Bug 286670 - www/firefox-esr 128.10.0,1 segmentation fault upon startup using X-forwarding in jail
Summary: www/firefox-esr 128.10.0,1 segmentation fault upon startup using X-forwarding...
Status: New
Alias: None
Product: Ports & Packages
Classification: Unclassified
Component: Individual Port(s) (show other bugs)
Version: Latest
Hardware: amd64 Any
: --- Affects Only Me
Assignee: freebsd-gecko (Nobody)
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2025-05-08 12:05 UTC by c0141
Modified: 2025-05-08 12:05 UTC (History)
0 users

See Also:
bugzilla: maintainer-feedback? (gecko)


Attachments
output from lldb bt all (7.87 KB, text/plain)
2025-05-08 12:05 UTC, c0141
no flags Details

Note You need to log in before you can comment on or make changes to this bug.
Description c0141 2025-05-08 12:05:33 UTC
Created attachment 260262 [details]
output from lldb bt all

OS: FreeBSD 14.2 amd64
Hardware: a number Lenovo laptops

Firefox-esr is installed in a vnet jail which is configured to use X-forwarding either through a socket (null-mounted), with xhost and DISPLAY set or through ssh -X.

When firefox-esr is started it dies immediately with segmentation fault.

Example:
$ /usr/local/bin/firefox
Crash Annotation GraphicsCriticalError: |[0][GFX1-]: Failed to initialize shared font list, falling back to in-process list. (t=0.742647) [GFX1-]: Failed to initialize shared font list, falling back to in-process list.
Segmentation fault

I get a core dump which shows pretty much nothing. I will attach the output from bt all.

This problem seems to happen on firefox 136.x and newer, firefox-esr 128.x and librewolf 136.x and newer.

I first experienced this with librewolf version 136.x and later with firefox 138.x.

Yesterday I reinstalled my Lenovo T14 gen4 laptop with a fresh 14.2. 
It happens on 3 laptops I have tested this on (Lenovo T14 gen4, Lenovo Thinkpad Yoga 14 and a Lenovo x230). I have tested RAM on all three machines with memtest without issues.

On the Yoga laptop I have a jail with firefox 135.x and firefox 136.x. Firefox 135.x does not die with segmentation fault 136.x does. 
On the same machine I built librewolf 138.x (from ports) with debugging enabled. This makes the problem go away.

I don't know how to debug this any further. I am willing to test suggestions and provide feedback if the person asking for feedback is a bit patient (I can usually test and provide feedback in a week or two but sometimes life happens and it takes longer).