Bug 1325391

Hi, 
Where is 'nvidia-modprobe' stored? 
I would say nvidia-modprobe should have xserver_t domain instead of xdm_t.

(In reply to Lukas Vrabec from comment #1)
> Hi, 
> Where is 'nvidia-modprobe' stored? 
> I would say nvidia-modprobe should have xserver_t domain instead of xdm_t.

Thanks for looking at this, Lukas. I recently hit a similar SELinux denial on my F22 machine with xorg-x11-drv-nvidia-340xx-340.96-1.fc22.x86_64 (and friends) installed from rpmfusion.

nvidia-modprobe itself actually lives in the xorg-x11-drv-nvidia-340xx-cuda-340.96-1.fc22.x86_64 RPM, not the main driver RPM, and it's not entirely clear to me whether it's even part of the normal GPU driver operation. It may be something CUDA-specific that isn't really necessary. (I don't do anything with CUDA.) But the CUDA RPM is a dependency for xorg-x11-drv-nvidia-340xx-devel, and that one I'd rather not remove.

The exact denial in my case was a write access request on /dev. My /dev/nvidiactl has the context system_u:object_r:xserver_misc_device_t:s0 which differs from the original report and may explain the difference. Regardless, the details on nvidia-modprobe match.

I'll do my best to answer your questions, at least from my perspective (which hopefully is close enough to Dmitry's to be applicable.)

1. nvidia-modprobe (at least in the rpmfusion install) lives in /usr/bin/, it's setuid root and has the context system_u:object_r:bin_t:s0.

2. The source context from the report may be explained by one additional piece of information from my report, which doesn't appear in the original. Note this syscall log:

type=SYSCALL msg=audit(1461818353.699:745): arch=x86_64 syscall=mknod success=no exit=EACCES a0=7fffd95d1f00 a1=21b6 a2=c3ff a3=1999999999999999 items=0 ppid=1829 pid=1841 auid=4294967295 uid=42 gid=42 euid=0 suid=0 fsuid=0 egid=42 sgid=42 fsgid=42 tty=tty1 ses=4294967295 comm=nvidia-modprobe exe=/usr/bin/nvidia-modprobe subj=system_u:system_r:xdm_t:s0-s0:c0.c1023 key=(null)

nvidia-modprobe is being executed by the gdm user (42), not by the logged-in user. AFAICT all of gdm's processes run under the system_u:system_r:xdm_t:s0-s0:c0.c1023 context, which I assume would then be propagated to /usr/bin/nvidia-modprobe when it's executed?

As for what nvidia-modprobe does, the one-line summary from its man page is, "Load the NVIDIA kernel module and create NVIDIA character device files." And the Description section reads (in part):

========
The  nvidia-modprobe utility is used by user-space NVIDIA driver components to make sure the NVIDIA kernel module is loaded and that the NVIDIA character device files are present.  These  facilities  are  normally provided  by  Linux  distribution configuration systems such as udev.  When possible, it is recommended to use your Linux distribution's native mechanisms for managing kernel module loading and  device  file  creation.  This utility is provided as a fallback to work out-of-the-box in a distribution-independent way.

When installed by nvidia-installer , nvidia-modprobe is installed setuid root.

Users  should  not  normally  need to run nvidia-modprobe by hand: the NVIDIA user-space driver components will fork(2)/exec(3) it when needed.
========

In my case, the SELinux alerts came AFTER a system error that involved a lengthy gnome-shell and/or Xorg freeze, followed by an eventual crash. The denials came AFTER this occurred; nvidia-modprobe definitely didn't cause it (nor did the denials), and in fact they appear to have been part of "the user-space driver components" attempting some sort of automated recovery. An unsuccessful one, since as I recall I ended up just rebooting the machine.

Problem still exists:

Fedora 24
Nvidia 960
SELinux Policy version 3.13.1-191.19.fc24

output of SEtroubleshoot:

SELinux is preventing nvidia-modprobe from create access on the chr_file nvidiactl.

*****  Plugin device (91.4 confidence) suggests   ****************************

If you want to allow nvidia-modprobe to have create access on the nvidiactl chr_file
Then you need to change the label on nvidiactl to a type of a similar device.
Do
# semanage fcontext -a -t SIMILAR_TYPE 'nvidiactl'
# restorecon -v 'nvidiactl'

*****  Plugin catchall (9.59 confidence) suggests   **************************

If you believe that nvidia-modprobe should be allowed create access on the nvidiactl chr_file by default.
Then you should report this as a bug.
You can generate a local policy module to allow this access.
Do
allow this access for now by executing:
# ausearch -c 'nvidia-modprobe' --raw | audit2allow -M my-nvidiamodprobe
# semodule -X 300 -i my-nvidiamodprobe.pp

Additional Information:
Source Context                system_u:system_r:xdm_t:s0-s0:c0.c1023
Target Context                system_u:object_r:device_t:s0
Target Objects                nvidiactl [ chr_file ]
Source                        nvidia-modprobe
Source Path                   nvidia-modprobe
Port                          <Unknown>
Host                          kerberos.fixitbear.com
Source RPM Packages           
Target RPM Packages           
Policy RPM                    selinux-policy-3.13.1-191.19.fc24.noarch
Selinux Enabled               True
Policy Type                   targeted
Enforcing Mode                Enforcing
Host Name                     kerberos.fixitbear.com
Platform                      Linux kerberos.fixitbear.com 4.8.4-200.fc24.x86_64
                              #1 SMP Tue Oct 25 13:06:04 UTC 2016 x86_64 x86_64
Alert Count                   252
First Seen                    2016-10-07 16:28:06 PDT
Last Seen                     2016-11-06 06:49:09 PST
Local ID                      93edcacc-ea99-43b9-b9f6-4fe38cb0ed44

Raw Audit Messages
type=AVC msg=audit(1478443749.356:208): avc:  denied  { create } for  pid=1298 comm="nvidia-modprobe" name="nvidiactl" scontext=system_u:system_r:xdm_t:s0-s0:c0.c1023 tcontext=system_u:object_r:device_t:s0 tclass=chr_file permissive=0


Hash: nvidia-modprobe,xdm_t,device_t,chr_file,create

(In reply to Mathew Alexander from comment #3)
> Problem still exists:
> 
> Fedora 24
> Nvidia 960
> SELinux Policy version 3.13.1-191.19.fc24
> 
> output of SEtroubleshoot:
> 
> SELinux is preventing nvidia-modprobe from create access on the chr_file
> nvidiactl.
> 
> *****  Plugin device (91.4 confidence) suggests  
> ****************************
> 
> If you want to allow nvidia-modprobe to have create access on the nvidiactl
> chr_file
> Then you need to change the label on nvidiactl to a type of a similar device.
> Do
> # semanage fcontext -a -t SIMILAR_TYPE 'nvidiactl'
> # restorecon -v 'nvidiactl'
> 
> *****  Plugin catchall (9.59 confidence) suggests  
> **************************
> 
> If you believe that nvidia-modprobe should be allowed create access on the
> nvidiactl chr_file by default.
> Then you should report this as a bug.
> You can generate a local policy module to allow this access.
> Do
> allow this access for now by executing:
> # ausearch -c 'nvidia-modprobe' --raw | audit2allow -M my-nvidiamodprobe
> # semodule -X 300 -i my-nvidiamodprobe.pp
> 
> Additional Information:
> Source Context                system_u:system_r:xdm_t:s0-s0:c0.c1023
> Target Context                system_u:object_r:device_t:s0
> Target Objects                nvidiactl [ chr_file ]
> Source                        nvidia-modprobe
> Source Path                   nvidia-modprobe
> Port                          <Unknown>
> Host                          kerberos.fixitbear.com
> Source RPM Packages           
> Target RPM Packages           
> Policy RPM                    selinux-policy-3.13.1-191.19.fc24.noarch
> Selinux Enabled               True
> Policy Type                   targeted
> Enforcing Mode                Enforcing
> Host Name                     kerberos.fixitbear.com
> Platform                      Linux kerberos.fixitbear.com
> 4.8.4-200.fc24.x86_64
>                               #1 SMP Tue Oct 25 13:06:04 UTC 2016 x86_64
> x86_64
> Alert Count                   252
> First Seen                    2016-10-07 16:28:06 PDT
> Last Seen                     2016-11-06 06:49:09 PST
> Local ID                      93edcacc-ea99-43b9-b9f6-4fe38cb0ed44
> 
> Raw Audit Messages
> type=AVC msg=audit(1478443749.356:208): avc:  denied  { create } for 
> pid=1298 comm="nvidia-modprobe" name="nvidiactl"
> scontext=system_u:system_r:xdm_t:s0-s0:c0.c1023
> tcontext=system_u:object_r:device_t:s0 tclass=chr_file permissive=0
> 
> 
> Hash: nvidia-modprobe,xdm_t,device_t,chr_file,create

Actually, may be an issue with the way the Nvidia drivers were installed, used the .bin file from Nvidia, not the repository.

This message is a reminder that Fedora 23 is nearing its end of life.
Approximately 4 (four) weeks from now Fedora will stop maintaining
and issuing updates for Fedora 23. It is Fedora's policy to close all
bug reports from releases that are no longer maintained. At that time
this bug will be closed as EOL if it remains open with a Fedora  'version'
of '23'.

Package Maintainer: If you wish for this bug to remain open because you
plan to fix it in a currently maintained version, simply change the 'version' 
to a later Fedora version.

Thank you for reporting this issue and we are sorry that we were not 
able to fix it before Fedora 23 is end of life. If you would still like 
to see this bug fixed and are able to reproduce it against a later version 
of Fedora, you are encouraged  change the 'version' to a later Fedora 
version prior this bug is closed as described in the policy above.

Although we aim to fix as many bugs as possible during every release's 
lifetime, sometimes those efforts are overtaken by events. Often a 
more recent Fedora release includes newer upstream software that fixes 
bugs or makes them obsolete.

Fedora 23 changed to end-of-life (EOL) status on 2016-12-20. Fedora 23 is
no longer maintained, which means that it will not receive any further
security or bug fix updates. As a result we are closing this bug.

If you can reproduce this bug against a currently maintained version of
Fedora please feel free to reopen this bug against that version. If you
are unable to reopen this bug, please file a new report against the
current release. If you experience problems, please add a comment to this
bug.

Thank you for reporting this bug and we are sorry it could not be fixed.