glisse: airlied: sighs so many regs to think about...
ptexpress: Does radeon support 1440x900 resolution on the IGP 340M card?
ptexpress: I'm trying to set up an external monitor for my laptop.
SunTzuTech: Should. my lattop with IGP340M runs 1400x1050 native
ghepeu: agd5f, this morning I rebooted the machine to try to find the commit that cause my problem with fonts and apprently after the reboot the problem's gone
ghepeu: the minor differences in colors (favicons, grey tones in antialiased fonts, icons borders) are still there but they're unnoticeable
ghepeu: that said, I think that the random lockups I was experiencing with my rv370 were due to a combination of commit 31815730732a5d2a446aa316a5b4d837766762e6 in drm and something that went in mesa between 7.0.3rc2 and 7.0.3 final
ghepeu: I upgraded to xf86-video-ati from git to check what happens
bridgman: airlied: 0x2184 is VAP_VSM_VTX_ASSM - Vertex state memory assemble control register
MostAwesomeDude: bridgman: Morning.
bridgman: Each bit in this field applies to the corresponding vector in the VSM memory (i.e. Bit 0 applies to VECTOR_0 (POSITION), etc). If the bit is set, then the corresponding 4-Dword Vector is output into the Vertex Stream.
bridgman: N.A.D.: oh good; spend enough time on IRC and you think nobody is in the same timezone ;)
bridgman: MAD, not NAD, sorry
MostAwesomeDude: bridgman: I don't know about you, but it's like five in the morning here. I couldn't sleep 'cause of the heat wave.
bridgman: 9:30 AM, couldn't sleep 'cause it was cold; last week was warm, switched heating off
bridgman: time to light a fire
MostAwesomeDude: Good plan.
bridgman: 5AM: Hawaii ?
MostAwesomeDude: Mistyped; should be 6.
bridgman: Same thing ;)
MostAwesomeDude: And now it's closer to 7. Go figure.
MostAwesomeDude: Well, no. They have indoor plumbing in Hawaii, I think.
bridgman: The curse of driver development; it sucks you in for hours at a time
bridgman: A heat wave in Oregon; that's unusual isn't it ?
MostAwesomeDude: Yeah. Weird temperatures all over.
MostAwesomeDude: I guess Southern Cali and Arizona are up to 110+ today and tomorrow.
bridgman: ok, maybe there is something to global warming then. Every year it gets colder here
bridgman: 110 - ouch. I'm staying home
bridgman: you guys are really making great progress on 5xx 3d
bridgman: I can see how this 2184 register would make a difference; masking the vertex stream and all ;)
MostAwesomeDude: Thanks. airlied and agd5f have done all the register work, I'm just working on the maths, really.
bridgman: I sometimes think we ought to have graphics developers go to schools and explain to
MostAwesomeDude: But we're mostly there. There's only a few FP insts left, and they're maths.
bridgman: kids that there really IS something interesting you can do with math ;)
MostAwesomeDude: Hehe. No kidding.
bridgman: or maybe it would just scare them away from their dreams of being a game developer ;)
MostAwesomeDude: Linus lives here, to the north in Beaverton. Every year, he comes down to Oregon State and gives a few talks on development to inspire the would-be comp scientists.
MostAwesomeDude: I've never met him, but he's apparently a chill guy.
bridgman: that's great; is the turnout good ?
MostAwesomeDude: It's pretty good. Lots of OSS guys up there workin' away at stuff.
MostAwesomeDude: I'd like to go back to school there, when I can.
bridgman: yep, seems like a great part of the world; nice mix of laid-back & smart people
bridgman: my VPN connection just crashed, so I'm gonna brain-dump the rest of the register info
bridgman: before I lose it
MostAwesomeDude: It is. I don't want to leave Western Oregon, if I can avoid it.
MostAwesomeDude: 'k. Go for it.
bridgman: field name is OUTPUT_VEC_CNTL, bits 21:0
bridgman: hmm. guess that's it ;)
MostAwesomeDude: 'k. That is an interesting reg to know.
bridgman: glisse: haven't forgotten your request; airlied just asks much easier questions
bridgman: yeah, I can see it wreaking havoc if not set properly
bridgman: re: oregon, there has to be a reason everyone moves to the west coast and nobody
bridgman: ever seems to move the other way
MostAwesomeDude: bridgman: It can't be the weather. Nine-month rainy season and no beaches.
bridgman: no beaches ?
bridgman: oh yeah, I remember... big rocks
bridgman: need another million years or so for them to break down to beaches
bridgman: yeah, it's starting to come back now; I worked out there for a few months, don't
MostAwesomeDude: There's some beach, but it's eternally cold and bad weather.
bridgman: remember ever seeing the sun. It was always this beautiful, slightly hazy kind of
bridgman: weather; I figured it was there to make the trees look nice
MostAwesomeDude: Huh, could be.
bridgman: Anyways, I still liked it. You just need a big fire to make up for the weather ;)
bridgman: speaking of fire, I'm gonna go haul in some firewood; back in a bit
MostAwesomeDude: New FP stuffs in git, for anybody who cares.
MostAwesomeDude: All the simple stuff.
stefan1: MostAwesomeDude: Argghhh more changes :-)
stefan1: I guess I have to recompile mesa again as you just committed more r500 stuff
glisse: stefan1: usualy you don't need to rebuild whole mesa
stefan1: MostAwesomeDude: But with the changes from the last 10 hours glxgear works correctly on Mobility X1300 now
glisse: if you built it one time just go in r300 directory and type make
stefan1: glisse: I use RPM, easier to manage for me :-)
stefan1: MostAwesomeDude: No triangle errors on the red gear anymore
stefan1: MostAwesomeDude: And also kwin4 composite looks better now. At least you can recognize window contents now. Still isn't correct and lots of crashes though
stefan1: I guess those added r500 opcodes in mesa means it is using HW acceleration instead of SW renderer for those?
stefan1: or does it add missing functionality/correctness?
stefan1: *phew* looks like when I did a git pull I got those changes in already
stefan1: So I guess it means those are responsible for fixing glxgear on r5xx
MostAwesomeDude: Ah, he left.
MostAwesomeDude: But it looks like together, we are improving the situation for r5xx greatly.
agd5f: airlied: what bridgman said :) I'll put the bits in git once I'm back up
MostAwesomeDude: agd5f: I have git again.
agd5f: MostAwesomeDude: cool. I'm not up yet
MostAwesomeDude: 'k. I have one more to push now, but I may have more later.
agd5f: sounds good
dli: drmRadeonCmdBuffer: -22
agd5f: dli: need a newer drm
dli: agd5f, thanks
dli: agd5f, glxgears works now
dli: -vo xv, fullscreen still crashes, the cursor is shown and moving, but have to power off to recover
MostAwesomeDude: Or evening.
MostAwesomeDude: Or very early morning.
MostAwesomeDude: Out of curiousity, what's SmoothFlag and why must we fall back on it?
MostAwesomeDude: I gather that it's GL_POLYGON_SMOOTH.
MostAwesomeDude: Nevermind. I see.
leio: and I don't see such an option in my driconf (to disable low impact fallbacks that was talked about), I assume r200 doesn't have that or something
MostAwesomeDude: leio: It's R300+.
MostAwesomeDude: radeon doesn't have that option.
MostAwesomeDude: agd5f: Do you know anything about AA lines, by any chance?
eboettcher: MostAwesomeDude: it's on my gsoc todo list :)
airlied: MostAwesomeDude: fglrx on r300 had to use the shader to do it.
airlied: MostAwesomeDude: maybe revenge dumps had the info..
airlied: bridgman: thanks (assuming you read the logs ;-)