captureTouchpads segfaults

[guiniol]

Describe the bug

facter panics in captureTouchpads. This happens both at v0.3.0 and HEAD. The trace below is from HEAD:

2025/01/03 21:18:29 INFO building report system=x86_64-linux version=1
2025/01/03 21:18:29 INFO scanning hardware features="[memory pci net serial cpu bios monitor scsi usb prom sbus sys sysfs udev block wlan]"
panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0xa pc=0x520f86]

goroutine 1 [running]:
github.com/numtide/nixos-facter/pkg/hwinfo.captureTouchpads(0x4a)
        github.com/numtide/nixos-facter/pkg/hwinfo/input.go:21 +0xc6
github.com/numtide/nixos-facter/pkg/hwinfo.Scan({0xc000098080, 0x10, 0x6f5780?})
        github.com/numtide/nixos-facter/pkg/hwinfo/hwinfo.go:82 +0x59c
github.com/numtide/nixos-facter/pkg/facter.(*Scanner).Scan(0x6d3c00)
        github.com/numtide/nixos-facter/pkg/facter/facter.go:76 +0x273
github.com/numtide/nixos-facter/cmd.Execute()
        github.com/numtide/nixos-facter/cmd/root.go:118 +0x2fd
main.main()
        github.com/numtide/nixos-facter/main.go:8 +0xf

To Reproduce

Steps to reproduce the behavior:

1. nix run --option experimental-features "nix-command flakes" #nixos-facter -- -o facter.json

Expected behavior

nixos-facter produces facter.json without segfaulting.

System information

Instead here are the list of parts, and I am also attaching a facter.json generated with v0.2.0.

• Asrock B550m Pro 4
• Ryzen 5 Pro 5650GE
• 2x Kingston KSM32ED8/32HC 32GB
• Kioxia Exceria G2 1TB (boot drive)
• 6x PM863A 3.84TB
• Connectx4-Lx 2*25Gb with a single 10G fiber transceiver
• HDPLEX 250W GaN Passive
• Streacom FC10 ST-FC10B-ALPHA

I git bisect-ed it down to b7473dd, which doesn't really help since it's the commit that adds input.go.
I thought maybe it would help if it had a peripheral, but plugging a keyboard didn't change anything. This is a headless and fanless system, so the keyboard and the bootable USB stick are the only things plugged in.

[brianmcgee]

Thanks for the report. I should have some time next week to look into this.

[brianmcgee]

I believe this will be resolved by #172 . I changed how I read udev data and it should account for the nil pointer happening here.

[guiniol]

I'm afraid I still get a SEGFAULT at HEAD (and yes, HEAD is after #172 ):

$ sudo nix run .#nixos-facter -- -o facts.json
2025/01/09 22:09:25 INFO building report system=x86_64-linux version=1
2025/01/09 22:09:25 INFO scanning hardware features="[memory pci net serial cpu bios monitor scsi usb prom sbus sys sysfs udev block wlan]"
panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0xa pc=0x520627]

goroutine 1 [running]:
github.com/numtide/nixos-facter/pkg/hwinfo.captureTouchpads(0x48)
        github.com/numtide/nixos-facter/pkg/hwinfo/input.go:34 +0x187
github.com/numtide/nixos-facter/pkg/hwinfo.Scan({0xc00008c080, 0x10, 0x6f4680?})
        github.com/numtide/nixos-facter/pkg/hwinfo/hwinfo.go:82 +0x59c
github.com/numtide/nixos-facter/pkg/facter.(*Scanner).Scan(0x6d2c00)
        github.com/numtide/nixos-facter/pkg/facter/facter.go:76 +0x273
github.com/numtide/nixos-facter/cmd.Execute()
        github.com/numtide/nixos-facter/cmd/root.go:118 +0x2fd
main.main()
        github.com/numtide/nixos-facter/main.go:8 +0xf

Could you reopen the issue?
How can I help you debug this issue?

[brianmcgee]

@guiniol I'll add some logging tomorrow and then have you run it again to work out what's going on.

I see it's failing after pulling udev data. I'll examine the underlying code to see if there's any obvious edge cases.

[guiniol]

Thanks! Happy to help however I can.

[brianmcgee]

@guiniol please try #174

[guiniol]

It works!!

2025/01/11 22:56:06 INFO building report system=x86_64-linux version=1
2025/01/11 22:56:06 INFO scanning hardware features="[memory pci net serial cpu bios monitor scsi usb prom sbus sys sysfs udev block wlan]"
2025/01/11 22:56:07 INFO reading IOMMU groups
2025/01/11 22:56:07 INFO processing devices count=67
2025/01/11 22:56:07 INFO processing smbios entries count=27
2025/01/11 22:56:07 INFO detecting virtualisation
2025/01/11 22:56:07 INFO report complete

(and sorry for the delay, couldn't make it to the computer yesterday)

[guiniol]

Ran it again with --log-level debug and I see the log line you add in #174:

2025/01/11 23:12:52 input.go:35: DEBUG udev data missing input name="ASRock LED Controller" sysfs=/devices/pci0000:00/0000:00:02.1/0000:02:00.0/usb1/1-8/1-8:1.0/0003:26CE:01A2.0002/input/input2

I didn't even know I had a LED controller, but that's what was causing the issue 😄

[Mic92]

Can be closed than.

[guiniol]

Awesome. Thanks again!