[Proposal] The dual head set-up criterion

[Lukas Ruzicka]

Hello friends of Fedora,

I would like to propose a new release covering criterion that was suggested
on the yesterday's Blocker Review Meeting. Please let me know, what you
think about it and perhaps suggest improvements.

Target:
*Preferably Beta*

*Proposal 1 (conservative criterion)*

*On computers with dual video cards (or with external video output) the
system must be able to display the video content on each connected device.
Any release blocking desktop must provide ways to set up various display
modes (mirroring, extension), scaling, resolution, frequency, and
orientation). *

*Proposal 2 (bold criterion)*
*On computers with dual video cards (or with external video output), as
well as with multiple video cards, the system must be able to display the
video content on each connected device. Any release blocking desktop must
provide ways to set up various display modes (mirroring, extension),
scaling, resolution, frequency, and orientation).*

-- 

Lukáš Růžička

FEDORA QE, RHCE

Red Hat

<https://www.redhat.com>

Purkyňova 115

612 45 Brno - Královo Pole

lruzicka(a)redhat.com
TRIED AND PERSONALLY TESTED, ERGO TRUSTED. <https://redhat.com/trusted>

[Richard Ryniker]

Should multiple displays be a release-blocking criterion?  I can imagine
a desire to ship a new release on-schedule because of significant,
content that works fine, but with multiple display support deferred for
an update because upstream problems are likely to take a while to fix.  I
am uncomfortable with the notion that a release must wait for all
desktops to work with multiple displays on erery architecture.  It is
desirable, yes, but this sounds like a QA attempt to control development.

Your criterion should be limited to hardware that works in single display
configurations.

I think two displays is a reasonable limitation.  Even with that, test
coverage will be extremely sparse - there are just too many devices to
make tests of different combinations practical. (If we cannot test it,
we should not claim a release meets a criterion.)

If I can install three graphics adapters in a machine, and each supports
four displays, would you require that I can run 12 displays on each
desktop?  Lovely if it works, but too rare a use case to be a blocker, or
even to test on a regular basis.

Slots capable of driving a graphics adapter are very limited, but what
about USB devices?  With a few hubs, I could connect scores of displays,
and your criterion asserts they will work.

[Lukas Ruzicka]

Thanks for your view.

On Tue, Mar 30, 2021 at 3:45 PM Richard Ryniker <ryniker(a)ryniker.org&gt; wrote:

>  desktops to work with multiple displays on erery architecture.  It is
>  desirable, yes, but this sounds like a QA attempt to control development.
>  
No, this is not a QA attempt to control development. This is a reaction to
a bug proposal that
aspires to become a blocker for Fedora 34 Final (see
https://bugzilla.redhat.com/show_bug.cgi?id=1942175).
Actually, behaviour like that does not violate any criteria, so
theoretically there is no way how this could
block the release, but we somehow feel (at least I do) that such behaviour
is a nasty thing and should not
happen on a mature system which I believe Fedora is.


>  Your criterion should be limited to hardware that works in single display
>  configurations.
>
>  I think two displays is a reasonable limitation.  Even with that, test
>  coverage will be extremely sparse - there are just too many devices to
>  make tests of different combinations practical. (If we cannot test it,
>  we should not claim a release meets a criterion.)
>  
I agree, therefore my conservative criterion only suggests dual video cards
(I hope
I have understood the term correctly and dual video cards are cards that
enable two monitors)
or an external monitor output (like a laptop would have).

You probably would agree that one external device (in case of a laptop) or
two monitors (in case of a desktop)
is a reasonable expectation where people need to connect projectors,
external monitors, etc.


>  If I can install three graphics adapters in a machine, and each supports
>  four displays, would you require that I can run 12 displays on each
>  desktop?  Lovely if it works, but too rare a use case to be a blocker, or
>  even to test on a regular basis.
>  
That is why I have put the draft here to collect views and suggestions, so
if you believe you could help
with the exact wording to exclude the above situation, please go ahead.

Thanks, Lukas


>  Slots capable of driving a graphics adapter are very limited, but what
>  about USB devices?  With a few hubs, I could connect scores of displays,
>  and your criterion asserts they will work.
>  _______________________________________________
>  test mailing list -- test(a)lists.fedoraproject.org
>  To unsubscribe send an email to test-leave(a)lists.fedoraproject.org
>  Fedora Code of Conduct:
>  https://docs.fedoraproject.org/en-US/project/code-of-conduct/
>  List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
>  List Archives:
>  https://lists.fedoraproject.org/archives/list/[e-mail]
>  Do not reply to spam on the list, report it:
>  https://pagure.io/fedora-infrastructure
>  

-- 

Lukáš Růžička

FEDORA QE, RHCE

Red Hat

<https://www.redhat.com>

Purkyňova 115

612 45 Brno - Královo Pole

lruzicka(a)redhat.com
TRIED AND PERSONALLY TESTED, ERGO TRUSTED. <https://redhat.com/trusted>

[Geoffrey Marr]

Hi Lukas,

*On computers with dual video cards (or with external video output), as
>  well as with multiple video cards*
>  
In regard to the above, I think what you intend to say is: "On computers
with video-out capability, as well as with multiple video cards...", or
something like that. As it reads now, "dual video cards" and "multiple
video cards" are a way of saying the same thing; the "dual" implies two
cards, where the "multiple" implies more-than-one card. I hope that makes
it clearer.

In all, I am skeptical that blocking a release based off of video-out
working is worth having. If one monitor works, be that the integrated
monitor in a laptop, or a single display connected to a desktop, I think
that's enough to have a basic user experience and have the ability to
submit a bug report. However, I do see your point when you say that you
believe that a mature system like Fedora should have multiple-display
capability; I agree with this. From my past experience, I know that
graphics driver issues (noveau, nVidia, etc) can render the entire graphics
ability unable to work in some fashion, and being that Fedora is not the
direct maintainer of these drivers, I wonder how long it would take to get
a fix should this occur in the wild.

I can appreciate the more conservative of the two criteria; my
interpretation of what you have written there is that any computer with one
video card that has video-out capability must work. Between the two, I
would pick this one. Having to test all of the different outputs on a
system with more than one video card would be extremely tedious and I
believe a corner case.

Geoff Marr
IRC: coremodule


On Tue, Mar 30, 2021 at 10:08 AM Lukas Ruzicka <lruzicka(a)redhat.com&gt; wrote:

>  Thanks for your view.
>
>  On Tue, Mar 30, 2021 at 3:45 PM Richard Ryniker <ryniker(a)ryniker.org&gt;
>  wrote:
>
> >
> > desktops to work with multiple displays on erery architecture.  It is
> > desirable, yes, but this sounds like a QA attempt to control development.
> >
>
>  No, this is not a QA attempt to control development. This is a reaction to
>  a bug proposal that
>  aspires to become a blocker for Fedora 34 Final (see
>  https://bugzilla.redhat.com/show_bug.cgi?id=1942175).
>  Actually, behaviour like that does not violate any criteria, so
>  theoretically there is no way how this could
>  block the release, but we somehow feel (at least I do) that such behaviour
>  is a nasty thing and should not
>  happen on a mature system which I believe Fedora is.
>
>
> >
> > Your criterion should be limited to hardware that works in single display
> > configurations.
> >
> > I think two displays is a reasonable limitation.  Even with that, test
> > coverage will be extremely sparse - there are just too many devices to
> > make tests of different combinations practical. (If we cannot test it,
> > we should not claim a release meets a criterion.)
> >
>
>  I agree, therefore my conservative criterion only suggests dual video
>  cards (I hope
>  I have understood the term correctly and dual video cards are cards that
>  enable two monitors)
>  or an external monitor output (like a laptop would have).
>
>  You probably would agree that one external device (in case of a laptop) or
>  two monitors (in case of a desktop)
>  is a reasonable expectation where people need to connect projectors,
>  external monitors, etc.
>
>
> >
> > If I can install three graphics adapters in a machine, and each supports
> > four displays, would you require that I can run 12 displays on each
> > desktop?  Lovely if it works, but too rare a use case to be a blocker, or
> > even to test on a regular basis.
> >
>
>  That is why I have put the draft here to collect views and suggestions, so
>  if you believe you could help
>  with the exact wording to exclude the above situation, please go ahead.
>
>  Thanks, Lukas
>
>
> >
> > Slots capable of driving a graphics adapter are very limited, but what
> > about USB devices?  With a few hubs, I could connect scores of displays,
> > and your criterion asserts they will work.
> > _______________________________________________
> > test mailing list -- test(a)lists.fedoraproject.org
> > To unsubscribe send an email to test-leave(a)lists.fedoraproject.org
> > Fedora Code of Conduct:
> > https://docs.fedoraproject.org/en-US/project/code-of-conduct/
> > List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
> > List Archives:
> > https://lists.fedoraproject.org/archives/list/[e-mail]
> > Do not reply to spam on the list, report it:
> > https://pagure.io/fedora-infrastructure
> >
>
>
>  --
>
>  Lukáš Růžička
>
>  FEDORA QE, RHCE
>
>  Red Hat
>
>  <https://www.redhat.com>
>
>  Purkyňova 115
>
>  612 45 Brno - Královo Pole
>
>  lruzicka(a)redhat.com
>  TRIED AND PERSONALLY TESTED, ERGO TRUSTED. <https://redhat.com/trusted>
>  _______________________________________________
>  test mailing list -- test(a)lists.fedoraproject.org
>  To unsubscribe send an email to test-leave(a)lists.fedoraproject.org
>  Fedora Code of Conduct:
>  https://docs.fedoraproject.org/en-US/project/code-of-conduct/
>  List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
>  List Archives:
>  https://lists.fedoraproject.org/archives/list/[e-mail]
>  Do not reply to spam on the list, report it:
>  https://pagure.io/fedora-infrastructure
>

[Felix Miata]

Lukas Ruzicka composed on 2021-03-30 13:22 (UTC+0200):

>  *On computers with dual video cards (or with external video output)
> the
>  system must be able to display the video content on each connected device. 
Above to me is rather ambiguous.

Laptops generally have one external output only.

Traditional PCs only have external outputs. One or more may be motherboard
outputs. One or more may be on one or more graphics "expansion" cards. A PC may
have both expansion and motherboard outputs.

Docking stations may be employed, confusing the concepts of outputs even more.

Better would be language of connected displays in lieu of video outputs or
"cards".

If the criteria were to be implemented, as others have mentioned, blocking would
have to be limited to two displays, however connected, on account of limited
facility to test more than two, and the many possible permutations of connectivity.
-- 
Evolution as taught in public schools, like religion,
	is based on faith, not on science.

 Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!

Felix Miata  ***  http://fm.no-ip.com/

[Kamil Paral]

On Tue, Mar 30, 2021 at 1:27 PM Lukas Ruzicka <lruzicka(a)redhat.com&gt; wrote:

>  Hello friends of Fedora,
>
>  I would like to propose a new release covering criterion that was
>  suggested on the yesterday's Blocker Review Meeting. Please let me know,
>  what you think about it and perhaps suggest improvements.
>
>  Target:
>  *Preferably Beta*
>
>  *Proposal 1 (conservative criterion)*
>
>  *On computers with dual video cards (or with external video output) the
>  system must be able to display the video content on each connected device.
>  Any release blocking desktop must provide ways to set up various display
>  modes (mirroring, extension), scaling, resolution, frequency, and
>  orientation). *
>
>  *Proposal 2 (bold criterion)*
>  *On computers with dual video cards (or with external video output), as
>  well as with multiple video cards, the system must be able to display the
>  video content on each connected device. Any release blocking desktop must
>  provide ways to set up various display modes (mirroring, extension),
>  scaling, resolution, frequency, and orientation).*
>
>  
As someone who hasn't been at the blocker meeting, I find the criterion
confusing and don't understand the terminology. As far as I know, the "dual
head" term is used to describe having two monitors. Not graphics cards,
monitors. That's the first source of my confusion, because the title
doesn't seem to match the criterion text.
Second, what is an "external video output"? Is it some particular
connector? Or an external graphics card?
Third, "dual video cards" (two) is a subset of "multiple video cards"
(many), the phrasing seems weird.
Fourth, is this criterion about graphics cards producing output in general,
or about all output connectors having to work properly? E.g. if a grapics
card 1 can display output on DP1, but not HDMI1 nor HDMI2, is that a
violation?
Fifth, what about graphics cards that share the same output connectors, and
pass through their data? This is common on laptops, you often have two
graphics cards inside (one inside the CPU and one dedicated), but only one
set of connectors. Does this mean that connectors must work ("on each
connected device") or that you must be able to switch between those two
cards (if possible) and the output must work in both cases?
Sixth, why are the blocking desktops required to provide ways to configure
display modes (mirroring, etc) only when you have 2+ graphics cards, but
not when you have just a single card? Again, this seems to mix up multiple
displays vs multiple graphics cards. The phrasing also doesn't say whether
the configuration must work properly, or just be available.

[Ben Cotton]

On Tue, Mar 30, 2021 at 7:27 AM Lukas Ruzicka <lruzicka(a)redhat.com&gt; wrote:
>  I would like to propose a new release covering criterion that was suggested on the
> yesterday's Blocker Review Meeting. Please let me know, what you think about it and
> perhaps suggest improvements.
>  
I like the idea of dual monitors being a blocker in general, but as
this thread shows, it gets complicated quickly. I think if we do this,
it should be very narrowly scoped. The two main use cases we need to
worry about are:

1. Laptop with an external monitor (and does it have to be directly
attached, or would we block if a monitor attached via a docking
station doesn't work?)
2. Desktop with a single video card that drives two monitors

I suspect that this would cover most usage without getting too caught
up in all of the possible hardware combinations, especially since one
major video card manufacturer isn't the best at supporting Linux.

-- 
Ben Cotton
He / Him / His
Senior Program Manager, Fedora & CentOS Stream
Red Hat
TZ=America/Indiana/Indianapolis

[Chris Murphy]

On Thu, Apr 1, 2021 at 1:12 PM Ben Cotton <bcotton(a)redhat.com&gt; wrote:
>  On Tue, Mar 30, 2021 at 7:27 AM Lukas Ruzicka <lruzicka(a)redhat.com&gt; wrote:
>  >
>  > I would like to propose a new release covering criterion that was suggested on the
> yesterday's Blocker Review Meeting. Please let me know, what you think about it and
> perhaps suggest improvements.
>  >
>
>  I like the idea of dual monitors being a blocker in general, but as
>  this thread shows, it gets complicated quickly. I think if we do this,
>  it should be very narrowly scoped. The two main use cases we need to
>  worry about are:
>
>  1. Laptop with an external monitor (and does it have to be directly
>  attached, or would we block if a monitor attached via a docking
>  station doesn't work?)
>  2. Desktop with a single video card that drives two monitors
>
>  I suspect that this would cover most usage without getting too caught
>  up in all of the possible hardware combinations, especially since one
>  major video card manufacturer isn't the best at supporting Linux.
>  
I agree.


-- 
Chris Murphy

[Adam Williamson]

On Thu, 2021-04-01 at 23:36 -0600, Chris Murphy wrote:
>  On Thu, Apr 1, 2021 at 1:12 PM Ben Cotton <bcotton(a)redhat.com&gt;
> wrote:
>  > 
>  > On Tue, Mar 30, 2021 at 7:27 AM Lukas Ruzicka <lruzicka(a)redhat.com&gt; wrote:
>  > > 
>  > > I would like to propose a new release covering criterion that was suggested on
> the yesterday's Blocker Review Meeting. Please let me know, what you think about it
> and perhaps suggest improvements.
>  > > 
>  > 
>  > I like the idea of dual monitors being a blocker in general, but as
>  > this thread shows, it gets complicated quickly. I think if we do this,
>  > it should be very narrowly scoped. The two main use cases we need to
>  > worry about are:
>  > 
>  > 1. Laptop with an external monitor (and does it have to be directly
>  > attached, or would we block if a monitor attached via a docking
>  > station doesn't work?)
>  > 2. Desktop with a single video card that drives two monitors
>  > 
>  > I suspect that this would cover most usage without getting too caught
>  > up in all of the possible hardware combinations, especially since one
>  > major video card manufacturer isn't the best at supporting Linux.
>  > 
>  
>  I agree. 
Yeah, I think even the 'conservative' version is a bit too ambitious
for a first movement towards blocking on this. I agree with Ben and
Chris about restricting it to these most-common and most-important
cases, and we might want to hedge it around a bit with reference to the
relevant FAQ sections:

https://fedoraproject.org/wiki/Blocker_Bug_FAQ#What_about_hardware_and_lo...
https://fedoraproject.org/wiki/Blocker_Bug_FAQ#Why_isn.27t_my_graphics_ca...

I think personally I'd probably only be inclined to take a bug in this
area as a blocker if it completely broke external output on a popular
laptop line or family (e.g. XPS 13, Thinkpad X1), or broke things
across a popular family or generation of desktop adapters...
-- 
Adam Williamson
Fedora QA
IRC: adamw | Twitter: adamw_ha
https://www.happyassassin.net

[Lukas Ruzicka]

Well, I agree that the terminology might not be the most exact, but I was
not expecting that this version will be taken as is, I rather wanted to
start a discussion. Since this one has been the only proposal that appeared
on the list, I think I did a good thing to start it no matter what
confusion it might have caused.

Those, who were present at the meeting, might remember that when we were
discussing the possible need of that criterion, there were two approaches:

   1. one card -> two monitors
   2. two cards -> two monitors

and it was not clearly decided which one should be the one.
Now, reading through the above posts, I think we do know it:  *one card ->
two monitors.*

In this rank, I see the following options and this is basically what I
wanted to suggest in the *conservative criterion* and this is also what I
fully support. TLDR I think it matches Ben's view:

   - a desktop computer has one graphical adapter with two outputs, so two
   monitors can be used
   - a laptop has an output for external monitor and it can be used (if the
   laptop has two cards and the external monitor can only be used when both
   cards are in operation, then I would rather not block on this or limit this
   to popular devices (as Adam suggests)).
   - With the gnome-control-center or KDE Settings, it should be possible
   to organize the monitors and to set-up their modes *independently for
   any of them. *I believe that we already have a criterion that could be
   used when these things do not work with a single (main) monitor.

Lukas

On Sat, Apr 3, 2021 at 1:53 AM Adam Williamson <adamwill(a)fedoraproject.org&gt;
wrote:

>  On Thu, 2021-04-01 at 23:36 -0600, Chris Murphy wrote:
>  > On Thu, Apr 1, 2021 at 1:12 PM Ben Cotton <bcotton(a)redhat.com&gt; wrote:
>  > >
>  > > On Tue, Mar 30, 2021 at 7:27 AM Lukas Ruzicka <lruzicka(a)redhat.com&gt;
>  wrote:
>  > > >
>  > > > I would like to propose a new release covering criterion that was
>  suggested on the yesterday's Blocker Review Meeting. Please let me know,
>  what you think about it and perhaps suggest improvements.
>  > > >
>  > >
>  > > I like the idea of dual monitors being a blocker in general, but as
>  > > this thread shows, it gets complicated quickly. I think if we do this,
>  > > it should be very narrowly scoped. The two main use cases we need to
>  > > worry about are:
>  > >
>  > > 1. Laptop with an external monitor (and does it have to be directly
>  > > attached, or would we block if a monitor attached via a docking
>  > > station doesn't work?)
>  > > 2. Desktop with a single video card that drives two monitors
>  > >
>  > > I suspect that this would cover most usage without getting too caught
>  > > up in all of the possible hardware combinations, especially since one
>  > > major video card manufacturer isn't the best at supporting Linux.
>  > >
>  >
>  > I agree.
>
>  Yeah, I think even the 'conservative' version is a bit too ambitious
>  for a first movement towards blocking on this. I agree with Ben and
>  Chris about restricting it to these most-common and most-important
>  cases, and we might want to hedge it around a bit with reference to the
>  relevant FAQ sections:
>
>
>
> https://fedoraproject.org/wiki/Blocker_Bug_FAQ#What_about_hardware_and_lo...
>
>
> https://fedoraproject.org/wiki/Blocker_Bug_FAQ#Why_isn.27t_my_graphics_ca...
>
>  I think personally I'd probably only be inclined to take a bug in this
>  area as a blocker if it completely broke external output on a popular
>  laptop line or family (e.g. XPS 13, Thinkpad X1), or broke things
>  across a popular family or generation of desktop adapters...
>  --
>  Adam Williamson
>  Fedora QA
>  IRC: adamw | Twitter: adamw_ha
>  https://www.happyassassin.net
>
>
>  _______________________________________________
>  test mailing list -- test(a)lists.fedoraproject.org
>  To unsubscribe send an email to test-leave(a)lists.fedoraproject.org
>  Fedora Code of Conduct:
>  https://docs.fedoraproject.org/en-US/project/code-of-conduct/
>  List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
>  List Archives:
>  https://lists.fedoraproject.org/archives/list/[e-mail]
>  Do not reply to spam on the list, report it:
>  https://pagure.io/fedora-infrastructure
>  

-- 

Lukáš Růžička

FEDORA QE, RHCE

Red Hat

<https://www.redhat.com>

Purkyňova 115

612 45 Brno - Královo Pole

lruzicka(a)redhat.com
TRIED AND PERSONALLY TESTED, ERGO TRUSTED. <https://redhat.com/trusted>

[Kamil Paral]

On Thu, Apr 1, 2021 at 9:17 PM Ben Cotton <bcotton(a)redhat.com&gt; wrote:

>  On Tue, Mar 30, 2021 at 7:27 AM Lukas Ruzicka
> <lruzicka(a)redhat.com&gt; wrote:
>  >
>  > I would like to propose a new release covering criterion that was
>  suggested on the yesterday's Blocker Review Meeting. Please let me know,
>  what you think about it and perhaps suggest improvements.
>  >
>
>  I like the idea of dual monitors being a blocker in general, but as
>  this thread shows, it gets complicated quickly. I think if we do this,
>  it should be very narrowly scoped. The two main use cases we need to
>  worry about are:
>
>  1. Laptop with an external monitor (and does it have to be directly
>  attached, or would we block if a monitor attached via a docking
>  station doesn't work?)
>  
In this case it needs to be clarified whether we're talking about a laptop
which has a single graphics card, or whether it might have multiple (as is
often the case nowadays, unfortunately most likely involving an Nvidia
card). In case of multiple, whether a single card working is sufficient for
passing that criterion (provided it is possible to toggle between them in
bios or using specialized tools) or whether both must work even if
toggle-able. If they can't be toggled, whether that situation is also
covered (this is the most complicated and most often broken scenario).
Docking stations also need to be mentioned, yes.