Note: This is a public test instance of Red Hat Bugzilla. The data contained within is a snapshot of the live data so any changes you make will not be reflected in the production Bugzilla. Email is disabled so feel free to test any aspect of the site that you want. File any problems you find or give feedback at bugzilla.redhat.com.
Bug 1934841 - [aarch64] Firefox quits with pci_generic_error
Summary: [aarch64] Firefox quits with pci_generic_error
Keywords:
Status: NEW
Alias: None
Product: Fedora
Classification: Fedora
Component: firefox
Version: 34
Hardware: aarch64
OS: Linux
unspecified
medium
Target Milestone: ---
Assignee: Gecko Maintainer
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks: ARMTracker
TreeView+ depends on / blocked
 
Reported: 2021-03-03 21:20 UTC by customercare
Modified: 2021-03-08 11:14 UTC (History)
13 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed:
Type: Bug
Embargoed:

Attachments (Terms of Use)

Description customercare 2021-03-03 21:20:36 UTC 
Description of problem:


Mär 03 22:16:50 fedorapine systemd-coredump[2902]: [🡕] Process 2900 (firefox) of user 1000 dumped core.
                                                    
                                                    Stack trace of thread 2900:
                                                    #0  0x0000007fac1b9544 _ZL14UnexpectedExitv (libxul.so + 0x40e8544)
                                                    #1  0x0000007fb80b7194 __run_exit_handlers (libc.so.6 + 0x3b194)
                                                    #2  0x0000007fb80b7324 exit (libc.so.6 + 0x3b324)
                                                    #3  0x0000007fb5074b20 pci_generic_error (libpci.so.3 + 0x2b20)
                                                    #4  0x0000007fb507630c pci_init (libpci.so.3 + 0x430c)
                                                    #5  0x0000007fac1c7cb0 childgltest (libxul.so + 0x40f6cb0)
                                                    #6  0x0000007fac1c8418 _Z20fire_glxtest_processv (libxul.so + 0x40f7418)
                                                    #7  0x0000007fac1bcd34 _ZN7XREMain12XRE_mainInitEPb (libxul.so + 0x40ebd34)
                                                    #8  0x0000007fac1c3944 _ZN7XREMain8XRE_mainEiPPcRKN7mozilla15BootstrapConfigE (libxul.so + 0x40f2944)
                                                    #9  0x0000007fac1c3d4c _Z8XRE_mainiPPcRKN7mozilla15BootstrapConfigE (libxul.so + 0x40f2d4c)
                                                    #10 0x000000558f10baf0 _ZL7do_mainiPPcS0_ (firefox + 0xbaf0)
                                                    #11 0x000000558f10aecc main (firefox + 0xaecc)
                                                    #12 0x0000007fb80a0adc __libc_start_main (libc.so.6 + 0x24adc)
                                                    #13 0x000000558f10b1b8 _start (firefox + 0xb1b8)


depending on the cmd arguments a window opens, or not and ~70 processes hanging around.

Version-Release number of selected component (if applicable):

Name        : firefox
Version     : 85.0
Release     : 11.fc34
Architecture: aarch64


How reproducible:

just starting ...
Comment 1 Eric Sandeen 2021-03-04 18:22:34 UTC 
firefox, not fio.
Comment 2 Martin Stransky 2021-03-05 10:48:38 UTC 
It's a problem with libpci:

#3  0x0000007fb5074b20 pci_generic_error (libpci.so.3 + 0x2b20)
#4  0x0000007fb507630c pci_init (libpci.so.3 + 0x430c)

Firefox uses that to detect VGA adapters. I see it's aarch64 - may that be a factor here?
Comment 3 Martin Stransky 2021-03-05 10:51:52 UTC 
Michal, I see you did some work on libpci, do you have any idea how to handle that?
Thanks.
Comment 4 customercare 2021-03-05 12:35:57 UTC 
The device was a pinephone, which ode snot even have the bus interface to query the pci devices.

The patch needs to check that the pci interface iteself exists, before querying the devices.
Comment 5 Peter Robinson 2021-03-05 13:40:28 UTC 
> Firefox uses that to detect VGA adapters. I see it's aarch64 - may that be a
> factor here?

That won't work for a lot of ARM devices because most display GPUs and related HW aren't attached by pci so lspci or associated API won't return a display adapter.
Comment 6 Michal Hlavinka 2021-03-08 10:45:27 UTC 
If there is no pci bus on that device it won't work. You should probably check it in advance if it exists or not. If you can't do that, you can replace pci_access->error method between pci_alloc(...) and pci_init(...) to set some global flag, test it if initialization failed and handle it the way you need.
Comment 7 Peter Robinson 2021-03-08 11:14:29 UTC 
> Firefox uses that to detect VGA adapters. I see it's aarch64 - may that be a
> factor here?

By "detect VGA adapters" do you mean the model of the GPU to work out some capability? What functionality is this used for?
Note You need to log in before you can comment on or make changes to this bug.