00:00:00 --- log: started osdev/03.10.07 00:17:00 --- join: trans (vmvnue@fatwire-201-147.uniserve.ca) joined #osdev 00:17:41 --- join: witten (~witten@adsl-gte-la-216-86-199-140.mminternet.com) joined #osdev 00:28:22 --- quit: pengo ("I could develop a system that would make money obsolete, but it would only make me rich") 00:38:40 hmm 00:39:15 what? 00:39:54 00:41:36 I think if I put too many debugging print()s in my timer code, my timer handler is too slow 00:41:48 and that causes issues because it's called so frequently 00:42:15 yep, I've had problems like that before 00:42:31 I had the same thing when i was hacking on an ohci driver he other day 00:42:50 you can turn a particular interrupt on that signals start-of-frame, which is once per ms 00:42:58 debugging it caused timing issues? 00:43:08 and the amount of printfs I had made it so slow the cpu spent all it's time in the interrupt handler 00:43:15 hehe 00:43:19 exactly 00:45:53 ok, this is really weird 00:46:26 if I put a single debugging print in my timer callback loop, the test checkpoint to disk works fine, and so does the first test load 00:46:28 it always is until you understand it 00:46:47 if I take out that print, it appears to hang somewhere in the timer handler upon the checkpoint 00:47:41 could be some stack corruption that calling the print routine fixes 00:47:47 that happens some time 00:48:59 could be 00:49:07 but I bet it's more a timing thing 00:49:38 or something funky where a spinlock is acquired and then not released, and then somehow something tries to reacquire it and spins forever 00:50:04 or it's acquired without interrupts being disabled 00:50:08 that's worth an ASSERT 00:50:34 why do interrupts have to be disabled to acquire a spinlock? 00:51:33 think about it for a minute 00:51:37 it's a good mental exercise 00:51:54 hmm 00:52:21 well I guess it's not really a global lock if it can be interrupted at any time 00:52:33 umm hmm 00:52:43 remember, a spinlock is not a thread lock 00:52:59 yeah 00:53:02 it's not there to keep other threads out of a critical section, it's there to keep other *cpus* out of a critical section 00:53:19 so if you didn't have interrupts disabled, what could happen while you have the spinlock acquired? 00:53:41 the cpu could very well pop an interrupt and go back to that critical section 00:53:54 true, that's one case 00:54:11 what other things can happen as a result of taking an interrupt? 00:54:23 triple fault? :) 00:55:05 nope, but it could result in a reschedule 00:55:15 which means the cpu will go off doing something else 00:55:15 well, yeah 00:55:26 and might not get back to the original thread for potentially ever 00:55:33 which means it wouldn't release the spinlock ever 00:55:43 ok, so I've got to add a bunch of enter_critical() / leave_critical() pairs around spinlocks 00:55:45 which means all of the other cpus are now dead in the water waiting for that spinlock 00:55:56 that would suck 00:56:00 you had better make sure interrupts are disabled around acquiring spinlocks 00:56:13 and I'd put code in the spinlock that asserts that interrupts are 00:56:40 you can do a few different tactics. you have just require that interrupts are disabled however it happesn when you enter the spinlock code 00:56:57 or you can disable interrupts if they aren't already automatically as part of the spinlock acquire and release functions 00:57:08 yeah 00:57:46 I'll try the latter, because my enter_critical() code keeps a counter 00:59:37 also, you can save some speed by just disabling the spinlock code if you only have a single cpu 00:59:54 I suppose so, yeah 00:59:57 I have a if(num_cpus >= 2) { ... } inside my spinlock code 01:00:07 spinlocks are absolutely useless on a single cpu machine 01:00:18 one of the reasons NT and linux have seperate builds for smp vs not 01:00:20 they just compile it out 01:00:36 spinlocks are useless, but disabling interrupts around a critical section are not 01:01:29 a spinlock is basically an extention of the enter_critical() and leave_critical() calls to encompass multiple cpus 01:01:45 where enter_critical and leavel critical basically just disable interrupts 01:01:56 yeah 01:02:41 there may be critical sections where you dont need to acquire a spinlock, but the inverse is never true 01:02:55 and now I've beat this dead horse, I think you understand. :) 01:03:03 yeah, I appreciate it :) 01:03:09 I hadn't really thought of it this way before 01:03:28 woot, and it works! 01:03:49 my floppy driver now works with the new multitasking code and semaphores and stuff 01:03:52 thanks! 01:04:13 which means I can now make my checkpointer a separate process.. 01:04:18 and my disk cleaner too 01:04:52 awesome! congratulations! 01:05:13 yeah 01:05:17 almost ready for a release 01:05:57 --- join: gab (~prfalken@nas-p19-3-62-147-223-53.dial.proxad.net) joined #osdev 01:10:56 can anyone give me a good example of needed pointers to pointers? 01:11:09 needing 01:11:12 air: 2d array 01:11:34 witten: eh? 01:12:23 well 01:12:41 the first point points to an array of rows 01:12:50 and each row is similarly an array of elements 01:12:53 invalid example, next 01:12:54 so that's a pointer to a pointer 01:12:59 why? 01:13:07 thats a C problem 01:13:19 something like crush would use an array of arrays 01:13:25 ok.. 01:13:42 oh shoot, I needed it the other day, but i forgot why 01:13:47 im thinking about removing pointers from crush 01:14:28 I'm suprised you had them at all 01:15:13 geist: without spinlocks, what primitive atomic locking mechanism do you use to implement semaphore locking? 01:15:20 on a single cpu system? 01:15:34 spinlocks aren't used to implement semaphores at all 01:15:45 they are only used to deal with critical sections across cpus 01:15:55 ok.. 01:16:03 but to do the atomic add off the semaphore count, you need atomic_add() 01:16:16 on x86 this is lock xadd I think 01:16:26 or some variant of cmpxchg 01:16:33 I see 01:18:36 or you could just enter a critical section, do the add, then exit, but that'd be more overhead 01:19:40 how does exchanging the operands guarantee atomicity? 01:19:51 anyone here have ops in #c? 01:19:59 air: not I 01:20:10 they are in need of one right now 01:20:34 on x86 the lock prefix generates a bus lock on the cpu bus, guaranteeing that no other cpu will interleave the multiple memory accesses 01:20:54 another reason x86's dont scale that well 01:21:08 the other cpus are dead in the water until that lock is released 01:21:57 hmm 01:22:36 anyone know what channel ircops hang out in? 01:23:23 #freenode? 01:23:30 yeah, I see 01:23:39 damn goatsex ascii art 01:24:47 arg 01:24:52 i cant talk in #freenod 01:26:01 a staffer klined him 01:26:27 i know 01:26:49 OBEY 01:26:57 anyway, time to sleep 01:27:10 nite geist 01:28:00 nite 01:31:11 --- quit: CLxyz (Read error: 104 (Connection reset by peer)) 01:32:45 --- join: CLxyz (~clxyz@209.53.54.1) joined #osdev 01:33:17 --- quit: CLxyz (Client Quit) 01:34:26 --- quit: CrayT3E (Remote closed the connection) 01:34:44 --- join: demise (~e9o@c-864272d5.01-94-7673741.cust.bredbandsbolaget.se) joined #osdev 01:40:35 --- join: quantis (quantis@quantis.demon.co.uk) joined #osdev 01:40:40 evening 01:41:40 hi 01:42:31 ello there how is you 01:43:06 fine here, quantis. how about ya 01:45:15 ahh not to bad 01:45:27 working on my uni projects 01:45:33 project even 01:45:39 student days are fun 01:46:12 ohh yeah 01:46:24 ... im rsiting my third year 01:48:48 do either of u have good examples for needing pointers to pointers? 01:49:03 no success in #c? 01:49:23 they gave some examples, but none good 01:50:21 im thinking the only reason u need them is for ugly imperative hacks like having a function modify a pointer from the caller 01:50:34 which can be solved by returning the pointer 01:50:38 hm, the guys there are good. zhivago and some others, they know what they talk. if none are good, tough luck 01:50:45 p = foo(p); 01:50:54 instead of foo(&p); 01:51:02 what about indexes? 01:51:12 or linked lists, 01:51:15 arrays are completely out 01:51:16 2 dim arrays? quantis 01:51:18 and lists 01:51:22 no htats crap, cause you pointing ot the datastructure 01:51:38 sorry me not thinking straight 01:51:46 two early in morning and im freezing 01:51:50 someone told me that pointers to pointers are a sign of a weak language 01:51:52 --- join: CrayT3E (~CrayT3E@finsler.math.uoa.gr) joined #osdev 01:51:56 and so far its true 01:52:21 lets say foo() needs to return an error/success value so u cant return a pointer 01:52:23 did ya try #c of efnet? there are many good folks there too 01:52:36 exceptions are one way of solving that, multiple return values is another 01:52:53 then there was the argv example 01:52:55 exceptions are a re pretty good way i think 01:53:12 but when array is a type u have arrays of arrays 01:53:30 cause have multiple return values can get quite complicated, and make learning the language a lot harder 01:54:03 whats wrong with arrays of arrays,? 01:54:17 you could always have a list of arrays much more memmroy efficent, but slower 01:54:23 quantis: yes but passing pointers to pointers is pure imperative and there are side effects 01:55:04 like? 01:55:15 argv is a pointer to a char * 01:56:03 its not so much a pointer as an alias 01:56:17 C has no array type 01:56:24 so u get stuff like pointers to pointers 01:56:37 true, but C is a systems language 01:56:40 but if arrays are real types then u wont 01:56:46 you need it for that 01:56:54 like the 2d arrays witten pointed out 01:56:57 why are you thinking of implmenting it in your system? 01:57:03 ya 01:57:09 i wanna remove pointers 01:57:28 you shouldnt need it unless you planning of using BRiX for driver dev in the system as well 01:57:36 simplifies the language and allows the compiler to make better optimizations 01:57:44 hmmm 01:57:49 cuz im also gonna get rid of stack variables 01:58:00 thats sounds like a plan 01:58:14 it will appear as if all objects are allocated on the heap but the compiler will optimize some to the stack 01:58:26 but pointerxs to pointer can bring into play a lot of problems cuase it can be very easy to lose track of what type you actually pointing to 01:58:48 right 01:59:08 of course C would catch any errors 01:59:10 one of the points where java wins out 01:59:16 er s/C/crush/ 01:59:17 everytrhing is an object 01:59:30 you could prehaps have an advanced mode 01:59:52 but if the programmer loses track of types then he is likely to create mistakes and introduce bugs 02:00:06 removing pointers solves that problem 02:00:49 i will just have references 02:01:11 yeah but for systems programing you need the most efficent ways of doing datastructures, otherwise it ill have a knock on effect for the rest of the system 02:01:30 i know 02:01:39 i think if you have it so the code has to marked as advanced or special, then it will make the programming think about it more 02:02:15 or you could have an expected type item when declaring the pointer 02:02:15 references shouldnt impact performance 02:02:23 not in general 02:02:45 the programmer can think of all objects as being pointers to the data 02:02:47 but low level datstructures benefit more cuase you dont want the overhead of full blow objects getting thrown around all the time 02:03:11 passing an object would clone it, but passing a reference would pass the pointer to the object 02:03:27 if you declare an expected type the pointers pointer should be pointing at the compiler can flag up a warning if its not 02:03:40 but the language will optimize out stuff like ints and floats 02:03:49 so no pointers will be involved 02:03:57 yeah but pointers are a lot finer detail, cause you still have to look up the part of the object you need 02:04:02 and it can optimize the pointers out of other types as well 02:04:16 just sounds like a lot more hard work thats all 02:04:16 eh? 02:04:40 with pointers you can point to the consitute parts of an object rather then the whole thing 02:04:52 well, it will remove a lot of code from the compiler but it will also require it to do more work 02:05:42 the fields in an object are also objects 02:05:53 u can reference them 02:06:10 foo.bar 02:06:45 air: what if you actually need to manipulate raw memory? 02:06:58 like, for implementing, say, a memory manager or malloc implementation? 02:07:02 example? 02:07:07 you could create a global memory object prehaps? 02:07:09 malloc() 02:07:34 objects are mapped over the memory 02:07:48 by what 02:07:58 i thought brix os was written in brix 02:08:00 crush has a special array type that doesnt have a size field 02:08:09 quantis: brix is written in crush 02:08:24 sorry yeah i forgot, i just knew they were closely related 02:08:39 * quantis gives himself a firm slap on the wrist 02:09:12 witten: and crush also has overlay types 02:09:50 witten: special types that must be overlaid on memory instead of created like other objects 02:10:03 quantis: heh 02:10:29 --- quit: lodda (Remote closed the connection) 02:11:09 intresting concept, how are you going to implment swap space and things like that 02:11:43 wait ive just realised what you ment 02:12:52 air: cool 02:12:53 lets say the memory manager has an array of base:size nodes 02:13:07 the base field would be a uint 02:13:10 not a pointer 02:13:23 cuz its really not 02:14:13 and the memory manager would grab the grab the base and then modify it for whatever memory was taken 02:14:39 it would be unsafe to do arithmetic on a pointer (something crush doesnt allow) 02:14:56 this is true 02:15:40 but pointer arimetic is another thing thats very handy in efficent systems programming 02:15:48 --- quit: petrusss (Read error: 113 (No route to host)) 02:15:50 and then malloc would map the base, if not already mapped and then do an unsafe cast 02:16:02 quantis: not really 02:16:59 quantis: walking arrays/lists? crush uses foreach methods on all collections 02:17:32 quantis: and crush has special iterators that can be used like a C pointer but safely 02:18:04 its an object that holds a pointer to the collection and a current pointer 02:19:42 sounds a bit like c# 02:19:53 all you need now are delgates 02:20:30 heh 02:20:37 its not C# 02:20:47 anyway, i need to sleep 02:21:09 thanks for going over this with me 02:21:20 i like c# anways so i wouldnt complain 02:21:32 sleep well 02:24:24 --- join: eniac (~jonasg@219.4-201-80.adsl.skynet.be) joined #osdev 02:25:06 yo eniac 02:25:10 hoi hoi 02:30:04 --- join: kyelewis (~kye@dsl-180.67.240.220.lns02-dryb-mel.dsl.comindico.com.au) joined #osdev 02:30:07 heyas 02:30:19 hi :) 02:30:38 kyelewis: did u drop frames yet? 02:30:57 heh.. i haven't worked on the site for a few days 02:31:10 been busy :/ 02:31:32 u gonna drop em? :) 02:31:51 probably, the demand is high enough 02:32:11 and i don't have frames on any of my other proper sites 02:32:27 yer making an opensource OS which means u will get a lot of lynx visitors :) 02:33:54 --- quit: demise () 02:34:43 u should also look into moving the sidebar div to the bottom of the html if it gets much larger 02:35:17 that way it wont be displayed before the content in text-only browsers 02:36:10 check out brix-os.sf.net in a graphical browser and then in a text-only browser 02:36:24 that top navbar moves to the bottom in text-mode 02:38:27 anyway, im really going to bed now 02:38:43 --- quit: air ("cria 0.2.9 -- http://cria.sf.net") 02:48:48 --- quit: eniac ("Lost terminal") 02:51:20 --- join: gab_ (~prfalken@80.65.224.254) joined #osdev 03:03:33 --- join: erlingre (~erlingre@algordo.ifi.uio.no) joined #osdev 03:05:08 --- quit: gab (Nick collision from services.) 03:06:09 --- nick: erlingre -> Pyld 03:06:48 with a preemptive kernel, how is a timer innterupt handled during a system call? 03:21:41 --- quit: trans (Read error: 110 (Connection timed out)) 03:37:47 --- quit: kyelewis (Read error: 110 (Connection timed out)) 03:39:57 --- quit: Pyld ("Leaving") 03:40:52 --- join: draq (ident@203-219-179-14-nsw.tpgi.com.au) joined #osdev 03:44:29 --- join: SIS-1650-01 (~sis@as11-2-1.rny.s.bonet.se) joined #osdev 03:57:32 --- join: erlingre (~erlingre@algordo.ifi.uio.no) joined #osdev 04:08:36 --- quit: erlingre ("Client Exiting") 04:09:19 --- join: kyelewis (~kye@220.240.67.180) joined #osdev 04:09:25 argh 04:09:28 stupid connection 04:13:59 --- quit: file (Read error: 110 (Connection timed out)) 04:28:27 --- quit: quantis (Read error: 104 (Connection reset by peer)) 04:32:51 --- join: petrusss (~petrusss@h178n2fls31o862.telia.com) joined #osdev 04:37:24 --- nick: gab_ -> gab 04:38:33 hi 04:51:25 --- join: trans (vuzicd@fatwire-201-147.uniserve.ca) joined #osdev 04:53:29 hi 04:53:34 :P 04:55:46 it is a weird day 04:56:15 hmm 04:56:33 yey, t-net has now pretty much completely changed to buzzpot :) 04:59:49 --- quit: redblue (Read error: 104 (Connection reset by peer)) 05:23:08 --- quit: Matzon (Read error: 104 (Connection reset by peer)) 05:27:41 --- join: lodda (~playgroun@h31n2fls31o965.telia.com) joined #osdev 05:31:41 --- quit: trans (Read error: 60 (Operation timed out)) 05:38:03 --- join: mrMister (~andri@ti122110a080-5011.bb.online.no) joined #osdev 05:44:00 --- join: yuriz (~yuriz@rcr.teraflops.com) joined #osdev 06:12:57 --- quit: mors (Remote closed the connection) 06:31:59 --- part: yuriz left #osdev 06:36:09 --- join: mur (~mur@uiah.fi) joined #osdev 06:36:13 privet 06:51:14 --- join: auto (~auto@174.80-202-49.nextgentel.com) joined #osdev 06:53:45 hello auto 06:53:57 more new nicks 06:54:03 wonder if auto talks 06:54:10 heh 06:54:38 kye, unside downside up 06:54:54 hello, how's blood in head there? 06:54:57 * kyelewis looks at mur 06:55:12 * mur looks kyelewis looking at mur -- upside down 06:55:18 ;) 06:55:35 remember, from here, *you're* upside-down :) 06:55:50 no 06:55:59 hi! 06:56:04 you just dont realise you are upside down 06:56:12 been here long time, auto? 06:56:34 no, just came home 06:56:44 i mean overall 06:56:54 :) 06:56:54 hehe 06:57:21 not really 06:57:51 first time? 06:58:00 or have you been here during nights 06:58:19 (shh.. dont tell kye, but australians irc only during nights and mornings) 06:58:35 what timezone are you in? I'm in norway. 06:58:42 haha 06:58:52 * mur is in homeland of irc 06:58:55 mur, you are so funny ;) 06:59:06 we have here so much free time, we need to invent irc to take some free time away 06:59:08 so funny i could just kick you up the backside :) 06:59:16 and i'd still be laughing my head off :) 06:59:40 i hugged in my dreams one girl. but then i had to get up 06:59:47 while sleeping 07:00:02 --- join: trans (eujhph@fatwire-201-147.uniserve.ca) joined #osdev 07:00:12 actaully she was my tutoring student so it might not be best to do that .. often 07:00:13 :) 07:08:09 ACTION! 07:08:09 --- quit: asm (Read error: 104 (Connection reset by peer)) 07:08:13 * mur requires action 07:08:49 muuuuuuuuuuur 07:08:54 is that enough action? 07:09:39 lol 07:10:14 no 07:10:17 more action 07:10:32 some girl shoudl come hug me now! 07:10:39 or large gorilla to fight against 07:10:46 (and i shoudl win that game) 07:11:46 hmm 07:12:02 girl 07:12:04 hmm 07:12:20 so much girls today 07:12:33 or then you coudl help me picking a laptop 07:12:39 you probably know about that one more 07:13:57 --- join: Matzon (Mazon@80.197.154.22) joined #osdev 07:14:09 mzzson 07:14:12 --- quit: auto (Remote closed the connection) 07:14:24 mur: i doubt it 07:14:47 if you meant me 07:14:50 hmm 07:14:53 kyelewis 07:14:59 --- join: auto (~auto@174.80-202-49.nextgentel.com) joined #osdev 07:15:00 or generally about PC 07:15:11 is studying harder than school? 07:15:23 aren't they the same thing? 07:15:41 well... high school and uni i mean :) 07:16:04 it's more broad kind of knowledge and information 07:16:10 and more specified 07:17:31 hmm 07:17:38 if i had enough money i'd buy a centrino notebook with 512 MB RAM, 60GB, WLAN and Ati Raedon 9000 mobility or so 07:17:46 that's kick ass 07:17:57 * lodda dreams again :) 07:18:04 lol 07:18:39 what's with centrino processors? 07:18:44 why do they cost more? 07:19:39 lodda 07:20:22 they have a lot of cache(1MB) and they are low energy 07:20:55 they're fast and dont spend energy? 07:21:02 how fast are they compared to celerons? 07:21:53 ugh no idea.. they're much likely much faster :) 07:22:00 just look for benchmarks or something 07:22:52 1.3 ghz centrino laptop costs 1550 eur, and celeron 2,4 1100 eur 07:23:33 http://www.kampusdata.fi/?ID=2003 07:23:38 which one woudl you choose 07:23:44 if it shoudl cost 1000-1500 07:24:02 it's for my cousin, he wants to play some games and compose music with some software that i dont know 07:24:19 but music software ought to work in all i think 07:24:52 --- nick: kyelewis -> kyelewis_z||z 07:24:55 night ppls 07:24:59 night 07:25:07 good night 07:25:17 i don't know... has one of it 3d accleration? 07:25:39 and btw sound of most notebooks is not so good... 07:26:02 with headphones neither? 07:26:44 dunno... guess that's better 07:26:53 he's in the army currently 07:27:00 so i doubt he'd play it outloud 07:27:19 poor boy 07:27:20 so travelmate, fujitsu amilo or hp business home work 07:27:35 HP Business Notebook nx9005 07:28:00 i think it would be used for more than year 07:28:06 or 2 07:28:37 --- join: asm (~user@dsl-082-082-159-109.arcor-ip.net) joined #osdev 07:29:17 reasm 07:29:39 hmm? 07:30:42 idle knows 07:30:43 tell me 07:30:51 * mur wants laptop 07:30:58 --- join: I440r (~foo@saturn.vcsd.com) joined #osdev 07:30:59 dell or sony :) 07:31:02 but has to send offer with cousins laptop offer too 07:31:07 why? 07:31:26 because I have experience with dell and sony's vaios look cool 07:31:35 lol 07:31:40 heh 07:31:58 i had bouth Dell latitude and Dimension years ago 07:32:02 *bought 07:32:41 (it shoudl cost around 1000-1500) 07:32:56 sony vaios are more expensive afaik :) 07:32:56 yeap, they are cheap now 07:33:50 --- quit: kyelewis_z||z (Read error: 60 (Operation timed out)) 07:34:26 hmm.. 07:35:56 hell, i'd need my cellular to call cousin and mom 07:36:02 but i have only keys and i'm in library 07:36:36 check this out 07:36:42 http://www.pcworld.com/reviews/article/0,aid,112033,00.asp 07:36:51 complete detailed review 07:38:26 * mur checks 07:41:47 second table is for budget laptops 07:41:54 :) 07:41:57 thanks 07:42:17 i never heard of their top 1.. 07:42:21 Chem USA ?!? 07:53:09 --- join: ide (~idle@208.27.1.9) joined #osdev 07:53:12 --- quit: idle (Nick collision from services.) 07:53:18 --- nick: ide -> idle 07:55:38 --- quit: draq (Remote closed the connection) 07:56:53 --- join: GTCoder (~dan@r40h63.res.gatech.edu) joined #osdev 07:57:03 --- quit: GTCoder (Read error: 54 (Connection reset by peer)) 07:57:04 --- join: GTCoder (~dan@r40h63.res.gatech.edu) joined #osdev 08:00:16 idle 08:00:17 weird 08:00:27 ? 08:00:29 only dell of value ones for sale in finland 08:00:39 hm 08:00:41 although i know gateway by other contacts 08:00:52 maybe toshiba too 08:01:11 oh there's ibm 08:01:17 but no toshiba. wrong lists :) 08:01:44 wow.. the toshiba on the first list is $3k 08:04:00 that list sounds like propaganda 08:04:03 :P 08:04:34 why not get desktop 08:06:05 because he needs to take it with him 08:06:17 I thought this was for you 08:06:23 and because they have desktop and it's died always 08:06:25 nah 08:06:26 sorry 08:06:41 * mur needs to pick cousin one before he can get powerbook :9~~ 08:06:44 drool 08:07:17 mmm 08:08:21 capitalism is stupid 08:14:18 No, it's the way to salvation. 08:15:11 okay 08:16:17 do you have opinoin about the following: Fujitsu Amilo 7600, HP Business Notebook nx9005 and Acer TravelMate 244LM 08:29:40 ugh 08:29:45 you dont know! 08:31:24 --- quit: gab ("leaving") 08:34:38 --- join: bedboi (~bedboi@host44-137.pool80181.interbusiness.it) joined #osdev 08:34:41 hi there 08:34:45 when i try to insmod a 2.6.0 kernel module (that i'm trying to port from 2.4.x) i get "insmod: error inserting 'nistnet.ko': -1 Invalid module format" and in dmesg i read: "nistnet: no version magic, tainting kernel." 08:35:20 salve 08:35:30 hehe 08:36:47 mur: hi there 08:36:59 mur: got some hints? 08:37:49 no none what so ever 08:37:55 but i need to get somethign to eat 08:38:08 dunno what that should be 08:38:16 bread might be okay 08:47:30 --- quit: trans (Read error: 110 (Connection timed out)) 08:53:16 --- join: gab (~prfalken@80.65.224.254) joined #osdev 08:53:36 mm garlic bread 08:55:10 sounds like a good idea! 09:09:23 bbl 09:16:16 --- quit: I440r (Read error: 54 (Connection reset by peer)) 09:25:03 --- quit: petrusss (Read error: 104 (Connection reset by peer)) 09:25:23 --- join: petrusss (~petrusss@h178n2fls31o862.telia.com) joined #osdev 09:51:17 --- join: Prophet__ (~Prophet@217.88.220.203) joined #osdev 09:53:37 --- quit: auto ("Client exiting") 09:57:23 --- quit: Prophet_ (Operation timed out) 10:01:23 --- join: demise (~e9o@c-a34272d5.01-94-7673741.cust.bredbandsbolaget.se) joined #osdev 10:07:02 --- join: Dr_Evil (DSLflat@p508FCFAE.dip.t-dialin.net) joined #osdev 10:10:39 --- join: kernel-panic (rewt@ANice-205-1-3-107.w81-53.abo.wanadoo.fr) joined #osdev 10:11:00 hello * 10:16:21 --- join: trans (edanuo@fatwire-201-147.uniserve.ca) joined #osdev 10:28:55 hello 10:34:59 --- quit: GTCoder (Remote closed the connection) 10:39:15 --- quit: CrayT3E (Remote closed the connection) 10:42:38 --- join: CrayT3E (~CrayT3E@finsler.math.uoa.gr) joined #osdev 10:57:00 --- join: pr1 (~pr1@ras04.ras.SBG.AC.AT) joined #osdev 11:03:24 --- join: redblue (star@ppp049.216-96-207.sherb.mt.videotron.ca) joined #osdev 11:13:31 heya folks 11:16:01 --- quit: HeavyJoost (Remote closed the connection) 11:18:06 hey geist 11:18:54 --- join: HeavyJoost (~heavyjoos@ditwilookwel.xs4all.nl) joined #osdev 11:20:03 --- quit: bedboi ("Client exiting") 11:21:26 --- join: GTCoder (~dan@r40h63.res.gatech.edu) joined #osdev 11:23:07 hi asm 11:23:29 geist, when will they announce the winner 11:23:56 I dunno, it's only noon here 11:24:01 so there's still some time left 11:24:18 I think they'll know basically who won, but it'll take a while to get official results 11:24:40 and if Arnold won, a new round of legal stuff will start, and he wont take office for a few months at least 11:24:45 hm 11:25:08 the secretary of state in CA has to approve the results, and the current one is one of Davis's buddies 11:25:25 and he has 30 or 45 days or so to approve it, so I'm sure he'd wait until the absolutely last minute 11:25:30 wtf approve ?! 11:25:52 well, he can't really disapprove it 11:26:16 but he has to make sure the electon went through, etc. he's the guy that signs the paper that says who was elected 11:26:30 who knows what stall tactics they'll use 11:26:53 I'm no fan of davis, but arnold is even worse 11:27:10 the least worst of.. 11:27:45 umm, governers 11:28:04 the real stuff behind the scenes are the folks that are driving arnold's campaign 11:28:24 arnold is going to be a puppet for sure 11:30:43 what they had to do a recount 11:30:57 *if 11:31:36 --- join: air (~brand@12-210-160-198.client.attbi.com) joined #osdev 11:31:49 I dont know CA's recount procedure 11:32:02 but if it's really close, then yeah someone might force a recount 11:32:51 yay 11:33:09 there are some strange news on cnn.com... "Moroccan teacher throws pupils out the window" 11:33:14 --- nick: Ricardo -> Rico 11:33:36 yeah, that's awesome 11:33:48 awesome? 11:33:50 lol 11:34:03 the teacher told them to shut up 11:34:05 hah 11:36:02 fractured shoulder 11:36:13 hehe 11:36:23 the kids should be lucky the classroom was on the first floor 11:36:32 haha 11:40:42 damn HL 2 is delayed till April now 11:41:29 * Odin- ponders the logic behind that. 11:42:10 --- quit: pr1 (Remote closed the connection) 11:42:23 yup 11:42:26 It doesnt make sense "... April 2004. Just the time to rewrite parts of the game".. their sources were on paper ?! 11:42:41 --- join: pr1 (~pr1@ras04.ras.SBG.AC.AT) joined #osdev 11:42:49 my guess is they needed a little more time anyway to polish it off, and this is a great excuse 11:43:06 heh 11:43:22 --- quit: redblue (Read error: 104 (Connection reset by peer)) 11:43:29 but they probably need to redo parts of the game that deals with multiplayer stuff to make it a little harder to cheat and hack it 11:43:38 the leaking of the source so widely makes it super easy to do that 11:49:20 --- join: swivelleroo (~kamasutra@Mix-Dijon-116-1-224.w193-250.abo.wanadoo.fr) joined #osdev 11:49:33 --- part: swivelleroo left #osdev 12:07:48 --- join: witten_ (~witten@adsl-gte-la-216-86-199-140.mminternet.com) joined #osdev 12:07:50 wow, CA's GDP is now the 5th largest 12:08:22 --- quit: pr1 (Remote closed the connection) 12:08:55 i forget what GDP stands for.. Gross... 12:09:31 Gross Departmental Product 12:09:38 Gold Down Payment 12:11:34 gross domestic product I believe 12:17:24 --- join: setre (~jussi@h136n1fls301o291.telia.com) joined #osdev 12:17:34 --- part: setre left #osdev 12:20:50 geist: yup 12:24:24 --- nick: abydos -> file 12:43:27 --- join: file[laptop] (~file@mctn1-2201.nb.aliant.net) joined #osdev 12:49:58 --- quit: witten_ ("Client exiting") 13:00:26 --- quit: trans (Read error: 110 (Connection timed out)) 13:01:25 --- join: eniac (~jonasg@181.180-136-217.adsl.skynet.be) joined #osdev 13:10:26 --- join: pr1 (~pr1@ras04.ras.SBG.AC.AT) joined #osdev 13:27:56 --- join: Pyld (~ere@sos-dhcp073.studby.uio.no) joined #osdev 13:28:48 --- join: Ward (ward@madoka.isw.student.khleuven.be) joined #osdev 13:28:53 anyone can explain to me what is the common practice, when an interrupt occurs during a system call? is the system call preempted like a regular user process? 13:31:38 --- part: Ward left #osdev 13:32:51 --- part: Pyld left #osdev 13:33:00 --- join: Pyld_ (~ere@sos-dhcp073.studby.uio.no) joined #osdev 13:40:21 --- join: dax_ (~dax@81.11.137.15) joined #osdev 13:41:06 hello? 13:42:45 yo Pyld_ 13:43:01 Pyld_: depends on how your kernel is constructed 13:43:18 a totally non-preempive kernel (ie all interrupts disabled) would become very slow 13:43:37 --- join: witten_ (~witten@ip-64-32-131-193.dsl.lax.megapath.net) joined #osdev 13:43:49 so aiming for as much preemptivity (is that even a word?) as possible seems to be a Good Thing 13:43:50 I, for one, welcome our new Terminator Overlord 13:44:06 I'm working on a small preemptive kernel ( assignment in an os course ) 13:44:08 --- quit: GTCoder (Remote closed the connection) 13:44:10 --- quit: witten (Read error: 54 (Connection reset by peer)) 13:44:13 --- nick: witten_ -> witten 13:45:06 but how is it done in for instance linux? 13:48:46 --- quit: demise () 13:52:18 --- quit: dax (Read error: 110 (Connection timed out)) 13:52:41 witten, he won ?! 13:55:00 --- quit: debug (Remote closed the connection) 13:57:47 --- quit: Dr_Evil () 14:00:32 idle: no, not yet 14:03:32 hmm 14:08:18 --- join: nullify (nullify@pool-141-153-207-132.mad.east.verizon.net) joined #osdev 14:08:30 --- quit: Smari (Remote closed the connection) 14:08:52 --- join: Smari (~spm@optimized.bitcode.org) joined #osdev 14:15:04 --- quit: Pyld_ (Read error: 110 (Connection timed out)) 14:17:28 --- join: Pyld_ (~ere@sos-dhcp073.studby.uio.no) joined #osdev 14:17:29 --- quit: idle ("Sic Sempere Tyrannis!") 14:20:02 --- quit: file[laptop] () 14:23:57 --- join: I440r (~foo@saturn.vcsd.com) joined #osdev 14:24:22 --- quit: pr1 ("Client Exiting") 14:24:28 --- join: trans (rtafss@fatwire-201-147.uniserve.ca) joined #osdev 14:25:41 --- quit: Pyld_ ("http://www.killozap.com/hmm/") 14:26:04 --- join: wossname (wossname@HSE-QuebecCity-ppp80489.qc.sympatico.ca) joined #osdev 14:28:53 --- quit: nullify () 14:31:39 --- quit: lynx (Read error: 110 (Connection timed out)) 14:31:57 --- join: jsr (www@du-15-134.ppp.telenordia.se) joined #osdev 14:34:06 --- join: lynx (~lodsb@p508093E3.dip.t-dialin.net) joined #osdev 14:34:35 --- join: ptlo (~senko@cmung2681.cmu.carnet.hr) joined #osdev 14:48:52 --- join: idle (~idle@208.27.1.9) joined #osdev 15:05:44 --- join: thib (~thib@bofh.bitcode.org) joined #osdev 15:06:02 Hola! 15:06:19 Bleesašur. :) 15:06:23 poke 15:06:27 -e 15:06:29 +s 15:06:30 :) 15:06:32 Hi 15:06:34 lol 15:06:45 Robert: ertu ķslendingur ? 15:07:04 --- quit: trans (Read error: 110 (Connection timed out)) 15:07:21 thib: nei 15:07:31 Pffff 15:07:36 Det vet du mycket väl. 15:07:45 thib: hann er sęnskur 15:07:56 :ž 15:08:15 --- quit: petrusss () 15:08:17 Aha 15:09:08 Ja. 15:09:11 Ser du inte det? 15:16:08 --- quit: jsr (Read error: 60 (Operation timed out)) 15:18:17 --- quit: I440r (Read error: 60 (Operation timed out)) 15:20:19 --- quit: mrMister ("gone") 15:23:02 mr blister 15:31:22 --- quit: ptlo ("Client exiting") 15:31:34 --- quit: SIS-1650-01 (Read error: 60 (Operation timed out)) 15:37:54 --- join: gianluca (~gl@ppp-165-133.28-151.libero.it) joined #osdev 15:42:57 --- join: wcstok (strtok_r@dialup-67.31.178.17.Dial1.Denver1.Level3.net) joined #osdev 15:43:34 --- join: newbs (~tumbler@ts1-illavl75.shawneelink.net) joined #osdev 15:45:20 --- join: jwesley (~jwesley@adsl-18-192-208.mem.bellsouth.net) joined #osdev 15:50:19 --- quit: air (Remote closed the connection) 15:50:48 --- join: air (~brand@12-210-160-198.client.attbi.com) joined #osdev 15:55:22 --- quit: thib ("leaving") 15:58:19 http://www.itworld.com/Man/2685/031006sgisco/ 15:58:30 praise sgi 16:01:16 hopefully now some firms don't have to buy sco licenses for their computers running linux 16:01:44 like mentioned from sco 16:03:15 --- join: glguida_ (~gl@ppp-228-135.28-151.libero.it) joined #osdev 16:17:35 --- quit: gianluca (Read error: 110 (Connection timed out)) 16:20:11 --- join: petrusss (~petrusss@h178n2fls31o862.telia.com) joined #osdev 16:21:52 --- quit: glguida_ (Read error: 60 (Operation timed out)) 16:22:06 blargh 16:23:43 --- quit: petrusss (Client Quit) 16:23:48 geist: ? 16:26:46 --- quit: jwesley (Read error: 104 (Connection reset by peer)) 16:27:04 witten: phooie! 16:27:36 watching cartoons? 16:28:24 nah, writing code 16:29:22 geist: having problems? 16:30:36 --- join: glguida_ (~gl@ppp-80-133.28-151.libero.it) joined #osdev 16:31:37 nah 16:31:42 just waiting for a build to complete 16:31:50 --- join: petrusss (~petrusss@h178n2fls31o862.telia.com) joined #osdev 16:31:53 --- quit: petrusss (Remote closed the connection) 16:31:57 --- join: jwesley (~jwesley@adsl-18-192-208.mem.bellsouth.net) joined #osdev 16:33:53 --- quit: glguida_ (Client Quit) 16:33:55 --- join: gianluca (~gl@ppp-80-133.28-151.libero.it) joined #osdev 16:34:41 --- join: trans (ebzeny@fatwire-201-147.uniserve.ca) joined #osdev 16:38:27 though builds are a lot faster on my new linux box at work 16:41:53 --- quit: wossname ("HAK THE PLANET, Hack THE PLAnet~!#") 16:50:51 geist: yay 16:55:09 --- join: GTCoder (~dan@r40h63.res.gatech.edu) joined #osdev 16:59:39 --- join: petrusss (~petrusss@h178n2fls31o862.telia.com) joined #osdev 17:00:45 --- quit: petrusss (Client Quit) 17:05:46 what the hay 17:12:48 --- quit: idle ("Sic Sempere Tyrannis!") 17:14:33 --- quit: wcstok () 17:15:21 --- join: idle (~idle@208.27.1.1) joined #osdev 17:48:57 --- quit: trans (Read error: 60 (Operation timed out)) 17:57:49 --- quit: gianluca ("Client exiting") 18:23:10 --- join: Mathis (~anyone@pD9EABFDE.dip.t-dialin.net) joined #osdev 18:54:22 --- quit: Mathis (Read error: 110 (Connection timed out)) 18:56:06 --- join: CLxyz (CLxyz@209.53.54.1) joined #osdev 19:09:19 --- quit: CrayT3E (Remote closed the connection) 19:12:03 --- join: Mathis (~anyone@pD9EABB31.dip.t-dialin.net) joined #osdev 19:12:11 --- quit: kernel-panic (Read error: 54 (Connection reset by peer)) 19:14:07 --- quit: Mathis (Client Quit) 19:17:19 --- join: CrayT3E (~CrayT3E@finsler.math.uoa.gr) joined #osdev 19:24:41 --- join: trans (vkmysm@fatwire-202-106.uniserve.ca) joined #osdev 19:29:06 --- join: kslc (~clsk@dsc07-mil-fl-199-183-233-37.rasserver.net) joined #osdev 19:56:16 --- join: bugx0r (~bugx0r@2001:470:1f00:782:2e0:7dff:fea9:b8f6) joined #osdev 20:04:48 --- quit: trans (Read error: 60 (Operation timed out)) 20:08:33 pretty quiet, anything interesting going on anywhere? 20:09:02 #debian-devel is being a bunch of dickweeds 20:09:04 that's all 20:09:13 why are they being assmunches? 20:09:46 they don't like the software I use to filter my email 20:13:29 --- quit: krux ("Client Exiting") 20:14:16 --- join: krux (~krux@66.14.118.85) joined #osdev 20:14:42 fuck em 20:14:50 yeah 20:15:54 cantankerous virgin stinkypants 20:15:56 the lot of them 20:17:44 --- quit: krux (Client Quit) 20:22:04 oh my god it's gonna be arnold 20:22:07 this is so sad 20:25:42 --- quit: newbs (Client Quit) 20:31:30 yep 20:50:49 --- quit: witten ("Client exiting") 21:01:44 --- join: mAnTeKiYa (~clsk@dsc07-mil-fl-199-183-233-37.rasserver.net) joined #osdev 21:04:14 --- quit: kslc (leguin.freenode.net irc.freenode.net) 21:04:14 --- quit: Robert (leguin.freenode.net irc.freenode.net) 21:04:14 --- quit: file (leguin.freenode.net irc.freenode.net) 21:04:14 --- quit: Rico (leguin.freenode.net irc.freenode.net) 21:04:14 --- quit: rdragon (leguin.freenode.net irc.freenode.net) 21:04:14 --- quit: Divine (leguin.freenode.net irc.freenode.net) 21:04:15 --- quit: jwesley-work (leguin.freenode.net irc.freenode.net) 21:05:19 --- join: kslc (~clsk@dsc07-mil-fl-199-183-233-37.rasserver.net) joined #osdev 21:05:19 --- join: Robert (~snofs@h31n2fls31o965.telia.com) joined #osdev 21:05:19 --- join: file (~joshnet@mctn1-2201.nb.aliant.net) joined #osdev 21:05:19 --- join: Rico (Rico@node-c-75bb.a2000.nl) joined #osdev 21:05:19 --- join: rdragon (~hi@m-j32.rh.sunyit.edu) joined #osdev 21:05:19 --- join: Divine (~john@12-246-112-182.client.attbi.com) joined #osdev 21:05:19 --- join: jwesley-work (~jwesley@tkgeisel.com) joined #osdev 21:05:19 --- mode: orwell.freenode.net set +b [*]!*@* 21:05:52 --- quit: Robert (Read error: 104 (Connection reset by peer)) 21:05:56 --- quit: jwesley-work (Remote closed the connection) 21:05:56 --- join: Robert (~snofs@h31n2fls31o965.telia.com) joined #osdev 21:06:28 --- quit: kslc (Success) 21:06:29 --- join: jwesley-work (~jwesley@tkgeisel.com) joined #osdev 21:09:12 --- nick: mAnTeKiYa -> clsk 21:12:23 --- quit: Odin- (Read error: 113 (No route to host)) 21:21:45 --- quit: bugx0r ("time to sleep") 21:29:18 --- join: trans (bjxfjr@fatwire-202-32.uniserve.ca) joined #osdev 21:35:52 --- quit: lodda (Remote closed the connection) 21:37:26 --- join: lodda (~playgroun@h31n2fls31o965.telia.com) joined #osdev 22:08:31 --- join: Odin- (~sbkhh@adsl-2-216.du.snerpa.is) joined #osdev 22:39:09 --- join: debug (~debug@tab.csbnet.se) joined #osdev 22:39:30 gmorgning wofld 22:39:38 * debug needs coffee 22:40:55 hiya 22:41:05 we elected arnold! isn't that great! (no it isn't) 22:42:48 :) 22:43:09 the american dream 22:43:22 to be ruled by an austrian 22:43:23 lived by an austrian 22:43:27 heh 22:44:12 geist: have you made progress on the usb stuff? 22:44:32 yeah, I checked in a bunch of code the other day 22:45:00 got control transfers working on ohci, managed to talk to devices a bit, so now I'm building up the bus manager, which is where the bulk of the work will go 22:45:52 ok 22:47:42 --- quit: trans (Read error: 60 (Operation timed out)) 22:47:42 --- quit: Matzon () 22:47:58 * debug added a simple graphical framebuffer device to his emulator this morning 22:48:02 small steps... 22:48:33 coolio 22:49:02 netbsd outputs pixels to it, writing text, but doesn't really get very far 22:49:20 now what are you emulating? 22:49:27 I seem to add support for new machines / devices before I have finished what I've already begun. 22:49:29 mips64 22:49:33 or any mips platform 22:49:45 this morning I added hpcmips, used in some handheld thingies 22:49:46 port newos to it! 22:49:56 perhaps later 22:50:03 :) 22:50:05 yay! 22:50:21 it's actually quite easy to port newos 22:50:38 partially because the hardware abstraction is pretty good, and partly because it doesn't do anything 22:50:49 have you ported newos to any mips platform yet? 22:51:07 no, I started to port it to my indy r5k, but didn't get too far 22:52:41 well, in time you probably will 22:53:11 it was mostly due to a lack of info about sgi boxes 22:53:15 perhaps caches and stuff like that can be a bit complex (as it varies with each CPU, not just which ISA it supports) 22:53:22 it wasn't terribly straightforward, so I didn't continue 22:53:26 well, that's a problem too 22:53:54 right, and netbsd folks dont have it working quite right on r5ks 22:54:03 so I didn't really even have any reference code to work from 22:54:16 there were no docs available on the web? 22:54:25 I've found "good" docs on r4k, and some on r10k 22:54:37 there are some on the 5k 22:54:43 good = they have errors, but seem to contain everything needed to write an OS or emulator 22:54:58 it's caching is different from 4k, and it supports the 64bit isa 22:55:08 r4k is 64bit 22:55:19 oh, is 4400 64? 22:55:20 but caches are usually unique for each cpu 22:55:22 yes 22:55:35 I don't remember if it is MIPS III or IV though 22:55:38 but it is 64bit 22:55:43 didn't know that at all, I was under the impression that all of the 4ks are 32 22:55:49 ah, maybe that's the difference for 5k 22:55:58 I seemed to remember that it had something that the 4ks didn't have 22:56:00 maybe it's IV 22:56:06 some of the 4ks are not multiprocessor capable, but they are all 64bit 22:56:13 --- quit: clsk (Read error: 110 (Connection timed out)) 22:56:21 I dont know too much about the mips line, there's bound to be a reference page somewhere 22:56:38 I seemed to remember to stay away from the 8k or one of the higher up ones 22:56:48 it's totally different from the others, and only used in a few places 22:56:59 the general feeling I get about mips is: for userspace (that only depends on the ISA), it's really really nice. for kernel stuff: you need to have support for each individual cpu 22:57:11 right 22:57:21 pretty pure RISC stuff 22:57:30 lots of cpus are that way, they just totally change the system mode stuff when they feel like it 22:57:45 powerpc is pretty consistant, thoughy you still have to work around bugs in each version, and the cache stuff is different 22:58:00 sparc can definitely change between revisions (v7, v8, v9) 22:58:22 I know that power varies bigtime between versions 22:58:31 I don't think mips can run usercode with caches disabled, so you _have_ to support the caches. kernel code is ok uncached though 22:59:06 I'm not sure about that, the netbsd guys have got netbsd booting on r5ks, but I think that involved disabling the caches, which means it runs like ass 22:59:13 hm 22:59:19 then it is cpu dependant 22:59:20 but it could be they didn't totally disable it, or they flush it all the time to work around other bugs, etc 22:59:36 --- join: Ninja_KtS (~kye@220.240.67.180) joined #osdev 22:59:45 --- nick: Ninja_KtS -> kyelewis 22:59:54 anyway, gotta go cover the bike. it's had to have cooled off by now 23:00:01 good afternoon boys and girls ;) 23:00:05 well, the r2k and r3k differ a lot from the r4k. on r2k/3k you access memory in special patterns to control the caches, on the r4k you have separate (new) instructions, not part of the MIPS ISA 23:00:15 gmornign :) 23:04:26 cpu0 at mainbus0: MIPS R4000 CPU (0x400) Rev. 0.0 with software emulated floating point 23:04:26 cpu0: 32KB/32B direct-mapped L1 Instruction cache, 48 TLB entries 23:04:26 cpu0: 32KB/32B direct-mapped write-back L1 Data cache 23:04:35 at least netbsd thinks it's working... 23:13:52 yeah, but it probably sees rev 0.0 and thinks 'I can't do anything with this shit, I'm outta here' and then spins 23:14:18 no no, it gets further 23:14:30 http://www.mdstud.chalmers.se/~md1gavan/mips64emul/ <-- the first "example" 23:16:53 if I had an INSTALL ramdisk kernel, that would be nice. but there are none of those precompiled, as the linux-aware firmware in those machines cannot handle that large kernel images :( 23:20:22 it only hangs right now because there's no boot device, and I haven't implemented console input yet :) 23:21:15 LAME! what kind of programmer are you? 23:22:12 :) 23:22:14 the lazy kind 23:22:40 damn, I keep trying to play day of defeat but it keeps crashing 23:41:25 --- quit: Divine ("My damn controlling terminal disappeared!") 23:55:20 --- quit: eniac ("Lost terminal") 23:59:59 --- log: ended osdev/03.10.07