Re: [gentoo-user] Gentoo sources
On Sun, 20 Mar 2016 10:52:12 + Peter Humphrey wrote: > Hello list, > > Does anyone know what's happened to > sys-kernel/gentoo-sources-4.1.15-r1? It was the latest stable version > until today, when it just disappeared. I can't see anything in the > change log, and Google doesn't help. > It is removed ("cleanup") with this commit https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=ac2e2c329145fe83037f8081018e502a953ab50d Looks like it was a mistake https://bugs.gentoo.org/show_bug.cgi?id=577822
[gentoo-user] Portage cleaning up 32bit preserved libs (x86_64/amd64)
Hi list, I have a problem with cleaning/removing 32bit libraries on my x86_64 system (amd64 multilib). At one point I needed skype for something, and it needs a lot of 32bit libs (abi_x86_32) everything was fine, little fight with freetype+harfbuz (circular dep). And after some time I'm happy again without need for proprietary closed source software, so I wanted to remove it, first thing I did was: # emerge -C skype It was fine, next I wanted to remove unneeded dependencies # emerge --depclean and after that I removed package.use file which contained all dependencies with USE="abi_x86_32", and did world update # emerge -uND world everything was "fine" except it left 'preserved-libs', and told me to run # emerge @preserved-rebuild which did nothing, so currently I don't need those libs, and don't know what is a proper way to get rid of them. Here is the list: app-arch/bzip2-1.0.6-r6 /usr/lib32/libbz2.so.1.0.6 /usr/lib32/libbz2.so.1 dev-libs/glib-2.44.1 /usr/lib32/libglib-2.0.so.0.4400.1 /usr/lib32/libglib-2.0.so.0 media-gfx/graphite2-1.2.4-r1 /usr/lib32/libgraphite2.so.3.0.1 /usr/lib32/libgraphite2.so.3 media-libs/freetype-2.5.5 /usr/lib32/libfreetype.so.6.11.4 /usr/lib32/libfreetype.so.6 media-libs/harfbuzz-0.9.41 /usr/lib32/libharfbuzz.so.0.941.0 /usr/lib32/libharfbuzz.so.0 media-libs/libpng-1.6.18 /usr/lib32/libpng16.so.16.18.0 /usr/lib32/libpng16.so.16 sys-libs/zlib-1.2.8-r1 /lib32/libz.so.1.2.8 /lib32/libz.so.1 and this is how they relate to each other: !!! existing preserved libs: >>> package: media-libs/harfbuzz-0.9.41 * - /usr/lib32/libharfbuzz.so.0 * - /usr/lib32/libharfbuzz.so.0.941.0 * used by /usr/lib32/libfreetype.so.6 (preserved) * used by /usr/lib32/libfreetype.so.6.11.4 (preserved) >>> package: dev-libs/glib-2.44.1 * - /usr/lib32/libglib-2.0.so.0 * - /usr/lib32/libglib-2.0.so.0.4400.1 * used by /usr/lib32/libharfbuzz.so.0 (preserved) * used by /usr/lib32/libharfbuzz.so.0.941.0 (preserved) >>> package: media-libs/freetype-2.5.5 * - /usr/lib32/libfreetype.so.6 * - /usr/lib32/libfreetype.so.6.11.4 * used by /usr/lib32/libharfbuzz.so.0 (preserved) * used by /usr/lib32/libharfbuzz.so.0.941.0 (preserved) >>> package: sys-libs/zlib-1.2.8-r1 * - /lib32/libz.so.1 * - /lib32/libz.so.1.2.8 * used by /usr/lib32/libfreetype.so.6 (preserved) * used by /usr/lib32/libfreetype.so.6.11.4 (preserved) * used by /usr/lib32/libpng16.so.16 (preserved) * used by /usr/lib32/libpng16.so.16.18.0 (preserved) >>> package: app-arch/bzip2-1.0.6-r6 * - /usr/lib32/libbz2.so.1 * - /usr/lib32/libbz2.so.1.0.6 * used by /usr/lib32/libfreetype.so.6 (preserved) * used by /usr/lib32/libfreetype.so.6.11.4 (preserved) >>> package: media-libs/libpng-1.6.18 * - /usr/lib32/libpng16.so.16 * - /usr/lib32/libpng16.so.16.18.0 * used by /usr/lib32/libfreetype.so.6 (preserved) * used by /usr/lib32/libfreetype.so.6.11.4 (preserved) >>> package: media-gfx/graphite2-1.2.4-r1 * - /usr/lib32/libgraphite2.so.3 * - /usr/lib32/libgraphite2.so.3.0.1 * used by /usr/lib32/libharfbuzz.so.0 (preserved) * used by /usr/lib32/libharfbuzz.so.0.941.0 (preserved) Use emerge @preserved-rebuild to rebuild packages using these libraries So I think this is because of that harfbuzz<->freetype circular dependency, which some people/devs on IRC confirm. " bitlord: Yes, it seems to be related to that circular dependency. libfreetype is being preserved due to libharfbuzz and vice-versa." So don't know if this is a portage bug (considering circular dependencies I guess it is really hard to manage them, so maybe not a bug really?). In the end I don't need those libs in /usr/lib32 (probably just those listed, others are there intentionally (multilib)), if I remove them I won't break anything (except maybe package manager being confused?). Does anyone know how to manage this? :-)
Re: [gentoo-user] Xfce Logout/Shutdown/Suspend
On Mon, 14 Sep 2015 19:36:53 +0200 "siefke_lis...@web.de" wrote: > Hello, > > In Xfce all buttons are disabled until log off. > > I try https://forums.gentoo.org/viewtopic-p-7803094.html but it help > nothing. > > I not know where should search mistakes and what make me crazy i has > it on both laptops. > > siefke ~ $ ck-list-sessions > Session1: > unix-user = '1000' > realname = '(null)' > seat = 'Seat2' > session-type = '' > active = FALSE > x11-display = ':0.0' > x11-display-device = '/dev/tty7' > display-device = '' > remote-host-name = '' > is-local = FALSE > on-since = '2015-09-14T16:50:41.656831Z' > login-session-id = '2' > > siefke ~ $ upower -d > > Daemon: > daemon-version: 0.99.2 > on-battery: yes > lid-is-closed: no > lid-is-present: yes > critical-action: PowerOff > > Has someone a idea? > > Silvio You use consolekit, so how do you start your sessions? on the wiki[1] it says 'ck-list-sessions' should display 'active' and 'is-local' as 'TRUE' (works here I use lightdm as a display manager). Also check the USE flags and kernel options needed for consolekit which are listed on the wiki link[1] [1] https://wiki.gentoo.org/wiki/ConsoleKit
Re: [gentoo-user] Java plugin in Midori web browser
On Thu, 27 Aug 2015 20:25:43 -0300 José Romildo Malaquias wrote: > I need help getting java plugin working in the Midori web browser. > > I have dev-java/oracle-jdk-bin-1.8.0.51 installed on my ~amd64 > system. > > The plugin works with Firefox. > > Romildo > Have you tried to go into 'Preferences > Extensions' it is probably listed there, and you can enable/disable it?
Re: [gentoo-user] Re: Anyone using xfce4 with compositing turned off?
On Sat, 2015-08-22 at 19:08 -0700, walt wrote: > On Sat, 22 Aug 2015 04:08:41 +0200 > wrote: > > > I'm using XFCE as DE and xfwm4 as WM. Since I bought a new GPU > > (Radeon > > R7 250), I don't use compositing any more because it causes tearing > > when I watch videos in fullscreen with 3840x2160. With this GPU I > > also had some random freezes when compositing was enabled. > > > > Beside this, performance is very good, regardless compositing is > > enabled or disabled. Scrolling text or moving windows around is a > > bit > > faster and smoother with compositing enabled, especially when other > > windows are in the foreground. > > > > With my old GPU (Radeon HD4550) I always had compositing enabled. > > Everything was smoother and I saw absolutely no glitches, but > > performance was also good with compositing disabled, just not quite > > as smooth as with > > I forgot about xf86-video-ati until you mentioned it, so I just > emerged > it and (I think) made all the changes needed to reconfigure Xorg to > use > it instead of fglrx. > > Maybe I'm just too tired right now to think straight, but the error > messages I see in Xorg.log tell me that my video chip is not > supported. > > But, in the process of switching to xf86-video-ati and then back > again > to fglrx I noticed this error message from xfwm4: > > "Error opening /dev/dri/card0: No such file or directory" > > Correct, I have no /dev/dri directory. Do you have one? > > For radeon (free driver) you need to configure more than Xorg, check wiki article about radeon driver [1], It needs in kernel support, also most cards especially "newer" (>=r600) need proprietary firmware. [1] https://wiki.gentoo.org/wiki/Radeon
Re: [gentoo-user] Mysterious library access...
On Sun, 21 Jun 2015 10:27:33 +0200 meino.cra...@gmx.de wrote: > Hi, > > After trying to play a flac-file with mpv I got this error message: > Playing: Track_01.flac > (+) Audio --aid=1 (flac) > File tags: > Artist: Unknown > Album: Unknown Disc > Genre: Alternative > Title: Track 01 > Track: 1 > ALSA > lib > /var/tmp/portage/media-libs/alsa-lib-1.0.29/work/alsa-lib-1.0.29/src/pcm/pcm_dmix.c:1024:(snd_pcm_dmix_open) > unable to open slave [ao/alsa] Playback open error: Device or > resource busy connect(2) call to /dev/shm/jack-0/default/jack_0 > failed (err=No such file or directory) attempt to connect to server > failed [ao/jack] cannot open server ALSA > lib > /var/tmp/portage/media-libs/alsa-lib-1.0.29/work/alsa-lib-1.0.29/src/pcm/pcm_dmix.c:1024:(snd_pcm_dmix_open) > unable to open slave Could not open/initialize audio device -> no > sound. Audio: no audio : 00:00:00 / 00:05:56 (0%) > > I reinstalled alsa-lib and mpv but the problem remains: > Why does alsa tries to load a library which is meant to be found > under /var/tmp/portage? > That probably means some debug message from alsa-lib where for example pcm_dmix.c:1024 is a source file 'pcm_dmix.c' and :1024 is a line where the "falling" function call is. It probably tries to open '/dev/shm/jack-0/default/jack_0' maybe your alsa configuration is a problem, or missing some kernel functionallity (just guessing). Using 'jack-audio-connection-kit'?
Re: [gentoo-user] having problems getting burned cds to work
On Mon, 25 May 2015 22:27:39 -0400 cov...@ccs.covici.com wrote: > Hi. I am trying to burn an iso image to a cd (maybe its actually a > dvd disk), and it boots without any errors, and I can mount the iso > with the loop option, but if I try to mount the cd, it complains > about ufs file systems, and if I try to mount with -t iso9660, it > just says bad file type or bad superblock and the disk is unusable. > Nothing interesting in the logs, it just says the same about ufs and > nothing about iso9660. > > I am using cdrecord from cdrtools-3.01_alpha28 and I am running the > unstable version of gentoo. > > Any assistance would be appreciated. > Sorry, I know nothing about standards, optical media and filesystems, and don't understand how it works, once I had some issues with mounting optical media, also missing a filesystem which I fixed with enabling CONFIG_UDF_FS in kernel. I tried to read about UDF... before replying, but it is too much for me, if I understand it correctly it is not just what is written to the media but also how (different modes, for adding more data later ...), so may be some option while burning media that allows you to use free space left on the disk. If this is not useful just ignore it. ;-)
Re: [gentoo-user] cryptsetup wont use aes-xts:plain64
On Sat, 18 Apr 2015 12:27:15 +0200 Marko Weber | 8000 wrote: > > hello list, > > i try to crypt a partition with cryptsetup. > Yes, in Kernel i had all need things i think. > > CONFIG_CRYPTO=y ... > # CONFIG_CRYPTO_HW is not set > > > but when i try to use cryptsetup i get this: > > # cryptsetup -c aes-xts:plain64 -y -s 256 luksFormat > /dev/mapper/VolGroup01-media2 > > WARNING! > > This will overwrite data on /dev/mapper/VolGroup01-media2 irrevocably. > > Are you sure? (Type uppercase yes): YES > Enter passphrase: > Verify passphrase: > device-mapper: reload ioctl on failed: Invalid argument > Failed to setup dm-crypt key mapping for device > /dev/mapper/VolGroup01-media2. > Check that kernel supports aes-xts:plain64 cipher (check syslog for > more info). > > > > Any ideas? > > i built cryptsetup with this useflags: > > nls openssl python udev urandom > > > > cryptsetup --help shows me i am able to use the options > > Default compiled-in device cipher parameters: > loop-AES: aes, Key 256 bits > plain: aes-cbc-essiv:sha256, Key: 256 bits, Password > hashing: ripemd160 > LUKS1: aes-xts-plain64, Key: 256 bits, LUKS header hashing: > sha1, RNG: /dev/random > > > any help / ideas or knowledge welcome. > > best regards > > marko > > > > > I read the whole tread, but will reply here. I use this mode on some devices, and for me works fine (gentoo systems), I have it enabled in kernel, also I have cryptsetup with sys-fs/cryptsetup-1.6.5 (gcrypt nls python_single_target_python2_7 python_targets_python2_7 python_targets_python3_3 udev) You can probably ignore python*, 'gcrypt' is probably important USE flag. Also something which maybe unrelated to you but is important about CONFIG_CRYPTO_XTS is "XTS: IEEE1619/D16 narrow block cipher use with aes-xts-plain, key size 256, 384 or 512 bits. This implementation currently can't handle a sectorsize which is not a multiple of 16 bytes."
Re: [gentoo-user] [OT] Strange behaviour of google certificates.
On Thu, 2 Apr 2015 11:57:26 +0300 Gevisz wrote: > On Thu, 2 Apr 2015 03:52:40 -0400 "Walter Dnes" > wrote: > > > On Wed, Apr 01, 2015 at 08:19:45PM +0300, Gevisz wrote > > > > > So, I am using Claws Mail that downloads e-mails from several > > > google mail accounts (all are mine :) and about once or twice > > > in a month get into the situation when Claws asks me to verify > > > and change the google certificates, first in one direction and > > > soon after that (usually during the next downloading of my > > > e-mails) > > > - in another. > > > ... > > > > The 2 servers probably have different certificates, which is why > > you get this behaviour. I suggest going into "apk mode" and > > putting an entry into your hosts file , like... > > > > 173.194.192.108 pop.gmail.com > > > > This will force your system to always use the same server, and > > avoid the re-validation every time you hit the other server from > > the one you used the previous time. > > Thank you for your advice. Added that line to my /etc/hosts file. > After that Claws asked to verify the google certificate once again, > but I hope that that was the last time this month and that that > madness with google certificates finally ends. (Because in the last 2 > days this situation repeated at least 20 or more times.) > > By looking at the screenshoots that is >=claws-mail-3.10.x (I think that is the version when it got support for validating certificate chains)? There is a option in Configuration > Edit Accounts ... then for every account you have "SSL" options, you can check to accept "unknown valid certificates" so it will do it automatically, won't ask if there is a new certificate and it is valid.
Re: [gentoo-user] [OT] Mysterious vanishing of DNS entry of www.youtube.com...was I hacked?
On Tue, 10 Mar 2015 19:16:12 +0100 meino.cra...@gmx.de wrote: > Hi, > > the following happens some minutes before: > I was searching on youtube for some reviews... > and suddenly BOOM: "Server not found: Unknow host" > > I restarted firefox...which did not help. > > I did a ping & traceroute to www.youtube.com from > the commandline...same results... > > Wireshark shows the DNS query to my DSL modem... > and the answer was that from above. > > I rebooted my Gentoo box...no help... > > The problem vanishes as I powercycled my DSL modem. > > Any other access was working the whole time. > > Was my DSL modem hacked? > Does anyone else noticed a glithc in the matrix? > > Best regards, > mcc > > Today everyone can make a dsl router/modem, and ISPs when they give you equipment they want it to be cheap. I saw exactly the same problem on cheap dsl routers, I don't have enough knowledge to debug it, or to be sure if it is safe or not, it worked for months with no issues, no reboots ..., one day it stopped working/resolving hosts, and its dhcp server config doesn't allow "forwarding" of DNS servers from WAN configuration, or manually setting one which will be given to the dhcp "clients", so there are two ways, it can work as a proxy (send its own IP as DNS server to clients), or it doesn't provide DNS, so you need manually to configure each client. If you restart the router, it works sometimes for few hours, sometimes few days, but it will fail again randomly. The only worrying thing is that it worked for months without an issue.
Re: [gentoo-user] Stroppy perl-cleaner
On Sat, 14 Feb 2015 11:33:49 + Peter Humphrey wrote: > Hello list, > > What is perl-cleaner objecting to here? It all seems pretty anodyne > to me: > > $ alias perl-cleaner > alias perl-cleaner='sudo perl-cleaner --all -- --ask --usepkg > --jobs=3 --keep-going' $ perl-cleaner > > *** > You are supplying additional command line options for the package > manager. This is NOT RECOMMENDED, not tested, and may lead to > incorrect, incomplete, confusing, and/or nonfunctional results. You > are on your own now. > *** > You are using something which is not recommended by developers, that is '-- ', telling perl-cleaner to pass additional package manager parameters, not sure why is not recommended, but it warns you, also in 'perl-cleaner --help' or man page it says it is not recommended to do that. There is probably case when you can do something wrong, but also it can be safe too passing some parameters which don't affect build like '--ask'. One case which may be problematic (or not, just guessing) is --jobs=N, what will happen if one perl module depend on the other, and they are building at the same time?
Re: [gentoo-user] perl-cleaner lerfovers
On Sat, 14 Feb 2015 13:13:25 +0200 Alexander Kapshuk wrote: > 'perl-cleaner --all' generated the following output. > > * Finding left over modules and header > > * The following files remain. These were either installed by hand > * or edited. This script cannot deal with them. > > /usr/lib/perl5/vendor_perl/5.16.3/XML/SAX/ParserDetails.ini > /usr/lib/perl5/vendor_perl/5.18.2/XML/SAX/ParserDetails.ini > /usr/lib/perl5/5.12.4/i686-linux/Encode/ConfigLocal.pm > > What's the recommended way to go about this? > > Thanks. As I understand this, it is safe to remove and that is what I do when they appear on my system, if you don't have perl 5.16.3, 5.18.2 or 5.12.4 ..., and updated/rebuild all perl modules with perl-cleaner. I also used 'qfile /path/to/file' (from portage-utils) to check if they belong to any installed package. (which is probably not needed, per-cleaner knows about this?)
Re: [gentoo-user] graphviz won't compile
On Sat, 14 Feb 2015 10:46:16 +0100 (CET) Alain Didierjean wrote: > Cannot emerge graphviz : > > ERROR: media-gfx/graphviz-2.26.3-r4::gentoo failed (prepare phase): > * (no error message) > * > * Call stack: > * ebuild.sh, line 93: Called src_prepare > * environment, line 5657: Called die > * The specific snippet of code: > * cp "${EPREFIX}"/usr/share/libtool/config/install-sh > config || die; > > libtool ebuild doesn't install /usr/share/ligtool/config directory. > > Known bug, or should I report it ? > Thx > Already reported, https://bugs.gentoo.org/show_bug.cgi?id=537850
Re: [gentoo-user] Switching off some linguas variables
On Wed, 11 Feb 2015 21:09:44 +0200 Gevisz wrote: > What is the elegant way to switch off all but one linguas variables > for a given package. > > I have tried all "obvious" solutions but they seem to do not work. > > For example, I have tried to put the following line > into /etc/portage/package.use file: www-client/chromium -nls > -linguas* linguas_en linguas_pl > > So far I am afraid to recompile everything with global -nls USE flag > and LINGUAS="en" in /etc/portage/make.conf. So, trying to cut the > cat's tail by parts. :) > For package specific env, check https://wiki.gentoo.org/wiki//etc/portage/env
Re: [gentoo-user] rebuild of graphviz failed
On Tue, 10 Feb 2015 00:55:00 +0100 wrote: > Hi folks, > > after the last update, rebuild of graphviz failed with: > > * Package:media-gfx/graphviz-2.26.3-r4 > * Repository: gentoo > * Maintainer: graph...@gentoo.org > * USE:abi_x86_64 amd64 cairo doc elibc_glibc examples gtk > kernel_linux nls perl python userland_GNU > * FEATURES: ccache preserve-libs sandbox userpriv usersandbox > >>> Unpacking source... > >>> Unpacking graphviz-2.26.3.tar.gz > >>> to /var/tmp/portage/portage/media-gfx/graphviz-2.26.3-r4/work > >>> Source unpacked > >>> in /var/tmp/portage/portage/media-gfx/graphviz-2.26.3-r4/work > >>> Preparing source > >>> in > >>> /var/tmp/portage/portage/media-gfx/graphviz-2.26.3-r4/work/graphviz-2.26.3 > >>> ... > * Applying graphviz-2.26.3-libtool.patch ... > [ ok ] > * Applying graphviz-2.26.3-automake-1.11.2.patch ... > [ ok ] > cp: cannot stat '/usr/share/libtool/config/install-sh': No such file > or directory > * ERROR: media-gfx/graphviz-2.26.3-r4::gentoo failed (prepare phase): > * (no error message) > * > * Call stack: > * ebuild.sh, line 93: Called src_prepare > * environment, line 5681: Called die > * The specific snippet of code: > * cp "${EPREFIX}"/usr/share/libtool/config/install-sh > config || die; > * > * If you need support, post the output of `emerge --info > '=media-gfx/graphviz-2.26.3-r4::gentoo'`, > * the complete build log and the output of `emerge -pqv > '=media-gfx/graphviz-2.26.3-r4::gentoo'`. > * The complete build log is located at > '/var/log/portage/media-gfx:graphviz-2.26.3-r4:20150209-234745.log'. > * The ebuild environment file is located at > '/var/tmp/portage/portage/media-gfx/graphviz-2.26.3-r4/temp/environment'. > * Working directory: > '/var/tmp/portage/portage/media-gfx/graphviz-2.26.3-r4/work/graphviz-2.26.3' > * S: > '/var/tmp/portage/portage/media-gfx/graphviz-2.26.3-r4/work/graphviz-2.26.3' > > Any recommendations how I can solve this problem? > > Regards > wabe > Already reported https://bugs.gentoo.org/show_bug.cgi?id=537850 I did a trick, downgraded libtool, installed graphviz, upgraded libtool ... (worked for me on two systems)
Re: [gentoo-user] Issues with USB memory sticks
On Mon, 02 Feb 2015 18:22:17 -0800 Daniel Frey wrote: > I've been having a strange issue with my USB flash drives. > > I can't seem to copy much anything to them. It appears to copy and > then hangs, then I get: > > [ 6841.490036] usb 2-6: reset high-speed USB device number 16 using > ehci-pci [ 6848.540029] usb 2-6: reset high-speed USB device number > 16 using ehci-pci [ 6855.590031] usb 2-6: reset high-speed USB device > number 16 using ehci-pci [ 6862.640026] usb 2-6: reset high-speed USB > device number 16 using ehci-pci [ 6869.690025] usb 2-6: reset > high-speed USB device number 16 using ehci-pci [ 6876.740023] usb > 2-6: reset high-speed USB device number 16 using ehci-pci > > > dmesg is full of these messages. > Not an expert, but check contacts (if ports are not used for a long time check if there is dust in them), if you use extension cable or a hub, try without it (power issues), or try different port (does it happen on all ports?) Also check other messages from kernel, they can be useful diagnosing this issue.
Re: [gentoo-user] [OT] ASUS Crosshair V Formula-Z MB
On Sunday, February 01, 2015 12:16:20 PM Dan Johansson wrote: > I just got me a new MB, a ASUS Crosshair V Formula-Z, after my old MB > fried due to a fan failure. > > Is anyone else (on this list) using this MB, with Gentoo of cause? > If you are, have you got all the sensors working? > I can not get information from all the sensors on this MB (compared with > what the BIOS shows) and some of the values that do show are wrong , > some temp-sensor shows as negative (and my room is not THAT cold). > > Regards, I don't have that motherboard, but most ASUS boards I used (still use one, really old board (p5b)) use SENSORS_ATK0110.
Re: [gentoo-user] Radeon driver - blank console
On Monday, January 05, 2015 02:42:46 PM Helmut Jarausch wrote: > Hi, > > the recent update of virtual/opengl forced me to switch from my ati > [legacy] driver to the open source driver (radeon) > > While I got X11 running, my console gets blanked before I can log in > on the command line. > > Which (framebuffer?) kernel parameters have to set? > It looks like the kernel or init process (openrc) tries to switch to a > higher resolution (on the text console) when the screen (console) gets > blanked. > > (Loggin in "blindly" and invoking startx, I get a running X-server). > > Many thanks for some pointers, > Helmut I'm not sure what is the problem there, I have amd/ati 4350 with free driver on my system, works fine with X or tty. I don't know how you configured your system, but you can always check gentoo_wiki/radeon [1], most modern amd/ati cards need non-free firmware to work properly, even for some "more basic" stuff, not only 3d (I think KMS needs non-free firmware too). [1] - http://wiki.gentoo.org/wiki/Radeon