On Fri, Mar 31, 2023 at 09:38:24PM +0000, Eric Sellman wrote: > So who do I need to go to next to get this fixed? It could be potentially be changed in the kernel, and linux-wireless@xxxxxxxxxxxxxxx is the right group of people. But there are a lot of systems around running older kernels where the wireless-regdb might still be getting updated. So we'd want to at least make an effor to see that the change makes it to all the supported stable and distro kernels, open AP firmware projects, etc before we updated the database. That's likely to take a while. I was about to propose an alternate idea, but then I went to check the kernel source and I think that it already does what I was going to propose. AUTO-BW means that a range can be combined with adjacent ranges for what you called "band straddling," and it also allows the kernel to automatically determine what bandwidth is supported. So all it should need to do is not just consider the bandwiths for the individual ranges but for the entire range which can be created by combining AUTO-BW rules. And I think it's already doing this. reg_get_max_bandwidth_from_range() is called to determine the max bandwidth for a channel when the matching rule has AUTO-BW set. This looks at the rules before and after the matching rule, and the rules before and after those, etc, until it reaches rules which are no longer adjacent. Then it returns a maximum bandwidth which is the span of all the adjacent rules it found. You haven't said if you're seeing an actual problem using 160 MHz channels or if you just noticed the bandwidth is not 160 MHz in the database. If you are having trouble, please provide details about what you observe versus what you expect. Also note that a lot of wireless hardware now comes with it's own in-firmware regulatory database which will override the one loaded from userspace; you can run 'iw reg get' and see if your hardware reports a "self-managed" set of rules. Thanks, Seth > > Get Outlook for Android<https://aka.ms/AAb9ysg> > ________________________________ > From: sforshee@xxxxxxxxxx <sforshee@xxxxxxxxxx> > Sent: Friday, March 31, 2023 2:36:10 PM > To: Eric Sellman <GAK_ACTION@xxxxxxxx> > Cc: wireless-regdb@xxxxxxxxxxxxxxxxxxx <wireless-regdb@xxxxxxxxxxxxxxxxxxx>; linux-wireless@xxxxxxxxxxxxxxx <linux-wireless@xxxxxxxxxxxxxxx> > Subject: Re: wireless-regdb: Update regulatory rules for USA on 5GHz > > On Fri, Mar 31, 2023 at 08:30:09PM +0000, Eric Sellman wrote: > > First, I would like to state that I live in the USA And that this is my first time submitting a correction to the wireless-regdb. > > > > According to this list here >> https://git.kernel.org/pub/scm/linux/kernel/git/sforshee/wireless-regdb.git/tree/db.txt#n1727 > > > > I only see VHT160 allowed to be utilized for 5500mhz to 5640mhz but there is no mention of VH160 being allowed for use with 5180 to 5320mhz. > > > > And according to this presentation here > https://transition.fcc.gov/oet/ea/presentations/files/may17/31-Part-15-Panel-UNII-UpdatesDT.pdf > > > > And this paragraph from this FCC document here > https://www.federalregister.gov/d/2014-09279/p-65 > > > > "Band Straddling" is allowed with U-NII-1 and U-NII-2A. 5150mhz to 5330mhz > > > > Therefore, I would like you to add to the db that 5180 to 5320 can be used with vht160. > > You are correct, however the kernel rejects the database if the max > bandwidth for a range exceeds the size of the range for a given rule. So > unfortunately we can't set the max bandwidth to 160 MHz. > > Thanks, > Seth _______________________________________________ wireless-regdb mailing list wireless-regdb@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/wireless-regdb
