Crash in [@ IPCError-browser | RecvCreateBrowsingContext Parent has different group object]
Categories
(Core :: DOM: Navigation, defect)
Tracking
()
People
(Reporter: aryx, Unassigned, NeedInfo)
References
Details
(Keywords: crash)
Crash Data
This signature is only observed with Nightly (92 crashes for 63 installations of 90.0a1 over the past weeks).
Crash report: https://crash-stats.mozilla.org/report/index/906aaf86-401d-4b3b-bdfb-1b5220210517
Reason: EXC_BREAKPOINT / EXC_I386_BPT
Top 10 frames of crashing thread:
0 libmozglue.dylib arena_t::DallocSmall memory/build/mozjemalloc.cpp:3288
1 libmozglue.dylib free memory/build/malloc_decls.h:54
2 XUL mozilla::ipc::URIParams::MaybeDestroy ipc/ipdl/URIParams.cpp:701
3 XUL void mozilla::ipc::WriteIPDLParam<nsIURI*> ipc/glue/IPDLParamTraits.h:64
4 XUL mozilla::ipc::IPDLParamTraits<mozilla::dom::DocShellLoadStateInit>::Write ipc/ipdl/DOMTypes.cpp:2149
5 XUL mozilla::ipc::IPDLParamTraits<mozilla::net::DocumentChannelCreationArgs>::Write ipc/ipdl/NeckoChannelParams.cpp:3390
6 XUL mozilla::net::PNeckoChild::SendPDocumentChannelConstructor ipc/ipdl/PNeckoChild.cpp:1331
7 XUL mozilla::net::DocumentChannelChild::AsyncOpen netwerk/ipc/DocumentChannelChild.cpp:162
8 XUL nsURILoader::OpenURI uriloader/base/nsURILoader.cpp:696
9 XUL nsDocShell::InternalLoad docshell/base/nsDocShell.cpp:9610
Comment 1•3 years ago
|
||
This crash will probably be fixed by bug 1535913, which landed yesterday.
Comment 2•2 years ago
|
||
There are still some cases, it seems. Recent report
Updated•2 years ago
|
Comment 3•2 years ago
|
||
I'm confused, the reports doesn't seem to have anything to do with ContentParent::RecvCreateBrowsingContext
, :jstutte do you see something that I'm missing?
Comment 4•2 years ago
|
||
In fact this looks a bit unactionable as a report. IIUC, this is a content process kill initiated from the parent process by ContentParent::RecvCreateBrowsingContext
via IPC_FAIL
. This seems to be caused by this call inside BrowsingContext::Attach
, introduced/modified by bug 1580565. Being this an async send, the stack traces in the reports at the time the crash happens are not helpful and should not be looked at, only the IPCError-browser | RecvCreateBrowsingContext Parent has different group object
is interesting here. This makes it hard to understand, why we can get into this situation and if it is even the child's fault or an error in the parent's book-keeping.
Comment 5•1 year ago
|
||
S3 - according to the very low crash volume and not a security issue.
Let me know if the severity rating doesn't fit.
Comment 6•1 year ago
|
||
There might be some relation to bug 1815480 here (at least also there we might end up with a misaligned book keeping around BrowsingContextGroup
s).
Description
•