The USB High-Speed negotiation root cause is confirmed on this unit. Piotr replaced both verified 100 Ω per-line D+/D− series resistors with matched 12 Ω parts; the immediate runtime speed changed from 12 to 480 Mbit/s. This direct A/B result proves that the original 100 Ω-per-line network blocked HS negotiation on this unit.
The document preserves the complete source audit and hardware investigation history. It does not claim that 12 Ω is a universal SSC338Q requirement or that long-term signal margin and WFB stability are already established. The historical parallel-resistor proposal is explicitly superseded and was never executed.
Acceptance still pending
Keep the installed 12 Ω parts unchanged and retain the 4096 kbit/s WFB limit until all three owner-run gates are recorded:
Fresh dmesg shows a new high-speed enumeration without relevant USB errors.
A full Sstar-ehci-1 platform-driver unbind/bind over Ethernet or console still reports 480.
The established WFB baseline operates stably without USB disconnect/reset events.
Repository validation
independent claim-by-claim review against pinned openipc/linux revision 97b7fc4e
each hardware-evidence follow-up independently reviewed
ci-matrix.py --self-test: pass
repository lint self-tests and workflow shell lint: pass
git diff --check: pass on the final follow-up commit
changed-lines credential/security scan, stale-instruction sweep, and overclaim/unsafe-instruction review: pass
full-tree lint-cli-paths.py: known pre-existing false positive reproduced at base; unchanged and out of scope
## Result
The USB High-Speed negotiation root cause is confirmed on this unit. Piotr replaced both verified `100 Ω` per-line D+/D− series resistors with matched `12 Ω` parts; the immediate runtime speed changed from `12` to `480` Mbit/s. This direct A/B result proves that the original 100 Ω-per-line network blocked HS negotiation on this unit.
The document preserves the complete source audit and hardware investigation history. It does not claim that `12 Ω` is a universal SSC338Q requirement or that long-term signal margin and WFB stability are already established. The historical parallel-resistor proposal is explicitly superseded and was never executed.
## Acceptance still pending
Keep the installed 12 Ω parts unchanged and retain the 4096 kbit/s WFB limit until all three owner-run gates are recorded:
1. Fresh `dmesg` shows a new high-speed enumeration without relevant USB errors.
2. A full `Sstar-ehci-1` platform-driver unbind/bind over Ethernet or console still reports `480`.
3. The established WFB baseline operates stably without USB disconnect/reset events.
## Repository validation
- independent claim-by-claim review against pinned `openipc/linux` revision `97b7fc4e`
- each hardware-evidence follow-up independently reviewed
- `ci-matrix.py --self-test`: pass
- repository lint self-tests and workflow shell lint: pass
- `git diff --check`: pass on the final follow-up commit
- changed-lines credential/security scan, stale-instruction sweep, and overclaim/unsafe-instruction review: pass
- full-tree `lint-cli-paths.py`: known pre-existing false positive reproduced at base; unchanged and out of scope
Closes #1
piotr
was assigned by grzesiu2026-10-02 16:53:20 +02:00
Reviewed the PR against the current evidence. No device-overlay change is justified: the active DT and builder delta contain no divergent USB configuration, the RTL driver only consumes the already-negotiated USB speed, and USB device rebind reproduces Full-Speed.
The investigation record is useful and correctly keeps kernel/analog candidates outside this builder overlay. No merge blocker found for the documentation.
Next discriminating experiment is not a bitrate/FEC or supply change. Piotr has Ethernet access and should run the documented full Sstar-ehci-1 platform-driver unbind/bind once. Unlike the prior USB-device rebind, it re-runs the vendor EHCI/UTMI initialization and CA calibration. Hardware acceptance remains Piotr's task; Grzesiu has no physical-device access.
## Review: documentation-only outcome is appropriate
Reviewed the PR against the current evidence. No device-overlay change is justified: the active DT and builder delta contain no divergent USB configuration, the RTL driver only consumes the already-negotiated USB speed, and USB device rebind reproduces Full-Speed.
The investigation record is useful and correctly keeps kernel/analog candidates outside this builder overlay. No merge blocker found for the documentation.
Next discriminating experiment is **not** a bitrate/FEC or supply change. Piotr has Ethernet access and should run the documented full `Sstar-ehci-1` platform-driver unbind/bind once. Unlike the prior USB-device rebind, it re-runs the vendor EHCI/UTMI initialization and CA calibration. Hardware acceptance remains Piotr's task; Grzesiu has no physical-device access.
grzesiu
changed title from hilink: document RTL8812EU USB High-Speed investigation to hilink: document RTL8812EU USB High-Speed root cause and rework2026-10-15 10:31:34 +02:00
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Result
The USB High-Speed negotiation root cause is confirmed on this unit. Piotr replaced both verified
100 Ωper-line D+/D− series resistors with matched12 Ωparts; the immediate runtime speed changed from12to480Mbit/s. This direct A/B result proves that the original 100 Ω-per-line network blocked HS negotiation on this unit.The document preserves the complete source audit and hardware investigation history. It does not claim that
12 Ωis a universal SSC338Q requirement or that long-term signal margin and WFB stability are already established. The historical parallel-resistor proposal is explicitly superseded and was never executed.Acceptance still pending
Keep the installed 12 Ω parts unchanged and retain the 4096 kbit/s WFB limit until all three owner-run gates are recorded:
dmesgshows a new high-speed enumeration without relevant USB errors.Sstar-ehci-1platform-driver unbind/bind over Ethernet or console still reports480.Repository validation
openipc/linuxrevision97b7fc4eci-matrix.py --self-test: passgit diff --check: pass on the final follow-up commitlint-cli-paths.py: known pre-existing false positive reproduced at base; unchanged and out of scopeCloses #1
Review: documentation-only outcome is appropriate
Reviewed the PR against the current evidence. No device-overlay change is justified: the active DT and builder delta contain no divergent USB configuration, the RTL driver only consumes the already-negotiated USB speed, and USB device rebind reproduces Full-Speed.
The investigation record is useful and correctly keeps kernel/analog candidates outside this builder overlay. No merge blocker found for the documentation.
Next discriminating experiment is not a bitrate/FEC or supply change. Piotr has Ethernet access and should run the documented full
Sstar-ehci-1platform-driver unbind/bind once. Unlike the prior USB-device rebind, it re-runs the vendor EHCI/UTMI initialization and CA calibration. Hardware acceptance remains Piotr's task; Grzesiu has no physical-device access.hilink: document RTL8812EU USB High-Speed investigationto hilink: document RTL8812EU USB High-Speed root cause and reworkView command line instructions
Checkout
From your project repository, check out a new branch and test the changes.