arch/tricore: add support for tc4xx series chips and boards by zhangyu-duck · Pull Request #18391 · apache/nuttx

zhangyu-duck · 2026-02-13T08:11:26Z

Note: Please adhere to Contributing Guidelines.

Summary

Two main updates:

Uses the official Infineon ILLD repository instead of a private repository.
Adds support for tc4xx series chips and boards

Impact

support tc4xx chip

Testing

Both cmake and makefile methods can compile tc3xx/tc4xx.
pass ostest, basic functions are normal

nsh> hello
Hello, World!! tc4 ostest
nsh>

i give tc4 ostest.log in attach files
tc4_ostest.log

arch/tricore/src/illd/tc3xx/Configurations/Ifx_Cfg_Ssw.c

jerpelea

please document all non Apache license file to
https://github.com/apache/nuttx/blob/master/LICENSE

jerpelea

please add SPDX license identifiers to all files not only Apache licensed ones

zhangyu-duck · 2026-02-13T11:45:37Z

please add SPDX license identifiers to all files not only Apache licensed ones

ok，thanks, i have updata code

zhangyu-duck · 2026-02-13T11:45:54Z

please add SPDX license identifiers to all files not only Apache licensed ones

get it, thanks

acassis

@zhangyu-duck please update the Documentation https://nuttx.apache.org/docs/latest/platforms/tricore/index.html to include these new chips.

Unfortunately our Documentation is very non-standardized, some Documentation arch are very good, others are very primitive (let say not helpful). So, I suggest you to use https://nuttx.apache.org/docs/latest/platforms/arm/imxrt/index.html as a good reference.

arch/tricore/include/tc3xx/irq.h

arch/tricore/src/illd/tc4xx/Libraries/doc/iLLD-TC4xx-Releasenotes.pdf

linguini1

I guess since this modifies the TC3xx logic as well, could you please include an OSTest log from a tc3 board? The tc4 log looks good.

arch/tricore/src/illd/tc4xx/Libraries/doc/iLLD-TC4xx-UM.chm

arch/tricore/src/tc3xx/Kconfig

arch/tricore/src/tc4da/CMakeLists.txt

arch/tricore/src/tc4da/Make.defs

arch/tricore/src/tc4xx/Kconfig

arch/tricore/src/tc4xx/tc4xx_libc.c

arch/tricore/src/tc4xx/tc4xx_dummy.c

linguini1 · 2026-02-13T18:10:10Z

Also I won't block the PR about this since I have pretty much no familiarity with the tricore architecture, but is it necessary to use the SDK? It seems like this PR needs to include a lot of workarounds for quirks with the SDK (like the libc stubs and patch files). Is it at all possible to avoid using the SDK directly? I know a few other archs in NuttX use this approach (hence why I don't think it's a dealbreaker or anything), but I also know the preferred method is to avoid using vendor SDKs/HALs entirely if possible.

acassis · 2026-02-14T11:47:45Z

@simbit18 do you know if LVGL v9.2 files were removed from their site?

====================================================================================
Configuration/Tool: raspberrypi-4b/lvgl
2026-02-14 04:31:13
------------------------------------------------------------------------------------
  Cleaning...
  Configuring...
  Building NuttX...
[v9.2.1.zip]
  End-of-central-directory signature not found.  Either this file is not
  a zipfile, or it constitutes one disk of a multi-part archive.  In the
  latter case the central directory and zipfile comment will be found on
  the last disk(s) of this archive.
unzip:  cannot find zipfile directory in one of v9.2.1.zip or
        v9.2.1.zip.zip, and cannot find v9.2.1.zip.ZIP, period.
make[3]: *** [Makefile:77: lvgl] Error 9
make[3]: Target 'context' not remade because of errors.
make[2]: *** [Makefile:57: /github/workspace/sources/apps/graphics/lvgl_context] Error 2
make[2]: Target 'context_all' not remade because of errors.
make[1]: *** [Makefile:187: context] Error 2
make: *** [tools/Unix.mk:458: /github/workspace/sources/apps/.context] Error 2
make: Target 'all' not remade because of errors.
/github/workspace/sources/nuttx/tools/testbuild.sh: line 385: /github/workspace/sources/nuttx/../nuttx/nuttx.manifest: No such file or directory
  [1/1] Normalize raspberrypi-4b/lvgl

We need to create our own mirror at github.com/nuttx/

parthitce · 2026-02-14T12:14:58Z

Also I won't block the PR about this since I have pretty much no familiarity with the tricore architecture, but is it necessary to use the SDK? It seems like this PR needs to include a lot of workarounds for quirks with the SDK (like the libc stubs and patch files). Is it at all possible to avoid using the SDK directly? I know a few other archs in NuttX use this approach (hence why I don't think it's a dealbreaker or anything), but I also know the preferred method is to avoid using vendor SDKs/HALs entirely if possible.

On this note, we are already in the phase of developing iLLD independent Tricore support. Although this isn't mature yet, plans are tracked here 1. Also you can find development tree with various branches 2 named dev/infineon/tricore-xxyy, for example working variant of UART 3. Our main goal is get to upstream NuttX during the course of this year.

linguini1 · 2026-02-14T14:46:37Z

On this note, we are already in the phase of developing iLLD independent Tricore support. Although this isn't mature yet, plans are tracked here 1. Also you can find development tree with various branches 2 named dev/infineon/tricore-xxyy, for example working variant of UART 3. Our main goal is get to upstream NuttX during the course of this year.

Woah awesome news! Thanks!

simbit18 · 2026-02-15T12:45:55Z

@simbit18 do you know if LVGL v9.2 files were removed from their site?

Hi @acassis There were temporary issues downloading the package.

acassis · 2026-02-15T13:01:14Z

@simbit18 do you know if LVGL v9.2 files were removed from their site?

Hi @acassis There were temporary issues downloading the package.

Thank you @simbit18

parthitce · 2026-02-17T15:57:21Z

@zhangyu-duck are these changes tested with Tasking only or OSS GCC can also be used?

zhangyu-duck · 2026-02-24T02:11:24Z

@zhangyu-duck are these changes tested with Tasking only or OSS GCC can also be used?
only gcc now, and for tasking we will pull up in subsequent work later.

arch/tricore/Kconfig

xiaoxiang781216 · 2026-02-24T02:57:35Z

arch/tricore/Kconfig

 	--help--
 		In addition to the 6 cores, tc4xx also has one more security core.

+config CPU_COREID


why need this option

chip’s code and linker script needs to be distinguished for different cores. Originally, this was under vendor, and I placed it in the kconfig under the chip for tc4xx/tc3xx in the community, but someone suggested extracting it to the common section. so i place it in arch/tricore/Kconfig

I would prefer remove it from Kconifg, since:

you can get the coreid at runtime directly

different core has the different defconfig, you can select the different link script directly.

boards/tricore/tc397/a2g-tc397-5v-tft/include/board_memorymap.h

xiaoxiang781216 · 2026-02-24T03:00:28Z

arch/tricore/src/tc4xx/tc4xx_libc.c

@@ -0,0 +1,86 @@
+/****************************************************************************
+ * arch/tricore/src/tc4xx/tc4xx_libc.c


should we move to libc/machine/tricore instead

i think noneed.

/* The implementation of libc is introduced by default in the Tricore

toolchain, in nuttx we made a fake libc_fpu.a library to bypass

this issue, but the linker will still generate markup code,

add a few definitions to fool the linker.
*/

how about protected mode, the code in arch can't be used by nuttx_user.bin

xiaoxiang781216 · 2026-02-24T03:14:42Z

arch/tricore/src/illd/tc3xx/Configurations/Ifx_Cfg_Ssw.c

+#endif
+
+#if (IFX_CFG_SSW_ENABLE_LBIST == 1)
+void Ifx_Ssw_Lbist(void)


why not use illd version directly?

This involves copyright issues, so I extracted the key and necessary parts and rewrote them so that they can be used with the SPDX license.

I mean using the code from illd package directly, and patch the origin file is real need

wangchdo · 2026-02-25T03:00:13Z

arch/tricore/src/common/tricore_irq.c

  for (i = 0; i < 6; i++)
    {
-#ifdef CONFIG_ARCH_TC3XX
+#ifdef CONFIG_ARCH_CHIP_TC3XX


why change CONFIG_ARCH_TC3XX to CONFIG_ARCH_CHIP_TC3XX?

TC3XX is not chip name, It should be like ARMV8R: ARCH_ARMV8R