--- Log opened Втр Авг 01 00:00:03 2006
00:15 < ddl> hey
00:21 -!- qfire [n=qfire@dsl093-010-213.cle1.dsl.speakeasy.net] has quit [Read error: 104 (Connection reset by peer)]
00:29 < lumag_offline> huh. And I wondered, why it didn't start :) 
00:29 -!- qfire [n=qfire@dsl093-010-213.cle1.dsl.speakeasy.net] has joined #nouveau
00:39 < pmdata> yeah, wrong window :)
00:44 < lumag_offline> yup.
00:45 < lumag_offline> Unfortunately still no luck with fifo programming.
00:52 -!- Gloubi [n=Gloubi@ABordeaux-253-1-54-146.w82-125.abo.wanadoo.fr] has joined #nouveau
00:52 < pmdata> hi gloubi
00:52 < Gloubi> hi/salut all
01:25 < pmdata> gloubi> what is your hw?
01:26 < Gloubi> my cg?
01:26 < Gloubi> geforce 4 ti 4200
01:26 < Gloubi> un bon cru :)
01:29 < pmdata> nv2x then
01:29 < pmdata> could you make some dump with latest renouveau cvs?
01:30 < Gloubi> you want i try the last cvs?
01:30 < pmdata> yep
01:30 < pmdata> I have some nv28 dumps, but many things have been added to cvs, so I want to see the progress
01:30 < Gloubi> i'm on gentoo amd64 with xorg 7.1 , you think it's good?
01:31 -!- nano [n=nano@h100n2fls33o1016.telia.com] has quit ["leaving"]
01:31 < pmdata> I hope, I don't know if it works on 64bit
01:31 < Gloubi> i will try to install the cvs
01:32 < Gloubi> what module must i install?
01:32 < Gloubi> The current code in CVS :
01:32 < Gloubi>     *
01:32 < Gloubi>       The DRI CVS holds the DRI module.
01:32 < Gloubi>     *
01:32 < Gloubi>       The DRM CVS holds the DRM module.
01:32 < Gloubi>     *
01:32 < Gloubi>       The DDX CVS holds the "nv" DDX modified to initialize 3D operation.
01:35 < pmdata> renouveau
01:35 -!- nano [n=nano@h100n2fls33o1016.telia.com] has joined #nouveau
01:36 < Gloubi> make: *** [renouveau] Erreur 1
01:36 < Gloubi> ^^
01:37 < Gloubi> NVCtrl.o: dans la fonction « find_display »:
01:37 < Gloubi> /home/abyss/renouveau/NVCtrl.c:50: référence indéfinie vers « XextCreateExtension »
01:37 < Gloubi> /home/abyss/renouveau/NVCtrl.c:50: référence indéfinie vers « XextFindDisplay »
01:37 < Gloubi> /home/abyss/renouveau/NVCtrl.c:50: référence indéfinie vers « XextAddDisplay »
01:37 < Gloubi> NVCtrl.o: dans la fonction « close_display »:
01:37 < Gloubi> /home/abyss/renouveau/NVCtrl.c:55: référence indéfinie vers « XextRemoveDisplay »
01:37 < Gloubi> NVCtrl.o: dans la fonction « XNVCTRLQueryVersion »:
01:37 < Gloubi> /home/abyss/renouveau/NVCtrl.c:122: référence indéfinie vers « XMissingExtension »
01:37 < Gloubi> /home/abyss/renouveau/NVCtrl.c:125: référence indéfinie vers « _XFlush »
01:37 < Gloubi> etc....
01:38 < pmdata> you need xlib dev files
01:39 < pmdata> do you have SDL installed?
01:39 < Gloubi> no
01:40 < pmdata> you need to compile it with X11 support
01:40 < nano> Here's from a "nVidia Corporation NV17 [GeForce4 MX 420] (rev a3)" if that gives you anything, http://debian.as/~nano/renouveau.txt
01:41 < Gloubi> why do you use SDL ?
01:41 < pmdata> to setup the opengl context
01:42 < pmdata> nano> interesting, d5c-d60 just after color/depth buffer setup, already seen that on nv20 and higher, not on nv15 and older
01:42 < Gloubi> ok , sorry mayb e i will  compile X with sdl support the next time
01:43 < airlied> Gloubi: you compile SDL standalonoe.. you are also missing xlibs devel packages..
01:44 < pmdata> nano> also 2bc with some enable stuff, and 3f8,3fc to disable/set something to 0
01:44 < pmdata> I need more nv17 dumps
01:45 < Gloubi> ok , will need to install sdl with x11 support and xlib dev files
01:45 < pmdata> nano> could you post a glxinfo?
01:46 < nano> pmdata: I'm a CS student, have studied some basic hardware/x86-asm, and it would be fun to dig down in something like this. Just wondering on where to start and read and so on. 
01:46 < nano> pmdata: sure
01:46 < nano> debian.as/~nano/glxinfo.txt
01:47 < pmdata> just read your dump, and try to find a command that trigger some changes in unknown stuff (the one still marked with its 0xnumber/4)
01:49 -!- tibbs|h is now known as tibbs
01:50 < nano> Ok.. so the documents on the wiki is for later. At this time it's only about trying diffrent combinations of functions and what happens. 
01:51 < nano> +see
01:51 < nano> "only"
01:55 < pmdata> according to the glxinfo, nv17 can have 16 bits depth/0 stencil with a 32 bits color buffer
01:59 < nano> Humm.. how to interpret glxinfo?
01:59 < Gloubi> so, good night all, next time i hope will have renouveau  installed
02:00 < nano> I guess you got that from the data table at bottom. But I haven't found any description for what it prints. Actually never cared, but I'm unable to find anything on google. 
02:00 < nano> ok.. oth it's pretty self explainatory. 
02:00 -!- Gloubi [n=Gloubi@ABordeaux-253-1-54-146.w82-125.abo.wanadoo.fr] has quit ["I'll be back"]
02:01 < pmdata> yep, 'dpth' is depth, 'stcl' is stencil, etc...
02:14 -!- pmdata [i=patrice@ANantes-154-1-93-251.w86-210.abo.wanadoo.fr] has quit ["using sirc version 2.211+KSIRC/1.3.11"]
03:16 -!- Duke` [n=gnu@ANantes-251-1-96-167.w86-203.abo.wanadoo.fr] has quit ["Fatal signal: Segmentation Fault"]
03:31 -!- stringfellow [n=stringfe@ip56503c9f.direct-adsl.nl] has quit [Remote closed the connection]
03:55 -!- swany [n=swany@81-234-181-143-o1108.tbon.telia.com] has quit []
04:46 -!- philv [n=bleep@cowpig.ca] has quit [Excess Flood]
04:47 -!- philv [n=bleep@cowpig.ca] has joined #nouveau
05:40 -!- nano [n=nano@h100n2fls33o1016.telia.com] has quit ["leaving"]
06:26 -!- hiyuh [n=hiyuh@ZL050248.ppp.dion.ne.jp] has joined #nouveau
06:30 -!- johill [n=johannes@p5487ED9D.dip.t-dialin.net] has joined #nouveau
06:55 -!- johill [n=johannes@p5487ED9D.dip.t-dialin.net] has quit [Read error: 113 (No route to host)]
08:52 -!- tibbs is now known as tibbs|h
10:25 -!- qfire [n=qfire@dsl093-010-213.cle1.dsl.speakeasy.net] has quit [Read error: 113 (No route to host)]
11:49 -!- EdB [n=EdB@ARennes-251-1-152-124.w86-214.abo.wanadoo.fr] has joined #nouveau
12:24 -!- Duke` [n=gnu@ANantes-251-1-96-167.w86-203.abo.wanadoo.fr] has joined #nouveau
14:31 -!- Shipon [n=Aexoden@207-118-75-166.dyn.centurytel.net] has joined #nouveau
14:33 -!- Aexoden [n=Aexoden@207-118-72-211.dyn.centurytel.net] has quit [Read error: 110 (Connection timed out)]
14:51 < airlied> is there an nvidia 2d regisster list anywhere yet?
14:52 < marcheu> the 2D registers are contained within the 2D objects
14:53 < marcheu> which we document at the same place as 3D objects here : http://nouveau.cvs.sourceforge.net/nouveau/renouveau/nv_objects.h?revision=1.154&view=markup
14:54 < airlied> marcheu: I mean like the modesetting ones ..
14:55 < marcheu> I think lumag_offline was working on the modesetting object ; not sure how it actually works yet (the open driver uses legacy vga registers for that)
14:57 < airlied> ah so there is another modesetting object.. where are we getting that info from? nvidia binary or otherwhere..
14:58 < marcheu> what do you mean "another" ?
14:58 -!- ElAngelo [n=ElAngelo@merlin.ugent.be] has left #nouveau []
14:58 < airlied> as in non legacy VGA register..
14:58 < marcheu> it's not easy to get information about it since it's at least partially kernel-side
14:58 < airlied> they have extended the legacy VGA..
14:58 < airlied> ssame as everyone did back in the day..
14:59 < marcheu> the legacy vga registers are not exposed as an object
14:59 < marcheu> since they're.. well ... the legacy vga regs :)
14:59 < airlied> I'm still not up to speed on what is meant by object in this case..
14:59 < airlied> is there not just a large flat register space?
15:00 < marcheu> no
15:00 < airlied> ah so what does the 16MB MMIO mapping contain
15:00 < airlied> freel free to tell me to read if this is written somewhere..
15:00 < marcheu> which one ? our renouvecau mapping ?
15:01 < airlied> the nvidia has a PCI BAR which is 16MB containing MMIO regs..
15:01 < airlied> or at least mine has..
15:01 < marcheu> yeah, that's the stuff I map in renouveau by hand
15:01 < airlied> ssso how to objects relate to that space?
15:01 < marcheu> so yes, that area contains all objects and all commands
15:01 < airlied> I'm on ppc so I can't run nvidia or renouveau presently..
15:02 < marcheu> I think an osx port might be pretty easy
15:02 < marcheu> plus it'd be very very interesting because 1. it's big endian 2. the driver is different and might show new commands
15:03 < airlied> well I'm going to try and get some 2d features working first..
15:03 < marcheu> do you have osx on the quad ?
15:03 < airlied> yup.. but I've no idea how it works :)
15:04 < airlied> but I do know how to write a register dumper I think and then I can at least dump the mode regs..
15:04 < airlied> and perhaps get dual-head ideas written down..
15:04 < marcheu> yeah
15:04 < marcheu> we're also interested in the fifo contents
15:05 -!- airlied_ [n=airlied@ppp12-128.lns2.syd7.internode.on.net] has joined #nouveau
15:05 < airlied_> just rebooting other server...
15:05 < marcheu> so, about the objects
15:05 < marcheu> each fifo can have 8 objects bound to it
15:06 < airlied_> how many fifos?
15:06 < marcheu> and then you can use the object's commands by issuing fifo writes at a given offset for one of these 8 objects
15:06 < marcheu> how many fifos depends on the card, I think geforces have 32
15:06 < marcheu> but the original riva128 has 8
15:06 < marcheu> but I'm not sure there's even a hw limit
15:07 < marcheu> since the switching code seems to just point the fifo at different places when switching
15:07 < airlied_> so what does an object consist off?
15:07 < marcheu> it's a set of registers for a given functionality, like drawing triangles
15:08 < marcheu> see nv_objects.h, for example the object NV04_DX5_TEXTURED_TRIANGLE
15:08 < airlied_> yes got it..
15:09 < marcheu> it's just about grouping together the commands that are used to achieved a textured triangle
15:10 < marcheu> with the added advantage that each card that has this object implemented will react the same, so that card versioning is made a bit easier
15:12 -!- swany [n=swany@81-234-181-143-o1108.tbon.telia.com] has joined #nouveau
15:13 < airlied_> so that DX5 textured triangle object, how big?? is it or what does it contain?
15:13 < marcheu> the object contains all the NV04_DX5_TEXTURED_TRIANGLE_* commands
15:13 < airlied_> I'm still a bit lost I think, but if I have say object DX5_TT bind to a fifo
15:13 < airlied_> ahh I think I get it now..
15:13 < marcheu> so you can judge its size for yourself by looking at the header :)
15:14 < marcheu> I think we've got that one pretty complete
15:14 < airlied_> so the info in nv_objects.h are the registers for that object..
15:14 < marcheu> well, there's also #define NV04_DX5_TEXTURED_TRIANGLE			0x00000054
15:14 < marcheu> which means NV04_DX5_TEXTURED_TRIANGLE is object #54
15:14 < airlied_> okay I don't suppose we could perhaps comment the object numbers separate from the internals of the object?
15:15 < airlied_> either in the object define or just //OBJECT
15:15 < marcheu> yeah, well I guess it's pretty clear if you're used to it
15:16 < airlied_> can you add me to the sf.net project
15:16 < airlied_> airlied is my sf.net username as well.
15:21 -!- shenki [n=shenki@ppp167-129.lns3.adl4.internode.on.net] has joined #nouveau
15:22 < marcheu> done
15:23 < airlied_> excellent... I'll try not to braek too much :-) but I expect I'll be mainly working on DDX things in git :-)
15:24 < marcheu> well, a renouveau osx port & some dumps would be wonderful :)
15:25 < airlied_> marcheu: I shall consider that at some point I just looked at the code and it probably isn't that impossible...but I'm no OSX developer :-)
15:26 < marcheu> if you can just map the fifo & some regs, that's enough. I guess the rest of the code is pretty portable
15:30 -!- airlied [n=airlied@skynet.skynet.ie] has quit [Read error: 110 (Connection timed out)]
15:35 -!- airlied [n=airlied@skynet.skynet.ie] has joined #nouveau
15:47 -!- airlied_ [n=airlied@ppp12-128.lns2.syd7.internode.on.net] has quit ["leaving"]
15:56 -!- nano [i=nano@debian.as] has joined #nouveau
17:52 -!- Duke` [n=gnu@ANantes-251-1-96-167.w86-203.abo.wanadoo.fr] has quit [Connection timed out]
17:53 -!- Duke` [n=gnu@ANantes-251-1-81-125.w86-203.abo.wanadoo.fr] has joined #nouveau
18:28 -!- hiyuh [n=hiyuh@ZL050248.ppp.dion.ne.jp] has quit ["Leaving"]
19:47 -!- stringfellow [n=stringfe@ip56503c9f.direct-adsl.nl] has joined #nouveau
20:28 -!- Gloubi [n=Gloubi@ABordeaux-253-1-59-152.w82-125.abo.wanadoo.fr] has joined #nouveau
20:40 < marcheu> darktama: ping
21:13 -!- pmdata [i=patrice@ANantes-154-1-44-18.w81-53.abo.wanadoo.fr] has joined #nouveau
21:16 < pmdata> hello
21:22 -!- EdB [n=EdB@ARennes-251-1-152-124.w86-214.abo.wanadoo.fr] has quit [Read error: 113 (No route to host)]
21:23 -!- EdB [n=EdB@ARennes-251-1-152-124.w86-214.abo.wanadoo.fr] has joined #nouveau
21:28 < pmdata> marcheu> when you have time, could you make dumps on nv18? I want to see it there is same unknown stuff as nv17 in nv10 tcl engine
21:29 < marcheu> sure, what dumps do you need ?
21:30 < marcheu> be aware though that nv17 and nv18 are identical
21:50 -!- shavenger [n=sha@swarm.ulb.ac.be] has joined #nouveau
21:51 < pmdata> if you can make a dump for all tests, it should be ok
21:51 < marcheu> alright
21:51 < pmdata> nano just made a few, and I saw unknown stuff
21:55 < nano> Should I make a new dump with all tests enabled?
21:55 < marcheu> we could both do one, but I bet they'll be the same
21:56 < marcheu> pmdata: btw the nv18 uses some nv17 object(s) so I actually expect differences with nv15
21:56 < marcheu> pmdata: OTOH I can try on the nv11, that might show interesting stuff
21:59 < marcheu> hmm what happened with the makefile
22:00 < pmdata> then yes, marcheu with nv11 and nano with its nv17
22:00 < pmdata> but I would like one file per test function please
22:01 < marcheu> hmm -lGL used to bring X stuff along
22:03 < pmdata> ??? dynamically loading libGL should not give any problem, SDL doing all the X stuff
22:04 < marcheu> yeah, except the nvctrl stuff needs Xext
22:04 < marcheu> and since SDL/GL/whatever used to bring it we didn't need -lXext, but now we do
22:04 < pmdata> ah
22:14 < nano> hmm.. maybe some automatic logging ?
22:15 < nano> Got to eat now though.
22:15 < marcheu> hmm one file per test...
22:15 < marcheu> life's hard :)
22:16 < nano> nah, redirect stdio to a file in the beginning of each test.
22:17 < marcheu> well, you've got to give meaningful names to each test file... so it's not as easy as it sounds
22:18 < pmdata> name of the function should suffice :)
22:18 < nano> test_enables.$(pid).txt
22:18 < nano> or timestamp
22:20 -!- gourgou [n=gourgou@dra38-2-89-80-66-67.dsl.club-internet.fr] has joined #nouveau
22:23 < marcheu> pmdata: I'll add common code that does that
22:23 -!- gourgou [n=gourgou@dra38-2-89-80-66-67.dsl.club-internet.fr] has left #nouveau ["bye"]
22:24  * pmdata was thinking about it :)
22:33 < marcheu> hey, I said it first :)
22:36 -!- shavenger is now known as shavengerAway
22:38 < marcheu> hehe, the nv11 is softquadroable
22:40 < pmdata> hum, with 1.0-5336 I can't load libGL, and when I put full path to it, I got floating point exception error
22:42 < pmdata> maybe it is broken on 1.0-5xxx, will have to install 6xxx then
22:53 < pmdata> is the X mouse cursor a hardware sprite?
22:53 < marcheu> yes
22:53 < marcheu> it's like a small overlay
22:54 < pmdata> there should be an object for it then
22:55 < marcheu> I don't know, we don't need it right now
22:59 < pmdata> is twinview only a second ramdac to output different stuff to 2 monitors?
22:59 < marcheu> a 2nd crtc
23:00 < pmdata> so an extra object maybe for this one?
23:01 < marcheu> the haiku driver does dual head just fine without an extra object
23:04 < pmdata> it does not mean there is none
23:04 < marcheu> yeah, just like for the cursor
23:12 -!- pmdata [i=patrice@ANantes-154-1-44-18.w81-53.abo.wanadoo.fr] has quit ["using sirc version 2.211+KSIRC/1.3.11"]
23:20 < nano> marcheu: did you take care of logging to diffrent files?
23:24 -!- pmdata [i=patrice@ANantes-154-1-44-18.w81-53.abo.wanadoo.fr] has joined #nouveau
23:25  * pmdata running 1.0-6106 now
23:27 < nano> stupid question, what's the reason of trying out older versions of drivers?
23:28 < pmdata> to compare stuff
23:29 < pmdata> but once again, renouveau can not load libGL
23:31 < pmdata> this time, even giving full path does not work
23:33 < pmdata> do you all run 1.0-7xxx ?
23:35 < nano> 1.0-8762
--- Log closed Срд Авг 02 00:00:04 2006