[sensei-hacker]

Summary

On F7 targets, scheduler() is FAST_CODE, and LTO inlines afatfs_poll() into it through taskRunRealtimeCallbacks(). The SD-card state machine therefore takes ITCM_RAM although it only runs on the idle path and waits on a card that is far slower than the code.

Design

afatfs_poll() is marked NOINLINE, so it stays in flash behind one call from scheduler(). NOINLINE expands to nothing on targets without ITCM_RAM, so only F7 and H7 change. Behaviour is unchanged; this only affects code placement.

This matters now because KAKUTEF7 has about 150 B of ITCM_RAM left, and data-ready reads on F7 (#12065) add about 190 B to the FAST_CODE functions and overflow it by 40 B.

Testing

Built with stock linker scripts, compared with the parent commit:

Target	ITCM_RAM before	ITCM_RAM after	FLASH1
KAKUTEF7	16232 B	15288 B (-944)	-84 B
MATEKF722SE	12360 B	11432 B (-928)	-80 B

MATEKF405, MATEKH743 and SITL also build. The firmware unit tests pass (620/620).

Not tested: SD-card logging on hardware, in SITL or otherwise. The change does not alter control flow, but nobody has run the SD path with it.

Code review

Reviewed with the inav-code-review agent: no critical or important findings.

[qodo-code-review]

ⓘ Qodo reviews are paused because the subscription is no longer active. Ask your workspace admin to reactivate the subscription to resume reviews. Manage billing

[qodo-free-for-open-source-projects]

Code Review by Qodo

🐞 Bugs (0) 📘 Rule violations (0) 📎 Requirement gaps (0)

Great, no issues found!

Qodo reviewed your code and found no material issues that require review

Tip of the day

💡 Did you know, you can move or hide resolved findings to focus on what still needs action

More tips ↗ | Customize Qodo ↗ | Qodo docs ↗

[qodo-free-for-open-source-projects]

PR Summary by Qodo

Keep SD-card polling out of scheduler ITCM on F7 and H7

✨ Enhancement 🕐 Less than 10 minutes

AI Description

• Prevent SD-card polling from being inlined into the FAST_CODE scheduler on F7 and H7.
• Keep the polling code in flash to free ITCM_RAM without changing filesystem behavior.

Diagram

graph TD
    scheduler["FAST_CODE scheduler"] --> callbacks["Realtime callbacks"] --> poll["Filesystem polling"] --> sdcard["SD card work"]
    scheduler --> itcm[("ITCM RAM")]
    poll --> flash[("Flash code")]

Loading

High-Level Assessment

Keep the targeted NOINLINE annotation: it prevents this large, idle-path function from consuming scheduler ITCM without constraining other realtime callbacks. Marking the entire callback function NOINLINE would impose a broader inlining restriction.

Files changed (1) +3 / -1

Enhancement (1) +3 / -1

asyncfatfs.cPrevent afatfs_poll from inlining into the scheduler +3/-1

Prevent afatfs_poll from inlining into the scheduler

• Adds NOINLINE to afatfs_poll and documents the ITCM rationale. The macro applies the attribute on F7 and H7 and expands to nothing on other targets.

src/main/io/asyncfatfs/asyncfatfs.c

[github-actions]

RAM / Flash usage vs. base commit 01a7aa3 — commit f7feede

Using the nearest available size baseline: this build contains base commit 29ec86e, which has no stored baseline yet, so the deltas also include what was merged to the base since 01a7aa3.

Target	Flash Δ	RAM Δ
MATEKF405	±0 B (±0.00%)	CCM: ±0 B (±0.00%) RAM: ±0 B (±0.00%)
MATEKF722	-80 B (-0.02%)	ITCM_RAM: -928 B (-7.40%) RAM: ±0 B (±0.00%) TCM: ±0 B (±0.00%)
MATEKF765	-72 B (-0.01%)	DTCM_RAM: ±0 B (±0.00%) SRAM1: ±0 B (±0.00%)
MATEKH743	+8 B (+0.00%)	D2_RAM: ±0 B (±0.00%) DTCM_RAM: ±0 B (±0.00%) ITCM_RAM: -952 B (-5.80%) RAM: ±0 B (±0.00%)

See RAM/flash optimization guide for techniques to reduce usage.

[github-actions]

Test firmware build ready — commit f7feede

Download firmware for PR #12169

251 targets built. Find your board's .hex file by name on that page (e.g. MATEKF405SE.hex). Files are individually downloadable — no GitHub login required.

Development build for testing only. Use Full Chip Erase when flashing.