--- Log opened Вск Июл 23 00:00:56 2006
00:26 -!- pmdata_tv is now known as pmdata
00:41 < pmdata> marcheu> when trying nvclock, it dumped the bios, do you need it?
00:41 < marcheu> pmdata: ddl needs it
00:42 -!- EdB [n=EdB@ARennes-251-1-39-55.w81-250.abo.wanadoo.fr] has quit ["Konversation terminated!"]
00:43 < pmdata> I   tried runnin nvclock -Q before X, but it does not change anything for me
00:43 < marcheu> hmm, strange
00:43 < marcheu> what's your pci id ?
00:45 < pmdata> 10de:0151
00:46 < marcheu> I've got a 0152 and it works fine
00:46 < pmdata> does the nv kernel driver must be loaded?
00:46 < marcheu> it doesn't change the glxinfo output for you ?
00:46 < pmdata> no
00:46 < marcheu> yes it has to
00:46 < marcheu> I think
00:47 < marcheu> at least, that's how I did it
00:48 < pmdata> I remember it displays a line about some fbdevext
00:49 < marcheu> well, nvclock needs the driver loaded or running as root
00:49 < marcheu> did you try running as root before X startup on a fresh boot ?
00:51 < pmdata> pextdev that's it
00:51 < pmdata> yes for root
00:51 < pmdata> I think I will try again in a few minutes
00:51 < marcheu> so, if you run it undex X, does it work ?
00:51 < pmdata> when I'm ready for a reset
00:51 < pmdata> nvclock and nvclock -i works, did not try -Q
00:51 < marcheu> you can do both, but if you run it once X is started it's too late to enable the quadro features from xorg.conf
00:52 < marcheu> but you should still get glxinfo seeing a quadro (and hopefully acting like a quadro)
00:57  * pmdata reading the quadro vs gf thread on nvnews
01:00 < pmdata> could you make a dump of your quadro glxinfo, so I see the diff I should get ?
01:00 < marcheu> hmm, that'd require me to switch cards
01:00 < marcheu> but maybe I've got it somewhere
01:05 < marcheu> but that _only_ changes the "OpenGL renderer string:"
01:07 -!- lumag_of1line [n=mitya@chimpanzee.school.ioffe.ru] has joined #nouveau
01:07 -!- Topic for #nouveau: http://nouveau.freedesktop.org | open source 3D acceleration for nvidia cards | http://perso.orange.fr/patrice.mandin/images/nv-kitten.jpg
01:07 -!- Topic set by marcheu [] [Fri Jul 14 19:33:43 2006]
01:07 [Users #nouveau]
01:07 [@ChanServ] [ airlied ] [ ddl          ] [ marcheu   ] [ pmdata      ] 
01:07 [ _Demo_  ] [ blx     ] [ etzel        ] [ megaloutre] [ qfire_away  ] 
01:07 [ Aexoden ] [ Colbert ] [ lumag_of1line] [ nano-     ] [ stringfellow] 
01:07 [ ag      ] [ darktama] [ lumag_offline] [ philv     ] 
01:07 -!- Irssi: #nouveau: Total of 19 nicks [1 ops, 0 halfops, 0 voices, 18 normal]
01:07 -!- Channel #nouveau created Sun Jun 25 07:52:56 2006
01:07 -!- Irssi: Join to #nouveau was synced in 6 secs
01:07 < marcheu> well, I'm impressed
01:07 < pmdata> I should try to write better test functions
01:08 < marcheu> nv30 still have registers for register combiners ??
01:08 < pmdata> yep
01:08 < pmdata> well, at least same as nv20
01:08 < marcheu> that sounds stupid to me, they should have used the fragment program stuff
01:09 < marcheu> but maybe it does it internally
01:11 < marcheu> lol the mail flow doesn't stop
01:12 < pmdata> yep, you'll got numerous 1-line diff/commits :)
01:13 < pmdata> checkout renouveau if you want the final result
01:13 < marcheu> no, I always read all the diffs
01:13 < marcheu> that keeps me up to date on all fronts, and that helps spotting potential mistakes
01:14 < pmdata> I made some, you'll find numerous "oops..." commits messages
01:14 < marcheu> yep
01:15 < marcheu> so if you can put your bios image somewhere for ddl 
01:15 < marcheu> that'd be nice
01:15 < marcheu> he's in US right now and seems he only has net acces in the evening (so during the french night)
01:18 < pmdata> ddl> http://pmandin.atari.org/download/10de_0151_bios.rom
01:21 < pmdata> besides the viewport size change, fsaa setup is done very early (maybe even before creating the context)
01:21 < pmdata> I think that maybe it lies with the video overlay stuff (which does resize also)
01:21 < marcheu> well, there's the BES (back end scaler)
01:22 < pmdata> yep, I think it is done with it
01:23 < pmdata> the 3d engine just renders a bigger image, which is then scaled down
01:23 < marcheu> yes, that's for supersampling
01:23 < marcheu> multisampling is a bit smarter
01:24 < marcheu> it only renders 4 pixels for each pixel internall, averages them and then writes it to a standard-sized buffer
01:25 < pmdata> why with renouveau we don't see those operations (back end scaler) ?
01:25 < marcheu> hmm what do you mean ?
01:26 < pmdata> well, if I enable fsaa for 1.5x1.5 or 2x2, dumps only show me the bigger size of the buffer, nothing more
01:26 < marcheu> ah
01:26 < marcheu> I think buffer swaps are queued in the kernel
01:26 < marcheu> and done asynchronously
01:26 < marcheu> which is the best way to do it, really
01:27 < marcheu> in our case rescaling is a kind of buffer swap
01:27 < pmdata> ah, so doing many swapbuffers() would show it or not?
01:28 < marcheu> no, that'd only queue more swaps
01:28 < pmdata> hm, I also did not try a fullscreen opengl context, to see if flip happens instead of blitcopy
01:29 < marcheu> did you see a buffer swap in the user space fifo already ?
01:29 < marcheu> I don't think you'd see any anyway
01:29 < pmdata> hum, when trying sdl_gl_swapbuffers(), I see some raster copy setup and done
01:29 < pmdata> in windowed mode
01:29 < marcheu> hmm ok
01:29 < marcheu> so maybe the backend scaler cannot be shared, so it has to be done from only one fifo
01:30 < marcheu> I don't know
01:30 < pmdata> hm, I don't know if you can have several hardware overlay opened
01:31 < pmdata> there are many things we don't know :)
01:31 < marcheu> I was thinking we could maybe clone you
02:24 -!- lumag_offline [n=mitya@chimpanzee.school.ioffe.ru] has joined #nouveau
02:24 -!- Topic for #nouveau: http://nouveau.freedesktop.org | open source 3D acceleration for nvidia cards | http://perso.orange.fr/patrice.mandin/images/nv-kitten.jpg
02:24 -!- Topic set by marcheu [] [Fri Jul 14 19:33:43 2006]
02:24 [Users #nouveau]
02:24 [@ChanServ] [ ag     ] [ Colbert ] [ etzel        ] [ megaloutre] [ pmdata      ] 
02:24 [ _Demo_  ] [ airlied] [ darktama] [ lumag_offline] [ nano-     ] [ qfire_away  ] 
02:24 [ Aexoden ] [ blx    ] [ ddl     ] [ marcheu      ] [ philv     ] [ stringfellow] 
02:24 -!- Irssi: #nouveau: Total of 18 nicks [1 ops, 0 halfops, 0 voices, 17 normal]
02:24 -!- Channel #nouveau created Sun Jun 25 07:52:56 2006
02:24 -!- [freenode-info] if you're at a conference and other people are having trouble connecting, please mention it to staff: http://freenode.net/faq.shtml#gettinghelp
02:24 -!- Irssi: Join to #nouveau was synced in 1 secs
03:02 < pmdata> I hate magic constants, especially when I don't know what they are used for
03:02 -!- pmdata [n=patrice@ANantes-154-1-70-172.w86-195.abo.wanadoo.fr] has quit [Remote closed the connection]
03:21 < Colbert> heve something to do this evening, bbl
03:21  * Colbert is away: Colbert
03:30 -!- _Demo_ [n=Demo@modemcable206.154-131-66.mc.videotron.ca] has quit ["Leaving"]
04:30 -!- Aexoden [n=Aexoden@207-118-72-211.dyn.centurytel.net] has quit ["WeeChat 0.1.9"]
05:09 -!- Aexoden [n=Aexoden@207-118-72-211.dyn.centurytel.net] has joined #nouveau
08:13 -!- Colbert [n=colby@adsl-69-211-27-8.dsl.chcgil.ameritech.net] has quit [Remote closed the connection]
13:28 -!- EdB [n=EdB@ARennes-251-1-12-240.w83-195.abo.wanadoo.fr] has joined #nouveau
17:23 -!- johill [n=johannes@p5487DE67.dip.t-dialin.net] has joined #nouveau
17:33 -!- johill [n=johannes@p5487DE67.dip.t-dialin.net] has quit ["Client exiting"]
18:04 -!- sturmflut [n=sturmflu@2001:6f8:992:1337:213:d4ff:feaa:8a9f] has joined #nouveau
18:12 -!- Lumag [n=Lumag@chimpanzee.school.ioffe.ru] has joined #nouveau
18:13 < Lumag> hi
19:10 < megaloutre> 000)))))mall stuff
19:10 < megaloutre> <marcheu> because doing it the other way would create fragmentation
19:10 < megaloutre> mall stuff
19:10 < megaloutre> <marcheu> because doing it the other way would create fragmentation
19:10 < megaloutre> mall stuff
19:10 < megaloutre> <marcheu> because doing it the other way would create fragmentation
19:10 < megaloutre> mall stuff
19:10 < megaloutre> <marcheu> because doing it the other way would create fragmentation
19:10 < megaloutre> mall stuff
19:10 < megaloutre> <marcheu> because doing it the other way would create fragmentation
19:25 -!- Netsplit leguin.freenode.net <-> irc.freenode.net quits: @ChanServ
19:34 -!- Netsplit over, joins: @ChanServ
19:41 < sturmflut> I am not able to compile renouveau because of missing declarations for NV10_TCL_PRIMITIVE_3D_MAGIC_WIDTH and NV10_TCL_PRIMITIVE_3D_MAGIC_HEIGHT
20:03 -!- _Demo_ [n=Demo@modemcable206.154-131-66.mc.videotron.ca] has joined #nouveau
20:17 < sturmflut> I just commented it out now
20:18 < sturmflut> There is an error in objects.c, line 2414 (also in the CVS version): should be "const data_print_store* dp = &field->data[0];"
20:25 -!- Myrizio [n=Myrizio@Linuz.sns.it] has joined #nouveau
21:48 -!- Myrizio [n=Myrizio@Linuz.sns.it] has quit ["I like core dumps"]
22:02 -!- Lumag [n=Lumag@chimpanzee.school.ioffe.ru] has left #nouveau []
22:19 -!- johill [n=johannes@p5487DE67.dip.t-dialin.net] has joined #nouveau
22:49 -!- Lumag [n=Lumag@chimpanzee.school.ioffe.ru] has joined #nouveau
23:05 < Lumag> Hmm... I examined X server startup FIFO. There are some more objects with context = 0x8000000. The problem is that the RAMIN with zero ofsset contains garbage :(
23:08 -!- pmdata [i=patrice@ANantes-154-1-81-58.w81-48.abo.wanadoo.fr] has joined #nouveau
23:10 < Lumag> hi, pmdata
23:13 < pmdata> hello
23:24 < Lumag> Also got some unknown types: 0x46, 0x47, 0x49, 0x4D :)
23:25 < Lumag> All those objects map to zero context offsets, so no context info :(
23:33 < pmdata> can you add them to the nvobjecttypes wiki page?
23:46 < Lumag> yes. I'm examining the rest of log then there will be a commit to objects.c and wiki page :)
23:47 < pmdata> what is your hardware?
23:48 < Lumag> nv06 ---  Vanta :)
23:50 < Lumag> and a PCI nv05 (TNT2 Pro)
23:50 < Lumag> Still searching plain nv04 and nv03 cards.
--- Log closed Пнд Июл 24 00:00:57 2006
