1341807 - Crash when resizing perf.html, in mozilla::ipc::FatalError | libxul.so@0xc84c80 | mozilla::layers::PLayerTransactionParent::Read, with FatalError "Error deserializing 'sem' (CrossProcessSemaphoreHandle) member of 'CrossProcessSemaphoreDescriptor'"

Status: RESOLVED DUPLICATE of bug 1340117

(Core :: Graphics: Layers, defect)

Description

[Daniel Holbert [:dholbert]]

This bug was filed from the Socorro interface and is 
report bp-b0d894fd-df62-4bbc-a7d4-3f23e2170222.
=============================================================
STR:
 1. Visit https://perfht.ml/2l0H1DR
 2. Resize your Firefox window horizontally. (Make it a bit wider)

ACTUAL RESULTS:
Crash.


This seems to reproduce pretty (entirely?) reliably for me, in current Nightly, with a fresh profile.

Comment 1

[Daniel Holbert [:dholbert]]

If I trigger this in a debug build with gdb attached, it shows that we're crashing from an explicit MOZ_CRASH, inside of this FatalError call (in $obj/ipc/ipdl/PLayerTransactionParent.cpp line 2797):
=========
auto PLayerTransactionParent::Read(
        CrossProcessSemaphoreDescriptor* v__,
        const Message* msg__,
        PickleIterator* iter__) -> bool
{
    if ((!(Read((&((v__)->sem())), msg__, iter__)))) {
        FatalError("Error deserializing 'sem' (CrossProcessSemaphoreHandle) member of 'CrossProcessSemaphoreDescriptor'");
=========

Comment 2

[Daniel Holbert [:dholbert]]

...and after my debug build hits the crash, I get this additional output printed to my terminal:

Crash Annotation GraphicsCriticalError: |[C0][GFX1-]: Failed 2 buffer db=0 dw=0 for 0, 0, 64, 6 (t=178.998) |[C406][GFX1-]: Failed 2 buffer db=0 dw=0 for -1, 0, 64, 6 (t=179.059) |[C407][GFX1-]: Failed 2 buffer db=0 dw=0 for -1, 0, 64, 6 (t=179.059) |[C408][GFX1-]: Failed 2 buffer db=0 dw=0 for 0, 0, 64, 6 (t=179.059) |[C409][GFX1-]: Failed 2 buffer db=0 dw=0 for 0, 0, 64, 6 (t=179.06) |[C410][GFX1-]: Failed 2 buffer db=0 dw=0 for 0, 0, 64, 6 (t=179.06) |[C411][GFX1-]: Failed 2 buffer db=0 dw=0 for -1, 0, 64, 6 (t=179.06) |[C412][GFX1-]: Failed 2 buffer db=0 dw=0 for 0, 0, 64, 6 (t=179.06) |[C413][GFX1-]: Failed 2 buffer db=0 dw=0 for -1, 0, 64, 6 (t=179.06) |[C414][GFX1-]: Failed 2 buffer db=0 dw=0 for -1, 0, 64, 6 (t=179.06) |[C415][GFX1-]: Failed 2 buffer db=0 dw=0 for 0, 0, 64, 6 (t=179.06) |[C416][GFX1-]: Failed 2 buffer db=0 dw=0 for 0, 0, 64, 6 (t=179.06) |[C417][GFX1-]: Failed 2 buffer db=0 dw=0 for 0, 0, 64, 6 (t=179.06) |[C418][GFX1-]: Failed 2 buffer db=0 dw=0 for 0, 0, 64, 6 (t=179.06) |[C419][GFX1-]: Failed 2 buffer db=0 dw=0 for 0, 45, 1341, 77 (t=179.06) |[C420][GFX1-]: Failed 2 buffer db=0 dw=0 for 0, 0, 64, 6 (t=317.979)

Comment 3

[Markus Stange [:mstange]]

Nical, do you know anything about this?

Comment 4

[Daniel Holbert [:dholbert]]

This is a pretty recent regression. Range:
https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=bcda57aa0c3b291dcd5de5d7fd048af69dbdff4d&tochange=a19ac5da60828cdcf374e6798dc7e07b8b4c1b7d

--> caused by Bug 1325227. Possibly a dupe of bug 1340117.

Comment 5

[Nicolas Silva [:nical]]

I'll be on PTO for a week and didn't have time to figure this one out so moving the needinfo to Matt.

Comment 7

[Gerry Chang [:gchang]]

FF54 was fixed in bug 1340117. Mark 54 fixed.

Comment 8

[Daniel Holbert [:dholbert]]

(This is already marked as a dupe of that bug, and that bug has "status-firefox54: fixed" already, so let's just clear the (mostly-incomplete) set of flags here & share the tracking flags on the dupe-target.)