00:00:00 --- log: started osdev/03.01.05 00:02:39 --- quit: spaz (Remote closed the connection) 00:05:05 --- quit: trans (Read error: 110 (Connection timed out)) 00:07:38 --- nick: zwanem -> zwane 00:16:34 --- join: redblue (star@ppp039.216-96-207.sherb.mt.videotron.ca) joined #osdev 00:20:57 --- quit: file (Read error: 113 (No route to host)) 00:29:47 --- join: Mathis (~Mathis@manz-d9b94c16.pool.mediaWays.net) joined #osdev 00:30:19 I've had a terrible cold/flu thing since dec 25 00:34:41 --- join: p253053 (~keke@ua114d6hel.dial.kolumbus.fi) joined #osdev 00:34:47 * p253053 murr! 00:34:51 no 00:34:54 not really 00:34:55 --- nick: p253053 -> mu 00:34:58 --- nick: mu -> mur 00:35:34 I SAID HELLO! 00:36:46 ho mur 00:36:54 * mur kicks Mathis 00:37:10 is that evil phrase? 00:37:37 I HATE THAT 00:37:38 Mathis Mathis Mathis 00:37:48 OKAY I GOT IT 00:38:22 and next time i know to greet you like that to get your attention >;) 00:38:39 disattenting... 00:39:13 "not paying attention" is what i woudl use 00:39:28 ;-) 00:39:37 anyways 00:39:47 do you have any idea what are applications "roboform" 00:39:52 "soulseeker" 00:40:03 soulseek 00:40:08 "winfingerprint" 00:40:11 and 00:40:15 unable to comply, not enough data to proceed 00:40:25 NCS 00:41:02 wtf is NCS 00:41:10 it's in start apps thing 00:41:30 or pop 00:43:02 C:\WINNT\crazy_ncs.exe 00:43:39 sound 1337 wannabe app 00:44:06 * mur installs opera, should i? 00:45:18 mur... 00:45:22 shut up! 00:45:27 Mathis what do you think, shoudl i install opera on public computer? 00:45:39 no i wont, dont you either 00:46:06 I'm not allowed to answer that question 00:46:19 sure you can tell your opinoin 00:47:02 DCL-E-FILENO, no such device or address.... 00:48:50 NCS has description "Toggles Num, Caps, Scroll Lock at specified intervals" 00:49:23 "C:\WINNT\wööö.exe.exe" ? 00:49:33 what the foo 00:53:13 * mur uses mirc :D 00:53:28 no other clients on computer :( 00:53:55 Robert wake up now, no time to sleep it's 10 am there already 00:55:54 * mur surfs around with office chair (which has wheels under it) 00:57:14 well, to surf around with a chair without wheels would be a bit tricky... 00:57:51 http://www.linuks.mine.nu/debian-worldmap/ 00:58:07 Mathis :) never underestimate my skills ;) 00:59:04 I dont 00:59:33 very good 00:59:38 * mur gives Mathis a cookie 00:59:54 I dont want any internet cookies... I want real cookies... 01:00:24 that is real cookie 01:03:39 --- nick: geist -> geist-sleep 01:04:25 mur: trying to understand people... 01:04:42 whom? 01:04:54 everyone 01:05:41 http://www.linuks.mine.nu/debian-worldmap/ is fancy 01:05:48 maybe... 01:15:31 --- join: lodda (htjonhigfg@p508FCC5A.dip.t-dialin.net) joined #osdev 01:18:12 loddddddda 01:18:32 --- join: trans (rppcia@fatwire-201-98.uniserve.ca) joined #osdev 01:18:38 trans 01:18:48 murrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr! 01:18:55 mur seems to be lonly 01:19:08 seems to have no job, no work 01:19:14 seems to be bored 01:19:31 the second one was correct 01:19:38 i am bored 01:19:56 no sane non-bored person joins #debian 01:20:52 http://www.calle.com/info.cgi?l&name=kunnas&cty=Finland&alt=282 01:20:59 umm wait! finland? 01:23:17 Mathis: where did you get your scancode information for german keyboard? 01:23:35 lodda: by trying out 01:23:59 Mathis: would you mind if i use your information? 01:24:08 no 01:24:15 thanks 01:24:17 they are in my sources 01:24:35 also if you would try them out you would get same results... 01:27:39 --- quit: zhware ("Client Exiting") 01:28:43 arghhabagblsblsblablabla. 01:28:47 it can NOtbe thar hard 01:29:20 to type correctly? 01:29:25 * mur is stupid. it said "Please enter longitude coordinate in decimal form only, without spaces. " i changed it again and again. it did have space in the beginning. it worked instantly when i removed it :P 01:34:34 daxos is messy :( 01:35:21 doesnt messer, uuh, matter... 01:37:36 dax: :( 01:39:23 "Since I still don't trust html editors, these pages are written by hand. But since I also do not believe in HTML anymore, these pages are written in XML. " 01:39:50 mur: dont write in ANY language... 01:40:27 Mathis be quiet 01:40:31 i will 01:41:04 I'm not quit 01:41:08 I'm not quiet 01:41:17 talk then 01:41:40 praglaglos 01:42:22 gtg 01:42:26 haha 01:42:28 read you later... 01:42:31 I'm not quit 01:42:32 I'm not quiet 01:42:39 argh, not the both lines! :O 01:42:50 k cya 01:42:57 --- quit: Mathis ("gotcha!") 02:13:02 --- quit: Kurt ("Connection reset by fear") 02:21:33 --- quit: trans (Read error: 60 (Operation timed out)) 02:47:34 boooring 02:53:10 --- quit: lodda (Read error: 104 (Connection reset by peer)) 02:55:02 cya later talk .. ah something 02:55:10 --- join: SnitraM (ams@speedy.ludd.luth.se) joined #osdev 02:55:26 --- quit: SnitraM (Client Quit) 02:56:12 --- nick: mur -> blabla345235 02:56:16 --- quit: blabla345235 () 03:08:20 --- quit: redblue (Read error: 60 (Operation timed out)) 03:21:02 --- join: lodda (htjonhigfg@p508FDEE1.dip.t-dialin.net) joined #osdev 03:43:03 --- join: trans (bafcjd@fatwire-201-98.uniserve.ca) joined #osdev 04:14:02 --- join: muphicks (~muphicks@earthforce.freeserve.co.uk) joined #osdev 04:28:12 moo 04:31:52 --- quit: trans (Read error: 110 (Connection timed out)) 05:43:13 --- join: trans (kzoklb@fatwire-201-98.uniserve.ca) joined #osdev 05:50:46 --- nick: prologic -> pl_zzzz 06:00:33 --- quit: SlowCoder (Read error: 110 (Connection timed out)) 06:15:09 --- quit: pl_zzzz ("KVIrc 2.1.2 'Monolith'") 06:19:09 --- join: zhware (~sto@218.221.192.191) joined #osdev 06:23:16 --- join: darkito (darkito@80-25-82-102.uc.nombres.ttd.es) joined #osdev 06:23:19 hihi 06:23:40 --- quit: trans (Read error: 60 (Operation timed out)) 06:29:05 --- join: Mathis (~Mathis@manz-d9b94c1b.pool.mediaWays.net) joined #osdev 06:29:41 hiyall 06:32:25 --- join: SlowCoder (~jja@194.239.250.254) joined #osdev 06:36:24 --- quit: zhware () 06:37:05 --- join: zhware (~sto@pddc0bf.osakac00.ap.so-net.ne.jp) joined #osdev 06:37:32 --- nick: zhware -> zh 06:41:56 --- nick: pav[sleep] -> pavlovski 06:41:58 hi zwane 06:41:59 * 06:42:07 hi all 06:42:07 hi zh 06:42:18 hi pavlovski 06:43:10 --- nick: minddog -> minddog[dead] 07:04:01 --- quit: zh () 07:05:05 --- join: file (jwired@mctn1-3777.nb.aliant.net) joined #osdev 07:05:44 Hey fi 07:05:46 le 07:06:36 Hello Robert. 07:07:46 hi 07:07:57 Hello file. 07:08:03 where is directory? *g* 07:09:20 --- quit: muphicks (Read error: 54 (Connection reset by peer)) 07:09:30 :) 07:09:56 --- quit: dh (Read error: 104 (Connection reset by peer)) 07:33:47 --- quit: asmodeus ("reboot. Fuck XMMS!") 07:35:57 --- join: asmodeus (~www@h125n2fls33o867.telia.com) joined #osdev 07:58:17 --- join: zh (~sto@pddc0bf.osakac00.ap.so-net.ne.jp) joined #osdev 08:00:39 --- join: trans (usquip@fatwire-201-98.uniserve.ca) joined #osdev 08:02:32 --- nick: minddog[dead] -> minddog 08:05:53 --- quit: zh () 08:12:24 --- join: neniu_ (~masm@cliente-217-70-66-35.bragatel.pt) joined #osdev 08:19:43 --- join: cray (cray@pc-80-195-121-33-sa.blueyonder.co.uk) joined #osdev 08:20:07 what is the size of the descriptors and also the descriptotrs tables? 08:22:17 anybody alive? 08:22:26 yep 08:22:34 wooohoooo! 08:22:43 what is the size of the descriptors and also the descriptotrs tables? 08:22:52 do u have any idea about this qusetion? 08:23:06 nope 08:23:13 cool! LOL 08:23:18 I also dont know what exact you mean 08:23:22 u can say dead 08:23:37 * file growls 08:23:40 which arch? 08:23:46 I hate the Yahoo Photos thing 08:24:03 cray? 08:24:09 i mean what is the lenght of the descriptors in intel cpus and also the length of the gdt and ldt 08:24:16 ah 08:24:18 8 byte 08:24:32 cool thanks and the table? 08:24:42 8192 entries max 08:25:09 first GDT entry is NULL 08:25:09 right so its 64 * 8192 for max size ? 08:25:25 8 byte x 8192, yep 08:25:36 cool thanks 08:25:56 you can also read it in the Intel docs... 08:26:11 which = 524288 = 80000h 08:26:32 well... 08:28:00 and? 08:28:11 the table doesnt have to be that long 08:28:13 nothing just needto know 08:28:59 ? it doesnt have to be that long? do u not need to set aside that much space for it even though some decriptotrs may be blank 08:29:05 you can also set a smaller one 08:29:32 by setting the offset in the GDT register you also set the size of your table 08:29:43 also IDT register 08:29:46 and LDT register 08:30:09 offset? u mean how many bits the offset is? e.g. for the 286 it is 16-bit ? 08:30:16 no 08:30:23 the offset of the table in memory 08:30:39 i dont understand 08:31:32 how does the CPU know where it can find the entries in memory? 08:31:54 no clue! do u set that up yourself? 08:32:00 yep 08:32:31 so what has that got to do with the size? i put it somewhere where there is 8000h free 08:33:15 with the position in memory you also set the size of the table (and the max entry count with it) 08:33:53 ohh right, so u specify how many descriptors u want? it does not have to be 8192, but that is the max that are avaible? 08:34:08 correct 08:34:35 struct GDT_reg { unsigned int offset; unsigned short size; }; 08:34:39 kinda... 08:34:50 cool thanks! uve been a great help 08:35:51 no problem, you would find these infos also in the Intel docs ;-) 08:36:51 true but u have to traul thorugh so much stuff before u get the info. i belive the whole reason of this irc channels is to share/help etc, anyway thanks again!!!!! 08:37:11 --- join: jsr (www@du-15-69.ppp.telenordia.se) joined #osdev 08:37:19 yep ;-) 08:39:00 1 byte = 8 bits? (intel) 08:39:04 yep 08:39:11 ok 08:39:30 --- join: q_work (quantis@130.88.157.186) joined #osdev 08:39:34 lo 08:39:38 --- nick: q_work -> quantis 08:39:44 ho quantis 08:40:05 hi mathias 08:40:07 how is you? 08:40:14 no Mathias here 08:40:41 doh.... 08:40:43 sorry 08:40:46 just got 08:40:48 up 08:41:22 me is fine 08:43:00 --- quit: file (Read error: 60 (Operation timed out)) 08:43:45 --- join: alphakiller|food (~1234@200.173.175.95) joined #osdev 08:43:53 --- nick: alphakiller|food -> alphakiller 08:44:29 --- join: Xeon (~ask@cs666823-113.austin.rr.com) joined #osdev 08:44:41 wee 08:44:45 hey guys ... 08:44:56 hey xeon ... how is going your card instalation ? 08:45:30 --- join: file (jwired@mctn1-1002.nb.aliant.net) joined #osdev 08:47:57 hey file 08:48:02 how is you 08:48:07 --- quit: cray () 08:48:08 hey alphakiller, how is your 'compiler' going? 08:48:40 I added ASCII suport for it ... So I can use : print "Hey my compiler" 08:48:47 ehehe 08:48:51 --- join: wannabedead (~HeavyJoos@a213-84-139-110.adsl.xs4all.nl) joined #osdev 08:49:05 --- quit: wannabedead (Read error: 104 (Connection reset by peer)) 08:49:33 but if I were you I would change postfix to midfix... 08:50:41 nah ... "( )" stuff sux ... I prefer without it .. 08:50:46 midfix? 08:51:16 c = a + b 08:51:17 --- join: HeavyJoost (~HeavyJoos@a213-84-139-110.adsl.xs4all.nl) joined #osdev 08:51:22 instead of c = a b + 08:51:33 ahh okay 08:51:34 ( ) stuff doesnt sux 08:52:04 talking or I so? 08:52:05 *g* 08:52:14 using stuff like () makes it a lot easier for the compiler to work out whats going on 08:52:24 without putting special cases in for everything 08:52:51 Mathis: how far has elf support gone? 08:53:13 --- quit: kemu (Read error: 54 (Connection reset by peer)) 08:53:26 lodda: just trying to figure out... 08:53:35 Mathis: :/ 08:53:40 can relocate variables now... 08:53:51 dunno if functions are relocated correctly too 08:54:08 coz it doesnt seem to matter correctly 08:54:20 aha 08:54:34 quantis: postfixing you define the priority of the operations. They don´t have defined priority. 08:54:43 --- join: zwane_ (User-10417@modemcable092.130-200-24.mtl.mc.videotron.ca) joined #osdev 09:00:01 --- quit: jsr (Read error: 110 (Connection timed out)) 09:06:15 * Xeon thinks hes going to start on libc :/ 09:06:32 * Mathis has libc++ 09:06:34 --- join: Madmonky1 (~leaveme@12-235-171-19.client.attbi.com) joined #osdev 09:06:54 Xeon: _alloca forever?@!#$?#$%^&& 09:08:46 --- quit: Xeon (Read error: 104 (Connection reset by peer)) 09:09:58 --- quit: Madmonky1 (Client Quit) 09:11:12 --- quit: zwane (Read error: 110 (Connection timed out)) 09:16:59 --- nick: alphakiller -> alphakiller|away 09:20:51 --- join: n0vice (~n0vice@pcp02102499pcs.towson01.md.comcast.net) joined #osdev 09:21:13 hey 09:22:51 ello novice 09:23:36 whats up 09:24:26 the sky.... 09:25:49 who wants to see my neat new toy? 09:26:12 me 09:26:13 * quantis puts his hand up.... 09:26:14 You have two sets of genitals? 09:26:14 what is it 09:26:15 http://www.ghettohardware.com/iopener/photos/DSCN0571.JPG 09:26:26 Robert: ha 09:26:47 hes not ready for that, his first havnt even dropped yet 09:27:15 you bought some christmas lights ? 09:27:26 quantis: no 09:27:43 a bunch of wires? 09:27:55 no :p 09:28:01 ohh i know, that picture with the nice frame 09:28:24 :) 09:28:29 that's an LCD :p 09:28:33 i like the matting, it goes nice with the blue picture and black frame 09:28:36 and a computer gah dang it! 09:28:36 you got some new hooks for you bondage fetish... 09:28:46 lol 09:28:51 wheres the comp? 09:29:00 is it an all in one ? 09:29:00 lol 09:29:01 n0vice: that's it - all components are behind the LCD 09:29:08 sweet 09:29:17 is it a gateway profile or something? 09:29:20 http://www.ghettohardware.com/iopener/ 09:29:21 go read!!! 09:29:55 ok 09:30:03 anyone know opengl or sdl or something 09:30:17 n0vice: nope 09:30:23 whats an iopener 09:30:48 quantis: that's the name of the device 09:30:57 quantis: some company sold them for $99 USD thinking you'd buy their dialup service 09:31:08 quantis: people found out it's a standard computer and a normal OS could be run on it :) 09:31:31 did you do that guide? 09:31:42 n0vice: no 09:31:55 n0vice: my I-Opener is getting sent out tomorrow, already hacked and running Redhat on a 2GB hardrive 09:32:20 oh 09:32:27 so that pic wasnt actually yours 09:32:31 affirmative 09:32:34 I was just using it as an example :p 09:32:47 i was gonna say, your either rich or insane 09:32:59 n0vice: why be rich? it's a bloody cheap thing 09:33:14 no i meant if you were the guy in the guide 09:33:20 oh ic 09:33:25 just to buy a 100$ comp and fuck around with it 09:33:26 lol 09:33:27 i like the carmp3 thingy 09:34:22 I'll take a picture once it's up and running here next week 09:34:51 what are the system specs? 09:35:04 of the I-Opener? 09:35:18 yea 09:35:25 by default it's 32MB of RAM with a 200MHz Winchip 09:35:33 --- quit: alphakiller|away () 09:35:34 but it's just SODIMM, and Socket 7 09:35:56 your gonna upgrade it/ 09:35:58 ?* 09:36:01 dunno 09:36:08 I might 09:36:14 put a k6 in there it ill be a dec ent mchine 09:36:16 gotta get my USB NIC first 09:36:50 god no one is in #sdl 09:36:51 this sucks 09:36:57 * file dances 09:37:32 --- nick: minddog -> zzzzdog 09:40:16 * file dances some more 09:40:19 what's everyone up to? 09:40:47 god irc people are assholes 09:40:57 some 09:40:57 not all 09:41:37 a lot of the people that know enough to help are lol 09:42:23 * file yawns 09:42:33 I was supposed to go network somebody's house today... but I'm lazy 09:44:14 --- quit: quantis (Read error: 104 (Connection reset by peer)) 09:44:22 --- join: quantis (quantis@130.88.157.186) joined #osdev 09:46:35 how does that guy on ghettohardware know all that shit lol 09:46:45 * spyck is back (gone 16:54:45) 09:46:48 about schematics and hacking up hardware 09:47:02 * file shrugs 09:47:06 you can do alot with I-Openers... 09:47:45 --- join: liesbeth (Simon@83.15-201-80.adsl.skynet.be) joined #osdev 09:47:47 ello 09:47:57 --- join: dh (~dh@80-235-52-24-dsl.kj.estpak.ee) joined #osdev 09:48:05 ello 09:48:21 seems that your h key is defective 09:48:25 *g* 09:49:28 My key doesn't work as it sould :/ 09:50:03 novice: its not different from high school electronics wise everthing works more or less same, only problem being messing with stuff on motherboards is that its all so small now :( 09:52:15 yea i dont think my high school has any electronic classes that do anything like that 09:54:46 mine didnt really but i learnt all the principles 09:55:34 THE KLINGONS ARE APPROACHING 09:55:50 plus most chip companies tend to put all there info on the web now 09:57:36 yea 09:58:22 --- nick: neniu_ -> neniu 10:12:40 --- join: jsr (www@du-15-99.ppp.telenordia.se) joined #osdev 10:21:08 --- nick: geist-sleep -> geist 10:21:34 yawn 10:21:51 --- join: Kurt (~kmw@gc-nas-01-s276.cinergycom.net) joined #osdev 10:22:13 Good morning. 10:22:25 Good Day 10:22:48 Salvete, pueri et puellae 10:29:55 --- quit: trans (Read error: 110 (Connection timed out)) 10:40:47 hi geist 10:41:05 geist: just wondering did that catching the triple fault thing that you were planning work ? 10:48:10 hmm 10:48:17 i can't come up with a name for my 3d engine 10:49:23 dax: ....something 10:49:24 3D.Engine? *g* 10:49:38 dax3d sucks as a name 10:50:02 Axes! 10:50:21 --- join: revanthn (revanthn@202.9.183.198) joined #osdev 10:50:22 quantis: sort of. it can only catch some cases where things are screwed up 10:51:05 it can catch a kernel stack overflow or a stack corruption, but can't catch a pgdir getting screwed up or a bad cr3 10:51:08 right 10:51:30 is there anyway of recovering from a triple fault 10:51:38 or resetting the mcachine without clearing the memory? 10:51:49 for a second there I was thinking it would catch tht as well, but if the pgdir is screwed up it can't read from the idt to take the fault 10:52:20 sort of, but depending on the cause it's most likely a screwed up state 10:52:50 back in a minute, gotta get coffee 10:53:04 --- join: kemu (~kemu@67.6-201-80.adsl.skynet.be) joined #osdev 10:53:27 well hopefully that would be a problem id just get rid of that process, just dont want it to bring down the resto f the machine, thats the main problem 10:53:57 why do i suck so much @ names 10:54:00 cause unless you store page info within each page theres not much way of being able to recoere from a bad page table is there 10:54:12 dax: just nick othere ppls 10:54:33 quantis: i suck @ that too 10:55:26 what about quantis 10:55:35 i think its a wonderful name for a 3d engine... 10:55:45 omg no then i would have to thing bout you 10:55:48 j/k 10:55:50 dax:what about darkmatter.3d 10:56:19 dax: ast least im not lynx 10:56:24 roygbiv ? 10:56:54 theli? 10:56:56 vovin 10:58:15 what do those words mean? 10:58:20 no idea 10:58:21 lol 10:58:48 it's a therion cd 10:59:24 * geist is back from coffee 11:00:43 geist: mmm 11:00:53 --- quit: Kurt ("Connection reset by beer") 11:01:03 terbium? 11:05:08 hmm 11:05:10 bleh 11:05:35 dont bleh round here! 11:05:52 darkmatter is a decent name i suppose 11:05:53 no? 11:05:59 I BLEH BLEH BLEH 11:12:26 but that could become dm3d 11:12:30 which sounds a bot better 11:12:34 or dm.3d 11:13:03 nobody knows what dm means... 11:13:03 i don't need the 3d part in it 11:13:12 maybe DM (Deutsche Mark) *g* 11:13:29 NOOOOOOOOOOOOOOO 11:13:49 just though you have all the old dm coins in it as primatives.... 11:13:57 oh no we don't want to link it to something sucky 11:14:16 :) 11:14:38 hows the actualyl coding for it going ? 11:17:29 --- join: jsr^03 (www@du-12-241.ppp.telenordia.se) joined #osdev 11:18:07 call it "ed" 11:19:05 --- quit: revanthn () 11:20:44 coding part is coming ok 11:20:51 having some probs with MSVC though 11:20:54 for it is crap 11:21:18 aw 11:21:26 you must become one with msvc 11:21:39 itas better for cutting code then borland 11:21:59 prolly but it's still crap 11:22:05 but its c implmentation is fucked and nearly cost me 50% of my marks in a programmign course once.... 11:22:06 --- quit: jsr (Read error: 60 (Operation timed out)) 11:22:11 best editor i know 11:22:24 it's c++ implementation is braindamaged :D 11:22:26 been trying to findsome thing like it on other platforms for ages 11:22:33 never done any c++ in it 11:23:30 visual slickedit 11:24:01 it's template implementation is broken 11:24:06 stl is broken too 11:24:18 (things not being in std:: as they are supposed to) 11:24:50 gives warning about it's own header files 11:24:52 dont like slickedit 11:24:53 kinda sucky 11:24:56 kept crashing on me 11:24:57 www.slickedit.com 11:25:03 wtf 11:25:09 try again? 11:25:12 or even better 11:25:15 VIM!!!! 11:25:35 copy con.... 11:25:45 i like x edit 11:25:53 xemacs doh 11:25:57 or just take the pins of your CPU and induct the bits directly *g* 11:26:00 but it seems buggered on win32 11:26:12 no syntax highlighing or autoindentiion 11:26:14 i like vim 11:26:17 can't help it 11:26:43 i have a new found respect for it since since it dawned on me why it works the way it does 11:27:13 jsut havent been arsed to memorise all the commands (or pretty much any except i, dd and ctrl-q q! etc) 11:27:38 oh i don't memorise many either 11:28:43 my neck hurts... 11:28:48 bbl 11:28:57 --- nick: quantis -> q_cleaning_Room 11:34:34 --- quit: liesbeth () 11:35:51 --- join: Ghiottone (~alex@ppp-236-17.27-151.libero.it) joined #osdev 11:36:12 --- part: Ghiottone left #osdev 11:58:08 --- join: alphakiller|away (~1234@200.190.23.227) joined #osdev 11:58:16 --- nick: alphakiller|away -> alphakiller 11:58:28 --- join: trans (cvluyj@fatwire-201-98.uniserve.ca) joined #osdev 12:01:29 --- quit: jsr^03 (Read error: 60 (Operation timed out)) 12:04:08 --- quit: file (Read error: 113 (No route to host)) 12:04:24 --- join: cray (~temp@pc-80-195-121-33-sa.blueyonder.co.uk) joined #osdev 12:04:39 hi mathis 12:04:42 hey ppl 12:05:03 a descriptor is 32 bit right? 12:08:10 ho cray 12:08:11 --- join: file (jwired@mctn1-1595.nb.aliant.net) joined #osdev 12:08:16 a descriptor is 64 bit 12:08:42 thanks 12:11:50 --- join: nov1ce (~root@pcp02102499pcs.towson01.md.comcast.net) joined #osdev 12:14:47 --- quit: file (Read error: 104 (Connection reset by peer)) 12:17:19 anybody here? 12:17:26 yo 12:17:28 no 12:17:34 pavlovski: No. 12:18:10 in body not mind 12:18:41 pavlovski: basically, if the kernel tracks allocations on a per-4KB page basis, you need 1 data structure per 4KB page 12:18:49 --- join: axeld (axeld@p50829533.dip.t-dialin.net) joined #osdev 12:18:52 pavlovski: on a 64GB box that's 16M 4KB pages 12:19:08 typical per-4KB page data structures are around 64B in size 12:19:12 wli: how can it overcome the limitations of the TLB? 12:19:20 axeld: It's not for the TLB 12:19:43 SVR4/i386 ABI says userspace gets >= 3GB of the 4GB virtualspace for itself, and the kernel gets 1GB 12:19:56 wli: OK, so far so good 12:20:00 But 64B/page * 16M pages == 1GB 12:20:02 why wli ? 12:20:20 which means just tracking things per-4KB page runs you out of space with the coremap alone 12:20:27 --- join: witten (~witten@adsl-gte-la-216-86-199-140.mminternet.com) joined #osdev 12:20:32 wli: so it's just a layer on top of the pages for the heap?? 12:20:35 "page clustering" increases the base page size. 12:20:45 yes, but think about it in terms of space per page 12:21:07 So instead of tracking allocations for every 4KB you track them at every 16KB or so (compile-time configurable). 12:21:13 i.e. 64 bytes per 4KB page is 1/64th of total memory dedicated to tracking memory 12:21:17 (which seems a lot to me) 12:21:18 right 12:21:29 --- quit: cray () 12:21:32 yay, actual osdev talk 12:21:37 and something I'm kind of interested in 12:21:41 --- join: marcus_o (DSLflat@p508FDCF7.dip.t-dialin.net) joined #osdev 12:21:44 if you track allocations for every 16KB you cut it by a factor of 4 12:21:47 on a 32MB machine you're still giving 512KB to tracking pages 12:21:50 hmm 12:21:54 wli: but you only need as much as mem available, you won't need 16M pages if you only have 1G 12:21:56 This actually doesn't cause much in the way of complications 12:22:15 axeld: Also true; this is just an optimization for smaller boxen. 12:22:15 speaking as an MVP, I feel I should introduce the way NT does things ;) 12:22:26 --- join: file (jwired@mctn1-4585.nb.aliant.net) joined #osdev 12:22:38 Win32 deals in 64KB regions, although AFAIK the NT kernel talks in 4KB pages 12:22:40 axeld: but you also have to account for all the virtual mem as well, they still need pagetable entrys 12:22:51 --- nick: q_cleaning_Room -> quantis 12:23:08 quantis: but those can paged out :-) 12:23:10 The way things normally work process' PTE's just increase reference counters and "point" with page frame numbers measured in 4KB units. 12:23:20 --- quit: alphakiller () 12:23:28 axeld: paging out VM structures? 12:23:43 pavlovski: not? :-) 12:23:45 axeld: you still need to keep track of them in case they get refferenced 12:23:58 so basically all you have to do is shift by 2 bits (for 16KB allocation units) or 4 bits (for 64KB allocation units) and you get a "software pfn" 12:24:06 axeld: well, if the VM manager does the paging, what happens if you page out the stuff the VMM needs? 12:24:10 quantis: but not always in memory 12:24:23 and nobody actually cares that the PTE is pointing "inside the struct page", just that it refers to it. 12:25:00 The only real constraint is just arranging mmap() so that it actually sets up the userspace mappings the right way. 12:25:06 wli: my solution to this is to keep as little data as possible associated with _physical_ pages 12:25:21 associate all the important memory management stuff with _ranges_ of _virtual_ pages 12:25:28 pavlovski: And it more or less works until things get "too" extreme. 12:25:38 wli: so the address range is still bound to the real pages, right? 12:25:54 I'm thinking 8 bytes: backward and forward pointers within a linked list 12:26:05 8 bytes per page? that's hardly anything 12:26:17 pavlovski: but there's no way to tag the things with flags or attach extra stuff to them 12:26:25 no need 12:26:28 (and it's not how pre-existing kernels are doing it) 12:26:32 free pages go in the free page list 12:26:38 zeroed pages go in the zeroed page list 12:26:48 dirty-to-be-written-to-disk pages go in the modified list 12:26:57 allocated pages don't go in any list 12:27:07 ^ the way NT does it 12:27:21 (except that, in NT, allocated pages go into a working set) 12:29:12 by and large most things keep more metadata per physical allocation 12:29:44 so beefing up the allocation unit trims down the fraction of memory required for the metadata. 12:29:57 No big surprises. 12:30:12 hmm... maybe 12:30:34 hmm... maybe allocating physical memory in multiples of the cache line size would solve two problems 12:31:06 if (sizeof(struct page) == 64) then you reduce the physical memory for the coremap from 1/64 of RAM to 1/4096 12:31:23 or from 64B per 4KB page to 1B per 4KB page. 12:32:41 And other things are easy to do also. 12:32:41 ok, so what if a process asks for 4KB of memory? 12:32:48 --- join: jbreker (~jbreker@HSE-Ottawa-ppp236097.sympatico.ca) joined #osdev 12:33:10 you have to request physical memory in 64-page blocks 12:33:14 pavlovski: You give it an allocation unit's worth of memory anyway. 12:33:27 pavlovski: i.e. "internal fragmentation" 12:33:29 --- quit: quantis (Read error: 104 (Connection reset by peer)) 12:33:45 hmm 12:34:00 so what if you ran 100 copies of an image 4KB in size? 12:34:24 ok, better: 12KB in size? 4KB .text, 4KB .data, 4KB .bss 12:34:28 Well, most of the time you're just going to point the PTE's at the same physical pages. 12:34:35 So that's not the worst case. 12:34:36 for .text, sure 12:35:01 you're allocating 256KB per physical allocation 12:35:17 which means 256KB minimum per image (assuming .data and .bss can be combined) 12:35:32 not necessarily 256KB 12:35:40 well, 64 pages 12:35:41 it's configurable 12:36:12 64KB would be a factor of 16, 256KB is a bit excessive even for 64GB (but might come in handy if a 40-bit version of PAE or something materialized) 12:36:47 64KB is actually not even that useful for 64GB, 32KB or 16KB are more common for it. 12:37:04 even then, if you had 64GB of memory, I guess you wouldn't mind if you wasted 256KB here or there ;) 12:37:26 my hard disks are smaller than 64GB in size, and I don't mind having 256KB files sitting aroudn 12:37:47 presumably on a smaller system (32MB?) you'd allocate with 4KB granularity? 12:38:11 A "small system kernel" would probably just go for 4KB pages yeah. 12:38:43 It's a tradeoff of "how much internal fragmentation are you willing to pay for cutting down on metadata?" 12:38:55 hmm 12:39:13 well, I maintain that the sensible approach is to minimise the amount of metadata 12:39:26 "Think smarter, not harder" :) 12:46:34 pavlovski: Sure 12:48:45 anyway, gtg 12:48:59 working out, please wait... 12:49:06 biceps... 12:50:10 pavlovski: It's bad to be bloated, OTOH mere bloat-consciousness isn't quite good enough for the huge bigboxen. 12:55:41 pavlovski: Also, there are prefaulting benefits since once you've handed a struct page to someone, you might as well point all the PTE's you can at it. 12:56:49 --- quit: nov1ce ("Client Exiting") 13:09:09 eek why is this giving me a linkin error 13:09:31 --- quit: marcus_o ("reboot("BeOS");") 13:11:14 --- quit: witten ("bye") 13:11:37 --- join: marcus_o (Marcus@p508FDCF7.dip.t-dialin.net) joined #osdev 13:11:48 ho marcus_o 13:12:05 hi 13:12:35 was gehtn? 13:15:14 gibt nix neues 13:15:21 aso? 13:15:46 schreibste jetz auch n OS? 13:16:44 sollst ja für Bernd's Zeta einiges gemacht haben 13:17:40 --- quit: trans (Read error: 60 (Operation timed out)) 13:19:07 das mit Zeta stimmt aber nicht, ich bastel für openbeos 13:19:14 jau 13:19:21 wees ick ja 13:20:12 biste ihm schonma begegnet? ich glaub, er hat da ma was erzählt von ner BeGeistert... 13:21:36 no german... 13:22:49 dax: ;-) 13:24:03 auf jeden schwärmt er von dir ("einer der besten Programmierer der Welt") *ggg* 13:25:14 Ja ich kenne ihn persönlich 13:25:25 moooomoo 13:26:37 aber ich glaub eher, dass Programmieren eher ne Kopfsache ist 13:27:07 der Mensch nutzt ja nur 10% oder so seines Gehirns, und wir wissen nicht, was der Rest gerade tut *ggg* 13:28:06 the foundation for my 3d engine is starting to work a bit 13:28:44 foundation? 13:29:04 base stuff 13:29:07 ah 13:29:10 things you build on later 13:29:36 log, logmanager, singleton, vectors, matrices, exceptions 13:29:45 ah 13:29:51 what is singleton? 13:30:20 it's a small template class here on which i can base other classes so that there can only exist one copy of that object 13:30:27 so perfect for a manager class 13:30:34 ah 13:31:09 will your 3d engine support hardware acceleration? 13:31:16 yes 13:31:25 vertex processing and so on... 13:31:35 built around opengl with a directx backend later on if i find the time 13:31:58 LogManager::getSingleton().logMessage(LML_NORMAL, "I is cow"); 13:32:01 so it's not really YOUR 3d engine? 13:32:01 hmm 13:32:02 OO galore 13:32:14 depends what you mean iwth 3d engine i suppose 13:32:49 the 3d engine processes the vertices and transformations... nearly everything that has to do with 3D 13:33:50 * spyck is away: I'm busy 13:33:57 * spyck is back (gone 00:00:03) 13:33:59 the "lower-level" stuff will be handled with opengl/directx, the engine will manage textures, scenes, shaders, meshes, ... 13:34:11 ah 13:34:34 I'm planning to use Mesa 5.0 with hardware support from XFree4 source tree 13:34:41 for my 3D.GUI 13:35:08 don't think on porting this beast yet 13:35:15 i'm being a STL whore 13:35:19 heh 13:35:44 and i really need a name 13:35:44 well, at first I have to implement harddisk support which itself needs modularity-support what I'm working on right now 13:37:06 heh, I'm not good in finding names for anything 13:37:12 same 13:37:14 thats why my OS is just called OS 13:37:46 also Shell and UI and so on... 14:00:41 --- quit: lodda (Read error: 104 (Connection reset by peer)) 14:09:37 --- quit: Mathis ("Shutdown sequence initiated...") 14:11:19 --- join: gianluca (~kernel@ppp-36-135.28-151.libero.it) joined #osdev 14:11:23 hello 14:14:14 --- join: damieng (~damieng@user111.res2-fi2.jtibs.net) joined #osdev 14:25:26 --- join: jsr (www@du-16-88.ppp.telenordia.se) joined #osdev 14:30:15 --- quit: neniu ("leaving") 14:32:36 --- quit: file (Read error: 110 (Connection timed out)) 14:38:00 --- quit: n0vice ("I like core dumps") 14:54:45 --- join: Kurt (~kmw@gc-nas-01-s43.cinergycom.net) joined #osdev 14:56:54 --- join: trans (axpeen@fatwire-201-98.uniserve.ca) joined #osdev 15:14:09 --- quit: gianluca (Read error: 110 (Connection timed out)) 15:18:15 --- quit: Kurt ("Connection reset by ear") 15:22:40 --- quit: trans (Read error: 60 (Operation timed out)) 15:25:01 --- quit: jsr (Read error: 110 (Connection timed out)) 15:28:15 --- join: spaz (~kreep@mkc-65-28-33-22.kc.rr.com) joined #osdev 15:28:21 howdy 15:30:04 --- join: gianluca (~kernel@ppp-217-139.28-151.libero.it) joined #osdev 15:36:23 --- quit: air (Excess Flood) 15:36:24 --- join: air (~brand@12-254-199-50.client.attbi.com) joined #osdev 15:49:41 excess flood?!?!?!?!? 15:59:48 hmmm 16:00:07 better than 'Success Flood' 16:00:09 :) 16:14:57 --- join: n0vice (~n0vice@pcp02102499pcs.towson01.md.comcast.net) joined #osdev 16:15:55 --- join: file (jwired@mctn1-4069.nb.aliant.net) joined #osdev 16:18:48 hi file 16:19:01 hiya 16:19:03 my ISP was down 16:19:22 hmm wireless? 16:20:20 no 16:20:21 ADSL 16:20:25 they did some maintenance I think 16:20:31 cause the service is incredibly faster :) 16:21:47 wow 16:21:56 8mbits? 16:22:57 nah - right now I'm at 1.6 16:24:46 oh 16:24:53 here in italy the max is 1.2 :-) 16:25:02 but just for business thigns :) 16:26:33 the max with my ISP is 2 16:26:39 base is 256kbits 16:26:47 if I was in the city I'd probably get a burstable rate to that 16:32:05 * spyck is away: sleeping 16:36:00 night everybody 16:38:57 --- quit: gianluca ("ircII EPIC4-1.1.7 -- Are we there yet?") 16:42:00 --- join: Aardappel (wvo96r@ipd54be3ab.olympus.planet.nl) joined #osdev 16:52:22 --- quit: jbreker ("Client Exiting") 16:56:03 --- join: DRF (Daniel@host213-121-70-116.surfport24.v21.co.uk) joined #osdev 16:57:23 geist: hiya 16:58:47 hrm, there is a burning plastic smell comming out of my monitor 16:58:52 damn 16:59:10 coming 17:00:13 --- quit: damieng ("Crikey!") 17:01:44 --- quit: Aardappel ("http://wouter.fov120.com/") 17:02:00 --- join: trans (akcftg@fatwire-201-98.uniserve.ca) joined #osdev 17:03:01 Hi trans 17:05:24 hi file 17:05:37 --- quit: darkito (Read error: 104 (Connection reset by peer)) 17:13:43 geist: what's up? 17:13:53 --- join: Kurt (~kmw@gc-nas-01-s6.cinergycom.net) joined #osdev 17:31:36 --- quit: pavlovski ("Client Exiting") 17:40:07 --- quit: trans (Read error: 60 (Operation timed out)) 17:48:27 --- join: bd32322 (bd32322@pool-141-157-80-34.balt.east.verizon.net) joined #osdev 17:49:23 anybody from purdue and know zephir? 17:50:35 he comes in here somewhat frequently 17:51:16 he was supposed to get back to purdue today but i cant get a hold of him 17:52:11 are you from purdue? 17:53:01 --- join: Ep (~EP@211.136.159.235) joined #osdev 17:53:13 --- join: alphakiller (~1234@200.190.21.237) joined #osdev 17:53:20 nupe 17:53:58 --- quit: bd32322 () 17:54:15 hey guys 17:54:35 hi 17:54:48 Hi alphakiller 17:57:40 "When lammers will be hackers and they will code their own code ? If you know when, you are probally an idiot" 17:58:13 Do you agree with it ? 18:00:03 :) 18:01:53 ....I just has upgraded my xchat.... 18:01:59 --- quit: Ep ("Client Exiting") 18:05:41 --- quit: Kurt ("Connection reset by rear") 18:07:43 --- quit: marcus_o ("Vision[0.9.5-0906]: i've been blurred!") 18:07:49 --- quit: axeld ("Vision[0.9.5-0906]: i've been blurred!") 18:10:32 --- join: nullify (~nullify@pool-138-89-44-221.mad.east.verizon.net) joined #osdev 18:24:18 --- join: Ep (~Ep@211.136.159.235) joined #osdev 18:37:03 --- join: Xeon (~root@cs666823-113.austin.rr.com) joined #osdev 18:37:08 WEEE!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 18:37:17 MWHAHAHAHAHA 18:37:32 guess what IM talking from!?!?!?!?!? 18:37:40 dh ????~~~~ 18:37:48 xchat!!!!!!!!!!!!!!!!!!!! 18:37:51 mwhahahahahahahahahahahaha 18:39:09 cool? eh? eh 18:39:27 after days of working i finally got my network card to work under LINUX 18:39:30 !yaya 18:40:41 Kewl 18:41:04 Xeon Congra. 18:43:06 Congra? 18:43:31 congratulations xeon 18:43:39 * Xeon doesn't know what to do with linux now............................... 18:43:58 im going to try that new wm xftc (sp?) 18:45:19 What dist of linux did you get? 18:47:36 debian 18:47:58 kewl :) Same as what I use :) 18:48:39 Are you new to debian or have you just run out of things to do generally? 18:49:18 im pretty much new to it 18:49:32 Ah I see 18:50:00 Take it you have used apt-get to update debian. 18:54:39 not yet 18:55:25 Ah, tends to be a good place to start if bored and got band width to spare. 18:57:18 'apt-get update' to update your list of available packages then 'apt-get upgrade' to grab and inst newer versions of the currently installed packages. ('apt-get dist-upgrade' for major upgrading to new dists not just security fixes) 18:57:58 * DRF loves atp. Where would I be without it. :) ;) 18:59:05 ok, ill be right vack 18:59:07 back 18:59:27 ive got to logoff and get back on with my user and not root 18:59:29 --- quit: Xeon ("Client Exiting") 19:02:40 should page tables be created automatically when mapping in pages? 19:03:18 or should the page mapper just require that all the necessary page tables be already present? 19:09:51 hmm? 19:10:08 if you preallocate all of the page tables, that's 4MB worth of memory 19:10:33 --- join: squrky (~squrky@pc2-oxfd3-3-cust133.oxfd.cable.ntl.com) joined #osdev 19:11:30 --- quit: Ep ("Client exiting") 19:11:50 ullo? 19:12:04 yo 19:12:10 pretty quiet in here people? 19:12:11 Hi squrky 19:12:19 well, there's been some stuff 19:12:22 alright DRF how tricks? 19:12:44 I'm ok thanks, yourself? 19:13:02 yeah ok considering. so you are all os development type people then 19:13:41 Pretty much. U interested in osdev? 19:14:14 well yeah. i'm pretty much into any programming, but for the moment i need to get into OS development to solve a particular problem 19:14:27 which is moer or less why i'm here 19:14:32 O? 19:15:05 I've noticed the learning curve is pretty much vertical man. Anyone else notice that one? 19:15:59 Errm, well it's a long process put it that way. But how long depends on what experience you have with which languages. 19:16:32 --- join: trans (ceezju@fatwire-201-98.uniserve.ca) joined #osdev 19:16:49 don't worry, I haven't come here for a C lesson. I'm a professional coder of 15yrs. BUT I am a bit puzzle. actually its started to hack me off a bit. 19:17:28 But it was only a year just over when I started taking an interest. Thing is not everyone is heavily into osdev so not as many resources as in some areas but you know the normal kinds of probs lol. 19:17:48 You have 15 years, or u´ve been coding for 15 years ?? 19:18:46 I've been coding for about 15 yrs. most system level for the last 7 or so. i've got a good graps of programming mostly in C but the usual others as well. 19:19:06 and yeah, its more or less impossible to find very good information on OS development. 19:19:25 there are a lot of good information ... 19:19:29 asm and c are the most popular here. (some C++ and very little of anything else) 19:19:36 here is the information ... ;) 19:20:16 indeed. actually I need some help if anyone would ablidge. I know its annoying but I'm really starting to pull my hair out here. its IDT related (sigh) 19:20:29 www.osdev.org <- Just browse the link page till you locate a decent site for tutorials lol. 19:21:46 do you have the x86 manual ?? 19:21:51 it´s a good resource ... 19:21:56 Well give us the details of the problem and I'm sure we'll see what we can do in the way of an answer. 19:21:59 thankyou DRF. I've been all for about a week now. I think I must be having trouble taking the information in or something man. One question: Am I on the right track trying to set up an IDT in protected mode, or is there something fundamentally wrong with that? 19:22:01 there u find information about IDT 19:22:29 squrky: nothing wrong with that 19:23:10 I set up mine on pmode .. 19:23:15 geist: what i asked was stupid in retrospect, ignore :) 19:23:53 nullify: ok right, so assuming my boot code is ok (fat12 reader then into protected mode), and my GDT is ok (1 code and 1 data segment 4GB), then I should just be able to go straight aghead and build my IDT in my C kernel right? 19:24:17 yes 19:25:01 right - so i'm not stupid after all! it just keeps throwing page faults man. i do a lidt with exceptions off of course. any idea where i could be going wrong? 19:25:22 nullify: :) 19:25:44 sorry people, i know how sad and annoying this type of thing is 19:25:52 kick me when it all gets too much :) 19:26:04 squrky: check how you initialize the idt entries 19:26:51 check. i combed the john fine docs very closely and I think i have that nailed. anything else you can think of? 19:27:32 I gotta get some coffee, but I'll help ya out when I get back 19:27:38 cheers 19:27:42 looks like you 19:27:48 you're getting close 19:28:04 squrky: check the format of the descriptor you're feeding to LIDT maybe? 19:28:22 i reckon too. i've made some stupid oversight somewhere. i'm not trying anything complex: no paging, single task. its for a sequencer actually. 19:28:44 nullify: right. word length then dword base right? 19:28:56 yes 19:29:20 packed together and then memory reference operand for lidt? 19:29:51 yeah 19:30:18 lidt [idtr] 19:30:56 yep ok... i'm a little bit confused about how exceptions work here. could they be interfering in some way. i am getting page faults 19:30:59 i ask because of the wierd way in which the exception numbers collide with the irqs 19:31:04 g2g people 19:31:14 --- quit: nullify ("Client Exiting") 19:35:30 you mean exception 14? 19:35:48 no bochs says 13 (3rd) exception... 19:36:03 which is that? 19:36:09 oh, the 14th one 19:36:11 page fault 19:36:12 yeah 19:36:16 i'm not sure how to interpret that. yep page fault. 19:36:34 where is it triggering the page fault? 19:36:43 what is it doing when it triggers it 19:36:44 ? 19:36:45 on an sti. how many entries should i have in my idt? one for all 256 iterrupts? 19:36:53 sure 19:37:05 are you programming the pic? 19:37:12 --- join: Xeon (~root@cs666823-113.austin.rr.com) joined #osdev 19:37:14 no. ahha should i be? 19:37:15 not sure what it's base is set to normally 19:37:19 wee 19:37:22 i can find out 19:37:22 linux is fun 19:37:26 that could be an external interrupt that 19:37:36 that's showing up as an exception 19:37:55 normally yould program the pic to set it's base vector to be something above 32 19:38:00 I set it to 32 itself 19:38:10 so that irq 0 == exception 0x20 19:38:30 irq 15 == exception 0x2f 19:38:35 right so i do have to map irq 0-15 to above 32? 19:38:42 yeah 19:38:52 otherwise it wont work> 19:38:53 ? 19:39:06 otherwise I dont know what happens if you get an external interrupt 19:39:18 i see man ok that makes vast amounts of sense 19:39:18 the pic may deliver it to an exception vector offset from 0 19:39:31 14th exception is the first ide controller IIRC 19:39:33 right. 19:39:49 normally you dont turn interrupts on until pretty late 19:40:01 and remember, sti doesn't enable exceptions, they're always there 19:40:02 does anyone have a sound blaster 16 ? 19:40:06 it just enables external interrupts 19:40:33 yeah.. i'm messing around on b8000 and i'm in protected mode... i just want to handle interrupts like mouse keyboard and exceptions etc... 19:40:34 I set up most of my kernel before doing sti, and certainly have the pic and vector table stuff set up before then 19:40:46 yeah, before you do that you have to program the pic 19:40:51 it's not that hard, just a few outs 19:41:06 yea 19:41:10 hold on and I'll find ya some code 19:41:11 ok i think thats the winning information geist. i'm missing out pics. thanks 19:41:57 http://www.newos.org/cgi-bin/fileViewer.cgi?FSPC=//depot/newos/kernel/arch/i386/arch%5fint.c&REV=52 19:42:09 look at arch_int_init 19:42:10 () 19:42:20 it sets up the two pics to be based at 0x20 19:42:27 with all the interrupts masked by default 19:42:33 right ok excellent. 19:43:26 cool 19:43:31 looks like you're pretty close 19:43:59 i've stumbled on lots of apic info also. if its a better chip to use would it be advisable to play with that? (AMD XP 2000+) see the thing is i'm trying to write an os for a dedicated midi sequencer for my studio so the added features will help lots (i.e. timer) 19:44:32 well, if you're talking about the local apic, it's been around for a long time 19:44:49 you *have* to use it for smp machines 19:45:05 otherwise it's got a very very high resolution timer (runs at bus speed, 32-bit counter) 19:45:30 wicked thats what I need. its does the job of the PIC tho right? 19:45:36 no 19:45:39 oh really? 19:45:51 so its just a high res interrupt timer? 19:45:51 no, it's a big chain of legacy stuff 19:46:03 it also lets to sent interrupts to other cpus 19:46:09 which is why you need it for smp machines 19:46:22 otherwise it operates by default in pass thru mode, which is why you haven't had to deal with it yet 19:46:32 the local apic is on the cpu 19:46:34 one per cpu 19:46:37 the pic is on the chipset 19:46:49 and handles isa or legacy interrupts 19:47:13 right i.e. hardware interrupts? 19:47:14 the pic handles priorities and whatnot of the 15 external isa interrupts 19:47:29 but it speaks to the cpu and hands the cpu the highest of it's 15 interrupts 19:47:38 the local apic is on the cpu and by default just passes it through 19:47:54 do i use the PIC to mask unwanted interrupts? 19:48:05 yes, unwanted external interrupts 19:48:25 where it gets confusing is when you can remap external interrupts to bypass the pic, but that's a per-chipset thing 19:48:36 when you say external you are refering to hardware interrupts or just any interrupt that isn't an int call right? 19:48:44 hardware interrupts 19:48:46 external to the cpu 19:48:51 right. so what about exceptions 19:49:04 they are internal to the cpu, you cannot remap them 19:49:16 thats was stupid question wasn't it. 19:49:35 well, it's pretty confusing 19:49:39 * n0vice is away: I'm busy 19:49:40 --- join: Ep (~Ep@211.136.159.235) joined #osdev 19:49:53 it's a bunch of stuff chained together, like a big multiplexor 19:49:54 * n0vice is away: sleeping 19:49:55 --- quit: Xeon (Remote closed the connection) 19:50:00 thanks. ok well i think you've given me enough info to solve my problem. i'll be back more confused in a few days probably :) 19:50:03 it's important to know at what level you're at 19:50:07 okay, sounds good! 19:50:18 cheers. later 19:50:22 --- quit: squrky () 19:50:39 ah, that was cool. it's always nice to help someone out 19:51:56 lol, thats good. Considering the number of times I've been tempted to direct peoples questions your way to save my brain power. ;) 19:53:28 :) 19:53:32 now I have to get coffee 19:53:38 lots of cleaning to do in the apartment 19:53:40 it's a mess 19:54:13 Ouch, yuk cleaning. (Even more yucky tho if not done) 19:55:26 --- quit: DRF (Read error: 54 (Connection reset by peer)) 19:55:37 --- join: DRF (~Daniel@host213-121-70-116.surfport24.v21.co.uk) joined #osdev 19:55:39 --- join: squrky (~squrky@pc2-oxfd3-3-cust133.oxfd.cable.ntl.com) joined #osdev 19:55:41 --- join: dax_ (dax@u212-239-163-194.adsl.pi.be) joined #osdev 19:55:58 geist: NewOS> is this your creation? 19:56:38 --- join: witten (~witten@adsl-gte-la-216-86-199-140.mminternet.com) joined #osdev 19:56:58 --- join: Xeon (~root@cs666823-113.austin.rr.com) joined #osdev 19:57:20 Hmm, as it looks like he's off cleaning his appartment, yep it is his project. 19:57:46 --- quit: alphakiller () 19:57:52 well, tell him its impressive work (not to suck his dick). its good to see good code now and again. 19:58:16 heh, there will be no dick sucking here 19:58:58 nope. my os is called NXOS Algorhythm. Its a more impressive title (IMHO:)) but its does shit all so you win. later. 19:59:12 lol, well geist your the nearest we have to a famous regular here. ;) 19:59:37 nah, air was here longer 19:59:42 he's been slashdotted too 20:00:00 excellent stuff. 20:00:07 --- quit: squrky () 20:00:21 How long have you been here geist? 20:00:50 probably about a year and a half 20:01:10 back in a minute 20:01:28 I've been slashdotted: http://torsion.org/witten/katzdot/ 20:02:39 ERR 20:02:52 KBD BROKEN<<< 20:03:43 lol 20:04:37 DOES ANYONE (SORRY) HAVE DAMN>>> 20:04:59 IVE BEEN TROJANED>>> 20:05:39 Hey, now you've lost me all together Xeon 20:06:06 --- quit: Xeon (Remote closed the connection) 20:13:43 --- quit: dax (Read error: 110 (Connection timed out)) 20:15:55 witten: wtf is that? 20:17:48 * geist is trying to figure it out too 20:20:47 ditto 20:24:09 else echo "  ".$i+1."; 20:24:15 whats wrong with that line? 20:24:24 nm 20:25:29 ok, whats wrong with it after u insert " before 20:26:26 Can you have a =\ like that? I've never seen it if so. 20:26:50 yes 20:27:01 O, sorry, ignore me 20:27:07 the \ is for the " after it 20:27:27 * DRF was thinking in .html terms not programming. 20:27:36 oh 20:29:43 What sort of error/prob do you get from it? 20:31:50 im moving my news pages to php 20:34:39 Ah, I've never needed to mess with php. Sorry can't think of any reason why it shouldn't work. (unless poss a reserved char in the variables?) 20:34:55 i think it was the $1+1 20:35:00 er $i+1 20:35:16 doesnt seem to like doing math inside an echo 20:36:07 Ah, ok 20:36:25 Anyways I think I'll be off now as I need to at least get some sleep 20:36:33 cya 20:36:51 Night 20:37:04 --- quit: DRF ("Good night all") 20:37:45 * geist finally removes all traces of uinderscores in his mp3 collection 20:38:09 heh 20:38:30 as a result, I have to download 5 GB of mp3s to my machine tonight 20:38:31 sigh 20:38:43 eh? 20:38:54 well, I updated my master copy of mp3s, which is on my server 20:39:07 I usually just rsync my machine at home with that one 20:39:12 and rsync my machine at work with it too 20:39:21 ah 20:39:22 but rsync can't tell if the file changed name 20:39:36 though it figures out pretty much everything else 20:40:12 #!/bin/bash 20:40:12 for file in *; do 20:40:12 mv "$file" `echo $file | perl -e "while (<>) { s/ /_/g; print;}"`; 20:40:15 done 20:40:38 in yer case u can swap the space and underscore :) 20:41:00 it's in a deep directory structure 20:41:06 that doesn't work as well as that 20:41:28 recursive execute that script 20:41:30 it's a bit harder, mostly due to the way stuff doesn't work worth a shit with files with spaces already in it 20:41:34 recursively 20:41:40 the for file in * part fails 20:42:01 you end up having to do 20:42:09 for file in `find . bl;ah blah` 20:42:22 and have the find command print the file with quotes around it 20:42:30 etc etc, it's a pain in the ass, but I got a script to do it 20:42:52 it's easier to just update the master copy and retransfer 20:42:57 should finish tonight 20:44:12 if it doesn't I just drop a tarball of the whole archive onto my usb hard drive and carry it home 20:49:54 * geist starts gutting the interrupt code on newos 20:55:48 geist: how do you represent irq 'descriptors' struct irq_desc { int irq, struct list_head *handlers...} 20:56:04 is it a static array of a structure of that kind? 20:56:23 --- quit: spaz ("Client Exiting") 20:56:58 it's a static list of pointers 20:57:10 to interrupt handler structures 20:57:32 how about the actual lines, not handlers 20:57:51 huh? 20:57:54 where the line struct might have info like the kind of PIC its connected to 20:58:17 I guess I dont track that 20:58:22 I'm not totally sure what you're asking 20:58:50 ok lets call a struct describing each irq line 20:58:54 an irq descriptor 20:58:58 or irq_desc for short 20:58:59 line? 20:59:11 is this configuration stuff? 20:59:19 yes say 0,1... 15 on cascaded 8259 20:59:30 I dont track that 20:59:45 not yet at least 20:59:52 hmm ok 21:00:05 i was thinking of switching to linked lists 21:00:24 dynamic data structures are always better than static if you can get away with it 21:00:35 tends to be more representative of the way some other systems have PICs, devices connected 21:00:43 yeah 21:00:53 yeah, an array with 10,000 members was looking ugly 21:01:06 heh 21:01:13 this is actually linux 21:01:25 since its indexed by interrupt number 21:01:32 and interrupt numbers are sparse on some platforms 21:01:59 so the poor sod had interrupt 2000 and nothing inbetween in some cases 21:02:26 i'm currently putting together a hack to dynamically allocated each descriptor 21:02:27 yeah, I'm sure I'll have to do redo my system if I hit something like that 21:02:46 so i was thinking perhaps the linked list stage would save all this pain early on 21:02:57 * geist nods 21:02:58 and just map between irq line and linked list entry somehow 21:05:00 now if only this crap would stop triple faulting =) 21:08:59 yeah 21:26:58 news system is online 21:35:18 --- join: zh (~stoyan@61.114.242.13) joined #osdev 21:37:54 hmm, so it builds 21:38:00 but I know I just introduced a ton of bugs 21:38:04 it'd be too easy 21:39:11 --- quit: zh (Remote closed the connection) 21:48:18 anyone know if i can record what my sound card is playing? (without looping line out into line in) 21:51:14 --- join: redblue (star@ppp058.216-96-207.sherb.mt.videotron.ca) joined #osdev 22:11:17 --- nick: zzzzdog -> minddog 22:33:32 --- quit: zwane_ (calvino.freenode.net irc.freenode.net) 22:33:32 --- quit: SlowCoder (calvino.freenode.net irc.freenode.net) 22:35:15 --- quit: trans (Read error: 60 (Operation timed out)) 22:37:40 --- join: zh (~stoyan@snapgear.silveregg.co.jp) joined #osdev 22:42:42 --- join: zwane_ (User-10418@modemcable092.130-200-24.mtl.mc.videotron.ca) joined #osdev 22:46:36 --- join: SlowCoder (~jja@194.239.250.254) joined #osdev 23:30:44 --- join: witten_ (~witten@adsl-gte-la-216-86-199-140.mminternet.com) joined #osdev 23:45:16 yarg 23:58:21 --- join: nothing (a-o-l@AC833460.ipt.aol.com) joined #osdev 23:59:59 --- log: ended osdev/03.01.05