2

I am stuck on virtual terminal 14. Starting some SDL-software brought me there. Now I cannot go anywhere else. Every "chvt" freezes (at the "VT_WAITACTIVE" ioctl) if it is not chvt 14 which succeeds. There is no chvt running. I can start shell and Xorg on vt14 (from remote ssh).

How to be able to switch vt again (other than rebooting)? How to kill any maybe pending and thus blocking vt switch? looking at the output of ps -A did not reveal any suspicious process (except maybe Xorg, killing it is not an option). Any getty I already killed (and they got restarted by init).

Graphics: uvesafb, Xorg uses fbdev, no kernel mode setting, sysvinit/openrc (no systemd).

Golar Ramblar
  • 1,831
  • Is it still a problem? Can you find with lsof which processes have /dev/tty14 open? – Incnis Mrsi Sep 14 '15 at 18:36
  • @ Incnis Mrsi: I tried lsof | grep tty14 and fuser /dev/tty14, no result. Now it is not a problem anymore since I just rebooted the system (I was not able to get to my runnung old X session on vt13, x11vnc does not work on the very same machine due to Xorg not updating if it's vt is not active), but it might happen again, although very seldom. – Golar Ramblar Sep 16 '15 at 09:46

1 Answers1

0

I don't know if it applies to your case, but I saw these symptoms when virtual terminal switching was turned off with physlock -l. Running physlock -L unlocked switching, both from the keyboard and with chvt.

Vladimir Panteleev
  • 1,696
  • Interesting, I even did not know about physlock and I don't had/ have it installed. – Golar Ramblar Apr 26 '22 at 17:52
  • I've seen X get stuck pretty hard - almost always due to some game doing something or other that was doubtless unnecessary - where chvt and the like will only work most of the time, but occasionally would get wedged so hard chvt would block (until it interrupted itself) in the VT_ACTIVATE call (and then loop), and IN THAT CASE - even phylock -L doesn't help (at least in my small test sample). – Alex North-Keys Sep 10 '23 at 17:32
  • That sounds like you might be running into GPU driver bugs. You can check the kernel-side stack trace of blocked tasks to confirm. – Vladimir Panteleev Sep 15 '23 at 06:48