Scott1979: can nayone help me with my xorg.conf
cxo: have you just tried doing it with xrandr and deleting the xorg.conf?
DanaG: I'm sane, and I got ATI specifically, on purpose -- and really haven't regretted it.
DanaG: Granted, I do all my gaming in Windows, though.
cxo: well that statement was conditional
cxo: and you know, sanity is all relative
raevol: cxo could one reasonably say that wine is awful
raevol: ah i see what you said earlier, nvm
Scott1979: thats silly talk danag
Scott1979: i am a amd/ati man but i still need to get a nvidia card
Scott1979: scarabeus canu help me with my xorg.conf
Scott1979: dual head advice required on ati
soreau: Scott1979: What are you trying to do exactly?
soreau: gets bored and wonders off
Scott1979: can anyone help me with my ubuntu xorg.conf
idletask: Hello everyone
idletask: My girlfriend has just received a new screen with an HDMI connector, so we set to try it out... But there is a problem: the edges of the display are nearly 100 pixels off on each side of the screen, uniformly
idletask: She has an nVidia 9600GT with the proprietary driver under Linux (Ubuntu) and Windows XP
idletask: So, I said I'd give it a try
idletask: But the problem is exactly the same with my machine
chithead: possibly a setting in the screen (overscan?) can influence that
idletask: The only thing I could see in my xorg.log was this message: "(WW) EDID preferred timing clock 154.00MHz exceeds claimed max 150MHz, fixing"
idletask: Is that related? Is the screen lying?
idletask: chithead: there's no problem with the VGA input though
idletask: And Windows isn't exempt from the problem either BTW
chithead: idletask: maybe it wants a mode with reduced blanking, try to use cvt and xrandr --newmode to create one
chithead: but I think it is the screen itself
idletask: Oh err... :p I guess I have to read about that
idletask: chithead: when I am connected to the screen again, I'll try that (I'm back on my good ole trusty Hyundai here)
Scott1979: can anyone help me with my xorg.conf for ati card dual heading
chithead: Scott1979: http://wiki.debian.org/XStrikeForce/HowToRandR12
chithead: Scott1979: read III.1 and .2
Scott1979: yeah im still lost on this my xorg show a virtualize resolution and i want independant resoulutions per screen
Scott1979: chithead
Scott1979: will u view my xorg.conf if i use paste bin chithead
chithead: Scott1979: you need to add two monitor-... options to the device section
Scott1979: yes and i am unsure
Scott1979: of this action
chithead: Scott1979: the exact names are the ones listed by xrandr
chithead: so pastebin xrandr output
Scott1979: ok will do
idletask: About the ddx/mesa dri module: the mesa module is in charge of OpenGL and the ddx in charge of all acceleration stuff not OpenGL, is that correct?
Scott1979: chithead my xrandr
Scott1979: http://paste.ubuntu.com/350379/
Scott1979: and my xorg.conf
Scott1979: http://paste.ubuntu.com/350381/
chithead: Scott1979: so you need two monitor statements in device section
Scott1979: can u show me how
chithead: you did not even attempt
Scott1979: can u alter my paste bin
Scott1979: no dont want to break system
Scott1979: im scared
idletask: Scott1979: try Xorg -configure as root, it will generate an xorg.conf file for you
Scott1979: i have compiz on expanded dual screens at present
idletask: Just swap them and try it out
Scott1979: cant sudo have to be root
chithead: just back up your working xorg.conf and restore if the new one does not work
Scott1979: so ctrl alt f1
idletask: There's nothing to be scared about, X will not destroy your monitors unless they are _really_ badly configured
uzi18: one question - KMS is working on X700 ?
chithead: removing gdm/kdm/xdm from runlevel is a good idea until it works
idletask: uzi18: yes, KMS works on all supported chipsets if I'm not mistaken
Scott1979: not destroy monitors just im not only one who uses system and i cant not have it working
chithead: uzi18: http://wiki.x.org/wiki/RadeonFeature your x700 is r400 generation
Scott1979: have u got a missus
Scott1979: i do and she goes off when i break the system
Scott1979: she cant facebook
chithead: Scott1979: you will not destroy anything. if you are concerned you can boot ubuntu livecd or something
uzi18: chithead: but it is mobile version, any clue how to be sure that kms works?
idletask: uzi18: see the page above
chithead: Scott1979: and on livecd you can safely modify xorg.conf until it works
uzi18: thinking about dmesg os something
idletask: uzi18: it _does_ work from what this page says, and the fact that it's a mobile chip doesn't change anything
uzi18: os=or
Scott1979: ok yes but are u not willing to walk me thru this
chithead: uzi18: mobile and desktop variants do not differ much
uzi18: yes need to know only what to check and if it's not i will fill bug report if needed
Scott1979: show me a ati configure dual head display with two screens with indepandant resolutions
chithead: Scott1979: the wiki page lists one entire example near the end
idletask: Scott1979: ah, don't be such a wuss, just try it out - and backup your old xorg.conf
idletask: When it works, it's highly rewarding, in a geeky sort of way
idletask: :p
Scott1979: SubSection "Display"
Scott1979: Virtual 2384 768
Scott1979: walk me thru so i know for next time
chithead: Scott1979: you just need to add two lines to device section and two monitor sections as in III.1 it can't be that difficult
chithead: Scott1979: do it as good as you can and we will look at the result
Scott1979: ok
idletask: wonders whether the "no xorg.conf" case will set up a dual screen correctly
idletask: I cannot test that unfortunately...
chithead: no xorg.conf will lead to cloned screens
Scott1979: yeah clone is default
idletask: OK, good to know
Scott1979: post the website i can follow
Scott1979: so i will try
chithead: [13:14:50]
idletask: And anyway I'd need to recompile everything with xinerama support
chithead: idletask: your window manager would be enough to cover most things
idletask: chithead: even if it's compiled without xinerama support? Somehow I doubt it :p
chithead: idletask: you don't need to recompile entire system, only the window manager
chithead: and maybe the display manager if you don't want login screen in the middle of two monitors
idletask: I don't use a display manager
Scott1979: i dont even know where to start
chithead: Scott1979: section III.1
idletask: Well, I miss one DVI-to-VGA adapter anyway
Scott1979: nah still lost
Scott1979: this is hard for me to understand
Scott1979: foo and bar
Scott1979: how would u change my xorg.conf
chithead: we are not going to spoon-feed you. I already said that xrandr lists your outputs, and you replace output names with what xrandr reports
chithead: if you backup a known working xorg.conf before modifying you should be good to go
Scott1979: this is a support chat room and i dont want spoon feed i just want to be taught the process so i can epand on it
Scott1979: im a fast leearner but this is stumping me
Scott1979: i have one dvi port and one vga port on my x1600 r530 chip vid card pciexpress
chithead: xrandr says you have DVI-0 and VGA-0, doesn't it?
Scott1979: monitor on right is 1024x768 monitor on left is 1362 x 768
Scott1979: well why will it only support 1362x768 when by itself this monitor is 1680x1050
idletask: Scott1979: lack of video RAM,maybe?
Scott1979: 256 on card only
Scott1979: ok so this is my limit
Scott1979: i can deal with this
Scott1979: i am being greedy i spose
chithead: 256 should be enough for any resolution even 2560x1600
Scott1979: well thats what i thought
idletask: I was just asking :p
chithead: probably monitor does not deliver correct edid, but you can take care of this after getting output configuration correct
Scott1979: can u please help me with this
Scott1979: spoon feed me and i will be forever in your debt
chithead: Scott1979: please say what you think are the two lines you have to add to the device section
idletask: Scott1979: look at what xrandr says about your monitors
Scott1979: DVI-0 connected 1360x765+0+0
Scott1979: and
Scott1979: VGA-0 connected 1024x768+1360+0
Scott1979: under section device
Scott1979: would this be ok
chithead: Scott1979: compare with the two lines in III.1 in the wiki
adamk_: Scott1979, Take a look at this xorg.conf file: http://pastebin.com/m2c615c7 This will drive one monitor at 1600x1200 and the other monitor at the resolution the EDID says it prefers.
Scott1979: thankyou
adamk_: The modeline was generated with gtf or cvt, so you would have to do the same if one of your monitors isn't picking the best resolution.
Scott1979: but im still lost
Scott1979: chithead there is so much to look at in these posts i still am lost
chithead: the section in the wiki is less than half a page long
chithead: please at least attempt to answer how the monitor-... entries in the device section should look like
Scott1979: i am using the opnsource ati driver
idletask: Scott1979: that's not the answer now, is it?
Scott1979: well is it possible to do this with the opensource driver
idletask: Yes it is
Scott1979: mine talks about configured video device and your posts says ati
Scott1979: dont be so hard on me as i am trying to learn nd if u had a problem i would make the time to help u
Scott1979: as i do if i can help people with there issuse
adamk_: Alright, I'm back...
adamk_: The thing is, if we tell people exactly what to write in their xorg.conf file, they don't actually learn anything :-)
raevol: dreams of the day xorg has a gui
adamk_: Take a look at the xorg.conf file I posted. See how Identify my monitors DVI-0 and DVI-1? That information came from 'xrandr'.
adamk_: Yours are named 'DVI-0' and 'VGA-0' in xrandr, so those are the Identifiers you want to use for two separate Monitor sections, very much like mine, but with those identifiers.
Scott1979: and change 1600x1200 to 1680x1050
Scott1979: are u using opnsorce driver adamk
Scott1979: i dont have a section monitor in my xorg.conf
adamk_: Well, except that your xrandr output doesn't show 1680x1050 as a supported resolution. You can try and create a modeline for 1680x1050 with the gtf or cvt program.
adamk_: I am. You likely don't need one.
Scott1979: why does it not show this when it is capable of this
Scott1979: the 2nd monitor is only capable of 1024x768
adamk_: Sorry, you do need a monitor section. You need 2.
adamk_: Scott1979, Either the monitor's EDID is broken, or the radeon driver has a bug.
Scott1979: 22inch widescreen for main and 15inch 4:3 for right
Scott1979: im going to delete my xorg.conf and see what happens
adamk_: I doubt very much anything will change, and I'm positive it's not just going to miraculously make things start working.
Scott1979: see u all soon
Scott1979: i had 3 xorg.conf to delete one was failsafe the other just conf and one had a series of number behind it
xming: hehe, got kms working just by downgrading the kernel from .32.2 to .32
xming: and now X uses software render, no DRI1 any more
xming: [dri] radeon kernel module version is 2.0.0 but version 1.17.0 or newer is needed.
xming: [dri] Disabling DRI.
xming: 2.0.0 *is* newer thatn 1.17.0, ins't it?
Scott1979: ok since deleting xorg.conf it came back on reboot as mirroring screens and my main monitor show 1680x1050 and 2nd was 1024x768 when i unmirroed it drops the res on main monitor and hides 1680x1050
idletask: Is anyone interested by a piglit result set for my setup?
idletask: I have run it twice already
xming: hmm this is strange, now with kms enable and working, my X (software rendered) is faster/snappier thatn before (no kms and DRI1 enabled)
xming: moving window uses far *less* CPU than before
idletask: xming: now don't you complain :p
xming: well this X is now workable (as work station) but not for gaming (yet)
idletask: xming: what chipset again?
idletask: I have an rv790 and many games work fine
xming: 3850 (r670) agp
idletask: Others don't, for various reasons
xming: .32.2 breakss KMS on 3850
idletask: xming: I use the drm-radeon-testing branch from the drm kernel tree
xming: I am on software render here so no opengl games will be good :p
idletask: xming: too old Mesa?
xming: 7.7 is that old?
idletask: Well, no actually
idletask: But I use the master branch...
xming: Mesa was working with DRI1 enabled
xming: when using UMS not KMS
idletask: I use git's master for libdrm, mesa and the ddx
idletask: I still don't understand the interaction between all of these
idletask: All I know is that 3D mostly works :p
idletask: The latest mesa master broke something though
idletask: chithead asked me to try crack-attack and it does have display problems, as predicted
idletask: No fix yet
xming: I tried git sources, my X starts up but kb/mouse freeze
idletask: And oblivion with wine has texture problems (I also use wine's master branch)
idletask: xming: what distribution do you have?
xming: gentoo
idletask: Do you use the x11 overlay?
Scott1979: http://imagebin.org/77946
xming: yes tried that
idletask: It's what I do
xming: with all -9999 kb/mouse froze
Scott1979: can u check this out idle and adam
xming: but of course that was with .32.2
Scott1979: http://imagebin.org/77946
idletask: xming: no need to use _all_ 9999 stuff
xming: kms hangs with it I might try with .32 now
idletask: xming: here's my package.unmask --> http://paste.pocoo.org/show/161310/
xming: idletask: well driver-ati, mesa, libdrm were
idletask: xming: no need for the rest
idletask: xming: and you should definitely try the drm-radeon-testing kernel
xming: ehum
idletask: It's not in portage though, you have to do it yourself
xming: :D of course :D
idletask: xming: there's the firmware problem though
xming: idletask: this one? http://git.kernel.org/?p=linux/kernel/git/airlied/drm-testing.git;a=summary
idletask: This one is a bitch
Scott1979: http://imagebin.org/77947
Scott1979: and here
idletask: xming: no
Scott1979: my desktop
xming: ok gtg now, be be in a few hours, then I will do that
idletask: xming: http://git.kernel.org/?p=linux/kernel/git/airlied/drm-2.6.git
idletask: xming: the drm-radeon-testing branch, NOT the master branch
idletask: needs to learn how to do an initramfs because of these f... blasted firmwares
Scott1979: im good now idle deleting the xorg.conf worked and now i have the resolutions i require i aslo change a setting on my monitor to display edid
Scott1979: and it allworks fine so thanks for nothing
Scott1979: 22inch running 1680 x 1050 and 15inch running 1024 x 768 with just mouse clicks no of this change two lines crap
idletask: Scott1979: well, didn't I tell you to just try it out?
idletask: You did and it worked
Scott1979: all i need now is a background to look good on desktops
Scott1979: the jetty in mine is acrros the monitors
Scott1979: so i need to look for a good desktop back ground
Scott1979: did u see my image bin posts
Scott1979: what deleting the xorg.conf
Scott1979: i dont think it was this i feal that the settings had been changedon monitor from missus as we switch between xbox 360 and pc via buttons on front of monitor
Scott1979: ddc/ci was on
Scott1979: i turned it of
Scott1979: im stoked now as i got the monitor for free today and all is well now
Scott1979: see u all im off to be productive in bussiness rather then be held up by demeaning peoples that cant really help maybe try to help more guys as we are here to help u maybe u are beyond help i dont know but curtesy will get u further then think the smug thoughts
Scott1979: i am here to help u should be also rather then trying to help everyone concentrate on actually fixing one persons needs
Scott1979: but anyways as if u listen to me
idletask: Scott1979: you've been given links to documentation, you've been given tips, what else did you want? What's more I don't even have a dual screen setup
Scott1979: all i can say is as if u listening to me
Scott1979: so u trying to help when u dont even kn ow the path just considering that u do
Scott1979: i am greatfull dont get me wrong but the art of teaching is a skill well worht its wheight
Scott1979: and if u can teach like i feal i can on the sujects i know u will breing a crowd of listeners to ur table and they will u respect u more
Scott1979: respect is the highest power and i have none for people that hinder operations
Scott1979: but thankyou anyways
Scott1979: seriuosly thankyou all for helping me use more downloads to look at pages i realy didnt need to
Scott1979: sudo rm -R is my frind
Scott1979: and to my typing i can explain this as i have only recently had a complete wrist rebuild and it is still very unstable
Scott1979: but the more i use it the more movment i get back
cire: thats weird: I am using libdrm, mesa and glx from master, kernel is drm-next. Everything works fine until 2.6.32 (RCs are fine too). Screen stays black, nothing in logs... have to do REISUB. Any ideas?
cire: s/glx/ddx, and I am using KMS
idletask: Yes
idletask: Firmware
idletask: Well, that's my first guess
idletask: I use the drm-radeon-testing branch though
idletask: And maybe, after all, this is not the issue
idletask: What chipset is it?
idletask: The IRQ firmware is only needed for r6xx and up
idletask: What is kdrive?
maligor: a minimal build I think
Ghworg: Bunch of Windows build fixes in mesa git log, do people really use mesa on Windows?
adamk_: Apparently, yes.
Reilithion: Hello Everyone
Reilithion: idletask: Did you get Planeshift installed?
soreau: adamk: I think when I asked that I was told that the windows stuff was for VM's
idletask: Reilithion: not yet
Reilithion: idletask: I'll be interested in hearing about it when you do.
idletask: Reilithion: it will take some time, I have a crappy bandwidth
chithead: idletask: did downgrading to libdrm 2.4.16 help or not?
idletask: Uh, I haven't tried yet... Will do it right now
idletask: Will I need to remerge mesa afterwards?
evil_core: s
soreau: as well as the ddx
idletask: Well, I'll retry without
idletask: If that doesn't work, I'll remerge
idletask: brb
idletask: Re
idletask: chithead: for some reasons, gentoo's x11 overlay doesn't want libdrm < 2.4.17 with git's mesa
idletask: So, I run 2.4.17
idletask: And no, the bug is still there
idletask: If I want to bisect, I'll have to compile libdrm by hand I believe
idletask: chithead: I can do that if you're interested
idletask: wonders whether gentoo's x11 overlay is right about the libdrm dependency
BioTube: it is
BioTube: 2.4.17 changed the radeon api
idletask: Hmm, OK, so I'd need to revert to an old mesa as well
idletask: The bug only appeared recently though
idletask: So I'd be inclined to think the bug is in mesa itself
idletask: BioTube: in case you're interested, I have noticed the bug with one of my OpenGL screensavers
idletask: After the patterns were drawn, they would "imprint" on the screen, and eventually the whole screen becomes white
idletask: The bug is still there
idletask: chithead told me to try crack-attack, and it does have rendering bugs too
idletask: Some other OpenGL screensavers don't have this problem though
idletask: Well, brb, rebooting with git's libdrm again
idletask: Re
idletask: OK, so the nVidia proprietary driver does not have the "save" bug at all with Dungeon Siege
idletask: The ATI proprietary driver does...
idletask: And radeon does too
xming: idletask: what does your package.keyword file look like?
idletask: .keywords, you mean?
xming: x11 overlay ebuilds are hard and testing masked
idletask: http://paste.pocoo.org/show/161364/ <-- here
idletask: So, you need to put them in packages.unmask
idletask: package.unmask, sorry
xming: yes I know that that will unmask but not keywords them to ~amd64
idletask: http://paste.pocoo.org/show/161366/
idletask: Oh, you use stable?
idletask: I use unstable
xming: well I am on funtoo stable
idletask: funtoo? I don't know that variant :p
xming: it's the project that drobbins started after he was kick out of gentoo :p
idle_task: Re :(
xming: idletask: could you paste output of this for me
xming: qlist -Iv |egrep "mesa|xf86-video-ati|libdrm"
idle_task: xming: garbled, but I can tell you they're all 9999
xming: this is mine
xming: http://pastebin.com/d5d2d6c03
idle_task: http://paste.pocoo.org/show/161371/ <-- here
xming: yeah then I need to keyword a lot of pkg, autounmask to the recue :D
xming: ty
idle_task: xming: you should first emerge libdrm, then mesa, then xorg, then rebuild $(qlist -I -C x11-drivers/)
idle_task: I didn't know you could autounmask
xming: this (kms + soft rendering) is actually pretty good now :D but I can't play ut2004 with this :(
xming: emerge -av autounmask :D
idle_task: I don't have ut2004, so I can't tell
Reilithion: I figured out my problem with Planeshift. It seems it was using some unsupported texture compression.
idletask: Isn't that something like s3tc?
Reilithion: eh?
Pallokala: hey, where should I go to complain when 2.6.33-rc1/2 crash whenever I try to enable irq-handling during firmware loading?
chithead: Pallokala: best report on bugzilla or in the mailing list, attach kernel log if you are able to obtain
idle_task: chithead: would you be interested if I bisected mesa to see when the bug appeared for the screensavers/crack attack?
xming: ut2k4 with software rendering won't be good :D
chithead: idle_task: you think it is a mesa issue? if you are able to identity the commit via bisect, then best report a bug on bugs.freedesktop.org
idle_task: chithead: I think so, because when I emerged it after the 20th but before the 30th of December, there wasn't a problem - and commits in libdrm since then don't affect radeon
idle_task: I haven't changed kernel since
xming: idle_task: 17:50 < idle_task> xming: you should first emerge libdrm, then mesa, then xorg, then rebuild $(qlist -I -C x11-drivers/)
Achil: HellOo
xming: xorg == xrog-server?
idle_task: xming: yes
chithead: xorg, xserver, xserver-xorg, xorg-server, distros invent funny names for it
xming: X
idle_task: Well, "the thing that allows GUIs to be displayed", meh
xming: idle_task: which xorg-server are you using? 1.7.4 RC 2 here
xming: SVGAlib does GUI too :p
idle_task: xming: same here - "X.Org X Server 1.7.3.902 (1.7.4 RC 2)"
idle_task: Hmm
idle_task: What role does pixman play?
Achil: i need to join french chanel please help me
idle_task: xming: but you must remerge xorg-server when you change mesa
xming: ok will do
xming: French? loads of Frnech here I see
idle_task: And you must also remerge the ddx
xming: ddx is the ati driver, isn't it?
idle_task: Yes
xming: n'est-ce pas?
xming: :p
Achil: yess french chanel
Achil: sorry my english is not good
idle_task: Achil: I am French but this is an English speaking channel, as you can see
Achil: Ok please i need information idle_task help me
xming: feeds Achil with all the info from google.fr :p
Achil: LOL
xming: Achil: just ask your question
Achil: Ok thx
idletask: brb
xming: bye, restarting X
xming_: idle_task: that froze my kb and corrupted display
idle_task: xming: did you remerge all drivers?
xming: idle_task: aha, all nope, just ddx, not input drivers
xming: should I?
idletask: Grrr, the net connection flicked out again
idletask: Hmm, bisecting is not going to be easy
idletask: Uh, the ddx is not even linked to mesa
idletask: So, how is the device dependent dri module from mesa loaded then?
soreau: idletask: What are you bisecting for?
idletask: soreau: the problem I see with some opengl screensavers which now let traces behind them
idletask: And which makes that eventually the whole screen turns white :p
soreau: If it used to work, it shouldnt be too difficult to bisect mesa
idletask: Well, it is when you let portage do the stuff for you
idletask: I don't want to damage my system too much :p
soreau: You wont
soreau: Just install to non standard prefix then set LIBGL_DRIVERS_DIR=$PREFIX/lib/dri
soreau: restart xscreensaver with that and test
idletask: soreau: err, you mean you can use two dri modules at the same time??
soreau: yea
idletask: But basically, what you say is that X (or Mesa?) obeys this environment variable?
soreau: LIBGL has to do with mesa
soreau: This is what I do to test gallium
idletask: I've had no idea that this could work at all
soreau: Install gallium to nonstandard prefix, set LIBGL_DRIVERS_DIR=$PREFIX/lib/dri in /etc/environment and .bashrc, then anything that isnt gallium friendly yet, I set the env var to /usr/lib/dri :)
anom01y: Hi, I have a ATI Radeon mobility 7000 igp
soreau: this way, gallium is default but if it breaks or anything, classic mesa is still installed to where it should be
anom01y: not sure if I need the open sources or the binary driver
anom01y: but I was told to downgrade to ubuntu 8.10
soreau: anom01y: The closed source drivers do not support that card so your only option is the open radeon driver
idletask: anom01y: by coming to this channel, you are bound to know what people here will answer you (see topic) :p
anom01y: soreau, what if I downgraded to ubuntu 8.10
anom01y: beacuse the driver that works by default after installing ubuntu seems really choppy
soreau: you can downgrade to whatever you want but we dont support fglrx here
soreau: which version of ubuntu do you currently have installed?
anom01y: 8.10
soreau: ...
anom01y: ubuntu 8.10
idletask: anom01y: see what you said just above
anom01y: yeah I see
anom01y: I just finished downgrading this morning,
anom01y: it had 9.10 on it prior to the downgrade
anom01y: but now after the downgrade the video quality is still choppy (especially flash)
soreau: If you want to use the open driver, you should use the very latest version of your distro possible. However if you want to use the proprietary fglrx driver, you should ask in #ati
anom01y: soreau, ok thanks
soreau: I highly recommend you use the open driver because fglrx does not support your card at all
idletask: Hmm
idletask: I don't have an r600_dri built
soreau: how did you configure?
idletask: Argh
idletask: Badly
idletask: My fault
soreau: :)
idletask: Why is there a radeon_dri.so at all? Is it for older cards or something? I have r200, r300 and r600 installed with mesa
idletask: Or is radeon_dri.so just a wrapper?
BioTube: I think that's either gallium or the one for the most ancient cards
idletask: BioTube: I --disable-gallium
BioTube: then it's the ancient card driver
idletask: is stumped that more than one mesa dri module can be used at the same time
soreau: idletask: I believe you want to rm r600_dri.so then rename radeon_dri.so to r600_dri.so
MostAwesomeDude: Um.
idletask: soreau: err, no, I don't think so :p
MostAwesomeDude: No, not if he didn't do Gallium.
BioTube: idletask: each instance of the mesa library can dlopen() its own instance of the dri module
soreau: Ok, that is gallium specific then
idletask: BioTube: but can't that wreak havoc with the video card somewhat?
MostAwesomeDude: The actual name of the DRI driver comes from the DDX. Long stupid story.
idletask: I guess the answer is probably no, but that's just so strange
BioTube: idletask: no, that's how mesa talks to the kernel module which talks to the card
MostAwesomeDude: No, the kernel will discard obviously stupid CS, like r100 CS on an r600.
idletask: BioTube: well, that makes mesa easier to bisect by far
soreau: idletask: That's what I was trying to tell you :)
idletask: soreau: indeed
idletask: I have understood what you said, I just didn't believe my ears :p
idletask: Well
idletask: My eyes
soreau: and with specific configure for your instance and ccache, it's even easier
soreau: MostAwesomeDude: So can you explain again what is the difference between radeon_dri.so and r*00_dri.so?
MostAwesomeDude: soreau: Well, r100's DRI name, for backwards compat, is radeon, not r100.
MostAwesomeDude: So its DRI module is radeon_dri.so.
MostAwesomeDude: But that's also the name of the Gallium DRI module. :T
idletask: soreau: I do use ccache
soreau: idletask: Well you have to set it if compiling manually
idletask: MostAwesomeDude: they should have named it radeong_dri.so, just like there is r300g_dri.so
idletask: soreau: I export PATH :p
soreau: idletask: I don't think setting PATH will set CC :P
idletask: Well, yes, if, like gentoo, you happen to set wrappers over gcc which use distcc
soreau: anyway, did you find a good commit yet?
idletask: Err no, I meant ccache
idletask: No, the one I picked doesn't compile, meh
MostAwesomeDude: idletask: I'm "they", and no.
soreau: lol
MostAwesomeDude: And there's not r300g_dri.so. (Or, there wasn't when I went on vacation.)
soreau: MostAwesomeDude: Did you commit the last patch set eosie sent?
MostAwesomeDude: soreau: Not yet. I haven't headed home yet.
soreau: MostAwesomeDude: Since when have you been away? oO
MostAwesomeDude: I thought that I could get at least some of them ready, but my home box isn't picking up the phone.
soreau: heh
MostAwesomeDude: soreau: This entire week.
soreau: MostAwesomeDude: Ahh, ok
soreau: wonders if eosie will finally get an access account
MostAwesomeDude: Maybe. His patches are hit-and-miss sometimes. OTOH, having airlied blame him directly for bad code might not be a bad thing; the same thing happened to me and zhasha.
soreau: MostAwesomeDude: btw, is the st supposed to work? I tried it once here and got swrast but instead of compiz WSOD, I got super slow mode compiz with all textures in tact
MostAwesomeDude: should probably start using Reviewed-by for patches he actually reads
MostAwesomeDude: soreau: Not sure, actually. I think it uses the wrong DRI name.
soreau: MostAwesomeDude: Can that be remedied? :)
MostAwesomeDude: soreau: Yeah.
soreau: MostAwesomeDude: Well the time before I tried with modesetting, X just segfaulted. So this is definitely a set up
soreau: s/set/step
soreau: and fwiw, I applied all eosie's patches but didn't notice a significant change at all except maybe an overall performance increase
soreau: of course I use my senses to judge performance, maybe I should start logging FPS of glxgears like some of the others lol
MostAwesomeDude: OA would be better.
soreau: I recently installed WOP to get an idea and it runs flawlessly afaict
soreau: et is looking better and better, runs flawlessly with r300c but been having trouble from the start with r300g
idletask: WOP == World Of Padman?
soreau: yea
idletask: Runs well here too
soreau: MostAwesomeDude: The gears resize window bug is still present too
soreau: I still dont know how to play wop, but the graphics are awesome ;D
idletask: Hmm, as chithead said, it seems to be a bug with libdrm, not mesa :( But plain libdrm 2.4.17 doesn't work either
soreau: idletask: I would downgrade libdrm first, then find a mesa commit that compiles and go from there
idletask: soreau: I try to start from b065aec840b4c07eb1db156c1100ef4004634549 but it does not compile, unfortunately
idletask: YESS
idletask: I HAVE found a good commit
idletask: ddd9729bc37f4b1098ef940da6e723743db3ded8 is OK
idletask: And crack-attack also renders fine!
soreau: Is it with old libdrm too?
idletask: No
soreau: Then it should only take a few passes I imagine, granted they compile ;)
idletask: soreau: I just git log'ged src/mesa/drivers/dri/r600
idletask: There weren't that many commits, so I took the closest possible to dec 21
idletask: OK, so, git tag good
eosie: MostAwesomeDude: hit-and-miss? thanks for that
MostAwesomeDude: eosie: Not all of your patches have been perfect, and airlied's bugged me on at least one of them.
MostAwesomeDude: If it makes you feel better, airlied's bugged me on dozens upon dozens of mine. :3
MostAwesomeDude: I did read through some of the patches. They look good.
soreau: sure, look at the bright side.. it's constructive criticism by nature :)
soreau: MostAwesomeDude: Do you know what it will take to get compiz alpha blur effect working on r3xx?
dim3000: my fonts for everything are messed up with some letters draw inproper and such, after adding a radeon.modeset=1 param. Any advice?
MostAwesomeDude: soreau: No idea.
anom01y: soreau, ok, well I don't think there is any driver for me to use except the open source one
anom01y: except for this Foss driver
anom01y: I just need to be able to speen up the driver a little so I can watch flash video or scroll through web pages without having the choppyness
anom01y: two questions: have you ever heard of the Foss driver ?, b) is there a way to speed up the open source driver ?
soreau: anom01y: 1) FOSS = free open source software 2) When you were on 9.10 did you try booting with
soreau: radeon.modeset=1
eosie: MostAwesomeDude: I know. I guess there were 3 patches of mine which were reverted/not approved and considered bad. However it looked like you meant there were more of them, and now it appears I misunderstood you.
MostAwesomeDude: eosie: Nope, I just meant that everybody has a couple bad patches when they start out.
MostAwesomeDude: And even the experienced guys sometimes talk about whether their patches are good.
idletask: verifies his script for git bisect run
idletask: anom01y: if you use ubuntu, consider the xorg-edgers repository
idletask: Heh
idletask: I love git bisect run
idletask: Make your script, verify it, git bisect run and you're good to go
idletask: 3728673bd1b974e54858fbab6ff62d3607b0d3f0 is the first bad commit
idletask: OK, so I've traced back my render bug
dim3000: my fonts for everything are messed up with some letters draw inproper and such, after adding a radeon.modeset=1 param. Any advice?
anom01y: soreau, sorry I just got back in
idletask: anom01y: if you use ubuntu, consider the xorg-edgers repository
idletask: Hmmm
idletask: MostAwesomeDude?
anom01y: thanks for the reply, so I am going to get this computer back to ubuntu 9.10
anom01y: (from 8.10), and then I will try those tweaks, and perhaps that xorg-edgers repo
anom01y: soreau, how do I boot with that option you mentioned ?
anom01y: just type it in at bootup ?
idletask: chithead: it turns out that this is indeed a bug in mesa, and not even in the dri driver itself
MostAwesomeDude: idletask: Yes?
idletask: MostAwesomeDude: 3728673bd1b974e54858fbab6ff62d3607b0d3f0 causes a bug for me, but just looking at the diff of it, I think mesa could do with some bit manipulation macros
Ford: Hi all!
idletask: And given the patch, I'm surprised that I'm the only one seeing the bug at all
Ford: I've got an rs690, running Gentoo on amd64, can anyone point me at a URL or something that might tell me what I need to do to play with the r300g driver?
idletask: But then in the diff comment the author says that some drivers test for Color.BlendEnabled
Ford: updated his box, and wishes he hadn't...
idletask: And r600_dri is such a driver
idletask: (specifically, src/mesa/drivers/dri/r600/r700_state.c:454)
MostAwesomeDude: idletask: Send an email to the list.
MostAwesomeDude: idletask: Yeah, r600 should probably grow drawbuffers and drawbuffers2 support soon.
chithead: Ford: set USE="gallium" and VIDEO_CARDS="radeon" and install mesa from ~arch
MostAwesomeDude: Ford: Don't bother; it's broken on RS690 for now.
Ford: craptacular.
MostAwesomeDude: Sorry.
Ford: I really am starting to wish I'd just left it with x 1.5.3, kernel 2.6.28, and fglrx :|
MostAwesomeDude: If it were trivial to fix, I'd already have it fixed.
MostAwesomeDude: Just use classic r300. It'll be a lot more performant.
Ford: I appreciate the work you guys are doing on this but there's this whole lapse in usability :\
Ford: Windoze has invaded my laptop again, because I can't game under wine anymore.....
Ford: ...that's pretty much the only reason I boot the Wintendo :P
Ford: MostAwesomeDude, could you take a shot in the dark, at when it might be working on rs690?
Ford: looks
Ford: ...hey wait a minute, I've *got* USE="gallium" and VIDEO_CARDS="radeon"
Ford: I mean, it halfway works, but games / 3d performance are still... well...
Ford: ...well, desktop comp works (kwin 4.3.4) anyway. I mean, what's good so far, is actually rather impressive :)
Ford: chithead, by ~arch, are you talking for e.g., mesa-7.7-r1, or mesa-9999 (overlay)?
chithead: Ford: mesa-7.7 comes will gallium support already. if you want to test the latest stuff you can use the live -9999 ebuild
Ford: hmm...
idletask: MostAwesomeDude: mesa-devel?
Ford: err.... ok, the gallium support is in the radeon_dri or r300_dri?
Ford: scratches his head
idletask: Well, mesa3d-dev, actually
Ford: note to self: NEVER AGAIN will I buy a laptop without MXM (or comparable) upgradability
amarsh04: what is MXM? (mini pci-express card for graphics?)
Ford: basically
Ford: Mobile pcieXpress Module
Ford: ATi and nV both make cards for it
dmb: Ford, what laptops has those?
dmb: dells have a weird one
Ford: some laptops have it, but it's kind of hard to figure out.
Ford: but unfortunately, my Toshiba Satellite A215 has no such capabilities :(
taiu: eosie: great work
Ford: do I need a really-really-new libdrm to take advantage of the newer radeon driver stuff?
Ford: I presently have libdrm-2.4.17...
Ghworg: 2.4.17 is the latest, there have been a handful of commits past that and nothing important to radeon
Ford: right on
idletask: MostAwesomeDude: sent
taiu: idletask: on intel i get the same effect
Ghworg: Most fixes are trivial, it's finding where the code needs to be trivially fixed that's hard
xming_: finally KMS and DRI2 running
xming_: what does this mean?
xming_: do_wait: drmWaitVBlank returned -1, IRQs don't seem to be working correctly.
Ghworg: xming_: You don't have a kernel with IRQ code in it
xming_: running vanilla 2.6.32 here, does it have it?
xming_: well is it bad not having it?
Ghworg: 2.6.33rc* needed (or one of the drm-* branches in the radeon trees)
xming_: ok thanks
Ghworg: You won't get vsync in opengl apps without it, but that's about it currently I believe
xming_: what performancee (in opengl) should I expect with my system (3850 AGP, 1600x1050)? I am mostly interested in UT2004.
xming_: any small opengl benchmark to run?
adamk_: ut2004 should be playable, at least with some reasonable settings.
xming_: adamk_: thanks
nemo: Hey guys. have a linux mint 8.0 user in #hedgewars (basically ubuntu 9.10) who has an ATI Xpress 200
nemo: hedgewars is crashing in SDL_SetVideoMode w/ the radeon driver
nemo: looking for the radeon bug system to see if others have run into this. also wondering if there is an easy way for him to try a newer radeon driver build, if that might help
xming_: flicker free and instant vt switch, waaoo. :D
_KAMI_: hi
nemo: hm. he disappeared. oh well
nemo: neeeevermind
nemo: wouldn't be able to get to him even if there was an option
nemo: wanders back to the game
_KAMI_: I tried KMS with my Mobility Radeon 3470 (33RC2 kernel)
_KAMI_: But It frozen
_KAMI_: (Shows black screen and the cursor is jumping line by line
_KAMI_: do you have idea?
eosie: would like to know how to debug hardlocks
adamk_: _KAMI_: Did you compile in support for the framebuffer console? Did you disable support for all other framebuffer drivers?
idletask: eosie: a netconsole at the very least
xming_: now compiz time :D
soreau: eosie: I'd say some really good guessing tactics ;)
_KAMI_: adamk_ I think no I did not
xming_: idletask: I have radeon as kmod and I don't have problems with firmware loading
_KAMI_: I did not compiled
_KAMI_: but used xorg edgers ppa
xming_: y
idle_task: Hmm
idle_task: Planeshift crashes if it uses its own libs
idle_task: Meh
xming_: [drm] Detected VRAM RAM=256M, BAR=256M
xming_: [drm] radeon: 256M of VRAM memory ready
xming_: drm] radeon: 512M of GTT memory ready.
xming_: what do all that mean?
xming_: my card has 512MB, only 256 is dectected?
xming_: and what is GTT memory?
idletask: xming_: kernel limit, on purpose
idletask: It seems that for now, it is unsafe to access more memory without running into serious memory access issues
idletask: As far as I could understand, this is being worked on
xming_: ehum, ok thanks
idletask: Some folk here will be able to give you a more precise view of the problem
xming_: restarting X, bbl
ossman: I'm having problems with what seems to be a hew texture sampler bug
ossman: anyone feeling knowledgeable enough to have insight on that? :)
ossman: I had the shader program output the {s,t} value, which gives 0x7f,0xaf
ossman: but when using that coord to sample, it behaves more like 0x80,0xb0
ossman: I've compared with the software rasteriser, and that behaves like I expect it to
ossman: pokes MostAwesomeDude
ossman: soreau?
ossman: airlied?
dileX: on vacation
mikala: hi,
ossman: dileX, lazy buggers :)
mikala: does an HD5770 is supported by the radeon driver?
dileX: make[5]: Entering directory `/home/sd/src/mesa/mesa/src/gallium/winsys/drm/radeon/dri'
dileX: ../../Makefile.template:126: depend: No such file or directory
anom01y: ok I am back to ubuntu 9.10 latest
BioTube: mikala: no
mikala: ok.
mikala: & neither by the radeonhd i guess?
BioTube: nope
BioTube: work's under way for the radeon driver, though
mikala: well currently is working without xorg.conf locally but @ 1280*1024 :/
idletask: Well, it's a start
mikala: if i start adding my monitor
mikala: it's failing :D
BioTube: well, it's using the VESA driver
mikala: i was supposing that
mikala: but since i read it was using a builtin configuration & it was marked ati i was not sure.
idletask: Well, your log file probably tells somewhere that it had to fallback to using vesa
mikala: idletask: right
mikala: i read too fast.
dileX: thats still the same problem
dileX: make[5]: Entering directory `/home/sd/src/mesa/mesa/src/gallium/winsys/drm/radeon/dri'
dileX: make[5]: *** No rule to make target `../../../../../../src/mesa/libmesagallium.a', needed by `radeon_dri.so'. Stop.
chithead: dileX: when did you last pull from git
dileX: $ grep "git log" -A5 build.log
dileX: + git log --pretty=short -1
dileX: + cat
dileX: commit c978ab017d103e60e7e69b6a7857a70baca29e7b
dileX: Author: José Fonseca
dileX: gallium: Hunt down all references to GALLIUM_AUXILIARY_DIRS.
evil_core: anybody plays bzflag?
evil_core: I get one frem per 10 seconds, isnt that impressing?
idletask: Not me, but I try a lot of games since I test the latest radeon drivers
idletask: Includine "wine" games
evil_core: so what with bzflag?
idletask: Well, I haven't tested it yet
evil_core: its damny simple and small game
evil_core: evn ut2k4 works for me
idletask: evil_core: yes, but the OpenGL API is vast, and not all is implemented yet
idletask: Nexuiz works perfectly here
idletask: Other FPSes do not...
dileX: chithead: I hoped
idletask: chithead: are you on the mesa3d-dev list?
chithead: no
dileX: this should be fixed by latest commits
idletask: OK, just to know
idletask: I have sent a mail two hours ago wrt the rendering problem and it looks like it didn't make it yet :/
idletask: Not the first time it happens with sf-hosted mailing lists but that's a pain
anom01y: idletask, what is xorg-edgers going to do for me ?
chithead: idletask: sf.net public mailing list archives are slow. try with gmane
idletask: anom01y: provide you with very recent builds of all components necessary to run an up-to-date graphics environment
anom01y: idletask, do they have a web site ? I didn't see anything in apt
anom01y: soreau, I use grub, how do you enter boot options ?
idletask: anom01y: no idea, but a quick Google search might tell you everything you need to know
idletask: chithead: indeed, it did make it to gmane
idletask: Strange
anom01y: idletask, xorg-edgers looks like something I would need for newer computers. I am trying to make an older video card less choppy on this old Toshiba Satellite a60
soreau: anom01y: I think 'E' at the grub prompt might work
idletask: anom01y: ah, I thought you had a recent card and wanted 3D support for it...
idletask: I have misread
anom01y: thats ok, thank god I didn't waste too much time with that one
idletask: Sorry :/
idletask: But you could try it out anyway... What's your card again?
anom01y: idletask, its a ATI radeon mobile 7000 igp
anom01y: the laptop is in great shape, but the video rendering is not so good, and I really don;t like using XP nowadays with all the viruses they get
idletask: anom01y: well, this is an rv360 chip and it is fully supported (KMS, DRI2 and 3D) by recent radeon drivers
idletask: So imho xorg-edgers is worth a try
dileX: same with intel-gallium
dileX: make[5]: Entering directory `/home/sd/src/mesa/mesa/src/gallium/winsys/drm/intel/dri'
dileX: mkdir -p ../../../../../../lib/gallium
dileX: make[5]: *** No rule to make target `../../../../../../src/mesa/libmesagallium.a', needed by `i915_dri.so'. Stop.
idletask: What does GART stand for exactly?
MostAwesomeDude: Graphics Aperture Relocation Table IIRC.
Droste1: http://en.wikipedia.org/wiki/Graphics_Address_Remapping_Table
dileX: good its not F...
idletask: Wikipedia says Graphics Address Remapping Table - and that's yet another IOMMU! Ohwell
idletask: Is that the reason for the 256MB VRAM limit in the kernel currently? Because it can't address such IOMMUs safely otherwise?
MostAwesomeDude: No.
MostAwesomeDude: The reason for the limit is because of PCI BARs.
MostAwesomeDude: Right now, we can't get at the rest of VRAM without acceleration, and so we can't really fallback if we need to.
MostAwesomeDude: glisse was working on the correct way to do it. Dunno if he's finished yet.
dileX: seems to be a problem with ccache and parallel-make-jobs
dileX: export CC="ccache gcc"
dileX: #export MAKEFLAGS="-j3"
dileX: ^^ this works
idletask: dileX: eh? I've never had a problem compiling mesa with make -j4
dileX: $ find mesa/ -name libmesa*
dileX: mesa/src/mesa/libmesa.a
dileX: mesa/src/mesa/libmesagallium.a
MostAwesomeDude: You can't build Mesa in parallel.
dileX: seems mesa/src/mesa/libmesagallium.a is not built to be ready for radeon-gallium
dileX: MostAwesomeDude: BTW, happy awesome years
dileX: MostAwesomeDude: why not?
Ford: hey MostAwesomeDude
Ford: Didn't you say rs690 support is broke?
Ford: is confused
Ford: or is that just for r300g?
Ford: eh... I'm not fussin', I'm sayin', mine seems to work pretty well, as long as I'm not tryin' to play games...
Ford: y'all seem to be doing a fairly good job, all things accounted for ;)
idletask: agrees
MostAwesomeDude: dileX: Because it doesn't work.
MostAwesomeDude: Ford: RS690 doesn't work in Gallium. Should work fine with the classic drivers.
dileX: MostAwesomeDude: so only recommended speedup is ccache and alternatives?
MostAwesomeDude: dileX: I don't recommend any speedups.
Ford: ahh...
Ford: MostAwesomeDude, can you make a wild-ass-guess as to when (and if) RS690 will work?
MostAwesomeDude: Ford: Well, it used to. I could look at it when I finally get back home, whenever that is.
MostAwesomeDude: It shouldn't be super-difficult.
dileX: MostAwesomeDude: whats your current configure/autogen.sh line?
MostAwesomeDude: dileX: Couldn't tell you. It's got r600 and swrast, and also --enable-gallium-radeon and --enable-debug.
dileX: MostAwesomeDude: thats mine
dileX: --with-state-trackers= ???
idletask: dileX: why do you clean ccache by hand? ccache has an option for that
dileX: idletask: ah OK
idletask: dileX: and why do you git clone every time??
idletask: Just cd into the repo and git pull
dileX: idletask: no - its unpacking tarball and doing git-pull
Ford: righteous :)
Ford: I'd really like to be able to play at least ET again =D
idletask: Err, that's not what I see
MostAwesomeDude: dileX: Man, you've got a more involved setup than i do.
dileX: idletask: whats the option of ccache to do a clean? dont say RTFM
idletask: dilex: ccache -C
dileX: hmm
dileX: with ccache
dileX: real 6m19.067s
dileX: user 5m31.836s
dileX: sys 0m32.396s
dileX: real 6m19.067suser 5m31.836s
dileX: sys 0m32.396s
dileX: grr
dileX: its around 5-6mins here to build
idletask: dileX: http://paste.pocoo.org/show/161489/ <-- here is the script I used to git bisect my rendering bug. Not even 2 minutes before the dri is ready to go :p ccache and make -j4 are maybe not recommended but they work
dileX: soreau: radeong is starting xorg. but corrupted fonts in rxvt and resizing did freeze X.
dileX: idletask: why "--enable-xcb"?
idletask: dileX: because I use xcb, probably :p
idletask: I don't really know why, but I needed it at some point
dileX: X-C-bindings
dileX: debian-x has no --enable-xcb for mesa set
dileX: anyway, parallel-make-jobs breaks mesa-build
dileX: *can break
soreau: dileX: mesa has an xcb flag?
BioTube: soreau: yes
soreau: BioTube: Actually, there shouldn't have been a question mark there :)
dileX: $ ./configure --help | grep xcb
dileX: --enable-xcb use XCB for GLX [default=disabled]
soreau: I compile with it
idletask: dileX: as to parallel make jobs breaking Mesa builds... this has never been a problem for me here, your problem is probably elsewhere
BioTube: -j4 works for me
dileX: idletask: no. I had -j3 and worked fine
dileX: w/o no problems
idletask: Then keep on using it, it accelerates things so much it's not worth doing without it
idletask: There _are_ some packages that break with it, but they are few
idletask: Argh
idletask: I just made the drm kernel module reset the GPU
idletask: By trying to launch Anno 1701 into wine :(
idletask: With "fence timeouts"
idletask: So, which is the culprit here? (running drm-radeon-testing kernel)
idletask: http://paste.pocoo.org/show/161498/ <-- the kernel messages
idletask: Well, I'm going to try again and see...
Ford: ...was that an ant fart I heard?
BioTube: anybody else getting a segfault with the ddx after upgrading to the new API?
Scott1979: anyone in
dileX: libdrm_radeon API?
BioTube: yes
dileX: works fine here with rv515
idletask: Segfault from? X?
BioTube: I upgraded libdrm, ddx and get a segfault in R600LoadShaders with my rv620
dileX: but ddx is (6.12.99+git20091221.4be451b)
BioTube: at least that's where it seems to be occurring
idletask: Not me, but I get kernel errors (fence errors) when trying to launch Anno 1701 under wine
idletask: Bad
dileX: BioTube: segfault is here with input-hotplug
dileX: Xorg
idletask: The GPU is soft reset, the machine is still alive but I have to collect messages from another machine
BioTube: well, the backtrace includes an unlabled address
idletask: I also bisected my rendering problem btw
BioTube: 0: /usr/bin/X(xorg_backtrace+0x3b) [0x81314cb]
BioTube: 1: /usr/bin/X(xf86SigHandler+0x51) [0x80c1df1]
BioTube: 2: [0xb780b400]
BioTube: 3: /opt/xorg2/lib/xorg/modules/drivers//radeon_drv.so(R600LoadShaders+0x80) [0xb749a980]
idletask: BioTube: when was your last update apart from that one you just did?
BioTube: everything's currently just one commit shy of the API change
idletask: libdrm API change?
BioTube: yes
idletask: Ah
idletask: Here it is "enabled"
idletask: Well, I'm going to try my hand at Anno 1701 again, which will certainly mean I won't see the next messages :p
idletask: Argh
idletask: I get a different error now :/
idletask: But at least no kernel messages
dim3000: Can anyone help me with some video issues?
BioTube: depends on what they are
dim3000: The fonts are drawing improperly sometimes for some letters/symbols
BioTube: card?
dim3000: r100 9Radeon 7200)
idletask: dim3000: I've had this problem some time ago
dim3000: *r100 (Radeon 7200)
dim3000: idletask: Any advice?
idletask: dim3000: it disappeared when I upgraded libdrm/mesa/ddx to master... Although I have other problems now :p
dim3000: As im typing right now for example, some letters are cutoff, etc...
idletask: dim3000: I have noticed also that not all fonts have the problem
idletask: Which kernel are you using?
dim3000: idletask: well actually this problem I have didnt happen before
dim3000: idletask: but now that I added a radeon.modset=1 param. it does
idletask: dim3000: it appeared for me at some point, and then disappeared
dim3000: idletask: the only reason I added that parameter is to have OpenGL support
Droste: dim3000 which kernel version?
dim3000: Droste: 2.6.31-15-generic
Droste: you should try a more recent kernel
dim3000: Droste: any speific recommendation?
Droste: if you know how to compile a kernel on your own, you can try drm-radeon-next from airlied, if not than the most recent thats avaible for your distro
dim3000: ok, thanks I update and check back
anom01y: soreau, that radeon.modeset=1 boot option seems to have helped
fractalis: So I followed the instructions outlined in the wiki for radeonBuildHowTo. However, glxinfo is still reporting Software Rendering instead of using DRI. Any suggestions on debugging why that is the case?
fractalis: dmesg | less
chithead: fractalis: Xorg.0.log might tell
idletask: Hmm
idletask: New mesa commits
idletask: tries and see whether remerging mesa solves his rendering problems
idletask: I'd like to see sauerbraten render correctly :/
idletask: Even with minimal settings, it's a nono
idletask: What is vsync BTW? Sync to what?
BioTube: syncing to the vertical blank of the screen
BioTube: ie, when the monitor redraws
idletask: BioTube: and what is it supposed to avoid? Choppy rendering?
BioTube: screen tearing
dim3000: Droste: updated a bit higher to 2.6.31-16-generic
anom01y: soreau, maybe it isn't helping
dim3000: Droste: I still have the same font problem, should I upgrade even more?
anom01y: anyone here know why flash videos (youtube) draw 100% of the CPU) ? I am using the open source drivers
Droste: dim3000: yes, thats not really higher
anom01y: ati radeon mobile 7000 igp
dim3000: Droste: thats all I have right now (ubuntu derivative)
Droste: at least 2.6.32, which distro do you use?
Droste: dim3000: http://kernel.ubuntu.com/~kernel-ppa/mainline/drm-next/current/
dim3000: Droste: thanks ill try those
Droste: dim3000: if the problems persist, you should also update your libdrm, ddx and mesa (I think there are recent packages for ubuntu)
dim3000: Droste: btw how do I add that link you gave me as a source
dim3000: or should I just download them
Droste: I never used ubuntu... maybe you have to use the parent directory?
BioTube: there's nothing there to form an apt repo with
Droste: just download the kernel :-)
dim3000: Droste: btw while im waiting for the download, the problem is worsening
dim3000: i barely see symbols now
Droste: I had the same problem. for me a new kernel and the current libdrm and ddx helped
dim3000: Droste: seemingly works right now
Droste: good :-)
dim3000: Droste: but we'll see as it progresses
dim3000: Droste: what else did you might I might need to upgrade?
dim3000: Droste: (just in case ;))
idletask: Apart from the DRI modules, does other code from mesa depend on libdrm?
Droste: libdrm and ddx. i think the ubuntu repository is called xorg-edgers where the most recent xorg packages are
idletask: If you emerge libdrm, be careful with mesa! See my message just above
dim3000: ok thanks
idletask: Must restart X, I've just upgraded the whole stuff
Droste: a new mesa is also the xorg-edgers repo :-)
dim3000: Droste: open synaptic, and... here come the upgrades ;)
dim3000: Droste: is there any new in the drivers that you know of and that i can use?
dim3000: Droste: or does my system just improve by itself
Droste: it "improve" by itself :-)
dim3000: ok thanks
dim3000: just upgraded it all
Droste: I think for r100 there is not much new stuff :-)
dim3000: Droste: doesnt matter, btw do I need to restart to see changes?
Droste: at least x, yes
dim3000: Droste:
dim3000: problems
dim3000: Droste: OpenGL renderer string: Software Rasterizer
dim3000: im not rendering anymore
Droste: did you update mesa?
dim3000: mesa-utils yes i think
dim3000: hmm
Droste: I think there is a package called "mesa". do you know how to find out what version is installed?
dim3000: found
dim3000: ibgl1-mesa-dri
dim3000: and ibgl1-mesa-glx
dim3000: no actually mesa
Droste: mesa and mesa-utils should be 7.7 or 7.8
Droste: maybe thats the problem :-)
dim3000: i dont have a "mesa" package though
Droste: not installed or not in the repository?
dim3000: "mesa-utils" is 7.7
dim3000: not in synaptic
Droste: could you give me the url to the xorg-edger repository?
dim3000: wait
dim3000: unless libglu1-mesa is mesa
dim3000: which is also 7.7
Droste: hm no... :-) could you give me the url to the xorg-edger repository? so I can check whats the exact name of the package?
dim3000: one sec
dim3000: two:
dim3000: http://ppa.launchpad.net/xorg-edgers/drivers-only/ubuntu
dim3000: and
dim3000: ttp://ppa.launchpad.net/xorg-edgers/ppa/ubuntu/
dim3000: this was my line when adding it:
dim3000: deb http://ppa.launchpad.net/xorg-edgers/ppa/ubuntu/ karmic main
Droste: there is a package called "mesa" :-)
Droste: maybe you have to refresh the repository infos?
Droste: apt-get update or aptitude update?
dim3000: did that
Droste: aptitude search pkgnames | grep -i mesa
Droste: and this lists no "mesa" package?
Droste: or just try "sudo aptitude install mesa" :-)
dim3000: Droste: search lists nothing
dim3000: Droste: ant install eiher
dim3000: either
dim3000: it seems my Mesa is hiding
Droste: could you pastebin the output of "LIBGL_DEBUG=verbose glxinfo"?
dim3000: here: http://paste.ubuntu.com/350644/
Droste: it never tried loading something different from software rasterizer... weird
Droste: ls -al /usr/lib/dri | grep radeon_dri.so <-- whats the output of this command?
dim3000: maybe i should remove that radeon.modeset=1 now that I updated?
dim3000: Output: -rw-r--r-- 1 root root 2506596 2009-12-30 16:49 radeon_dri.so
Droste: ok... pastebin the output of dmesg | grep -E "(radeon|drm)"
Droste: and xorg.log :-)
dim3000: the first: http://paste.ubuntu.com/350647/
dim3000: (BTW when I did the kernel manually I think I only did the image not the headers, is that a problem?
Droste: no
dim3000: ok wait ill also get my xorg log
dim3000: http://paste.ubuntu.com/350649/
dim3000: thats the log
dim3000: (WW) RADEON(0): Direct rendering disabled
dim3000: this is new
Droste: yes.. and thats the problem :-) but why...
dim3000: nothing else seems to indicate a problem
Droste: did you reboot or just restarted x?
Droste: and do you have a xorg.conf?
dim3000: i rebooted
dim3000: my xorg.conf: http://paste.ubuntu.com/350652/
Droste: mybe try setting "NoAccel" to false and "DRI" to true in the device section
dim3000: ok ill try
dim3000: btw does the kernel param. radeon.modeset=1 have any effect on this?
Droste: could be, you can try with radeon.modeset=0
dim3000: i had it at one to have OpenGL support
dim3000: ill try first the xorg chnage
dim3000: then the parameter change
dim3000: be back
dim3000: Droste: back
Droste: and I found the real problem :-)
dim3000: Droste: changing xorg.cong did nothing radon.modeset=0 worked
dim3000: but I have suckky OpenGL now and DRI instead of DRI2
dim3000: Droste: whats the "real" problem?
Droste: what version of libdrm do you have installed?
dim3000: one sec
dim3000: 2.4.17
dim3000: for libdrm2
dim3000: and also for libdrm-radeon1
Droste: ok... I thought that was the problem.. but that's the right version
dim3000: well radeon.modeset=0 did do it but that brings me to my old problems
Droste: http://paste.ubuntu.com/350649/ <-- the problem is line 375 to 381
Droste: but the versions are right... and with the other kernel it worked, so the kernel module can't be the problem
dim3000: how come without KMS it did work though?
Droste: without kms it doesn't need a recent libdrm
dim3000: ohh
dim3000: yeah the DRIVersionMismatch is bothering me
Droste: this error shows if the lidrm is too old
Droste: but yours isn't too old :-D
Droste: maybe reinstall libdrm
dim3000: ok
dim3000: maybe reinstall mesa also?
Droste: https://wiki.ubuntu.com/X/RadeonKMS
Droste: mesa looks fine
Droste: the second topic "module loading issues" could be an answer too
dim3000: yeah im reading that, thanks for the link
dim3000: so by inserting that line reinstall libdrm and rebooting we should see something
Droste: hopefully :-)
dim3000: ok i think ill try that then
Droste: good luck
dim3000: be back :)
dim3000: Droste: :) :) :)
dim3000: Droste: im back! with my DRI!
dim3000: Droste: thanks a lot dude
Droste: no problem :-)
dim3000: btw before when i was without KMS i always did HyperZ to speed up things
dim3000: with KMS enabled i get this when running apps:
dim3000: DRM version 1.0 too old to support HyperZ, disabling.
Droste: is this an option in the xorg.conf?
dim3000: no it was an option I read to enable somewhere using DRCon tool
dim3000: DRIConf*
Droste: ah never used it, sorry :-)
dim3000: nvm mind then
Droste: perhaps you need a newer version of driconf to :-)
dim3000: i think the new mesa and stuff also speed things up a bit
dim3000: well i gtg but thanks for the help again!
AStorm: hello there
AStorm: or, maybe the question I was going to ask is better placed in another channel :)
anom01y: soreau, maybe it isn't helping (that option you told me radeon.modeset=1).
anom01y: its hard to tell, what exactly is this option used for ?