1 |
On 2016-12-17 21:35, Ian Zimmerman wrote: |
2 |
|
3 |
> > > [ 2890.175] (++) using VT number 2 |
4 |
> > > [ 2890.175] (WW) xf86OpenConsole: VT_ACTIVATE failed: Operation not permitted |
5 |
> > > [ 2890.175] (EE) xf86OpenConsole: Switching VT failed |
6 |
|
7 |
I made some progress, but now I am blocked in an even tougher place :-( |
8 |
|
9 |
The error above had a simple and amusing cause. My shell code that run |
10 |
startx redirected both stdout and stderr to a logfile. And Xorg |
11 |
inspects those file descriptors to guess what the current vt is, so that |
12 |
when the _requested_ vt is the same, it can avoid the switch, which does |
13 |
require root. So, when I left stdout and stderr going to the terminal |
14 |
this part started working. [1] |
15 |
|
16 |
But now I hit the error chronicled at [2]. A look at kernel source, in |
17 |
drivers/gpu/drm/drm_ioctl.c, shows that the relevant kernel ioctl is |
18 |
declared thus: |
19 |
|
20 |
DRM_IOCTL_DEF(DRM_IOCTL_SET_MASTER, drm_setmaster_ioctl, DRM_ROOT_ONLY), |
21 |
|
22 |
at least in the kernel version I run (4.4.26) which doesn't inspire much |
23 |
hope :-( |
24 |
|
25 |
All this really does seem to depend on the precise GPU model one has. |
26 |
|
27 |
[1] |
28 |
This trick by Xorg is completely undocumented! I really had to wallow |
29 |
in the source code to discover it. |
30 |
|
31 |
[2] |
32 |
https://unix.stackexchange.com/questions/178883/running-x-without-root |
33 |
|
34 |
-- |
35 |
Please *no* private Cc: on mailing lists and newsgroups |
36 |
Personal signed mail: please _encrypt_ and sign |
37 |
Don't clear-text sign: http://cr.yp.to/smtp/8bitmime.html |