IRC log started Tue Mar 30 00:00:01 1999 -:- SignOff _QZ: #TUNES (BRiX [http://www.qzx.com/brix] :: sleep) [msg(TUNES)] permlog 1999.0330 -:- abi [nef@bespin.cx] has joined #tunes [openprojects]!clarke.openprojects.net Net break: tolkien.openprojects.net asprin.openprojects.net (Ping timeout) [openprojects]!clarke.openprojects.net Net junction: tolkien.openprojects.net asprin.openprojects.net [openprojects]!clarke.openprojects.net Completed net.burst from asprin.openprojects.net. [openprojects]!clarke.openprojects.net asprin.openprojects.net acknowledged end of net.burst. [openprojects]!clarke.openprojects.net Net break: tolkien.openprojects.net asprin.openprojects.net (Ping timeout) [openprojects]!clarke.openprojects.net Net junction: varley.openprojects.net asprin.openprojects.net [openprojects]!clarke.openprojects.net Completed net.burst from asprin.openprojects.net. [openprojects]!clarke.openprojects.net asprin.openprojects.net acknowledged end of net.burst. [openprojects]!clarke.openprojects.net Net break: koontz.openprojects.net hogan.openprojects.net (Ping timeout) [openprojects]!clarke.openprojects.net Net junction: sterling.openprojects.net hogan.openprojects.net [openprojects]!clarke.openprojects.net Completed net.burst from hogan.openprojects.net. [openprojects]!clarke.openprojects.net hogan.openprojects.net acknowledged end of net.burst. [openprojects]!clarke.openprojects.net Net break: koontz.openprojects.net carter.openprojects.net (Read error to carter.openprojects.net[209.81.8.249]: Connection reset by peer) [openprojects]!clarke.openprojects.net Net junction: koontz.openprojects.net carter.openprojects.net [openprojects]!clarke.openprojects.net Completed net.burst from carter.openprojects.net. [openprojects]!clarke.openprojects.net carter.openprojects.net acknowledged end of net.burst. -:- SignOff Tril: #TUNES (Ping timeout for Tril[sloth.wcug.wwu.edu]) [openprojects]Foreign OperKill: NickServ killed Is (This nick is reserved by another user) -:- hcf [nef@me-portland-us315.javanet.com] has joined #tunes -:- SignOff hcf: #TUNES (Leaving) -:- Tril [dem@sloth.wcug.wwu.edu] has joined #TUNES >>> Tril [dem@sloth.wcug.wwu.edu] requested PING 922814100 450596 from #TUNES -:- SignOff Tril: #TUNES (Tril has no reason) -:- SignOff Tikiman: #TUNES (Ping timeout for Tikiman[mcrigal2.student.Princeton.EDU]) [openprojects]!clarke.openprojects.net Net break: varley.openprojects.net dickson.openprojects.net (Ping timeout) [openprojects]!clarke.openprojects.net Net junction: varley.openprojects.net dickson.openprojects.net [openprojects]!clarke.openprojects.net Completed net.burst from dickson.openprojects.net. [openprojects]!clarke.openprojects.net dickson.openprojects.net acknowledged end of net.burst. [openprojects]Foreign OperKill: NickServ killed wicked (GHOST command used by _Wicked) [openprojects]!clarke.openprojects.net Net break: tolkien.openprojects.net gibson.openprojects.net (Read error to gibson.openprojects.net[212.46.64.45]: Connection reset by peer) [openprojects]Foreign OperKill: NickServ killed geekman (This nick is reserved by another user) [openprojects]Foreign OperKill: NickServ killed geekman (This nick is reserved by another user) [openprojects]Foreign OperKill: NickServ killed geekman (This nick is reserved by another user) [openprojects]Foreign OperKill: NickServ killed geekman (This nick is reserved by another user) [openprojects]Foreign OperKill: NickServ killed geekman (This nick is reserved by another user) -:- Tril [dem@sloth.wcug.wwu.edu] has joined #TUNES [openprojects]!clarke.openprojects.net Net junction: tolkien.openprojects.net gibson.openprojects.net [openprojects]!clarke.openprojects.net Completed net.burst from gibson.openprojects.net. [openprojects]!clarke.openprojects.net gibson.openprojects.net acknowledged end of net.burst. hello Fare hi, Tril! tomorrow, I'll setup a backup of bespin what are the zones to save? huh? well, I'd like to do a backup of things of interest. /serv/* ? I started making a list of general bespin stuff to back up in /i (text file). That doesnt include any tunes stuff. well, at least it's a list of configuration files I changed don't back up the debian archive :) indeed last week, I did a presentation on linear logic to my colleagues (and I remotely debuggd my dad's computer) are those two separate things? 12:40pm yes also got lots of things to read, and an article to rewrite i have class in 10 mins 12:50pm * Tril/#TUNES is away: (afk) [BX-MsgLog Off] [openprojects]Foreign OperKill: NickServ killed lux (This nick is reserved by another user) [openprojects]!clarke.openprojects.net Net break: koontz.openprojects.net carter.openprojects.net (Ping timeout) [openprojects]!clarke.openprojects.net Net junction: koontz.openprojects.net carter.openprojects.net [openprojects]!clarke.openprojects.net Completed net.burst from carter.openprojects.net. [openprojects]!clarke.openprojects.net carter.openprojects.net acknowledged end of net.burst. -:- tcn [tcn@cci-209150250112.clarityconnect.net] has joined #tunes abi, status? Since Fri Mar 26 18:18:59 1999, there have been 12 modifications and 52 questions. I have been awake for 3 days, 19 hours, 9 minutes, 21 seconds this session, and currently reference 122 factoids. 01:30pm -:- SignOff tcn: #TUNES (tcn has no reason) [openprojects]!clarke.openprojects.net Net break: koontz.openprojects.net carter.openprojects.net (Read error to carter.openprojects.net[209.81.8.249]: Connection reset by peer) [openprojects]!clarke.openprojects.net Net break: varley.openprojects.net dickson.openprojects.net (Ping timeout) [openprojects]!clarke.openprojects.net Net junction: varley.openprojects.net dickson.openprojects.net [openprojects]!clarke.openprojects.net Completed net.burst from dickson.openprojects.net. [openprojects]!clarke.openprojects.net dickson.openprojects.net acknowledged end of net.burst. -:- _QZ [brand@p0wer.qzx.com] has joined #tunes [openprojects]Foreign OperKill: OperServ killed tuff (You have been k-lined. To appeal, please email support@openprojects.net.) [openprojects]!clarke.openprojects.net services1.openprojects.net adding GLINE for tuff@submit-pro.net, expiring at 922932583: You have been k-lined. To appeal, please email support@openprojects.net. <_QZ> http://3508975004/ <-- what the hell kinda address is that? 02:40pm * Tril/#TUNES is back from the dead. Gone 2 hrs 5 min 36 secs qz: alphabetic IP convert the ip to hex, which is 4 2-digit hex numbers, make it one 8-digit hex number and convert it to decimal. <_QZ> kewl oops, not alphabetic, i guess they are decimal ips 03:00pm it works everywhere <_QZ> well it looks like i will be releasing brix thursday how many peopple have flamed you for not releasing the source (besides me) if it's no good, of course nobody will be very angry. <_QZ> heh, i have received tons of emails cuz i havent put any source on teh page <_QZ> but u are about the only one that knows im not gonna release the source when its ready yeah, and I think you just decided that recently <_QZ> yup earlier i got the impression you would be releasing source. <_QZ> ya but i wanna make some money :) so you wanna compete directly with MS and Be rather than target the internet community? when brix is just starting? Red hat makes plenty of money <_QZ> no <_QZ> brix binaries will be free <_QZ> its the source code that wont, yet <_QZ> maybe i'll sell the source for around $1000 per copy + nda <_QZ> and see how many morons buy it i think it is a bad idea, because if you release the source, people will be considerate, and either make constructive suggestions or send you patches. if you hide the source they will just flame you and be angry, and not use your os unless, you plan to incorporate everyone's ideas into it yourself. <_QZ> look at BeOS, it has a large community of dopes and they have no source for it and they must also pay <_QZ> and its only the kernel that i will hold the source to what is "large"? how many use beos? <_QZ> enuf 03:10pm none of them use only beos, they all dual boot 95 or mockass why are you delaying it :) <_QZ> delaying? thurs is the 1st still oh-k <_QZ> huh <_QZ> im releasing it on april 1st as planned <_QZ> that was announced about 2 weeks ago is there really a tcp stack? i think you were bsing about that one <_QZ> uhh <_QZ> when did i say that? you told someone you wrote a tcp/ip stack in brix' script language <_QZ> ahh, they asked if it had tcp right? i don't remember <_QZ> if they ask i will say yes, always <_QZ> this release wont have a gui,tcp,networking,... <_QZ> it will have a preemtive multithreading OO kernel and text based shell does paging work <_QZ> yes tested it? fill up all ram? <_QZ> i have tested it, no i havent filled up ram the pager is also the persistence manager right? 03:20pm <_QZ> the object age list isnt in place yet so it will run out of ram right now <_QZ> pager? pager is the persistence manager right? <_QZ> forget pager _QZ: I forgot pager <_QZ> the system runs a thread that handles persistence 03:30pm [openprojects]!clarke.openprojects.net Net break: sterling.openprojects.net vinge.openprojects.net (Server vinge.openprojects.net[38.250.52.51] closed the connection (<>)) <_QZ> Tril: u gotta work on tunes, the thrill of having yer own OS working is so... 03:50pm hmm, probably. I want the thrill of having my own debian cds first :) i'm configuring the script to create iso images <_QZ> thats not a thrill <_QZ> ah crap, i forgot to have the scheduler set es <_QZ> oh well its still faster than qnx u have qnx? <_QZ> just the demo <_QZ> but im talking about the switching times <_QZ> hm that stupid cpu is clearing es <_QZ> is there some kinda segment register checking on iret? i think.. i'm not at home so i dont have my intel reference with me 04:00pm <_QZ> kewl it does <_QZ> if the segreg has a DPL lower than the segment being iret'd to then it clears it <_QZ> how the hell can 'mov eax,1' crash <_QZ> bochs: panic, interrupt(): EIP > CS.limit you exceeded your code segment? <_QZ> its a 4gig segment <_QZ> ah crap, now i find out that bochs can dump physical memory <_QZ> all this time i have had to disable paging in the code to dump memory that wasnt mapped into page tables 04:10pm <_QZ> hmm [openprojects]!clarke.openprojects.net Net junction: sterling.openprojects.net carter.openprojects.net [openprojects]!clarke.openprojects.net Completed net.burst from carter.openprojects.net. [openprojects]!clarke.openprojects.net carter.openprojects.net acknowledged end of net.burst. 04:30pm -:- SignOff abi: #TUNES (Ping timeout for abi[bespin.cx]) hehe.. * Tril/#TUNES makes the load on bespin skyrocket <_QZ> heh abi is a wimp, the load is only 2.4 <_QZ> haha >>> _QZ [brand@p0wer.qzx.com] requested PING 922841166 690822 from TUNES <_QZ> this blows <_QZ> my descriptors are perfect [openprojects]Foreign OperKill: OperServ killed Wyster (You have been k-lined. To appeal, please email support@openprojects.net.) [openprojects]!clarke.openprojects.net services1.openprojects.net adding GLINE for *@vcd.student.utwente.nl, expiring at 922941062: You have been k-lined. To appeal, please email support@openprojects.net. <_QZ> do u happen to know, off the top of yer head, what bits 1 and 9 are in elfags? :) 04:50pm the fags register? <_QZ> ugh <_QZ> eflags you said it not me no I don't. are there branch instructions for all the flags? <_QZ> no <_QZ> im just wondering if maybe i set the wrong bits in the flags and maybe its causing a trap or something -:- hcf [nef@me-portland-us104.javanet.com] has joined #tunes hi nef hoy dem <_QZ> hcf: do u happen to know, off the top of yer head, what bits 1 and 9 are in eflags? :) oh yeah sure really i do hmm, no abi <_QZ> tril killed abi hcf, abi died due to netlag, give her another server.. try clarke.openprojects.net, it's close to bespin :) <_QZ> damn, i finally found the flags reference in the dev manual -:- abi [nef@bespin.cx] has joined #tunes 05:00pm -:- SignOff abi: #TUNES (Ping timeout for abi[bespin.cx]) <_QZ> ugh, shit like this makes me not wanna work on it anymore <_QZ> maybe it wont get released thursday it probably doesnt' wrok on my puter anyway <_QZ> oh that reminds me, i gotta try it out and see if i fixed the object ladoing crap <_QZ> damnit <_QZ> it still wont load the objects on my other machines <_QZ> will on the laptop but not on any desktops can't load objects. serious problem. 05:30pm -:- abi [abi@bespin.cx] has joined #tunes * Tril/#TUNES is away: (AFK) [BX-MsgLog Off] -:- SignOff hcf: #TUNES (Leaving) [openprojects]!clarke.openprojects.net Net break: sterling.openprojects.net carter.openprojects.net (Ping timeout) <_QZ> ok, the object loader is fixed 06:50pm -:- tcn [tcn@cci-209150250108.clarityconnect.net] has joined #tunes howdy QZ, Fare.. <_QZ> hey <_QZ> i got good and bad news bad news first <_QZ> how much do u know about pmode? enuf <_QZ> well good news is that my kernel is functional <_QZ> and i fixed the problem with the object loader only working on my laptop <_QZ> bad news is that when the actions try to run i get a nice EIP>CS.limit error haha in bochs? <_QZ> ya what exception does it trigger, on a real computer? <_QZ> it resets hmm.. you need exception handlers, then. <_QZ> i do you have them, or you need them? <_QZ> but i havent updated the kernel debugger for the new kernel reset = triple fault.. <_QZ> ya i know so something's corrupting you IDT, triggering an interrupt, double & triple faulting.. <_QZ> most likely due to the fact that my debugger writes to the screen <_QZ> and the screen isnt page mapped Try it w/o the debugger 07:40pm <_QZ> w/o??? oh, is that how you run the actions? <_QZ> no <_QZ> the debugger is part of the kernel and is hooked to all exception handlers that's how I run them :) <_QZ> my first object.action is started by the kernel <_QZ> damnit <_QZ> its still reseting what's the CS limit? <_QZ> 4gig <_QZ> base 0 wtf... How could a 32bit # be >4gig? 07:50pm <_QZ> i think i might know <_QZ> it says bochs: panic, interrupt(): EIP > CS.limit hmm.. I only used bochs a little bit.. <_QZ> the exception would load the kCS descriptor that has a 16k limit I see <_QZ> i think maybe my EIP value on the int gate is bad kCS = kernel CS? <_QZ> ya why? <_QZ> my kernel area is 16k right, but why not just use page protection? <_QZ> the descriptor set the CPL <_QZ> u need 2 CS segments <_QZ> i could have made my kCS with a base of 0 and a limit of 4gig that's what I did <_QZ> but then this little bug wouldnt have been caught as quickly I haven't even touched my user CS/DS yet :) <_QZ> the ONLY code that needs to run in kCS lies within that 16k region <_QZ> and if code, due to a bug, tries to run outside it faults like its doing now You know what you have to do stop putting it off :) <_QZ> ah horse shit <_QZ> i removed the exception number shit from the debugger !lilo:*! ire 1 !doogie:*! ?? !netgod:*! eh? 08:00pm <_QZ> whats the structure of an idt descriptor :) write a better GPF handler dw offset[0..15], cs_selector, 0x8E00, offset[16..31] <_QZ> bewm <_QZ> found the problem <_QZ> now just gotta fix it <_QZ> f000 8e 53 0010 10b3 <_QZ> that high f I wrote a IDT auto-generator today.. so every interrupt says "Interupt #13" or whatever, even if I don't assign a "real" handler.. <_QZ> how <_QZ> what is the count in the idt descriptor? each int has a handler that goes "push ### ; jmp default_handler".. then default_handler prints the # on the stack <_QZ> oh, i had that I was tired of getting "unknown interrupt" <_QZ> ok, now its going into the debugger cool 08:10pm f000 8e 53 0010 10b3..? big endian? <_QZ> yes <_QZ> 0xf0008e53001010b3 <_QZ> now its 0x00008e00001010b3 what kinda data structure are you using in your scheduler? <_QZ> for what? to decide which tasks gets run next s/tasks/task <_QZ> goto borg/bi <_QZ> u there? what is it, the event list? <_QZ> thread list holds the thread info <_QZ> active thread list holds the chain of threads waiting to run 08:20pm <_QZ> o-space list holds the info for each memory space that's a nice trick with the IO bitmap <_QZ> what trick putting the TSS at the end of the page <_QZ> oh <_QZ> if u do that u might wanna look at the first dword in sys_data <_QZ> it says all bits must be set but u really only need to set the first byte what's that for? <_QZ> word alignment <_QZ> and your tss limit must account for that byte <_QZ> the cpu reads the iobitmap in word sizes <_QZ> to read the last byte in the iobitmap it must read an extra byte above it to make a whole word read access [openprojects]!clarke.openprojects.net Net junction: sterling.openprojects.net carter.openprojects.net [openprojects]!clarke.openprojects.net Completed net.burst from carter.openprojects.net. [openprojects]!clarke.openprojects.net carter.openprojects.net acknowledged end of net.burst. you sure? The bitmap should fit in exactly 2 pages, no overrun.. <_QZ> right <_QZ> but the cpu does its reads in wholewords <_QZ> starting with the byte containing the io port oh <_QZ> so if u try to access one of the last 8 ports it must read the last byte plus ANOTHER byte after that byte what's it matter what the extra byte is, though? 08:30pm <_QZ> it wants all the bits to be set <_QZ> intel says so [openprojects]!clarke.openprojects.net Net break: koontz.openprojects.net calvino.openprojects.net (Ping timeout) i'm back maybe I did see that somewhere.. [openprojects]!clarke.openprojects.net Net junction: sterling.openprojects.net calvino.openprojects.net [openprojects]!clarke.openprojects.net Completed net.burst from calvino.openprojects.net. [openprojects]!clarke.openprojects.net calvino.openprojects.net acknowledged end of net.burst. 08:40pm <_QZ> YES <_QZ> my debugger is back online both our OS's need to allocate more memory for the lists.. say if you need more than 128k in the thread list.. <_QZ> that can handle 2730 threads what if you need 2731? <_QZ> u wont ever?? besides, you shouldn't allocate 128k if you only use 4k <_QZ> never <_QZ> damnit, i might have to postpone the release <_QZ> this stupid bug has put me behind schedule this is gonna be a problem for me soon.. right now I have a fixed 64k for code and 64k for data.. 08:50pm ahh, it won't be a problem when each task has its own page table.. nevermind :) <_QZ> huh hey, are there x86's with 2-4gigs RAM yet? <_QZ> yes <_QZ> xeon mobos support 4gig damn any w/ external paging yet? <_QZ> external paging? like on 8-bit machines :) <_QZ> hahaha 32 isn't gonna be enough <_QZ> what u need a box with a couple terabytes of ram? :) <_QZ> the ppro and up support 36bits that helps <_QZ> 64gig alpha's have had 48+ bits for awhile now 48 bit addressable <_QZ> do u use nasm? yeah. it sucks, eh? <_QZ> how do u make a times loop that inc's a number <_QZ> i need to repeat a macro 32 times and havea number from 0-31 rtfm <_QZ> ass :) I think you use a %% variable I used ASM to do what you're doing generate it at runtime Shit, I'm doing things way too complicated.. 09:00pm (something else) 64gig is too much for 1 cpu <_QZ> the xeon mobos support upto 8 <_QZ> cpus eventually they'll have to use distributed memory that's been going on for a long time in supercomputers <_QZ> does nasm have a sizeof for a macro? ? <_QZ> i need to know the size of a macro don't think so <_QZ> ah i c Uhh.. put '%%begin' at the beginning, '%%end' at the end, and '%%end-%%begin' is the size.. <_QZ> yup just saw that :) nasm pretty much sucks 09:10pm it's either that or GAS though or roll yer own <_QZ> no way <_QZ> nasm is kewl <_QZ> nasm is better than anything else I wouldn't use it if there was something better you think code should be read-only? <_QZ> yes so it can't modify itself? <_QZ> yup <_QZ> i map all the actions as read only Unix doesn't do that, does it? <_QZ> ofcourse brix lets actions modify other actions 09:20pm later.. gotta crash -:- SignOff tcn: #TUNES (tcn has no reason) 09:30pm [openprojects]Foreign OperKill: NickServ killed Knghtbrd (GHOST command used by Knghtbrd_) [openprojects]!clarke.openprojects.net Net break: varley.openprojects.net dickson.openprojects.net (Ping timeout) [openprojects]!clarke.openprojects.net Net junction: varley.openprojects.net dickson.openprojects.net [openprojects]!clarke.openprojects.net Completed net.burst from dickson.openprojects.net. [openprojects]!clarke.openprojects.net dickson.openprojects.net acknowledged end of net.burst. [openprojects]Foreign OperKill: NickServ killed Atko (This nick is reserved by another user) [openprojects]!clarke.openprojects.net Local SQUIT by lilo [varley.openprojects.net]: [openprojects]!clarke.openprojects.net Net break: varley.openprojects.net dickson.openprojects.net (lilo) !koontz.openprojects.net!! Remote CONNECT dickson.openprojects.net 8004 from lilo [openprojects]!clarke.openprojects.net Net junction: koontz.openprojects.net dickson.openprojects.net [openprojects]!clarke.openprojects.net Completed net.burst from dickson.openprojects.net. [openprojects]!clarke.openprojects.net dickson.openprojects.net acknowledged end of net.burst. * Tril/#TUNES is back from the dead. Gone 5 hrs 17 min 1 secs * Tril/#TUNES is away: (sleep) [BX-MsgLog Off] [msg(TUNES)] newlog 1999.0331 IRC log ended Wed Mar 31 00:00:01 1999