Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.user > #230713 > unrolled thread

Open source Atheros wifi? (ath10k)

Back to article view | Back to linux.debian.user

Contents

  Open source Atheros wifi? (ath10k) Steven Mainor <steve@degga.net> - 2021-01-16 18:30 +0100
    Re: Open source Atheros wifi? (ath10k) Reco <recoverym4n@enotuniq.net> - 2021-01-16 19:00 +0100
      Re: Open source Atheros wifi? (ath10k) Steven Mainor <steve@degga.net> - 2021-01-16 19:20 +0100
        Re: Open source Atheros wifi? (ath10k) "Andrew M.A. Cater" <amacater@einval.com> - 2021-01-16 19:30 +0100
          Re: Open source Atheros wifi? (ath10k) Brian <ad44@cityscape.co.uk> - 2021-01-16 20:40 +0100
            Re: Open source Atheros wifi? (ath10k) "Andrew M.A. Cater" <amacater@einval.com> - 2021-01-16 22:20 +0100
              Re: Open source Atheros wifi? (ath10k) Brian <ad44@cityscape.co.uk> - 2021-01-17 20:30 +0100
            Re: Open source Atheros wifi? (ath10k) Stefan Monnier <monnier@iro.umontreal.ca> - 2021-01-16 23:00 +0100
              Re: Open source Atheros wifi? (ath10k) The Wanderer <wanderer@fastmail.fm> - 2021-01-16 23:10 +0100
                Re: Open source Atheros wifi? (ath10k) Stefan Monnier <monnier@iro.umontreal.ca> - 2021-01-17 06:10 +0100
                  Re: Open source Atheros wifi? (ath10k) Darac Marjal <mailinglist@darac.org.uk> - 2021-01-17 13:50 +0100
                    Re: Open source Atheros wifi? (ath10k) Andrei POPESCU <andreimpopescu@gmail.com> - 2021-01-18 08:20 +0100
                Re: Open source Atheros wifi? (ath10k) Andrei POPESCU <andreimpopescu@gmail.com> - 2021-01-18 08:00 +0100
              Re: Open source Atheros wifi? (ath10k) didier gaumet <didier.gaumet@gmail.com> - 2021-01-17 10:00 +0100
        Re: Open source Atheros wifi? (ath10k) <tomas@tuxteam.de> - 2021-01-16 21:00 +0100
          Re: Open source Atheros wifi? (ath10k) Brian <ad44@cityscape.co.uk> - 2021-01-17 21:20 +0100
            Re: Open source Atheros wifi? (ath10k) Andrei POPESCU <andreimpopescu@gmail.com> - 2021-01-18 08:20 +0100
              Re: Open source Atheros wifi? (ath10k) <tomas@tuxteam.de> - 2021-01-18 10:30 +0100
        Re: Open source Atheros wifi? (ath10k) Reco <recoverym4n@enotuniq.net> - 2021-01-16 21:50 +0100
      Re: Open source Atheros wifi? (ath10k) Celejar <celejar@gmail.com> - 2021-01-17 02:50 +0100
        Re: Open source Atheros wifi? (ath10k) <tomas@tuxteam.de> - 2021-01-17 10:20 +0100
    Re: Open source Atheros wifi? (ath10k) Gene Heskett <gheskett@shentel.net> - 2021-01-16 22:50 +0100

Page 1 of 2  [1] 2  Next page →

#230713 — Open source Atheros wifi? (ath10k)

[Multipart message — attachments visible in raw view] — view raw

I have a laptop with a QCA9377 Atheros AC wifi card. When I install 
debian 10.7 amd64 I get a warning about missing non-free firmware and 
the wifi doesn't work. It was my understanding that ath10k was open 
source and included in linux. Is this not the case?

If not is there any wireless AC card that doesn't require binary 
drivers? And barring that, what is the best wifi(N?) card I can buy that 
doesn't require closed source drivers/firmware?

-- 
Steven Mainor

[toc] | [next] | [standalone]

#230714

	Hi.

On Sat, Jan 16, 2021 at 12:25:10PM -0500, Steven Mainor wrote:
> I have a laptop with a QCA9377 Atheros AC wifi card. When I install
> debian 10.7 amd64 I get a warning about missing non-free firmware and
> the wifi doesn't work. It was my understanding that ath10k was open
> source and included in linux. Is this not the case?

The driver is free software. The firmware it requires is not.
Thank US FCC for that - [1].


> If not is there any wireless AC card that doesn't require binary
> drivers?

To the best of my knowledge - there are none. Again, [1].


> And barring that, what is the best wifi(N?) card I can buy
> that doesn't require closed source drivers/firmware?

Anything that's ath9k supports does not require any firmware at all.
Ath9k is limited to 802.1n though. See [2] for an example list.

Reco

[1] https://hackaday.com/2016/02/26/fcc-locks-down-router-firmware
[2] https://wireless.wiki.kernel.org/en/users/drivers/ath9k/products/external

[toc] | [prev] | [next] | [standalone]

#230715

[Multipart message — attachments visible in raw view] — view raw

My main concern for the laptop in question is security. So from a 
security standpoint, what is the difference between using a wifi card 
with built in closed source firmware, and closed source firmware that is 
loaded by the kernel like ath10k. Either way the firmware is only 
running on the card, Not the CPU, correct?

Or am I misunderstanding the terminology?

Obviously neither is ideal but is either less secure?

---
Steven Mainor

---
Steven Mainor

On 2021-01-16 12:52, Reco wrote:
> Hi.
> 
> On Sat, Jan 16, 2021 at 12:25:10PM -0500, Steven Mainor wrote:
>> I have a laptop with a QCA9377 Atheros AC wifi card. When I install
>> debian 10.7 amd64 I get a warning about missing non-free firmware and
>> the wifi doesn't work. It was my understanding that ath10k was open
>> source and included in linux. Is this not the case?
> 
> The driver is free software. The firmware it requires is not.
> Thank US FCC for that - [1].
> 
> 
>> If not is there any wireless AC card that doesn't require binary
>> drivers?
> 
> To the best of my knowledge - there are none. Again, [1].
> 
> 
>> And barring that, what is the best wifi(N?) card I can buy
>> that doesn't require closed source drivers/firmware?
> 
> Anything that's ath9k supports does not require any firmware at all.
> Ath9k is limited to 802.1n though. See [2] for an example list.
> 
> Reco
> 
> [1] https://hackaday.com/2016/02/26/fcc-locks-down-router-firmware
> [2] 
> https://wireless.wiki.kernel.org/en/users/drivers/ath9k/products/external

[toc] | [prev] | [next] | [standalone]

#230716

On Sat, Jan 16, 2021 at 01:19:21PM -0500, Steven Mainor wrote:
> My main concern for the laptop in question is security. So from a security
> standpoint, what is the difference between using a wifi card with built in
> closed source firmware, and closed source firmware that is loaded by the
> kernel like ath10k. Either way the firmware is only running on the card, Not
> the CPU, correct?
> 

There's a thread going on in debian-devel mailing list at the moment on more
or less exactly the problem of getting wifi to work for nonfree firmware.

You may find that the firmware-atheros package from non-free in Debian will do the trick. You may need to put the .deb onto a USB stick. I can't quite
remember whether the installer including the non-free firmware CD includes 
firmware-atheros by default but that's found at 
https://cdimage.debian.org/debian-cd/current/amd64/iso-cd/debian-10.7.0-amd64-netinst.iso 

Non-free: From the manufacturer, freely distributable but we can't fix it easily if it goes wrong and we're reliant on the manufacturers to produce updates. Wifi is the most significant stuff to require non-free 
firmware in general.


> Or am I misunderstanding the terminology?
> 
> Obviously neither is ideal but is either less secure?

Hope this helps - all the very best, as ever,

Andy C

> 
> ---
> Steven Mainor
> 
> ---
> Steven Mainor
> 
> On 2021-01-16 12:52, Reco wrote:
> > Hi.
> > 
> > On Sat, Jan 16, 2021 at 12:25:10PM -0500, Steven Mainor wrote:
> > > I have a laptop with a QCA9377 Atheros AC wifi card. When I install
> > > debian 10.7 amd64 I get a warning about missing non-free firmware and
> > > the wifi doesn't work. It was my understanding that ath10k was open
> > > source and included in linux. Is this not the case?
> > 
> > The driver is free software. The firmware it requires is not.
> > Thank US FCC for that - [1].
> > 
> > 
> > > If not is there any wireless AC card that doesn't require binary
> > > drivers?
> > 
> > To the best of my knowledge - there are none. Again, [1].
> > 
> > 
> > > And barring that, what is the best wifi(N?) card I can buy
> > > that doesn't require closed source drivers/firmware?
> > 
> > Anything that's ath9k supports does not require any firmware at all.
> > Ath9k is limited to 802.1n though. See [2] for an example list.
> > 
> > Reco
> > 
> > [1] https://hackaday.com/2016/02/26/fcc-locks-down-router-firmware
> > [2]
> > https://wireless.wiki.kernel.org/en/users/drivers/ath9k/products/external

pub   RSA 4096/9477C19B 2020-03-14 Steven Mainor <steve@degga.net>
> sub   RSA 4096/A4F717B9 2020-03-14
> > 

[toc] | [prev] | [next] | [standalone]

#230718

On Sat 16 Jan 2021 at 18:27:58 +0000, Andrew M.A. Cater wrote:

> On Sat, Jan 16, 2021 at 01:19:21PM -0500, Steven Mainor wrote:
> > My main concern for the laptop in question is security. So from a security
> > standpoint, what is the difference between using a wifi card with built in
> > closed source firmware, and closed source firmware that is loaded by the
> > kernel like ath10k. Either way the firmware is only running on the card, Not
> > the CPU, correct?
> > 
> 
> There's a thread going on in debian-devel mailing list at the moment
> on more or less exactly the problem of getting wifi to work for
> nonfree firmware.

See

  https://lists.debian.org/debian-devel/2021/01/msg00151.html

Very informative.

> You may find that the firmware-atheros package from non-free in Debian
> will do the trick. You may need to put the .deb onto a USB stick. I
> can't quite remember whether the installer including the non-free
> firmware CD includes firmware-atheros by default but

It does include it.

>                                                      that's found at
> https://cdimage.debian.org/debian-cd/current/amd64/iso-cd/debian-10.7.0-amd64-netinst.iso

That isn't the non-free, unofficial installer.

-- 
Brian.

[toc] | [prev] | [next] | [standalone]

#230722

On Sat, Jan 16, 2021 at 07:30:16PM +0000, Brian wrote:
> On Sat 16 Jan 2021 at 18:27:58 +0000, Andrew M.A. Cater wrote:
> 
> > On Sat, Jan 16, 2021 at 01:19:21PM -0500, Steven Mainor wrote:
> > > My main concern for the laptop in question is security. So from a security
> > > standpoint, what is the difference between using a wifi card with built in
> > > closed source firmware, and closed source firmware that is loaded by the
> > > kernel like ath10k. Either way the firmware is only running on the card, Not
> > > the CPU, correct?
> > > 
> > 
> > There's a thread going on in debian-devel mailing list at the moment
> > on more or less exactly the problem of getting wifi to work for
> > nonfree firmware.
> 
> See
> 
>   https://lists.debian.org/debian-devel/2021/01/msg00151.html
> 
> Very informative.
> 
> > You may find that the firmware-atheros package from non-free in Debian
> > will do the trick. You may need to put the .deb onto a USB stick. I
> > can't quite remember whether the installer including the non-free
> > firmware CD includes firmware-atheros by default but
> 
> It does include it.
> 
> >                                                      that's found at
> > https://cdimage.debian.org/debian-cd/current/amd64/iso-cd/debian-10.7.0-amd64-netinst.iso
> 
> That isn't the non-free, unofficial installer.
> 

As ever, you realise your mistake as soon as you've made it: 

http://cdimage.debian.org/cdimage/unofficial/non-free/cd-including-firmware/ 
will point you to it:

http://cdimage.debian.org/cdimage/unofficial/non-free/cd-including-firmware/10.7.0+nonfree/amd64/iso-cd/firmware-10.7.0-amd64-netinst.iso

https://wiki.debian.org/Firmware is the page giving full details.

All best, as ever,

Andy C.

> -- 
> Brian.
> 

[toc] | [prev] | [next] | [standalone]

#230757

On Sat 16 Jan 2021 at 21:12:54 +0000, Andrew M.A. Cater wrote:

[...]

> https://wiki.debian.org/Firmware is the page giving full details.


That is quite a smart page, explaining the need for firmware during
a Debian installation succinctly. It should have some prominance for
users.

The -devel thread is essentially about the need to make a wireless
chipset usuable when the Debian images do not support this. It is not
really about revisiting the argument about firmware as software of
many years ago. The OP simply wants to get something that works on
his machine without having to jump through Google hoops.

The Debian home page has "Download" as a prominent feature. It leads
to an amd64 netinst iso. Why can it not link to some documentation
(as above), and from there to what is available?

Pragmatism goes a long way to easing the path to installing Debian.

-- 
Brian.

[toc] | [prev] | [next] | [standalone]

#230725

>   https://lists.debian.org/debian-devel/2021/01/msg00151.html
> Very informative.

Yes, I agree with Ted Tso's suggestion.

I think the `non-free` repository needs to be significantly improved, so
that the user is properly warned/informed when something is installed
from there (in both senses: "is being installed" or "is currently
installed").

I often enable `non-free` because I need it for one package or another
(GFDL docs for example), but I strongly dislike the fact that as soon as
it's enabled, all its packages become "silently" installable.

I wish non-free packages were always be carefully flagged as such,
ideally with hyperlinks to appropriate documentation about what's
at stake.


        Stefan

[toc] | [prev] | [next] | [standalone]

#230726

[Multipart message — attachments visible in raw view] — view raw

On 2021-01-16 at 16:55, Stefan Monnier wrote:

>>   https://lists.debian.org/debian-devel/2021/01/msg00151.html
>> Very informative.
> 
> Yes, I agree with Ted Tso's suggestion.
> 
> I think the `non-free` repository needs to be significantly improved, so
> that the user is properly warned/informed when something is installed
> from there (in both senses: "is being installed" or "is currently
> installed").

Nothing immediate for the former, but for the latter, you may want to
look at the 'vrms' package.

It stands for 'virtual Richard M. Stallman', in honor of RMS' role as
the orienting compass of the Free Software movement (although its
criteria are based on the DFSG instead of RMS' own views), and what it
does is to scan the installed packages and report any which are from
contrib or non-free.

On my system (and I believe by default), it's set up with a monthly cron
job to send an E-mail with the output of the program. I review that list
every time, and there's rarely anything I feel I can clean up from the
list, but it has happened a time or two. To the best of my recollection,
I've never yet felt the need to run the program by hand.

-- 
   The Wanderer

The reasonable man adapts himself to the world; the unreasonable one
persists in trying to adapt the world to himself. Therefore all
progress depends on the unreasonable man.         -- George Bernard Shaw

[toc] | [prev] | [next] | [standalone]

#230731

> Nothing immediate for the former, but for the latter, you may want to
> look at the 'vrms' package.

Yes, I use that as well.  But that's an "opt-in", and it's rather crude
(and operates after the fact).  I think Debian would benefit from having
this kind of mechanism be much more "in your face" (and more
informative/educational).  It would make it more acceptable (to me,
anyway) to distribute install-media with non-free firmwares, knowing that
those firmware won't be enabled silently).

> On my system (and I believe by default), it's set up with a monthly cron
> job to send an E-mail with the output of the program. I review that list
> every time, and there's rarely anything I feel I can clean up from the
> list, but it has happened a time or two. To the best of my recollection,
> I've never yet felt the need to run the program by hand.

FWIW, I only ever run it by hand.  I can't remember disabling any
monthly cron job nor receiving any email from it, OTOH.  I guess it
somehow guessed that I'd prefer to run it by hand ;-)


        Stefan

[toc] | [prev] | [next] | [standalone]

#230746

[Multipart message — attachments visible in raw view] — view raw

On 17/01/2021 05:03, Stefan Monnier wrote:
>> Nothing immediate for the former, but for the latter, you may want to
>> look at the 'vrms' package.
> Yes, I use that as well.  But that's an "opt-in", and it's rather crude
> (and operates after the fact).  I think Debian would benefit from having
> this kind of mechanism be much more "in your face" (and more
> informative/educational).  It would make it more acceptable (to me,
> anyway) to distribute install-media with non-free firmwares, knowing that
> those firmware won't be enabled silently).

In theory, I imagine it would be possible to make some sort of hook for
apt here. That is, in the same way that apt-listbugs and apt-listchanges
fire off before apt runs and as "Are you still sure you want to install
these packages?", it might be possible to do the same with vrms.

Oh, yes, here's a wishlist bug from 2014 requesting this functionality:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=742854


>
>> On my system (and I believe by default), it's set up with a monthly cron
>> job to send an E-mail with the output of the program. I review that list
>> every time, and there's rarely anything I feel I can clean up from the
>> list, but it has happened a time or two. To the best of my recollection,
>> I've never yet felt the need to run the program by hand.
> FWIW, I only ever run it by hand.  I can't remember disabling any
> monthly cron job nor receiving any email from it, OTOH.  I guess it
> somehow guessed that I'd prefer to run it by hand ;-)
>
>
>         Stefan
>

[toc] | [prev] | [next] | [standalone]

#230774

[Multipart message — attachments visible in raw view] — view raw

On Du, 17 ian 21, 12:43:26, Darac Marjal wrote:
> 
> In theory, I imagine it would be possible to make some sort of hook for
> apt here. That is, in the same way that apt-listbugs and apt-listchanges
> fire off before apt runs and as "Are you still sure you want to install
> these packages?", it might be possible to do the same with vrms.
> 
> Oh, yes, here's a wishlist bug from 2014 requesting this functionality:
> https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=742854

A slightly different approach (Bcc'd to the bug):


    Explanation: disable automatic installation of packages from non-free
    Package: *
    Pin: release c=non-free
    Pin-Priority: -1


    Explanation: allow installation and updates (USB ethernet adapter)
    Package: firmware-realtek
    Pin: release *
    Pin-Priority: 100


Kind regards,
Andrei
-- 
http://wiki.debian.org/FAQsFromDebianUser

[toc] | [prev] | [next] | [standalone]

#230772

[Multipart message — attachments visible in raw view] — view raw

On Sb, 16 ian 21, 17:04:54, The Wanderer wrote:
> 
> Nothing immediate for the former, but for the latter, you may want to
> look at the 'vrms' package.
> 
> It stands for 'virtual Richard M. Stallman', in honor of RMS' role as
> the orienting compass of the Free Software movement (although its
> criteria are based on the DFSG instead of RMS' own views), and what it
> does is to scan the installed packages and report any which are from
> contrib or non-free.

aptitude search '?installed?section(contrib~|non-free)'
 
Kind regards,
Andrei
-- 
http://wiki.debian.org/FAQsFromDebianUser

[toc] | [prev] | [next] | [standalone]

#230733

Le 16/01/2021 à 22:55, Stefan Monnier a écrit :
[...]
> I often enable `non-free` because I need it for one package or another
> (GFDL docs for example), but I strongly dislike the fact that as soon as
> it's enabled, all its packages become "silently" installable.
[...]

On this particular point, perhaps is it possible to forbid this behavior 
by setting an apt preference with the "component" keyword?
  https://linux.die.net/man/5/apt_preferences

[toc] | [prev] | [next] | [standalone]

#230720

[Multipart message — attachments visible in raw view] — view raw

On Sat, Jan 16, 2021 at 01:19:21PM -0500, Steven Mainor wrote:
> My main concern for the laptop in question is security. So from a
> security standpoint, what is the difference between using a wifi
> card with built in closed source firmware, and closed source
> firmware that is loaded by the kernel like ath10k. Either way the
> firmware is only running on the card, Not the CPU, correct?

The difference between burnt-in and loadable firmware is that the
second can change under you at the manufacturer's whim [1]. This
may have advantages and disadvantages, but you don't get to control
those :)

Cheers

[1] Of course, you can try to freeze the firmware and decline
updates.

 - t

[toc] | [prev] | [next] | [standalone]

#230759

On Sat 16 Jan 2021 at 20:57:19 +0100, tomas@tuxteam.de wrote:

> On Sat, Jan 16, 2021 at 01:19:21PM -0500, Steven Mainor wrote:
> > My main concern for the laptop in question is security. So from a
> > security standpoint, what is the difference between using a wifi
> > card with built in closed source firmware, and closed source
> > firmware that is loaded by the kernel like ath10k. Either way the
> > firmware is only running on the card, Not the CPU, correct?
> 
> The difference between burnt-in and loadable firmware is that the
> second can change under you at the manufacturer's whim [1]. This
> may have advantages and disadvantages, but you don't get to control
> those :)

In terms of control, the burnt-in firmware cannot cannot be controlled
either. I do not know where this gets us.

-- 
Brian.

[toc] | [prev] | [next] | [standalone]

#230773

[Multipart message — attachments visible in raw view] — view raw

On Du, 17 ian 21, 20:12:38, Brian wrote:
> On Sat 16 Jan 2021 at 20:57:19 +0100, tomas@tuxteam.de wrote:
> 
> > On Sat, Jan 16, 2021 at 01:19:21PM -0500, Steven Mainor wrote:
> > > My main concern for the laptop in question is security. So from a
> > > security standpoint, what is the difference between using a wifi
> > > card with built in closed source firmware, and closed source
> > > firmware that is loaded by the kernel like ath10k. Either way the
> > > firmware is only running on the card, Not the CPU, correct?
> > 
> > The difference between burnt-in and loadable firmware is that the
> > second can change under you at the manufacturer's whim [1]. This
> > may have advantages and disadvantages, but you don't get to control
> > those :)
> 
> In terms of control, the burnt-in firmware cannot cannot be controlled
> either. I do not know where this gets us.

Besides, the manufacturer might even fix some bugs in the firmware.

Kind regards,
Andrei
-- 
http://wiki.debian.org/FAQsFromDebianUser

[toc] | [prev] | [next] | [standalone]

#230785

[Multipart message — attachments visible in raw view] — view raw

On Mon, Jan 18, 2021 at 09:19:10AM +0200, Andrei POPESCU wrote:
> On Du, 17 ian 21, 20:12:38, Brian wrote:
> > On Sat 16 Jan 2021 at 20:57:19 +0100, tomas@tuxteam.de wrote:

[...]

> > >  This
> > > may have advantages and disadvantages, but you don't get to control
> > > those :)
> > 
> > In terms of control, the burnt-in firmware cannot cannot be controlled
> > either. I do not know where this gets us.

I don't know what was ambiguous in what I said. I said "you don't get
to control when (and what kind of) an update arrives".

No more, no less. In the case of a burnt-in firmware, you get to work
with a "known xxx" [1] state. You lear to know (and possibly circumvent)
bugs, backdoors and other old fellas.

> Besides, the manufacturer might even fix some bugs in the firmware.

Or to insert new and nastier backdoors, as mandated by their
local government (or by their new, ad-industry fueled corporate
overlords, or by whomever whose interests might not align with
yours).

Did I say "advantages and disadvantages?

Cheers

[1] for "xxx" in "good", "bad" and any other suitable value in
   between.

 - t

[toc] | [prev] | [next] | [standalone]

#230721

Please do not top post.

On Sat, Jan 16, 2021 at 01:19:21PM -0500, Steven Mainor wrote:
> My main concern for the laptop in question is security. So from a
> security standpoint, what is the difference between using a wifi card
> with built in closed source firmware, and closed source firmware that
> is loaded by the kernel like ath10k.

The manufacturer's ability to change of said firmware the most arbitrary
way possible, like others said.

For instance (purely theoretical, no known cases) how would you like
your WiFi adapter to transform into a WiFi adapter + Wifi Access Point
combo with well-known pre-shared key?


> Either way the firmware is only running on the card, Not the CPU,
> correct?

Yes, but to pump all those 802.11 frames to and from the kernel's queue
the driver needs to map a certain amount of host memory. Such
allocations can be audited (the driver's source is available, after
all), but it won't be the first case of a kernel's module doing memory
allocations without care.


> Obviously neither is ideal but is either less secure?

If you need your own, controlled 802.11 hardware, you need something
like [1]. I won't call it consumer-friendly, and the price of the needed
hardware is orders of magnitude higher than a usual PCI-X WiFi card, and
of course there are multiple interoperability questions, but it's here
already.


[1] https://www.nuand.com/bladeRF-wiphy/

Reco

[toc] | [prev] | [next] | [standalone]

#230730

On Sat, 16 Jan 2021 20:52:25 +0300
Reco <recoverym4n@enotuniq.net> wrote:

> 	Hi.
> 
> On Sat, Jan 16, 2021 at 12:25:10PM -0500, Steven Mainor wrote:
> > I have a laptop with a QCA9377 Atheros AC wifi card. When I install
> > debian 10.7 amd64 I get a warning about missing non-free firmware and
> > the wifi doesn't work. It was my understanding that ath10k was open
> > source and included in linux. Is this not the case?
> 
> The driver is free software. The firmware it requires is not.
> Thank US FCC for that - [1].

I think you're conflating two senses of wireless 'firmware' - the kind
that runs on the wireless chipset itself (i.e., the stuff that Debian
ships in its free and non-free 'firmware' packages), and the kind that
runs on the system containing wireless hardware (e.g., WiFi OEM's stock
firmware, OpenWrt, etc.).

The article you link to was a scare about the latter, which never
really materialized:

https://arstechnica.com/information-technology/2016/08/fcc-forces-tp-link-to-support-open-source-firmware-on-routers/

The fact that the former is usually closed source these days is
apparently due to a variety of reasons, including manufacturers' desire
to keep their hardware IP secret. There are other computer parts that
also require firmware and don't have open source firmware available,
including GPUS and even Ethernet adapters.

> > If not is there any wireless AC card that doesn't require binary
> > drivers?
> 
> To the best of my knowledge - there are none. Again, [1].
> 
> 
> > And barring that, what is the best wifi(N?) card I can buy
> > that doesn't require closed source drivers/firmware?
> 
> Anything that's ath9k supports does not require any firmware at all.
> Ath9k is limited to 802.1n though. See [2] for an example list.
> 
> Reco
> 
> [1] https://hackaday.com/2016/02/26/fcc-locks-down-router-firmware
> [2] https://wireless.wiki.kernel.org/en/users/drivers/ath9k/products/external

Celejar

[toc] | [prev] | [next] | [standalone]

Page 1 of 2  [1] 2  Next page →

Back to top | Article view | linux.debian.user

csiph-web