Nightwulf|work: hi all
osiris__: mmp: does your machine hang when starting X?
osiris__: sorry, wrong chan
dragonstear: Works like a charm.
dragonstear: Don't get it but it works. Cool.
wizir-: hello, i have a question, does x.org support hd4850?
Thargor: wizir-: both drivers (radeon and radeonhd) should work
Thargor: but without any acceleration
wizir-: Thargor, thanks
Thargor: neither 2d nor 3d
Nightwulf|work: important to say, that only the git tree supports 7xx!
Ferocanis: how do i tell if my 2d acceleration is working properly?
MostAwesomeDude: Ferocanis: Turn on a compositing manager, drag stuff around. :3
Ferocanis: i could try that with kwin
Ferocanis: does aiglx need to be enabled?
MostAwesomeDude: KDE3 or KDE4?
MostAwesomeDude: I have KDE3 with kompmgr, and I can *definitely* tell when it's not working.
marcheu: /win 14
Ferocanis: that was failtastic, i don't really think compositing even started
Ferocanis: i enabled it and the screen went a couple different solid colors and i had to beat kwin into submission
MostAwesomeDude: Ferocanis: kwin4 is 3D-accelerated, I thought...
Ferocanis: never mind then
Ferocanis: i think you're right
Ferocanis: any other, better ways to test if 2d accel is working properly?
MostAwesomeDude: Mm, dunno.
Ferocanis: well that sucked. has anyone else had a (2d rendering related?) problem that causes the computer to go unresponsive and the mouse to lag around the screen?
Ferocanis: it's happened to me a few times now, mostly when scrolling
Ferocanis: for example, scrolling over large images on a webpage
yangman: not since 2d acceleration was added ages ago
Ferocanis: could it have something to do with my acceleration method?
Ferocanis: i think right now it's xaa
dragonstear: Hm, but as far as I understand it depends on what card you use... Or?
dragonstear: am I wrong
yangman: somewhat, yes
libv: dragonstear: yes, if you use R6xx and up then there's nothing yet
Ferocanis: well, i've got an rs690
dragonstear: Then it looks bad^^
libv: Ferocanis: then you should be reasonably happy
Ferocanis: libv, i am reasonably happy!
libv: Ferocanis: no, i meant that you then should have reasonable performance :)
Ferocanis: well, openarena is playable ;)
Ferocanis: libv, i was just wondering if anyone had this problem or if you knew about it or if i'm doing something wrong :)
yangman: DRI is enabled?
yangman: and how recent is the driver?
libv: Ferocanis: what happens if you switch to exa?
Ferocanis: dri is enabled and i pulled it from git a couple days ago
Ferocanis: libv, i'll give it a try
Ferocanis: libv, it's working so far
Ferocanis: i'll just have to use it and see if something funky happens
libv: Ferocanis: does this feel faster?
Ferocanis: weird crash... libv, i'll tell you if it feels faster once i get to facebook XD
libv: Ferocanis: so it crashes on you?
Ferocanis: same issue as with xaa
Ferocanis: laggy mouse, unresponsive system
libv: Ferocanis: oh, so the crash was not related to using exa
Ferocanis: maybe it isn't 2d accel related
libv: Ferocanis: run top and see what's causing it
Ferocanis: libv, the issue is that it happens pretty randomly and i can't even switch vts or anything
libv: Ferocanis: dmesg giving you warnings?
Ferocanis: X:7644 conflicting memory types c0000000-c8000000 write-combining<->uncached-minus
Ferocanis: reserve_memtype failed 0xc0000000-0xc8000000, track write-combining, req write-combining
Ferocanis: does that count?
libv: i was more looking for a drm warning
Ferocanis: i'll fire up sshd for the next time it happens and see if i can tell
Ferocanis: and i'm looking at /var/log/messages now
Ferocanis: nothing from drm, there's an unhandled acpi event before i start pushing magic sysrq keys, but i think that's a coincidence
Zajec2: does radeonhd uses drm module?
Ferocanis: i'll just wait for something weird to happen again
Temujin_: Zajec2: it wouldn't compile without libdrm-dev on Debian/Ubuntu, so my "spidey sense" says yes
Ferocanis: Temujin_: I'm pretty sure spidey senses tingle, not speak o_O
Temujin_: tinginling is a form of communication
Temujin_: tingling rather
Ferocanis: i'll give you that
Temujin_: thank you
Ferocanis: libv: i've found the longest, most obnoxious facebook page i could and scrolling feels very smooth with exa...
Ferocanis: frozen again, i tried to change focus to another window this time
libv: hardlocked or just stalled?
Ferocanis: well, the mouse lags around the screen, alt+ctr+backspace doesn't do anything
Ferocanis: the sysrq keys work, but i can't get control of the machine again without rebooting
Ferocanis: the hard drive light flashes occasionally, so i'm thinking it's just hardcore stalled
Ferocani1: hrm, net cut out for the past 2 minutes
Ferocani1: what do you think, libv?
libv: Ferocanis: try disabling acceleration
libv: go for shadowfb and see whether this changes anything
Ferocani1: okay, i'm using shadowfb now
PeterHofer: Good evening
PeterHofer: I thought I'd ask here before writing to the mailing list
PeterHofer: I'm using FreeBSD 7.0 with Xorg 7.3 (Xorg server 1.4.2)
PeterHofer: I have previously used my onboard Radeon HD 3200 (AMD 780G chipset), which worked fine with the radeonhd driver obtained from git
PeterHofer: My 24" monitor with 1920x1200 native resolution was attached to HDMI and my 19" with 1280x1024 to normal VGA output
PeterHofer: With this setup, I had my desktop extended to the 19" display, with the 24" being the primary display with panels and all
ndim: PeterHofer: Given the time of day... you'll probably be better off sending to the mailing list. That is quite a lot of text as an exposition, you can just use that in the mail :)
PeterHofer: I suppose you're right
ndim: I can't help you on dual screen stuff (my second display is broken), so I could only recite stuff you could get from the channel logs or from the mailing list archives yourself.
PeterHofer: Guess I'll just make it short: Does anyone have a Sapphire Radeon HD 3470 or are there any known problems with the card and/or chip (RV620)?
ndim: No one else appears to be around. :)
yangman: so, what's the issue?
yangman: 3470 seems to be the source of a lot of issues
PeterHofer: Both displays always show exactly the same
yangman: always, or at the beginning?
PeterHofer: At the beginning and always afterwards
yangman: xrandr has no effect?
PeterHofer: xrandr --auto does not help anything
ndim: The xrandr defaults have changed somewhate.
PeterHofer: No, I'm afraid it does not
yangman: the default is to clone screens
yangman: so there's no problems there
yangman: what xrandr command are you using to get dual-head?
PeterHofer: Xorg/RandR seem to believe that everything is correct
PeterHofer: They show that both monitors have their native resolutions, but both have 1280x1024
PeterHofer: yangman: I have tried using LeftOf and RightOf in xorg.conf, of couse
yangman: I'm talking about xrandr
yangman: not xorg.conf
PeterHofer: I have tried --left-of and --right-of there as well
PeterHofer: and also --auto for both outputs and without specifying an output
yangman: and you're sure the output names are correct?
PeterHofer: Yes, absolutely
PeterHofer: They're "DVI-I_1/digital" and "DVI-I_2/digital"
adamk: PeterHofer, Check the 'xrandr --verbose' output and make sure they have different CRTC values.
ndim: BTW, do we still need "Virtual" in xorg.conf to make non-cloned multi-displays work?
yangman: ndim: yes
PeterHofer: For DVI-I_2/digital, it says: CRTC: 0; for DVI-I_1/digital, it says: CRTC: 1
adamk: OK, so that's not it.
ndim: I presume PeterHofer has Virtual set to something large enough.
PeterHofer: I have tried a lot of settings, including Virtual
PeterHofer: With everything I've tried, the best outcome would be that both displays would show the same at 1280x1024
PeterHofer: All other outcomes were that either one or both display would be blank
PeterHofer: yangman: That sounds similar to the problem I have.
yangman: yeah, seems to be a recent regression
yangman: I misread the card model you were talking about, and didn't occur to bring that up
PeterHofer: It's a Sapphire Radeon HD 3470 with 256 MB GDDR3 and 2x DVI
PeterHofer: I just found this in the Xorg.0.log:
PeterHofer: 113-B40802-X05 RV620 GDDR3_16Mx32 64bit 256MB 800e/700m
PeterHofer: yangman: Since this appears to be a regression, could you point me to an earlier revision for which it's likely the problem did not exist?
yangman: PeterHofer: try something before the massive atombios_support branch merge
yangman: commit messages are sent to the list, so you can look through the archives to find the hash
PeterHofer: yangman: Alright, I think I got it now -- the last commit should be: "XAA: More uint*_t fixes."
PeterHofer: Let's see if it works...
PeterHofer: Still no luck :-|
yangman: I've been looking at the AtomBIOS code in radeonhd
yangman: is it correct to assume that it's not using any of the type definitions from atombios.h directly, but parses everything for radeonhd specifically?
yangman: and that the convention is to only access AtomBIOS table data through RHDAtomBiosFunc()?