Date: Tue, 07 Apr 2026 12:16:44 +0000 From: bugzilla-noreply@freebsd.org To: ports-bugs@FreeBSD.org Subject: [Bug 291239] Kernel panic when running most wayland WMs (nvidia) Message-ID: <bug-291239-7788-lch34hUyMh@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-291239-7788@https.bugs.freebsd.org/bugzilla/>
index | next in thread | previous in thread | raw e-mail
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=291239 --- Comment #7 from Tomoaki AOKI <junchoon@dec.sakura.ne.jp> --- (In reply to Adriaan de Groot from comment #5) Try setting hw.nvidiadrm.modeset=1 in your /boot/loader.conf. This is NOT a variable for rc scripts, so setting it in your /etc/rc.conf shouldn't work, thus, nvidia-drm.ko is NOT actually working even though it's loaded. Cannot recommend to set it in /etc/sysctl.conf, too. This is because hw.nvidiadrm.modeset is a tunable that nvidia-drm.ko fetches on load, means, need to be set BEFORE nvidia-drm.ko is loaded. If you're lucky enough, setting it via /etc/sysctl.conf is done before nvidia-drm.ko is loaded via rc script WRT kld_list variable in your /etc/rc.conf, but would be in race condition. Tunalbe hw.nvidia.registry.EnableGpuFirmware basically matters only for Turing and later generation of architectures which has GSP in it. But setting it to 1 in your /boot/loader.conf for GPUs that are pre-Turing generation of architectures could possibly cause crash (no fallback is allowed and forcibly attempts to load even for pre-Turing GPUs). 17 (0x11) would be safe (0x10=allow fallback, 0x01=forcibly enable). -- You are receiving this mail because: You are the assignee for the bug.home | help
Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-291239-7788-lch34hUyMh>