boards/nucleo-h743zi: Add nxboot bootloader support

[neilberkman]

Summary

Add nxboot bootloader board integration for nucleo-h743zi.

nxboot uses a three-slot layout (primary, secondary, tertiary) with no scratch partition.

• add STM32_APP_FORMAT_NXBOOT board support and nxboot-specific OTA slot defaults
• gate scratch partition configuration on MCUboot, where it applies
• select nxboot linker scripts in the board build logic
• add tertiary OTA partition support in stm32_progmem.c
• add nxboot-loader and nxboot-app defconfigs and linker scripts
• validate the fixed nxboot flash layout used by this board so mismatched config values fail at build time instead of producing a silently broken image

Impact

• affects only boards/arm/stm32h7/nucleo-h743zi
• no expected impact on non-nxboot board configurations
• no expected impact on runtime behavior outside the new nxboot configurations
• invalid nxboot layout settings now fail at build time instead of generating a bad image layout

Testing

Host machine:

• macOS 26.3.1 (arm64)
• apps tree: ../nuttx-apps-wt

Board/builds tested:

1. nucleo-h743zi:nxboot-loader

   • make distclean
   • ./tools/configure.sh -a ../nuttx-apps-wt nucleo-h743zi:nxboot-loader
   • make -j8
   • result: build succeeded
   • log excerpt:
     • CPP: .../flash-nxboot-loader.ld -> .../flash-nxboot-loader.ld.tmp
     • LD: nuttx

2. nucleo-h743zi:nxboot-app

   • make distclean
   • ./tools/configure.sh -a ../nuttx-apps-wt nucleo-h743zi:nxboot-app
   • make -j8
   • result: build succeeded
   • log excerpt:
     • CPP: .../flash-nxboot-app.ld -> .../flash-nxboot-app.ld.tmp
     • LD: nuttx

Negative validation:
3. changed CONFIG_NXBOOT_HEADER_SIZE from 0x400 to 0x200

• make olddefconfig
• make -j8
• result: build failed during linker script preprocessing as expected
• log excerpt:
  • flash-nxboot-app.ld:117:4: error: #error "CONFIG_NXBOOT_HEADER_SIZE must be 0x400 for flash-nxboot-app.ld"

4. restored CONFIG_NXBOOT_HEADER_SIZE=0x400 and rebuilt nucleo-h743zi:nxboot-app
   • result: build succeeded

Hardware testing:

• not yet tested on target hardware

[linguini1]

Hi @neilberkman , thanks for the patch!

Please take a look at the contributing guide :) Your PR description is a good summary, but still needs the impact & testing sections from the template.

[neilberkman]

Force-pushed an update to add the missing Signed-off-by on the top commit and fix the cmake-format issue in boards/arm/stm32h7/nucleo-h743zi/src/CMakeLists.txt. The PR description still includes the detailed Impact and Testing sections.

[neilberkman]

CI failure in Linux (arm-13) is due to pre-existing Clang warnings in the nxboot app code (this PR is the first to build nxboot with Clang). Fix submitted: apache/nuttx-apps#3419

[acassis]

@neilberkman nice work, please fix this issue:

Configuration/Tool: nucleo-h743zi/nxboot-loader,CONFIG_ARM_TOOLCHAIN_CLANG
2026-03-08 22:18:56
------------------------------------------------------------------------------------
  Cleaning...
  Configuring...
  Disabling CONFIG_ARM_TOOLCHAIN_GNU_EABI
  Enabling CONFIG_ARM_TOOLCHAIN_CLANG
  Building NuttX...
Error: loader/flash.c:140:54: error: format specifies type 'long' but the argument has type 'off_t' (aka 'int') [-Werror,-Wformat]
  140 |       syslog(LOG_ERR, "Could not seek to %ld: %s\n", off, strerror(errno));
      |                                          ~~~         ^~~
      |                                          %d
Error: loader/flash.c:148:15: error: format specifies type 'long' but the argument has type 'off_t' (aka 'int') [-Werror,-Wformat]
  147 |       syslog(LOG_ERR, "Write to offset %ld failed %s\n",
      |                                        ~~~
      |                                        %d
  148 |               off, strerror(errno));
      |               ^~~
Error: loader/flash.c:198:54: error: format specifies type 'long' but the argument has type 'off_t' (aka 'int') [-Werror,-Wformat]
  198 |       syslog(LOG_ERR, "Could not seek to %ld: %s\n", off, strerror(errno));
      |                                          ~~~         ^~~
      |                                          %d
Error: loader/flash.c:205:59: error: format specifies type 'long' but the argument has type 'off_t' (aka 'int') [-Werror,-Wformat]
  205 |       syslog(LOG_ERR, "Read from offset %ld failed %s\n", off,
      |                                         ~~~               ^~~
      |                                         %d
4 errors generated.
make[2]: *** [/github/workspace/sources/apps/Application.mk:239: loader/flash.c.github.workspace.sources.apps.boot.nxboot.o] Error 1
Error: nxboot_main.c:132:17: error: passing an object that undergoes default argument promotion to 'va_start' has undefined behavior [-Werror,-Wvarargs]
  132 |   va_start(arg, type);
      |                 ^
nxboot_main.c:127:43: note: parameter of type 'enum progress_type_e' is declared here
  127 | void nxboot_progress(enum progress_type_e type, ...)
      |                                           ^
1 error generated.
make[2]: *** [/github/workspace/sources/apps/Application.mk:330: nxboot_main.c.github.workspace.sources.apps.boot.nxboot.o] Error 1
make[2]: Target 'all' not remade because of errors.
make[1]: *** [Makefile:54: /github/workspace/sources/apps/boot/nxboot_all] Error 2
make[1]: Target 'all' not remade because of errors.
make: *** [tools/LibTargets.mk:248: /github/workspace/sources/apps/libapps.a] 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 nucleo-h743zi/nxboot-loader

@neilberkman do you have plans to participate of the GSoC and contribute the two partitions support to NXBoot?

https://cwiki.apache.org/confluence/display/COMDEV/GSoC+2026+Ideas+list

[neilberkman]

Thanks @acassis! These are pre-existing Clang warnings in the nxboot app code. This PR is the first defconfig that builds nxboot with Clang, so they hadn't been caught before. Fix submitted: apache/nuttx-apps#3419. CI on that PR has one failed job (risc-v-05) which died at Docker Login (infra issue, not a build failure). All jobs that actually built passed. Will rebase this PR once that merges.

Re two-partition support for nxboot, would be great to see that happen. I won't be able to take it on myself though.

[linguini1]

not yet tested on target hardware

Does this mean the patch hasn't been tested on the nucleo yet? If so, I think we should hold off on merge until it can be tested.