Research window: Past 24 hours (2026-09-29 11:00 ~ 2026-09-30 11:00 Beijing Time) Sources: RT-Thread official news API, community forum article interface, GitHub (org: RT-Thread main repo commits / PR / Issue / Releases, env tool repo, package index, device support packages and RT-Thread Studio component organization, vendor downstream organizations, topic:rt-thread ecosystem search), Google News aggregation (Chinese and English, via proxy), WeChat official account channel verification

In This Issue

  • I. Kernel and Version Updates
    • Next RISC-V architecture refactoring proposal gains founder merge momentum, submitter rebases and refreshes overnight (09-29 ~ 09-30)
    • Kernel calloc overflow fix proposal passes review (09-29 ~ 09-30)
    • Three driver and BSP fix proposals submitted in batch: ofw hash out-of-bounds, ESP32-C6 SPI, ESP32-C3 linker script (09-29)
    • Official news API sees no new entries for the 41st consecutive day (09-30)
  • II. Official and Community Events
    • No new content on forums or public channels within the window (09-30)
  • III. Ecosystem and Vendor Adaptation
    • RT-Thread Studio Infineon OpenOCD debugger component upgraded to 5.50.0.3639 (09-29)
    • No new pushes from vendors or package side within the window (09-30)
  • IV. Industry Applications
    • Universiti Sains Malaysia capstone project update: RA8P1 + RT-Thread edge acoustic event detection firmware validation (09-30)
  • V. Summary and Trend Observations
    • Recent one-day trends (09-29 ~ 09-30)
    • Limitations and deduplication notes
  • Appendix: Source Verification Table

I. Kernel and Version Updates

1.1 Next RISC-V Architecture Refactor Proposal Gains Founder’s Merge Push, Submitter Rebases and Refreshes Overnight (09-29 ~ 09-30)

Date: 2026-09-29 21:09 (founder’s comment) ~ 09-30 00:01 (latest commit)　Source: PR #10320

The long-stalled main-repo RISC-V architecture refactor proposal “Next RISC-V” (#10320, 171 files, +15018/-3490) received a clear push forward within the window: founder Bernard Xiong commented at 09-29 21:09, “This PR should be mergeable soon; meanwhile, try enabling device tree on some RISC-V SoCs as well,” and named maintainers and the submitter to follow up. Subsequently, submitter GuEe-GUI triggered three consecutive refresh cycles for the proposal from that evening into the early hours of the next day (21:22, 23:54, 00:01, with the formatting and memory-comparison pipelines rerunning accordingly), and the latest head commit “Refactor the RISC-V architecture port” was rebased at 09-30 00:01 onto the current mainline head dc8aaa73f. Key points of the proposal: ISA management interfaces, RV32E/32I/64I unification, CBOM-based cache handling, Sv32~Sv64 MMU and SVPBMT, merged M/S-mode CSR operations, SMP, AIA (APLIC/IMSIC), CLIC, ACLINT/CLINT, PLIC, device tree (DM) support, T-HEAD vendor code merge, qemu-virt64-riscv64 device tree port, and booting from a non-zero core. The plan checklist shows S-mode RT-Thread, RT-Smart, and the device tree path have been verified, while remaining combinations (including M-mode and no-MMU) are still in progress. Since its submission in 2026-05, the proposal has accumulated 74 review comments.

1.2 Kernel calloc Overflow Fix Proposal Passes Review (09-29 ~ 09-30)

Date: 2026-09-29 22:34 (submitted) ~ 09-30 08:30 (LGTM received)　Source: PR #11848

Contributor luchenxu73 submitted kernel fix #11848 (single commit, 3 files, +57/-3): rt_calloc does not check whether count×size overflows during allocation; when the limit is exceeded the result wraps around, creating an out-of-bounds risk from inconsistent allocation and zeroing. The fix precomputes the allocation size and rejects allocation when the limit is exceeded, touching include/rtdef.h, src/kservice.c, and the memory unit test src/utest/mem_tc.c (new test case), verified on qemu-vexpress-a9. The proposal received a community review “LGTM” at 08:30 on the morning of 09-30, making it the fastest-advancing new proposal within the window.

1.3 Three Driver and BSP Fix Proposals Submitted Together: ofw Hash Out-of-Bounds, ESP32-C6 SPI, ESP32-C3 Linker Script (09-29)

Date: 2026-09-29 20:11 ~ 21:53　Source: PR #11845, PR #11846, PR #11847

Within roughly an hour and a half on the evening of 09-29, the community submitted three fix proposals in quick succession, all with test evidence, still pending review. The first (#11845) fixes an out-of-bounds access in the device tree (ofw) phandle hash table: the number of allocated entries uses the difference of the range, while the lookup index is computed from the difference between the phandle and the start value, causing an out-of-bounds access when the upper bound of the range is accessed; during expansion, the zeroing position also fails to subtract the start value. The fix adds one to the entry count and, during expansion, zeroes the newly added region starting from the end of the old table. The author completed function-level verification of 6 test cases with GCC plus AddressSanitizer/UBSan (start phandle values of 1, 5, and 128, plus a test expanding the upper bound of 128 to 256). The second (#11846) enables on-demand compilation of SPI for Espressif ESP32-C6: after selecting BSP_USING_SPI, it hooks into the IDF SPI master, fixes issues where the shared ESP driver did not return the number of transferred bytes and the chip-select pin value, and returns -RT_EINVAL when the configuration is empty (6 files, +466/-19). The third (#11847) fixes wildcards in the ESP32-C3 linker script under Windows paths: wildcards written with “/” in sections.ld fail to match the riscv directory object files in Windows target paths, causing vector jump relocations to exceed the limit; switching to “?” makes it compatible with both separators. This is the same change as #11816, redone this time on the latest mainline, with scons build passing under Windows with GCC 11.2.0.

1.4 Official News API Sees No New Entries for the Forty-First Consecutive Day (09-30)

Date: 2026-09-30 (check time)　Source: RT-Thread official news API

The latest entry in the archive returned by the official news API (rt_org category) is still dated 2026-06-17, with no new entries within the window, marking the forty-first consecutive day without updates; product and event information continues to be first released via the WeChat official account, community forum, and GitHub channels.

II. Official and Community Activities

2.1 No New Content on the Forum or Public Channels Within the Window (09-30)

The community forum article API shows no new articles within the window (the latest remains the morning of 09-29’s “MCU Linker Script ‘Upgrade’,” already covered in yesterday’s daily report); no new posts were seen on public aggregation channels, and the most recent valid coverage remains the 09-28 article on the “virtual-physical integrated logistics factory” teaching platform. Official external information distribution maintains the pattern of “news API paused, community channels first.”

III. Ecosystem and Vendor Adaptation

3.1 RT-Thread Studio Infineon OpenOCD Debugger Component Upgraded to 5.50.0.3639 (09-29)

Date: 2026-09-29 14:35 (merge time)　Source: sdk-debugger-openocd-infineon PR #2

RT-Thread Studio’s SDK component repository sdk-debugger-openocd-infineon (the OpenOCD debugger package for Infineon and former Cypress boards) merged a version update within the window (#2): the bundled OpenOCD was upgraded to 5.50.0.3639 (committed on 09-26). The repository was created on 08-31 and completed its first merge round on 09-01; this is the second merge, with the Studio-side debugger component continuing to track upstream versions.

3.2 No New Pushes on the Vendor and Package Side Within the Window (09-30)

The package index (RT-Thread/packages, latest 09-24), the env tool repository (09-28), the GD32 downstream repository (09-28), and the device support package organization (latest 09-24) all saw no new pushes within the window; the esp-idf package has remained without pushes since late 2023.

IV. Industry Applications

4.1 Universiti Sains Malaysia Final-Year Project Update: RA8P1 + RT-Thread Edge Acoustic Event Detection Firmware Verification (09-30)

Date: 2026-09-30 05:49 ~ 05:50 (commit time)　Source: KhaiFaw/ai-acoustic-event-detection

The third-party project under the topic:rt-thread ecosystem, “Acoustic event detection at the edge” (a mechatronics engineering final-year project at Universiti Sains Malaysia), submitted two consecutive updates within the window: reconciling the acoustic firmware build with the model contract, fixing an input scaling inconsistency between the embedded model and the firmware, and adding software verification records (16 host tests and one full BSP local build passed; the corrected firmware has not yet been evaluated via on-device flashing). The project is based on the Renesas RA8P1 Titan board: RT-Thread provides the OS framework, Renesas FSP provides device support, CMSIS-DSP implements the FFT, and TensorFlow Lite Micro runs the INT8 CNN; a MAX9814 microphone captures 16000 audio samples, which after MFCC feature extraction are used to classify six sound categories on-device: baby crying, glass breaking, alarm/siren, knocking, doorbell, and background.

V. Summary and Trend Observations

5.1 Trends Over the Past Day (09-29 ~ 09-30)

  • A clear merge signal has emerged on the RISC-V architecture line: the founder personally pushed #10320 to “merge soon,” and the submitter refreshed it three times overnight and rebased onto the latest mainline—if this refactor (171 files) is merged, it will mark a systematic generational change for RISC-V support.
  • Kernel memory safety fixes submitted intensively: the calloc overflow fix (#11848) received LGTM the day after submission, and fixes such as the ofw hash out-of-bounds (#11845) all come with reproducible tests; the direction of fixes centers on boundary and overflow checks.
  • Espressif-family BSPs enter the post-merge cleanup phase: C6 SPI (#11846) and C3 linker script (#11847) are being advanced in succession by the same contributor, forming a complete sequence with the 09-28 C6 BSP merge and S3 BSP proposal.
  • Mainline merge gap and proposal backlog: master has seen no new merges since 09-28 22:22, with all activity within the window concentrated on the proposal side; attention is recommended on the follow-up merge actions for #10320 and #11848.
  • Official channels remain quiet: the official news API has entered its forty-first day without updates; the forum and aggregation channels saw no new content within the window.

5.2 Limitations and Deduplication Notes

  • Deduplication: #11844 (ESP32-S3 BSP, already covered on 09-29, no updates within the window), #11834 (vsnprintf fix, already covered on 09-25; a revision push was completed at 09-29 21:53 within the window, still under review), and #11816 (no updates) are not reported again; #11797, #11841, #11132, and several already-merged older proposals only triggered automated rechecks (formatting and memory-comparison bots) within the window, with no substantive progress.
  • Excluded: zlgopen/awtk (general GUI documentation-type commits) and h120d-protocol (readme updates, no substantive connection to RT-Thread) matched by topic:rt-thread are not included.
  • Verification: Google News shows no new valid coverage in either Chinese or English over the 3-day window; matched industry annual conference promotional content and old articles are not included.

Appendix: Source Verification Table

Source Verification Result
RT-Thread official news API Latest entry 2026-06-17, no new items in window, stalled for the forty-first consecutive day
Official website news list page Still front-end dynamically rendered, cannot be directly verified (data same as official news API)
Community forum article API No new articles in window; latest remains the 09-29 morning article (already included)
WeChat official account and aggregation channels No new posts observed (verified via aggregation channels)
GitHub org: RT-Thread (main repo commits / PR / Issue) No new merges in main repo within window (last was 09-28 22:22); 4 new PRs opened (#11845 ~ #11848); #10320 refreshed three times within window, receiving merge push from founder; #11848 passed review; no non-PR updates among Issues within window
GitHub RT-Thread/env No new pushes in window (latest 09-28)
GitHub RT-Thread/packages No new pushes in window (latest 09-24)
GitHub RT-Thread-Studio component org sdk-debugger-openocd-infineon merged #2 (see 3.1); no pushes in other repos within window
GitHub device support package org and vendor downstream (GD32-MCU-IOT, etc.) No new pushes in window (latest 09-24 / 09-28)
GitHub topic:rt-thread (pushes in the past week) Within window: ai-acoustic-event-detection (see 4.1); awtk, h120d-protocol not included (see 5.2)
GitHub releases / tags (main repo) No new releases in window, latest remains v5.3.0 from 09-10
Google News aggregation (Chinese and English, via proxy) No new valid coverage (hits are industry annual conference promotions and old articles)

Complete Source List