1246997 - Assertion failure: nsLayoutUtils::HasDisplayPort(content), at /home/billm/moz/gecko/gfx/layers/apz/util/APZCCallbackHelper.cpp:256

Status: RESOLVED DUPLICATE of bug 1245925

(Core :: Panning and Zooming, defect)

Description

[Bill McCloskey [inactive unless it's an emergency] (:billm)]

I'm hitting this a lot with a proprietary test case. I can supply it on request.

Here is the stack. The line numbers are a little off since this is an opt build.

#7  mozilla::layers::APZCCallbackHelper::UpdateRootFrame (aMetrics=...)
    at /home/billm/moz/gecko/gfx/layers/apz/util/APZCCallbackHelper.cpp:229
#8  0x00007f5f0e6c2a0d in mozilla::dom::TabChildBase::ProcessUpdateFrame (aFrameMetrics=..., this=<optimized out>)
    at /home/billm/moz/gecko/dom/ipc/TabChild.cpp:268
#9  0x00007f5f0e6cb535 in mozilla::dom::TabChildBase::UpdateFrameHandler (this=0x7f5ef3170000, aFrameMetrics=...)
    at /home/billm/moz/gecko/dom/ipc/TabChild.cpp:246
#10 0x00007f5f0d2356ed in mozilla::dom::PBrowserChild::OnMessageReceived (this=<optimized out>, msg__=...)
    at /home/billm/moz/gecko/objdir-ff-dbgopt/ipc/ipdl/PBrowserChild.cpp:2920

Comment 1

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

Is this a recent regression? It might be from bug 990916. See if setting the apz.displayport_expiry_ms pref to 0 helps. Note that bug 1246676 modified the pref to different values on different platforms.

Comment 2

[Bill McCloskey [inactive unless it's an emergency] (:billm)]

Setting apz.displayport_expiry_ms to 0 seems to make the problem go away. I've only started looking at this testcase today, so I don't know if it's recent. I'm not sure if mozregression works with debug builds, and I'm afraid I don't have time to do a manual regression range right now.

Comment 3

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

If setting apz.displayport_expiry_ms to 0 fixes the problem that's likely the issue. I'll take a look at this at some point in the near future, after we get telemetry data from bug 1246676.

Comment 4

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

I have a patch posted in bug 1245925, I would be interested to know if you still hit this assertion with that patch applied (and apz.displayport_expiry_ms set back to a nonzero value; it used to be 15000).

Comment 5

[Bill McCloskey [inactive unless it's an emergency] (:billm)]

Yes, seems fixed. Thanks!