Under my ppc64 with mesa-7.8.2, glxgears doesn't run as expected: there are some unexpected polygons being flashed irregularly and clipping the running gears in the window. Scorched3d also does not run properly: only noising polygons shown when you play a game. Compiz has the same problem of flashing polygons. Reproducible: Always Steps to Reproduce: 1. run glxgears in a terminal 2. run scorched3d in a terminal 3. Actual Results: 1. Running gears appeared in a window with flashing and clipping polygons 2. Just various flashing polygons shown when play a game. Expected Results: 1. clean running gears 2. a game emerge --info Portage 2.1.8.3 (default/linux/powerpc/ppc64/10.0/64bit-userland, gcc-4.4.3, glibc-2.11.2-r0, 2.6.34-gentoo-r1 ppc64) ================================================================= System uname: Linux-2.6.34-gentoo-r1-ppc64-PPC970,_altivec_supported-with-gentoo-1.12.13 Timestamp of tree: Thu, 30 Sep 2010 01:00:01 +0000 app-shells/bash: 4.1_p7 dev-java/java-config: 2.1.10 dev-lang/python: 2.6.5-r3 dev-util/cmake: 2.8.1-r2 sys-apps/baselayout: 1.12.13 sys-apps/sandbox: 2.3-r1 sys-devel/autoconf: 2.13, 2.65-r1 sys-devel/automake: 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r2, 1.10.3, 1.11.1 sys-devel/binutils: 2.20.1-r1 sys-devel/gcc: 4.3.4, 4.4.3-r2 sys-devel/gcc-config: 1.4.1 sys-devel/libtool: 2.2.10 sys-devel/make: 3.81-r2 virtual/os-headers: 2.6.33 ACCEPT_KEYWORDS="ppc64" ACCEPT_LICENSE="* -@EULA" CBUILD="powerpc64-unknown-linux-gnu" CFLAGS="-O2 -mcpu=970 -mtune=970 -maltivec -mabi=altivec -mpowerpc64 -fno-strict-aliasing -fomit-frame-pointer -pipe" CHOST="powerpc64-unknown-linux-gnu" CONFIG_PROTECT="/etc /usr/share/X11/xkb" CONFIG_PROTECT_MASK="/etc/ca-certificates.conf /etc/env.d /etc/env.d/java/ /etc/eselect/postgresql /etc/fonts/fonts.conf /etc/gconf /etc/revdep-rebuild /etc/sandbox.d /etc/terminfo" CXXFLAGS="-O2 -mcpu=970 -mtune=970 -maltivec -mabi=altivec -mpowerpc64 -fno-strict-aliasing -fomit-frame-pointer -pipe" DISTDIR="/usr/portage/distfiles" FEATURES="assume-digests distlocks fixpackages news parallel-fetch protect-owned sandbox sfperms strict unmerge-logs unmerge-orphans userfetch" GENTOO_MIRRORS="http://ftp.ucsb.edu/pub/mirrors/linux/gentoo/ http://open-systems.ufl.edu/mirrors/gentoo http://lug.mtu.edu/gentoo/ http://gd.tuwien.ac.at/opsys/linux/gentoo/ " LANG="en_US.UTF-8" LDFLAGS="-Wl,-O1 -Wl,--as-needed" LINGUAS="en_US zh_CN" MAKEOPTS="-j3" PKGDIR="/usr/portage/packages" PORTAGE_CONFIGROOT="/" PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --compress --force --whole-file --delete --stats --timeout=180 --exclude=/distfiles --exclude=/local --exclude=/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/usr/local/portage" SYNC="rsync://rsync.gentoo.org/gentoo-portage" USE="X a52 aac aalib accessibility acl aiglx alsa altivec asf aspell automount berkdb browserplugin bzip2 cairo cdr chardet cjk cli consolekit cpdflib cracklib crypt cups curl cxx dba dbus dri dts dvd dvdnav dvdr eds emerald encode esd exif extras faac faad2 fb ffmpeg flac foomaticdb fortran gd gd-external gdbm gimpprint glitz gmp gnome gnomecd gnutls gpm gstreamer gtk gtk2 hal ibm iconv imagemagick imlib ipv6 java jpeg jpeg2k lcms ldap mad mbox mng modules mp3 mpeg mpg mudflap mysql nas ncurses nls nptl nptlonly ogg opengl openmp oss pam pcre pdf perl png ppc64 ppds pppd python qt qt3support readline reflection samba saslradeon sdl secure-delete session slang spell sqlite ssl svg sysfs tcpd theora threads tiff unicode v4l v4l2 vcd vidix vorbis webkit wma xanim xcb xcomposite xine xml xml2 xmms xorg xulrunner xv xvid zlib" ALSA_CARDS="aoa" ALSA_PCM_PLUGINS="adpcm alaw asym copy dmix dshare dsnoop empty extplug file hooks iec958 ioplug ladspa lfloat linear meter mmap_emul mulaw multi null plug rate route share shm softvol" APACHE2_MODULES="actions alias auth_basic authn_alias authn_anon authn_dbm authn_default authn_file authz_dbm authz_default authz_groupfile authz_host authz_owner authz_user autoindex cache cgi cgid dav dav_fs dav_lock deflate dir disk_cache env expires ext_filter file_cache filter headers include info log_config logio mem_cache mime mime_magic negotiation rewrite setenvif speling status unique_id userdir usertrack vhost_alias" COLLECTD_PLUGINS="df interface irq load memory rrdtool swap syslog" ELIBC="glibc" GPSD_PROTOCOLS="ashtech aivdm earthmate evermore fv18 garmin garmintxt gpsclock itrax mtk3301 nmea ntrip navcom oceanserver oldstyle oncore rtcm104v2 rtcm104v3 sirf superstar2 timing tsip tripmate tnt ubx" INPUT_DEVICES="evdev keyboard mouse" KERNEL="linux" LCD_DEVICES="bayrad cfontz cfontz633 glk hd44780 lb216 lcdm001 mtxorb ncurses text" LINGUAS="en_US zh_CN" QEMU_SOFTMMU_TARGETS="i386 x86_64" QEMU_USER_TARGETS="i386 x86_64" RUBY_TARGETS="ruby18" USERLAND="GNU" VIDEO_CARDS="ati radeon fbdev" XTABLES_ADDONS="quota2 psd pknock lscan length2 ipv4options ipset ipp2p iface geoip fuzzy condition tee tarpit sysrq steal rawnat logmark ipmark dhcpmac delude chaos account" Unset: CPPFLAGS, CTARGET, EMERGE_DEFAULT_OPTS, FFLAGS, INSTALL_MASK, LC_ALL, PORTAGE_COMPRESS, PORTAGE_COMPRESS_FLAGS, PORTAGE_RSYNC_EXTRA_OPTS
Does downgrading to mesa-7.7 make the problem disappear? Can you try mesa-7,9_rc2 from the x11 overlay too?
Is this with Mesa software rendering (without hardware accelleration) or with some drivers?
You didn't say what video card you've got, but it's probably a Radeon. If so, it's probably this bug: https://bugs.freedesktop.org/show_bug.cgi?id=26496 I haven't had time to work on it past bisecting.
(In reply to comment #1) > Does downgrading to mesa-7.7 make the problem disappear? Can you try > mesa-7,9_rc2 from the x11 overlay too? > No, downgrading to mesa-77 doesn't make any difference. I'll try 7.9 version when having time.
(In reply to comment #2) > Is this with Mesa software rendering (without hardware accelleration) or with > some drivers? > Hardware accelleration with r300. Tha card is Radeon 9800XT.
(In reply to comment #1) > Does downgrading to mesa-7.7 make the problem disappear? Can you try > mesa-7,9_rc2 from the x11 overlay too? > Just tried mesa-7.9 in x11 overlay, but the problem's same as before. I tried Radeon KMS in kernel 2.6.34-gentoo-r1 too, it simply didn't work for hardware acceleration with error "RADEONDRIGetVersion failed to open the DRM" in Xorg.0.log.
Do you by chance use radeonfb or offb? Those are incompatible with modern X.org and will cause problems.
Please answer Chí-Thanh's question in comment #7 and please also attach your Xorg.0.log and your xorg.conf (if you have one). Thanks
I'm not sure if you just missed my response, or you're ignoring me or what? :p He's reporting the exact same symptoms as the Xorg bug I've posted, polygons appearing in random places, etc, it's mostly likely an endian bug in the Buffer Object implementation. According to his use flags, he's using a Radeon card (I didn't notice it at first). And I'm not sure what you're talking about wrt to the older driver not working with radeonfb, but I'm currently running with an accelerated X server(1.8.2) and radeonfb. KMS isn't feature complete on PowerPC yet (dimming the display and sleep are not yet implemented) so laptop and iMac users are unlikely to switch at this time.
And, if you're suggesting a change to KMS on PowerPC, as far as I know this is the latest information: http://www.mail-archive.com/dri-devel@lists.sourceforge.net/msg50265.html For what it's worth, on 2.6.36, it doesn't work on my PowerBook because the AGP driver locks up before the console even shows. We don't have a fallback device because we don't have a VGA console like x86.
(In reply to comment #9) > I'm not sure if you just missed my response, or you're ignoring me or what? :p > > He's reporting the exact same symptoms as the Xorg bug I've posted, polygons > appearing in random places, etc, it's mostly likely an endian bug in the Buffer > Object implementation. According to his use flags, he's using a Radeon card (I > didn't notice it at first). And I'm not sure what you're talking about wrt to > the older driver not working with radeonfb, but I'm currently running with an > accelerated X server(1.8.2) and radeonfb. > I am sorry T did miss your response. It seems the problem is same as yours, thank you. > KMS isn't feature complete on PowerPC yet (dimming the display and sleep are > not yet implemented) so laptop and iMac users are unlikely to switch at this > time. >
(In reply to comment #7) > Do you by chance use radeonfb or offb? Those are incompatible with modern X.org > and will cause problems. > I used radeonfb driver for my card, but I don't know how to disable it. I tried to comment CONFIG_FB_RADEON out and enable KMS in .config file before compilation, but rebooted to a blank screen. can I boot to a working screen without using radeonfb? Xorg.0.log and xorg.conf files follow.
Created attachment 249993 [details] Xorg.0.log file
Created attachment 249995 [details] xorg.conf file
(In reply to comment #10) > And, if you're suggesting a change to KMS on PowerPC, as far as I know this is > the latest information: > > http://www.mail-archive.com/dri-devel@lists.sourceforge.net/msg50265.html > > For what it's worth, on 2.6.36, it doesn't work on my PowerBook because the AGP > driver locks up before the console even shows. We don't have a fallback device > because we don't have a VGA console like x86. > Thank you for your info, but somehow here the link cann't be accessed right now. I'll try later.
offb will _maybe_ work. radeonfb will definitely cause problems. Do not use radeonfb.
Can you please tell me what problems radeonfb will cause? As I've outlined above, we cannot reliably use the new KMS driver on PowerPC yet due to a number of reasons. Offb cannot be used for the boot console because it currently does not handle switching off to the KMS driver (as outlined in the second link I posted), which would lock up due to an AGP bus corruption issue, even if it did work. As I already said, I'm using a 2.6.36 kernel with radeonfb and have an accelerated X server. The only issue at the moment is with OpenGL, which is the same issue the original reporter has. It has already been bisected and the issue is known, but nobody has found time to work on it yet.
(In reply to comment #17) > Can you please tell me what problems radeonfb will cause? Random lockups: bug 232280, bug 328389 DRI fails to initialize: bug 298219 OpenGL slowdown: bug 315367 On bugs.freedesktop.org you will find some more bugs related to radeonfb. In short, the memory management in radeonfb and recent X.org will collide. > Offb cannot be used for the boot console because it currently does > not handle switching off to the KMS driver (as outlined in the second link I > posted), which would lock up due to an AGP bus corruption issue, even if it did > work. Recent KMS should implement offb handover already. AGP troubles can be avoided with radeon.agpmode=-1 kernel parameter (this will reduce performance though). > same issue the original reporter has. It has already been bisected and the > issue is known, but nobody has found time to work on it yet. If you had mentioned that you are using radeonfb in the upstream bug, it would probably have been closed as INVALID.
I have not seen any of the reported issues with RadeonFB, but since I don't use OpenGL, this might explain why. In the freedesktop bug that I linked, it's clearly explained that the bug is present when not using KMS. In the interest of finding a solution, I tried KMS again on my Powerbook. The issues that are listed in the email still remain. I cannot put the laptop into sleep. The backlight cannot be dimmed. As for OpenGL, it does not have the same issues as with using RadeonFB, but it is so slow as to be completely unusable. The driver is definitely using Direct Rendering, I can post logs if you'd like. I really can't recommend that on PowerPC, that our users use KMS. It's not feature complete and is still more functional with the old RadeonFB driver. Although the KMS driver is in theory better, it really isn't in practice, at least on PowerPC.
radeonfb causes problems both with KMS enabled and KMS disabled, in 2D and 3D operation. The problems range from slowdown to high CPU load and instability (not all have to manifest in all configurations). On Intel systems, acpi_video is used for backlight. On systems with no ACPI, patches have been posted to linux-kernel[1] but the inclusion status is not clear to me. [1] http://kerneltrap.org/mailarchive/linux-kernel/2010/9/8/4617074
(In reply to comment #16) > offb will _maybe_ work. radeonfb will definitely cause problems. Do not use > radeonfb. > Could you tell me how to NOT use radeonfb? In kernel configuration, or other conf files?
I tried KMS again with mesa-7.9, xorg-server-1.9 and kernel 2.6.36, now it works for me, though having some problems with 3D applications: -- glxgears, compiz, and celestia worked but with wrong colors, when classic driver used. -- glxgears worked properly, compiz crashed X, celestia with texture problems, when gallium driver used. -- scorched3d did not work properly.
(In reply to comment #22) > I tried KMS again with mesa-7.9, xorg-server-1.9 and kernel 2.6.36, now it > works for me, good :) > though having some problems with 3D applications: > -- glxgears, compiz, and celestia worked but with wrong colors, when classic > driver used. > -- glxgears worked properly, compiz crashed X, celestia with texture problems, > when gallium driver used. > -- scorched3d did not work properly. Comment #23 already. May I suggest you open a new bug and detail this there? And close this one as FIXED
