EruditeHermit: airlied: is there a set of patches to a normal kernel that I can apply rather than using fedora rawhide kernels for kms?
EruditeHermit: and in your blog instructions does it mean "a kernel with GEM bits in it" ?
EruditeHermit: *what does it mean
airlied: EruditeHermit: there is the drm-rawhide tree
airlied: EruditeHermit: branch rather of my drm-2.6.git tree
EruditeHermit: and what do you mean a kernel with GEM bits?
airlied: that tree or a tree with the GEM patches applied, like anholts drm-gem-merge tree
EruditeHermit: so 2.6.27 vanilla doesn't count
airlied: really the drm-rawhide branch is the easiest place
airlied: well Fedora kernels are easiest then that branch :)
airlied: you could also just apply the drm-modesetting-radeon.patch from Fedora kernels
EruditeHermit: I can apply that patch to 2.6.27?
airlied: should be able..
EruditeHermit: I'll try that
airlied: its just the drm-rawhide branch in one big patch
airlied: probably easier to just use git to do it all :)
EruditeHermit: ok, I'll try to use the drm-rawhide git repo but MAD said he had trouble getting it to play nicely with the debian kernel building tools
EruditeHermit: well let me try
MostAwesomeDude: EruditeHermit: I just went ahead and built the kernel "manually."
MostAwesomeDude: I've never had good luck with make-kpkg.
MostAwesomeDude: airlied: Have you pushed your r6xx fixes? My workstation will be here tomorrow. :3
EruditeHermit: btw will radeon work with r6xx and r7xx right now?
airlied: MostAwesomeDude: I need to push them intor my rawhide tree.
airlied: MostAwesomeDude: with the pciids.
MostAwesomeDude: airlied: I figured. No rush.
MostAwesomeDude: Was it the FB location, or something completely different?
airlied: MostAwesomeDude: no parser for the bios table for the outputs
airlied: MostAwesomeDude: and fb location bits
MostAwesomeDude: Ah. Well, now I don't feel *quite* so stupid. :3
airlied: I didn't realise the parser wasn't ported until I wrote it :)
JamesLast: Hi All
JamesLast: i would like to test modeset on an x700, however booting the kernel with modeset on only gives me a white screen on my laptop display
airlied: JamesLast: doesn't sound nice.. what kernel you running?
JamesLast: 2.6.27-0.352.rc7.git1.fc9.x86_64 (running fc-9 with recompiled koji-rpms)
airlied: have you ssh access?
JamesLast: nope behind a router, sorry
airlied: no I mean for yourself locally?
JamesLast: network is wireless and i havent figured out how to work without networkmanager, give me some time and i set up static networking :)
airlied: if you can ssh in, you can try the debugging bits in http://airlied.livejournal.com/
airlied: at the end of the top entry.
airlied: can you post a normal Xorg.0.log from that machine to pastebin or somewhere?
JamesLast: i typed blindly and have a copy of messages with drm debug=1, its here http://pastebin.com/m32b0b60e
JamesLast: xorg.log from startx (still white screen) : http://pastebin.com/m40abbfe6
JamesLast: xorg.log with nomodeset : http://pastebin.com/m7b96571d
JamesLast: ok setup static network, here comes dmesg from ssh (debug and modeset on): http://pastebin.com/m7ebed474
airlied: JamesLast: could you open a bug on bugzilla.redhaty.com
airlied: and attach all that.. I think agd5f might be changing the r400 code soon.
JamesLast: ok will do that, thank you
JamesLast: airlied: opened bug on redhats bugzilla : https://bugzilla.redhat.com/show_bug.cgi?id=463895
osiris_: MrCooper: I have a question about patch that sets FRAG_ATTRIB_FOGC. it tells us that fog color is used in FP, right?
MrCooper: osiris_: not sure but think so
Rioting_pacifist: is it possible to test the latest drivers without messing up my X (e.g i can fall back to old ones easily, bearing in mind im a noob)
osiris_: MrCooper: shouldn't it be cleared for fixed function fog then?
MrCooper: no idea
z3ro: ok, so that was interesting.
z3ro: I just paid a nz taxi in coins and a 5 euro note. :P
z3ro: he wasn't too happy, even though it's like 3x exchange rate.
z3ro: goes to pass out.
z3ro: the floor is not soft =/
MostAwesomeDude: osiris_: FRAG_ATTRIB_FOGC means that fragment.fogcoord is used in the FP.
MostAwesomeDude: It either comes from the SC, or from the fogcoord in VPS/SW TCL.
spstarr_work: hullo MAD
MostAwesomeDude: spstarr_work: Mornin'.
spstarr_work: afternoon by 10 mins ;)
MostAwesomeDude: I just woke up and haven't eaten, plus it's only 9 AM here. :3
spstarr_work: don't you work? :)
spstarr_work: or you're doing the school thing
MostAwesomeDude: agd5f: I know you're quite busy, and that everybody's bugging you about docs. With that in mind, do you have the magic numbers for ATI HW fog -> OGL fog? If not, I'll RE from revenge, since fglrx works on my workstation now, but I thought I'd ask. :3
MostAwesomeDude: spstarr_work: I'm in school, and I *should* be working.
agd5f: MostAwesomeDude: what do mean?
MostAwesomeDude: agd5f: Well, for each fog type, there's a magic multiplier for fog start and fog density. Right now, it works. But, if I switch over to the fog coord path, all of a sudden it just, well, doesn't.
MostAwesomeDude: Was just wondering if you guys have a book somewhere: "Fog in OpenGL and how we abuse the r3xx chipset to make it work right."
agd5f: MostAwesomeDude: ah ok. I'll see if I can find anything
agd5f: MostAwesomeDude: is this for the fog blends?
MostAwesomeDude: agd5f: Yeah. Basically, I don't want to move *all* the fog processing into the FP, because the fog HW is faster.
spstarr_work: didn't recognize (or see) any ATI/AMD people at the food court
terracon: spstarr_work: you're in markham?
spstarr_work: terracon: yep
spstarr_work: your @ AMD dude too?
terracon: cool I used to work for a company on west beaver
terracon: so I would pass the building's every day
spstarr_work: ATI's HQ is like mins from here
zhick: does radeon ignore modelines?
zhick: cause i just cant get my monitor to use 1280x1024@75Hz at X startup...
agd5f: zhick: should work http://wiki.debian.org/XStrikeForce/HowToRandR12
zhick: agd5f: i'm pretty sure i already tried what's mentioned there, but ill try again now. :>
zhick: well, it doesn't work. x still start with 60 hz.
zhick: xrandr -r 75 works fine though... well, it sometimes crashes X, but who cares ;)
zhick: would somebody have a look at my xorg.conf and tell me if i'm doing something wrong?
agd5f: zhick: pastebin it
agd5f: and your log
zhick: agd5f: xorg.conf: http://pastebin.com/m2e001eab
zhick: agd5f: and my log: http://pastebin.com/m4b9ee965
agd5f: zhick: you need to associate your monitor section with the output you want to use it with or change the monitor indentifier to the name of the output
zhick: agd5f: so for example i should change the identifiere to DVI-0 ?
agd5f: zhick: yeah
zhick: agd5f: thanks alot :) will try right now :>
agd5f: zhick: although 60 hz is the perferred mode for your panel
agd5f: it'll probably look better at 60 than 75
zhick: agd5f: well, that would probably usualy be the case, but my monitor seems to be somewhat defect. at 60 hz there are blue horizontal lines and red dots everywhere.
zhick: zhick: but they dissapear at 75hz
zhick: agd5f: yeah, i'm pretty sure it's a problem with the lcd.
zhick: yay, it works :) thanks alot again agd5f
zhick: and you guys are doing a great job at the driver, btw :)
kdekorte: I just updated F9 to the latest stuff and for some reason when I run mplayer with the gl vo I need to add swapinterval=0. I don't recall having to do that before
kdekorte: mplayer is giving me this message
kdekorte: VO: [gl] 640x480 => 640x480 BGRA [zoom]
kdekorte: do_wait: drmWaitVBlank returned -1, IRQs don't seem to be working correctly.
kdekorte: Try adjusting the vblank_mode configuration parameter.
kdekorte: if I don't have the swapinterval set to 0
MrCooper: agd5f: was your last comment to #17767 meant for #17723? And if so, do you have any reason to believe 6.9.0 would behave differently? I'm not sure how it could
agd5f: I agree
agd5f: MrCooper: I don't see how it could, but maybe I missed something subtle
MrCooper: if #17767 is indeed the same problem, isn't it funny how those things tend to appear in gangs
orkid: ding dong. any news on xv for r6xx? :)
orkid: and as always, thanks for your efforts up to this point
airlied: agd5f: so those pll changes with the 120000 and the 200000 bust my r600 + 30" monitor
spstarr_coding: airlied: temporary i hope :)