Re: S1G and 6GHz channels in Estonia (and EU)

Re: S1G and 6GHz channels in Estonia (and EU)

[Chen-Yu Tsai]

(Adding linux-wireless ML, Kalle and Johannes)

On Mon, Jun 3, 2024 at 1:13 AM Taavi Eomäe <taavieomae@gmail.com> wrote:
>
> Hi!
>
> I have a few questions about the current regdb values for Estonia/EE and EU:
>
> 1) There are no allowed S1G/HaLow/802.11ah EE (or EU) bands:
>
> Based on resources online and Estonian law
> (https://www.riigiteataja.ee/akt/117052024018?leiaKehtiv - Extra 6)
> these two additions should be correct:
>
> + (863 - 868 @ 1), (25 mW)
> + (915.8 - 919.4 @ 1), (25 mW)
>
> (Note: I don't see a way to define duty-cycle limitations, but that
> should be okay?)

I'm not sure if there are plans to support S1G in Linux. If not then
the information would go unused and untested. If it is supported, then
patches are more than welcome.

> 2) There's seems to be a missing 160MHz channel (50):
>
> This combines the restrictions of the two existing defined bands it
> overlaps with (https://ttja.ee/eraklient/side-ja-meediateenused/raadioside/wifi-seadmete-kasutus):
>
> + (5150 - 5350 @ 160), (100 mW), NO-OUTDOOR, DFS, AUTO-BW, wmmrule=ETSI

The AUTO-BW flag lets the kernel know that it can combine neighboring
rules for wider bandwidth at the lowest power limit within the rules.

Thus merging the rules is not needed. Indeed we would prefer rules not
be merged, as the database should convey information closest to the
regulations.

> 3) The 6GHz band should also have an outdoor limit and probably also
> allow 320MHz channels (at least in EE):
>
> - (5945 - 6425 @ 160), (23), NO-OUTDOOR, wmmrule=ETSI
> + (5945 - 6425 @ 320), (25 mW), wmmrule=ETSI
> + (5945 - 6425 @ 320), (200 mW), NO-OUTDOOR, wmmrule=ETSI

The kernel currently can't deal with duplicate rules for the same
channel. Thus we have to make a decision on whether to have a rule
with higher power but indoor use only, or the opposite.

Within our database, we have the former for some jurisdictions
and the latter for others. It really depends on the submitter.

> Would a patch for these be accepted? Should I update all EU countries
> (considering it seems harmonized within the EU)?

I would expect patches to be per country referencing local regulations,
not just the EU one. IIUC the EU sets one harmonized regulation but then
it's up to each member country to implement / incorporate them?

Whether you would like to update other EU countries is up to you.


Thanks
ChenYu

Re: S1G and 6GHz channels in Estonia (and EU)

[Chen-Yu Tsai]

(Resend to Kalle's correct email address)

On Mon, Jun 3, 2024 at 2:36 PM Chen-Yu Tsai <wens@kernel.org> wrote:
>
> (Adding linux-wireless ML, Kalle and Johannes)
>
> On Mon, Jun 3, 2024 at 1:13 AM Taavi Eomäe <taavieomae@gmail.com> wrote:
> >
> > Hi!
> >
> > I have a few questions about the current regdb values for Estonia/EE and EU:
> >
> > 1) There are no allowed S1G/HaLow/802.11ah EE (or EU) bands:
> >
> > Based on resources online and Estonian law
> > (https://www.riigiteataja.ee/akt/117052024018?leiaKehtiv - Extra 6)
> > these two additions should be correct:
> >
> > + (863 - 868 @ 1), (25 mW)
> > + (915.8 - 919.4 @ 1), (25 mW)
> >
> > (Note: I don't see a way to define duty-cycle limitations, but that
> > should be okay?)
>
> I'm not sure if there are plans to support S1G in Linux. If not then
> the information would go unused and untested. If it is supported, then
> patches are more than welcome.
>
> > 2) There's seems to be a missing 160MHz channel (50):
> >
> > This combines the restrictions of the two existing defined bands it
> > overlaps with (https://ttja.ee/eraklient/side-ja-meediateenused/raadioside/wifi-seadmete-kasutus):
> >
> > + (5150 - 5350 @ 160), (100 mW), NO-OUTDOOR, DFS, AUTO-BW, wmmrule=ETSI
>
> The AUTO-BW flag lets the kernel know that it can combine neighboring
> rules for wider bandwidth at the lowest power limit within the rules.
>
> Thus merging the rules is not needed. Indeed we would prefer rules not
> be merged, as the database should convey information closest to the
> regulations.
>
> > 3) The 6GHz band should also have an outdoor limit and probably also
> > allow 320MHz channels (at least in EE):
> >
> > - (5945 - 6425 @ 160), (23), NO-OUTDOOR, wmmrule=ETSI
> > + (5945 - 6425 @ 320), (25 mW), wmmrule=ETSI
> > + (5945 - 6425 @ 320), (200 mW), NO-OUTDOOR, wmmrule=ETSI
>
> The kernel currently can't deal with duplicate rules for the same
> channel. Thus we have to make a decision on whether to have a rule
> with higher power but indoor use only, or the opposite.
>
> Within our database, we have the former for some jurisdictions
> and the latter for others. It really depends on the submitter.
>
> > Would a patch for these be accepted? Should I update all EU countries
> > (considering it seems harmonized within the EU)?
>
> I would expect patches to be per country referencing local regulations,
> not just the EU one. IIUC the EU sets one harmonized regulation but then
> it's up to each member country to implement / incorporate them?
>
> Whether you would like to update other EU countries is up to you.
>
>
> Thanks
> ChenYu

Re: S1G and 6GHz channels in Estonia (and EU)

[Johannes Berg]

On Mon, 2024-06-03 at 14:39 +0800, Chen-Yu Tsai wrote:
> (Resend to Kalle's correct email address)
> 
> On Mon, Jun 3, 2024 at 2:36 PM Chen-Yu Tsai <wens@kernel.org> wrote:
> > 
> > (Adding linux-wireless ML, Kalle and Johannes)
> > 
> > On Mon, Jun 3, 2024 at 1:13 AM Taavi Eomäe <taavieomae@gmail.com> wrote:
> > > 
> > > Hi!
> > > 
> > > I have a few questions about the current regdb values for Estonia/EE and EU:
> > > 
> > > 1) There are no allowed S1G/HaLow/802.11ah EE (or EU) bands:
> > > 
> > > Based on resources online and Estonian law
> > > (https://www.riigiteataja.ee/akt/117052024018?leiaKehtiv - Extra 6)
> > > these two additions should be correct:
> > > 
> > > + (863 - 868 @ 1), (25 mW)
> > > + (915.8 - 919.4 @ 1), (25 mW)
> > > 
> > > (Note: I don't see a way to define duty-cycle limitations, but that
> > > should be okay?)
> > 
> > I'm not sure if there are plans to support S1G in Linux. If not then
> > the information would go unused and untested. If it is supported, then
> > patches are more than welcome.

There's some S1G support, but perhaps only in hwsim right now.

johannes