You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@nuttx.apache.org by "Brijesh-Thakkar (via GitHub)" <gi...@apache.org> on 2026/10/17 18:27:02 UTC
[PR] arch/rp2040: fail build when boot2 cannot be generated [nuttx]
Brijesh-Thakkar opened a new pull request, #20730:
URL: https://github.com/apache/nuttx/pull/20730
*Note: Please adhere to [Contributing Guidelines](https://github.com/apache/nuttx/blob/master/CONTRIBUTING.md).*
## Summary
Fixes #17516.
Flash builds for RP2040 (`CONFIG_RP2040_FLASH_BOOT=y`) need the 256-byte
second stage bootloader (boot2), which is built from the pico-sdk sources.
Since cf72133d3c ("rp2040: Continue to build even if PICO_SDK_PATH is not
set", #2887), the Make build silently skips boot2 when `PICO_SDK_PATH` is
not set:
- `arch/arm/src/rp2040/Make.defs` skips `include chip/boot2/Make.defs`, so
`rp2040_boot_stage2.S` is never built and the `.boot2` section is empty.
- The vector table then lands at `0x10000000`, where the BOOTROM expects
boot2. The boot2 CRC check fails and the board falls back to BOOTSEL.
- If `picotool` is in `$PATH`, a `nuttx.uf2` is still produced without any
warning. That's exactly what the reporter hit.
#2887 made this non-fatal because CI had no pico-sdk at the time. CI now
provides `PICO_SDK_PATH` everywhere RP2040 boards are built (Docker image,
`tools/ci/platforms/linux.sh`, `ubuntu.sh`, `windows.ps1`), and the CMake
build already fails when it is unset. So the original reason no longer
applies.
Changes:
- `tools/rp2040/Config.mk`: for flash builds, a PREBUILD check that
`PICO_SDK_PATH` is set and that the boot2 source for
`CONFIG_RP2040_FLASH_CHIP` exists under it. Otherwise it stops with an
error explaining how to fix it. The check runs in PREBUILD rather than at
parse time, because `Config.mk` is also read by `configure.sh` and
`make distclean`, and those must keep working without the SDK.
- `arch/arm/src/rp2040/Make.defs`: include `boot2/Make.defs` unconditionally
for flash builds, so the two paths cannot drift apart again.
- `arch/arm/src/rp2040/boot2/rp2040_boot_stage2.cmake`: add the same
file-exists check. CMake previously only checked that the variable was
set, so a wrong path failed later with an unhelpful message.
- `Documentation/platforms/arm/rp2040/index.rst`: document that
`PICO_SDK_PATH` is required for flash builds.
## Impact
- **Users:** building an RP2040 flash config without the pico-sdk now fails
immediately with a clear message, instead of producing an image that
does not boot. Builds with `PICO_SDK_PATH` set are unchanged.
- **Boards:** all RP2040 boards share this path (raspberrypi-pico,
raspberrypi-pico-w, adafruit-feather-rp2040, adafruit-kb2040,
adafruit-qt-py-rp2040, pimoroni-tiny2040, seeed-xiao-rp2040,
w5500-evb-pico, waveshare-rp2040-lcd-1.28, waveshare-rp2040-zero).
rp23xx / rp23xx-rv are not affected: RP2350 has no boot2.
- **SRAM boot configs** (`CONFIG_RP2040_FLASH_BOOT=n`, e.g. `nshsram`):
not affected, still build without the SDK.
- **`configure.sh` / `make distclean`:** still work without the SDK.
- **CI:** no change expected, since CI already exports `PICO_SDK_PATH`.
- **Compatibility:** anyone who built flash configs without the SDK was
getting a non-booting image, so no working setup breaks.
- **Security:** none.
## Testing
Host: Ubuntu 26.04 LTS (x86_64), arm-none-eabi-gcc 14.2.1 20241119
(14.2.rel1), pico-sdk 2.2.0 (same version as CI), picotool 2.2.0 (built
from source).
"Before" = master at c6e866578e. "After" = this PR. All runs use stock
defconfigs with the Make build unless noted. Images were checked by
reconstructing the flash image from `nuttx.uf2` and verifying the boot2 CRC32
in its first 256 bytes, the same check the BOOTROM performs. No board was
flashed: the bug is in the generated image, so it can be shown at the
image level.
**raspberrypi-pico:usbnsh:**
| Check | Before, SDK unset | Before, SDK 2.2.0 | After, SDK unset | After, SDK 2.2.0 |
|---|---|---|---|---|
| build | succeeds, no warning | succeeds | **fails early with error** | succeeds |
| nuttx.uf2 | produced (659 blocks) | produced (660) | not produced | produced (660) |
| UF2 family ID | 0xE48BFF57 | 0xE48BFF56 | — | 0xE48BFF56 |
| boot2 CRC32 | **FAIL** (stored 0xC302E7FE, calc 0x7CD56999) | OK (0xD58F0B07) | — | OK (0xD58F0B07) |
| at 0x10000000 | vector table | boot2 code | — | boot2 code |
**Other cases (after this PR):**
| Case | Result |
|---|---|
| `raspberrypi-pico:usbnsh`, `PICO_SDK_PATH` = non-existent path | fails early: boot2 source not found |
| `raspberrypi-pico:usbnsh`, `PICO_SDK_PATH` = existing dir that is not an SDK | fails early: boot2 source not found |
| `raspberrypi-pico:nshsram`, no SDK | builds (no boot2, as expected) |
| `raspberrypi-pico-w:telnet`, SDK set | builds, boot2 CRC OK |
| `adafruit-feather-rp2040:nsh`, SDK set | builds, boot2 CRC OK |
| `configure.sh` + `make distclean`, no SDK | work |
**Error output (Make, SDK unset):**
```
$ unset PICO_SDK_PATH
$ make distclean && ./tools/configure.sh -l raspberrypi-pico:usbnsh && make
...
Loaded configuration '.config'
Configuration saved to '.config'
tools/Unix.mk:535: *** "PICO_SDK_PATH is not set, it is needed to build the boot2 stage, the image will not boot without it. Run: export PICO_SDK_PATH=<path to pico-sdk>". Stop.
```
**CMake (`cmake -B build -DBOARD_CONFIG=raspberrypi-pico:usbnsh`):**
- SDK unset: configure fails (existing check, unchanged).
- Bogus path: configure fails with the new boot2-source-not-found message.
- SDK set: builds, `nuttx.uf2` has family 0xE48BFF56, boot2 CRC OK.
**checkpatch:**
```
$ ./tools/checkpatch.sh -g HEAD~..HEAD
✔️ All checks pass.
```
## Notes
- The pico-w build has a fallback that writes a dummy CYW43 firmware file to
`/tmp` when `PICO_SDK_PATH` is unset. For flash configs this is now
unreachable, because the build stops earlier. That's fine, since those
configs could not produce a bootable image anyway. It could be cleaned
up in a follow-up.
- Possible follow-up: an `ASSERT` in the RP2040 flash linker scripts that
`.boot2` is exactly 256 bytes. That would catch an empty boot2 regardless
of how the build got there. I kept it out of this PR to keep the change
small.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] arch/rp2040: fail build when boot2 cannot be generated [nuttx]
Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on PR #20730:
URL: https://github.com/apache/nuttx/pull/20730#issuecomment-6238657717
<!-- membrowse-pr-comment -->
## <img src="https://membrowse.com/membrowse-logo.svg" height="24" align="top"> MemBrowse Memory Report
*No memory changes detected for:*
- [arduino-mega2560](https://membrowse.com/public/apache/nuttx/target/arduino-mega2560/compare?head=052b4b9e2751330f84bd05fc0f222b13c0a3d031&base=17747ded0aaba11af1dd07422c15354200e65e3e)
- [esp32-devkitc](https://membrowse.com/public/apache/nuttx/target/esp32-devkitc/compare?head=052b4b9e2751330f84bd05fc0f222b13c0a3d031&base=17747ded0aaba11af1dd07422c15354200e65e3e)
- [hifive1-revb](https://membrowse.com/public/apache/nuttx/target/hifive1-revb/compare?head=052b4b9e2751330f84bd05fc0f222b13c0a3d031&base=17747ded0aaba11af1dd07422c15354200e65e3e)
- [mirtoo](https://membrowse.com/public/apache/nuttx/target/mirtoo/compare?head=052b4b9e2751330f84bd05fc0f222b13c0a3d031&base=17747ded0aaba11af1dd07422c15354200e65e3e)
- [qemu-armv8a](https://membrowse.com/public/apache/nuttx/target/qemu-armv8a/compare?head=052b4b9e2751330f84bd05fc0f222b13c0a3d031&base=17747ded0aaba11af1dd07422c15354200e65e3e)
- [qemu-intel64](https://membrowse.com/public/apache/nuttx/target/qemu-intel64/compare?head=052b4b9e2751330f84bd05fc0f222b13c0a3d031&base=17747ded0aaba11af1dd07422c15354200e65e3e)
- [rx65n-rsk2mb](https://membrowse.com/public/apache/nuttx/target/rx65n-rsk2mb/compare?head=052b4b9e2751330f84bd05fc0f222b13c0a3d031&base=17747ded0aaba11af1dd07422c15354200e65e3e)
- [s698pm-dkit](https://membrowse.com/public/apache/nuttx/target/s698pm-dkit/compare?head=052b4b9e2751330f84bd05fc0f222b13c0a3d031&base=17747ded0aaba11af1dd07422c15354200e65e3e)
- [stm32-nucleo-f103rb](https://membrowse.com/public/apache/nuttx/target/stm32-nucleo-f103rb/compare?head=052b4b9e2751330f84bd05fc0f222b13c0a3d031&base=17747ded0aaba11af1dd07422c15354200e65e3e)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] arch/rp2040: fail build when boot2 cannot be generated [nuttx]
Posted by "Brijesh-Thakkar (via GitHub)" <gi...@apache.org>.
Brijesh-Thakkar commented on PR #20730:
URL: https://github.com/apache/nuttx/pull/20730#issuecomment-6248983160
@raiden00pl @masayuki2009 @linguini1 Please review this PR
All ci are passing and @xiaoxiang781216 already approved this PR
so we can get this merged
Thank youu
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] arch/rp2040: fail build when boot2 cannot be generated [nuttx]
Posted by "linguini1 (via GitHub)" <gi...@apache.org>.
linguini1 commented on PR #20730:
URL: https://github.com/apache/nuttx/pull/20730#issuecomment-6249612082
> I didn't know that I shouldn't ping maintainers directly. Sorry for that, I'll make sure to request reviews through the reviewers section from now on.
No problem! I should mention I can't speak for everybody, some people might be alright with pings (and it's good with me when you're responding in a direct conversation). It's just because there are only a few active maintainers so they usually get _a lot_ of inbox messages. Thanks again for the patch :)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] arch/rp2040: fail build when boot2 cannot be generated [nuttx]
Posted by "Brijesh-Thakkar (via GitHub)" <gi...@apache.org>.
Brijesh-Thakkar commented on PR #20730:
URL: https://github.com/apache/nuttx/pull/20730#issuecomment-6249195303
I didn't know that I shouldn't ping maintainers directly. Sorry for that, I'll make sure to request reviews through the reviewers section from now on.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] arch/rp2040: fail build when boot2 cannot be generated [nuttx]
Posted by "linguini1 (via GitHub)" <gi...@apache.org>.
linguini1 commented on PR #20730:
URL: https://github.com/apache/nuttx/pull/20730#issuecomment-6249151198
Hi @Brijesh-Thakkar,
Thank you for the patch and I appreciate your enthusiasm! However, please do not ping maintainers to review your PRs, just request their review under the sidebar widget titled "reviewers". There are many new PRs per day, and pings really add up in our inboxes. CI passing and 1 review is not enough to merge, NuttX requires two reviewers to approve and merge. Your PR is only about 16 hours old, please have a little patience :)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] arch/rp2040: fail build when boot2 cannot be generated [nuttx]
Posted by "acassis (via GitHub)" <gi...@apache.org>.
acassis merged PR #20730:
URL: https://github.com/apache/nuttx/pull/20730
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org