Edit:
solved with an xorg.conf containing framebuffer (and empty line) as suggested by brummer. Thanks a lot!
Hi,
as described here
http://forum.siduction.org/index.php?topic=2149&highlight=iceweasel (http://forum.siduction.org/index.php?topic=2149&highlight=iceweasel)
(german) in April, i have the very same problem now, again:
some lines or words in web-pages are "pixeled" or shown as a block, unreadable. But it goes away, if i hover with the mouse, scroll or move the iceweasel-window over the edges of the screen.
I did already switch off hardware accelleration in iceweasel (and have the impression, it is less frequent since, now) and checked for libcairo2 (already newest):
# dpkg -s libcairo2
Package: libcairo2
Architecture: amd64
Version: 1.12.2-2
Anybody else has seen this? Or maybe a solution?
Thanks!
What version of Iceweasel is that?
greetz
devil
Hi Devil,
thanks for taking care!
It is 10.0.7, apt claims, it's the newest.
Greets
Edit:
here is a link how it does look like
http://debianforum.de/forum/viewtopic.php?f=2&t=136313&hilit=Schrift+Fehler
I use (as before, nothing changed here) nouveau for a nvidia GF108 [Quadro 1000M]
I faintly remember that. I run 17.0a2 (2012-09-21).
That and some inbetween are available at http://mozilla.debian.net. Pick your poison.
greetz
devil
Here no problems with:
pol iceweasel
iceweasel:
Installiert: 10.0.7esr-2
Installationskandidat: 10.0.7esr-2
Versionstabelle:
Hi mylo,
thanks for having a look and sharing, but i had the very same version.
Hi devil,
thanks for the hint! So i tried the next, experimental:
# apt-cache policy iceweasel
iceweasel:
Installiert: 15.0.1-1
Installationskandidat: 15.0.1-1
But: same (bad) behaviour :-(
Additionally i realized, that the pixelized rendering also occurs in iceweasel's page header tabs, not only inside the web pages.
you can try this:
http://forum.siduction.org/index.php?topic=2623
Hi brummer,
thanks, but it seems, shadowfb is already active (and glxgears fast).
[ 24.553] (II) Loading sub module "shadowfb"
[ 24.553] (II) LoadModule: "shadowfb"
[ 24.553] (II) Loading /usr/lib/xorg/modules/libshadowfb.so
[ 24.554] (II) Module shadowfb: vendor="X.Org Foundation"
[ 24.554] compiled for 1.12.4, module version = 1.0.0
[ 24.554] ABI class: X.Org ANSI C Emulation, version 0.4
and no unload...
Well, that it is loaded doesn't mean that the shadow framebuffer is "on", look for this lines in your Xorg.0.log:
Quote[ 19.710] (**) NOUVEAU(0): Option "ShadowFB" "true"
[ 19.710] (==) NOUVEAU(0): Using HW cursor
[ 19.711] (**) NOUVEAU(0): Using "Shadow Framebuffer" - acceleration disabled
[
@wannek:
apt-cache policy iceweasel
iceweasel:
Installiert: 17.0~a2+20120928042009-1
Installationskandidat: 17.0~a2+20120928042009-1
Versionstabelle:
*** 17.0~a2+20120928042009-1 0
500 http://mozilla.debian.net/ experimental/iceweasel-aurora amd64 Packages
100 /var/lib/dpkg/status
v15 to old ... ;)
@brummer
no, i don't have any of your three lines...
(I don't even use a /etc/X11/xorg.conf)
@reddark
experimental still has 15.0.1-1. And since nothing changed sid->experimental, i don't believe in iceweasel's fault.
Well, the shadow framebuffer is off by default, so it's no wonder.
here is my etc/X11/xorg.conf.d/20-nouveau.conf
Section "Device"
Identifier "Device 0"
Driver "nouveau"
Option "ShadowFB" "true"
#Option "GLXVBlank" "true"
EndSection
you can try it if you like,
I have the same problem that wannek3 has, for all practical intents and purposes. I usually use Chromium, so I haven't spent a lot of time looking for a solution. All the work wannek3 and others have put into this spurred me to look at it.
1 - I have the same software versions wannek3 has. I went to the version in experimental when it was suggested above with no change. I should add that I even see the problem in the address bar, it isn't limited to tabs or web pages.
2 - it has not always been this way. Since I do not use it regularly, I can't be sure of the exact time it happened.
3 - if I boot with a 64bit Desperado Reloaded iso, I see the identical problem.
4 - I grabbed an iso of Mint KDE 64bit and tried it - there is no problem in Mint, and there is no /etc/X11/xorg.conf.d/20-nouveau.conf by default.
@brummer
thanks, but unfortunately your suggestion did not work. If Your lines are the only ones in my xorg.conf then x does not start at all:
no screens found
And Xorg -configure dies with something like "number of created screens do not match".
By the way: as ghstryder told, i can see this on a brand new installation of Desperado Reloaded, too.
just in case:
Quote...
EndSection
# Leerzeile muss sein
and what's the name of your xorg.conf file (full path)
@michaa7
Thanks a lot, i forgot that! Ok, i added an empty line to /etc/X11/xorg.conf (as mentioned above), rebooted and voila.
X comes up with
(**) NOUVEAU(0): Option "ShadowFB" "true"
(**) NOUVEAU(0): Using "Shadow Framebuffer" - acceleration disabled
And since 5 Minutes now, no rendering errors!
glxgears is still quite fast.
-> solved (at least for me)
Is there a reason why you have a xorg.conf, when using nouveau?
greetz
devil
@ devil
because it's the solution (see brummers postings) and presumably a Debian bug (as this behavior does not show up in other distros with same versions of IW and Xorg).
Anyone cares to file a bug report?
Yes, michaa7, i had neither a xorg.conf nor a xorg.conf.d before.
Bugreport is #689817
It isn't a bug in iceweasel, actual it is a bug in xorg-server, first reported on 31 Mar 2012:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=666468
The bug is triggered by libcairo, and first come visible on ATI Cards.
Unfortunately the fix for ATI move the bug to the nouveau driver.
Other Distros (manly ubuntu) aren't bitten by this bug, because they ship a patched version of libcairo, were they switch of the related functions in cairo. I didn't believe that this is the right workaround, because cairo didn't do it wrong, it is xorg related, so the workaround to use a nouveau.conf and switch off acceleration in the "Xorg-driver" is closer to the real Bug. Well it isn't a solution to use a nouveau.conf, but it is somewhat clearer, then to patch a library for disable a function, which do it right, but trigger a bug in a other "library".