--- Log opened Птн Авг 18 00:00:05 2006
--- Day changed Птн Авг 18 2006
00:00 < pmdata> can you post a dump url?
00:01 < KoalaBR> I need to redump, because I split the whole test in smaller parts
00:01 < pmdata> "yesterday: 23:57 < pmdata> nv40[0x1ee8] seems to be the point sprite parameter"
00:01 < KoalaBR> Yes
00:02 < KoalaBR> regl.TexEnvi(GL_POINT_SPRITE_NV, GL_COORD_REPLACE_NV, GL_FALSE); leads to
00:02 < KoalaBR> NV30_TCL_PRIMITIVE_3D      [0x1ee8/4] = 0x00000801 | UNKNOWN = 00000801
00:03 < KoalaBR> NV30_TCL_PRIMITIVE_3D      [0x1ee8/4] = 0x00000001 | UNKNOWN = 00000801
00:03 < KoalaBR> Last one is the correct one
00:03 < KoalaBR> 8 is seen if you use GL_TRUE
00:03 < KoalaBR> Bit 0 is only 0 if setup runs / reset perhaps
00:05 < KoalaBR> Bit 1 is set for GL_R and Bit 2 = GL_S
00:05 < KoalaBR> That's what I found out, but wait I'll post a complete test
00:07 < KoalaBR> pmdata: http://sh.nu/p/2759
00:08 -!- sturmflut [n=sturmflu@2001:6f8:992:1337:213:d4ff:feaa:8a9f] has joined #nouveau
00:09 < KoalaBR> pmdata: was that all you were looking for in that test?
00:15 < KoalaBR> pmdata: Please look here: http://sh.nu/p/2760
00:15 < KoalaBR> pmdata: the first dump, was with GL_FALSE and did not have 1ee8 commands ...
00:15 < KoalaBR> This one is with GL_TRUE and has 1ee8 in it
00:19 < KoalaBR> Well, that's it for today. Good night all
00:20 -!- KoalaBR [n=KoalaBR@port-83-236-15-108.dynamic.qsc.de] has quit ["ChatZilla 0.9.61 [Mozilla rv:1.7.13/20060417]"]
00:22 < pmdata> don't forget that 1ee8 maybe set to disable in test_startup
00:24 < pmdata> try to copy the nv20_tcl_primitive_3d_point_sprite in nv_objects.h/objects.c and put in in nv30_tcl as 1ee8 to see if it matches, I think so
00:24 < pmdata> arg, he's gone
00:25 < Duke`> I'll read the log
00:26 < pmdata> duke> do you have nv30 or nv40?
00:29 < Duke`> nv30 but not mounted at this time (nv10 installed)
00:37 < pmdata> can you softquadro it?
00:41 < Duke`> the nv10 I failed, the nv30 I haven't tried yet
00:55 < atcl> hi there, does somebody here have a dump, in which the program is rendering a texture with a fragment shader, i like to find out how that works, and may be how the gpu is moved to use the fp. 
00:55 < pmdata> there is a test_nv_texture_shader in renouveau
00:55 < atcl> ok
00:55 < atcl> ill check
00:55 < pmdata> just look in the main.c test_* calls
00:56 < pmdata> there is also a glsl one
00:59 < atcl> ok iam probably to stupid finding it.., 
01:02 < pmdata> there are also some test_fragprog in it
01:02 < atcl> yes, i found them in the main,...
01:03 < darktama> if you're on a >=nv40 card, just run test_tex()
01:03 < atcl> oki
01:03 < atcl> i do, actually, (nv43)
01:04 < darktama> ok, then it'll be using a fragment program regardless of whether you tell it to or not :)
01:06 < atcl> got it now,
01:06 < atcl> thx
01:07 < darktama> np
01:09 -!- sturmflut [n=sturmflu@2001:6f8:992:1337:213:d4ff:feaa:8a9f] has quit [Remote closed the connection]
01:12 -!- stringfellow [n=stringfe@ip56503c9f.direct-adsl.nl] has quit [Remote closed the connection]
01:17 -!- atcl [n=Universe@Lb465.l.pppool.de] has quit [Read error: 104 (Connection reset by peer)]
01:20 -!- pmdata [i=patrice@ANantes-154-1-68-160.w86-195.abo.wanadoo.fr] has quit ["using sirc version 2.211+KSIRC/1.3.11"]
01:23 -!- predatorfreak [n=predator@adsl-69-213-80-253.dsl.sfldmi.ameritech.net] has joined #nouveau
01:31 -!- maxtoo [n=maxtoo@berryx.homedns.org] has quit [Remote closed the connection]
01:54 -!- Duke` [n=gnu@ANantes-251-1-142-44.w86-210.abo.wanadoo.fr] has quit [Remote closed the connection]
02:22 -!- K [i=hazel@tor/session/external/x-bb78ab9175d1edbb] has quit [Remote closed the connection]
02:55 < ddl> hey!
02:55 < ddl> home again :)
03:00 < marcheu> hey ddl, nice to have you back
03:02 < ddl> thanks :)
04:36 -!- hiyuh [n=hiyuh@ZL050248.ppp.dion.ne.jp] has joined #nouveau
05:14  * Myrizio will now try to softquadro its card...
05:15 -!- Myrizio [n=Myrizio@host222-102.pool80104.interbusiness.it] has quit ["Goodbye Ruby Tuesday (Perl Friday)"]
05:26 -!- Myrizio [n=Myrizio@host124-101.pool80104.interbusiness.it] has joined #nouveau
05:36  * Myrizio was unable to softquadro his card (10de:0110 (rev a1)), MISSION FAILED.
07:07 -!- Myrizio [n=Myrizio@host124-101.pool80104.interbusiness.it] has quit ["Goodbye Ruby Tuesday (Perl Friday)"]
07:08 -!- predatorfreak [n=predator@adsl-69-213-80-253.dsl.sfldmi.ameritech.net] has quit ["<blank>"]
08:43 -!- stillunknown_ [n=madman20@82-168-177-167.dsl.ip.tiscali.nl] has joined #nouveau
08:46 -!- Netsplit calvino.freenode.net <-> irc.freenode.net quits: stillunknown, Aexoden, ddl, cptn, hiyuh
08:47 -!- Netsplit over, joins: ddl
08:55 -!- Aexoden [n=Aexoden@207-118-82-54.dyn.centurytel.net] has joined #nouveau
08:55 -!- cptn [n=jw@217-162-119-116.dclient.hispeed.ch] has joined #nouveau
09:03 -!- hiyuh [n=hiyuh@ZL050248.ppp.dion.ne.jp] has joined #nouveau
09:57 -!- predatorfreak [n=predator@adsl-69-213-80-253.dsl.sfldmi.ameritech.net] has joined #nouveau
10:46 -!- EdB [n=EdB@ARennes-251-1-16-116.w83-195.abo.wanadoo.fr] has joined #nouveau
10:47 -!- predatorfreak [n=predator@adsl-69-213-80-253.dsl.sfldmi.ameritech.net] has quit ["<blank>"]
11:59 -!- Unavowed [n=silent@host81-151-27-126.range81-151.btcentralplus.com] has joined #nouveau
12:16 -!- Duke` [n=gnu@ANantes-251-1-102-30.w86-203.abo.wanadoo.fr] has joined #nouveau
12:40 -!- maxtoo [n=maxtoo@berryx.homedns.org] has joined #nouveau
13:29 -!- Duke` [n=gnu@ANantes-251-1-102-30.w86-203.abo.wanadoo.fr] has quit ["Fatal signal: Segmentation Fault"]
13:33 -!- Duke` [n=gnu@ANantes-251-1-102-30.w86-203.abo.wanadoo.fr] has joined #nouveau
14:06 -!- Myrizio [n=Myrizio@host10-101.pool80104.interbusiness.it] has joined #nouveau
17:26 -!- Myrizio_ [n=Myrizio@host115-101.pool80104.interbusiness.it] has joined #nouveau
17:32 -!- Myrizio [n=Myrizio@host10-101.pool80104.interbusiness.it] has quit [Read error: 113 (No route to host)]
18:13 -!- K [i=hazel@tor/session/external/x-fd67a977a03b6c16] has joined #nouveau
19:37 -!- hiyuh [n=hiyuh@ZL050248.ppp.dion.ne.jp] has quit ["Leaving"]
19:37 -!- Unavowed [n=silent@host81-151-27-126.range81-151.btcentralplus.com] has quit ["leaving"]
19:40 -!- stringfellow [n=stringfe@ip56503c9f.direct-adsl.nl] has joined #nouveau
20:00 -!- littlesniper [n=guillaum@68.174.97-84.rev.gaoland.net] has joined #nouveau
20:04 -!- pmdata [i=patrice@ANantes-154-1-58-15.w81-53.abo.wanadoo.fr] has joined #nouveau
20:43 -!- Myrizio_ [n=Myrizio@host115-101.pool80104.interbusiness.it] has quit ["Goodbye Ruby Tuesday (Perl Friday)"]
20:43 -!- pmdata [i=patrice@ANantes-154-1-58-15.w81-53.abo.wanadoo.fr] has quit ["using sirc version 2.211+KSIRC/1.3.11"]
20:45 < darktama> mm, quiet in here tonight :)
20:45 < marcheu> hey darktama 
20:46 < marcheu> darktama: are you currently doing anything with drm/ddx ? otherwise I'll start to break as much stuff as I can as quickly as possible
20:46 < marcheu> in as many ways as I can find
20:47 < darktama> nope, I've started work on >=NV30 EXA Composite stuff; but until we get 3D going it's useless anyway
20:47 < darktama> so, break away :)
20:48 < marcheu> do you plan to do it with the 3D engine ?
20:48 < darktama> yup
20:48 < marcheu> because the alpha blit available on nv10+ is premultiplied
20:48 < marcheu> I was hoping we could find one that isn't, and extend support down to nv10
20:49 < marcheu> btw does my latest drm work for you ?
20:49 < darktama> I recall reading on the Xorg ML at some stage that the original writer (Lars?) had ideas on how to do it, maybe we should contact him
20:49 < darktama> I haven't tried it yet actually
20:49 < marcheu> yeah, I met lars at fosdem
20:50 < marcheu> maybe we could fire up a couple of emails, he doesn't hang on irc much those days..
20:51 < darktama> does the current composite hook work for you?  it hangs e17 very quickly on my card
20:51 < marcheu> it's deadly slow
20:51 < marcheu> and it timeouts very often
20:51 < marcheu> and, yeah, crashes after some seconds
20:52 < marcheu> we still have to fix that stoopid timeout
20:52 < darktama> hm, perhaps we should figure out the FIFO error reporting ASAP.. might give us a better idea of where these issues are
20:53 < marcheu> hmm, it's fixed in cvs ?
20:53 < marcheu> darktama: you did that ?
20:53 < darktama> error reporting?
20:54 < marcheu> the NVDoSync
20:54 < darktama> oh, I just made it kill X if the timeout is too long.. I got sick of rebooting
20:54 < marcheu> at least the timeout is not CPU-dependent anymore
20:55 < marcheu> I really think my drm update is the culprit for my libdri crashes
20:58 < darktama> mm, I'll update from cvs and see what happens here
21:07 < darktama> <darktama> mm, something is borked
21:07 < darktama> <darktama> XAA shows a garbled screen, then hangs
21:07 < darktama> <darktama> EXA shows the X grid thingo, then hangs
21:07 < darktama> <darktama> oops, wrong channel :S
21:07 < darktama> just typed that in #winehackers by mistake...
21:07 < marcheu> hehe, yeah I can't even get that far
21:08 < marcheu> I'll se what I did wrong in there
21:14 -!- pmdata [i=patrice@ANantes-154-1-58-15.w81-53.abo.wanadoo.fr] has joined #nouveau
21:17 < pmdata> hello
21:18 < Duke`> hi
21:24 < darktama> marcheu: [87155.258559] [drm] DMA command buffer is 1024KiB at 0xfff00000(VRAM)
21:24 < darktama> that could be a problem, my vram is at 0xE0000000.
21:26 < marcheu> yeah, same happens here
21:27 < marcheu> but
21:29 < darktama> hm, it looks like the code that initialises fb_usable_size is #if 0'd out.. the cmdbuf base is calculated using it
21:32 < marcheu> yeah, that's what I was thinking :)
21:33 < marcheu> it was dumb on my part to update everything at once - drm, ddx, Xorg...
21:34 < darktama> mm, makes it hard to find bugs that way :)
21:37 < darktama> so, I guess for now we should leave it at BIOS setup mem_amount()-(560*1024) etc. until we can use the mman to alloc it
21:38 < marcheu> hmm ?
21:39 < marcheu> ah ok
21:44 < darktama> ok, that fixed it for me
21:45 < darktama> EXA at least, didn't bother trying XAA
21:45 < marcheu> cool, I'll try to move X to use the new memory manager in the next couple of hours
21:45 < marcheu> see what happens...
21:45 < darktama> can we do that yet?  I didn't think EXA had the hooks to let the driver alloc memory itself?
21:46 < marcheu> well... I'll put in a hack to allocate a big exa chunk
21:46 < marcheu> and also have the front buffer, cursor and friends allocated through drm
21:46 < darktama> ah, good idea.
21:46 < marcheu> btw any idea why our DDX has that 256Mb limit ?
21:47 < darktama> which limit?
21:47 < marcheu> on vidram
21:47 < darktama> no idea, where is that?
21:47 < marcheu> that's from memory let me grep..
21:47 < marcheu> nv_setup.c
21:48 < marcheu>     if(pNv->RamAmountKBytes > 256*1024)
21:48 < marcheu>         pNv->RamAmountKBytes = 256*1024;
21:49 < marcheu> oh also, I'm making all offsets 64 bit
21:49 < marcheu> that'll help a lot in the future, for a number of reasons
21:49 < darktama> 2005-08-04  Alan Coopersmith  <alan.coopersmith@sun.com>
21:49 < darktama> 	* programs/Xserver/hw/xfree86/drivers/nv/nv_setup.c:
21:49 < darktama> 	Bug #3987 <https://bugs.freedesktop.org/show_bug.cgi?id=3987>
21:49 < darktama> 	Patch #3264 <https://bugs.freedesktop.org/attachment.cgi?id=3264>
21:49 < darktama> 	Workaround for 512 Meg video cards (can only map 256 Meg of it).
21:49 < darktama> 	(Mark Vojkovich)
21:50 < marcheu> "can only map 256 Meg of it" ??
21:50 < darktama> yes, it's a bit vague.. possible like radeon's max-128MB CPU accessible limit?
21:50 < darktama> possibly*
21:50 < marcheu> wel... radeon has a hw limitation
21:50 < marcheu> maybe geforces have one as well..
21:51 < marcheu> if that's the case, we have to take it into account in the memory allocator, because X likes to have pixmaps mmaped, while we don't care for 3D
21:52 < marcheu> hmm we can get around by adding a 3rd heap with non-mapable vram
21:53 < darktama> With EXA direct access is wrapped by PrepareAccess()/FinishAccess()?  Can we use that to move pixmaps around if it's in a non-accessible area?
21:54 < marcheu> yeah, if you like moving things around, sure :)
21:54 < marcheu> but the other way is allocating mmapable ram if possible
21:55 < darktama> that's probably the better way actually :)
21:55 < marcheu> I'm not starting to add bounce buffers for VRAM anyway :)
21:55 < darktama> hehe
21:56 < darktama> so, on the topic of the ddx.. can we kill XAA? *hopeful glare*
21:56 < marcheu> hmm, you don't have any EXA crashes ? :)
21:57 < marcheu> did you try to VT switch ? ;)
21:57 < marcheu> ah wait, you've got a 6600 right ?
21:57 < darktama> 6800GT.. I *think* VT switching worked
21:57 < marcheu> wow
21:57  * darktama tries
21:57 < marcheu> I have to fix stuff here before this works
21:58  * marcheu says bye bye to darktama 
21:58 < darktama> yup, I'm still here.. survived one VT switch..
21:58  * darktama tries again
21:58 < marcheu> ah well
21:58 < darktama> just survived a couple.. now back to nvidia's driver so I can have both screens back :(
21:59 -!- predatorfreak [n=predator@adsl-69-213-80-253.dsl.sfldmi.ameritech.net] has joined #nouveau
22:08 -!- lumag_offline [n=mitya@chimpanzee.school.ioffe.ru] has joined #nouveau
22:08 -!- Topic for #nouveau: http://nouveau.freedesktop.org | open source 3D acceleration for nvidia cards | card dumps at http://nouveau.sourceforge.net/tests/ | http://perso.orange.fr/patrice.mandin/images/nv-kitten.jpg
22:08 -!- Topic set by sturmflut [] [Sun Aug 13 17:02:24 2006]
22:08 [Users #nouveau]
22:08 [@ChanServ] [ ajmitch ] [ Duke`        ] [ marcheu] [ pmdata       ] [ tibbs] 
22:08 [ Aexoden ] [ cptn    ] [ EdB          ] [ maxtoo ] [ predatorfreak] 
22:08 [ ag      ] [ dagb    ] [ K            ] [ nano-  ] [ stringfellow ] 
22:08 [ airlied ] [ darktama] [ lumag_offline] [ philv  ] [ swany        ] 
22:08 -!- Irssi: #nouveau: Total of 21 nicks [1 ops, 0 halfops, 0 voices, 20 normal]
22:08 -!- Channel #nouveau created Sun Jun 25 07:52:56 2006
22:08 -!- Irssi: Join to #nouveau was synced in 1 secs
22:09 < marcheu> hmm, libdri still broken :)
22:09 < darktama> not good :(
22:13 -!- gajownik_away [i=gajownik@zspswidwin.pl] has joined #nouveau
22:18 -!- ddl [i=erikw@montezuma.acc.umu.se] has joined #nouveau
22:21  * pmdata can not connect to nouveau.cvs.sf.net: "ssh: connect to host nouveau.cvs.sourceforge.net port 22: Connection refused"
22:22 -!- [freenode-info] help freenode weed out clonebots, please register your IRC nick and auto-identify: http://freenode.net/faq.shtml#nicksetup
22:22 < marcheu> webcvs doesn't work for me
22:22 < marcheu> maybe that's related
22:23 < darktama> mm, I checked out a fresh copy of the drm and ddx about 1/2 hour ago
22:23 < marcheu> although I don't have X up & running now so I can't check the sf.net page about errors
22:23 < marcheu> but on the upper right corner, there is a green/yellow ball that has the site status
22:25 < darktama> hmm, the status says that CVS is working fine.. it isn't though :)
22:26 < marcheu> yeah, that's what I could gather from running links on their page..
22:29 -!- gajownik_away [i=gajownik@zspswidwin.pl] has quit [Read error: 110 (Connection timed out)]
22:34 < darktama> sometimes it's easier to read the HTML than to use links :)
22:41 -!- KoalaBR [n=KoalaBR@port-83-236-13-53.dynamic.qsc.de] has joined #nouveau
22:42 < KoalaBR> Hi all
22:42 < darktama> hey KoalaBR
22:43 < KoalaBR> Does anyone here know, whether the Wiki can use precreated html pages?
22:43 < KoalaBR> So that I can simply upload them?
22:44 < KoalaBR> pmdata: Read your suggestions and questions you offered after I logged off. If  you haven't done it yourself, I will have a look..
22:45 < KoalaBR> And finally: Yes, I can't currently connect to cvs on SF too
22:46 < pmdata> koala> I can leave some work for other people, I prefer to finish nv10 tcl
22:46 < pmdata> I wait for nv1x quadro dumps
22:48 < KoalaBR> pmdata: Ok, I will just do the tests you asked me to do and will report back 
22:48 < marcheu> pmdata: you mean you're waiting for me ?
22:48 < marcheu> pmdata: or someone else ?
22:49 < marcheu> pmdata: I could plug the nv1x into my machine if you like, I just have to remove the nv28
22:50 -!- K [i=hazel@tor/session/external/x-6d6e619018345bca] has quit [Remote closed the connection]
22:51 < KoalaBR> CVS works again, at least for me 
22:55 -!- Netsplit over, joins: marcheu, airlied
22:56 < marcheu> pmdata: which one you prefer ? nv11, nv15 or nv18 ?
22:57 -!- stillunknown [n=madman20@82-168-177-167.dsl.ip.tiscali.nl] has joined #nouveau
23:02 < KoalaBR> pmdata: You were right, except for the different address, NV30_TCL_PRIMITIVE_3D_POINT_SPRITE works the same as NV20_TCL_PRIMITIVE_3D_POINT_SPRITE
23:04 < marcheu> darktama: yeah, EXA without composite is stable here
23:04 < marcheu> maybe we can slash XAA already
23:05 < marcheu> darktama: also, maybe we could slash fbdev support, I don't know, that's up for discussion
23:05 < darktama> sounds good to me :)
23:05 < stillunknown> there is actually a half working driver with a new interface?
23:06 < marcheu> darktama: I think the new sync loop is helping a lot here
23:06 < darktama> I'm a little worried right now.. I've either gotten the 3D engine to work, or somehow my code is using the normal EXA codepath and reporting otherwise in Xorg.0.log
23:06 < marcheu> I wouldn't say "worried"
23:06 < marcheu> but maybe you ran the proprietary driver before, which setup some magic state
23:07 < marcheu> nvidia cards have some state that survive even reboots (like the objects)
23:07 < darktama> well, I was doing that before.. and it still didn't work.. I think this must be the normal EXA path.. it just doesn't make sense that it'd work now "magically"
23:08 < KoalaBR> darktama, marcheu: If you want me to write some docs (how to get "test ready") tell me how to do it and I will 1. try it out myself 2. write a small intro doc
23:08 < KoalaBR> Whenever you feel it would be helpful. Just ask me at that time
23:09 < pmdata> marcheu> nv18 , in softquadro mode please (and a cup of coffee)
23:10 < marcheu> KoalaBR: ok
23:11 < marcheu> pmdata: ok, I'll eat and switch the cards when I'm back. but... I don't have coffee since I never ever drink coffee
23:11 < darktama> you don't drink coffee!!!
23:12 < KoalaBR> He is not alone :)
23:12 < KoalaBR> I prefer tea, and I'm not british :)
23:12 < darktama> well, it's 5am here.. I *need* coffee
23:12 < darktama> :)
23:12 < stillunknown> there are even somewhat normal people who don't drink coffee :-)
23:13 < KoalaBR> Darktama: You are excused :)
23:13 < stillunknown> darktama: why aren't you in bed?
23:13 < KoalaBR> He drank too much coffee 
23:13 < darktama> haha :)
23:13 -!- Myrizio [n=Myrizio@host180-101.pool80104.interbusiness.it] has joined #nouveau
23:13 < KoalaBR> SCNR
23:14 < darktama> no, I was deciding on how I'm going to implement the shader compiler.. then I ended up messing with the ddx somehow :S
23:15 < stillunknown> sleep is good, because we need someone to maintain the driver later on:-)
23:15 < pmdata> marcheu> oh I was waiting for t0ma7 dumps in fact, but I need to check some stuff on nv1xgl (softquadro or not)
23:29 < darktama> KoalaBR: what docs did you want to write?
23:32 < KoalaBR> darktama: If you think the code is advanced enough, what you need to check out, build, how to configure the system to use your new code (Driver, DRM?) 
23:33 < KoalaBR> In the same vein, as the intro to renouveau I wrote
23:34 < KoalaBR> So that we have a standard answer for that code too "Read docs blabla and report back" or something like that
23:35 < darktama> ahh, ok
23:37 < KoalaBR> Just ask me, if you think others (devs at first) can try it too. I will play the guinea pig and if I survive, I will write some HOWTO doc :)
23:42 < pmdata> yep, would be nice to know if the nouveau stuff is barely tryable
23:44 < darktama> ok, so it wasn't the 3D engine.. I didn't fill exa_major/exa_minor and it fell back to software :(  but, I do get some rendering and then a hang with them filled in.. I wish I knew if they were sw-fallbacks or not
23:51 < marcheu> pmdata: ok, I've made it into a nVidia Corporation NV18GL [Quadro4 380 XGL] (rev c1)
23:51 < pmdata> thanks
23:52 < marcheu> however, the nv18 doesn't use single-vertex stuff
23:52 < marcheu> nv18hl
23:53 < marcheu> nv18gl
23:54 < marcheu> hmm, maybe that variant does actually
23:54 < marcheu> it looks like quadro NVS cards behave like geforces, though
--- Log closed Сбт Авг 19 00:00:17 2006
