Version:            (using KDE 4.4.2)
OS:                Linux
Installed from:    Debian testing/unstable Packages

When I hover the mouse pointer over the last opened file menu entry in kwrite X.org with Radeon KMS reliably crashes here. I tried three times - I get slightly different backtraces, but Radeon DRI /usr/lib/dri/r300_dri.so is always involved. I attach the logs in few moments.

I think this is a DRI bug, so I reported it upstream: 
https://bugs.freedesktop.org/show_bug.cgi?id=27416

Please mark as upstream if you agree. I also reported it as debian bug with upstream reference. I did not yet get back back number, will add a link later.

It is triggered by a certain (long) recent file entry in kwriterc:

Its a certain entry in recent files:

[Recent Files]
File1[$e]=nepomuksearch:/nepomuk_3A_2Fres_2Fde660a90-9ad7-47db-86d8-7aca791dc104?sparql=select%20distinct%20%3Fr%2
······
0%20where%20%7B%20%7B%20%3Fr%20%3Fv1%20%3Fv2%20.%20%3Fv2%20bif%3Acontains%20%22%27Linux%2A%27%22%20.%20%7D%20U
   NION···
%20%7B%20%3Fr%20%3Fv1%20%3Fv3%20.%20%3Fv3%20%3Fv4%20%3Fv2%20.%20%3Fv4%20%3Chttp%3A%2F%2Fwww.w3.org%2F2000%
   2F01%2Fr···
df-schema%23subPropertyOf%3E%20%3Chttp%3A%2F%2Fwww.w3.org%2F2000%2F01%2Frdf-schema%23label%3E%20.%20%3
   Fv2%20bif%3A    contains%20%22%27Linux%2A%27%22%20.%20%7D%20.%20%20%20%20%7D
Name1[$e]=nepomuk_3A_2Fres_2Fde660a90-9ad7-47db-86d8-7aca791dc104

This bug is critical: I took one hour to write a mail, I copied the text to kwrite to have it there for later reference
without the hard line breaks that kmail adds on sending the mail. Then I tried to save it by hovering to the save menu entry. Unfortunately I came across the recent file menu entry, X.org crashed taking the mail with it. KMail's crash crash recovery mechanism did not trigger.

I think choosing the save menu entry to save a text in kwrite is common practice and while I do not know how I got that crazy looking Recent file entry into kwrite, I believe it was some clicking around or trying to open Nepomuk search results
elsewhere.

It may not be that easy to trigger, but this also makes it
more unpredictable as to when serious data loss may happen.

Packages installed on last attempts to reproduce the bug:
kdelibs5/experimental uptodate 4:4.4.2-1
kwrite/experimental uptodate 4:4.4.2-1
libdrm-intel1/squeeze uptodate 2.4.18-3
libdrm-radeon1/squeeze uptodate 2.4.18-3
libdrm-radeon1-dbg/squeeze uptodate 2.4.18-3
libdrm2/squeeze uptodate 2.4.18-3
libdrm2-dbg/squeeze uptodate 2.4.18-3
libgl1-mesa-dri/sid uptodate 7.7.1-1
libgl1-mesa-dri-dbg/sid uptodate 7.7.1-1
libkwineffects1a/experimental-snapshots uptodate 4:4.4.2-0rc1
libqt4-gui/experimental uptodate 4:4.6.2-2
xserver-xorg/squeeze uptodate 1:7.5+5
xserver-xorg-core/sid uptodate 2:1.7.6-2
xserver-xorg-core-dbg/sid uptodate 2:1.7.6-2
xserver-xorg-video-radeon/sid uptodate 1:6.13.0-1
xserver-xorg-video-radeon-dbg/sid uptodate 1:6.13.0-1
xserver-xorg-video-radeonhd/squeeze uptodate 1.3.0-2

I am using Radeon KMS on ThinkPad T42 with

martin@shambhala:~> lspci -nn | grep VGA
01:00.0 VGA compatible controller [0300]: ATI Technologies Inc RV350 [Mobility Radeon 9600 M10] [1002:4e50]