588298 - During mochitest-5 SSL mixedcontent tests, we spam tons of "###!!! ASSERTION: QueryInterface needed: 'query_result.get() == mRawPtr'" in nsCOMPtr<nsIX509Cert>::Assert_NoQueryNeeded

Status: RESOLVED FIXED

(Core :: Security: PSM, defect)

Description

[Daniel Holbert [:dholbert]]

We spam tons of this assertion-failure during SSL-mixedcontent mochitests:
> ###!!! ASSERTION: QueryInterface needed: 'query_result.get() == mRawPtr', file ../../../../dist/include/nsCOMPtr.h, line 544

These are all from this mochitest directory:
  security/ssl/mixedcontent/

For example, the (green) mochitest log linked in URL field contains 896 instances of this assertion-failure.

I don't have a lot of experience with this particular assertion, but it looks a bit scary to me (and probably easily fixable...).

Here's a sample backtrace of the failure:
{
###!!! ASSERTION: QueryInterface needed: 'query_result.get() == mRawPtr', file ../../../../dist/include/nsCOMPtr.h, line 544
nsCOMPtr<nsIX509Cert>::Assert_NoQueryNeeded [nsCOMPtr.h:544]
nsCOMPtr<nsIX509Cert>::operator= [nsCOMPtr.h:665]
nsNSSSocketInfo::Read [security/manager/ssl/src/nsNSSIOLayer.cpp:710]
nsBinaryInputStream::ReadObject [xpcom/io/nsBinaryStream.cpp:768]
NS_DeserializeObject [netwerk/base/src/nsSerializationHelper.cpp:106]
nsDiskCacheEntry::CreateCacheEntry [netwerk/cache/nsDiskCacheEntry.cpp:90]
nsDiskCacheDevice::FindEntry [netwerk/cache/nsDiskCacheDevice.cpp:484]
nsCacheService::SearchCacheDevices [netwerk/cache/nsCacheService.cpp:1369]
nsCacheService::ActivateEntry [netwerk/cache/nsCacheService.cpp:1273]
nsCacheService::ProcessRequest [netwerk/cache/nsCacheService.cpp:1153]
nsCacheService::OpenCacheEntry [netwerk/cache/nsCacheService.cpp:1238]
nsCacheSession::OpenCacheEntry [netwerk/cache/nsCacheSession.cpp:98]
nsHttpChannel::OpenCacheEntry [netwerk/protocol/http/nsHttpChannel.cpp:1865]
nsHttpChannel::Connect [netwerk/protocol/http/nsHttpChannel.cpp:217]
nsHttpChannel::AsyncOpen [netwerk/protocol/http/nsHttpChannel.cpp:3245]
nsURILoader::OpenURI [uriloader/base/nsURILoader.cpp:863]
nsDocShell::DoChannelLoad [docshell/base/nsDocShell.cpp:8862]
nsDocShell::DoURILoad [docshell/base/nsDocShell.cpp:8704]
nsDocShell::InternalLoad [docshell/base/nsDocShell.cpp:8374]
nsDocShell::LoadURI [docshell/base/nsDocShell.cpp:1414]
nsLocation::SetURI [dom/base/nsLocation.cpp:317]
nsLocation::SetHrefWithBase [dom/base/nsLocation.cpp:591]
nsLocation::Assign [dom/base/nsLocation.cpp:871]
}

Despite the mention of "nsNSSIOLayer.cpp:710" in the backtrace, I'm pretty sure it's the line of code before that one (line 706) that triggers the failure.  (I caught an instance of this assertion failure, and stepping up put me at that line.):

>706  mCert = reinterpret_cast<nsNSSCertificate*>(serializable.get());
http://mxr.mozilla.org/mozilla-central/source/security/manager/ssl/src/nsNSSIOLayer.cpp?mark=706-706#706

This is in a block of code added just last week in bug 570616.

Comment 1

[Daniel Holbert [:dholbert]]

(In reply to comment #0)
> (I caught an instance of this assertion failure, and stepping up
(caught it in GDB, I mean)

Comment 2

[Honza Bambas (:mayhemer)]

-> me

Comment 3

[Robert O'Callahan (:roc) (email my personal email if necessary)]

I get zillions of these while I'm dogfooding debug builds. It's super-annoying. Please fix!

Comment 4

[Honza Bambas (:mayhemer)]

Attachment: v1 [Check-in comment 6]

Very very simple fix.

Comment 5

[Kai Engert (:KaiE:)]

Comment on attachment 469160 [details] [diff] [review]
v1 [Check-in comment 6]

r=kaie Thanks Honza

Comment 6

[Honza Bambas (:mayhemer)]

Comment on attachment 469160 [details] [diff] [review]
v1 [Check-in comment 6]

http://hg.mozilla.org/mozilla-central/rev/d906347e3a36