[PATCH 3/5] wireless-regdb: Fix 6 GHz LPI power for Russia (RU)

[Chen-Yu Tsai]

On Thu, Apr 16, 2026 at 10:18 PM Louis Kotze <loukot at gmail.com> wrote:
>
> On Mon, 13 Apr 2026 at 07:03, Chen-Yu Tsai <wens at kernel.org> wrote:
> > This page doesn't actually contain any useful information.
>
> You are right — the eleoscompliance.com reference was a secondary
> source and not useful. I have since obtained the primary SCRF
> documents (shared by Valeriy Kosikhin) and verified them directly.
>
> > I did find https://www.gccorp.com.tw/en/?p=2173 using the document
> > name you gave, but that also mentions that "automatic power control
> > system" is required. I assume that is TPC, which Linux basically
> > doesn't have. So we still have a problem here. If the rules were
> > based on FCC or EU regulations, then we sort of adopt a 3 dBm
> > reduction in the limit.
>
> The TPC concern is resolved. SCRF Decision 23-68-03 (September 27,
> 2023) replaced the phrase "With the use of an automatic power
> control system" with "With the use of LBT mode" across all entries
> in Annex 2 of SCRF 07-20-03-001. This is a literal text
> substitution (clause 3.1 of 23-68-03). There is no TPC requirement
> in the current SCRF allocation, so the -3 dBm reduction does not
> apply here.
>
> LBT as required by ETSI is satisfied by standard 802.11 CCA (Clear
> Channel Assessment) — the mandatory listen-before-transmit mechanism
> in the 802.11 MAC layer. ETSI EN 301 893 clause 4.8 establishes
> this for 5 GHz RLAN, and EN 303 687 for 6 GHz. Linux mac80211
> implements CCA in the standard transmit path.
>
> (The SCRF text actually says "LBT2 mode" — Valeriy traced this to a
> clerical error where a footnote marker "2" was merged into the term
> during a copy-paste between SCRF decisions 21-60 and 23-68-03. There
> is no "LBT2" in any technical standard. He has reported the error to
> the SCRF.)
>
> Verified SCRF allocation values (from primary PDFs):
>
>   2400-2483.5 MHz: 100 mW EIRP, no restriction
>   5150-5350 MHz:   200 mW EIRP, indoor, LBT
>   5650-5850 MHz:   200 mW EIRP, indoor, LBT
>   5925-6425 MHz:   200 mW EIRP (LPI), indoor, LBT
>   57-66 GHz:       40 dBm EIRP, indoor
>
> The current db.txt has 100 mW for all three 5/6 GHz bands. The SCRF
> allocation shows 200 mW. For 5650-5850 and 5925-6425, all applicable
> Russian regulations agree on 200 mW — these entries do not reflect
> the current regulatory limits.
>
> For 5150-5350, there is a complication: the SCRF allocation says
> 200 mW, but the equipment certification rules (per-standard Orders)
> cap power at 100 mW for 802.11n (Order 124) and 802.11ax (Order
> 321). Only 802.11ac (Order 129) allows 200 mW with TPC.
>
> This raises a question I would appreciate your guidance on:
>
> When a country has separate spectrum allocation (200 mW) and
> per-standard equipment certification rules (100 mW for most
> standards), should the regdb encode the spectrum allocation ceiling
> or the most restrictive equipment limit? The ETSI model puts both
> values in the same standard (EN 301 893: 23 dBm with TPC, 20 dBm
> without), but Russia has them in different bodies of law.

The most restrictive one. Basically we cannot give the user device any
chance of violating the limits (or any other restrictions).

> Full source URLs are in Valeriy's reply in this thread. My v1 was
> incomplete — it only addressed 5925-6425 and used an inadequate
> reference. Valeriy is preparing a comprehensive update with lawyer
> review. We are coordinating on the scope of a v2.

Cool. A full document (such as Valeriy's reply) in the commit log is
probably too long. Please try to make it more concise. You can also
provide a link to the other reply on lore.kernel.org to provide more
context.


ChenYu