1418556 - Ship a hotfix (system add-on) to fix bug 1417442

Status: RESOLVED FIXED

(Core :: Audio/Video: Playback, enhancement, P1)

Description

[Ritu Kothari (:ritu) (Inactive, please n-i to RyanVM, jcristau, or pascal)]

We'll use this bug to track the hotfix push. Rhelmer has the system add-on ready. We will need this deployed to release-cdntest for Softvision team to be able to test it. Once QA sign off is done, Sylvestre/Julien/Liz can sign off on the push via email, here and in balrog.

Comment 2

[Robert Helmer [:rhelmer]]

Attachment: [e-mail]

Could you please sign this system add-on update? Thanks!

Comment 3

[Jean-Yves Avenard [:jya]]

This isn't going to work for 57

Comment 4

[Jean-Yves Avenard [:jya]]

Attachment: [e-mail]

update preference name

Comment 5

[Jean-Yves Avenard [:jya]]

The install.rdf states "This disables NV12 format on Windows for AMD drivers earlier than 21.19.411.0"

there's no such test in the file. It's a blanket application for all AMD devices.. Is this intended?

telemetryEnv.system.gfx contains the driver version, can't we use that? a tad more complicated I admit, seeing that the version number is a string.

Comment 6

[Robert Helmer [:rhelmer]]

(In reply to Jean-Yves Avenard [:jya] from comment #5)
> The install.rdf states "This disables NV12 format on Windows for AMD drivers
> earlier than 21.19.411.0"
> 
> there's no such test in the file. It's a blanket application for all AMD
> devices.. Is this intended?

Good point... I originally got that description and version # from earlier in the email thread, but from later discussions I wasn't sure if this was desired.

> telemetryEnv.system.gfx contains the driver version, can't we use that? a
> tad more complicated I admit, seeing that the version number is a string.

We can, let's confirm via email what's needed here. Thanks for catching this!

Comment 7

[Andrei Vaida [:avaida]]

Are there any updates here for QA? A test plan [1] has been drafted for this request, but testing has not yet started as we don't have all the specifics and the required update rules are not configured yet on the update channel.

[1] https://docs.google.com/document/d/11ESpnvxexaNWgLh_C_jlMmmeJyQ6N_c0SHLTexL-XBk/edit#

Comment 8

[Jason Thomas [:jason]]

Attachment: [e-mail]

Please see attached.

Comment 9

[Robert Helmer [:rhelmer]]

(In reply to Andrei Vaida [:avaida], Desktop Release QA – please ni? me from comment #7)
> Are there any updates here for QA? A test plan [1] has been drafted for this
> request, but testing has not yet started as we don't have all the specifics
> and the required update rules are not configured yet on the update channel.
> 
> [1]
> https://docs.google.com/document/d/11ESpnvxexaNWgLh_C_jlMmmeJyQ6N_c0SHLTexL-
> XBk/edit#

Rehan, could you please host the signed hotfix add-on attached to this bug (attachment 8930122 [details]) on a test channel for QA? Thanks!

Comment 10

[Rehan Dalal (Inactive)]

This has been added to the "release-sysaddon" channel for 57.*

Comment 11

[Andrei Vaida [:avaida]]

It's 7 PM in Romania, we won't be able to test the system add-on today. 

Once we clarify the following questions, we should be able to test the 'release-sysaddon' channel early tomorrow:

    1. What are the exact changes made by this system add-on and how can QA verify them?
    2. What Firefox versions and what operating systems are we planning to push this system add-on to?


Feel free to comment on the gdoc directly, if that's easier.

Comment 12

[Robert Helmer [:rhelmer]]

(In reply to Andrei Vaida [:avaida], Desktop Release QA – please ni? me from comment #11)
> It's 7 PM in Romania, we won't be able to test the system add-on today. 
> 
> Once we clarify the following questions, we should be able to test the
> 'release-sysaddon' channel early tomorrow:
> 
>     1. What are the exact changes made by this system add-on and how can QA
> verify them?
>     2. What Firefox versions and what operating systems are we planning to
> push this system add-on to?
> 
> 
> Feel free to comment on the gdoc directly, if that's easier.

This hotfix add-on sets the pref "media.windows-media-foundation.use-nv12-format" to false on Windows systems with AMD graphics.

On such a system, it should be enough to confirm the pref is set to true initially and then changes to false after the add-on is installed.

needinfo'ing :jya to confirm that I am not missing anything here.

Comment 13

[Robert Helmer [:rhelmer]]

(In reply to Robert Helmer [:rhelmer] from comment #12)
> (In reply to Andrei Vaida [:avaida], Desktop Release QA – please ni? me from
> comment #11)
> > It's 7 PM in Romania, we won't be able to test the system add-on today. 
> > 
> > Once we clarify the following questions, we should be able to test the
> > 'release-sysaddon' channel early tomorrow:
> > 
> >     1. What are the exact changes made by this system add-on and how can QA
> > verify them?
> >     2. What Firefox versions and what operating systems are we planning to
> > push this system add-on to?
> > 
> > 
> > Feel free to comment on the gdoc directly, if that's easier.
> 
> This hotfix add-on sets the pref
> "media.windows-media-foundation.use-nv12-format" to false on Windows systems
> with AMD graphics.
> 
> On such a system, it should be enough to confirm the pref is set to true
> initially and then changes to false after the add-on is installed.
> 
> needinfo'ing :jya to confirm that I am not missing anything here.

Well I suppose it'd be ideal to also test that bug 1417442 is reproducible on such a system and that applying the hotfix and the pref being flipped causes the problem to go away.

Comment 14

[Jean-Yves Avenard [:jya]]

Yes, testing that the pref is set to false in 57 should be enough.

Comment 15

[Jean-Yves Avenard [:jya]]

Oh yes, the pref should only be set to false on amd devices, not Intel and not nvidia.

Also verify on machines supporting the AMD vp9 decoders, that YouTube still works. We haven't tested the VP9 decoder with RGBA32 surfaces. Like AMD A10, or AMD FX 4xx or 5xx

Comment 16

[Ciprian Georgiu, Desktop QA]

We encountered an issue while testing the SSO on "release-sysaddon" update channel on FF 57.0 (2017111212534).
The associated SSO - Disable Media WMF NV12 format 1.0 has been installed, but it seems that the pref "media.windows-media-foundation.use-nv12-format" was not set to false. The GPU used was an AMD Radeon R7 360 with the following driver version: 16.300.2311.0.

Can you please take a look at this, :jya?

(In reply to Jean-Yves Avenard [:jya] from comment #15)
> Oh yes, the pref should only be set to false on amd devices, not Intel and
> not nvidia.
> 
> Also verify on machines supporting the AMD vp9 decoders, that YouTube still
> works. We haven't tested the VP9 decoder with RGBA32 surfaces. Like AMD A10,
> or AMD FX 4xx or 5xx

Unfortunately we don't have those GPUs so we won't be able to test this scenario.

Comment 17

[Jean-Yves Avenard [:jya]]

(In reply to Ciprian Georgiu, QA [:ciprian_georgiu] from comment #16)
> We encountered an issue while testing the SSO on "release-sysaddon" update
> channel on FF 57.0 (2017111212534).
> The associated SSO - Disable Media WMF NV12 format 1.0 has been installed,
> but it seems that the pref "media.windows-media-foundation.use-nv12-format"
> was not set to false. The GPU used was an AMD Radeon R7 360 with the
> following driver version: 16.300.2311.0.
> 
> Can you please take a look at this, :jya?

I'm not familiar on how those hotfix work.
I'll take a look.

> 
> (In reply to Jean-Yves Avenard [:jya] from comment #15)
> > Oh yes, the pref should only be set to false on amd devices, not Intel and
> > not nvidia.
> > 
> > Also verify on machines supporting the AMD vp9 decoders, that YouTube still
> > works. We haven't tested the VP9 decoder with RGBA32 surfaces. Like AMD A10,
> > or AMD FX 4xx or 5xx
> 
> Unfortunately we don't have those GPUs so we won't be able to test this
> scenario.

you need to get some...

We can't really ship code we can't test !

Comment 18

[Jean-Yves Avenard [:jya]]

I followed the instructions provided in the Test Plan document.

I can see the addon being installed and run:

1511259755861	addons.xpi	DEBUG	Addon [e-mail] will be installed as a packed xpi
1511259755865	addons.xpi	INFO	Mapping [e-mail] to C:\Users\developer\AppData\Roaming\Mozilla\Firefox\Profiles\iihxqiz5.release\features\{3409ba2a-24b7-4dc7-991d-0fb2106987e2}\[e-mail]
1511259755866	addons.xpi-utils	DEBUG	Make addon app-system-addons:[e-mail] visible
1511259755867	DeferredSave.extensions.json	DEBUG	Save changes
1511259755868	DeferredSave.extensions.json	DEBUG	Save changes
1511259755868	addons.xpi	DEBUG	XPIStates adding add-on [e-mail] in app-system-addons: C:\Users\developer\AppData\Roaming\Mozilla\Firefox\Profiles\iihxqiz5.release\features\{3409ba2a-24b7-4dc7-991d-0fb2106987e2}\[e-mail]
1511259755869	addons.xpi	DEBUG	Updating XPIState for {"id":"[e-mail]","syncGUID":"{387cd395-e28c-4872-b7bb-4ea65037f1fa}","location":"app-system-addons","version":"1.0","type":"extension","internalName":null,"updateURL":null,"updateKey":null,"optionsURL":null,"optionsType":null,"aboutURL":null,"defaultLocale":{"name":"Disable Media WMF NV12 format","description":"This disables NV12 format on Windows for AMD drivers earlier than 21.19.411.0, see bug 1417442","creator":null,"homepageURL":null},"visible":true,"active":true,"userDisabled":false,"appDisabled":false,"installDate":1511259755824,"updateDate":1511259755824,"applyBackgroundUpdates":1,"bootstrap":true,"path":"C:\\Users\\developer\\AppData\\Roaming\\Mozilla\\Firefox\\Profiles\\iihxqiz5.release\\features\\{3409ba2a-24b7-4dc7-991d-0fb2106987e2}\\[e-mail]","skinnable":false,"size":7013,"sourceURI":"file:///C:/Users/DEVELO~1/AppData/Local/Temp/tmpaddon-387b02","releaseNotesURI":null,"softDisabled":false,"foreignInstall":false,"hasBinaryComponents":false,"strictCompatibility":false,"locales":[],"targetApplications":[{"id":"{ec8030f7-c20a-464f-9b0e-13a3a9e97384}","minVersion":"57.0","maxVersion":"57.0"}],"targetPlatforms":[],"multiprocessCompatible":true,"signedState":3,"seen":true,"dependencies":[],"hasEmbeddedWebExtension":false,"mpcOptedOut":false,"userPermissions":null,"icons":{},"iconURL":null,"icon64URL":null,"blocklistState":0,"blocklistURL":null}
1511259755869	DeferredSave.extensions.json	DEBUG	Save changes
1511259755870	DeferredSave.extensions.json	DEBUG	Save changes
1511259755871	addons.xpi	DEBUG	Loading bootstrap scope from C:\Users\developer\AppData\Roaming\Mozilla\Firefox\Profiles\iihxqiz5.release\features\{3409ba2a-24b7-4dc7-991d-0fb2106987e2}\[e-mail]
1511259755873	addons.xpi	DEBUG	Calling bootstrap method install on [e-mail] version 1.0
1511259755873	addons.xpi	DEBUG	Install of file:///C:/Users/DEVELO~1/AppData/Local/Temp/tmpaddon-387b02 completed.
1511259755874	addons.xpi	DEBUG	Registering manifest for C:\Users\developer\AppData\Roaming\Mozilla\Firefox\Profiles\iihxqiz5.release\features\{3409ba2a-24b7-4dc7-991d-0fb2106987e2}\[e-mail]
1511259755874	addons.xpi	DEBUG	Calling bootstrap method startup on [e-mail] version 1.0
1511259755875	DeferredSave.extensions.json	DEBUG	Starting timer
1511259755875	addons.xpi	DEBUG	removeTemporaryFile: file:///C:/Users/DEVELO~1/AppData/Local/Temp/tmpaddon-387b02 removing temp file C:\Users\DEVELO~1\AppData\Local\Temp\tmpaddon-387b02
Use of getAttributeNode() is deprecated. Use getAttribute() instead.

I can confirm that the preference media.windows-media-foundation.use-nv12-format isn't set to false...

I unzipped the file C:\Users\developer\AppData\Roaming\Mozilla\Firefox\Profiles\iihxqiz5.release\features\{3409ba2a-24b7-4dc7-991d-0fb2106987e2}\[e-mail] and check that it was using the right preference and it does call defaultBranch.setBoolPref("media.windows-media-foundation.use-nv12-format", false);

(I suspected that it was the wrong addon that got signed).

how I can debug this code to see what's going on?

Comment 19

[Jean-Yves Avenard [:jya]]

So after googling a bit...

I went into about:debugging, checked the "enable add-on debugging", clicked on Debug for the "Disable Media WMF NV12 Format"

that opened a debugger... but I don't see how I can force the code to run again...

Comment 20

[Jean-Yves Avenard [:jya]]

ok... I don't know how to debug, but I see what the problem is...

It uses the deviceID rather than the vendorID... that's not going to work.

Comment 21

[Jean-Yves Avenard [:jya]]

Attachment: [e-mail]

Comment 22

[Jean-Yves Avenard [:jya]]

Comment on attachment 8930421 [details]
[e-mail]

Could you resign this add-on please?

thank you

Comment 23

[Jean-Yves Avenard [:jya]]

Ok... I tested on Firefox Developer Edition 57b14 (I can't load unsigned xpi on release) and I can confirm that the new code works as expected.

Comment 24

[Jason Thomas [:jason]]

Attachment: [e-mail]

Please see attached.

Comment 25

[Andrei Vaida [:avaida]]

Is there an estimate for when the 'release-sysaddon' update channel will be ready again for testing? (with the updated xpi)

Comment 26

[Julien Cristau [:jcristau]]

(In reply to Andrei Vaida [:avaida], Desktop Release QA – please ni? me from comment #25)
> Is there an estimate for when the 'release-sysaddon' update channel will be
> ready again for testing? (with the updated xpi)

Comment 27

[Rehan Dalal (Inactive)]

Sorry I need the version number to be updated. We've already uploaded a v1.0 XPI.

Comment 28

[Jean-Yves Avenard [:jya]]

Attachment: [e-mail]

Please sign ...

(one day we'll get there :) )

Comment 29

[Jason Thomas [:jason]]

Attachment: [e-mail]

Please see attached.

Comment 30

[Rehan Dalal (Inactive)]

I don't think you updated the version number correctly. It is still set to 1.0. Please make sure you update the install.rdf file and not just the filename.

Comment 31

[Jean-Yves Avenard [:jya]]

damn... it was my zip command that was incorrect :(

Comment 32

[Jean-Yves Avenard [:jya]]

Attachment: [e-mail]

Comment 33

[Jean-Yves Avenard [:jya]]

3rd time lucky ? signing required please...

Comment 34

[Jason Thomas [:jason]]

Attachment: [e-mail]

Please see attached.

Comment 35

[Rehan Dalal (Inactive)]

This is now shipping to the test channel

Comment 36

[Sylvestre Ledru [:Sylvestre]]

Ryan, could we get help to test that?
We could push today if green!

Comment 37

[Jean-Yves Avenard [:jya]]

I followed the steps in the test document..

1511294648373	addons.update-checker	DEBUG	Requesting https://versioncheck-bg.addons.mozilla.org/update/VersionCheck.php?reqVersion=2&id=[e-mail]&version=&maxAppVersion=%ITEM_MAXAPPVERSION%&status=userEnabled,incompatible&appID={ec8030f7-c20a-464f-9b0e-13a3a9e97384}&appVersion=57.0&appOS=WINNT&appABI=x86_64-msvc&locale=en-US&currentAppVersion=57.0&updateType=%UPDATE_TYPE%&compatMode=%COMPATIBILITY_MODE%
1511294648376	addons.productaddons	INFO	downloadXHR File download. status=200
1511294648382	addons.productaddons	INFO	Downloaded file will be saved to C:\Users\DEVELO~1\AppData\Local\Temp\tmpaddon-61872c
1511294648567	addons.xpi	INFO	Starting system add-on update check from https://aus5.mozilla.org/update/3/SystemAddons/57.0/20171112125346/WINNT_x86_64-msvc-x64/en-US/release-sysaddon/Windows_NT%2010.0.0.0%20(x64)/default/default/update.xml.
1511294648567	addons.productaddons	INFO	sending request to: https://aus5.mozilla.org/update/3/SystemAddons/57.0/20171112125346/WINNT_x86_64-msvc-x64/en-US/release-sysaddon/Windows_NT%2010.0.0.0%20(x64)/default/default/update.xml
1511294648777	addons.productaddons	INFO	Completed downloading document
1511294649302	addons.productaddons	INFO	downloadXHR File download. status=200
1511294649304	addons.productaddons	INFO	Downloaded file will be saved to C:\Users\DEVELO~1\AppData\Local\Temp\tmpaddon-f957ae
1511294649339	addons.xpi	INFO	Installing new system add-on set
1511294649347	addons.repository	DEBUG	cacheAddons: enabled true IDs ["[e-mail]"]
1511294649348	addons.repository	DEBUG	Requesting https://services.addons.mozilla.org/en-US/firefox/api/1.5/search/guid:disable-media-wmf-nv12%40mozilla.org?src=firefox&appOS=WINNT&appVersion=57.0
1511294649550	addons.repository	DEBUG	Got metadata search load event
1511294649550	DeferredSave.addons.json	DEBUG	Save changes
1511294649550	DeferredSave.addons.json	DEBUG	Starting timer
1511294649601	DeferredSave.addons.json	DEBUG	Starting write
1511294649603	DeferredSave.addons.json	DEBUG	Write succeeded
1511294649607	addons.xpi	DEBUG	Starting install of [e-mail] from file:///C:/Users/DEVELO~1/AppData/Local/Temp/tmpaddon-f957ae
1511294649608	addons.manager	DEBUG	Background update check complete
1511294649611	addons.xpi	DEBUG	Addon [e-mail] will be installed as a packed xpi
1511294649615	addons.xpi	INFO	Mapping [e-mail] to C:\Users\developer\AppData\Roaming\Mozilla\Firefox\Profiles\8j5vbkwi.test\features\{cb8c61a8-f1e1-4b91-9ead-43ddf7268ac0}\[e-mail]
1511294649616	addons.xpi-utils	DEBUG	Make addon app-system-addons:[e-mail] visible
1511294649617	DeferredSave.extensions.json	DEBUG	Save changes
1511294649618	DeferredSave.extensions.json	DEBUG	Save changes
1511294649618	addons.xpi	DEBUG	XPIStates adding add-on [e-mail] in app-system-addons: C:\Users\developer\AppData\Roaming\Mozilla\Firefox\Profiles\8j5vbkwi.test\features\{cb8c61a8-f1e1-4b91-9ead-43ddf7268ac0}\[e-mail]
1511294649618	addons.xpi	DEBUG	Updating XPIState for {"id":"[e-mail]","syncGUID":"{d40ec170-b31b-4123-a598-e5ea825dfa8f}","location":"app-system-addons","version":"1.1","type":"extension","internalName":null,"updateURL":null,"updateKey":null,"optionsURL":null,"optionsType":null,"aboutURL":null,"defaultLocale":{"name":"Disable Media WMF NV12 format","description":"This disables NV12 surface format on Windows for AMD graphic adapters, see bug 1417442","creator":null,"homepageURL":null},"visible":true,"active":true,"userDisabled":false,"appDisabled":false,"installDate":1511294649607,"updateDate":1511294649607,"applyBackgroundUpdates":1,"bootstrap":true,"path":"C:\\Users\\developer\\AppData\\Roaming\\Mozilla\\Firefox\\Profiles\\8j5vbkwi.test\\features\\{cb8c61a8-f1e1-4b91-9ead-43ddf7268ac0}\\[e-mail]","skinnable":false,"size":7005,"sourceURI":"file:///C:/Users/DEVELO~1/AppData/Local/Temp/tmpaddon-f957ae","releaseNotesURI":null,"softDisabled":false,"foreignInstall":false,"hasBinaryComponents":false,"strictCompatibility":false,"locales":[],"targetApplications":[{"id":"{ec8030f7-c20a-464f-9b0e-13a3a9e97384}","minVersion":"57.0","maxVersion":"57.0"}],"targetPlatforms":[],"multiprocessCompatible":true,"signedState":3,"seen":true,"dependencies":[],"hasEmbeddedWebExtension":false,"mpcOptedOut":false,"userPermissions":null,"icons":{},"iconURL":null,"icon64URL":null,"blocklistState":0,"blocklistURL":null}
1511294649619	DeferredSave.extensions.json	DEBUG	Save changes1511294649620	addons.xpi	DEBUG	Loading bootstrap scope from C:\Users\developer\AppData\Roaming\Mozilla\Firefox\Profiles\8j5vbkwi.test\features\{cb8c61a8-f1e1-4b91-9ead-43ddf7268ac0}\[e-mail]
1511294649623	addons.xpi	DEBUG	Calling bootstrap method install on [e-mail] version 1.1
1511294649623	addons.xpi	DEBUG	Install of file:///C:/Users/DEVELO~1/AppData/Local/Temp/tmpaddon-f957ae completed.
1511294649623	addons.xpi	DEBUG	Registering manifest for C:\Users\developer\AppData\Roaming\Mozilla\Firefox\Profiles\8j5vbkwi.test\features\{cb8c61a8-f1e1-4b91-9ead-43ddf7268ac0}\[e-mail]
1511294649623	addons.xpi	DEBUG	Calling bootstrap method startup on [e-mail] version 1.1
1511294649624	DeferredSave.extensions.json	DEBUG	Starting timer
1511294649624	addons.xpi	DEBUG	removeTemporaryFile: file:///C:/Users/DEVELO~1/AppData/Local/Temp/tmpaddon-f957ae removing temp file C:\Users\DEVELO~1\AppData\Local\Temp\tmpaddon-f957ae
1511294649649	DeferredSave.extensions.json	DEBUG	Starting write
1511294649651	DeferredSave.extensions.json	DEBUG	Write succeeded
1511294649651	addons.xpi-utils	DEBUG	XPI Database saved, setting schema version preference to 23


checking the value of the preference in about:config:
media.windows-media-foundation.use-nv12-format;false

So TO-1 is verified.

I can't do TO-2 as my hardware doesn't suffer from the original problem (even with older amd drivers)

The confidence level that setting the preference to false will fix all those issues is extremely high however, it reverts to the code path used in FF 56 (nv12 was only used on intel devices in 56)

Comment 38

[Robert Helmer [:rhelmer]]

I've just verified this as well on a VM without the AMD vendor ID on the release-sysaddon channel, and media.windows-media-foundation.use-nv12-format remains true as expected.

Comment 39

[Liz Henry (:lizzard) (relman/hg->git project)]

That sounds promising, let's see if QE can do more testing (I'm hoping they have the hardware for it) and I can check back later today for signoff.

Comment 40

[Tom Grabowski [:TomGrab]]

QE was assigned to this bug and will post an update soon.

Comment 41

[Justin [:JW_SoftvisionQA]]

Attachment: Installing the Addon

When installing the signed addon we are seeing this error.

We are on the release channel.

Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:57.0) Gecko/20100101 Firefox/57.0

Comment 42

[Rehan Dalal (Inactive)]

You need to change "extensions.legacy.enabled" to true in about:config.

Comment 43

[Rehan Dalal (Inactive)]

Alternately you can just change to release-sysaddon and you should get the latest version (v1.1).

Comment 44

[Justin [:JW_SoftvisionQA]]

We have tried setting extensions.legacy.enabled to true in about:config on Release and also changing to release-sysaddon. Neither of these is showing the addon or allowing us to install the addon.

Comment 45

[Justin [:JW_SoftvisionQA]]

We were able to download the addon now. sorry for the miscommunication!

Comment 46

[Rehan Dalal (Inactive)]

So the "release-sysaddon" channel worked?

Comment 47

[Justin [:JW_SoftvisionQA]]

correct. At first we were not running the command

Components.utils.import("resource://gre/modules/AddonManager.jsm"); AddonManagerPrivate.backgroundUpdateCheck();

We are a bit confused on what exactly to test with this issue. Are we verifying what is stated in Comment 12 or the google sheet provided in Comment 7?

Comment 48

[Rehan Dalal (Inactive)]

That would be a question for rhelmer or Jean-Yves.

Comment 49

[u554753]

Per Comment 48.

Comment 50

[Justin [:JW_SoftvisionQA]]

Our testing can be found below:

https://public.etherpad-mozilla.org/p/WMF

We failed the test case where the driver AMD Radeon(TM) R3 Graphics 22.19.162.4 caused media.windows-media-foundation.use-nv12-format to flip to false. It is our understanding that media.windows-media-foundation.use-nv12-format is only suppose to flip to false if the AMD driver is earlier than 21.19.411.0.

Comment 51

[Justin [:JW_SoftvisionQA]]

Per Rhelmer: The version requirement was dropped. This is a go. Everything looks good from QA.

Comment 52

[Liz Henry (:lizzard) (relman/hg->git project)]

OK, sounds like we are ready to push this live. Thanks very much for the fast testing!

Comment 53

[Rehan Dalal (Inactive)]

The rule has been created in Balrog and pending signoff from QA/relman

Comment 54

[Liz Henry (:lizzard) (relman/hg->git project)]

Stefan and I signed off on balrog a few hours ago, so this should be live.

Comment 55

[Florian Quèze [:florian]]

It's unfortunate that this system add-on accesses TelemetryEnvironment.currentEnvironment during early startup (at a time when it's not ready yet, and has to be computed) for each startup.

The code in the add-on is:
function startup() {
    if (AppConstants.platform == "win") {
        let telemetryEnv = TelemetryEnvironment.currentEnvironment;
        let vendorID = telemetryEnv.system.gfx.adapters[0].vendorID;

        if (vendorID == BAD_VENDOR_ID) {
            let defaultBranch = Services.prefs.getDefaultBranch("");
            defaultBranch.setBoolPref("media.windows-media-foundation.use-nv12-format", false);
        }
    }
}

It would be nicer to cache the result and query the telemetry environment only when the add-on is installed. Is this add-on going to be disabled as soon as users update to 58?

Comment 56

[Robert Helmer [:rhelmer]]

(In reply to Florian Quèze [:florian] from comment #55)
> It's unfortunate that this system add-on accesses
> TelemetryEnvironment.currentEnvironment during early startup (at a time when
> it's not ready yet, and has to be computed) for each startup.
> 
> The code in the add-on is:
> function startup() {
>     if (AppConstants.platform == "win") {
>         let telemetryEnv = TelemetryEnvironment.currentEnvironment;
>         let vendorID = telemetryEnv.system.gfx.adapters[0].vendorID;
> 
>         if (vendorID == BAD_VENDOR_ID) {
>             let defaultBranch = Services.prefs.getDefaultBranch("");
>            
> defaultBranch.setBoolPref("media.windows-media-foundation.use-nv12-format",
> false);
>         }
>     }
> }
> 
> It would be nicer to cache the result and query the telemetry environment
> only when the add-on is installed.


Good point, thanks... we should really make it easier to run these through Talos (and the rest of the tests), since there are lots of pitfalls like this (accidentally being the first to initialize a lazy component). Right now it must be done manually by the developer, and it's easy to miss things like this.

I have some ideas for how we can change our process so doing try runs is automatic, would you mind filing a bug and assigning to me?

> Is this add-on going to be disabled as
> soon as users update to 58?

Yes. This is the default for system add-on updates (build ID change causes updates to be invalidated)

Comment 57

[Florian Quèze [:florian]]

(In reply to Robert Helmer [:rhelmer] from comment #56)

> I have some ideas for how we can change our process so doing try runs is
> automatic, would you mind filing a bug and assigning to me?

Filed bug 1427858.