Bug#888952: nvidia-driver and opencl

Andreas Beckmann anbe at debian.org
Mon Feb 5 17:12:19 UTC 2018


I still cannot reproduce it ...

On 2018-02-05 16:11, Hiromasa YOSHIMOTO wrote:
> modprobe: INFO: ../libkmod/libkmod.c:364 kmod_set_log_fn() custom logging function 0x559425f77b90 registered
> modprobe: INFO: ../libkmod/libkmod-module.c:886 kmod_module_insert_module() Failed to insert module '/lib/modules/4.14.8-custom/updates/dkms/nvidia-current.ko': Operation not permitted
> modprobe: ERROR: could not insert 'nvidia_current_uvm': Operation not permitted
> modprobe: INFO: ../libkmod/libkmod.c:331 kmod_unref() context 0x5594264a8400 released

 882         err = init_module(mem, size, args);
 883 init_finished:
 884         if (err < 0) {
 885                 err = -errno;
 886                 INFO(mod->ctx, "Failed to insert module '%s': %m\n", path);
 887         }
 888         return err;

init_module is from glibc ...

and (from the manpage) these are the two errors occuring:

       EPERM  The caller was not privileged (did not have the CAP_SYS_MODULE capability), or module loading is disabled (see /proc/sys/kernel/modules_disabled in proc(5)).

       EEXIST A module with this name is already loaded.

I think you are hitting EPERM as user and EEXIST under sudo.

So perhaps we need to apply some capabilities to nvidia-modprobe, try this:

setcap cap_sys_module+ep /usr/bin/nvidia-modprobe


Andreas



More information about the pkg-nvidia-devel mailing list