Nightwulf|work: hi all
edman007: hello
Nightwulf|work: hi edman007
Soul_keeper: excellent work
dto: should i enable the compositor in XFCE? it looks nice but will it be slow? i'm on debian lenny
dto: it doesn't seem terribly slow when i move transparent windows around etc.
loki_666_: hi
loki_666_: i've a question out of topic
loki_666_: doest directfb support radeonhd cards?
dragonstear: Is today the big merge day... Or why is CS now the git master branch?
peter3: hi all, is radeonhd expected to work in 2D mode with 3 moinitrs (xinerama on, no SLI)? one monitor on Radeon HD 4850 and the two others on Mobilitiy Radeon HD 3450; I used a config which works with catalyst and replaced fglrx with radeonhd, but I
peter3: ... but I
peter3: ... shit; I'm gettin "Requested entity in already in use error.
libv: m82 combined with a rv770?
peter3: yes
libv: what kind of hw is this :)
peter3: an intell based machine
libv: are m82s sold as standard pci-express cards?
peter3: the goal is to have more monitors not to have some kind of SLI; one quick card is enough for games; and for work I prefer to have more screan space
peter3: yes; I did not build the machine myself but the cards are in pcie slots
peter3: Intel x48 chipset; two pcie 16 slots
libv: peter3: are you certain that this is a mobility chip?
libv: peter3: and not just a standard hd 3450?
peter3: maybe they did put mobile versionthere because it is pasively cooled (that was a requirement, so that the machine is quiet)
peter3: let me check once again
peter3: lspci reports:
peter3: 01:00.0 VGA compatible controller: ATI Technologies Inc Device 9442 (I found this should be the Radeon HD 485)
peter3: 02:00.0 VGA compatible controller: ATI Technologies Inc Mobility Radeon HD 3450 (and for this I found somehere the M
ndim: lspci schould also have the device ids, just to make completely sure what chips you have.
libv: peter3: lspci -n -s 2:0.0
peter3: libv: just figured it out myslef: the id is 95c5
peter3: for the one which reports as Mobilitiy
dto: hi.
dto: i am about to paste my xrandr -q output and my xorg.conf
dto: (on lisppaste.org)
dto: http://paste.lisp.org/display/65586
dto: when i issue the command shown after xrandr -q output,
libv: RHD_DEVICE_MATCH( 0x95C5, RHD_RV620 ), /* ATI Radeon HD 3450 */
libv: standard rv620 apparently :)
libv: pci.ids seems off :)
dto: the left screen looks fine, the right screen is all garbage and unusable.
peter3: libv: ok; anyway is this supposed to work or is it a "feature" that I cannot get this working with 3 monitors? (it works with catalist; and I'm not interested in 3D yet)
libv: peter3: it should work, could you send in your (-logverbose 7) log and conf file to the mailinglist
ndim: dto: what radeonhd version? -logverbose 7 Xorg.0.log?
peter3: libv: yes; can you tell mi the mailing list address
libv: http://wiki.x.org/wiki/radeonhd
dto: how do I tell my radeon version
peter3: libv: ok; thanks; I'll post it there
ndim: dto: Xorg.0.log contains it, so no need
dto: one moment.
dto: (II) RADEONHD: version 1.2.1, built from non-git sources
dto:
dto: (II) Primary Device is: PCI 01:00:0
dto: (--) Assigning device section with no busID to primary device
dto: (--) Chipset RV630 found
dto:
dto: i have some nice hardware to test this stuff on. should i start using the git version of the driver?
dto: (--) RADEONHD(0): Detected an RV630 on an unidentified card
dto:
dto: let me paste the whole thing
ndim: Urgh.
dto: http://dto.mamalala.org/xorglog
dto: hardware: http://dto.mamalala.org/images/podpan-small.jpg
ndim: "non-git sources". Then you'll need to give more information. Yeah, the git version will probably give you better results, even if there is a slight chance it does something worse
dto: this is whatever comes with debian lenny.
ndim: OK, that is old.
dto: awright. now, i prefer to rebuild deb packages whenever i can.
dto: instead of just making my own with checkinstall or whateverr.
dto: or, should i just make && make install?
ndim: dto: git clone, autogen.sh, make distcheck, rebuild deb package with that tarball
dto: ndim: awesome :) gimme 5 minutes
ndim: or "make dist" should do the tarball generation job much more quickly
ndim: dto: Or, for quick testing, just configure --prefix=/path/to/someplace and tell Xorg to look for drivers there.
dto: ok... apt-get source xf86-video-radeonhd
dto: oops wait/.
dto: wait, so i don't need to download the source?
dto: i mean, from apt-get source
ndim: For the deb stuff, you need it.
ndim: I can't quite remember how Debian packages worked, though.
dto: i'm not sure what to do next. i cd'ed into the xserver-xorg-video-radeonhd-1.2.1 directory
dto: and typed git pull
ndim: IIRC upstream source tarball + one patch creating all the stuff in the debian/ subdir.
dto: but it says it's not a git repo.
dto: what i usually do is apt-get source foo, cd foo-n.m, make my changes (edit makefile or whatever) then issue "debuild -us -uc" and it rebuilds the deb.
dto: what i'm not clear on is how to update what i just downloaded into the latest git.
ndim: I haven't built Debian packages in years.
ndim: With RPM, this would be quite easy.
dto: ok, what's my alternative? just uninstall the debian radeonhd driver and make my own package?
dto: that might be easier
ndim: dto: configure --prefix=/some/place
dto: ah. then how do i tell xorg to look there first?
ndim: Section "Files"
ndim: ModulePath "/some/place/lib/xorg/modules,/usr/lib/xorg/modules"
ndim: EndSection
ndim: Or, with one-off testing, just "Xorg -modulepath"
dto: ok :)
dto: one moment
ndim: (Make sure /some/place is mounted when X is started :-)
dto: haha
dto: compiling
dto: i love to watch this dual-core run
dto: ok, brb, got to reboot x
dto: hello. i have verified that the new driver is running.
dto: in Xorg.0.log
dto: now. going to try xrandr... if you see me crash you'll know it didn't worik....
dto: work
dto: uploading images of the ensuing chaos
dto: good screen on left: http://dto.mamalala.org/images/hpim0935.jpg
ndim: anticipates smashed guitars
dto: bad screen on right
dto: http://dto.mamalala.org/images/hpim0936.jpg
ndim: Cute.
ndim: Xorg.0.log?
dto: shit. i think it's overwritten now right?
ndim: You should probably set Virtual in xorg.conf
ndim: Uhm.
dto: i do.
dto: i restarted X after the crash
dto: http://dto.mamalala.org/images/hpim0936.jpg
dto: oops, i mean
dto: Virtual 3040 1050
ndim: Where 3040 is?
ndim: 1400 + 1650 = 3050...
dto: http://paste.lisp.org/display/65586
dto: 1680x1360
ndim: 1360? Never heard of that.
dto: i mean, 1680+1360
dto: it's a toshiba HDTV.
dto: hold on.
dto: yeah the TV OSD confirms that res
dto: also i notice with my more recent kernel that i have an extra ALSA device that says Intel HDMI
dto: will that mean that i can play sound out to the other speakers?
dto: (attached to the hdtv)
dto: in the gaming/theater area
dto: i would like to emulate classic systems on that monitor
dto: well i'm pretty stuck.
dto: :(
dto: i also have a playstation3 and recently watched 2001 on bluray
dto: ndim: any suggestions
dto: bizarrely enough i have the soundtrack on an old vinyl lp
dto: so i hooked up my usb turntable
dto: and used the new version of ardour to record the usb audio while monitoring
dto: thru the delta 1010
dto: i really want to get video working in this setup and use my PC as a multimedia jukebox thing and buy an extra terabyte drive
ndim: dto: There is an hdmi-audio branch of radeonhd with a few patches which are supposed to enable audio output over HDMI in some way. You'd need to merge that onto master, probably, and play with it some.
ndim: NO idea whether it works.
ndim: dto: The Xorg.N.log with the broken screen could help. Just starting "Xorg :1 -logverbose 7 -modulepath ..." from a root console and then hitting ctrl-alt-backspace after X has come up should do the job (giving you an Xorg.1.log)
dto: brb
dto: ndim: ok, trying
ndim: dto: As to the garbled second screen... When you have that log file of current git, file a bug with it or mail the list, so that one of the rhd devs can take a look.
dto: ndim: http://dto.mamalala.org/notebook/Xorg.1.log
dto: but that isn't with the crash
dto: ok
dto: thanks for all your help ndim
ndim: crash? I thought you just had a scrambled screen?
dto: oh, well that's true.
dto: i was thinking crash i.e. can't use the main screen.
dto: anyway let me trigger the behavior and send it to the list
dto: thank you
dto: :)
dto: did you see my panoramic photo?
ndim: Yupp. Looks like a computer and music nerd's wet dream :)
dto: ok the mouse pointer is not distorted. just the rest of it.
dto: on the main screen.
dto: how do i reset this without killing X?
dto: is there an xrandr reset command?
ndim: --auto?
ndim: has only a single screen.
dto: didn't have much effect. a flicker
dto: ok :)
dto: i'll mail the list
dto: soon
dto: ndim: hey
dto: ndim: i did what you suggested and ran a virtual console as root with logverbose on. but xrandr worked fine!!! it still doesn't work when i use the "main" gdm-managed X server that stays on Ctrl-Alt-F7
ndim: Cute.
ndim: Some gdm/gnome/foo magic being broken, or just triggering driver bugs.
dto: it was the XFCE compositor.
dto: i turned that off, ran xrandr, now it works fine.
dto: what file do i have to put that in to make it automatic?
dto: ah i can use xfce's feature. never mind!
ndim: I don't quite understand what the "XFCE compositor" is supposed to do and what it was actually doing wrong.
dto: It enables drop shadows around windows and menus, transparent dragging, etc
dto: i had it turned on.
dto: from the root console, nothing but an xterm and no wm ran, so i figured maybe it was the presence of weird alpha blending stuff
dto: i also seem to be able to drag windows from one monitor to another. nice.
ndim: OK... so we have that composite stuff triggering (at least) one bug in radeonhd, or is that bug somewhere in that composite part?
dto: not sure :(
dto: hi. things are working smoothly now.
dto: i have a button in my menu that turns on XRANDR when i need it
dto: for the 2nd monitor :)
lupine_85: hopefully awaits r600 magic
libvde: lupine_85: if only it was magic :(
lupine_85: 'twould make like easier :)
lupine_85: ho hum, ati snuck out catalyst 8.8 while I wasn't looking