I am a bystander. Got intrigued by this.

To condense the reddit thread: Steam would crash when the user touched the keyboard. With troubleshooting, they diagnosed as the keyboard resetting when suddenly jolted with static electricity, which disconnects the USB and crashes steam. The steam crashing was reproduced by unplugging the keyboard.

Lines 716 and 717 give

Dec 15 16:57:53.978685 TumbleweedSystem kernel: usb 1-6.2: USB disconnect, device number 6
Dec 15 16:57:54.506273 TumbleweedSystem steam[4058]: free(): double free detected in tcache 2
Dec 15 16:57:54.548599 TumbleweedSystem crash_20241215165754_36.dmp[5821]: Uploading dump (out-of-process)
                                                                           /tmp/dumps/crash_20241215165754_36.dmp
Dec 15 16:57:54.548871 TumbleweedSystem steam[4058]: crash_20241215165754_36.dmp[5821]: Uploading dump (out-of-process)

Hello,

Please reproduce this with the latest beta client and provide a crash ID so we can investigate on our end.

Please reproduce this with the latest beta client and provide a crash ID so we can investigate on our end.

Hello, I will be attempting to complete this in about 4 hours and then I will update you. Where can I get the crash ID you are referring to?

They are printed to stdout when you run the steam client in a terminal.

You had a crash above but you didn't give the ID (crash_20241215165754_36.dmp[5821]: Uploading dump (out-of-process)). It would have printed a few lines below. That's why it's important to always give full logs :)

I'd prefer having a crash ID against the latest beta also, as it'll be more current with our development process.

How do I get that to print?

On Mon, Dec 16, 2024, 5:25 p.m. TTimo @.***> wrote:

They are printed to stdout when you run the steam client in a terminal.

You had a crash above but you didn't give the ID (crash_20241215165754_36.dmp[5821]:
Uploading dump (out-of-process)). It would have printed a few lines
below. That's why it's important to always give full logs :)

I'd prefer having a crash ID against the latest beta also, as it'll be
more current with our development process.

—
Reply to this email directly, view it on GitHub
https://github.com/ValveSoftware/steam-for-linux/issues/11577#issuecomment-2546980391,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/AAIECHH5MD4QTE6E2C5ZDU32F5HOHAVCNFSM6AAAAABTU7YZJSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKNBWHE4DAMZZGE
.
You are receiving this because you are subscribed to this thread.Message
ID: @.***>

It would have printed a few lines below. That's why it's important to always give full logs :)

the pastebin I posted had the entire logs I got from the command I was given (sudo journalctl -b 0 --no-pager --output=short-precise).
Same question as the above comment. How do I print that?

@eadavani can you reproduce with the beta version of the Steam client and provide an ID? There is indeed an crash ID for the stable release in the pastebin you provided.

I recreated the Steam crash while on the beta version. Here is the output from that command.
https://pastebin.com/MfH5JHXt
I believe the crash is somewhere around line 1900.

Or possibly line 1635

Thanks, I'm checking to see if there is some useful information we can extract from the crash report.

The steam crashing was reproduced by unplugging the keyboard.

FWIW I can unplug and replug my keyboard without crashing Steam, so there's gotta be something more specific to OP's system than that.

You may also want to run steam with MALLOC_CHECK_=0 (e.g. launch steam from the terminal as MALLOC_CHECK_=0 steam). Mostly out of curiosity if you still crash (likely not).

I will give that a try in about 4 hours and let you know.

Actually, try with MALLOC_CHECK_=1 that may even be better (will print a diagnostic message and attempt to continue).

so the full command to run is MALLOC_CHECK_=1 steam ?

Yes. Quite possibly it won't crash but it may also report some useful diagnostic info.

After launching Steam with that command, and while it was still running, I reproduced the crash. Everything should be in the link.
https://pastebin.com/N60KDwPK

Nothing useful then. Unfortunate. I thought those errors came strictly from glibc's malloc checking and that we'd at the very least get a somewhat better diagnostic. Thanks for trying this!

should I close this issue, or was there still more you would like to try?

That's ok we don't need to close it. It's still a known, active issue. We don't have reproduction and the crash dumps aren't much useful, so it's unlikely we'll make progress on this at this time, but still.