Steam gets stuck in a loop when starting

[PaddyMac]

I downloaded the latest git code from 4 days ago, compiled, and installed it. I also removed the FreeBSD package. I completely removed ~/.steam just to make sure nothing stale was in there. The complete output is in the attached text file. But, once the initial download and setup is completed, it repeats the following message ad infinitum:

[[sh]] [[-c]] [[exec '/usr/home/backstreet/.steam/steam/ubuntu12_64/steamwebhelper.sh.patched' '-lang=en_US' '-cachedir=/usr/home/backstreet/.steam/steam/config/htmlcache' '-steampid=16628' '-buildid=1689034492' '-steamid=0' '-logdir=/usr/home/backstreet/.steam/steam/logs' '-uimode=7' '-startcount=19' '-steamuniverse=Public' '-realm=Global' '-clientui=/usr/home/backstreet/.steam/steam/clientui' '-steampath=/usr/home/backstreet/.steam/steam/ubuntu12_32/steam' '-launcher=0' '--enable-media-stream' '--enable-smooth-scrolling' '--password-store=basic' '--log-file=/usr/home/backstreet/.steam/steam/logs/cef_log.txt' '--disable-quick-menu' '--disable-features=SameSiteByDefaultCookies' '--enable-blink-features=ResizeObserver,Worklet,AudioWorklet' '--disable-blink-features=Badging' --no-sandbox --in-process-gpu]] steamwebhelper.sh[17657]: Runtime for steamwebhelper: defaulting to /usr/home/backstreet/.steam/steam/ubuntu12_64/steam-runtime-heavy steamwebhelper.sh[17657]: CEF sandbox already disabled

steam-error-messages.txt

[LukeHack]

try to compile the port version

[PaddyMac]

try to compile the port version

I already tried the one in ports. It failed to start. I emailed the maintainer of the port about the issue. He replied saying that it was a known issue and that a fix had already been applied in the repo but it would be a couple of weeks until the port was updated. So that's why I decided to try the one straight from the repo. The other issue no longer is a problem. But now I have this issue.

[shkhln]

Perhaps you want to try a few different commits (like ca8654e6e0e5f610d78d0058bb78f435e234ce37 and 57fd4c1808d32b5a4a2029199ba8f169ec6dedc8).

[PaddyMac]

Perhaps you want to try a few different commits (like ca8654e and 57fd4c1).

I tried compiling those commits, but the error was the same. I even tried a few other commits -- including older ones. But newer commits had the same problem, and older commits tended to outright fail with some other problem.

[PaddyMac]

It appears this problem may only occur on 14-CURRENT because I tried the most recent git code, and it worked fine on another system running 13.2-RELEASE.

[glebius]

Observing same problem on 14-CURRENT.

[LukeHack]

i use to run only stable branches, so 13.2 is the right choice.. maybe it is totally untested on 14-current

[shkhln]

maybe it is totally untested on 14-current

I don't specifically test CURRENT, but it's not really unsupported per se — the expectation is that the Linux emulation code never breaks backward compatibility, so any issues with CURRENT are likely FreeBSD bugs. It's still useful to keep track of them there until they are properly analyzed.

[LukeHack]

maybe it is totally untested on 14-current

I don't specifically test CURRENT, but it's not really unsupported per se — the expectation is that the Linux emulation code never breaks backward compatibility, so any issues with CURRENT are likely FreeBSD bugs. It's still useful to keep track of them there until they are properly analyzed.

yes i agree.. and was thinking it could be useful to debug on CURRENT and break the loop with a

IF #ENDIF block

EDIT: i did read better the bug reported and it has nothing to do with loops.. so it's like you said a bug

[PaddyMac]

While I don't know enough to contribute code, I'm happy to test and try to help troubleshoot this. I also participate on the CURRENT mailing list. What can I do to help pin this down? Unfortunately I can't just use 13.2-RELEASE because there's a bug in drm-510-kmod which prevents me from using my dual-monitor setup. I have to run 14-CURRENT in order to use drm-515-kmod which doesn't have this bug and therefore allows me to use two monitors. So if I can help test and debug, I will. And yeah, my describing this issue as a loop may have been misleading. It just seemed appropriate at the time to describe an error message that repeats infinitely until the program is killed.

[glebius]

We have this patch against FreeBSD 14.0 kernel to fix the problem. This could be not the final version of the patch. cache_fplookup.diff.txt

[glebius]

Note: we also have in the works patch to fix kernel panics with CS:GO https://reviews.freebsd.org/D40481

[agrajag9]

Looks like a fix is in the works:

https://reviews.freebsd.org/D41806

I can confirm this patch fixes the issue for me.