1550023 - (wr-intel-mvp) [meta] wr-intel-mvp

Status: VERIFIED FIXED

(Core :: Graphics: WebRender, enhancement, P3)

Comment 1

[Kartikaya Gupta (email:[e-mail])]

Just to be clear, this bug ("Intel MVP") is for users who are:

• On Windows10,
• Using one of the whitelisted Intel graphics cards,
• Have no battery, and
• Have small screens (smaller than 1920 x 1200)

Comment 3

[Kartikaya Gupta (email:[e-mail])]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=ca36d7b93c4390796d7aab8f4c42493deaaee60a

Comment 4

[Kartikaya Gupta (email:[e-mail])]

Attachment: Bug 1550023 - Allow WR on a subset of Intel to ride into early beta on Windows. r?jrmuizel

This uses positive-but-empty #if conditions with else clauses rather than
inverted conditions for better readability and documentation.

Comment 5

[Kartikaya Gupta (email:[e-mail])]

Attachment: Bug 1550023 - Move variable into function. r?jrmuizel

It doesn't need to be an argument since it's not used by the caller.

Depends on D37083

Comment 6

[Kartikaya Gupta (email:[e-mail])]

Attachment: Bug 1550023 - Don't allow the qualified pref to disable WR for already released populations. r?jrmuizel

Depends on D37084

Comment 7

[Kartikaya Gupta (email:[e-mail])]

I'm not really that happy with the third patch but I couldn't think of a better way to do it.

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/db4c8cbf463e
Allow WR on a subset of Intel to ride into early beta on Windows. r=jrmuizel

Comment 9

[Kartikaya Gupta (email:[e-mail])]

Comment on attachment 9076241 [details]
Bug 1550023 - Allow WR on a subset of Intel to ride into early beta on Windows. r?jrmuizel

Beta/Release Uplift Approval Request

• User impact if declined: Intel graphics users on Windows desktop machines won't get WR in early beta.

We have WR enabled for this population on Nightly, would like to release it to early beta as well in 69. Then in 70 we can try to push it to release assuming no major issues crop up.

• Is this code covered by automated tests?: Yes
• Has the fix been verified in Nightly?: No
• Needs manual test from QE?: Yes
• If yes, steps to reproduce: After uplift, ensure WR is enabled by default on beta builds when running on Windows desktop machines with modern Intel graphics cards and sub-4k displays.
• List of other uplifts needed: None
• Risk to taking this patch: Low
• Why is the change risky/not risky? (and alternatives if risky): Low risk, just fiddling with some ifdefs to allow early beta windows users to get WR enabled.
• String changes made/needed:

Comment 10

[Kartikaya Gupta (email:[e-mail])]

Comment on attachment 9076241 [details]
Bug 1550023 - Allow WR on a subset of Intel to ride into early beta on Windows. r?jrmuizel

Clearing uplift request for now. Apparently we might want this to ride to release in 69 and not 70. Will re-request once we confirm.

Comment 11

[Narcis Beleuzu [:NarcisB]]

https://hg.mozilla.org/mozilla-central/rev/db4c8cbf463e

Comment 12

[Jeff Muizelaar [:jrmuizel]]

Yes, we want this to ride into release 69.

Comment 13

[Kartikaya Gupta (email:[e-mail])]

In that case we should:

• tweak the condition to use the smaller display size
• land the remaining patches up for review
• request uplift on all three/four patches

Then, just before 69 goes to release, we should land a patch in 69 to turn off gfx.webrender.all.qualified and use normandy to do the gradual rollout of turning it back on.

Comment 14

[Kartikaya Gupta (email:[e-mail])]

Attachment: Bug 1550023 - Qualify whitelisted Intel cards up to WUXGA screen size for Windows in release 69. r?jrmuizel

Comment 15

[Kartikaya Gupta (email:[e-mail])]

Once these patches land WR will be riding to release on small-screen intel Windows desktop devices, which effectively completes this bug. So dropping leave-open keyword.

Comment 16

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/b7cb3e295ca7
Move variable into function. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/b6505d3d38c1
Qualify whitelisted Intel cards up to WUXGA screen size for Windows in release 69. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/7aa8046c0902
Don't allow the qualified pref to disable WR for already released populations. r=jrmuizel

Comment 17

[Arthur Iakab [arthur_iakab]]

https://hg.mozilla.org/mozilla-central/rev/b7cb3e295ca7
https://hg.mozilla.org/mozilla-central/rev/b6505d3d38c1
https://hg.mozilla.org/mozilla-central/rev/7aa8046c0902

Comment 18

[Kartikaya Gupta (email:[e-mail])]

Comment on attachment 9076241 [details]
Bug 1550023 - Allow WR on a subset of Intel to ride into early beta on Windows. r?jrmuizel

Beta/Release Uplift Approval Request

• User impact if declined: Firefox 69 users on whitelisted Intel cards will not get WR enabled in beta/release. Also we won't be able to do progressive rollout of WR to more populations without impacting existing users who have WR enabled in release.
• Is this code covered by automated tests?: Yes
• Has the fix been verified in Nightly?: Yes
• Needs manual test from QE?: Yes
• If yes, steps to reproduce: On 69 beta with the patches applied, verify that:

• Windows 10 desktop machines with modern Intel graphics cards and small displays (less than 1920 x 1200) get WR enabled by default, and WR gets disabled if gfx.webrender.all.qualified (hidden pref) is set to false via about:config.
• Windows 10 desktop machines with modern Nvidia and AMD graphics cards get WR enabled by default, and WR remains enabled even if gfx.webrender.all.qualified is set to false via about:config.

• List of other uplifts needed: None
• Risk to taking this patch: Medium
• Why is the change risky/not risky? (and alternatives if risky): Small risk that we accidentally enable/disable WR on the wrong set of users as the conditions for enabling are getting more complex. QA verification should mitigate the risk, and we are early enough in the beta cycle that we should be able to catch any problems.
• String changes made/needed:

Comment 19

[Ryan VanderMeulen [:RyanVM]]

Comment on attachment 9076241 [details]
Bug 1550023 - Allow WR on a subset of Intel to ride into early beta on Windows. r?jrmuizel

Enables WebRender on a targeted set of Intel graphics chipsets as planned for 69. As was noted, it's early enough in the Beta cycle that we should have time to be on the lookout for any fallout. Thanks for including QA testing instructions. Approved for 69.0b4.

Comment 20

[Ryan VanderMeulen [:RyanVM]]

Looks like D37085 needs rebasing for Beta. The other 3 graft cleanly.

Comment 21

[Kartikaya Gupta (email:[e-mail])]

Thanks for the heads-up. I can rebase and push the uplift.

Comment 22

[Kartikaya Gupta (email:[e-mail])]

remote: https://hg.mozilla.org/releases/mozilla-beta/rev/ebb122d3a9039666139745370d309f4b7739c098
remote: https://hg.mozilla.org/releases/mozilla-beta/rev/1079a6582a0687c1b8457e9ab1417de11cd557aa
remote: https://hg.mozilla.org/releases/mozilla-beta/rev/c69f2d22e51ef9131843e2806ff20aa16a0de2f8
remote: https://hg.mozilla.org/releases/mozilla-beta/rev/ea96619eb18345b62d6280e61f877bc358af1c91

Comment 23

[Jeff Muizelaar [:jrmuizel]]

Reopened because it's a meta

Comment 24

[Kartikaya Gupta (email:[e-mail])]

But.. it's fixed in that lack of any further action will result in WR being released to the intel mvp population in 69. Open blockers should either be fixed or moved to a different meta. For tracking/QA purposes it's better to leave this bug as fixed.

Comment 25

[Jeff Muizelaar [:jrmuizel]]

Ok. Seems reasonable.

Comment 26

[Liz Henry (:lizzard)]

I don't think this happened for 69, but it looks to be potentially riding the train for 70.

Comment 27

[Jessie [:jbonisteel] pls NI]

(In reply to Liz Henry (:lizzard) from comment #26)

I don't think this happened for 69, but it looks to be potentially riding the train for 70.

Correct

Comment 28

[Jessie [:jbonisteel] pls NI]

[Tracking Requested - why for this release]: we would like to ship our Intel MVP with 70

Comment 29

[Timea Cernea [:tbabos][inactive]]

This feature was completed and Sign-off was sent with Green color. Removing qe flags.

Comment 30

[Liz Henry (:lizzard)]

This shipped and we don't need to track the bug.

Comment 31

["Saptarshi Guha[:joy]"]

Report for this experiment: https://metrics.mozilla.com/~sguha/mz/webrender/bug1580913/report.1.html