Bug 127877 – Allow the OpenGL Profiler to get through the sandbox

• Status: RESOLVED
• Resolution: FIXED
• Component: WebKit2

[Dean Jackson]

We need to open the sandbox slightly so that the OpenGL profiler can attach to WebProcess and the PluginProcess.

<rdar://problem/14817250>

[Dean Jackson]

Created attachment 222611 [details]
Patch

[Sam Weinig]

I thought we decided not to do this? What has changed?

[Build Bot]

Comment on attachment 222611 [details]
Patch

Attachment 222611 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://webkit-queues.appspot.com/results/6717990879887360

New failing tests:
compositing/checkerboard.html
compositing/absolute-inside-out-of-view-fixed.html
animations/3d/matrix-transform-type-animation.html
http/tests/cache/cancel-multiple-post-xhrs.html
animations/3d/state-at-end-event-transform.html
animations/added-while-suspended.html
animations/animation-add-events-in-handler.html
animations/additive-transform-animations.html
animations/3d/replace-filling-transform.html
accessibility/alt-tag-on-image-with-nonimage-role.html
compositing/bounds-in-flipped-writing-mode.html
accessibility/accessibility-node-reparent.html
animations/animation-border-overflow.html
accessibility/accessibility-object-detached.html
animations/animation-controller-drt-api.html
animations/3d/change-transform-in-end-event.html
http/tests/cache/content-type-ignored-during-revalidation.html
compositing/absolute-position-changed-in-composited-layer.html
http/tests/cache/cancel-during-revalidation-succeeded.html
canvas/philip/tests/2d.canvas.readonly.html
http/tests/cache/cancel-during-failure-crash.html
canvas/philip/tests/2d.canvas.reference.html
canvas/philip/tests/2d.clearRect+fillRect.alpha0.5.html
animations/3d/transform-origin-vs-functions.html
accessibility/accessibility-node-memory-management.html
http/tests/cache/cached-main-resource.html
accessibility/adjacent-continuations-cause-assertion-failure.html
canvas/philip/tests/2d.clearRect+fillRect.basic.html
compositing/absolute-position-changed-with-composited-parent-layer.html
canvas/philip/tests/2d.clearRect+fillRect.alpha0.html

[Build Bot]

Created attachment 222619 [details]
Archive of layout-test-results from webkit-ews-13 for mac-mountainlion-wk2

The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews.
Bot: webkit-ews-13  Port: mac-mountainlion-wk2  Platform: Mac OS X 10.8.5

[Dean Jackson]

(In reply to comment #2)
> I thought we decided not to do this? What has changed?

We talked to the CoreOS and OpenGL teams at Apple today, and worked out it was ok.

[Dean Jackson]

(In reply to comment #5)
> (In reply to comment #2)
> > I thought we decided not to do this? What has changed?
> 
> We talked to the CoreOS and OpenGL teams at Apple today, and worked out it was ok.

See the radar for more details.

[Dean Jackson]

(In reply to comment #3)
> (From update of attachment 222611 [details])
> Attachment 222611 [details] did not pass mac-wk2-ews (mac-wk2):
> Output: http://webkit-queues.appspot.com/results/6717990879887360

This patch has no code changes - I think this is a mistake.

[Sam Weinig]

(In reply to comment #6)
> (In reply to comment #5)
> > (In reply to comment #2)
> > > I thought we decided not to do this? What has changed?
> > 
> > We talked to the CoreOS and OpenGL teams at Apple today, and worked out it was ok.
> 
> See the radar for more details.

Ok. Let's chat about this tomorrow.

[Alexey Proskuryakov]

Comment on attachment 222611 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=222611&action=review

EWS is right - the current patch breaks launching WebProcess on Mountain Lion. Here is what I see in the uploaded results archive:

WebProcess: Couldn't initialize sandbox profile [/Volumes/Data/EWS/WebKit/WebKitBuild/Release/WebKit2.framework/Resources/com.apple.WebProcess.sb], error 'line 88: unbound variable: mach-register '

We should make this conditional on low level OS support with (if (defined? 'mach-register) ...).

> Source/WebKit2/ChangeLog:5
> +        <rdar://problem/14817250>

This patch also fixes <rdar://problem/14271180>, please add it to ChangeLog.

> Source/WebKit2/WebProcess/com.apple.WebProcess.sb.in:88
> +    (global-name-regex #"_oglprof_attach_.*"))

This rule is not quite what was suggested to us. Let's chat offline about why it is so.

[Dean Jackson]

Created attachment 222682 [details]
Patch

[Dean Jackson]

Thanks Alexey.

[Alexey Proskuryakov]

Comment on attachment 222682 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=222682&action=review

r=me

Please upload a fixed patch though to let EWS chew on it and see what happens on Mountain Lion.

> Source/WebKit2/Resources/PlugInSandboxProfiles/com.apple.WebKit.plugin-common.sb:233
> +    (allow mach-register (global-name-regex #â^_oglprof_attach_<[0-9]+>$")))

Looks like you've got a smart quote here, which will break everything.

> Source/WebKit2/WebProcess/com.apple.WebProcess.sb.in:88
> +    (allow mach-register (global-name-regex #â^_oglprof_attach_<[0-9]+>$")))

Ditto.

[Alexey Proskuryakov]

Comment on attachment 222682 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=222682&action=review

>> Source/WebKit2/Resources/PlugInSandboxProfiles/com.apple.WebKit.plugin-common.sb:233
>> +    (allow mach-register (global-name-regex #â^_oglprof_attach_<[0-9]+>$")))
> 
> Looks like you've got a smart quote here, which will break everything.

Also, you changed the string to what was suggested to us, but I never said it was correct! I'm not sure if it is. Did you verify that everything works on Mavericks?

[Alexey Proskuryakov]

Comment on attachment 222682 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=222682&action=review

>>> Source/WebKit2/Resources/PlugInSandboxProfiles/com.apple.WebKit.plugin-common.sb:233
>>> +    (allow mach-register (global-name-regex #â^_oglprof_attach_<[0-9]+>$")))
>> 
>> Looks like you've got a smart quote here, which will break everything.
> 
> Also, you changed the string to what was suggested to us, but I never said it was correct! I'm not sure if it is. Did you verify that everything works on Mavericks?

Actually, I misread something. Looks like it is correct.

[Build Bot]

Comment on attachment 222682 [details]
Patch

Attachment 222682 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://webkit-queues.appspot.com/results/5152626429657088

New failing tests:
compositing/checkerboard.html
compositing/absolute-inside-out-of-view-fixed.html
animations/3d/matrix-transform-type-animation.html
http/tests/cache/cancel-multiple-post-xhrs.html
animations/3d/state-at-end-event-transform.html
animations/added-while-suspended.html
animations/animation-add-events-in-handler.html
animations/additive-transform-animations.html
animations/3d/replace-filling-transform.html
accessibility/alt-tag-on-image-with-nonimage-role.html
compositing/bounds-in-flipped-writing-mode.html
accessibility/accessibility-node-reparent.html
animations/animation-border-overflow.html
accessibility/accessibility-object-detached.html
animations/animation-controller-drt-api.html
animations/3d/change-transform-in-end-event.html
http/tests/cache/content-type-ignored-during-revalidation.html
compositing/absolute-position-changed-in-composited-layer.html
http/tests/cache/cancel-during-revalidation-succeeded.html
canvas/philip/tests/2d.canvas.readonly.html
http/tests/cache/cancel-during-failure-crash.html
canvas/philip/tests/2d.canvas.reference.html
canvas/philip/tests/2d.clearRect+fillRect.alpha0.5.html
animations/3d/transform-origin-vs-functions.html
accessibility/accessibility-node-memory-management.html
http/tests/cache/cached-main-resource.html
accessibility/adjacent-continuations-cause-assertion-failure.html
canvas/philip/tests/2d.clearRect+fillRect.basic.html
compositing/absolute-position-changed-with-composited-parent-layer.html
canvas/philip/tests/2d.clearRect+fillRect.alpha0.html

[Build Bot]

Created attachment 222688 [details]
Archive of layout-test-results from webkit-ews-14 for mac-mountainlion-wk2

The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews.
Bot: webkit-ews-14  Port: mac-mountainlion-wk2  Platform: Mac OS X 10.8.5

[Dean Jackson]

Created attachment 222696 [details]
Patch

[Dean Jackson]

Chew!

[Dean Jackson]

Committed r163109: <http://trac.webkit.org/changeset/163109>