intel atom D525MW w/i915 onboard gfx vs 16x9 monitor

[Gene Heskett]

On Saturday 18 July 2015 02:41:44 Felix Miata wrote:
> Gene Heskett composed on 2015-07-18 01:51 (UTC-0400):
> > Felix Miata wrote:
> >> What, if any, video parameters do you have on your kernel cmdline?
>
> ???
>
> ...
>
> > This isn't TCC, but the XFCE used for a linuxcnc install on those
> > machines.  Its debian wheezy, 7.8, with a real time patched kernel.
> >
> > TCC has no such choices here on this machine, but since it works
> > reasonably well on this dell 16x9 23"er, I haven't looked that hard
> > either.
> >
> > I have looked at xrandr, but see no controls that would effect that.
>
> I don't know about controls. I just put
>
> 	http://fm.no-ip.com/Share/setup
>
> where it will get run when X starts, with whatever mode I want used
> uncommented. It works as expected on Jessie with TDE.
>
> Or, use
>
> 	http://fm.no-ip.com/Share/Linux/xorg.conf-minimal-EDID-workaround
>
> or one of the others in that directory as a skeleton to create an
> xorg.conf file, uncommenting or creating the appropriate PreferredMode
> line. I don't recall trying this on Jessie, but it works in
> Kubuntu/TDE,
> Linuxmint/Cinnamon, Mageia/KDE4 (with kscreen disabled), Fedora/KDE4&5
> (with kscreen disabled) and openSUSE/KDE4&5 (with kscreen disabled).
>
> > The default seems to be 1280x768.
>
> I don't recall running across that one before. 16:9 @768 is 136X.
>
> > There is a 13nnx800 mode but the
> > images are bigger than the screen and circles are still smunched
> > vertically. 17" AOC monitor on the machine doing the squashing.
>
> A quick search for 17" AOC 16:9 suggests it should most likely be
> preferred/native 1600x900. What modes are listed by xrandr and/or
> Xorg.0.log?

Xorg.0.log is in my other reply, Felix.  I'm just up, need a quart of 
coffee I haven't made yet to function correctly. :-\

In any event, the machines in question do not have either 
an /etc/X11/xorg.conf, nor an /etc/X11/xorg.conf.d

Here is what is there:
gene at shop:/var/log$ cd /etc/X11
gene at shop:/etc/X11$ ls -l
total 72
drwxr-xr-x 2 root root  4096 Jul 16 08:35 app-defaults
-rw-r--r-- 1 root root    18 Nov  5  2014 default-display-manager
drwxr-xr-x 6 root root  4096 May 18 12:11 fonts
-rw-r--r-- 1 root root 17394 Sep 29  2009 rgb.txt
lrwxrwxrwx 1 root root    13 May 18 12:11 X -> /usr/bin/Xorg
drwxr-xr-x 2 root root  4096 May 18 12:11 xinit
drwxr-xr-x 2 root root  4096 Dec 25  2012 xkb
-rwxr-xr-x 1 root root   709 Oct 13  2010 Xreset
drwxr-xr-x 2 root root  4096 May 18 12:11 Xreset.d
drwxr-xr-x 2 root root  4096 May 18 12:11 Xresources
-rwxr-xr-x 1 root root  3517 Apr  8  2009 Xsession
drwxr-xr-x 2 root root  4096 May 18 13:52 Xsession.d
-rw-r--r-- 1 root root   265 Jan 15  2009 Xsession.options
-rw-r--r-- 1 root root    13 May 24  2013 XvMCConfig
-rw-r--r-- 1 root root   601 Nov  5  2014 Xwrapper.config

Sort of off-topic, but this is the real world out here.

These installs are mostly debian wheezy (7.8) based and use the wheezy 
repo's for updates, but the install iso came from linuxcnc.org, is 
latest version.  The kernel is a special RTAI patched version & pinned 
so the update-manager doesn't muck with it. Linuxcnc will not run except 
in simulation mode w/o that kernel. Currently 3.4-9-rtai-686-pae, but 
the pae does not work.  The RTAI patches are VERY invasive.

I have the same situation with a machine I am attempting to make run a 
newer, bigger mill I am just now brining to life, but it doesn't have 
enough memory at 384 megs, and is a bit short of cpu as its a 1Ghz 
Athlon.  An HP Pavilion darned near old enough to vote, with an ATI 
video card running a 26" Samsung tv with a failed tuner as a monitor. 
TBR when a PSU for a Tyan S2721 gets here.  But that one I'll leave 
running in cold weather as supplemental heat in my well insulated 
garage.  The S2721 is a dual cpu monster server board.

Yet TBD is if isolcpus=1 will work, which leaves linuxcnc the total run 
of one core, improving its IRQ performance quite a bit if doing software 
stepping on stepper motors.  But I've a hardware card for that, so the 1 
millisecond "servo" thread is the main thread on machines with the 
hardware card in them.

NVidia drivers, FWTW, can even screw that up with their hundreds of 
milliseconds IRQ disabling.  So they get kicked into the next drainage, 
the curb isn't nearly far enough.

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>