00:00:00 --- log: started osdev/03.11.02 00:15:23 howdy folks 00:15:29 * geist finishes Usual Suspects 00:15:36 highly highly recommended film 00:24:26 my goddamned college roommate ruined the ending of that for me before I ever saw it 00:26:01 the first time I saw it I had never even heard of the movie, so it was the best possible way 00:27:31 nice 00:32:58 --- quit: CrayT3E (Read error: 110 (Connection timed out)) 00:42:08 --- join: redb1ue (~star@ppp053.216-96-207.sherb.mt.videotron.ca) joined #osdev 00:44:13 --- join: karingo (karingo@11.portland-03-04rs.or.dial-access.att.net) joined #osdev 00:50:34 --- join: dresden (~djdresden@ip-wv-68-119-143-145.charterwv.net) joined #osdev 00:58:27 --- join: wl (philipp@pD954EA21.dip.t-dialin.net) joined #osdev 01:00:02 --- quit: dresden ("BitchX-1.0c20cvs -- just do it.") 01:03:00 --- join: dresden (~djdresden@ip-wv-68-119-143-145.charterwv.net) joined #osdev 01:07:10 --- quit: jwesley (Read error: 113 (No route to host)) 01:09:00 --- join: jwesley (~chatzilla@adsl-155-139-10.mem.bellsouth.net) joined #osdev 01:12:19 --- join: Mathis_ (~anyone1@pD9EAA9B4.dip.t-dialin.net) joined #osdev 01:12:36 hi 01:23:33 yo 01:57:07 * dresden cringes 02:01:57 --- join: [Mathis] (~anyone1@pD9EAA9B4.dip.t-dialin.net) joined #osdev 02:01:57 --- quit: Mathis_ (Read error: 54 (Connection reset by peer)) 02:02:11 <[Mathis]> hi 02:05:09 --- join: frank (~frank@e143139.upc-e.chello.nl) joined #osdev 02:05:12 hiya 02:06:46 --- quit: karingo (""rebooot"") 02:39:17 --- join: Nayxx (Natron@d213-103-75-9.cust.tele2.fr) joined #osdev 02:41:40 --- join: karingo (karingo@120.portland-15-20rs.or.dial-access.att.net) joined #osdev 02:45:55 * dresden ponders rewriting the menuconfig stuff to use a better language... 02:51:04 brainfuck is an elegant language.. 02:51:23 <[Mathis]> and how about assfuck? 02:53:06 --- quit: air ("cria 0.2.9cvs4 -- http://cria.sf.net") 02:54:19 har har 02:55:41 <[Mathis]> stupid names, gell? *g* 03:04:41 --- quit: karingo () 03:24:17 --- quit: redb1ue (Read error: 110 (Connection timed out)) 03:44:08 --- join: quantis (quantis@quantis.demon.co.uk) joined #osdev 03:49:51 --- nick: Prophet__ -> Prophet_ 04:01:34 --- join: eniac (~eniac@38.203-201-80.adsl.skynet.be) joined #osdev 04:04:27 --- join: ToreSB (~tore@062016148138.customer.alfanett.no) joined #osdev 04:04:40 --- quit: ToreSB (Client Quit) 04:04:43 --- join: ToreSB (~tore@062016148138.customer.alfanett.no) joined #osdev 04:11:49 --- quit: ToreSB ("Leaving") 04:11:59 --- join: ToreSB (~tore@062016148138.customer.alfanett.no) joined #osdev 04:25:13 --- join: demise (HjuT@c-854272d5.01-94-7673741.cust.bredbandsbolaget.se) joined #osdev 04:30:37 whys it always so quite in heere these days? 04:31:03 quantis who knows 04:35:22 quantis, I stopped making OT conversation? :) 04:36:22 * quantis runs around slapping everyones arse with a wet fish in an effort to give them something to talk about... 04:37:39 i've spent the morning cleaning up my header file hell in my kernel 04:37:52 also rewrote a bit of the ia32 machdep shit in assembly for the hell of it 04:38:06 masochist 04:38:22 quantis no.. trying for small footprint ;) 04:38:38 later i'm going to thief the vmm from fbsd 04:38:48 and somehow smash that biatch into my kernel.. 04:38:56 did i meantion, freebsd is sexy? :) 04:39:37 funnily enough i dont think you did.... 04:41:05 so dresden what do you think of the heathen freebsd ? 04:41:25 i think freebsd gives me a woody :P~ 04:41:26 --- quit: frank (Read error: 104 (Connection reset by peer)) 04:41:41 or maybe that's the bottle of stoli? :) 04:42:15 sicko 04:42:32 i can write assembly only when drinking vodka 04:43:23 yeah the best code iever seem to write is when im pissed 04:43:30 heh 04:43:32 weird, huh? 04:44:27 i rewrote my message passing code in less than 500 bytes, after being assembled 04:45:10 msgpass-old.o: 5149 04:45:13 msgpass.o: 498 04:45:30 jesus 04:45:37 nows thats pretty damn impressive 04:45:40 first is C, second is asm 04:46:31 what compiler you using 04:46:40 gcc and gas 04:46:51 intel syntax is the antichrist :) 04:46:52 have you tried the intel one? 04:47:02 it works in about half the size 04:47:08 i love intel syntax 04:47:13 Hmm, I have the intel system reference manual somewhere 04:47:29 yeah i got the whole collection 04:47:29 quantis i avoid the antichrist where possible :) 04:47:34 intel sent them free 04:47:43 i liek the anticchrist hes my friend 04:48:02 * dresden tosses quantis a copy of ispell :) 04:48:46 even ispell wont save my lack of typing ability... 04:49:02 i'm listening to weird music, hiding under a blanket in front of this evil machine 04:49:15 hmm 04:49:22 you a nine inch nails fan? 04:49:31 *hugs the CRT for warmth* i really need to a dual athlon for use as a space heater.. 04:49:36 quantis yea 04:49:43 quantis listening to rhea's obsession right now 04:49:58 never heard of them 04:50:33 quantis got dcc's! :) 04:50:46 sorry as browsing 04:50:47 holy hell, it works 04:50:55 np 04:51:07 n like.. if you like it, buy their cd, ok? :) 04:51:24 will do 04:51:40 --- quit: asm2 (Read error: 110 (Connection timed out)) 04:51:48 heh.. i'm in a band.. selling stickers and cd's == lunch money :) 04:52:22 im havign a guitar lesson later 04:52:36 i think i may sleep at some point today 04:52:47 it's been a while :) 04:52:49 <[Mathis]> firmware update for playing guitar completed 2 months ago 04:52:49 just found out the guy living below me is some sort of guitar teacher 04:53:09 sublime == good shit 04:53:27 * quantis inserts a new variant of the CIH virus into mathis's firmware.... 04:53:31 some folks say that smoking herb is a crime.. if they catch you smoking, they're bound to drop the dime.. 04:53:56 ganja! ganja! rah! rah! rah! 04:54:33 lack of sleep is interesting.. 04:57:38 i feel like an old granny 04:57:54 got a blanket wrapped all around me now i feel cosy and at one with the world 04:58:46 quantis, You too? 04:59:43 Except that few grannies code ASM 05:00:25 --- join: frank (~frank@e143139.upc-e.chello.nl) joined #osdev 05:01:50 there must have been quite a few wusing the old mainframes, back i nthe day 05:03:59 dres:sounds like tool but with god-forbid a GIRL!!!!! 05:04:04 quantis :) 05:04:07 quantis, Yeah, but that was FORTRAN 05:04:07 --- quit: idle (Read error: 54 (Connection reset by peer)) 05:04:21 quantis got dcc's! :) 05:04:39 Sometimes machinecode, in BCD or octal 05:04:44 mmmm 05:04:48 bcd .... 05:04:52 haha 05:04:53 <[Mathis]> binary 05:04:56 * quantis creams himself 05:04:57 "eww" 05:05:04 * dresden hugs gcc 05:05:11 * ToreSB hugs PDP7 05:05:17 SHIT! 05:05:29 i found the bug ;) 05:05:38 * ToreSB gets his cheek on a capacitor 05:05:42 BZT 05:05:44 ow 05:06:06 init_free() was releasing a full page... but i forgot to have the linker script page align __init_end 05:06:06 ive blown capcitors up before 05:06:07 blork! 05:06:13 cracked a floescent light one 05:07:00 --- join: Dr_Evil (DSLflat@p508FD551.dip.t-dialin.net) joined #osdev 05:07:45 * dresden feeds this code to indent 05:08:32 * ToreSB huglz the linux 05:08:37 --- join: lodda (~lodda@p508FE99B.dip.t-dialin.net) joined #osdev 05:08:59 --- quit: frank ("code..") 05:09:32 --- quit: ToreSB ("I gotst to reboot biatch!") 05:10:33 quantis got dcc's! :) 05:11:20 now.. i need to write a _good_ console driver.. non of this shitty non-ansi capable, single vt crap i've been tolerating for months 05:13:10 you got a website for you os yet ? 05:13:42 quantis no 05:13:49 write code, then publish :) 05:14:12 write now, the ia32 code is being completely redone and the kernel is getting some long needed housekeeping done to it 05:14:17 this shits good 05:14:57 quantis Check out Raison D'etre (or something similar) good stuff too 05:15:04 im just working on building up an object orientated modle system for mine 05:15:15 and looking at how i can multithread everything 05:15:22 quantis my main goal is portability 05:15:27 and small footprint 05:15:40 i have quite a nice build system ;) 05:15:57 still has a few bugs and dependancy tracking isn't in place for headers, yet. 05:16:04 but it's quite nice other than that 05:16:27 including a modified menuconfig (from linux) which supports using KERNEL= to chose a config file 05:17:01 going to rewrite the menuconfig stuff soon to use a more intelligent scripting system 05:17:11 cool 05:17:33 most of my userland stuff i intend to impment in bytecode 05:18:06 heh, most of my userland will either be a gnu userland of minimal versions of tools (depending on what platform :) 05:18:06 been trying to find out more information on the amiga/de tao virtual processor asm 05:18:22 not planning on having posix support as such 05:18:45 more likely have a smalltalk type interface to it 05:19:11 quantis I'm going for unix-style programming interfaces, with a gui pretty much builtin 05:19:25 unix like APIs, but that's where the similarity will end heh 05:20:13 beos style then ? 05:20:39 quantis dunno, never used it 05:21:16 actually, the more i think about it.. it's a mix of unix and plan9 :) 05:21:27 * quantis comes again.... 05:21:31 i like plan 9 05:21:36 everything, from cpu0's, to ram, to files, to network resources will be objects 05:21:37 doesnt seem to boot in vmware 05:21:47 which can be referenced via the fs 05:21:50 actually thats kind of how im doing mine 05:21:55 s/cpu0/cpu/ 05:22:06 module system im trying to build 05:22:16 I'm using multiboot for such :) 05:22:26 eventually i want absolutely EVERYTHING to be a module 05:22:35 yeah same here 05:22:41 but im concentratingo bnhwo the modules interact 05:22:59 they each export lists of contained objects/methods etc 05:23:05 i've come up with a scheme for module swapping 05:23:24 kernel the has teh register that holds it all, basiclly its just going to be a scheduler and book keeper 05:23:29 modules have ref counts and flags, the flags can include MOD_ACTIVE, MOD_EXCLUSIVE, etc 05:23:35 got bits and pieces of it working 05:23:59 using grub as a loader for all the modules, t hen the kernel cycles through each module and runs its factory/init method 05:24:05 The new module is loaded and flagged active (each module fits a specific subsystem id) 05:24:21 then as the old module losers consumers, it's ref cnt hits 0, it's unloaded 05:24:44 Now in some cases, MOD_EXCLUSIVE is set in the module, to disallow the entire scheme (somethings have to remain exclusive, like ide drivers :) 05:24:58 s/losers/loses/ 05:25:03 you could jsut implment a handover method, 05:25:17 quantis requires too much overhead really 05:25:29 why? 05:25:33 Remember, i have to be concerned much so with size 05:25:57 on most of the targets i intend to support, modules will actually be in pages mapped from rom 05:26:17 but the hadover method would be a single routine that triggers all the cleanup code (which should already exist) and passes all the stanadrised variables for that module over to the new module 05:26:22 in s simple data structure 05:26:55 thats how i was planning on doing something similar in mine but then though thers not that much need for that stuff, and i could jsutrun to libs sie by side for the most pasrt unless they were realyl low level 05:26:56 quantis problem is, most of the modules in my case deal with extremely abstract data types 05:27:20 cants you contian them in a shared libary or is that twso much of an overhead> 05:27:47 plus most modules would be working over generally similar inputs and outputs etc 05:27:56 quantis perhaps once i get the message passing working better, it'll be possible :) 05:28:36 you going to have network message passing 05:28:53 quantis once i get things running nicely, it's on the todo list :) 05:29:06 and since i intend to treat network resources identical to local, it won't be too difficult 05:29:11 still a ways off 05:29:23 right now, you change modules by rebooting ;) 05:29:28 networking is a big task 05:29:37 hey thats another thig you can nick from fbsd 05:29:43 yeup 05:29:52 vmm and net :P~ 05:29:55 (ha.. good luck :) 05:29:58 what are you modules anmade yup as 05:30:33 quantis just an elf object file w/ a __init() symbol 05:30:38 which is called 05:30:43 The rest is up to the module right now 05:31:05 using flat binary at teh moment 05:31:36 what language you writing it in? 05:32:50 c 05:33:03 going to write my own language at some point 05:34:20 <[Mathis]> quantis: bad idea aslong as you have no design goals for it 05:34:32 what writing my own language ? 05:34:38 <[Mathis]> right 05:34:44 i do have design goals for it 05:35:03 its major overall goal is to make it easy for low level system programing 05:35:04 <[Mathis]> those which are not already implemented? 05:35:32 yes because a: it ill be object orientated b: it ill be architecture aware 05:35:36 quantis i eventually intend to also add guaranteed bandwidth crap eventually 05:35:55 s/eventually// 05:35:58 <[Mathis]> C/C++ already implements these goals 05:36:04 i.e you have a model for each architecture which you can overide parts of and things 05:36:36 c++ kind of does, but not very easily, plus tehrs a lot of overhead to get c++ working at as low level as im planning 05:36:48 <[Mathis]> it does very easy 05:37:02 <[Mathis]> also C++ is not for lowlevel 05:37:07 <[Mathis]> it is for application level 05:37:21 <[Mathis]> C can also be used to create object oriented code 05:37:34 yes i know 05:37:40 <[Mathis]> and it is usable for lowlevel coding 05:37:49 quantis you like those bands? 05:37:54 <[Mathis]> I see no need to create a new language 05:37:56 and the compiler i was working on a while back emits c code cause i cant be arsed writing my own bgackend 05:38:50 not so keen on line and distance 05:39:18 time and distance? they're local.. heh pretty cool guy tho 05:40:04 mmm rob zombie... 05:40:12 heh that remix is yummy... 05:40:16 * quantis comes again.. hist pants are getting vey sticky indeed.... 05:40:33 --- join: kernel-panic (rewt@ANice-205-1-22-6.w81-250.abo.wanadoo.fr) joined #osdev 05:40:44 mathis: but im hoping to complete negate teh need to use assembler 05:40:48 that is my main goal 05:40:56 one language to do all the low level programming 05:41:07 <[Mathis]> what does 'teh' mean? 05:41:08 then have a higher level application orientated language 05:41:16 it means i cant spell the 05:41:22 :) 05:41:34 <[Mathis]> does it matter to 'spell' on IRC? 05:41:53 --- join: steve-shaw (steve@164.b.004.brs.iprimus.net.au) joined #osdev 05:41:58 no as a long as people can understand you but opeople seem to be havign increasing dificulty understanding me these dayas 05:42:05 <[Mathis]> your main goal is already solved with the languages C and C++ 05:42:34 <[Mathis]> also there are many other application languages 05:42:36 well i jsut one a single language 05:42:41 i know 05:42:57 <[Mathis]> you want one universal tool for everything? 05:43:14 <[Mathis]> one tool that is unuseable because of its complexity? 05:43:15 but since ill be implmentating a brand new api as part of my os, i need to either port or create from scratch so i though i might create from scratch 05:43:27 imathis, yes but jsut for low level stuff 05:43:50 mathis: it wont be i utilise object orientated technology at a low level 05:44:08 <[Mathis]> dont you lose the overall goal from your mind? 05:44:23 for instance it would have models of processors, so you then write data structures that qwould interface with say the page directory etc 05:44:35 thats why you write them down 05:44:41 <[Mathis]> OS is for servicing applications, applications are tools for the humans 05:44:54 so you can look back and fix new ideas into the old framework 05:44:58 yes 05:45:02 this is true 05:45:26 <[Mathis]> you are fixing on lowlevel stuff, creating new APIs and so on 05:45:29 but since im creating a virtual machine based os theres a big seperation between what the user can use and the system its running on 05:45:32 good morning. 05:45:42 hi elvestone 05:45:47 quantis like that song? 05:46:01 <[Mathis]> why arent you using existing technology? 05:46:02 well that mix of it? :) 05:46:03 sorry hadt realised it id finished 05:46:12 quantis np :) 05:46:21 <[Mathis]> they were invented to solve these problems 05:46:24 mathis i may well do but at the moment its annoying me doing stuff the way im doing it 05:46:29 heh.. that remix made a local club very popular ;) 05:46:37 i can image 05:46:43 doesnt go down so welll over here 05:46:55 <[Mathis]> I dont think so that you can reinvent something much more better than many other people did before 05:47:05 they stop playing it when the club fills up and then start on linken park and bizkit and stuff like that 05:47:12 yucky 05:47:13 i never said it was going to be better 05:47:21 btu it will do things in a way i feel comfortable with 05:47:23 <[Mathis]> but you intended to do so 05:47:34 heh i used ot play a lot of dark trance and such at the underground :) 05:47:41 look at unix, thers god knows how many different shells scripting langugaes etc 05:47:46 most are largely interchangeable 05:47:58 but little quirks etc mae them slightly mroe suitable for one thing 05:48:11 or a kludge if you try doing other sutff with them 05:48:57 mathis:id love it to be beter and for other people to use it and praise me as a god (which is never never going to happen 05:48:59 ) 05:49:13 anyways im off for a bath 05:49:41 <[Mathis]> it is all up to you 05:49:44 hmm 05:49:55 got completly monged last night after the seargents mess do so i think i better clean up 05:50:09 <[Mathis]> warning: this is taking alot of time 05:50:12 dresden: not a godd as a nin remix 05:50:15 mathis: i know 05:50:23 but i never said it was going to be quick 05:50:27 quantis check that mp3 out :) 05:50:28 <[Mathis]> which could be used for much more productivity 05:50:30 it's good NIN 05:50:36 apparently still unreleased 05:50:42 what is ? 05:50:51 Maybe Just Once 05:50:55 nin got a new album coming soon 05:51:04 quantis i've had this track since.. 98? 05:51:06 :) 05:51:13 haven't seen it on any cd's yet.. :) 05:51:28 mathis: i know but i only do this for fun and personal enjoyment 05:51:49 i gotta get the boot rom working on PiXIe 05:52:01 that's going to be a lovely fun process 05:52:34 and the port that'll run on PiXIe is likely going to end up 99% assembly 05:52:56 pixle? 05:53:26 quantis pixie, a 4 processor 8051 machine (using the dallas/maxim ds89c420 @ 33mhz) 05:53:34 heh 05:53:43 hahah 05:53:47 i found out recently the ds89c420 can be pushed to 50mhz tho 05:54:39 bloody hell 05:54:49 there eight bit processors though arnt they? 05:55:06 yeup :) 05:55:07 --- join: mrMister (~andri@ti122110a080-6601.bb.online.no) joined #osdev 05:55:30 bloody hell 05:55:34 you really are a a masochist 05:55:58 rightim off ofor a bath and to change my sticky pants 05:56:02 --- nick: quantis -> q_bath 05:57:05 --- join: CrayT3E (~CrayT3E@finsler.math.uoa.gr) joined #osdev 05:58:14 --- quit: lodda () 06:02:35 --- join: idle (~idle@208.27.0.86) joined #osdev 06:02:54 hello 06:03:03 olleh 06:05:28 hmm 06:05:59 mmh 06:08:04 * dresden hugs his nifty lil syscall interface 06:11:07 rum 06:12:19 0 06:14:32 <3 unix syscalls :) 06:14:39 abuse that stack baybee! 06:15:24 puts your hands up this is a stack up..... 06:15:36 the stack is yo biatch man. 06:15:54 * gab didn't suspect a syscall interface could make an erection to some ppl 06:16:23 * q_bath comes in his fresh stet of pants... 06:16:40 karggg i ve got to get away from here otherwise my washing machine ill end up douign over time 06:16:47 hi wnatis 06:16:47 ¨! 06:17:42 what's up quantis? 06:17:50 you mur 06:18:03 apparently syscalls are the menu of the day :) 06:19:54 --- nick: q_bath -> q_work 06:20:31 <[Mathis]> quantis: you're in your bath, so dont talk here ;-) 06:22:35 oh jesus.. staring at intel syntax has confused me! 06:22:43 source left, dest right for intel syntax, right? 06:25:11 --- join: Ubel (~Ubel@194-144-42-191.xdsl.is) joined #osdev 06:27:05 err no :) 06:29:02 <[Mathis]> mov eax, 12 06:29:11 <[Mathis]> whats confusing on this? 06:30:16 [Mathis] just used to at&t :) 06:30:30 <[Mathis]> you're not flexible enough ;-) 06:30:59 --- nick: Ricardo -> Rico 06:31:22 Am I dead or alive? 06:31:35 <[Mathis]> none of both *g* 06:31:43 undead? 06:31:52 <[Mathis]> and unalive *g* 06:37:19 --- quit: steve-shaw (Connection timed out) 06:44:35 what's a good script for doing 'source webs' on a page, with links that take you to the definition of each function/type/etc, i've seen em before and wanna set up such a thing 06:47:38 --- quit: Nayxx ("Illegal call") 06:47:52 dresden you might want to use doxygen 06:49:05 hmmm 06:49:19 will check it out, thanks :) 06:50:31 bbiab, gonna go shower 06:50:43 then start on adding ansi crap to the console driver ;) 06:51:15 --- nick: cookin-fucked -> cookin 06:51:35 hmmmm 06:54:57 nethack installation fscked 06:55:32 i can't remove the misinstalled files, damn installed 06:56:06 --- join: wcstok (strtok_r@dialup-67.31.178.95.Dial1.Denver1.Level3.net) joined #osdev 06:59:51 --- join: ToreSB (~tore@062016148138.customer.alfanett.no) joined #osdev 07:04:29 anyone know where i can get a program that ill format disks to 1.76mb / 07:04:41 <[Mathis]> VGACOPY 07:04:57 <[Mathis]> freeware 07:05:15 cool 07:05:19 * q_work hugs mathis 07:05:33 * mur tries playing nethack (graphical version) 07:05:34 <[Mathis]> you know Star Trek - Deep Space 9 ? 07:05:45 ahh.. it works :) 07:05:46 yeah 07:05:53 <[Mathis]> it is running right now on TV 07:05:59 <[Mathis]> the title music 07:06:14 <[Mathis]> I have an extended version of it on my harddisk 07:06:53 <[Mathis]> seems I am able to press the play button at exact the same time the title music begins 07:07:20 <[Mathis]> it then runs perfectly synchronized 07:07:44 does vga copy work on windows 2000? 07:07:56 <[Mathis]> dunno 07:08:02 <[Mathis]> it is a very old program 07:08:27 !seen slowcoder 07:08:37 Hmm, no such bot. 07:08:44 ToreSB try /msg seenserv seen 07:08:45 :) 07:09:15 Hmm... Slowcoder is on IRC right now! 07:09:32 <[Mathis]> ToreSB: try /whois 07:09:41 <[Mathis]> every other info is not senseful 07:09:48 thanks mathis 07:09:56 dresden, you suck. ;) 07:10:44 <[Mathis]> it is also not senseful to know if someone has been here hours/days ago 07:22:27 [Mathis] to jmp back 1 instruction, $-1? 07:22:35 say halt: cli; hlt; jmp $-1 07:22:37 ? 07:23:22 <[Mathis]> depends if the assembler knows $ 07:23:31 gas 07:23:49 cli hlt die: jmp die 07:24:06 <[Mathis]> wcstoks code works 07:24:10 In PDP-7 asm it's .-1 07:24:13 no it doesn't 07:24:27 That crashes the comp 07:24:34 <[Mathis]> heh 07:24:44 it won't crash in supervisor mode =P 07:24:48 die:hlt jmp die 07:25:01 <[Mathis]> this is okay 07:25:07 blah.. i'm trying to save a byte here :) 07:25:15 <[Mathis]> heh 07:25:49 actually 4 bytes iirc :) 07:26:02 <[Mathis]> really... 07:26:16 <[Mathis]> memory is so expensive today... 07:26:25 better memorise everythign self 07:26:27 [Mathis] embedded things :) 07:26:55 <[Mathis]> no reason 07:27:11 <[Mathis]> my embedded thing has 64 MB RAM and 48MB Flash ROM 07:28:04 <[Mathis]> saving one byte doesnt matter... 07:28:06 mine has 256kb flash 07:28:11 and 16mb of nvram 07:28:13 <[Mathis]> buy a new one ;-) 07:28:19 i built mine! 07:28:26 <[Mathis]> rebuild it 07:28:29 ;) 07:29:18 mathis how do you use vgacopy to format 175mb discs? 07:29:26 1.76 even 07:29:36 <[Mathis]> I start it 07:29:40 <[Mathis]> insert the floppy disk 07:29:45 <[Mathis]> set the floppy drive 07:29:50 <[Mathis]> set the format size 07:29:53 <[Mathis]> and start formatting 07:29:56 yeah but wheres the option to select oramt size ? 07:30:06 <[Mathis]> button Format 07:30:41 I wanna build a computer in 74xx lofig 07:30:47 logic* 07:30:55 <[Mathis]> 7400 ? 07:31:12 A huge rack *dr00l0r* 07:31:37 it doesnt seem to do anything 07:31:44 <[Mathis]> how about mechanical relais instead of transistors? 07:32:03 <[Mathis]> you might also overclock it to 10kHz *g* 07:32:37 --- join: steve-shaw_ (steve@098.a.002.brs.iprimus.net.au) joined #osdev 07:32:43 <[Mathis]> you always have an accoustic feedback of every relay :-) 07:33:16 --- quit: demise () 07:33:43 bbl 07:33:51 <[Mathis]> babbl 07:33:54 going to pet store :) 07:34:09 [Mathis] don't forget to randomly abuse people for me while i'm gone, ok? 07:34:15 <[Mathis]> no 07:34:21 yes 07:34:27 <[Mathis]> no 07:34:30 yes 07:34:32 <[Mathis]> no 07:34:33 idle ok, you do it for me, k? 07:34:33 <[Mathis]> no 07:34:39 yes 07:34:39 yes 07:34:40 <[Mathis]> no 07:34:42 yes 07:34:42 yes 07:34:42 <[Mathis]> no 07:34:43 <[Mathis]> no 07:34:43 no 07:34:44 <[Mathis]> no 07:34:44 yes 07:34:45 <[Mathis]> no 07:34:45 yes 07:34:45 yes 07:34:46 <[Mathis]> no 07:34:47 yes 07:34:48 <[Mathis]> no 07:34:48 yes 07:34:48 <[Mathis]> no 07:34:49 yes 07:34:50 <[Mathis]> no 07:34:50 yes 07:34:50 lol. 07:34:51 <[Mathis]> no 07:34:51 yes 07:34:52 <[Mathis]> no 07:34:53 yes 07:34:54 <[Mathis]> no 07:34:54 yes 07:34:54 yes, no, yes, yes, no. 07:34:54 yes 07:34:54 moo 07:34:55 <[Mathis]> no 07:34:56 yes 07:34:57 <[Mathis]> no 07:34:57 yes 07:34:57 CUT IT! 07:34:57 yes 07:34:58 :p 07:34:58 yes 07:34:59 <[Mathis]> no 07:35:00 yes 07:35:01 <[Mathis]> no 07:35:01 hehe. 07:35:02 yes 07:35:03 <[Mathis]> no 07:35:04 yes 07:35:04 shut up! 07:35:05 <[Mathis]> no 07:35:06 yes 07:35:07 <[Mathis]> no 07:35:08 yes 07:35:09 <[Mathis]> no 07:35:09 yes 07:35:10 ok.. this has gone too far.. 07:35:11 <[Mathis]> no 07:35:11 yes 07:35:12 *farts( 07:35:12 retards. 07:35:13 <[Mathis]> no 07:35:15 <[Mathis]> no 07:35:15 yes 07:35:16 yes 07:35:17 <[Mathis]> no 07:35:18 yes 07:35:19 <[Mathis]> no 07:35:20 yes 07:35:21 <[Mathis]> no 07:35:22 MAYBE! 07:35:22 yes 07:35:23 <[Mathis]> no 07:35:24 MAYBE! 07:35:24 stop :( 07:35:24 yes 07:35:25 <[Mathis]> no 07:35:26 yes 07:35:27 <[Mathis]> no 07:35:27 pleeease.. stop! 07:35:28 yes 07:35:29 <[Mathis]> no 07:35:31 yes 07:35:31 <[Mathis]> no 07:35:33 yes 07:35:33 <[Mathis]> no 07:35:34 dresden: It involves Mathis. Of course it's gone too far. 07:35:34 elvstone: agreed 07:35:35 yes 07:35:35 <[Mathis]> no 07:35:36 yes 07:35:37 <[Mathis]> no 07:35:38 yes 07:35:39 * dresden busts out some madd /ignore 07:35:39 <[Mathis]> no 07:35:41 <[Mathis]> no 07:35:41 yes 07:35:42 yes 07:35:43 <[Mathis]> no 07:35:44 yes 07:35:45 <[Mathis]> no 07:35:46 yes 07:35:47 <[Mathis]> no 07:35:48 yes 07:35:49 <[Mathis]> no 07:35:50 yes 07:35:51 <[Mathis]> no 07:35:53 yes 07:35:53 <[Mathis]> no 07:35:54 yes 07:35:54 :( this makes me feel sad inside :( 07:35:55 <[Mathis]> no 07:35:55 yes 07:35:57 <[Mathis]> no 07:35:58 yes 07:35:59 <[Mathis]> no 07:36:00 yes 07:36:01 <[Mathis]> no 07:36:02 yes 07:36:03 <[Mathis]> no 07:36:04 yes 07:36:05 <[Mathis]> no 07:36:06 yes 07:36:07 <[Mathis]> no 07:36:08 yes 07:36:09 <[Mathis]> no 07:36:10 yes 07:36:11 <[Mathis]> no 07:36:12 yes 07:36:13 <[Mathis]> no 07:36:14 yes 07:36:15 <[Mathis]> no 07:36:16 yes 07:36:17 <[Mathis]> no 07:36:18 yes 07:36:19 <[Mathis]> no 07:36:19 idle! 07:36:20 yes 07:36:21 <[Mathis]> no 07:36:22 yes 07:36:23 <[Mathis]> no 07:36:24 yes 07:36:25 <[Mathis]> no 07:36:26 yes 07:36:27 <[Mathis]> no 07:36:28 *!* Ignoring ALL messages from idle, [Mathis] 07:36:29 idle and [Mathis], idle 07:36:29 <[Mathis]> nope, nope, nope, nope 07:36:31 <[Mathis]> :-) 07:36:33 <[Mathis]> buy 07:36:39 me too, have they stopped? 07:36:41 fly 07:37:04 i dunno but that was eating precious bytes on my log partition 07:37:05 silence 07:37:11 >:) 07:37:13 They've stopped. 07:37:19 ok, great. 07:37:25 you really are anal about saving every bit arnt you dresden :) 07:37:29 who logs this crap? 07:37:41 q_work it was an embedded joke :) 07:37:52 but my home dir is getting near quota :( too much pr0n 07:37:59 too much info 07:39:28 --- quit: steve-shaw_ ("ChatZilla 0.9.35 [Mozilla rv:1.5/1]") 07:40:19 --- quit: Dr_Evil () 07:57:03 --- join: lodda (~lodda@p508FE99B.dip.t-dialin.net) joined #osdev 08:01:07 --- join: frank_ (~frank@a238068.upc-a.chello.nl) joined #osdev 08:10:28 --- quit: wcstok (Read error: 54 (Connection reset by peer)) 08:39:09 --- quit: dresden (Read error: 60 (Operation timed out)) 08:41:08 --- quit: lodda (Read error: 110 (Connection timed out)) 08:42:59 --- nick: frank_ -> frank__ 08:43:15 --- nick: frank__ -> frank___ 08:44:13 --- nick: frank___ -> frank 08:55:11 --- join: vitamin-X (vitamin-X@213.138.100.231) joined #osdev 09:05:34 anything decent on telly tonight ? 09:08:40 * file[bed] arghs 09:09:42 q_work, yes brand new simpsons and Packers vs. Vikings 09:13:41 --- join: grrrrr (~jim@1Cust234.tnt5.sfo8.da.uu.net) joined #osdev 09:14:54 --- join: asm (~asm@dsl-082-082-149-221.arcor-ip.net) joined #osdev 09:19:13 packersvs vikings ? 09:21:44 --- quit: Ubel () 09:22:53 yes sunday night game 09:24:56 --- join: Kurt (weberk@mrdh1395pc2.ics.purdue.edu) joined #osdev 09:27:06 --- quit: frank (Read error: 104 (Connection reset by peer)) 09:30:40 --- join: clovis (~clovis@lns-p19-4-82-65-15-59.adsl.proxad.net) joined #osdev 09:43:55 good night 09:45:19 nite 09:46:32 --- quit: ToreSB (Remote closed the connection) 09:48:57 --- join: dresden (~djdresden@ip-wv-68-119-143-145.charterwv.net) joined #osdev 10:00:07 --- join: demise (HjuT@c-a64272d5.01-94-7673741.cust.bredbandsbolaget.se) joined #osdev 10:07:44 --- join: Nayxx (Natron@d213-103-119-133.cust.tele2.fr) joined #osdev 10:15:56 --- join: frank (~frank@a238068.upc-a.chello.nl) joined #osdev 10:17:17 gmorning 10:17:54 gevening 10:22:45 --- join: ToreSB (~tore@062016148138.customer.alfanett.no) joined #osdev 10:24:08 nak 10:25:20 ENQ 10:28:09 --- quit: Kurt (Client Quit) 10:28:40 --- quit: wl ("Quit") 10:30:27 --- nick: file[bed] -> file[desk] 10:31:57 --- quit: Nayxx (Read error: 54 (Connection reset by peer)) 10:32:14 --- quit: demise (Read error: 54 (Connection reset by peer)) 10:37:44 --- join: demise (HjuT@c-a64272d5.01-94-7673741.cust.bredbandsbolaget.se) joined #osdev 10:42:39 --- quit: Smari (Excess Flood) 10:43:11 --- join: air (~brand@12-210-160-198.client.attbi.com) joined #osdev 10:43:50 --- join: Smari (~spm@optimized.bitcode.org) joined #osdev 11:02:22 --- quit: debug (Remote closed the connection) 11:03:01 --- join: debug (~debug@tab.csbnet.se) joined #osdev 11:07:44 CBS 11:13:30 --- join: dresden_ (~djdresden@ip-wv-68-119-143-145.charterwv.net) joined #osdev 11:14:22 --- quit: dresden (Read error: 104 (Connection reset by peer)) 11:15:46 --- nick: dresden_ -> dresden 11:18:55 --- join: lodda (~lodda@p508FE99B.dip.t-dialin.net) joined #osdev 11:21:37 --- quit: lodda (Client Quit) 11:24:50 --- quit: air (orwell.freenode.net irc.freenode.net) 11:24:50 --- quit: Odin- (orwell.freenode.net irc.freenode.net) 11:25:11 --- join: air (~brand@12-210-160-198.client.attbi.com) joined #osdev 11:25:11 --- join: Odin- (~sbkhh@adsl-2-216.du.snerpa.is) joined #osdev 11:25:39 --- join: newbs (newbs@ts1-illavl77.shawneelink.net) joined #osdev 11:31:31 --- quit: newbs (Client Quit) 11:33:17 --- join: grrrrr_ (~jim@1Cust192.tnt35.sfo8.da.uu.net) joined #osdev 11:34:11 --- quit: eniac ("Lost terminal") 11:34:25 --- quit: grrrrr (Nick collision from services.) 11:35:35 --- join: newbs (newbs@ts1-illavl77.shawneelink.net) joined #osdev 11:40:28 --- nick: grrrrr_ -> grrrrr 11:48:09 hi 12:01:45 debug, YO! what up fresh homie straight fr'da hood!?!?!? 12:19:26 --- join: lynx (~lodsb@pD9544A3B.dip.t-dialin.net) joined #osdev 12:27:02 --- quit: vitamin-X (Read error: 104 (Connection reset by peer)) 12:37:17 --- quit: dresden (Read error: 110 (Connection timed out)) 12:42:26 --- join: Dr_Evil (~dos4gw@p508FD551.dip.t-dialin.net) joined #osdev 12:42:33 --- join: lodda (~lodda@p508FE99B.dip.t-dialin.net) joined #osdev 12:51:51 --- quit: witten ("Client exiting") 13:18:12 --- quit: lynx ("Client Exiting") 13:26:44 --- quit: lodda (Read error: 60 (Operation timed out)) 13:28:00 --- join: dresden (~djdresden@ip-wv-68-119-143-145.charterwv.net) joined #osdev 13:28:21 this is crazy.. 13:31:29 --- quit: idle ("Sic Sempere Tyrannis!") 13:33:38 --- join: idle (~idle@ppp-208-27-1-162.kinex.net) joined #osdev 13:33:45 --- quit: Dr_Evil (Nick collision from services.) 13:34:03 --- join: Dr_Evil (DSLflat@p508FF250.dip.t-dialin.net) joined #osdev 13:40:07 --- join: Kurt (~weberk@mrdh1395pc4.ics.purdue.edu) joined #osdev 13:48:38 --- join: witten (~witten@adsl-gte-la-216-86-199-140.mminternet.com) joined #osdev 13:49:13 --- join: mur_ (~mur@kyberias.uiah.fi) joined #osdev 13:49:17 ARGH 13:49:27 #command "Who are you? (Unable to look up login name)" 13:49:38 sounds scary! 13:50:20 --- join: voider (~voider@24.203.10.226) joined #osdev 13:50:23 * mur_ just found out he has next week off from school! vacation! :D 13:50:32 no, not :D but :/ 13:50:50 --- join: vitamin-X (vitamin-X@213.138.100.231) joined #osdev 13:51:25 anyone here? 13:51:27 :P 13:53:09 --- quit: ToreSB (Remote closed the connection) 13:53:37 --- join: ToreSB (~tore@062016148138.customer.alfanett.no) joined #osdev 13:53:58 vitamin-X :) 13:54:02 you're here too? 13:54:14 hehe 13:54:31 i came here for aros 13:54:35 :)) 13:54:52 what is aros? 13:55:40 oh 13:55:42 it look nice 13:55:45 can paste url here / 13:55:50 http://www.aros.org/ 13:55:55 hello voider and yet unknown vitamin-X 13:56:11 hello mur_ 13:56:32 yes it is but i cannot get it working with me yet lol 13:58:10 hi mur 13:58:20 i need some vitamin-X 13:58:48 idle, try nethack 13:58:54 lol 13:59:49 --- join: lynx (~lodsb@pD9544A3B.dip.t-dialin.net) joined #osdev 14:02:07 --- join: jsr (www@du-12-62.ppp.telenordia.se) joined #osdev 14:04:18 i think i'll start idling here from now :) 14:04:28 heh 14:04:40 anyone played nethack? 14:04:41 voider, only idle can do that 14:05:01 --- join: df (~yakumo@host81-132-91-12.in-addr.btopenworld.com) joined #osdev 14:05:06 mur_, maybe 14:05:21 a problem 14:05:27 what the hell? 14:05:33 eves 14:05:33 i have some damn heavy chest, what the hell shoudl i do with it to get it open? 14:05:48 idle: does it mean that nobody can use void but voider? 14:05:48 hack it 14:06:01 shoudl i get a key? or is there way to open it with pick-axe? 14:06:14 private volatile ulong32_t *machdep_setup(void); 14:06:14 voider, void doesnt apply to this channel 14:06:16 mur: practice being houdini and lock yourself in it... 14:06:16 private? :) 14:06:53 apparently no nethackers here 14:06:57 --- quit: witten ("Client exiting") 14:07:05 nethack! 14:07:08 * df the tourist 14:07:11 mur_, play real games! what are these dos based crap 14:07:28 * dresden smokes a joint and resumes fixing the gdt crud :) 14:08:13 mur_, #nethack 14:09:10 --- join: kyelewis (~kye@dsl-47.71.240.220.lns02-dryb-mel.dsl.comindico.com.au) joined #osdev 14:09:36 well i died anyways, so who cares anymore ;) 14:09:40 should go sleep 14:09:46 hi 14:09:51 debug 14:10:03 ToreSB: not coding right now, watching voyager 14:10:06 mur_: hi 14:10:08 --- join: mjh (isildur@tel22.newn.cam.ac.uk) joined #osdev 14:10:11 mur_: why the _? 14:10:13 heh 14:10:15 heyas 14:10:35 clone 14:10:38 yo 14:10:45 * debug continues watching 14:10:46 ho 14:12:36 debug? 14:12:41 there should be 1 universal compiler with no competitors. therefore no differeing syntax, etc!!! dammit.. 14:12:49 :P 14:13:10 there should be 1 univeral toaster 14:13:15 that woudl be visual basic of course, and there woudlnt be but one day until new compilers existed 14:13:25 :P 14:13:35 oh btw, hi key lewis 14:13:36 hmm 14:13:43 lol 14:13:46 key lewis? 14:13:55 going to asm for lots of stuff is fun :) 14:14:09 heh.. saving lots of spare bytes :) 14:14:15 hehe 14:14:24 especially for thingsl ike machdep_setup(), it returns nothing and takes no args 14:14:28 and it makes you seem more 1337 too ;) 14:14:35 kyelewis lol 14:14:39 sleep required in a second 14:14:51 i just decided all the machdep stuff is eventually going to end up in asm 14:15:01 since it's a worthwhile place to save some spare bytes :) 14:15:44 by going with nicely optimized asm in the machdep code, i should at least makeup for the few extra bits of overhead for calling the non-machdep wrappers :) 14:16:04 * kyelewis times "a second" to mean "5 hours" just for mur_ :) 14:16:08 what do you all think? :) 14:16:24 since the machdep stuff isn't portable anyways, optimize the hell out of it? :) 14:16:31 * mur_ runs away 14:16:39 in a second 14:16:40 ;) 14:17:18 heh 14:17:25 --- quit: frank ("sleep") 14:17:55 machdep_setup() should be an alias to a python module 14:18:22 df no thanks :) 14:18:58 i've managed to shrink the kernel down to 45kb for everything but network 14:18:59 hmm sounded good to me 14:19:12 well net/sound :) 14:20:44 sound? in a kernel?? what do you think your doing? making some multimedia home entertainment control center?? 14:21:08 will it support my adlib gold?? :) 14:21:17 df nope, just my cmi pci (ac97) 14:21:31 --- join: witten (~witten@adsl-gte-la-216-86-199-140.mminternet.com) joined #osdev 14:21:50 mur_: I play nethack 14:22:09 witten: how do you open the chests and boxes? 14:22:15 df the kernel needs to eventually be usable enough to where it can do self-hosted development 14:22:21 which means i have to have mp3s ;) 14:22:33 mur_: o 14:22:39 no music, no code :) 14:22:41 like doors? 14:22:46 o . ? 14:22:52 dres: hahahah 14:22:56 mur_: the "o" key 14:22:58 df what? :) 14:23:09 followed by the direction 14:23:13 or "." for here 14:23:14 witten: and what are the statues for? 14:23:17 mp3 startup tune? ;) thatd be cool 14:23:33 mur: remember to feed your lil doggy too! 14:23:40 mur_: various things. sometimes they come to life. sometimes if you destroy them you can find a spellbook inside them. if you're a wizard I think you can cast stone-to-flesh on them 14:24:08 so apply e.g. pick-axe to statue to try and see what's inside? 14:24:36 sure, if you're looking for spellbooks 14:24:56 even if you're not a magic user, you can sell the spellbooks 14:25:01 i'd be more worried bout finding suitable dog food 14:25:06 than spellbooks 14:25:14 df one in cans? ;) 14:25:33 find a store and the dog is your BESTEST EVER friend ;) 14:26:12 hmm i think i shoudl go sleep 14:26:25 and AVOID nethack, because i shoudl read 200 pages :/ 14:26:28 sleep is for the week 14:26:33 play nethack all weekend 14:26:34 in 3 days and make notes for 14:26:44 it's alreaedy monday here :) 14:26:55 :P 14:27:12 nethack sucks. angband is bad.. crawl is shitzors. 14:27:25 is shitzors good or bad? 14:27:35 shitting razors = nasty bad 14:27:46 so what game is good, then? 14:28:04 underdark. cough by some git called df ;) only its not finished. hahah 14:28:14 heh 14:28:18 roguelike? 14:28:18 some of the angband variants are ok. 14:28:21 omega is cool 14:28:55 too many bugs in crawl 14:28:57 df :) i could write a module for that :) 14:29:03 i need a pretty splash screen :) 14:29:22 pretty splash screen?! 14:29:26 i have a box which has a nasty HP logo at boot 14:29:28 and i wanna replace that 14:29:43 perhaps with a big fbsd daemon :) 14:30:00 16 colour and 100x100 pix? 14:30:16 was it anyone here that was working on the vmware fix for newos? I think it was slowcoder 14:30:23 I forget who 14:30:38 some new? 14:30:55 eh? 14:31:11 i go sleep 0.30 am and need to get up before 8 am :) 14:31:20 (7.30 am is the plan) 14:31:32 good night 14:31:34 --- quit: mur_ ("zzz") 14:31:54 mur it's full screen, no idea what res, most likely 16 colour yes 14:31:58 it'd say 320x240 14:32:09 it's a hp pavilion 14:32:40 can't complain for a free 366 that i slapped a 550 upgrade in 14:33:13 --- quit: witten ("Client exiting") 14:33:26 wow.. i was a bit lagged for a minute :) 14:33:34 me too 14:33:54 my downstream was max'd out, cvsup'ing 14:34:08 laters peeps. time to spend a few mins coding to keep up my monthly code pace ;) 14:34:09 --- quit: df () 14:35:16 lol 14:50:20 --- quit: Kurt (Client Quit) 14:51:00 wow 14:58:15 --- quit: clovis (Remote closed the connection) 15:02:30 --- quit: jsr (Read error: 110 (Connection timed out)) 15:06:34 --- join: damieng (~damien@user51.res-fi2.jtibs.net) joined #osdev 15:07:47 --- quit: newbs ("printf("%d\n", EOF);") 15:08:17 --- quit: q_work (Read error: 104 (Connection reset by peer)) 15:09:00 SVGAFifo: Capabilities violation for RECT_BITMAP_COPY (skipping it) 15:09:06 mean anything to anyone re: VMWare 4? 15:10:11 --- quit: Divine (Read error: 104 (Connection reset by peer)) 15:10:24 hmm 15:11:06 Specifically with regards to implementation of a frame buffer. 15:11:12 well, calls on top of one. 15:11:55 no idea :( 15:12:59 --- quit: demise ("wtf r u lookin at moron? YEH U !!!") 15:13:50 --- join: newbs (newbs@ts1-illavl77.shawneelink.net) joined #osdev 15:30:07 yo 15:30:13 <[Mathis]> no 15:30:25 me got chain lube 15:31:38 --- join: coredump (~ben@h00105a271fda.ne.client2.attbi.com) joined #osdev 15:31:47 --- quit: coredump (Read error: 104 (Connection reset by peer)) 15:33:22 --- join: Divine (~john@12-246-112-182.client.attbi.com) joined #osdev 15:35:06 --- join: grrrrr_ (~jim@1Cust179.tnt5.sfo8.da.uu.net) joined #osdev 15:35:40 --- quit: grrrrr (Nick collision from services.) 15:35:48 --- nick: grrrrr_ -> grrrrr 15:37:49 --- quit: [Mathis] ("User pushed the X - because it's Xtra, baby") 15:44:52 --- join: kryckan (~kryckan@h170n3fls20o1062.bredband.comhem.se) joined #osdev 15:50:00 --- quit: debug (Remote closed the connection) 15:50:24 --- join: debug (~debug@tab.csbnet.se) joined #osdev 16:00:12 * kyelewis looks around 16:09:45 --- quit: Dr_Evil () 16:10:24 hiya kyelewis 16:23:05 --- quit: kernel-panic ("zzz") 16:28:22 --- quit: mrMister ("gone") 16:30:33 * ToreSB throws herring leftovers at kyelewis 16:42:09 --- join: redblue (~star@ppp151.216-96-207.sherb.mt.videotron.ca) joined #osdev 16:42:56 --- quit: kryckan ("kernel panic!!!") 16:57:45 --- quit: ToreSB ("Leaving") 17:00:28 --- join: ToreSB (~tore@062016148138.customer.alfanett.no) joined #osdev 17:02:32 * debug does some experiments 17:03:12 * debug replaces the TLBWR (tlb write random entry) instruction with an implementation which seeks out the not-last-used entry instead of a random one 17:03:23 to see what the actual performance gain would be 17:05:23 indeed 17:05:40 do you mean not-last-inserted? 17:05:48 how would you know which one was last used? 17:06:20 ohhhh this is the emulator 17:06:25 duh, the emulator can know 17:06:47 the emulator, yes. 17:07:02 I've just read a paper which says that Alpha's PALcode does just that 17:07:08 (keeps track of not-last-recently accessed) 17:07:11 in other words, when they issue tlbwr, you will replace the oldest used one 17:07:19 not the order in which they are inserted, but accessed 17:07:20 hmm, I wonder how they do that? 17:07:30 * debug too 17:07:41 replace the oldest one, yes. 17:07:44 unless the hardware has some way to do that 17:08:58 if the difference is large enough, it would probably be worth it. a "linked list" of some sort? 17:09:09 accessing an entry would put it first (or last) in that list 17:09:13 * debug really has no idea :) 17:09:20 sure, but how do you update the linked list? 17:09:30 those would be special parts of the TLB entries 17:09:34 updated by hardware, perhaps 17:09:35 I don't know 17:09:53 maybe hardware can keep a generation number that it increments each access 17:10:00 pointer that's mucked to wrap around? :) 17:10:13 and updates the entries. so to find the oldest one you find the smallest generation number in each entry 17:10:20 my test now is to: boot NetBSD/pmax, drop out of the sysinst program, run ls -lRa, and run halt. (all commands entered before they are executed, so no delays waiting for keyboard input). I'll compare the number of executed instruction. 17:10:21 s. 17:10:52 actually that generation number thing would work well for you too, since it involves optimizing the common case (tlb access) 17:11:44 right now I'm just setting a last_used field of each tlb entry to the current COUNT register value 17:11:53 (MIPS has a count register, updating for each clock tick or instruction) 17:11:57 and it'd be easy to optimize that further where you only update the generation number when you access a a different tlb 17:12:06 debug: the generation number thing I was talking about is precisely the same 17:12:29 *sighs* 17:13:17 yes 17:13:50 this is not for performance in the emulator, though, just to see what such a "smart" random replacement would do 17:13:53 i can't figure out why the hell this compile is breaking.. 17:14:08 obj/GENERIC/include/machine/ipc.h:79: syntax error before `IPC_Options' 17:14:19 79: typdef unsigned long IPC_Options; 17:14:23 dresden: probably some sort of syntax error 17:14:31 err typedef 17:14:41 i typo'd manual c&p'ing it over there :) 17:14:44 no mouse heh 17:16:20 --- quit: mjh () 17:17:52 R2000: R4000: 17:17:53 old (decrementing) RANDOM: 79062880 70132971 (nr of cycles) 17:17:53 new (smart) replacement: 78840223 69872476 17:18:03 so, the performance gain is not very large 17:19:01 * file[desk] disco dances 17:19:07 or course, perhaps 15M instructions were run in kernel space only, before booting into userland, and the kernel doesn't use the TLB as much as userland does 17:20:10 right 17:20:29 you should run a case where a ton of memory is touched, and thus the tlb miss happens a lot 17:24:01 perhaps later 17:26:35 it actually makes the emulator slower, and it makes the emulation less correct. so I've disabled it 17:36:09 an emulator? oooh 17:49:02 --- join: Savio (~nak@200.152.61.88) joined #osdev 17:49:09 --- nick: Savio -> dm__ 17:49:43 --- nick: dm__ -> dmmmmmm 17:51:35 --- part: dmmmmmm left #osdev 17:57:57 i wonder why im so lagged.. 18:06:35 * kyelewis goes off to do some coding 18:06:38 haven't coded in so long :) 18:09:42 --- join: witten (~witten@adsl-gte-la-216-86-199-140.mminternet.com) joined #osdev 18:18:42 perhaps I've gotten it all wrong 18:19:03 what would you guys say is the difference between a simulator and an emulator? 18:22:34 --- join: steve-shaw (steve@055.b.003.brs.iprimus.net.au) joined #osdev 18:22:35 semantics 18:25:08 because if I google for mips emulator, I find hardware debuggers and stuff like that. if I search for mips simulator, I get a couple of entries regarding what I'm interested in 18:25:13 hm 18:26:32 now time for Coffee 18:27:00 Hi. I'm very much a newbie to os development. I'd like to write an operating system primarily for insights. I'm wondering if there is any reason to favour cooperative multitasking over preemptive multitasking? 18:28:06 * debug suggests preemptive multitasking 18:28:26 suggest you concentrate on that, that is 18:28:47 heh i like preemptive w/ a yield() call 18:29:09 That's what I was thinking. I was just wondering if anyone considers cooperative multitasking a real option these days...? 18:29:41 for some rtos stuff yes 18:30:11 personally i prefer preemptive, just in case a process dies 18:31:09 how does cooperative multitasking work for rtos? what if process doesn't yield before some other important event happens? 18:31:45 steve-shaw: then the time scheduling was flawed, or the scheduling failed. 18:31:46 my kernel has 'supervisory' feature for tasks w/ priority at infinity 18:32:00 in which case the entire rtos would be considered crashed, or something like that 18:32:13 * debug is not into realtime stuff, though 18:32:56 dresden: your kernel is preemptive mt? 18:33:45 somewhat :) not finished w/ tasking 18:34:23 but basicly, if a process has infinite priority (it has to yield() or exit()) if it doesn't call watchdog() every X seconds, it's reaped (at least that's my goal) 18:35:31 ok 18:35:49 that's only because i have an application i need to run that needs such a scheduling model 18:36:12 It absolutely has to be running at all times, and it's the most important task on the box :) 18:36:18 --- join: Prophet__ (~Prophet@pD9FF6E12.dip.t-dialin.net) joined #osdev 18:37:37 I have found the following resource that's not listed for this channel: http://my.execpc.com/~geezer/osd/. Also that sabre? site doesn't seem to be listed. 18:38:35 sabre is here: http://www.nondot.org/~sabre/os/articles 18:39:36 I am currently working through the grub booloader kernel example. 18:39:51 Are there any other useful resources that aren't listed? 18:41:52 debug: i'm not really into rtos either. it sounds like with a rtos an application may say to the os "I'd like to run at exactly x and I'll need only 500ms to execute". Using a guesstimate of processing time to help the scheduler. Is this true? 18:44:13 --- quit: Prophet_ (Read error: 60 (Operation timed out)) 18:49:17 in the OS course I took, there was some stuff about real-time scheduling algorithms 18:49:46 iirc, there's an algorithm which, if there is a way to schedule a set of processes, always will succeed 18:50:03 (assuming no overhead, and that the set of processes is guaranteed to be schedulable to start with) 18:50:21 but I'm not that interested in it to learn more right now 18:51:03 sure. no worries. 18:53:05 i've been viewing the mit operating system course and reading the operating system concepts book. I haven't played with nachos yet and haven't read much about rtos. 18:54:06 does anyone know a good resource for lisp-based operating systems. I'm interested in how that works. 19:09:28 premptive multitasking is the only thing that makes any sense nowadays 19:09:47 * dresden wonders.. 19:09:48 one can implement a mostly cooperative system on top of it anyway 19:10:11 and most user interactive systems are mostly cooperative for the most part 19:10:20 yield() and watchdog()! :) 19:14:55 geist: can you expand on what you mean by "implement a mostly cooperative system on top of it" 19:16:20 create a premptive system, then if you want to build a purely cooperative system, just create one thread 19:16:32 but there's little point nowadays 19:16:45 debug: rate monotonic scheduling has that property 19:16:51 ok 19:16:52 note that even in a fully multithreaded systems, typically tghe threads are basically cooperative 19:17:23 ie, one generates an event, another was waiting for an event, processes it, gives the cpu back to someone else when it waits for the next one 19:17:31 (basically, the processes that need to run most frequently have highest priority) 19:17:44 * debug tries to emulate turbochannel graphics option cards... 19:17:53 the only difference is that in a preemptive system that's not the only way to get a thread switched 19:18:28 and you have the option to create a thread that just goes on processing, relying on the os to switch threads to keep the system working 19:19:24 not having to Idle() is nice. 19:19:40 the fact that most systems work in basically a cooperative way is not super obvious unti you really look at how things work 19:19:46 geist: yes, I guess so. I wonder what the ratio of switches in a preemptive system would be caused by the process yeiding to wait for i/o or on a lock or something (compared to been switched when it's in the middle of something). 19:19:58 depends on the process. 19:20:02 steve-shaw: depends on the situation 19:20:15 but in a modern desktop system, spending most of it's time drawing and respoinding to user events 19:20:24 most of the context switches are basically yields 19:20:47 where a yield is a switch caused by the thread doing something that causes it to block 19:21:09 I remember once that someone broke the preemption timer in beos, and it took a little while to notice it 19:22:17 since most things were cooperative (mouse is moved, interrupt generated, wakes up the mouse reader thread, mvoes the cursor, sends events to the appropriate windows, schedules window threads, etc) 19:22:29 we haven't even enabled the preemptor in our OS, so far we have no real need for it :) 19:22:31 geist: absolutely - exactly what I was thinking. preemptive system just save you from the occasional "greedy" process 19:22:38 where you'd notice it was long term things like compiling or encoding a mp3 19:22:43 steve-shaw: right 19:22:56 it also greatly simplifies things to the programmer 19:23:08 since you no longer have to care about any other threads in the system necessarilly 19:23:20 and it means your OS isn't susceptible to something simply doing an infinite loop 19:23:44 heh 19:23:56 I <3 PRIO_INFINITY :) 19:24:08 my pr0n grabber will run at it 19:24:15 and just yield() and call watchdog() as needed 19:24:29 usually there is a priority level in systems that well run forever 19:24:39 realtime in NT, beos, etc 19:25:22 they are just such high priority threads that they will not be preempted by anything at a lower priority, and are exempt from any sort of dynamic priority adjustment 19:25:50 create a thread at that level and put it in an infinite loop and you have a locked up system (unless you have another cpu) 19:26:10 heh 19:26:27 we did that at be once 19:26:37 there was an app where you can turn off each cpu individually 19:26:55 if you tried to turn off the last one, it created a super high priority thread, and that thread spun for about 15 seconds 19:27:04 which made the system look like it totally locked up 19:27:11 i want to play wif beos :( 19:27:22 then it quit and popped uyp a dialog that said 'dont turn off the last cpu' 19:27:29 hah 19:27:30 we didn't ship with it though, but it was fun 19:28:24 so it sounds like preemptive is the way to go for safety and simplicity for application programmers. But not forgetting that most systems will be pretty cooperative in their behaviour. 19:29:44 Would there be anything to keep in mind when writing a preemptive os which will run mostly cooperatively? 19:30:49 no 19:34:52 what about a longer time slice? I think linux uses 10ms. 19:36:28 get it working, then worry about the quantum 19:36:34 that's a pretty small detail 19:36:38 and easy to change 19:37:20 yep. getting way ahead of myself :-) 19:37:44 * file[desk] waves to everyone 19:37:49 *whines* 19:38:03 cvs server at work is still down... 19:38:14 --- quit: newbs ("printf("%d\n", EOF);") 19:40:10 dresden: sacrifice a goat 19:40:47 file[desk] i thought it was a black chicken 19:40:52 and the problem is the admin is being a bitch 19:41:01 i emailed her 3 days ago.. she read it according to the reciept 19:41:05 and she's just put it off 19:46:09 --- quit: witten ("Client exiting") 19:48:36 --- join: witten (~witten@adsl-gte-la-216-86-199-140.mminternet.com) joined #osdev 19:49:37 --- quit: ToreSB (Remote closed the connection) 19:51:07 wow, you must not work at a software company 19:51:25 because otherwise the source depot should be the most important sserver in the company 19:52:21 geist Nope, i work for a company that does custom car navigation systems :( 19:56:49 --- quit: grrrrr (Read error: 113 (No route to host)) 20:06:21 hm 20:06:29 6.5714285~ weeks til i move ;) 20:06:43 <3 mental math 20:06:51 OBEY 20:07:20 "<3" just doesn't have the same charm in a font with a hooked 3 20:07:27 I want that "Become one with your inner daemon. http://www.freebsd.org" shirt :) 20:20:18 what's a good windoze ppp server w/ nat? :) got a palm pilot n wanna use it to get online at my friend's house.. he has *twitch* aol (only isp available there tho..) 20:20:59 maybe wait til you get home to get your internet fix and save yourself a lot of trouble? :) 20:21:33 heh.. 20:22:01 winroute! that might work 20:22:15 * dresden goes for win9x pppd :) 20:22:26 Step 1) Lock friend in basement 20:22:35 Step 2) Install GNU/Linux or BSD or something 20:23:02 you know it's the right thing to do in the long run. 20:24:53 heh.. problem is, no choice of isp's in his shittown :) 20:25:09 that's what long distance is for 20:25:23 sdt too expensive heh 20:25:32 verizon is only telco choice there 20:25:35 no dsl, no cable.. 20:25:38 just aol hah 20:25:39 you're at your friend's house, right? and he's locked in the basement? soo.... 20:25:46 sdt heh... 20:25:52 i'm trying to sex0r his sister.. :P~ 20:26:31 too much info 20:26:40 ;) 20:26:46 ah. well. impressing her with by connecting your palm with a ppp server is certainly the right way to go. 20:26:56 sdt with her? it is :) 20:27:10 she's a geekgrrl 20:27:38 heh 20:27:45 well, g'night. 20:27:51 damn.. working 9-5 is gonna be crappy 20:28:10 9-5, 6 days a week 20:28:10 heh 20:30:06 sucka 20:30:31 it's for 6.5 weeks :) 20:30:40 jus gotta get $ for the big move 20:31:03 go from 8$/hr to 120k$/yr 20:31:13 and not working any weekends or nights 20:34:05 nice 20:34:21 dresden: is your new employer hiring more people? :) 20:35:00 witten heh, i got the job through my sister's bf 20:35:49 He Who Hesitates, Dies. 20:36:51 dresden: getting a job through someone you know is usually the best way 20:40:42 --- quit: air (Remote closed the connection) 20:41:04 --- join: air (~brand@12-210-160-198.client.attbi.com) joined #osdev 20:44:14 bye air 20:44:23 hi air 20:46:56 --- quit: dresden ("BitchX-1.0c19 -- just do it.") 20:51:04 --- join: sleep- (~ivan@adsl-64-173-24-122.dsl.sntc01.pacbell.net) joined #osdev 21:00:22 anyone know where i can find a good looking black camera icon about the same size as those used in apache file listings? 21:01:02 how about the apache source tree? :P 21:01:14 eh? 21:01:18 lol 21:01:29 i dont have one in my apache icons dir 21:04:56 oh - i thought you meant "same size as the camera icon in apache" :) 21:05:14 but you're saying there isnt one.. 21:06:29 heh 21:08:46 wireless is too slow 21:09:05 sleep-: old wireless point/card? 21:09:18 802.11b 21:09:25 11mbps? 21:09:52 well, im getting 539KB/s 21:09:55 :( 21:10:10 * geist finally fully groks the weird ass interrupt queue tree that ohci exports 21:13:48 * debug has gotten TURBOchannel graphics option cards to work 21:13:57 no 3D-accelerated once, though :-/ 21:13:59 hm 21:14:02 time for more coffee 21:22:07 --- quit: steve-shaw (Read error: 104 (Connection reset by peer)) 21:22:57 mmm..coffee 21:22:59 brb 21:29:38 --- quit: pengo ("I could develop a system that would make money obsolete, but it would only make me rich") 21:30:10 --- join: grrrrr (~jim@1Cust38.tnt36.sfo8.da.uu.net) joined #osdev 21:39:39 --- quit: jwesley ("ChatZilla 0.8.23 [Mozilla rv:1.3.1/20030525]") 21:48:30 --- join: steve-shaw (steve@040.b.002.brs.iprimus.net.au) joined #osdev 22:09:13 --- quit: sleep- (Remote closed the connection) 22:16:50 --- join: sleep- (~ivan@adsl-64-173-24-122.dsl.sntc01.pacbell.net) joined #osdev 22:41:18 --- quit: idle (Read error: 60 (Operation timed out)) 22:58:26 --- quit: redblue (Connection timed out) 22:59:15 hrm...time to re-learn gnuplot again 23:02:29 --- join: redblue (~star@ppp051.216-96-207.sherb.mt.videotron.ca) joined #osdev 23:27:01 --- quit: sleep- ("Client Exiting") 23:47:59 --- quit: Prophet__ ("#define PROFTPD_MAGIC 0xDEADBEEF") 23:48:24 --- join: Prophet_ (~Prophet@p50813DD0.dip.t-dialin.net) joined #osdev 23:59:59 --- log: ended osdev/03.11.02