Graphics adapter bugs blocker status rationale

[Adam Williamson]

Hey folks. So this morning I remembered that
https://fedoraproject.org/wiki/Blocker_Bug_FAQ exists - it's a rather
useful page for explaining bits of the blocker process that we should
probably refer to more often. Given that the question keeps coming up, I
added a section to it which explains the precedent we've established for
deciding blocker status for graphics hardware bugs:

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

it's a bit long, but a) I wrote it, what do you expect? and b) it's an
important topic so I wanted to have it really comprehensively covered
somewhere we can point people to. So any time you need an explanation of
how we decide whether graphics adapter bugs are blockers - point to
that. Of course, if anyone thinks I got anything wrong or the text could
be improved, go ahead! I think there's probably other 'common
conditional blocker' cases we could add to the page, which is why I
created that section, but just right now I'm blanking on them. I'm sure
some will come up in subsequent blocker meetings, I'll try to remember
to add them to the page as we go.
-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
http://www.happyassassin.net

[drago01]

On Fri, Oct 5, 2012 at 9:03 PM, Adam Williamson <awilliam(a)redhat.com&gt; wrote:
>  Hey folks. So this morning I remembered that
>  https://fedoraproject.org/wiki/Blocker_Bug_FAQ exists - it's a rather
>  useful page for explaining bits of the blocker process that we should
>  probably refer to more often. Given that the question keeps coming up, I
>  added a section to it which explains the precedent we've established for
>  deciding blocker status for graphics hardware bugs:
>
>
> https://fedoraproject.org/wiki/Blocker_Bug_FAQ#Why_isn.27t_my_graphics_ca...
I disagree with "and affect at least a few different adapters"  ... if
it is just one GPU that a lot of people use it should be sufficent to
be a blocker (common laptop model, an APU or ironlake / ivy / snb gpu
as those are part of the CPU and thus likely have a large userbase).

So if it is a single but commonly used GPU (large userbase) it should
be no different than a bug that affects 3 GPUs that has a userbase as
large as the other one.

[Adam Williamson]

On Fri, 2012-10-05 at 21:24 +0200, drago01 wrote:
>  On Fri, Oct 5, 2012 at 9:03 PM, Adam Williamson
> <awilliam(a)redhat.com&gt; wrote:
>  > Hey folks. So this morning I remembered that
>  > https://fedoraproject.org/wiki/Blocker_Bug_FAQ exists - it's a rather
>  > useful page for explaining bits of the blocker process that we should
>  > probably refer to more often. Given that the question keeps coming up, I
>  > added a section to it which explains the precedent we've established for
>  > deciding blocker status for graphics hardware bugs:
>  >
>  >
> https://fedoraproject.org/wiki/Blocker_Bug_FAQ#Why_isn.27t_my_graphics_ca...
>  
>  I disagree with "and affect at least a few different adapters"  ... if
>  it is just one GPU that a lot of people use it should be sufficent to
>  be a blocker (common laptop model, an APU or ironlake / ivy / snb gpu
>  as those are part of the CPU and thus likely have a large userbase).
>  
>  So if it is a single but commonly used GPU (large userbase) it should
>  be no different than a bug that affects 3 GPUs that has a userbase as
>  large as the other one. 
In theory that's correct but I'm not sure there's actually such a thing
as a single adapter with enough users to constitute a blocker on its
own. The more popular Intel ones would be closest to qualifying, I
guess. Of course, it's hard to prove this by smolt, see my narrative on
https://bugzilla.redhat.com/show_bug.cgi?id=847818 . Still, I'll see if
I can come up with clearer wording, thanks for the note.
-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
http://www.happyassassin.net

[drago01]

On Fri, Oct 5, 2012 at 10:28 PM, Adam Williamson <awilliam(a)redhat.com&gt; wrote:
>  On Fri, 2012-10-05 at 21:24 +0200, drago01 wrote:
> > On Fri, Oct 5, 2012 at 9:03 PM, Adam Williamson <awilliam(a)redhat.com&gt; wrote:
> > > Hey folks. So this morning I remembered that
> > > https://fedoraproject.org/wiki/Blocker_Bug_FAQ exists - it's a rather
> > > useful page for explaining bits of the blocker process that we should
> > > probably refer to more often. Given that the question keeps coming up, I
> > > added a section to it which explains the precedent we've established for
> > > deciding blocker status for graphics hardware bugs:
> > >
> > >
> https://fedoraproject.org/wiki/Blocker_Bug_FAQ#Why_isn.27t_my_graphics_ca...
> >
> > I disagree with "and affect at least a few different adapters"  ... if
> > it is just one GPU that a lot of people use it should be sufficent to
> > be a blocker (common laptop model, an APU or ironlake / ivy / snb gpu
> > as those are part of the CPU and thus likely have a large userbase).
> >
> > So if it is a single but commonly used GPU (large userbase) it should
> > be no different than a bug that affects 3 GPUs that has a userbase as
> > large as the other one.
>
>  In theory that's correct but I'm not sure there's actually such a thing
>  as a single adapter with enough users to constitute a blocker on its
>  own. The more popular Intel ones would be closest to qualifying, I
>  guess. 
My point is the number of affected GPUs (i.e whether 1 vs. 1000) is
irrelevant, the number of affected users (10 vs. 100000) is what
matters.

[Adam Williamson]

On Sat, 2012-10-06 at 17:25 +0200, drago01 wrote:
>  On Fri, Oct 5, 2012 at 10:28 PM, Adam Williamson
> <awilliam(a)redhat.com&gt; wrote:
>  > On Fri, 2012-10-05 at 21:24 +0200, drago01 wrote:
>  >> On Fri, Oct 5, 2012 at 9:03 PM, Adam Williamson <awilliam(a)redhat.com&gt;
> wrote:
>  >> > Hey folks. So this morning I remembered that
>  >> > https://fedoraproject.org/wiki/Blocker_Bug_FAQ exists - it's a rather
>  >> > useful page for explaining bits of the blocker process that we should
>  >> > probably refer to more often. Given that the question keeps coming up, I
>  >> > added a section to it which explains the precedent we've established
> for
>  >> > deciding blocker status for graphics hardware bugs:
>  >> >
>  >> >
> https://fedoraproject.org/wiki/Blocker_Bug_FAQ#Why_isn.27t_my_graphics_ca...
>  >>
>  >> I disagree with "and affect at least a few different adapters"  ...
> if
>  >> it is just one GPU that a lot of people use it should be sufficent to
>  >> be a blocker (common laptop model, an APU or ironlake / ivy / snb gpu
>  >> as those are part of the CPU and thus likely have a large userbase).
>  >>
>  >> So if it is a single but commonly used GPU (large userbase) it should
>  >> be no different than a bug that affects 3 GPUs that has a userbase as
>  >> large as the other one.
>  >
>  > In theory that's correct but I'm not sure there's actually such a thing
>  > as a single adapter with enough users to constitute a blocker on its
>  > own. The more popular Intel ones would be closest to qualifying, I
>  > guess.
>  
>  My point is the number of affected GPUs (i.e whether 1 vs. 1000) is
>  irrelevant, the number of affected users (10 vs. 100000) is what
>  matters. 
Yep, as I said you're correct on that, and I'll try and come up with
better wording.
-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
http://www.happyassassin.net
adamw
adamw