--- Log opened Птн Июл 21 00:00:55 2006
00:02 < marcheu> http://icps.u-strasbg.fr/~marchesin/nvdri/nv28_rc_loop.txt
00:03 -!- Gloubi [n=Gloubi@ABordeaux-253-1-49-108.w82-125.abo.wanadoo.fr] has joined #nouveau
00:03 < marcheu> I forgot one ), so each printf is lacking it. but it takes so much time to run it that you'll add it in your mind :)
00:03 < Gloubi> hi
00:04 < marcheu> hi Gloubi
00:04 < marcheu> Gloubi: http://fr.wikipedia.org/wiki/Gloubiboulga ?
00:04 < Gloubi> nan :)
00:05 < marcheu> pmdata: so if you need any more loop, just ask
00:06 < marcheu> hmm, forgot the GL_ALPHA ones...
00:06 < marcheu> hmm, forgot the GL_RGB ones...
00:09 < pmdata> could you run a very simple test: just define different rc constant colors (there are 2)
00:10 < marcheu> with glCombinerParameter ?
00:13 < pmdata> I don't remember
00:14 < pmdata> it seems you only have 2 input combiners
00:14 < marcheu> we're talking about setting up the GL_CONSTANT_COLOR0_NV & GL_CONSTANT_COLOR1_NV here right ?
00:14 < marcheu> that's done with glCombinerParameter
00:15 < pmdata> yep
00:15 < pmdata> correction: 1 input combiner
00:15 < marcheu> hmm, you mean my loop failed ? or input remapping happens ,
00:15 < marcheu> ?
00:16 < pmdata> no, only input combiner 0 has stuff modified
00:17 < marcheu> hmm, my files contains stuff for GL_COMBINER{0->7}_NV
00:18 < pmdata> yep, but nothing is changed for combiners 1 to 7, only the latest values for combiners 0 are present and keep
00:18 < marcheu> so maybe input remapping happens
00:19 < pmdata> what do you mean? an extra command to select the combiner?
00:20 < marcheu> no, internally in the driver
00:20 < marcheu> maybe it renames the first used combiner to 0
00:22 < pmdata> ah
00:22 < pmdata> anyway, what about the constant color?
00:22 < marcheu> http://icps.u-strasbg.fr/~marchesin/nvdri/nv28_rc_col.txt
00:23 < marcheu> doesn't output anything
00:23 < pmdata> which color did you set?
00:23 < pmdata> what argb values?
00:23 < marcheu> defining GL_CONSTANT_COLOR0_NV as {0.1,0.2,0.3,0.4} and GL_CONSTANT_COLOR1_NV as {0.5,0.6,0.7,0.8}
00:24 < marcheu> it's written at the end
00:24 < marcheu> I'll add it before the test loop
00:27 < marcheu> ok, I'll run the test and go eat, it should upload the file at the end, at the same location
00:28 < marcheu> hopefully with GL_RGB this time
00:43 -!- nano- [i=nano@exodus.xmms.se] has joined #nouveau
00:44 < pmdata> for lighting, it would be useful to have the maths formula and results, to compare between chips
01:05 < marcheu> ok, the file is up now
01:07 < marcheu> pmdata: btw, 1st, 2nd, 3rd and 4th :)
01:10 < pmdata> hum, I see nothing, so they are setup elsewhere
01:10 < pmdata> and it is easier to compare dumps using 'diff -Naur', that's why I prefer them separate
01:36 -!- Gloubi [n=Gloubi@ABordeaux-253-1-49-108.w82-125.abo.wanadoo.fr] has quit ["I'll be back"]
01:42 < marcheu> hey, you can cut them if you like :)
01:43 < marcheu> it'd really be a lot of work to create those diffs separately, and that would be error prone
01:44 < airlied> marcheu: Alan cox was saying that ages ago nvidia released a kernel driver that was obfuscated..
01:44 < marcheu> airlied: the nv04/nv05 one ?
01:44 < airlied> and that he has a start at de-obfuscating it done...
01:46 < marcheu> well, I think nv04/nv05 is mostly with the leaked nv10 header, but if you've got a link I'm interested
01:46 < marcheu> is mostly done
01:47 < airlied> well Alan has a copy at home, nvidia made it disappear at some stage..
01:47 < airlied> I'll get him to get me a copy of what he has..
01:47 < marcheu> well, if it's the nvsdk we all have it I think
01:47 < marcheu> yeah, get his copy, we never know it might be something els
01:47 < marcheu> else
01:53 < darktama> airlied: nice slides, wish I could've seen the presentation :)
01:53 < marcheu> darktama: you should have traveled in airlied's luggage since you're both from .au :)
01:54 < darktama> hehe
01:58 -!- EdB [n=EdB@ARennes-251-1-58-194.w81-53.abo.wanadoo.fr] has quit ["Konversation terminated!"]
02:07 -!- pmdata [i=patrice@ANantes-154-1-108-209.w86-214.abo.wanadoo.fr] has quit ["using sirc version 2.211+KSIRC/1.3.11"]
02:19 < darktama> hmm, whatever address I pass to 0x0210 (frontbuffer offset?) the data appears at that address+0x804 in the fb..
03:10 < darktama> whoops, my bad.. was rendering a quad from (1,1->511,511) instead of (0,0->511,511)
04:07 -!- Lumag [n=Lumag@chimpanzee.school.ioffe.ru] has left #nouveau []
04:29 -!- hiyuh [n=hiyuh@ZL050248.ppp.dion.ne.jp] has joined #nouveau
06:13 -!- johill [n=johannes@p5487E56D.dip.t-dialin.net] has quit [Read error: 113 (No route to host)]
07:05 < airlied> marcheu: if darktama was in my luggage he might not have made it, my luggage lagged me by quite a few hours :-)
07:05 < ddl> heh :)
07:05 < ddl> better than ending up in honolulu ;)
07:06 < darktama> hehe, give me a laptop in the luggage and I'd have been fine :)
07:07 < ddl> heh, last time i was at the luggage band in the airport there where several boxes with biohazard signs standing by the band 
07:07 < ddl> went up to them and read the label: CONTAINS HUMAN BLOOD
07:10 < ddl> whats nice with US airports is that they almost always provide free wlan
07:23 < darktama> ddl: how goes the bios work?
07:25 < ddl> well, im working on it.. but theres alot of stuff to figure out 
07:26 < darktama> yeah I imagine :)  if you need any tests done let me know
08:18 -!- philv [n=bleep@cowpig.ca] has quit ["Lost terminal"]
08:19 -!- philv [n=bleep@cowpig.ca] has joined #nouveau
10:31 -!- xaid [n=Xaid@d199-126-153-40.abhsia.telus.net] has quit ["leaving"]
11:24 -!- EdB [n=EdB@ARennes-251-1-10-113.w83-195.abo.wanadoo.fr] has joined #nouveau
11:30 -!- johill [n=johannes@p5487E56D.dip.t-dialin.net] has joined #nouveau
11:35 -!- johill [n=johannes@p5487E56D.dip.t-dialin.net] has quit [Remote closed the connection]
11:37 -!- johill [n=johannes@p5487E56D.dip.t-dialin.net] has joined #nouveau
13:27 -!- stringfellow [n=stringfe@a80-100-96-220.adsl.xs4all.nl] has joined #nouveau
14:48 -!- Colbert [n=colby@adsl-69-211-27-8.dsl.chcgil.ameritech.net] has joined #nouveau
14:49 < Colbert> Looks  like I've missed a lot of fun the last couple of days.
14:51 < Colbert> changed over to M.K.'s glut instead of freeglut. freeglut is really broken once i looked at the source
14:59 < Colbert> I've made dumps of all test programs on my nv-17 pci if anyone needs info.
15:01 < Colbert> havnt worked on the colorbuffers since my last post a couple of days ago.
15:05 -!- johill [n=johannes@p5487E56D.dip.t-dialin.net] has quit [Remote closed the connection]
15:08 -!- EdB [n=EdB@ARennes-251-1-10-113.w83-195.abo.wanadoo.fr] has quit [Remote closed the connection]
15:11 -!- johill [n=johannes@p5487E56D.dip.t-dialin.net] has joined #nouveau
15:14 -!- johill [n=johannes@p5487E56D.dip.t-dialin.net] has quit [Client Quit]
15:14 -!- johill [n=johannes@p5487E56D.dip.t-dialin.net] has joined #nouveau
15:56 -!- EdB|w [n=EdB@212.234.68.206] has joined #nouveau
16:03 -!- Netsplit zelazny.freenode.net <-> irc.freenode.net quits: philv
16:12 -!- Netsplit over, joins: philv
16:40 -!- hiyuh [n=hiyuh@ZL050248.ppp.dion.ne.jp] has quit ["Leaving"]
17:07 -!- johill_ [n=johannes@p5487FBA9.dip.t-dialin.net] has joined #nouveau
17:14 -!- johill [n=johannes@p5487E56D.dip.t-dialin.net] has quit [Nick collision from services.]
17:14 -!- johill_ is now known as johill
17:17 < Colbert> Something I dont understand. the code I've been playing around with is not that different from the code in test_cull_face()
17:30 < Colbert> I get nice clean dumps from test_cull_face()  but chaos from my code
17:30 < Colbert> forgot to hit enter
17:33 < Colbert> actually, took a closer look at the dumps. the dump from  test_cull_face() is identical to the output from my test_draw_buffer() when using any buffer that dosent exist
18:22 -!- johill_ [n=johannes@p5487FBA9.dip.t-dialin.net] has joined #nouveau
18:26 -!- johill_ [n=johannes@p5487FBA9.dip.t-dialin.net] has quit [Client Quit]
18:27 -!- johill_ [n=johannes@p5487FBA9.dip.t-dialin.net] has joined #nouveau
18:28 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has quit [Nick collision from services.]
18:31 -!- johill_ [n=johannes@p5487FBA9.dip.t-dialin.net] has quit [Client Quit]
18:31 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has joined #nouveau
18:31 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has quit [Read error: 104 (Connection reset by peer)]
18:32 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has joined #nouveau
18:33 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has quit [Read error: 104 (Connection reset by peer)]
18:34 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has joined #nouveau
18:42 -!- johill_ [n=johannes@p5487FBA9.dip.t-dialin.net] has joined #nouveau
18:45 -!- johill_ [n=johannes@p5487FBA9.dip.t-dialin.net] has quit [Remote closed the connection]
18:45 -!- johill_ [n=johannes@p5487FBA9.dip.t-dialin.net] has joined #nouveau
18:51 -!- johill_ [n=johannes@p5487FBA9.dip.t-dialin.net] has quit [Remote closed the connection]
18:51 -!- johill_ [n=johannes@p5487FBA9.dip.t-dialin.net] has joined #nouveau
18:52 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has quit [Nick collision from services.]
18:52 -!- johill_ is now known as johill
19:01 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has quit ["Client exiting"]
19:01 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has joined #nouveau
19:02 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has quit [Remote closed the connection]
19:03 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has joined #nouveau
19:03 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has quit [Read error: 104 (Connection reset by peer)]
19:04 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has joined #nouveau
19:05 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has quit [Remote closed the connection]
19:05 -!- johill [n=johannes@p5487FBA9.dip.t-dialin.net] has joined #nouveau
19:25 -!- johill_ [n=johannes@p5487FBA9.dip.t-dialin.net] has joined #nouveau
19:25 -!- johill_ [n=johannes@p5487FBA9.dip.t-dialin.net] has quit [Remote closed the connection]
20:34 -!- Netsplit zelazny.freenode.net <-> irc.freenode.net quits: ag, stringfellow, blx, philv
20:34 -!- stringfellow [n=stringfe@80.100.96.220] has joined #nouveau
20:34 -!- Netsplit over, joins: ag
20:35 -!- blx [n=x@217.208.244.132] has joined #nouveau
20:48 -!- lumag_offline [n=mitya@chimpanzee.school.ioffe.ru] has joined #nouveau
20:48 -!- Topic for #nouveau: http://nouveau.freedesktop.org | open source 3D acceleration for nvidia cards | http://perso.orange.fr/patrice.mandin/images/nv-kitten.jpg
20:48 -!- Topic set by marcheu [] [Fri Jul 14 19:33:43 2006]
20:48 [Users #nouveau]
20:48 [@ChanServ] [ Colbert] [ johill       ] [ nano-     ] [ stringfellow] 
20:48 [ ag      ] [ EdB|w  ] [ lumag_offline] [ philv     ] [ xaid|work   ] 
20:48 [ blx     ] [ etzel  ] [ marcheu      ] [ qfire_away] 
20:48 -!- Irssi: #nouveau: Total of 14 nicks [1 ops, 0 halfops, 0 voices, 13 normal]
20:48 -!- Channel #nouveau created Sun Jun 25 07:52:56 2006
20:48 -!- Irssi: Join to #nouveau was synced in 3 secs
20:48 -!- airlied [n=airlied@skynet.skynet.ie] has joined #nouveau
20:50 -!- Aexoden [n=Aexoden@207-118-72-211.dyn.centurytel.net] has joined #nouveau
20:54 -!- darktama [n=darktama@gentoo/contributor/darktama] has joined #nouveau
20:54 -!- ddl [i=erikw@montezuma.acc.umu.se] has joined #nouveau
20:58 -!- pmdata [i=patrice@ANantes-154-1-21-130.w81-53.abo.wanadoo.fr] has joined #nouveau
20:58 < pmdata> hello
20:58 < EdB|w> hi
21:02 < pmdata> I wonder if the feature matrix on the wiki should be done per object type instead of per function
21:12 < pmdata> I also see from marcheu nv28 dumps that nv20 tcl engine has most of same commands as nv10 ones, just with a different number (to make room for extra units)
22:11 -!- tuXXX [n=damien@dyn-83-152-190-234.ppp.tiscali.fr] has joined #nouveau
22:12 < tuXXX> hi
22:13 < pmdata> hi
22:13 < tuXXX> I have some C/OpenGL skills and two nvidia cards (6600GT AGP and Go7600 PCIe). Maybe I could help?
22:16 < pmdata> yep, maybe. These cards are nv3x or nv4x?
22:16 < tuXXX> the first is nv4x, the second G70
22:17 < xaid|work> hi, I know C and I did some openGL programming as well, I have a GFX5200 (nv34) and I'd like to help as well.
22:18 < pmdata> so, grab renouveau from cvs, and start running some test function to figure out which hw command triggers corresponding opengl stuff
22:19 < tuXXX> ok, seems to be a good start for me :)
22:19 < xaid|work> pmdata: is there a list of whats already done and for which chipsets?
22:20 < xaid|work> like a grid of what commands were ran on which chipsets and what hw command triggers corrospond to them =)
22:21 < pmdata> yep, see nv_objects.h and objects.c in renouveau source
22:21 < pmdata> hint: start running the test_draw_pixels() only
22:22 < pmdata> at least on my nv15, it setup all default values for an opengl context before drawing
22:23 < xaid|work> I'll try that when I get home today, been using the nv driver for a while since the upgrade to xorg 7.0, so if all goes well when switching to the proprietry nvidia driver, I'll be able to run some tests today
22:23 < xaid|work> thanks for the help.
22:25 < pmdata> example dump for draw_pixels: http://pmandin.atari.org/download/nv10_draw_pixels.txt
22:25 < pmdata> renouveau is the tool we use while running the nvidia driver, to figure out the hw commands
22:27 < pmdata> you just modify renouveau/main.c to uncomment the open stuff you want to test (or even write a new one)
22:27 < pmdata> then you run it, redirecting to a file
22:27 < pmdata> then you add a command or change a parameter in a command
22:28 < pmdata> then you compare the outputs
22:30 < xaid|work> that doesn't sound too bad, but it looks like it'll take a while to understand some of that output
22:34 < pmdata> a part of nv30 tcl engine is already done (also valid for nv40 and g70)
22:35 < tuXXX> what do you mean by 'tcl' ?
22:36 < pmdata> the thing that do math in gpu instead of the cpu
22:36 < pmdata> transform,clipping,lighting
22:36 < tuXXX> ok
22:45 < tuXXX> I see some output I don't understand with test_draw_pixels : I'm modifying the first two args (width and height). I can see NV30_TCL_PRIMITIVE_3D_TX_XY_DIM_UNIT (width=, height=), value is changing is but the width value is only multiple of 2 : 4,5 -> 4,5 : 5,5 -> 6,5 : 6,5 -> 6,5 : 7,3 -> 8,3 : 7,4 -> 8,4 : 7,5 -> 8,5
22:46 < tuXXX> but another value is right
22:47 < pmdata> TX means it is a texture, and maybe the hardware can only supports texture with a width multiple of 2
22:48 < pmdata> be careful that the hardware may not always suits opengl
22:49 < pmdata> so there are sometimes conversion
22:51 < pmdata> could you make your dump available at some url? on your website or using http://pastebin.ca for example?
22:51 < tuXXX> yes
22:59 < tuXXX> http://pastebin.ca/95083 http://pastebin.ca/95086 http://pastebin.ca/95087 http://pastebin.ca/95088 http://pastebin.ca/95091 http://pastebin.ca/95092
23:00 < pmdata> so many? :)
23:01 < tuXXX> yes :) changing width and height 4,5 5,5 6,5 7,5 7,4 7,3
23:02 -!- EdB|w [n=EdB@212.234.68.206] has quit ["Parti"]
23:05 < pmdata> hum, I thought the real 3d stuff would be easier for you, because glDrawPixels() is done using a texture
23:05 < pmdata> so take for example http://pastebin.ca/95092
23:05 < pmdata> line 114,115 you have the viewport size in the 16 high bits of the long (0x200 = 512 for both)
23:06 < darktama> viewport is at 0xA00 on >=nv30 :)
23:06 < pmdata> yep, I just want to explain him
23:06 < pmdata> and just above, lines 103-104, you have commands 200,204 with same values
23:07 < pmdata> so obviously you also have w and h there
23:07 < pmdata> try change the size in main.c, in the call to sdl_setvideomode, and compare the output
23:08 < tuXXX> ok
23:10 < pmdata> or line 172, you have a 4 * 8bit value, being 0 or 1
23:10 < pmdata> well I think it is the glColorMask()
23:10 < pmdata> were you can enable/disable each of r,g,b,a writing to the color buffer
23:12 < pmdata> so you can add a glColorMask(GL_FALSE,GL_FALSE,GL_TRUE,GL_TRUE) before the glDrawPixels() call, and compare the output
23:13 < pmdata> opengl says it is GL_TRUE by default for all values 
23:15 < pmdata> another possibility to find stuff, take for example line 299 (it says cull_face, whatever, think we don't know)
23:15 < pmdata> the value is 0x0405
23:16 < pmdata> if you search 0405 in GL/gl.h header file (or GL/glext.h if it is from an extension), you'll find GL_BACK
23:17 < pmdata> and GL_BACK is related to some polygon stuff
23:17 < pmdata> then you read in the opengl doc where GL_BACK is used for polygon stuff
23:18 < pmdata> and you find glPolygonMode() and glCullFace()
23:19 < pmdata> then a simple test with either command will tell you which it is
23:20 < pmdata> am I not flooding the chan?
23:21 < tuXXX> lol not really :)
23:23 < pmdata> was I clear enough?
23:23 < tuXXX> For example, I see some cells with "Yes" on the matrix, each of these is a categorie we need to find informations about?
23:24 < pmdata> yes
23:24 < pmdata> however, I think the feature matrix should be updated, for example to match the nv object types
23:27 < pmdata> it's better to run renouveau, and check if there is any unknown command
23:29 < tuXXX> ok I'll try :)
23:34 < pmdata> well, just with the dump you did, there is quite some
23:34 < tuXXX> yeah
23:34 < tuXXX> ^^
--- Log closed Сбт Июл 22 00:00:55 2006
