This bug is for crash report bp-9658ba5d-1d9e-43fd-bff1-3e2d40190912.

Top 10 frames of crashing thread:

0 xul.dll mozilla::ipc::FatalError ipc/glue/ProtocolUtils.cpp:165
1 xul.dll mozilla::ipc::IProtocol::HandleFatalError ipc/glue/ProtocolUtils.cpp:404
2 xul.dll static bool mozilla::ipc::IPDLParamTraits<mozilla::dom::WebProgressSecurityChangeData>::Read ipc/ipdl/PBrowser.cpp
3 xul.dll static bool mozilla::ipc::IPDLParamTraits<mozilla::Maybe<mozilla::dom::WebProgressSecurityChangeData> >::Read ipc/glue/IPDLParamTraits.h:259
4 xul.dll mozilla::dom::PBrowserParent::OnMessageReceived ipc/ipdl/PBrowserParent.cpp:5262
5 xul.dll mozilla::dom::PContentParent::OnMessageReceived ipc/ipdl/PContentParent.cpp:5853
6 xul.dll mozilla::ipc::MessageChannel::DispatchMessage ipc/glue/MessageChannel.cpp:2107
7 xul.dll mozilla::ipc::MessageChannel::MessageTask::Run ipc/glue/MessageChannel.cpp:1985
8 xul.dll nsThread::ProcessNextEvent xpcom/threads/nsThread.cpp:1225
9 xul.dll NS_ProcessNextEvent xpcom/threads/nsThreadUtils.cpp:486

these browser crash reports with ipc fatal error msg Error deserializing 'securityInfo' (nsITransportSecurityInfo) member of 'WebProgressSecurityChangeData' are newly showing up on windows in the firefox 70 beta cycle, perhaps related to the changes from bug 1510569.

crashes are more often happening in 32bit builds of the browser, the crashing url in most cases is pointing to mail.google.com and the modules section is contining references to kasparsky's dlls hooking into the browser.

on 70.0b this problem is accounting for ~40 crashes per day currently.

the crash signature might cover various different issues - this crash-stats query is just covering the particular type of crash this bug was filed for:
https://crash-stats.mozilla.com/search/?ipc_fatal_error_msg=%3DError%20deserializing%20%27securityInfo%27%20%28nsITransportSecurityInfo%29%20member%20of%20%27WebProgressSecurityChangeData%27&date=%3E%3D2019-06-01&_facets=signature&_facets=version&_facets=platform_pretty_version&_facets=cpu_arch&_facets=url#crash-reports

I installed a trial version of Kaspersky inside a VM with both a recent debug build and an official beta build. However, I could not reproduce the crash on either version of Firefox.

Andrei, can you find someone in QA who can give this a try? The crash volume is fairly high for beta. Thanks!

(In reply to Liz Henry (:lizzard) from comment #3)

Andrei, can you find someone in QA who can give this a try? The crash volume is fairly high for beta. Thanks!

Cristian is currently investigating this, he'll follow up here as soon as possible.

Unfortunately to no success.

Attempted to get the crash with the following configs:

• Win 8.1 - ESET smart security - PC with AMD_FX 8320 with ATI Radeon 3000 - Ffx 70.0b5;
• Win 10 x86 - ESET smart security & Kaspersky - PC with AMD_FX 8320 with ATI Radeon 3000 - Ffx 70.0b4, Ffx 70.0b8;
• Win 10 x86 - no AV -PC with AMD_FX 8320 with ATI Radeon 3000 - Ffx 70.0b8;
• Win 7 x86 - KasperskyEnterprise - PC with AMD_FX 8320 with ATI Radeon 3000 - Ffx 70.0b5;
• Win 10x64 with Kaspersky - (PC with)R& 370 & (laptop with)GTX 1060 - Ffx 70.0b8(x86).

Since the initial report had a mention of crashes happening around mail.google.com most of the investigation was around the mail/website.
Attempts of opening links from mails or normal use scenarios had no success in identifying a crash/fail.

can we reach out to kaspersky? maybe they have more insight if and how their software is modifying the behaviour of common sites used for communication (gmail, outlook, whatsapp, twitter)...

Url
1 https://mail.google.com/mail/u/0/#inbox 189 37.88 %
2 https://mail.google.com/mail/u/0/ 96 19.24 %
3 https://mail.google.com/mail/u/0/?tab=wm#inbox 28 5.61 %
4 https://web.whatsapp.com/ 17 3.41 %
5 https://outlook.live.com/mail/ 15 3.01 %
6 https://mail.google.com/mail/u/0/#inbox?compose=new 13 2.61 %
7 https://mail.google.com/mail/u/0/?tab=wm&ogbl 13 2.61 %
8 https://twitter.com/ 11 2.20 %

Fwiw, did a quick check with the provided links on Windows 7 with Kaspersky Endpoint 10 + Firefox 70.0b9 but still nothing from my side.

I'm emailing our Kaspersky contacts today.

The priority flag is not set for this bug.
:jimm, could you have a look please?

For more information, please visit auto_nag documentation.

(In reply to Liz Henry (:lizzard) from comment #8)

I'm emailing our Kaspersky contacts today.

They replied and are looking into the issue.

We believe this issue might be related not to loaded modules, but to certificate installation method used in our older unsupported product KAV/KIS 2017. In this version we use nss3.dll to install our certificate in the browser. Newer product versions no longer rely on this method and enable system certificate store use in browser settings instead.

Here is a list of APIs we use:

PR_Malloc
PR_Open
PR_Close
PR_Read
PR_GetOpenFileInfo

NSS_Initialize
NSS_Shutdown

PORT_ZAlloc
PORT_Free
PORT_GetError

CERT_GetDefaultCertDB
CERT_DecodeTrustString
CERT_DestroyCertificate
CERT_ChangeCertTrust

PK11_ImportCert
PK11_FindCertFromNickname
PK11_FreeSlot
PK11_GetInternalKeySlot
PK11_NeedUserInit
PK11_InitPin

SECITEM_AllocItem
SECITEM_FreeItem

SEC_DeletePermCertificate

CERT_DecodeCertFromPackage
CERT_CompareCerts
CERT_GetCertTrust

We open database in the following way:
NSS_Initialize("sql:cert9.db", "", "", "secmod.db", NSS_INIT_OPTIMIZESPACE | NSS_INIT_NOROOTINIT);

We believe it might be a new bug in NSS code, possibly some concurrency issue or a change in firefox APIs that breaks this installation method. It seems that it won’t be possible to fix this problem by blocking Kaspersky modules in the Firefox process.

Could you please:

1. Assess how many users are impacted by this problem with your data so we see if it’s critical to release a patch for unsupported product.
2. See if there are any changes in APIs we use in firefox 70.
3. Investigate a possible regression in NSS and other problems related to serialization and certificate use.

Hi Maxim,

Could you please provide more information on the type of certs being imported? Better yet, if you can provide a cert9.db sample with your certificates imported, we can investigate and test with recent NSS versions. Thanks.

User comments are all related to not being able to get to their email.

Maxim, is there a particular version you could provide to our QA team for testing?

How to reproduce

1. Install Firefox 77.0b11 (32-bit)
2. Launch startup.exe from KIS2017 folder here to Install KIS 2017. Do not agree with “Download newest version”. Agree with everything else and continue everywhere.
3. Open KIS window, skip promo, refuse to activate.
4. Start Firefox and try to navigate for some https site (https://google.com for example)
   Security issue warning will appear

You can find our cert example here

Security warning means that we failed to install our certificate. We believe that failed certificate installation may, under some unknown conditions, lead to observed crashes. We have failed to reproduce firefox crash itself.

Adding back Cristian to see if he can repro with the STR in Comment 15.

Thanks Marcia for the ping and Maxim for the reply!

I was able to notice the warning blocking the page access as well, was completely locked out of the google websites due to the warn pages.
With a fresh profile I wasn't even able to access the mail/links.
Unfortunately, no crash from my side as well.

Tried several use scenarios with the websites listed in Comment 6 but at no success:
restart, reload, multiple tab + pinTab /navigation, incognito, AV on/off toggle, installed the addon.

There where some stalls due to the Workstation/OS architecture, but those were pretty much expected.

Crash rate is moderate on 70.0 release and it is unlikely we'll fix this in a dot release.
It would still be great to fix in 71 or 72.

The priority flag is not set for this bug.
:marcia, could you have a look please?

For more information, please visit auto_nag documentation.

I have just hit this twice trying to get the text in https://gist.github.com/AutomatedTester/1616894e0912b4846968d2440bb51d72 into https://paste.mozilla.org

I gave up and went to github

(In reply to David Burns :automatedtester from comment #21)

I have just hit this twice trying to get the text in https://gist.github.com/AutomatedTester/1616894e0912b4846968d2440bb51d72 into https://paste.mozilla.org

I gave up and went to github

Crash report is https://crash-stats.mozilla.org/report/index/cee1d3e4-f11a-4c79-8c94-a374f0191115

Hey :automatedtester, this bug is specifically about failing to deserialize an nsITransportSecurityInfo, which I don't see in your crash report. You should probably file a new bug for thiat crash

(In reply to Barret Rennie [:barret] (he/him) from comment #23)

Hey :automatedtester, this bug is specifically about failing to deserialize an nsITransportSecurityInfo, which I don't see in your crash repoet. You should probably file a new bug for thiat crash

my bad!

(In reply to David Burns :automatedtester from comment #22)

(In reply to David Burns :automatedtester from comment #21)

I have just hit this twice trying to get the text in https://gist.github.com/AutomatedTester/1616894e0912b4846968d2440bb51d72 into https://paste.mozilla.org

I gave up and went to github

Crash report is https://crash-stats.mozilla.org/report/index/cee1d3e4-f11a-4c79-8c94-a374f0191115

Raised https://bugzilla.mozilla.org/show_bug.cgi?id=1596789

We have shipped our last beta for 71, I am keeping the status as fix-optional for 71 in case a safe uplift is possible in a future dot-release given that the volume of crashes is noticeable.

Likely wontfix for 71/72 at this point, volume is low.

Since the crash volume is low (less than 15 per week), the severity is downgraded to S3. Feel free to change it back if you think the bug is still critical.

For more information, please visit auto_nag documentation.