Date: Tue, 7 Jun 2022 23:15:36 -0400 From: Theron <theron.tarigo@gmail.com> To: x11-list freebsd <freebsd-x11@freebsd.org> Subject: Re: Nvidia Optimus Crash When Starting Xorg Message-ID: <4b01d2a9-61b9-bd6b-143b-a19870e6127e@gmail.com> In-Reply-To: <20220607144249.beknikii3x3j34ol@topanga> References: <20220604231525.ybdydr4edglnnsb4@topanga.nomadlogic.org> <20220604234552.lomrfz4sngtkxxpu@colony.nomadlogic.org> <CALH631nDDH9eunqgeOiwZeMiN12X__OrPyMmkupQNaF0xg3Psw@mail.gmail.com> <20220607144249.beknikii3x3j34ol@topanga>
next in thread | previous in thread | raw e-mail | index | archive | help
On 6/7/22 10:42, Pete Wright wrote: > On Tue, Jun 07, 2022 at 09:04:33AM +0300, Gleb Popov wrote: >> Hi Pete, I also have this problem. Did you manage to debug this further? > hey there, Theron and I had a chat privately Oops, I had meant to respond on the list: > I see roughly the same: > > nvidia-secondary-driver-510.60.02_1 > > [ 222.989] (EE) Backtrace: > [ 222.990] (EE) 0: /usr/local/bin/Xorg (?+0x0) [0x41853a] > [ 222.991] (EE) unw_get_proc_name failed: no unwind info found [-10] > [ 222.991] (EE) 1: /lib/libthr.so.3 (?+0x0) [0x8273185be] > [ 222.992] (EE) unw_get_proc_name failed: no unwind info found [-10] > [ 222.992] (EE) 2: /lib/libthr.so.3 (?+0x0) [0x827317b6f] > [ 222.992] (EE) 3: ? (?+0x0) [0x7ffffffff2d3] > [ 222.993] (EE) 4: ? (?+0x0) [0x0] > [ 222.994] (EE) 5: /usr/local/lib/xorg/modules/drivers/nvidia_drv.so > (?+0x0) [0x83cca750f] > [ 222.995] (EE) 6: ? (?+0x0) [0x0] > [ 222.995] (EE) unw_step failed: unspecified (general) error [-1] > [ 222.995] (EE) > [ 222.995] (EE) Segmentation fault at address 0x0 > [ 222.995] (EE) > Fatal server error: > [ 222.995] (EE) Caught signal 11 (Segmentation fault). Server aborting > [ 222.995] (EE) > [ 222.995] (EE) > > $ uname -a > FreeBSD theron-xps 14.0-CURRENT FreeBSD 14.0-CURRENT #44 main-809922b010 > > The same kernel and Xorg build work with the Nvidia port reverted to > 470.86_1. On 6/7/22 10:42, Pete Wright wrote: > and the workaround seems to be to > use the older Nvidia 470 driver. I think what happened is that support > to my > Optimus card was dropped in the 5xx series, as once I loaded this older > driver dmesg reported the card as detected and Xorg-headless no longer > crashes. In my case (GTX 960M / PCI ID: 139B) the chip is still officially supported in the 510.60.02 readme, yet I see the same crash (up to 0x0 vs. 0x5A5A... difference, but I believe that is merely different alloc debug configuration: the driver and/or stack tracer is reading junk either way). Nvidia driver download website recommends 515.48.07 for my chip - I'll try to update to that and test if the crash is fixed. https://us.download.nvidia.com/XFree86/FreeBSD-x86_64/510.60.02/README/supportedchips.html https://us.download.nvidia.com/XFree86/FreeBSD-x86_64/515.48.07/README/supportedchips.html > i can share pkg's if you'd like, i'm hoping today, time permitting to > submit a bug > report with diff to add a 470 flavor to nvidia-hybrid-graphics along > with a > nvidia-secondary-driver-470 pkg. If I recall correctly, everything's in place since there is already a -390 flavor for the oldest Optimus systems, adding -470 should be trivial.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4b01d2a9-61b9-bd6b-143b-a19870e6127e>