Home Phoronix Phoronix Forums X.Org Videos From FOSDEM 2008

Radeon IRC Logs For 2009-11-13

Search This Log:


soreau: notice the upper right corner is supposed to be a cashew and the panel has corruption
soreau: Testing now without S-video plugged in at all
soreau: (btw when I tried to right-of the tv-out display was black with cursor)
soreau: brb
soreau: Even with S-video unplugged, corruption persists.
soreau: Testing s/r without S-video now
soreau: More corruption on the kpanel but more or less basically the same as the screenshot I posted
soreau: no change in s/r without S-video plugged.
soreau: Upon resume, things work like audio but screen is totally fubart
soreau: fubar :P
soreau: everything is corrupted including cursor
Nightwulf|work: hi all
dileX: hi Nightwulf|work :-)
Nightwulf|work: hi dileX
glisse: airlied: btw the radeon vga connector and scaler patch are both fixing bugs in fedora12
glisse: well at least i think they are user can only test rpm
dileX: 2.6.32-rc7: "...I have to admit that there's some bigger-than-I-would-have-liked patches to the Radeon KMS driver. Hopefully it's nearing stability now."
twnqx: hm
twnqx: so if an app i try to use crashes, and the stackdump looks like this: 000 - 0x40500C8B: memcpy [001- 003 unidentified] 004 - 0x375544CB: radeonTexSubImage2D 005 - 0x375BD07E: _mesa_TexSubImage2D
twnqx: would you expect a bug in mesa, radeon or the app?
glisse: likely radeon mesa driver
twnqx: ok, my git is already 9 days old, could try an update
glisse: yeah
twnqx: 0x00007fffe8e358b0 in memcpy_texture (ctx=0xf32fd0, dimensions=, dstFormat=, dstAddr=, dstXoffset=,
twnqx: hm
airlied: glisse: I'll do rpms on Monday, when I sober up, had work team buidl + drinks not able to do patches ;-)
glisse: airlied: no problem, i just wanted you to know
glisse: i don't fedora account working yet
glisse: i need to go through the process
twnqx: do i need to restart X after the update?
dileX: dileX testing glsl-pp-rework-2
dileX:
dileX: INFO:
Nightwulf|work: hmm....mesa 7.6 doesn't support shaders, is that right? or could it be a misconfiguration if a GL app claims that to be the case?
dileX: Nightwulf|work: you have an output (e.g. if you start from cli)? or try "LIBGL_DEBUG=verbose $GLapp"
Nightwulf|work: dileX: no, the app complaining is "YoFrankie", a game...and it complains via a message box
glisse: Nightwulf|work: mesa 7.6 with r300/r500 support arb shader not glsl
Nightwulf|work: glisse: the card in question is a rv790 running with the most up to date mesa package for arch linux
dileX: hehe, short trip with glsl-pp-rework-2
dileX: (hmm, maybe need a new libxext-dev)
glisse: Nightwulf|work: dno't remember when r600 support went in but it does support arb shader
glisse: no glsl yet thought
dileX: "Make library headers compatible with old server builds"
Nightwulf|work: glisse: so yofranke probably only supports glsl and therfore moans?
glisse: yeah
glisse: very few apps support the arb asm shader
Nightwulf|work: ah..ok...thanks for clearing that up
Nightwulf|work: glisse: this is the message, yofrankie prints out: http://paste.org/pastebin/view/11781
dileX: Nightwulf|work:
dileX: good
dileX: with libxext6 (2:1.1.1+git20091106.28d16f8-1~dileX+1) glsl-pp-rework-2 mesa GIT-branch is compileable
Nightwulf|work: dileX: that means my mesa is "too old"? ;-)
dileX: Nightwulf|work: unfortunately
dileX: Nightwulf|work: which distro do you use?
Nightwulf|work: dileX: archlinux
Nightwulf|work: my glxinfo says: OpenGL version string: 1.4 Mesa 7.6
dileX: Nightwulf|work: any, here debian/sid. but you can check with glxinfo your supported GL extensions.
Nightwulf|work: dileX: the only extensions with GLX_ARB at the beginning are GLX_ARB_multisample and GLX_ARB_get_proc_address
Nightwulf|work: dileX: is there a date when mesa 7.7 should arrive in stable?
chithead: first mesa 7.7 would have to be released
Nightwulf|work: chithead: sure..but archlinux is quite fast in releasing stable packets after the project did the release
dileX: Nightwulf|work: hmm, remember Brian Paul gave some status info on mesa-7.7, look into dri-devel and mesa3d-dev mailing-list archives.
Nightwulf|work: dileX: thanks!
dileX: Nightwulf|work: and .
dileX: Nightwulf|work: sorry, writing a bug-report.
Nightwulf|work: np
dileX: Nightwulf|work: you could try latest fedora-12 nightly-build ISO-image for testing?
dileX: $ openarena +exec anholt 2>&1 | egrep -e '[0-9]+ frames'
dileX: 840 frames 43.9 seconds 19.1 fps 9.0/52.3/157.0/26.1 ms
dileX: hmm, thats not a performance boost (but me doesnt know if oa benefits from GLSL)
Nightwulf|work: dileX: i can do..but not today or on weekend...earliest date would be monday
new2life: hi all. I'm using debian squeeze. Just dropped by to see if anyone has gotten 3d accel to work in squeeze with their r6/7xx card
new2life: I've got an hd4805
new2life: *4850
new2life: is it possible to use a debian package, or must I follow radeonhd wiki on building it myself. I was hoping for a way to generate a squeeze package, and avoid using unstable repos
new2life: (I've got drm next kernel installed fine
adamk: I don't use Debian, but I've not heard of any repos that contain the necessary packages.
adamk: So I think you'll have to follow the wiki in this channel.
adamk: (Not the radeonhd wiki).
new2life: yep. was afraid of that. I am aware of some ubuntu repos, but I'd like to steer clear of those
new2life: adamk: yeah.
new2life: adamk: regarding /configure --prefix=/opt/xorg --with-dri-drivers= step, I'd be after just r600, right?
new2life: for my hd4850
adamk: Yep.
new2life: unnaturally slow git clone of mesa
new2life: adamk: cool thanks man
adamk: mesa is about 450 megs, iirc :-)
Nightwulf|work: adamk: archlinux repositories (stable) contain all necessary packages...only drm-next is missing
new2life: adamk: christ
new2life: bandwidth-limited here
new2life: i was fairly sure that debian's radeonhd driver (1.2.5) supported xv for r600 chipsets
new2life: I'm only taking the git- route because I haven't the foggiest how to use backports from unstable. considering my bandwidth, I'm gonna have to learn :/
gimzo: is it normal that piglit readPixSanity fails on r400 mesa 7.7-devel (built a week or two ago from this howto) ?
hoo-hah: hi guys. if I'm going to use my distros radeonhd driver (1.2.5), can my kernel be > 2.6.30 as mentioned in the wiki?
hoo-hah: I don't need to enable staging drivers, right? but I will need to enable radeon dri?
Zajec: hoo-hah: it's radeon driver
Zajec: hoo-hah: and you didn't tell what you want to achieve
Zajec: EXA and Xv? 3D?
Zajec: KMS?
hoo-hah: xv
hoo-hah: exa and xv
Zajec: hoo-hah: for EXA and Xv you need 2.6.30 or newer, without any special options
Zajec: just standard 2.6.30 or newer
hoo-hah: Zajec: but there would be no probs in using newer kernel than 2.6.30?
hoo-hah: Zajec: okay thanks
Zajec: radeonhd 1.2.5 by default doesn't enable EXA or Xv
Zajec: you have to use xorg.conf options
Zajec: or use 1.3.0 which enables EXA and Xv by default
Zajec: hoo-hah: 2.6.30, 2.6.31, 2.6.32-rcX, all are fine
Zajec: hoo-hah: for 1.2.5 you need Option "DRI" and AccelMethod "EXA"
Zajec: or just upgrade to 1.3.0 and ignore xorg.conf options
Zajec: afk
hoo-hah: I only need CONFIG_DRM_RADEON right?
kdekorte: hoo-hah, I think that for XV you need to use the radeon driver not the radeonhd driver, but I could be mistaken. Also depending on the card you have, you may need more up to date components.
adamk: radeonhd supports Xv.
soreau: I think radeonhd is just confusing to users. It's name obviously implies it's specifically for HD cards
twnqx: both have working xv anyway
twnqx: the only reason i switched is the support for displayport >_>
glisse: radeonhd support displayport ?
ferr: radeonhd driver?
glisse: yup
ferr: #radeonhd ?
glisse: i know the hw
ferr: u're on wrong channel:)
glisse: ferr: i am just asking to twnqx as he was talking about that
twnqx: no, radeon does
twnqx: that's why i switched TO radeon :P
glisse: twnqx: radeon works with displayport ?
twnqx: mostly
glisse: you get picture ?
twnqx: and most of time.
twnqx: one i fix up the modelines, yes
twnqx: it always wants to give me 1920x540 >_>
twnqx: so i run a custom 1920x1080@24 modeline, then it works
glisse: sounds like bit shifting somewhere
twnqx: no, interlaced mode failures i guess
twnqx: or broken edid
glisse: anyway i missed the displayport commit
glisse: i think we have issue on displayport to properly get edid
glisse: but i am just not uptodate on this displayport stuff
agd5f: it was a bug in parsing interlaces modes in edids. fixed in xserver
twnqx: i only run 1.7.1
twnqx: not git there :P
agd5f: tvs tend to list field size rather than frame size
twnqx: can be, but this projector supports 24p/50p/60i
twnqx: and with progressive modes, framesize == fieldsize, right?
frmbl: hi all
frmbl: are irqs supposed to work on RS880 ? they are not on mine, kms or not
jcristau: frmbl: not yet, afaik
frmbl: everything is working though, but 3D is slow
Ghworg: IRQ support is waiting approval from amd to get released, code is written just needs to be reviewed
frmbl: ok, thanks
skatteola: Hello. Running "x11perf -wdcircle100 -time 1 -repeat 1 -rop GXxor" crashes my X.. is this a known problem or something on my end? This has been the case for a while, but sadly I never got around to reporting it.
skatteola: Running xorg-server 1.7.1 and xf86-video-ati from git at the moment.
muep_: now that we have KMS, I wonder how much is missing to allow running opengl apps directly in a Linux framebuffer, without X
legume: skatteola: Crashes for me as well using git xserver (though not quite up to date)
devh: skatteola: just tried it and here it crashes too
skatteola: Not just me then, I guess I'll try reporting a bug
skatteola: I reported it at https://bugs.freedesktop.org/show_bug.cgi?id=25078
legume: glisse: Is there supposed to be a way to enable tv-out after your patch on the dri list. (if there is what is it - I have failed so far)
pixelhaze: Hi, ATi lists the X1100 chipset to have pixel shaders and vertex shaders. From my searches I have found that there are 2 pixels shaders and 2 vertex shaders on this chipset.
pixelhaze: Does this mean that HW TCL can be enabled ?
pixelhaze: or SW TCL is the only option?
zhasha: if there are vertex shaders, you can have HW TCL
pixelhaze: ok
glisse: legume: xrandr add move 800x600 to s-video
glisse: mode
pixelhaze: zhasha: Ispci reports that I have ATI Technologies Inc RS482 [Radeon Xpress 200M]
pixelhaze: However, for that family the screen->chip_flags is not set with RADEON_CHIPSET_TCL flag
pixelhaze: Any ideas?
zhasha: no vertex shaders on that card
pixelhaze: oh...I am a bit confused ...so the Xpress 1100 series has vertex shader ...as per my laptop specs...but not as per lspci
zhasha: the Xpress 1100 doesn't have vertex shaders
zhasha: http://en.wikipedia.org/wiki/Comparison_of_AMD_graphics_processing_units#Radeon_R300_PCIe_.28X3xx.2C_X5xx.2C_X6xx.29_series
pixelhaze: Ok, I read it on the TomsHardware ..sure ...will follow the wiki
legume: glisse: Doesn't work - still shows as disconnected. (This is DIN on AVIVO card using 1024x768)
pixelhaze: this was my source http://www.tomshardware.com/forum/211077-33-xpress-1100-nvidia-6150 .... :-(
soreau: legume: Can you try to force load detection on/
legume: soreau: How do you do that with kms?
glisse: legume: pastebin output of xrandr
legume: glisse: http://www.pastebin.ca/1670207
pixelhaze: zhasha: http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118_14603_14624%5E14627,00.html states that I have Vertex Shader v2.0 and Pixel Shader v2.0 !!!! Strange !
zhasha: pixelhaze: magic of software processing
zhasha: it has no vertex shaders on board. the driver can easily emulate them
pixelhaze: ok
pixelhaze: Thanks for clearing my doubts.
zhasha: it has 2 fragment shaders, 2 texture units and 2 render target units on board though
legume: glisse: I know it's cloning 1 but --right-of doesn't help.
agd5f: actually 4 render targets
ksc654: I have Radeo Mobility 9600. Just upgraded to Ubuntu 9.10. I've noticed when I reboot its 50/50 chance dri will be enabled. Dri on this reboot, software rendering on the next. Any idea where to start troubleshooting this?
glisse: legume: what patch are you talking about ? did it worked before ?
pixelhaze: zhasha ...where can I find such information ... would like to know more about my chipset.
adamk: ksc654: I'd start by checking /var/log/Xorg.0.log from one of the sessions where DRI is not enabled.
glisse: legume: http://forum.ubuntu-fr.org/viewtopic.php?id=352361
agd5f: pixelhaze: wikipedia
glisse: list of command that should do the trick
pixelhaze: :-)
legume: glisse: Yes it worked before - the patch on the dri list - http://article.gmane.org/gmane.comp.video.dri.devel/40015
soreau: legume: xrandr --output S-video --set load_detection 1
pixelhaze: agd5f: Found it. Thanks.
glisse: legume: this patch is NAK
glisse: won't get in
legume: soreau: Will try - but it already reports load detection as 1
glisse: i have another on which disable just on rs480
soreau: legume: It shows connected?
legume: glisse: OK np then - just thought I would test when I saw it.
zhasha: pixelhaze: I linked you to wikipedia
legume: soreau: No - but it's irrelevant now as it's not going in - it works normally for me, (well ntsc is hardcoded - but I can change that) I just didn't want it to regress.
[Enrico]: mhm i see that even without KMS i get (II) RADEON(0): Color tiling disabled . should i enable it manully in xorg.conf? and btw is it usefull ?
Ghworg: [Enrico]: (WW) RADEON(0): Color tiling is not yet supported on R600/R700
pixelhaze: zhasha: Thanks. I am refering it. Unfortunately I have been referring to http://www.notebookcheck.net/ATI-Radeon-Xpress-200M.2175.0.html which mentioned 4 / 2 Pixel- / Vertexshader .....hence the confusion.
[Enrico]: Ghworg: doesn't appear in my xorg.0.log
Ghworg: [Enrico]: Guess we are running different versions then, but that is the reason nonetheless
[Enrico]: Ghworg: mhm i will try to enable it in xorg.conf and let's see
legume: glisse: Re. the ubuntu link - that won't work on AVIVO - I just get X Error of failed request: BadName (named color or font does not exist)
soreau: airlied: With latest updates using drm-next and all other latest components from yesterday, tv-out works in clone mode but when using --right-of the tv goes black and only displays the mouse cursor, even if I drag a window to it
soreau: Which component would be most suspect so I might bisect?
soreau: I wouldn't want to waste time blindly bisecting the wrong component
legume: glisse: On the load detect or tv_standard - the latter is hardcoded AFAICT anf load detect was 1 already - I have a bug open already about tv_standard
glisse: legume: jus try the addmode setmode command
glisse: ignore the rest
legume: glisse: I already did that with my 1024x768 mode - but I tried 800x600 and it still won't work. http://www.pastebin.ca/1670226
Vinky: if I wanna try kms on r700 do I need some branch of libdrm and the driver?
biotube: Vinky: it should all be in their respective masters
Vinky: ok trying again
[Enrico]: likes ForceLowPowerMode eheh
[Enrico]: btw what's the status of powersave in kms? here http://www.x.org/wiki/RadeonFeature is says MOSTLY for all but doesn't seems it
[Enrico]: has missunderstood he thinks
roysjosh: glisse, woohoooo! I got it working (kind of). re: bug 537140, rs200 blank screen. I have also been debugging suspend and resume via ssh (which has never worked on that laptop, until now). I added "acpi_sleep=s3_bios" to kernel args and after a suspend/resume cycle I can see the screen!
soreau: Can someone please verify this happens to them too or any idea how to fix it? after running compiz-manager & in a terminal to start compiz, I get spammed with these messages when and it continues when compiz effects happen http://pastebin.com/m52748572
agd5f: [Enrico]: no power options with kms yet
lordheavy: wow xserver (drm) crash with kernel 2.6.32-rc7 (with kms) on a HD4200 : http://pastebin.fr/5994
lordheavy: when trying to activate openGL effects with kde4 :-p
[Enrico]: agd5f: thank you :D
agd5f: lordheavy: need a newer mesa
lordheavy: ok
lordheavy: thks
agd5f: lordheavy: 7.6 branch or master
lordheavy: got mesa 7.6 .....
lordheavy: but will try git
agd5f: lordheavy: the 7.6 branch has ongoing fixes
lordheavy: so they fixes should be in mesa 7.6.1/2
lordheavy: *the*
agd5f: lordheavy: yes
lordheavy: ok thanks
quintela: hi
quintela: any hope getting an FireGL V3100 to work at 2048x1152 using VGA?
quintela: 2048x1152 already works well with DVI cable, but I wanted to use two monitors
soreau: adamk_: Perhaps you can confirm or deny this happens assuming you have a box available with latest drm-next and user components? I get spammed with these messages when and it continues when compiz effects happen http://pastebin.com/m52748572
adamk_: soreau: Sorry, all my radeon machines are on FreeBSD at the moment.
adamk_:
adamk_: In fact, they are at 2.6.31.5 at the moment sans drm-next.
soreau: adamk_: Ok thanks
soreau: This just started happening about a week ago and it's bugging the crap out of me
soreau: Even if I start compiz from say Alt+F2 run dialog, that terminal where I initially ran compiz will continually be spammed when effects happen
soreau: I have to start a new terminal to get it to stop
soreau: Maybe I just have to 2>/dev/null..
soreau: but I'd like to find the cause of the spam
adamk_: If I get a chance, I might try updating to drm-next this weekend. Last time i tried, there were unresolved conflicts.
Kaapa: adamk_: is that on the kernel already?
adamk_: Hmmm? Yes, this was when I tried to pull airlied_drm_remote drm_next
Kaapa: I don't mean git
Kaapa: I mean 32-rc
adamk_: KMS is available in the rc kernels, I believe.
adamk_: But airlied mentioned yesterday to me that he only recome
adamk_: sorry. He only recommends drm-next for KMS.
soreau: adamk_: The conflicts were fixed last I checked but that was a while ago now
Kaapa: kms is, I meant drm-next
soreau: Fortunately, piping stderr to /dev/null works for now
adamk_: Errr.
soreau: I don't understand the reason for all the spam since things seem to be working fine.
adamk_: Kaapa: drm-next is what will be merged into the kernel at some point down the road.
soreau: suspects mesa
adamk_: Kaapa: If it is merged at one point, it is then populated, so to speak, with more changes that will be merged in at some other point.
adamk_: So it's not like you can just say that drm-next is available in a particular kernel.
Kaapa: adamk_: ok. I thought airlied said that "somehwere down the road" would be in the next rc
Kaapa: k. I'd like to enable userspace and not have my X crash every 10 seconds
Kaapa: :)
adamk_: Right, and then more development happens in drm-next.
adamk_: For KMS, airlied always seems to recommend drm-next.
adamk_: Well, at least at this time.
spreeuw: you're famous
spreeuw: "there's some bigger-than-I-would-have-liked
spreeuw: patches to the Radeon KMS driver. Hopefully it's nearing stability now.
spreeuw: "
[Enrico]: eheheh
Ghworg: We all hope that :-)
cxo: I recorded a timedemo on openarena, and i got 100fps with a hd4870, anyone want the timedemo?
spreeuw: theres a reference on the benchmark page
cxo: wheres the benchmark page?
spreeuw: http://dri.freedesktop.org/wiki/Benchmarking
spreeuw: if only they could be arsed to include a demo and a configfile in the openarena distribution
spreeuw: but I asked that 2 years ago already
spreeuw: great this isnt available for download
spreeuw: cxo: shall we make a wiki page?
spreeuw: with results
kkk: If I want to debug X, do I need to run gdb with ssh on another pc?
spreeuw: yep if you have io eater bugs
cxo: spreeuw, yeah we could
cxo: that demo doesnt work on the newer versions of openarena
cxo: openarena is at 0.8.1 right now, and the oldest demo you can use with that is 0.7.1
kkk: I just tried to use gdb in the same machine
cxo: spreeuw, I have my demo and score for a baseline
kkk: and when X crashed, the system freezed
cxo: I think i just crashed the radeon driver
cxo: here you can see the kernel log http://en.pastebin.ca/1670367
cxo: line 890 onwards
spreeuw: cxo: put it online together with the config you use
cxo: where do i upload it to?
cxo: its a 0.5mb binary file
adamk_: So when I run 'git pull airlied_drm_remote drm-next' I'm getting conflicts in intel_display.c and intel_sdvo.c
adamk_: Any thoughts?
adamk_: I'm following these steps in the wiki for using drm-next: http://pastebin.com/m5689571
soreau: adamk_: There is a way to fix it with git merge IIRC
soreau: Where is rnoland when you need him :)
adamk_: Well consdering I'm on Slackware at the moment, I don't think he'd be able to help much :-)
adamk_: I'm guessing all the radeon guys are on/at the gallium3d talk/presentation.
soreau: adamk_: Try 'git merge'. I'll bbiab
adamk_: 'git merge' just gives me 'fatal: You have not concluded your merge. (MERGE_HEAD exists)
adamk_: bbiab.
soreau: adamk_: The reason I mentioned rnoland was because he's been extremely helpful to me with git issues in the past
soreau: wants to know if the minutes of the gallium3d talk/presentation will be available anywhere
soreau: bbl
Wood_Warrior: HI, my card is radeon x1200 (RS690M), xvinfo says: "Radeon Textured Video", Xorg.0.log says: "(II) RADEON(0): Direct rendering enabled"
Wood_Warrior: but watching HD videos with "mplayer -vo xv" results in tearing picture due to "X" (not mplayer) waste much cpu
Wood_Warrior: top says mplayer 27% cpu , and X 100% cpu (its dual core processor)
biotube: sounds like a vsync issue
Wood_Warrior: this means that mplayer do not use direct rendering ?
biotube: no, it means that it's not timing its frames to the monitor
Wood_Warrior: is there way to solve this problem ?
biotube: mplayer should have an option to enable vsync
soreau: Wood_Warrior: Are you using kms?
Wood_Warrior: tryed vsync - does not helps - X still 100% cpu
Wood_Warrior: was it direct rendering X should not use more cpu than mplayer i think ?
Wood_Warrior: soreau: not
Wood_Warrior: but i tryed kms in the past - not resulting on mplayer performance
soreau: are you running a compositing manager?
Wood_Warrior: dunno whats this, i use mesa - glxinfo says: "GL_RENDERER: Mesa DRI R300 (RS690 791F) 20090101 NO-TCL"
biotube: Wood_Warrior: do you use compiz or kwin effects?
Wood_Warrior: i think not - its slackware64 instaled - desktop efects disabled
Wood_Warrior: there is executable named compiz, but it is not ran i think - there no such process compiz
Wood_Warrior: there is process kwin
[Enrico]: Wood_Warrior: well kwin can run with both desktop effects enabled or disabled
soreau: which version of X and mesa?
[Enrico]: Wood_Warrior: you should check in systemsetting (kde config) to check if desktop effects are enabled
Wood_Warrior: soreau: latest libdrm, mesa, xf86-video-ati - got from cgit yesterday
Ghworg: X uses a lot of cpu with mplayer for me too (not 100%, but significant), but only in windowed mode. Fullscreen it's fine
Wood_Warrior: [Enrico]: desktop efects are not enabled
biotube: how does mplayer behave with the opengl driver?
Wood_Warrior: Ghworg: my case is mplayer -vo xv -fs
cxo: biotube, it crashes sometimes when i play bluray rips, but otherwise its ok
cxo: XV works fine though
[Enrico]: is Option "DynamicPM" working on r600 (without kms) by any chance ?
soreau: maybe he needs to define the textured video port from the output of xvinfo with -vo xv:port=57 or whatever/
Wood_Warrior: biotube: the first run mplayer -vo gl -fs - bad performance but working - second run gave me the message "do_wait: drmWaitVBlank returned -1, IRQs don't seem to be working correctly"
biotube: Wood_Warrior: sounds like the first one tried the software rasterizer
biotube: as for IRQs, the code for that's not in yet
agd5f: Irqs should work for pre-r6xx
Wood_Warrior: well mplayer option -nodouble reduces significantly X cpu usage, about twice , also helps with gl driver - no more this message "do_wait: drmWaitVBlank returned -1, IRQs don't seem to be working correctly"
Wood_Warrior: "mplayer -vo xv -nodouble" is better but stil there is problems with performance - mplayer uses about 25 % cpu, X - 67%cpu
Wood_Warrior: soreau: -vo xv:port=57 changes nothing
soreau: Wood_Warrior: You have to see which textured video port is yours in the output of xvinfo
Wood_Warrior: it is 57
Wood_Warrior: port base: 57
soreau: oh well that was my only idea
Wood_Warrior: [Enrico]: i have enabled "DynamicPM" in the xorg.conf , my kernel module "radeon" supports kms, but it is not enabled
soreau: but I assume irq's should be working for you *shrug*
chithead: kms and power management do not work together atm
[Enrico]: Wood_Warrior: well but is it working? here it is not with r600 and i guess i need to wait for kms powermanagement support :D
[Enrico]: indeed i disabled kms
Wood_Warrior: [Enrico]: my kernel is 2.6.32-rc6-git5, but option "Staging drivers" is disabled
Wood_Warrior: with this kms is not enabled by default at boot time (module built in kernel)
[Enrico]: Wood_Warrior: mine is .31 zen7 + drm-next with staging drivers enabled (but nomodeset in the kernel command line when i want powersave forced to low mode via xorg.conf
[Enrico]: )
[Enrico]: you know the pc gone hot here without low powermode eheheh
[Enrico]: so basically i can't use kms until powersave is done
[Enrico]: at least the static one
[Enrico]: i can wait for the dinamyc
[Enrico]: well good night people i go to sleep :D
_Groo_: hi/2 all
_Groo_: airlied: ping
_Groo_: agd5f: ping
cxo: Whats the difference between GLX and SDL?
cxo: nexuiz has two executables, with those as suffixes
MostAwesomeDude: Then those would be two different ways of getting a GL context.
MostAwesomeDude: Both should work nearly equivalently.
_Groo_: MostAwesomeDude: isnt sdl a 2d context? it can do 3d but it started has a 2d game engine
MostAwesomeDude: _Groo_: You can tell SDL to create a GL context instead of a mmap'd surface.
_Groo_: MostAwesomeDude: yeah i know but usually games have two binaries, glx , pure 3d and sdl, 2d context...
_Groo_: MostAwesomeDude: can you take a look at tghe google earth bug btw?
cxo: Not bad, Nexuiz plays nice on the r700
cxo: i really want to give MW2 a go, but i think it will need a really good GLSL setup
cxo: This game was only released 2 days ago and there are already numerous appdb entries for it, amazing http://appdb.winehq.org/objectManager.php?sClass=version&iId=18348
Wood_Warrior: biotube: "mplayer -vo gl:yuv=2:rectangle=1:force-pbo:ati-hack -dr -nodouble" works - "X" cpu usage is zero now, mplayer cpu usage is about 50%, but still slow video !
cxo: this is so cool, its an updated Quake 1, http://icculus.org/twilight/darkplaces/screenshots.html
quaker: anyone have second hand of Radeon FirePro 2450?
quaker: 5 cent is accepted. lol
cxo: Do you have Quake 3?
EruditeHermit: hello
hifi: cxo: more amazing is that the game was playable with wine without any fixes
soreau: EruditeHermit: When running compiz from your terminal, do you get spammed with all kinds of messages to stderr even though it runs ok/
soreau: ?
EruditeHermit: soreau, it crashed me
soreau: I guess you dont use compiz regularly then :P
EruditeHermit: nope
EruditeHermit: let me try again
EruditeHermit: if I sign off you'll know why
soreau: What happens?
soreau: It crashes X?
EruditeHermit: hangs computer
soreau: wow
EruditeHermit: or atleast it did last time
EruditeHermit: this time its fine
soreau: oh cool
EruditeHermit: no it doesn't spam the console
EruditeHermit: what are you using to run it
EruditeHermit: I did compiz --replace&
EruditeHermit: but I am using DRI1
soreau: are you using drm-next and all latest libdrm, mesa and ddx with x 1.7.1?
soreau: oh dri1