[Bug 331609] New: Bad DPI with Intel driver is uncorrectable via xorg.conf
https://bugzilla.novell.com/show_bug.cgi?id=331609
Summary: Bad DPI with Intel driver is uncorrectable via xorg.conf
Product: openSUSE 10.3
Version: Final
Platform: PC
OS/Version: openSUSE 10.3
Status: NEW
Severity: Major
Priority: P5 - None
Component: X.Org
AssignedTo: sndirsch@novell.com
ReportedBy: mrmazda@ij.net
QAContact: xorg-maintainer-bugs@forge.provo.novell.com
Found By: ---
http://qa.mandriva.com/show_bug.cgi?id=33935 is the same basic problem.
When Display DDC and/or EDID are missing or broken, there's no way I can find
to get KDM/KDE to start with a correct DPI except by the accident of physical
display DPI and screen resolution computing to 96. Once X is started, xrandr
can be used to set a good DPI, but this is of no use in KDM, Kicker and other
autostarted apps. Appending '-dpi
https://bugzilla.novell.com/show_bug.cgi?id=331609#c1
--- Comment #1 from Felix Miata
https://bugzilla.novell.com/show_bug.cgi?id=331609#c2
--- Comment #2 from Felix Miata
https://bugzilla.novell.com/show_bug.cgi?id=331609#c3
--- Comment #3 from Felix Miata
https://bugzilla.novell.com/show_bug.cgi?id=331609#c4
--- Comment #4 from Felix Miata
https://bugzilla.novell.com/show_bug.cgi?id=331609#c5
--- Comment #5 from Felix Miata
https://bugzilla.novell.com/show_bug.cgi?id=331609#c6
--- Comment #6 from Felix Miata
https://bugzilla.novell.com/show_bug.cgi?id=331609#c7
Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c8
--- Comment #8 from Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609
Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c9
--- Comment #9 from Felix Miata
These new RANDR 1.2 drivers are *very* special. Try xrandr --dpi <dpi>/<output> on the running Xserver.
Is there something I'm not understanding about the "/<output>" part of what you
wrote? Is that different from "...xrandr can be used to set a good DPI..." from
comment 0? It was late and I was tired and frustrated by failed workaround
attempts and rereading bug 257746 and the many other other Intel video bugs by
the time I was writing this bug last night. Maybe I failed to clearly spell out
impact of 'xrandr --dpi <dpi>', though indirectly it seems I did via
http://qa.mandriva.com/show_bug.cgi?id=33935#c35. It may show up in attachment
176641, as I ran 'xrandr -dpi 144' just before closing that X session.
As implied in comment 0, I can run 'xrandr -dpi <dpi>' after X starts, e.g. in
Konsole, but it doesn't survive closing the session and so post-session start
it is apparently useless for Kicker and autostarted apps from saved sessions.
How can it be run before KDM/KDE starts so that they are affected by the result
of it too?
To rephrase this bug's summary as if it were an enhancement request, can X
startup be made to automatically call 'xrandr --fbmm' using DisplaySize
dimensions or 'xrandr -dpi' from something like 'Option "DPI"
"
https://bugzilla.novell.com/show_bug.cgi?id=331609#c10
--- Comment #10 from Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c11
--- Comment #11 from Felix Miata
My understanding was that you're talking about the Xserver -dpi option and not the xrandr --dpi option.
That commentary was just about what didn't work to try to solve this Intel-only DPI problem.
Resolution/DPI chosing in RANDR 1.2 is just so broken. :-(
The i810 driver still obeys DisplaySize to produce correct DPI. Why does sax2 prefer the broken intel/randr driver set over i810 for the final 10.3 release? With so many bugs they seem too far from ready to live as a default in a public release.
So does using xrandr for setting DPI help?
As stated in comment 9, only partially: 1-Does not work for Kicker and other autostarted apps 2-Works for apps started after running it
I suggest to open an enhancement request for krandrtray to set also the DPI.
That shouldn't be necessary except as workaround for current brokenness. i810 seems a better user workaround until intel/randr work right. Proper DPI should be at least semi-automatic, not something users need to routinely choose off some toolbar that doesn't help for apps already running anyway. xorg.conf is supposed to be the mother config file that controls how X starts. If xrandr is the only way to do something that needs doing on X startup on Intel, then something in xorg.conf should at least trigger it even if some other config file must be used for xrandr to know what to do. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=331609#c12
Stefan Dirsch
Why does sax2 prefer the broken intel/randr driver set over i810 for the >final 10.3 release? With so many bugs they seem too far from ready to live as >a default in a public release. You would have expected the old VESA BIOS based i810 driver to support resolutions like 2048x1536? i810 was very limited in its functionality and suffered from different bugs. I'm not satisfied by the quality of the intel driver either. Even Intel is not. If we would have chosen i810 by default others would have complained, why we don't use intel as default ...
I think the wrong DPI setting is just a result of the wrong mode being selected by default. It should be set correctly when you use Option "PreferredMode" "2048x1536" in your monitor section. See Bug #257746. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=331609#c13
Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c14
Felix Miata
https://bugzilla.novell.com/show_bug.cgi?id=331609#c15
--- Comment #15 from Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c16
--- Comment #16 from Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c17
--- Comment #17 from Felix Miata
I suggest to add such a xrandr call to /etc/X11/xdm/Xsetup, right after the other xrandr calls.
This had no impact: --- Xsetup.01 2007-03-15 11:41:31.000000000 -0400 +++ Xsetup 2007-10-07 17:34:11.000000000 -0400 @@ -93,6 +93,7 @@ x*) echo "xrandr for $ext: known mode $line" xrandr --output $ext --mode $mode +xrandr --fbmm 281x211 ;; esac done (In reply to comment #16 from Stefan Dirsch)
So what's the response to the question in comment #13?
Paragraph 2 of comment 14 (no impact). -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=331609#c18
--- Comment #18 from Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c19
--- Comment #19 from Felix Miata
https://bugzilla.novell.com/show_bug.cgi?id=331609#c20
--- Comment #20 from Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c21
--- Comment #21 from Felix Miata
Are you sure you're using a displaymanager? It won't work with startx/xinit.
I was using startx as a short cut for these frequent login/logout tests. Comment 18 does not work for KDM. The first time I entered my password to login to KDE, the screen went black and the keyboard totally ceased to function. After power off restart I tried KDM again. 'xrandr --fbmm <widthXheight>' in Xsetup according to comment 18 works for KDE Desktop. 'xrandr --dpi <bar>' in Xsetup has equivalent desirable effect. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=331609#c22
Stefan Dirsch
Comment 18 does not work for KDM. The first time I entered my password to login to KDE, the screen went black and the keyboard totally ceased to function. Strange.
After power off restart I tried KDM again. 'xrandr --fbmm <widthXheight>' in > Xsetup according to comment 18 works for KDE Desktop. 'xrandr --dpi <bar>' in Xsetup has equivalent desirable effect. ok. So you have a workaround now.
-- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=331609#c23
--- Comment #23 from Felix Miata
https://bugzilla.novell.com/show_bug.cgi?id=331609#c24
--- Comment #24 from Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c25
Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c26
--- Comment #26 from Felix Miata
So you are comparing the behaviour of i815 and i845? AFAIK for 810/815 the intel driver still uses the VESA BIOS for modesetting so it just behaves like the old i810 driver. This is different for 830 and higher.
You're probably right. I trashed my Knoppix on 845G in order to install Gutsy below hda16. NAICT, its behavior is identically broken as 10.3. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=331609#c27
Matthias Hopf
https://bugzilla.novell.com/show_bug.cgi?id=331609#c28
--- Comment #28 from Felix Miata
https://bugzilla.novell.com/show_bug.cgi?id=331609#c29
--- Comment #29 from Matthias Hopf
https://bugzilla.novell.com/show_bug.cgi?id=331609#c30
--- Comment #30 from Matthias Hopf
https://bugzilla.novell.com/show_bug.cgi?id=331609
Matthias Hopf
https://bugzilla.novell.com/show_bug.cgi?id=331609
Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c31
Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c32
Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c33
Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c34
--- Comment #34 from Felix Miata
https://bugzilla.novell.com/show_bug.cgi?id=331609#c35
Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c36
Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c37
--- Comment #37 from Anja Stock
https://bugzilla.novell.com/show_bug.cgi?id=331609#c38
Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c39
Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c40
--- Comment #40 from Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c41
Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c43
Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c45
Ruediger Oertel
https://bugzilla.novell.com/show_bug.cgi?id=331609#c46
Stefan Dirsch
https://bugzilla.novell.com/show_bug.cgi?id=331609#c47
Anja Stock
https://bugzilla.novell.com/show_bug.cgi?id=331609#c48
Stefan Dirsch
Matthias, can you push this upstream? - if (output->conf_monitor->mon_width > 0 || - output->conf_monitor->mon_height > 0) + if (output->conf_monitor && + (output->conf_monitor->mon_width > 0 || + output->conf_monitor->mon_height > 0))
The code was not in get yet, so I updated the patch in the upstream bugreport. https://bugs.freedesktop.org/show_bug.cgi?id=10304 Finally closing as fixed. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=331609#c49
--- Comment #49 from Matthias Hopf
+ if (output->conf_monitor && + (output->conf_monitor->mon_width > 0 && + output->conf_monitor->mon_height > 0))
A monitor section with either being 0 or less should not be honored. No actual effect on our product, so no update needed. Fixed upstream. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply@novell.com