Development discussion of WireGuard
 help / color / mirror / Atom feed
* Should we sunset Windows 7 support?
@ 2020-11-10 12:27 Jason A. Donenfeld
  2020-11-10 12:47 ` Laslo Hunhold
                   ` (4 more replies)
  0 siblings, 5 replies; 21+ messages in thread
From: Jason A. Donenfeld @ 2020-11-10 12:27 UTC (permalink / raw)
  To: WireGuard mailing list

Hi,

Windows 7 has been EOL'd by Microsoft since January of this year. It
is no longer receiving security updates or fixes. This email is to get
the conversation started about doing the same with WireGuard for
Windows.

Supporting Windows 7 is an ongoing maintenance burden. For example, we
use SHA2 signatures instead of SHA1 signatures for our drivers, which
is not something we want to compromise on, and as a result Windows 7
users must have KB2921916 installed. But Microsoft never supplied
KB2921916 via Windows Update and it removed all Windows 7 hotfixes
from its webpage last year. So in order to keep supporting this, we're
forced to add clunky disgusting code like this:
https://git.zx2c4.com/wireguard-windows/commit/?id=b63957dc830e39c94844d2f0d32ba29575991e44
Keen readers will wince at all the layering violations there. Do we
really want to keep maintaining gross stuff like this? It makes me
uncomfortable to have kludges like that sitting around in the code.
Shouldn't I write an auto-downloader that then checks hashes?
Shouldn't I build this into the installer? Shouldn't I.... waste tons
of time supporting Windows 7 better?

Probably not.

But I know so many users are still using Windows 7. I'd like to hear
from you to understand why, in order to assess when is the right
moment to sunset our Windows 7 support.

So, if you care for Windows 7, please pipe up! We're not going to
remove support for it overnight, and we're not prepared yet to
announce any sort of formal deprecation plan, but the world is moving
on at some point.

Jason

^ permalink raw reply	[flat|nested] 21+ messages in thread

* Re: Should we sunset Windows 7 support?
  2020-11-10 12:27 Should we sunset Windows 7 support? Jason A. Donenfeld
@ 2020-11-10 12:47 ` Laslo Hunhold
  2020-11-10 12:56   ` samuel.progin
  2020-11-10 12:57 ` Isaac Boukris
                   ` (3 subsequent siblings)
  4 siblings, 1 reply; 21+ messages in thread
From: Laslo Hunhold @ 2020-11-10 12:47 UTC (permalink / raw)
  To: wireguard

On Tue, 10 Nov 2020 13:27:20 +0100
"Jason A. Donenfeld" <Jason@zx2c4.com> wrote:

Dear Jason,

> Windows 7 has been EOL'd by Microsoft since January of this year. It
> is no longer receiving security updates or fixes. This email is to get
> the conversation started about doing the same with WireGuard for
> Windows.
> [...] Do we really want to keep maintaining gross stuff like this? It
> makes me uncomfortable to have kludges like that sitting around in
> the code. Shouldn't I write an auto-downloader that then checks
> hashes? Shouldn't I build this into the installer? Shouldn't I....
> waste tons of time supporting Windows 7 better?
> 
> Probably not.
> 
> But I know so many users are still using Windows 7. I'd like to hear
> from you to understand why, in order to assess when is the right
> moment to sunset our Windows 7 support.
> 
> So, if you care for Windows 7, please pipe up! We're not going to
> remove support for it overnight, and we're not prepared yet to
> announce any sort of formal deprecation plan, but the world is moving
> on at some point.

this is a really difficult judgement to make, which comes up every time
Microsoft EOLs an operating system, because it really often is still
heavily used.

My stance is that we as open source developers don't owe anybody
anything, and if Windows 7 users really care about WireGuard they can
create, share and maintain a patchset that implements the fixes
themselves. You shouldn't be the one paying the price (i.e. time spent)
because people insist on using an EOL'd operating system, which
presents a security issue in many other aspects as well. If they can't
do it themselves, they could pay somebody to deal with such a patchset,
or just keep running the last supported version of WireGuard.

In an utilitarian sense, because you're losing time over Windows 7
support, everyone else is negatively affected, because it's time you
could spend on aspects of WireGuard everyone benefits from, and not
only those running an EOL'd operating system.

To put it shortly, I'm completely in support of sunsetting Windows 7
support, or even just keeping the Windows-7-changes in the next release
for one last time and then immediately dropping them right afterwards
in the git-master. I'm not sure what you exactly mean with sunsetting,
which is why I've given the above "drastic" proposal in case sunsetting
means dealing with this nonsense for another year or something.

With best regards

Laslo Hunhold

^ permalink raw reply	[flat|nested] 21+ messages in thread

* RE: Should we sunset Windows 7 support?
  2020-11-10 12:47 ` Laslo Hunhold
@ 2020-11-10 12:56   ` samuel.progin
  2020-11-10 13:06     ` Jason A. Donenfeld
  0 siblings, 1 reply; 21+ messages in thread
From: samuel.progin @ 2020-11-10 12:56 UTC (permalink / raw)
  To: wireguard

Dear all,

FWIW, Microsoft sells extended support (Windows 7 ESU) to corporate
customers using Pro or Enterprise editions. It can be extended until Jan
10th 2023.

https://docs.microsoft.com/en-us/troubleshoot/windows-client/windows-7-eos-f
aq/windows-7-extended-security-updates-faq

Kind regards.

Samuel

-----Original Message-----
From: WireGuard <wireguard-bounces@lists.zx2c4.com> On Behalf Of Laslo
Hunhold
Sent: mardi, 10 novembre 2020 13:48
To: wireguard@lists.zx2c4.com
Subject: Re: Should we sunset Windows 7 support?

On Tue, 10 Nov 2020 13:27:20 +0100
"Jason A. Donenfeld" <Jason@zx2c4.com> wrote:

Dear Jason,

> Windows 7 has been EOL'd by Microsoft since January of this year. It
> is no longer receiving security updates or fixes. This email is to get
> the conversation started about doing the same with WireGuard for
> Windows.
> [...] Do we really want to keep maintaining gross stuff like this? It
> makes me uncomfortable to have kludges like that sitting around in the
> code. Shouldn't I write an auto-downloader that then checks hashes?
> Shouldn't I build this into the installer? Shouldn't I....
> waste tons of time supporting Windows 7 better?
>
> Probably not.
>
> But I know so many users are still using Windows 7. I'd like to hear
> from you to understand why, in order to assess when is the right
> moment to sunset our Windows 7 support.
>
> So, if you care for Windows 7, please pipe up! We're not going to
> remove support for it overnight, and we're not prepared yet to
> announce any sort of formal deprecation plan, but the world is moving
> on at some point.

this is a really difficult judgement to make, which comes up every time
Microsoft EOLs an operating system, because it really often is still heavily
used.

My stance is that we as open source developers don't owe anybody anything,
and if Windows 7 users really care about WireGuard they can create, share
and maintain a patchset that implements the fixes themselves. You shouldn't
be the one paying the price (i.e. time spent) because people insist on using
an EOL'd operating system, which presents a security issue in many other
aspects as well. If they can't do it themselves, they could pay somebody to
deal with such a patchset, or just keep running the last supported version
of WireGuard.

In an utilitarian sense, because you're losing time over Windows 7 support,
everyone else is negatively affected, because it's time you could spend on
aspects of WireGuard everyone benefits from, and not only those running an
EOL'd operating system.

To put it shortly, I'm completely in support of sunsetting Windows 7
support, or even just keeping the Windows-7-changes in the next release for
one last time and then immediately dropping them right afterwards in the
git-master. I'm not sure what you exactly mean with sunsetting, which is why
I've given the above "drastic" proposal in case sunsetting means dealing
with this nonsense for another year or something.

With best regards

Laslo Hunhold


--
L'absence de virus dans ce courrier électronique a été vérifiée par le logiciel antivirus Avast.
https://www.avast.com/antivirus


^ permalink raw reply	[flat|nested] 21+ messages in thread

* Re: Should we sunset Windows 7 support?
  2020-11-10 12:27 Should we sunset Windows 7 support? Jason A. Donenfeld
  2020-11-10 12:47 ` Laslo Hunhold
@ 2020-11-10 12:57 ` Isaac Boukris
  2020-11-10 15:06 ` Reiner Karlsberg
                   ` (2 subsequent siblings)
  4 siblings, 0 replies; 21+ messages in thread
From: Isaac Boukris @ 2020-11-10 12:57 UTC (permalink / raw)
  To: Jason A. Donenfeld; +Cc: WireGuard mailing list

Hello,

New to the list :)

On Tue, Nov 10, 2020 at 1:30 PM Jason A. Donenfeld <Jason@zx2c4.com> wrote:
>
> Hi,
>
> Windows 7 has been EOL'd by Microsoft since January of this year. It
> is no longer receiving security updates or fixes. This email is to get
> the conversation started about doing the same with WireGuard for
> Windows.
>
> Supporting Windows 7 is an ongoing maintenance burden. For example, we
> use SHA2 signatures instead of SHA1 signatures for our drivers, which
> is not something we want to compromise on, and as a result Windows 7
> users must have KB2921916 installed. But Microsoft never supplied
> KB2921916 via Windows Update and it removed all Windows 7 hotfixes
> from its webpage last year. So in order to keep supporting this, we're
> forced to add clunky disgusting code like this:
> https://git.zx2c4.com/wireguard-windows/commit/?id=b63957dc830e39c94844d2f0d32ba29575991e44
> Keen readers will wince at all the layering violations there. Do we
> really want to keep maintaining gross stuff like this? It makes me
> uncomfortable to have kludges like that sitting around in the code.
> Shouldn't I write an auto-downloader that then checks hashes?
> Shouldn't I build this into the installer? Shouldn't I.... waste tons
> of time supporting Windows 7 better?
>
> Probably not.
>
> But I know so many users are still using Windows 7. I'd like to hear
> from you to understand why, in order to assess when is the right
> moment to sunset our Windows 7 support.

I've had some experience in the past (~5 years ago) with VPN on ATMs,
and we had to support Windows XP at the time since the ATM software
didn't support anything else and was very slow to get updated (several
vendors).

Regards.

^ permalink raw reply	[flat|nested] 21+ messages in thread

* Re: Should we sunset Windows 7 support?
  2020-11-10 12:56   ` samuel.progin
@ 2020-11-10 13:06     ` Jason A. Donenfeld
  0 siblings, 0 replies; 21+ messages in thread
From: Jason A. Donenfeld @ 2020-11-10 13:06 UTC (permalink / raw)
  To: samuel.progin; +Cc: WireGuard mailing list

On Tue, Nov 10, 2020 at 2:05 PM <samuel.progin@gmail.com> wrote:
> FWIW, Microsoft sells extended support (Windows 7 ESU) to corporate
> customers using Pro or Enterprise editions. It can be extended until Jan
> 10th 2023.
>
> https://docs.microsoft.com/en-us/troubleshoot/windows-client/windows-7-eos-f
> aq/windows-7-extended-security-updates-faq

I thought about this... One thing I wonder is: are purchasers of ESU
the same group of people that are likely to run newfangled security
software like WireGuard? How likely is there to be an overlap? I
realize my question implies, "not very likely," but I really don't
know too much about who Microsoft sells this stuff to, so if you have
any insights, I'm all ears.

^ permalink raw reply	[flat|nested] 21+ messages in thread

* Re: Should we sunset Windows 7 support?
  2020-11-10 12:27 Should we sunset Windows 7 support? Jason A. Donenfeld
  2020-11-10 12:47 ` Laslo Hunhold
  2020-11-10 12:57 ` Isaac Boukris
@ 2020-11-10 15:06 ` Reiner Karlsberg
  2020-11-12  8:34   ` Jason A. Donenfeld
  2020-11-10 17:38 ` Andrew Fried
  2020-11-12 17:38 ` Jeffrey Walton
  4 siblings, 1 reply; 21+ messages in thread
From: Reiner Karlsberg @ 2020-11-10 15:06 UTC (permalink / raw)
  To: wireguard

Am 10.11.2020 um 14:27 schrieb Jason A. Donenfeld:
> Hi,
> 
> Windows 7 has been EOL'd by Microsoft since January of this year. It
> is no longer receiving security updates or fixes. This email is to get
> the conversation started about doing the same with WireGuard for
> Windows.
> 
> Supporting Windows 7 is an ongoing maintenance burden. For example, we
> use SHA2 signatures instead of SHA1 signatures for our drivers, which
> is not something we want to compromise on, and as a result Windows 7
> users must have KB2921916 installed. But Microsoft never supplied
> KB2921916 via Windows Update and it removed all Windows 7 hotfixes
> from its webpage last year. So in order to keep supporting this, we're
> forced to add clunky disgusting code like this:
> https://git.zx2c4.com/wireguard-windows/commit/?id=b63957dc830e39c94844d2f0d32ba29575991e44
> Keen readers will wince at all the layering violations there. Do we
> really want to keep maintaining gross stuff like this? It makes me
> uncomfortable to have kludges like that sitting around in the code.
> Shouldn't I write an auto-downloader that then checks hashes?
> Shouldn't I build this into the installer? Shouldn't I.... waste tons
> of time supporting Windows 7 better?
> 
> Probably not.
> 
> But I know so many users are still using Windows 7. I'd like to hear
> from you to understand why, in order to assess when is the right
> moment to sunset our Windows 7 support.
> 
> So, if you care for Windows 7, please pipe up! We're not going to
> remove support for it overnight, and we're not prepared yet to
> announce any sort of formal deprecation plan, but the world is moving
> on at some point.
> 
> Jason
> 

I fully understand the intention, to drop support for Win 7. Although I am still a happy user of it.
I would appreciate a solution, in which further developments of wireguard still remain backward compatible,
to keep the latest official wireguard for Win 7 still functionable with peers, running newer versions of wireguard.
At least for "quite some time".

Cheers,

Reiner


^ permalink raw reply	[flat|nested] 21+ messages in thread

* Re: Should we sunset Windows 7 support?
  2020-11-10 12:27 Should we sunset Windows 7 support? Jason A. Donenfeld
                   ` (2 preceding siblings ...)
  2020-11-10 15:06 ` Reiner Karlsberg
@ 2020-11-10 17:38 ` Andrew Fried
  2020-11-12  8:38   ` Jason A. Donenfeld
  2020-11-12 21:56   ` Panagiotis Kalogiratos
  2020-11-12 17:38 ` Jeffrey Walton
  4 siblings, 2 replies; 21+ messages in thread
From: Andrew Fried @ 2020-11-10 17:38 UTC (permalink / raw)
  To: wireguard

We recently began deploying clusters of recursive DNS "firewalls" that
use wireguard to secure and authenticate all traffic between the client
and servers.   What we quickly learned was that virtually the entire
customer base in India uses Windows 7 almost exclusively.

I can certainly understand the desire to streamline development and
focus on current versions of client operating systems, but by
deprecating support for Windows 7 you would be reducing the number of
potential Wireguard deployments by hundreds of millions of users,
particularly in Asian and underpopulated communities in Africa.  Most of
the end users there simply can't afford the cost of updating to the
latest version of Windows.  I personally wish this were not the case,
but it is what it is.

Andrew

On 11/10/20 7:27 AM, Jason A. Donenfeld wrote:
> Hi,
>
> Windows 7 has been EOL'd by Microsoft since January of this year. It
> is no longer receiving security updates or fixes. This email is to get
> the conversation started about doing the same with WireGuard for
> Windows.
>
> Supporting Windows 7 is an ongoing maintenance burden. For example, we
> use SHA2 signatures instead of SHA1 signatures for our drivers, which
> is not something we want to compromise on, and as a result Windows 7
> users must have KB2921916 installed. But Microsoft never supplied
> KB2921916 via Windows Update and it removed all Windows 7 hotfixes
> from its webpage last year. So in order to keep supporting this, we're
> forced to add clunky disgusting code like this:
> https://git.zx2c4.com/wireguard-windows/commit/?id=b63957dc830e39c94844d2f0d32ba29575991e44
> Keen readers will wince at all the layering violations there. Do we
> really want to keep maintaining gross stuff like this? It makes me
> uncomfortable to have kludges like that sitting around in the code.
> Shouldn't I write an auto-downloader that then checks hashes?
> Shouldn't I build this into the installer? Shouldn't I.... waste tons
> of time supporting Windows 7 better?
>
> Probably not.
>
> But I know so many users are still using Windows 7. I'd like to hear
> from you to understand why, in order to assess when is the right
> moment to sunset our Windows 7 support.
>
> So, if you care for Windows 7, please pipe up! We're not going to
> remove support for it overnight, and we're not prepared yet to
> announce any sort of formal deprecation plan, but the world is moving
> on at some point.
>
> Jason

-- 
Andrew Fried
afried@spamteq.com



^ permalink raw reply	[flat|nested] 21+ messages in thread

* Re: Should we sunset Windows 7 support?
  2020-11-10 15:06 ` Reiner Karlsberg
@ 2020-11-12  8:34   ` Jason A. Donenfeld
  2020-11-12  9:13     ` Roman Mamedov
  0 siblings, 1 reply; 21+ messages in thread
From: Jason A. Donenfeld @ 2020-11-12  8:34 UTC (permalink / raw)
  To: Reiner Karlsberg; +Cc: WireGuard mailing list

Hi Reiner,

On Thu, Nov 12, 2020 at 9:29 AM Reiner Karlsberg
<karlsberg@softart-ge.com> wrote:
> I fully understand the intention, to drop support for Win 7. Although I am still a happy user of it.
> I would appreciate a solution, in which further developments of wireguard still remain backward compatible,
> to keep the latest official wireguard for Win 7 still functionable with peers, running newer versions of wireguard.
> At least for "quite some time".

Could you let me know the rationale for your continued use of Windows
7? Is it economic? Is it just UI preference, and security isn't a
priority to you? Something else?

Jason

^ permalink raw reply	[flat|nested] 21+ messages in thread

* Re: Should we sunset Windows 7 support?
  2020-11-10 17:38 ` Andrew Fried
@ 2020-11-12  8:38   ` Jason A. Donenfeld
  2020-11-12  8:46     ` Phillip McMahon
  2020-11-12 21:56   ` Panagiotis Kalogiratos
  1 sibling, 1 reply; 21+ messages in thread
From: Jason A. Donenfeld @ 2020-11-12  8:38 UTC (permalink / raw)
  To: Andrew Fried; +Cc: WireGuard mailing list

Hi Andrew,

On Thu, Nov 12, 2020 at 9:29 AM Andrew Fried <afried@spamteq.com> wrote:
>
> We recently began deploying clusters of recursive DNS "firewalls" that
> use wireguard to secure and authenticate all traffic between the client
> and servers.   What we quickly learned was that virtually the entire
> customer base in India uses Windows 7 almost exclusively.
>
> I can certainly understand the desire to streamline development and
> focus on current versions of client operating systems, but by
> deprecating support for Windows 7 you would be reducing the number of
> potential Wireguard deployments by hundreds of millions of users,
> particularly in Asian and underpopulated communities in Africa.  Most of
> the end users there simply can't afford the cost of updating to the
> latest version of Windows.  I personally wish this were not the case,
> but it is what it is.

Thanks for that info. This is by far the most compelling point brought
up so far and is one likely to keep us supporting Windows 7 for some
time. I assume your deployment to those parts of the world is quite
large?

Are there any websites or other sources you used to find these
statistics about usage per region? It sounds like that'd be useful to
look at.

Jason

^ permalink raw reply	[flat|nested] 21+ messages in thread

* Re: Should we sunset Windows 7 support?
  2020-11-12  8:38   ` Jason A. Donenfeld
@ 2020-11-12  8:46     ` Phillip McMahon
  2020-11-12  8:50       ` Jason A. Donenfeld
  2020-11-12  9:03       ` Berge Schwebs Bjørlo
  0 siblings, 2 replies; 21+ messages in thread
From: Phillip McMahon @ 2020-11-12  8:46 UTC (permalink / raw)
  To: Jason A. Donenfeld; +Cc: Andrew Fried, WireGuard mailing list

Take a look here

https://gs.statcounter.com/os-version-market-share/windows/desktop/worldwide

Would seem to suggest Windows 7 is in decline but still represents
around 25% of OS footprint out there.

On Thu, 12 Nov 2020 at 09:39, Jason A. Donenfeld <Jason@zx2c4.com> wrote:
>
> Hi Andrew,
>
> On Thu, Nov 12, 2020 at 9:29 AM Andrew Fried <afried@spamteq.com> wrote:
> >
> > We recently began deploying clusters of recursive DNS "firewalls" that
> > use wireguard to secure and authenticate all traffic between the client
> > and servers.   What we quickly learned was that virtually the entire
> > customer base in India uses Windows 7 almost exclusively.
> >
> > I can certainly understand the desire to streamline development and
> > focus on current versions of client operating systems, but by
> > deprecating support for Windows 7 you would be reducing the number of
> > potential Wireguard deployments by hundreds of millions of users,
> > particularly in Asian and underpopulated communities in Africa.  Most of
> > the end users there simply can't afford the cost of updating to the
> > latest version of Windows.  I personally wish this were not the case,
> > but it is what it is.
>
> Thanks for that info. This is by far the most compelling point brought
> up so far and is one likely to keep us supporting Windows 7 for some
> time. I assume your deployment to those parts of the world is quite
> large?
>
> Are there any websites or other sources you used to find these
> statistics about usage per region? It sounds like that'd be useful to
> look at.
>
> Jason



-- 
Use this contact page to send me encrypted messages and files

https://flowcrypt.com/me/phillipmcmahon

P.S. Drowning in email? Try SaneBox and take back control:
http://sanebox.com/t/old3m. I love it.

^ permalink raw reply	[flat|nested] 21+ messages in thread

* Re: Should we sunset Windows 7 support?
  2020-11-12  8:46     ` Phillip McMahon
@ 2020-11-12  8:50       ` Jason A. Donenfeld
  2020-11-12  9:03       ` Berge Schwebs Bjørlo
  1 sibling, 0 replies; 21+ messages in thread
From: Jason A. Donenfeld @ 2020-11-12  8:50 UTC (permalink / raw)
  To: Phillip McMahon; +Cc: Andrew Fried, WireGuard mailing list

Hi Philip,

On Thu, Nov 12, 2020 at 9:47 AM Phillip McMahon
<phillip.mcmahon@gmail.com> wrote:
>
> Take a look here
>
> https://gs.statcounter.com/os-version-market-share/windows/desktop/worldwide
>
> Would seem to suggest Windows 7 is in decline but still represents
> around 25% of OS footprint out there.

Thanks for sending. Indeed that seems pretty illustrative, and you can
adjust the graph to show it per-region, where indeed Africa and Asia
have ~22ish percent of the market share. So, I suppose we'll continue
to support Windows 7 until that dips much lower.

The good news is that we needn't worry about Windows 8 at all. :-P

Jason

^ permalink raw reply	[flat|nested] 21+ messages in thread

* Re: Should we sunset Windows 7 support?
  2020-11-12  8:46     ` Phillip McMahon
  2020-11-12  8:50       ` Jason A. Donenfeld
@ 2020-11-12  9:03       ` Berge Schwebs Bjørlo
  2020-11-13  2:56         ` Jeffrey Walton
  1 sibling, 1 reply; 21+ messages in thread
From: Berge Schwebs Bjørlo @ 2020-11-12  9:03 UTC (permalink / raw)
  To: wireguard

On Thu, Nov 12, 2020 at 09:46:32AM +0100, Phillip McMahon wrote:
> Take a look here
> 
> https://gs.statcounter.com/os-version-market-share/windows/desktop/worldwide
> 
> Would seem to suggest Windows 7 is in decline but still represents
> around 25% of OS footprint out there.

I read 16% from that graph, but that's still considerable.

The Wikimedia stats might also provide some insight:

https://analytics.wikimedia.org/dashboards/browsers/#desktop-site-by-os/os-family-and-major-hierarchical-view

The last 30 days, 8.5% of the Wikimedia "desktop sites" (ie. not mobile)
visits were from Windows 7, which in turn is 12% of their Windows clients
visits.

I imagine Wikimedia stats are pretty geographically biased, since its by far
biggest chunks of content is in either English or other European languages.

Cheers,
-Berge

-- 
Berge Schwebs Bjørlo
Try SCE to AUX.

^ permalink raw reply	[flat|nested] 21+ messages in thread

* Re: Should we sunset Windows 7 support?
  2020-11-12  8:34   ` Jason A. Donenfeld
@ 2020-11-12  9:13     ` Roman Mamedov
  0 siblings, 0 replies; 21+ messages in thread
From: Roman Mamedov @ 2020-11-12  9:13 UTC (permalink / raw)
  To: Jason A. Donenfeld; +Cc: Reiner Karlsberg, WireGuard mailing list

On Thu, 12 Nov 2020 09:34:43 +0100
"Jason A. Donenfeld" <Jason@zx2c4.com> wrote:

> Could you let me know the rationale for your continued use of Windows
> 7? Is it economic? Is it just UI preference, and security isn't a
> priority to you? Something else?

For me, the UI preference absolutely; but security *is* certainly a priority,
which is exactly why I will not run Windows 10: 
https://www.networkworld.com/article/2956574/windows-10-privacy-spyware-settings-user-agreement.html
https://www.forbes.com/sites/gordonkelly/2015/11/02/microsoft-confirms-unstoppable-windows-10-tracking/
https://blog.emsisoft.com/en/18770/the-truth-about-windows-10-spying-on-almost-everything-you-do/

Don't want to wrestle with disabling all of that, to find out that the next
update rolled back my settings, and then added more spying that can't be
disabled this time.

-- 
With respect,
Roman

^ permalink raw reply	[flat|nested] 21+ messages in thread

* Re: Should we sunset Windows 7 support?
  2020-11-10 12:27 Should we sunset Windows 7 support? Jason A. Donenfeld
                   ` (3 preceding siblings ...)
  2020-11-10 17:38 ` Andrew Fried
@ 2020-11-12 17:38 ` Jeffrey Walton
  2020-11-12 17:42   ` Phillip McMahon
  2020-11-12 18:11   ` Neal Gompa
  4 siblings, 2 replies; 21+ messages in thread
From: Jeffrey Walton @ 2020-11-12 17:38 UTC (permalink / raw)
  To: Jason A. Donenfeld; +Cc: WireGuard mailing list

On Tue, Nov 10, 2020 at 7:29 AM Jason A. Donenfeld <Jason@zx2c4.com> wrote:
>
> Windows 7 has been EOL'd by Microsoft since January of this year. It
> is no longer receiving security updates or fixes. This email is to get
> the conversation started about doing the same with WireGuard for
> Windows.

Microsoft does not determine when an operating system is obsolete. The
market does.

Windows 7 still has a 24.79% market share.
https://www.netmarketshare.com/operating-system-market-share.aspx.

Jeff

^ permalink raw reply	[flat|nested] 21+ messages in thread

* Re: Should we sunset Windows 7 support?
  2020-11-12 17:38 ` Jeffrey Walton
@ 2020-11-12 17:42   ` Phillip McMahon
  2020-11-12 18:11   ` Neal Gompa
  1 sibling, 0 replies; 21+ messages in thread
From: Phillip McMahon @ 2020-11-12 17:42 UTC (permalink / raw)
  To: noloader; +Cc: Jason A. Donenfeld, WireGuard mailing list

actually, they do. it's obsolete.

sadly it's still in use at high enough volume that people have to care about it.

On Thu, 12 Nov 2020 at 18:41, Jeffrey Walton <noloader@gmail.com> wrote:
>
> On Tue, Nov 10, 2020 at 7:29 AM Jason A. Donenfeld <Jason@zx2c4.com> wrote:
> >
> > Windows 7 has been EOL'd by Microsoft since January of this year. It
> > is no longer receiving security updates or fixes. This email is to get
> > the conversation started about doing the same with WireGuard for
> > Windows.
>
> Microsoft does not determine when an operating system is obsolete. The
> market does.
>
> Windows 7 still has a 24.79% market share.
> https://www.netmarketshare.com/operating-system-market-share.aspx.
>
> Jeff



-- 
Use this contact page to send me encrypted messages and files

https://flowcrypt.com/me/phillipmcmahon

P.S. Drowning in email? Try SaneBox and take back control:
http://sanebox.com/t/old3m. I love it.

^ permalink raw reply	[flat|nested] 21+ messages in thread

* Re: Should we sunset Windows 7 support?
  2020-11-12 17:38 ` Jeffrey Walton
  2020-11-12 17:42   ` Phillip McMahon
@ 2020-11-12 18:11   ` Neal Gompa
  1 sibling, 0 replies; 21+ messages in thread
From: Neal Gompa @ 2020-11-12 18:11 UTC (permalink / raw)
  To: noloader; +Cc: Jason A. Donenfeld, WireGuard mailing list

On Thu, Nov 12, 2020 at 12:39 PM Jeffrey Walton <noloader@gmail.com> wrote:
>
> On Tue, Nov 10, 2020 at 7:29 AM Jason A. Donenfeld <Jason@zx2c4.com> wrote:
> >
> > Windows 7 has been EOL'd by Microsoft since January of this year. It
> > is no longer receiving security updates or fixes. This email is to get
> > the conversation started about doing the same with WireGuard for
> > Windows.
>
> Microsoft does not determine when an operating system is obsolete. The
> market does.
>
> Windows 7 still has a 24.79% market share.
> https://www.netmarketshare.com/operating-system-market-share.aspx.

Microsoft is absolutely the one determining obsolescence for
*Microsoft Windows*. They are the sole proprietor of the operating
system. If you want it to be different, don't use Windows.




--
真実はいつも一つ!/ Always, there's only one truth!

^ permalink raw reply	[flat|nested] 21+ messages in thread

* Re: Should we sunset Windows 7 support?
  2020-11-10 17:38 ` Andrew Fried
  2020-11-12  8:38   ` Jason A. Donenfeld
@ 2020-11-12 21:56   ` Panagiotis Kalogiratos
  1 sibling, 0 replies; 21+ messages in thread
From: Panagiotis Kalogiratos @ 2020-11-12 21:56 UTC (permalink / raw)
  To: wireguard

It's not really a matter of cost of updating. Windows 10 *STILL* 
activates using
Win7, 8 and 8.1 keys (including OEM licenses implemented via SLIC in 
BIOS/UEFI
by using the COA key on the sticker or directly upgrading an activated 
Win7 installation).
The update is actually free.

Stubborn users of Windows 7, fall down to two main categories:
a) Those who are afraid of the telemetry in Win10, which they think it's 
spying onto them.
b) Those who can't stand the "Metro" aka "Modern" interface and have not 
discovered
Open-Shell (the FOSS continuation of Classic Shell).

That's always excluding the enterprise where there are other factors 
involved regarding
TCO, training employees, etc etc.

Frankly, the internals of 10 are such an upgrade from 7 that it's about 
time people let
7 die in peace. We should not encourage the use of antiquated tech, when 
there's
actually no cost factor involved and it's being done only because of 
stubbornness
(and sometimes misinformation).

Since 7 has been effectively EOL'd, let's stop wasting resources on it.
People can upgrade for free. If they're so fixed onto their obsession 
with 7 and wish
to keep using an EOL'd OS with no security updates, they can also live 
using old
versions of wireguard or another solution. Although I believe most 
developers will
drop support as well. I know I will for all applications I'm working on 
that I have a say
in doing so or not. I have no intention of supporting deprecated OS 
versions. It's
literally a waste of resources.
Soon it will be impossible to find drivers for any new hardware for it. 
It's dead. Let's
just accept that and move on.

Panagiotis

On 10/11/2020 19:38, Andrew Fried wrote:
> We recently began deploying clusters of recursive DNS "firewalls" that
> use wireguard to secure and authenticate all traffic between the client
> and servers.   What we quickly learned was that virtually the entire
> customer base in India uses Windows 7 almost exclusively.
>
> I can certainly understand the desire to streamline development and
> focus on current versions of client operating systems, but by
> deprecating support for Windows 7 you would be reducing the number of
> potential Wireguard deployments by hundreds of millions of users,
> particularly in Asian and underpopulated communities in Africa.  Most of
> the end users there simply can't afford the cost of updating to the
> latest version of Windows.  I personally wish this were not the case,
> but it is what it is.
>
> Andrew
>
> On 11/10/20 7:27 AM, Jason A. Donenfeld wrote:
>> Hi,
>>
>> Windows 7 has been EOL'd by Microsoft since January of this year. It
>> is no longer receiving security updates or fixes. This email is to get
>> the conversation started about doing the same with WireGuard for
>> Windows.
>>
>> Supporting Windows 7 is an ongoing maintenance burden. For example, we
>> use SHA2 signatures instead of SHA1 signatures for our drivers, which
>> is not something we want to compromise on, and as a result Windows 7
>> users must have KB2921916 installed. But Microsoft never supplied
>> KB2921916 via Windows Update and it removed all Windows 7 hotfixes
>> from its webpage last year. So in order to keep supporting this, we're
>> forced to add clunky disgusting code like this:
>> https://git.zx2c4.com/wireguard-windows/commit/?id=b63957dc830e39c94844d2f0d32ba29575991e44
>> Keen readers will wince at all the layering violations there. Do we
>> really want to keep maintaining gross stuff like this? It makes me
>> uncomfortable to have kludges like that sitting around in the code.
>> Shouldn't I write an auto-downloader that then checks hashes?
>> Shouldn't I build this into the installer? Shouldn't I.... waste tons
>> of time supporting Windows 7 better?
>>
>> Probably not.
>>
>> But I know so many users are still using Windows 7. I'd like to hear
>> from you to understand why, in order to assess when is the right
>> moment to sunset our Windows 7 support.
>>
>> So, if you care for Windows 7, please pipe up! We're not going to
>> remove support for it overnight, and we're not prepared yet to
>> announce any sort of formal deprecation plan, but the world is moving
>> on at some point.
>>
>> Jason


^ permalink raw reply	[flat|nested] 21+ messages in thread

* Re: Should we sunset Windows 7 support?
  2020-11-12  9:03       ` Berge Schwebs Bjørlo
@ 2020-11-13  2:56         ` Jeffrey Walton
  2020-11-19 16:59           ` Jason A. Donenfeld
  0 siblings, 1 reply; 21+ messages in thread
From: Jeffrey Walton @ 2020-11-13  2:56 UTC (permalink / raw)
  To: Berge Schwebs Bjørlo; +Cc: WireGuard mailing list

On Thu, Nov 12, 2020 at 6:36 PM Berge Schwebs Bjørlo <berge@trivini.no> wrote:
>
> On Thu, Nov 12, 2020 at 09:46:32AM +0100, Phillip McMahon wrote:
> > Take a look here
> >
> > https://gs.statcounter.com/os-version-market-share/windows/desktop/worldwide
> >
> > Would seem to suggest Windows 7 is in decline but still represents
> > around 25% of OS footprint out there.
>
> I read 16% from that graph, but that's still considerable.

My bad, try this:
https://www.netmarketshare.com/operating-system-market-share.aspx?options=%7B%22filter%22%3A%7B%22%24and%22%3A%5B%7B%22deviceType%22%3A%7B%22%24in%22%3A%5B%22Desktop%2Flaptop%22%5D%7D%7D%5D%7D%2C%22dateLabel%22%3A%22Trend%22%2C%22attributes%22%3A%22share%22%2C%22group%22%3A%22platformVersion%22%2C%22sort%22%3A%7B%22share%22%3A-1%7D%2C%22id%22%3A%22platformsDesktopVersions%22%2C%22dateInterval%22%3A%22Monthly%22%2C%22dateStart%22%3A%222019-11%22%2C%22dateEnd%22%3A%222020-10%22%2C%22segments%22%3A%22-1000%22%7D

I tried to trim the long URL but it lost detail.

The view with the long URL selected "Operating Systems" → "Desktop
Versions" in the left side bar.

Jeff

^ permalink raw reply	[flat|nested] 21+ messages in thread

* Re: Should we sunset Windows 7 support?
  2020-11-13  2:56         ` Jeffrey Walton
@ 2020-11-19 16:59           ` Jason A. Donenfeld
  2020-11-19 17:16             ` akloster
  2021-10-07 23:35             ` Jason A. Donenfeld
  0 siblings, 2 replies; 21+ messages in thread
From: Jason A. Donenfeld @ 2020-11-19 16:59 UTC (permalink / raw)
  To: Jeffrey Walton; +Cc: Berge Schwebs Bjørlo, WireGuard mailing list

So, the emphatic takeaway here seems to be that we're not going to
suddenly ditch 25% of people. So, the recent v0.2.y series released
has amped up support for Windows 7, including automatic detection of
KB2921916.

We'll revisit the discussion in some amount of time if the market
share seems to be decreasing.

I did notice, in the course of using Windows 7 to develop for it, that
every once in a while, the whole screen will fill up with a notice
saying WARNING, WINDOWS 7 IS EOL'D or something to that extent. We'll
see what effect that has.

Jason

^ permalink raw reply	[flat|nested] 21+ messages in thread

* Re: Should we sunset Windows 7 support?
  2020-11-19 16:59           ` Jason A. Donenfeld
@ 2020-11-19 17:16             ` akloster
  2021-10-07 23:35             ` Jason A. Donenfeld
  1 sibling, 0 replies; 21+ messages in thread
From: akloster @ 2020-11-19 17:16 UTC (permalink / raw)
  To: wireguard


[-- Attachment #1.1: Type: text/plain, Size: 659 bytes --]

unsubscribe

On 11/19/20 10:59 AM, Jason A. Donenfeld wrote:
> So, the emphatic takeaway here seems to be that we're not going to
> suddenly ditch 25% of people. So, the recent v0.2.y series released
> has amped up support for Windows 7, including automatic detection of
> KB2921916.
>
> We'll revisit the discussion in some amount of time if the market
> share seems to be decreasing.
>
> I did notice, in the course of using Windows 7 to develop for it, that
> every once in a while, the whole screen will fill up with a notice
> saying WARNING, WINDOWS 7 IS EOL'D or something to that extent. We'll
> see what effect that has.
>
> Jason


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 236 bytes --]

^ permalink raw reply	[flat|nested] 21+ messages in thread

* Re: Should we sunset Windows 7 support?
  2020-11-19 16:59           ` Jason A. Donenfeld
  2020-11-19 17:16             ` akloster
@ 2021-10-07 23:35             ` Jason A. Donenfeld
  1 sibling, 0 replies; 21+ messages in thread
From: Jason A. Donenfeld @ 2021-10-07 23:35 UTC (permalink / raw)
  To: WireGuard mailing list
  Cc: Jeffrey Walton, Berge Schwebs Bjørlo, Laslo Hunhold,
	Panagiotis Kalogiratos, Neal Gompa, Phillip McMahon,
	Roman Mamedov, Reiner Karlsberg, samuel.progin, paul.montgomery

Hey again,

About a year later, WireGuard on Windows keeps becoming more advanced
and integrated into the operating system, with better service
notifications, high speed multi-packet transmission, device arrival
notifications, software device management, and so on... the common
theme being that these are all made possible by Windows 8+ APIs. So
we're now carrying well over 1000 lines of Windows 7 compatibility
code, polyfills, and sometimes outright reimplementations. This is
nobody's idea of fun, and having lots of extra code (that receives
less and less testing) is usually a recipe for things to go wrong. The
added burden of Windows 7 slows progress on newer platforms.

So I really would like to sunset Windows 7 support at some point. The
party cannot go on forever.

However, according to [1], Windows 7 still makes up about 15% of
Windows installs. And I know for a fact that some pretty large
WireGuard deployments are running on Windows 7 boxes, sometimes
installed inside of ATMs... Whether or not that's a good idea, it is
happening. What to do?

My current thinking is to follow the Chromium project, or at least
watch attentively and see the fallout from their decision. After a 6
month extension due to the pandemic, Google is set to retire Windows 7
support in Chrome on January 15, 2022, which is exactly 100 days from
now. There are still some open questions, though: will Microsoft
follow by sunsetting Windows 7 support for Edge, or will they keep it
on life support even longer? Will Mozilla follow Chromium's decision
as well? Is tracking Chromium's decision sensible? It might be: web
standards tend to move awfully fast these days, and without an up to
date web browser, will Windows 7 users upgrade to Windows 11 [or
Linux]? It might not: the most important Windows 7 holdouts may well
be embedded machines that don't care about browsers anyway?

If anybody's thinking on this has evolved, I'd love to hear thoughts.

Thanks,
Jason

[1] https://gs.statcounter.com/windows-version-market-share/desktop/worldwide/

^ permalink raw reply	[flat|nested] 21+ messages in thread

end of thread, other threads:[~2021-10-07 23:35 UTC | newest]

Thread overview: 21+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-11-10 12:27 Should we sunset Windows 7 support? Jason A. Donenfeld
2020-11-10 12:47 ` Laslo Hunhold
2020-11-10 12:56   ` samuel.progin
2020-11-10 13:06     ` Jason A. Donenfeld
2020-11-10 12:57 ` Isaac Boukris
2020-11-10 15:06 ` Reiner Karlsberg
2020-11-12  8:34   ` Jason A. Donenfeld
2020-11-12  9:13     ` Roman Mamedov
2020-11-10 17:38 ` Andrew Fried
2020-11-12  8:38   ` Jason A. Donenfeld
2020-11-12  8:46     ` Phillip McMahon
2020-11-12  8:50       ` Jason A. Donenfeld
2020-11-12  9:03       ` Berge Schwebs Bjørlo
2020-11-13  2:56         ` Jeffrey Walton
2020-11-19 16:59           ` Jason A. Donenfeld
2020-11-19 17:16             ` akloster
2021-10-07 23:35             ` Jason A. Donenfeld
2020-11-12 21:56   ` Panagiotis Kalogiratos
2020-11-12 17:38 ` Jeffrey Walton
2020-11-12 17:42   ` Phillip McMahon
2020-11-12 18:11   ` Neal Gompa

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).