH_ESW valuesMark Hatle2024-09-131-3/+0
| | | | | | | | It is now up to the machine .conf files to define the way the ESW software will be built. 'xsct' or 'sdt' is the usual setting. Subsequent changes will adjust the default BSPs with correct defaults. Signed-off-by: Mark Hatle <mark.hatle@amd.com>
* layer.conf: Add LAYERBASE_xilinx-standalone variableSandeep Gundlupet Raju2024-08-201-0/+3
| | | | | | | Add LAYERBASE_xilinx-standalone variable so that files from meta-xilinx-standalone can be accessed from other layers. Signed-off-by: Sandeep Gundlupet Raju <sandeep.gundlupet-raju@amd.com>
* meta-xilinx-standalone: xilinx-standalone: Remove FULL_OPTIMIZATION changesMark Hatle2024-07-171-6/+0
| | | | | | | | | | | | | Yocto Project has changed the optimization settings in poky commit 1cf0974ad242f7eb2815a4ef0e3e5b6507ca56ea Since the settings match the Yocto Project defaults, remove them. If we need to default for size in the future, we can re-implement this using the current approach. Signed-off-by: Mark Hatle <mark.hatle@amd.com> (cherry picked from commit b7708778180849caa972a6ccebc01aa60d97ddfb) Signed-off-by: Mark Hatle <mark.hatle@amd.com>
* meta-xilinx-standalone: Move to YP best practicesMark Hatle2024-05-282-9/+24
| | | | | | | | | | | Follow the examples from meta-virtualization on how to enable bbappends and set PREFERRED_VERSIONS. This will require the user to set XILINX_WITH_ESW to some value. Any value will enable the components in this layer at this time. We may choose to be more selective in the future. Signed-off-by: Mark Hatle <mark.hatle@amd.com>
* meta-xilinx-standalone: Create 2024.2 recipesMark Hatle2024-05-211-4/+9
| | | | | | | Adjust the way the PREFERRED_VERSION happens to allow us to use the same login in layers that depend on this one. Signed-off-by: Mark Hatle <mark.hatle@amd.com>
* Merge remote-tracking branch 'xilinx/rel-v2024.1' into master-nextMark Hatle2024-03-261-3/+5
|\ | | | | | | Signed-off-by: Mark Hatle <mark.hatle@amd.com>
| * meta-xilinx-standalone-experimental: open-amp: pick up multiconfig recipeBen Levinsky2024-02-131-2/+1
| | | | | | | | | | | | | | Update to use latest open-amp and MC recipe. Signed-off-by: Ben Levinsky <ben.levinsky@amd.com> Signed-off-by: Mark Hatle <mark.hatle@amd.com>
| * standalone: Add support for r5/r52 hf tunesOnkar Harsh2024-02-091-0/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | c45298aaf4fe commit (machine/include/arm: Backport tune files from master) in meta-xilinx layer added the support for cortex-r5 and cortexr-r52 hard float ABI tunes. Update the conditions that generate the CMAKE_SYSTEM_PROCESSOR to incorporate these two new tunes. Moreover, some additional cflags (-DARMR5 -DARMR52) are needed to compile Cortex R52 specific boot code in embeddedsw. Update the same. Signed-off-by: Onkar Harsh <onkar.harsh@amd.com> Signed-off-by: Mark Hatle <mark.hatle@amd.com>
| * standalone: Add support for Versal Net specific CMAKE variablesOnkar Harsh2024-02-071-0/+2
| | | | | | | | | | | | | | | | | | | | | | | | CMAKE_MACHINE and CMAKE_SYSTEM_PROCESSOR values are different for Versal Net than that of Versal. Update the same using the SOC_VARIANT as the SOC_FAMILY is same for both. In addition, update the SOC specific flags (defined in-house to segragate execution of some part of the boot code and the drivers) for versal and versal-net. Signed-off-by: Onkar Harsh <onkar.harsh@amd.com> Signed-off-by: Mark Hatle <mark.hatle@amd.com>
| * xilinx-standalone.inc: Set default libmetal providerMark Hatle2023-11-271-1/+1
| | | | | | | | Signed-off-by: Mark Hatle <mark.hatle@amd.com>
* | layer.conf: Move from nanbield to scarthgapMark Hatle2024-03-191-1/+1
| | | | | | | | Signed-off-by: Mark Hatle <mark.hatle@amd.com>
* | layer.conf: Set to nanbieldMark Hatle2023-11-081-1/+1
| | | | | | | | Signed-off-by: Mark Hatle <mark.hatle@amd.com>
* | Merge remote-tracking branch 'origin/rel-v2023.2' into nanbieldMark Hatle2023-10-231-4/+2
|\| | | | | | | Signed-off-by: Mark Hatle <mark.hatle@amd.com>
| * xilinx-standalone: Fix invalid-packageconfig QA IssueSandeep Gundlupet Raju2023-08-031-0/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | libxil warnings are obversed in multiconfig builds. WARNING: mc:cortexa53-0-zynqmp-fsbl-baremetal:libxil-2023_sdt_experimental+gitAUTOINC+742a608800-r0 do_configure: QA Issue: libxil: invalid PACKAGECONFIG: rtc [invalid-packageconfig] WARNING: mc:cortexa53-0-zynqmp-fsbl-baremetal:libxil-2023_sdt_experimental+gitAUTOINC+742a608800-r0 do_configure: QA Issue: libxil: invalid PACKAGECONFIG: qemu-usermode [invalid-packageconfig] This is due to rtc and qemu-usermode was enabled for all targets in MACHINE_FEATURES_BACKFILL, Hence disable rtc and qemu-usermode for multiconfig targets by adding to MACHINE_FEATURES_BACKFILL_CONSIDERED list in inclusion file. Signed-off-by: Sandeep Gundlupet Raju <sandeep.gundlupet-raju@amd.com> Signed-off-by: Mark Hatle <mark.hatle@amd.com>
| * xilinx-standalone.inc: Remove security_flags.incMark Hatle2023-04-131-4/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Inclusion of security_flags.inc lead to a number of problems with the baremetal distributions: *ARM v8* aarch64-xilinx-elf-ld: testA53.elf: error: PHDR segment not covered by LOAD segment aarch64-xilinx-elf-ld: warning: testA53.elf has a LOAD segment with RWX permissions *ARM v7* arm-xilinx-eabi-ld: .../aarch32-xilinx-eabi/usr/lib/thumb/v7-a+fp/hard/crtbegin.o: relocation R_ARM_THM_MOVW_ABS_NC against `a local symbol' can not be used when making a shared object; recompile with -fPIC The configuration that works: export CFLAGS=" -O2 -pipe -g -feliminate-unused-debug-types " export CXXFLAGS=" -O2 -pipe -g -feliminate-unused-debug-types " export LDFLAGS="-Wl,-O1 -Wl,--hash-style=gnu -Wl,--as-needed -Wl,-z,relro,-z,now" export CPPFLAGS="" Configuration that provides the warnings/errors: export CFLAGS=" -O2 -pipe -g -feliminate-unused-debug-types " export CXXFLAGS=" -O2 -pipe -g -feliminate-unused-debug-types " export LDFLAGS="-Wl,-O1 -Wl,--hash-style=gnu -Wl,--as-needed " export CPPFLAGS="" The LDFLAGS was introdued by the security_flags.inc, the SECURITY_CFLGAS:class-target ?= "" and SECURITY_LDFLAGS:class-target ?= "" was supposed to prevent this, however the cross compiler is built in the cross or canadian-cross or other context and NOT target. Additionally the SDK environment is configured in the nativesdk environment, so the default values may not match the configured (target) compiler and CFLAGS. Removing security_flags.inc resolves both of these issues, as the security flags don't really do anything on a baremetal configuration, by default, anyway. Signed-off-by: Mark Hatle <mark.hatle@amd.com>
| * machines: Move special TC machines to meta-petalinuxMark Hatle2023-02-245-1083/+0
| | | | | | | | | | | | | | The special toolchains are really only applicable to a PetaLinux/Vitis toolchain build. Signed-off-by: Mark Hatle <mark.hatle@amd.com>
| * microblaze-tc.conf: Disable pie support, not available on microblazeMark Hatle2023-02-091-0/+8
| | | | | | | | Signed-off-by: Mark Hatle <mark.hatle@amd.com>
| * xilinx-standalone.inc: Further sync with upstream disro settingsMark Hatle2023-01-162-15/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | Many of the settings/workaround have been replaced by changes in the upstream OE-Core tclibc-newlib. Additionally adding the DISTROVERRIDES with an append was potentially changing -native recipes hashes. Move this only where needed to avoid this behavior. Also sync settings with meta-petalinux, such as the buildhistory. This is now moved to the local.conf file. Signed-off-by: Mark Hatle <mark.hatle@amd.com>
| * xilinx-standalone.inc: Sync with poky distro settingsMark Hatle2023-01-161-0/+18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Certain settings appear to have an overall affect on recipe hashes. Use petalinux (poky) distro as an example for base settings to ensure native items built in one config SHOULD also be re-usable in the bareemetal config. This has a side effect of NOT disabling static binaries in baremetal, but this seems like a reasonable default, since there really isn't a concept of dynamically loaded libraries. If necessary, the user can still override this, but will need to specifically select class-target to do so. Similarly SECURITY_CFLAGS and SECURITY_LDFLAGS are available by empty by default. The default cflags/ldflags for baremetal usage are likely different from a Linux system, so can be used in the future if requried. Signed-off-by: Mark Hatle <mark.hatle@amd.com>
| * *-tc.conf: Update machines to define qemu-xilinx as preferredMark Hatle2022-12-144-0/+4
| | | | | | | | | | | | | | Each Xilinx machine.conf should include the qemu file to ensure the correct PREFERRED_PROVIDERS_qemu (and related) are defined properly. Signed-off-by: Mark Hatle <mark.hatle@amd.com>
* | layer.conf: Enable mickledore supportMark Hatle2023-04-121-1/+1
| | | | | | | | Signed-off-by: Mark Hatle <mark.hatle@amd.com>
* | machines: Move special TC machines to meta-petalinuxMark Hatle2023-04-105-1083/+0
| | | | | | | | | | | | | | The special toolchains are really only applicable to a PetaLinux/Vitis toolchain build. Signed-off-by: Mark Hatle <mark.hatle@amd.com>
* | microblaze-tc.conf: Disable pie support, not available on microblazeMark Hatle2023-04-101-0/+8
| | | | | | | | Signed-off-by: Mark Hatle <mark.hatle@amd.com>
* | xilinx-standalone.inc: Further sync with upstream disro settingsMark Hatle2023-04-102-15/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | Many of the settings/workaround have been replaced by changes in the upstream OE-Core tclibc-newlib. Additionally adding the DISTROVERRIDES with an append was potentially changing -native recipes hashes. Move this only where needed to avoid this behavior. Also sync settings with meta-petalinux, such as the buildhistory. This is now moved to the local.conf file. Signed-off-by: Mark Hatle <mark.hatle@amd.com>
* | xilinx-standalone.inc: Sync with poky distro settingsMark Hatle2023-04-101-0/+18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Certain settings appear to have an overall affect on recipe hashes. Use petalinux (poky) distro as an example for base settings to ensure native items built in one config SHOULD also be re-usable in the bareemetal config. This has a side effect of NOT disabling static binaries in baremetal, but this seems like a reasonable default, since there really isn't a concept of dynamically loaded libraries. If necessary, the user can still override this, but will need to specifically select class-target to do so. Similarly SECURITY_CFLAGS and SECURITY_LDFLAGS are available by empty by default. The default cflags/ldflags for baremetal usage are likely different from a Linux system, so can be used in the future if requried. Signed-off-by: Mark Hatle <mark.hatle@amd.com>
* | *-tc.conf: Update machines to define qemu-xilinx as preferredMark Hatle2023-04-104-0/+4
|/ | | | | | | Each Xilinx machine.conf should include the qemu file to ensure the correct PREFERRED_PROVIDERS_qemu (and related) are defined properly. Signed-off-by: Mark Hatle <mark.hatle@amd.com>
* Merge remote-tracking branch 'origin/kirkstone-next' into master-nextMark Hatle2022-10-191-2/+2
|\ | | | | | | Signed-off-by: Mark Hatle <mark.hatle@kernel.crashing.org>
| * xilinx-standalone.inc: Move to xlnx specific versions of libmetal and open-ampkirkstone-nextMark Hatle2022-10-191-2/+2
| | | | | | | | Signed-off-by: Mark Hatle <mark.hatle@kernel.crashing.org>
* | Merge branch 'kirkstone-next'Mark Hatle2022-09-131-1/+5
|\| | | | | | | Signed-off-by: Mark Hatle <mark.hatle@amd.com>
| * Merge branch 'honister' into kirkstone-nextMark Hatle2022-09-121-1/+5
| |\ | | | | | | | | | Signed-off-by: Mark Hatle <mark.hatle@amd.com>
| | * User can now set XILINX_RELEASE_VERSION to either 'v2022.1' or 'v2022.2'Mark Hatle2022-09-121-1/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This will in-turn select the preferred version for various components to match a given release. It is advise that the user use the latest version whenever possible, but if they are stuck on an older version of Vivado they may need to select the older release. Signed-off-by: Mark Hatle <mark.hatle@amd.com>
| | * various: Move from 2022.1 to 2022.2Mark Hatle2022-02-051-1/+1
| | | | | | | | | | | | Signed-off-by: Mark Hatle <mark.hatle@xilinx.com>
| | * Remove hardknott from LAYERSERIES_COMPATJohn Toomey2022-02-021-1/+1
| | | | | | | | | | | | | | | | | | | | | We do not maintain support for hardknott so it is being removed Signed-off-by: John Toomey <jtoomey@xilinx.com> Signed-off-by: Mark Hatle <mark.hatle@xilinx.com>
| | * Remove gatesgarth from LAYERSERIES_COMPATJohn Toomey2022-01-311-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | We do not build, test or support gatesgarth in any of our layers so it makes sense to remove this for the various layer.conf files Signed-off-by: John Toomey <jtoomey@xilinx.com> Signed-off-by: Mark Hatle <mark.hatle@xilinx.com>
| | * meta-xilinx-standalone: Remove obsolete machinesMark Hatle2022-01-237-37/+0
| | | | | | | | | | | | | | | | | | | | | These machines are no longer necessary, as the regular "generic" machine with an alterantive DEFAULTTUNE should be used instead. Signed-off-by: Mark Hatle <mark.hatle@xilinx.com>
| | * Revert "microblaze-tc: Remove bigendian variants per tc group"Mark Hatle2022-01-071-5/+180
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This reverts commit a6d0eac8564fd02425b1260e58c552589586cf3d. The toolchain itself has internal expectations of big-endian libraries being available, even when building little-endian components. It uses relative paths that include certain big-endian path locations. Signed-off-by: Mark Hatle <mark.hatle@xilinx.com>
| | * meta-xilinx-standalone: Move from cortexr5f to cortexr5Mark Hatle2021-12-202-2/+2
| | | | | | | | | | | | | | | | | | | | | After moving from Xilinx specific cortex r5 turning file, the 5f is no longer defined. Move to the now standard 'cortexr5' definition. Signed-off-by: Mark Hatle <mark.hatle@xilinx.com>
| | * Remove cortexrm tune, as Yocto Project now has cortex-r tunesMark Hatle2021-12-201-27/+37
| | | | | | | | | | | | | | | | | | | | | armrm override is now replaced by armv7r and armv8r overrides. Adjust recipes to match the new usage. Signed-off-by: Mark Hatle <mark.hatle@xilinx.com>
| | * Move baremetal include to meta-xilinx-standaloneMark Hatle2021-12-201-0/+7
| | | | | | | | | | | | | | | | | | | | | This was missed in the prior move of the components to meta-xilinx-standalone layer. Signed-off-by: Mark Hatle <mark.hatle@xilinx.com>
| | * microblaze-tc: Remove bigendian variants per tc groupMark Hatle2021-12-071-180/+5
| | | | | | | | | | | | | | | | | | | | | | | | The big endian variants as they are not used much anymore. This preserved all of the little endian variants, and saves roughly 90 MiB of space and speeds up the build time considerably. Signed-off-by: Mark Hatle <mark.hatle@xilinx.com>
| | * aarch32-tc.conf: Fix reference to cortex-a9 tune fileMark Hatle2021-12-071-1/+1
| | | | | | | | | | | | Signed-off-by: Mark Hatle <mark.hatle@xilinx.com>
| | * move layer.conf defaults to 2022.1Mark Hatle2021-11-151-1/+1
| | | | | | | | | | | | | | | | | | Signed-off-by: Mark Hatle <mark.hatle@xilinx.com> (cherry picked from commit 77aea7992df133d04a5cef4d40878fbe372bfaa3) Signed-off-by: Mark Hatle <mark.hatle@xilinx.com>
* | | meta-xilinx-standalone: Switch layer to support langdaleMark Hatle2022-09-011-1/+1
|/ / | | | | | | Signed-off-by: Mark Hatle <mark.hatle@kernel.crashing.org>
* | meta-xilinx-standalone: Switch layer to kirkstoneMark Hatle2022-08-311-1/+1
| | | | | | | | Signed-off-by: Mark Hatle <mark.hatle@kernel.crashing.org>
* | Move from 2021.2 to 2022.1Mark Hatle2022-05-021-1/+1
| | | | | | | | | | | | | | | | | | | | Rename recipes from 2021.2 to 2022.1 as necessary. Move internal references to 2022.1 (layer.conf and local.conf.sample) Move kernel to 2022.1 branch, as well as various kernel modules. Signed-off-by: Mark Hatle <mark.hatle@kernel.crashing.org>
* | Remove hardknott from LAYERSERIES_COMPATJohn Toomey2022-02-041-1/+1
| | | | | | | | | | | | | | We do not maintain support for hardknott so it is being removed Signed-off-by: John Toomey <jtoomey@xilinx.com> Signed-off-by: Mark Hatle <mark.hatle@xilinx.com>
generated by cgit v1.2.3-54-g00ecf (git 2.39.0) at 2025-07-20 00:40:22 +0000