Obtain authenticated Swarm75 evidence without interrupting normal input #5
Notifications
Due Date
No due date set.
Blocks
Reference: xavierk/voidkontrol#5
Reference in New Issue
Block a user
Question
Collect the evidence needed to decide which controls v1 can safely implement for the current USB presentation: confirm device signature/descriptor and precise physical layout, perform a read-only official Kontrol session, record authentication/readback framing, and document readback coverage for a full backup. Preserve input kernel ownership. No guessed packets, feature-report-5 probes, firmware operations, or configuration writes. Trace only the vendor-control interface and separate/redact unrelated device data; a recording must be scoped before sharing. Raw configuration captures can contain sensitive macro content: retain raw captures privately, review contents before publishing fixtures, and preserve protocol structure in sanitized fixtures. Record kernel/hidraw/permissions checks and exact firmware/transport identity. This is HITL where interaction with the physical keyboard or vendor UI is needed; close only when the required evidence actually exists, otherwise retain concrete blockers.
Parent map: Find the way to VoidKontrol v1: complete verified Swarm75 control on Void Linux
Progress — discovery only; capture remains open
A read-only recheck confirms
CONFIG_HIDRAW=y, both Swarm HID interfaces remain bound tohid-generic, and interface 1 has the vendor usage and 519-byte report-6 feature payload described in the spec. Its currentroot:root/0600node is inaccessible to the desktop user. A failedmodprobe hidrawwas not evidence of absent support because the kernel option is boolean.Read-only host and descriptor evidence records the measurements and limits. The original spec now contains the dated correction.
No HID feature reports or setting writes were sent. Authentication, firmware/layout identity, readback coverage, and protocol captures are still missing; this ticket is not resolved. Release the exploratory claim so a later session can take the capture work with the human.
Progress - authenticated capture remains a human-operated blocker
Read-only discovery establishes the USB signature match, intact
hid-genericinput ownership, interface-1 vendor collection, and report-6 size. It does not establish authenticated device identity, firmware/layout identity, command framing, checksums, configuration-read coverage, or any safe write.Required capture, with no settings writes
/dev/input/event*.The next decision tickets remain blocked until this evidence exists. This task is not resolved.
Progress — awaiting the human-operated capture
I verified that the current evidence and the existing safe-capture checklist remain sufficient to run this prerequisite. No device command, capture, configuration write, input grab, driver change, or browser session was initiated in this pass.
The ticket remains blocked on the physical-keyboard interaction described in the prior progress comment: a scoped, read-only official Kontrol capture on the observed USB presentation, followed by private review and a sanitized structural fixture plus coverage table. It must remain open until that evidence (or a concrete safe blocker) is recorded.
Progress — official read-only attempt blocked before device access
At 2026-09-21, I revalidated the attached USB signature (
258a:010c) and the intacthid-genericbinding. The vendor interface remains/dev/hidraw5,root:rootmode0600, so the desktop Chrome process cannot open it.I opened the official Kontrol site, kept its optional analytics, crash reporting, and session replay disabled, and selected Connect your Kreo device. I did not run the page-proposed
curl | shudev installer. The flow remains visibly at Connecting…; it did not expose a device selector, browser permission request, authenticated session, firmware/layout identity, or any configuration readback.No HID feature report, configuration write, input-device read/grab, driver change, host configuration change, or raw packet capture occurred. The host has
usbhid-dump, but notshark/dumpcapand no user-accessibleusbmondebugfs endpoint, so there is no available scoped capture path for the required authentication/readback framing.Concrete blockers: (1) explicit authorization and a reviewed narrowly scoped udev access rule, rather than the third-party pipe-to-shell installer; (2) an authorized capture method that isolates only the Swarm vendor-control traffic; and (3) acceptance of the browser's device-permission prompt once it can enumerate the device. This ticket remains open.
Progress — reviewed desktop access installed; capture remains pending
With explicit authorization, I installed
/etc/udev/rules.d/70-voidkontrol-swarm75.rules. It invokes udev'susb_idbuiltin and grantssoubarna:soubarnamode0600access only when the hidraw device identifies as258a:010con USB interface01.udevadm testand a targeted change event verified that/dev/hidraw5is readable/writable bysoubarna, while the boot-keyboard/dev/hidraw4remainsroot:root 0600.I also prepared a bus-003/device-004 control-transfer-only usbmon collector in a root-only directory. Kontrol did not reach the device: the collector recorded zero lines. The collector, debugfs mount, and
usbmonmodule were then cleanly stopped and removed; the empty root-only artifact is retained as evidence of zero traffic. No feature report or configuration write occurred.The official page is now left at Connecting… after requesting the device. The remaining next action is accepting Chrome's device permission for
kontrol.kreo-tech.comto access the identified Kreo Swarm. That permission is required before restarting the scoped collector and performing the read-only authentication/readback capture. This ticket remains open.