[MrScothh]

A P frame logs up to 9 optional fields: vbat, amperage, mag x3, baro, pitot, surface and rssi. writeInterframe() collected them in int32_t deltas[8], and blackboxWriteTag8_8SVB() packs them under one 8-bit header. With all 9 present:

• the rssi delta is written one past the end of deltas, on the stack, in every P frame;
• its header bit is shifted out and the value is written only when it isn't zero, while decoders read a run of these fields as groups of 8 and a lone ninth as a plain signed VB that is always there. So every P frame where rssi didn't change throws the decoder off until the next I frame.

All 9 are there with VBAT, a current meter, a mag (logged by default, blackbox MAG), a baro, a pitot (virtual included), a rangefinder (MSP included) and an RSSI source. My T-28 has all of them, and its logs decode about 3 % of iterations: 172,619 frames failed in a 3 minute log. It goes back to when the pitot delta was added, so maintenance-9.x has it too.

The array now holds 9, and the writer splits a run longer than 8 into groups of 8, the way decoders read it, so a lone ninth field is a plain signed VB again. Logs with 8 or fewer of these fields are byte for byte the same, and existing decoders read the new ones without changes.

Until this is in, blackbox -MAG and save (or dropping any other of the nine) avoids it.

Testing

TBS Lucid H7 Wing on my T-28, with all nine present (VBAT, current, an IST8310 logged with blackbox MAG, a DPS310, the virtual pitot, an MSP rangefinder and RSSI), decoded with the current blackbox_decode, unchanged:

	log	P frames decoded	iterations missing
maintenance-10.x	3 min, blackbox_rate_denom 1	3	96.9 %
this PR	40 s, blackbox_rate_denom 2	18,555	0

The header is the same in both, nine fields with this encoding. With 8 or fewer the writer goes through the same code as before, as a single group.

A new unit test checks the bytes written for 0, 1, 3, 8, 9 and 17 fields; the 9 and 17 cases fail on maintenance-10.x.

Flash: MATEKF722SE +0 B, MATEKF405SE +80 B, KAKUTEF7/KAKUTEF7HDV +64 B, MATEKH743 +96 B; ITCM unchanged on KAKUTEF7 and KAKUTEF7HDV.

Logs already recorded with all 9 fields carry no marker for whether the ninth value is there, so a decoder could only recover them by trying each P frame both ways and keeping the reading that ends on the next frame marker. I can look at that in blackbox-tools if it's useful.

[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ⓘ Qodo reviews are paused because the subscription is no longer active.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]

PR Summary by Qodo

Fix Blackbox P-frame encoding with nine optional fields

🐞 Bug fix 🧪 Tests 🕐 20-40 Minutes

AI Description

• Expand P-frame delta storage to prevent a stack write when all nine optional fields are logged.
• Encode fields in decoder-compatible groups, preserving bytes for eight or fewer fields.
• Add byte-level tests for empty, single-field, eight-field, nine-field, and longer runs.

Diagram

graph TD
  A["Optional sensors"] --> B["P-frame deltas"] --> C{"Over eight?"} -->|Yes| D["Eight-field group"] --> E["Tail group"] --> F["Existing decoder"]
  C -->|No| G["Single group"] --> F

Loading

High-Level Assessment

The following are alternative approaches to this PR:

1. Special-case the ninth field in the P-frame writer

• ➕ Limits the encoder change to the immediate nine-field case.
• ➖ Duplicates encoding rules at the call site.
• ➖ Does not handle longer runs through the shared encoder.

Recommendation: Keep the shared encoder's grouping approach: it matches how existing decoders read the format, retains the old output for eight or fewer fields, and handles longer runs without P-frame-specific encoding logic.

Files changed (4) +118 / -8

Bug fix (2) +18 / -8

blackbox.cAllocate space for all nine optional P-frame deltas +2/-1

Allocate space for all nine optional P-frame deltas

• Increases the shared delta buffer from eight to nine entries so logging every optional sensor field cannot write past the stack array.

src/main/blackbox/blackbox.c

blackbox_encoding.cSplit tagged signed-VB runs into decoder-compatible groups +16/-7

Split tagged signed-VB runs into decoder-compatible groups

• Moves the existing encoding into a group helper and makes the public writer process runs eight fields at a time. A lone remainder is always written as a plain signed VB, including when its value is zero.

src/main/blackbox/blackbox_encoding.c

Tests (1) +97 / -0

blackbox_encoding_unittest.ccVerify tagged signed-VB output at group boundaries +97/-0

Verify tagged signed-VB output at group boundaries

• Adds byte-level tests for zero, one, three, eight, nine, and seventeen fields. The nine-field cases verify both a zero-valued ninth field and a nonzero ninth field.

src/test/unit/blackbox_encoding_unittest.cc

Other (1) +3 / -0

CMakeLists.txtRegister Blackbox encoding unit-test dependencies +3/-0

Register Blackbox encoding unit-test dependencies

• Links the new unit test to the Blackbox and common encoding sources and enables USE_BLACKBOX for its build.

src/test/unit/CMakeLists.txt

[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 route each severity your way: inline, summary, both, or drop

More tips ↗ | Customize Qodo ↗ | Qodo docs ↗

[github-actions]

RAM / Flash usage vs. base commit e87050f — commit 65b3234

Using the nearest available size baseline — the PR's exact base commit has no stored baseline yet.

Target	Flash Δ	RAM Δ
MATEKF405	+48 B (+0.01%)	CCM: ±0 B (±0.00%) RAM: +8 B (+0.01%)
MATEKF722	+56 B (+0.01%)	ITCM_RAM: ±0 B (±0.00%) RAM: ±0 B (±0.00%) TCM: ±0 B (±0.00%)
MATEKF765	+200 B (+0.03%)	DTCM_RAM: ±0 B (±0.00%) SRAM1: ±0 B (±0.00%)
MATEKH743	+128 B (+0.02%)	D2_RAM: ±0 B (±0.00%) DTCM_RAM: ±0 B (±0.00%) ITCM_RAM: ±0 B (±0.00%) RAM: ±0 B (±0.00%)

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

[github-actions]

Test firmware build ready — commit 65b3234

Download firmware for PR #12126

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.

[MrScothh]

Decoder side, for logs already recorded with nine fields: iNavFlight/blackbox-tools#102.

[sensei-hacker]

Thanks!