irclog2html for blob on 2001.10.08

00:54:01Sammyhello TauDog
02:20:09Russheh
08:45:14erikmmorning
08:45:21Russmorning
08:46:08ed__'tag!
08:52:10erikmyawns
10:28:34WeirdArmsHey erik!
10:29:07erikmWeirdArms!
10:29:39erikm*sigh* "How can I download my program/EXE in that board?"
10:29:47erikmto the linux-arm mailing list...
10:29:57erikmpolitely flames
10:30:13WeirdArmsHehehe
10:30:16WeirdArmsGuess what!
10:30:25WeirdArmsI got the first bit of the FASS ARM Linux stuff to work!
10:30:31erikmcool!
10:30:39erikmand? improvements?
10:30:47WeirdArmsHaven't measured
10:31:01WeirdArmsgota remove all the extra TLB/Cache flushes first
10:31:06erikmhehe
10:31:06WeirdArmsthen a bit more work to do
10:31:21WeirdArmsbut the basic CPD structure works
10:31:23WeirdArmsfinally
10:31:44erikmbtw, did davem ever merge your superfluos TLB/Cache flushes patch?
10:31:44WeirdArmsone little bit of code missing coz of the evil way Russel uses the Domain Access Control Register
10:31:59WeirdArmserikm: It was a patch-reversion
10:32:10WeirdArmsDavem talked to Linus
10:32:17erikmand linus ignored him?
10:32:20WeirdArmsthey found the patch-reversion and fixed it to what it should have been
10:32:25WeirdArmsit got fixed
10:32:29erikmah, good
10:33:12erikmwell, I completely restructured blob :)
10:33:56erikmI'm currently putting a good fix for the pleb serial port in
10:34:14erikmneeds coffee
10:35:23WeirdArmscool
10:38:56WeirdArmsBasicall for starters it will only reduce the number of TLB flushes
10:39:09WeirdArmsI gota look at the caching stuff again to see how to go about that
10:39:30WeirdArmsvirtual aliases on shared memory could be a problem
11:02:50WeirdArmserikm: where is koen?
11:25:40erikmreturns
11:26:02erikmwell, not in my room
11:26:20erikmoh, yeah, he's in berlin at the moment, AFAIK
11:34:28WeirdArmsok
11:34:38WeirdArmswonders why Russel does the things he does
11:34:47WeirdArmshey is there an arm linux # somewhere?
11:35:10erikmoh, koen just jumps in
11:35:27WeirdArms?
11:35:40WeirdArmsinto your room?
11:35:55WeirdArmsI guess so
11:36:18koenI am back from berlin but still have to change my .forward
11:36:36koenmaybe I will not. It is very quiet.
11:36:59WeirdArmshehehe
11:37:07WeirdArmsso you got my email then?
11:37:12koenyes
11:37:39WeirdArmsSorry I've been so quiet
11:37:57koenlike I said : i don't mind silence :-)
11:37:59WeirdArmsI got stuck Subject adminning till my Supervisor spat the dummy at the school
11:38:19WeirdArmsLike I said, I'm making progress again
11:38:20koenExcuses accepted: get to work
11:38:24WeirdArms:)
11:38:32WeirdArmsAimming for Usenix Technical
11:38:53koenme too with another paper
11:39:01WeirdArmsok
11:39:16WeirdArmsWe will be trying to pair it up with our microkernel work
11:39:38WeirdArmsand or working with IA-64 Linux along the same line for a OSDI or SOSP paper too
11:39:52koenwe are targetting OSDI too.
11:39:59WeirdArmsWe just got a grant for 100,000+ in IA-64 SMP systems
11:39:59koenLooks like we will meet again
11:40:21WeirdArmsI will probably be in Germany in December
11:40:22koenErik likes o work down under, when can he apply for a job?
11:40:23ed__so you get like 2 ?
11:40:25ed__:D
11:40:29WeirdArmsI'm not sure what my travel well be
11:40:33WeirdArmsI might be able to pop it
11:40:38WeirdArms1 QUAD
11:40:43WeirdArms2 dual's I think
11:40:49WeirdArmsand 3 single processor workstations
11:41:11ed__setup a quake server
11:41:20WeirdArmsNah I got my own machine for that
11:41:34WeirdArms1GHz Athlon with 266Mhz bus and DDR ram ;)
11:41:47WeirdArmskoen: He would like to inturn in Sydney?
11:42:29koenI won't let him go!
11:42:32WeirdArmshehehe
11:42:37WeirdArmsfair enough
11:42:50WeirdArmsWell if he wants to he just has to ask
11:43:04ed__oh be nice to erik and let him go to oz!
11:43:26WeirdArmswho's ed?
11:43:50ed__random person
11:43:52WeirdArmshehehe
11:43:59WeirdArmsI think I"m going to have to head home soon
11:44:13WeirdArmsI'm going to miss the last ferry and if I don't
11:44:31WeirdArmskoen: Feel free to send a nag email
11:44:46koenpromise
11:44:47WeirdArmsafter next weekend that is ;)
11:45:08WeirdArmsTill tomorrow
11:45:14koenbye
12:14:36sammy_koen ? what's that mean ?
12:15:07erikmit doesn't mean anything, it's a name
12:20:47sammyjust like sammy_wms
12:21:27erikmwell, in this case it was koen who came in five minutes after WeirdArms asked for him
12:22:00sammyohh... ^_^
12:22:39sammyI got your jffs2 and ramdisk, thanx ...I'll try it at home...
12:23:51sammyand in my case sammy_wms is the nick with a computer in my room... :-)
12:25:08sammyerikm: I saw your pic on your web site before , and curious about what's youe age ?
12:25:40erikmcalculates
12:25:42erikm29
12:26:02sammyreally ? so young...
12:26:53erikmsammy: young? people start calling me "sir" :)
12:27:21sammymarry ?
12:27:26erikmnope
12:27:37sammygilefriend ?
12:27:47erikmcurrently not
12:28:24erikmbut you could have known that from my night long hacking sessions :)
12:30:02ed__heheh
12:30:16sammyyeh , but that's not mean "girlfriend" or "wife" are not important ....*_*
12:32:11sammyerikm : do you live in the wilderness ...? ;-D
12:32:38erikmlooks outside
12:32:47erikmnot really
12:33:17sammyit should be have some girl ...
12:33:48ed__yeah erik!
12:33:50erikmplays "you can't hurry love"
12:34:39sammyI think computer is a good girl friend , but ....
12:36:02sammyeven you can't hurry , but you still need a target right?
12:36:28sammybut not a computer ....
13:02:16Sammyis go home ...
13:07:12seletzhi all.
13:09:08seletzerikm: hi erik. Still angry 'bout my cmd-line history :) ?
13:09:47erikmwakes up
13:09:53erikmhi seletz, prpplague
13:10:19erikmseletz: I think I just solved it in a nice way
13:10:45seletzerikm: removed it, eh?
13:10:56erikmno, it's in the separate diagnostics tool
13:11:55seletzerikm: Very good. Thank you. any comments 'bot the code?
13:12:16erikmblob> boot
13:12:17erikmStarting kernel ...
13:12:17erikmdiag version 2.0.5-pre1
13:12:17erikmCopyright (C) 2001 Stefan Eletzhofer and Erik Mouw
13:12:17erikmdiag comes with ABSOLUTELY NO WARRANTY; read the GNU GPL for details.
13:12:18erikmThis is free software, and you are welcome to redistribute it
13:12:20erikmunder certain conditions; read the GNU GPL for details.
13:12:21erikmdiag>
13:12:39erikmbot code?
13:12:53seletzerikm: (beeing at home doing house-keeping)s/bot/about/g
13:13:15sammy_all guys ...see you later...
13:13:20erikmbye sammy_
13:13:29seletzsammy_: bye!
13:14:53erikmseletz: well, the diag code is very clean right now
13:15:13seletzerikm: well, cool. Did you have to clean up much?
13:15:33erikmseletz: after I split up blob in a library and a core part, diag was done in 30 minutes. and it even worked first time right
13:16:04seletzerik: ugh. library part? Dll-sort of thing?
13:16:32erikmseletz: diag is currently about 550 lines of code, much of it is standard GPL disclaimer
13:16:36prpplaguelater sammy
13:16:51erikmseletz: no, no DLL. just a static library with code
13:17:04erikmseletz: so both blob and diag link against -lblob
13:17:59erikmseletz: the library contains things like serial port IO, LEDs, string handling, timer, etc.
13:18:45seletzerikm: ahh. a little code-branch. hmmm. well, i think thats quite nice, brings in best of all to everyone!
13:19:02erikmseletz: library is about 1600 lines of code,
13:20:26seletzerikm: hmm. just another .section of cmds i suppose?
13:20:55erikmseletz: oh, yeah, the commandlist and initlist stuff is also in the library
13:21:33seletzerikm: well, this thingy just gets nicer every week, does'nt it? :))
13:21:58seletzerikm: well, this thingy just gets nicer every week, does'nt it? :))
13:22:11seletzddoouubbllee :::)))
13:22:26erikmseletz: so I cleaned up your command history by adding an __initlist() entry for cmdhist_init()
13:22:46prpplagueerikm: did you ever get anything usefull from that mizi blob?
13:22:57erikmprpplague: yeah, I got their code
13:23:04erikmbut they switched to bootldr
13:23:20seletzerikm: great! removes the need of the cmdhist_init() calls in push() next() and prev()
13:24:08erikmseletz: well, my personal style is not to put that kind of stuff in. if it fails, it's a programmers error
13:24:29erikmseletz: for userland code I would have put an assert() over there, but nothing more
13:26:29seletzerikm: well, me too. i usually write xxx_init() and xxx_destroy() functions in modules i write. blob did not do do this, so i coded it that way. forgot about your cool init-lists. definitely the right thing to do, isnt it?
13:28:52seletzerikm: i wish i could get OS 10.1, so i could just cvs update to see the code ... :(
13:29:27erikmoh, you have a mac at home
13:29:48seletzerikm: yes. i got to hate PCs.
13:30:06erikmseletz: well, I love PCs, but only if they run linux
13:30:23seletzerikm: :) my mac runs linux PPC :)
13:30:36erikmseletz: (or another unix flavour with enough sane userland tools)
13:31:20seletzerikm: but dual-boot is a pain as of now. (erver saw forth? then you saw mac's boot-prom user IF)
13:31:43erikmseletz: like the boot prom from a Sparc
13:32:01erikmseletz: actually, I like OpenPROM. it's sane compared to a PC's BIOS
13:32:19seletzerikm: well, IIRC you just sys boot, don't you? (on SPARC)
13:32:37erikmseletz: depends if you set the right aliases
13:32:42seletzerikm: BIOS === YUCK
13:32:47prpplagueerikm: was there anything in the mizi code worth keeping?
13:33:04seletzwhat's MIZI ?
13:33:29erikmseletz: otherwise you can boot /IOMMU@0xff000000/SBUS/3/ESP@0xfe000000/disk/3
13:34:06seletzerikm: hmmmm. looks _exactly_ like the MAC way. (wondering ...)
13:34:27erikmprpplague: it's not a unified diff, so hard to read the patch. I need to rediff it
13:34:57erikmseletz: powermac and sparc use the same kind of PROM: IEEE 1275 aka OpenPROM
13:35:53seletzerikm: hmm. _maybe_ it's worth that i take a closer look ...
13:36:54erikmseletz: unfortunately the PCI standardisation comittee missed the opportunity to make OpenPROM mandatory for PCI devices
13:37:09prpplagueerikm: btw, you want diffs sent to you exactly as described at the end of the readme right?
13:37:37erikmprpplague: yeah, unified diff is the easiest to read
13:37:54seletzsame as for kernel diffs i suppose.
13:38:08erikmprpplague: -u == unified, -r == recurse through directories, -N == also new files
13:38:39erikmprpplague: be sure to run "make distclean" in both trees, or otherwise diff will also include backup files and object files
13:39:04seletzwell, guys. I'm off again now. (house-keeping, as i said :( )
13:39:10erikmbye
13:39:16prpplaguethat was my next question..., could you hear the squeak gears?
13:39:26prpplaguelater seletz
13:39:45seletzbye all!
13:39:51erikmsqeak gears?
13:40:21erikmprpplague: btw, it's ok if you only send the diff to me, I know you'll send it anyway :)
13:43:15prpplagueerikm: i'm new to the collabrative work, i'm used to being the only one to use my "bastardized" code
13:46:36prpplaguebbiaf
14:02:51prpplagueis back, like a bad case athlete's foot
15:29:08Russdo people try to make us laugh on linux-arm?
15:30:42prpplaguethe mailing list?
15:31:30prpplagueRuss: you talking about the winblozes ce question?
15:40:16Russya
15:41:21prpplagueyes, one made me laugh this morning, almost choked on my orange juice
15:54:15sammyhello all ...
15:55:55prpplaguehowdy sammy
15:56:20prpplaguemaking or breaking good code today?
15:58:04sammynope , still reading ;-|
15:59:06prpplaguesammy: me too, pcmcia driver stuff is a nightmare
16:02:08sammycheer up : )
16:09:15seletzy
16:09:26seletzhi all!
16:41:52erikmre
16:42:26seletzerikm: hi erik.
16:43:49seletzerikm: did a little code-browsing via sf & web-cvs. Looks to me if blob and diag are completely different objects, and thus cannot be installed at the same time, right?
16:46:03erikmseletz: they are different objects, yes
16:46:13erikmseletz: you download diag as if it is a kernel
16:46:38seletzerikm: ah, that's different.
16:46:54erikmseletz: but they share quite some code
16:47:05seletzerikm: this way everything is ok. :))
16:47:47erikmseletz: btw, shoot if you have a better name for diag :)
16:47:58seletzerikm: hmm. quite some time lag here .... my server is brin.openprojects.net
16:48:33seletzerikm: diag is fine IMHO. nice & short
16:48:43erikmhmm, yeah. brin is in .au
16:49:19seletzerikm: ugh. cant do dig or host or nslookup un my mac :((
16:49:37seletzerikm: some public server in nl?
16:49:39erikmtry nl.linux.org aka merril.openprojects.net. it's very well connected to the dutch university network, which has a very good connectivity to the amsterdam internet exchange
16:50:10seletzpart blob
16:52:23seletz_at_homewell, here i am.
16:52:49erikmbetter like that?
16:53:00sammyis back
16:53:03seletz_hmmm. ping
16:53:29seletz_say pong :)
16:53:35erikmoh, sorry
16:53:37erikmpong
16:53:40seletz_:)
16:53:44seletz_good.
16:54:12erikmsammy: shouldn't you be sleeping? ;)
16:56:09seletz_erikm: any plans to put diag on a separate flash partition?
16:56:54prpplagueseletz_: i.e. >boot diag  ?
16:56:57prpplaguevs.
16:56:59seletz_erikm: so blob :: diag :: kernel :: ramdisk :: jffs2 layout (or similar)
16:57:09sammyyep, but I still reading about some MTD and jffs2 's doc , and test my new mointer
16:57:37seletz_erikm: hmm. want diag reside on jffs2 fsys?
16:57:51sammy< Sync Master 560V TFT >
16:58:32erikmnot yet plans to put it in a separate flash partition, but it would be possible as soon as we get proper partitioning support
16:58:43erikmand jffs2 should also be possible
16:58:55seletz_define "proper" :)
16:59:07seletz_"working" i suppose :)
16:59:08erikmwell, I haven't yet decided what to use
16:59:58seletz_erikm: how about using i.e. redboot (yuck!) partition scheme?
17:00:26seletz_erikm: quite an overkill, maybe
17:00:56erikmI haven't looked at it yet. bootldr partition support is also nice, but jamey hicks says they dropped it
17:01:31seletz_erikm: what, bootldr dropped or partition support?
17:01:54erikmthey want to pass the partition information to the kernel with new ATAG nodes, which is a really stupid idea IMHO
17:03:01erikmseletz_: I'm not sure if they dropped it. looking in the bootldr CVS tree it looks like they didn't
17:03:30seletz_erikm: agreed. flash should look more like a block dev.... those nodes get really bloated. rmk agreed about that?
17:04:44erikmwell, rmk didn't get a patch from jamey yet, so he didn't complain yet
17:04:56erikma patch for new nodes, that is.
17:05:06seletz_erikm: oh. lets see :)
17:06:13erikmseletz_: are you subscribed to linux-arm-kernel?
17:06:14prpplagueseletz_: the ELF board from inhand has a nice partitioning scheme, IMHO
17:06:40seletz_erikm: well, of course.
17:07:03erikmseletz_: ok, just read the rant I send to l-a-k sunday afternoon
17:07:12seletz_erikm: no way without the help of The List :)
17:08:02seletz_erikm: well, i'm not in my office now, so tomorrow. I guess its a nice thing to read :)
17:08:22erikmseletz_: here's the URL from the archive: http://lists.arm.linux.org.uk/pipermail/linux-arm-kernel/2001-October/005224.html
17:08:53seletz_erikm: shame on me. should have thought about that :/
17:08:58seletz_looking
17:14:26seletz_erikm: hmmm. I think the real problem is that one has to store the partition info in some non-volatile memory. Thats flash IMHO. so, they build a parser in the bootloader to read this info, build tags out of it and the kernel does pares this tags and builds the partition info. Nonsense IMHO. Just store it in flash and let one instance do the whole work. Kernel is the one instance (and the only one) that can do the whole work.
17:15:05seletz_erikm: correct?
17:15:31erikmexactly my point
17:15:49seletz_erikm: then i understood it :)
17:16:33prpplaguei'm a little slower, but i got the same
17:17:12seletz_erikm: btw, rmk is always quite funny to read: "Please explain in detail!" Reminds me always on certain teachers :))
17:17:49seletz_erikm: of course __not__ when you are to explain :P
17:18:43erikmseletz_: well, his experiences with the handhelds.org patches are not too good
17:19:03erikmseletz_: they usually code before they think
17:19:29seletz_erikm: ugh. better not do that with rmk twice ...
17:19:55seletz_erikm: did you met him in real life once?
17:20:57erikmseletz_: yes, we met in new york januari this year
17:21:14prpplagueerikm: what is it about the redboot partitoning that people don't like?
17:22:44erikmseletz_: we == rmk, nico, phil blundell, john dorsey, ralph siemsen, /me, jeff sutherland
17:23:12erikmprpplague: I don't know, haven't looked at it yet
17:23:35prpplagueseletz_: what about you?
17:24:02erikmprpplague: but I can imagine dwmw2 and rmk would like blob to use an existing partitioning format instead of inventing its own
17:24:16erikmprpplague: unless there is a very good reason to do it, of course
17:24:56prpplagueerikm: yes i'm not big on re-inventing the wheel either
17:28:16prpplaguei'm off to lunch, before i hit the pcmcia stuff again
17:31:06sammynight all  ^_^ ...
17:33:30seletz_prpplague: sorry, phone. well, IMHO redboot is a big bunch of code. Horribly code. Full of macros of the worst form. No makefiles. arm-elf gnu compiler needed to compile it. Thats one reason.
17:34:03erikmseletz_: well, even rmk didn't manage to compile redboot, so...
17:34:38seletz_prpplague: and the partition scheme sound quite complex, too. they have partitions, config areas, redboot backup areas and so on. Who needs that?
17:36:16seletz_erikm: i got it to compile, but my god, what a mess. i t seems that you need to have a windows configurator (or use a poory supported wxwindows clone) to configure packages (or you hack CML2 configs by hand). and you have to use arm-elf-gcc.
17:38:05seletz_what i think what we need is quite a simle directory-like block of start/stop offsets, possibly with  description. just like the hard-coded partitions in the kernel. perhaps even those structs.
17:38:25seletz_s/simle/simple/g
17:41:13seletz_i do'nt know if we need some sort of free-block list to be able to delete partitions. On one kind of system there should IMHO one kind of partition layout. But perhaps thats a too simple view.
17:42:14erikmseletz_: well, the bootldr partition table layout didn't look too bad
17:42:38erikmseletz_: it's basically a linked list
17:42:40seletz_erikm: maybe we should just "grab" it.
17:43:04seletz_erikm: linked list of occupied blocks or partitions?
17:43:12erikmlinked list of partitions
17:43:19erikmI'll ask rmk what he thinks
17:43:32seletz_erikm: hmm. why not use the same structs as the kernel?
17:43:50erikmseletz_: you mean the struct al viro just wants to change? ;)
17:44:29seletz_erikm: i dont know, i mean those hard-coded example structs in the partition support code.
17:51:49erikmhey Russ|werk
17:53:24Russ|werkhello
18:10:56erikmseletz_: for a particular bad example of some code from handhelds.org CVS, have a look at this patch:
18:11:05erikmhttp://www.arm.linux.org.uk/developer/patches/?action=viewpatch&id=597/1
18:11:32erikma PCMCIA driver that depends on... MTD???
18:27:30erikmhmm, that must have been the most busy day for blob CVS ever: 83 CVS changes yesterday :)
18:28:22prpplagueerikm: i see that as a sign that blob is rapidly evolving
18:29:10erikmwell, it needed a cleanup anyway. at least in the include directory
18:33:31seletz_back and looking at patch
18:34:37prpplagueseletz_: i read your statements regarding redboot partitioning
18:37:22seletz_prpplague: sorry, talket to you while you are at lunch, bad behavior :)
18:38:05prpplaguenp
18:38:16seletz_erikm: well. strange. what was the reason for this dependency? problem between the ears or what?
18:38:56erikmseletz_: there is some flash memory in that device
18:39:20prpplagueseletz_ / erikm : here is the pdf on the elf which has a section on page 29 about the flash partition they use - http://www.abcsinc.com/small-linux/elf.pdf
18:39:43erikmseletz_: but that still doesn't justify the use of MTD in a PCMCIA driver, of course
18:39:46erikmwgets
18:41:43prpplaguehmm, that patch appears to break several of the omnimeter/elf pcmcia items
18:43:02erikmprpplague: structure looks nice, but unfortunately they put it at the start of a flash block
18:43:30erikmprpplague: which means that you can't point linux to and tell "this is a jffs2 filesystem"
18:43:46seletz_prpplague: hmm. _very_ lowlevel. Identifying contents of blocks from assembly language mnemonics ...
18:44:39prpplaguewhat can i say, i'm a K.I.S.S. kinda guy...
18:45:08erikmprpplague: well, they could have made it a lot simpler
18:46:40seletz_erik: what about tehe structs i mentioned? Are that those structs to be changed?
18:46:56erikmseletz_: sorry, let me have a look
18:47:04erikmseletz_: could you point me to the correct file?
18:47:22prpplagueerikm: by putting it at the end, so that programs would not have to be linked oddly, and by cutting it down to name,size,type of data,cksum?
18:47:37seletz_erikm: hmm. give me a moment, have to web-cvs it ...
18:48:06erikmprpplague: by putting it in a separate flash block
18:48:19erikmprpplague: so it doesn't clutter with the contents of the partitions
18:49:05erikmseletz_: are you talking about the linux kernel, or about bootldr CVS?
18:49:38seletz_erikm: linux kernel. there is somewhere a partition map hardcoded for each system.
18:50:37seletz_erikm: hmmm. sf doesnt have web-cvs for linux kernel ??? searching ....
18:50:52erikmseletz_: lxr.linux.no
18:51:07erikmseletz_: do you know the name of the structure?
18:51:36erikmseletz_: cause in that case lxrbot can tell you which file to use :)
18:51:56erikmlxrbot: whereis kmalloc?
18:51:58lxrbotkmalloc is Defined as a function in: mm/slab.c, line 1534 Defined as a function prototype in: include/linux/slab.h, line 60 include/asm-i386/pgalloc.h, line 23 arch/ia64/sn/sn1/synergy.c, line 30 Defined as a preprocessor macro in: include/asm-s390/debug.h, line 259 include/asm-s390x/debug.h, line 259 Referenced (in 1038 files total) in: fs/read_write.c, line 230 fs/super.c, 5 times fs/pipe.c, line 451 fs/select.c, 2 times fs/proc/kcore.c, 2 times fs/proc
18:52:01erikmsee?
18:52:01lxrboti don't know, erikm
18:52:15erikmlxrbot: yeah, good. nice bot. have a botsnack
18:52:16lxrbotthanks erikm :)
18:52:26seletz_erikm: ??? no, sorry. im not that geek yet ... :)
18:52:44erikmlxrbot: is a bot I run on #kernelnewbies
18:52:49seletz_stunned :)
18:53:06erikmit's a frontend for lxr.linux.no
18:53:30seletz_lxrbot: whereis partition_map
18:53:31lxrbotpartition_map is Not used
18:54:09erikmJALH cleaned up the perl code, I run it on my desktop
18:54:39seletz_cursing "why dont have the source here f*** s***"
18:55:11erikmhands seletz_ ssh so he can connect to a linux box
18:56:50seletz_erikm: wellllll yes. I cant as of now. i currently only have a office at one of my customers. soon i'll have a office for me alone & a linux box on dls all day&night long. ARGH.
18:58:11seletz_downloading source tarballs
19:00:47seletz_UARGH. de.kernel.org offers 4.5 kb/s to me .... OK. We'll do it tomorrow, ok?
19:00:53erikmok
19:02:41seletz_erikm: well, thinking of it i _could_ bypass my not-so-gifted sysadmin and change some routes & see if i could get ssh up & running tomorrow.
19:02:57erikmhehe
19:03:25seletz_erikm: yes. that's exactly what i'll do tomorrow. This sort of things make me angry.
19:03:34erikmseletz_: like the not-so-brilliant sysadmins in the marriot hotel in new york :)
19:04:04erikmseletz_: pop and smtp were free, all other ports were firewalled unless you paid
19:04:32seletz_erikm: :^D ipfwadm redirect
19:04:54erikmseletz_: except that they forgot to firewall port 22.... so I tunneled everything over ssh to my desktop at the university :)
19:05:13seletz_erikm: sweet lovely VPN, eh?
19:05:18seletz_:)
19:05:31erikmyeah
19:06:22erikmand what they also didn't know is that we actually connected *two* machines to the same ethernet port :)
19:06:32seletz_erikm: to my shame i must admit that i never used ssh like that, got to try it out ... never had the need, tough.
19:06:34erikmmasquerading is nice :)
19:07:20seletz_erikm: yes, nice. Deutsche Telekom also does not know of my boxen running behind one DSL cable ...
19:08:08erikmI usually try to tunnel most things over ssh to my desktop. pop, smtp, irc, http (to the proxy on my desktop)
19:08:28seletz_erikm: thats a nice thing about those macs: using an airport wireles lan is just easy and works :) sitting in my flat besides my girfriend and chatting & surfing all wireless & mobile ...
19:08:42erikm*nod*
19:08:59erikmI use a lucent orinoco card with my laptop
19:09:06seletz_erikm: why do you tunnel? security?
19:09:13erikmyes
19:09:31erikmbut also to circumvent stupid firewalls
19:09:49erikm(which was necessary at IBM watson)
19:09:50seletz_erikm: hmmm. your'e at a university, right? no big chef watching, yes?
19:10:31erikmseletz_: no. it pays less, but I get a lot more freedom
19:11:37seletz_erikm: btw, what do you need in nl to get an PhD? is an MSc needed?
19:12:33erikmofficially nothing is needed. but an MSc surely helps
19:13:38seletz_erikm: hmmm. in germany quite so. "oficially" but practically when you dont have one you have to start over at an univerity.
19:14:34seletz_erikm: I "only" got an "Dipl.-Ing.(FH)" The (FH) is the problem.
19:15:33seletz_erikm: means 9 semester computer science at an university compared to 12.
19:15:42erikmwe have the same over here: FH == TH.
19:16:40seletz_erikm: so, do you get an PhD with an "TH" degree? (after additional studying, o'course)
19:16:41prpplaguewishes he could have gone to school for cs
19:17:33seletz_prpplague: well, i actually _learned_ not mutch about cs. i learned everything "on the job".
19:18:02seletz_prpplague: learned to learn, so to say.
19:18:28erikmseletz_: same for me. I did electrical engineering, but that's a lot of CS nowadays as well
19:20:17prpplaguei do ok, i'm head of r&d and make about $75k, but i wonder how much better i could be with some more training
19:22:26seletz_prpplague: You're in USA, are you?
19:22:49prpplagueyes
19:23:14seletz_prpplague: $75k, how mutch tax rate do you pay on average in USA?
19:23:33erikmseletz_: a *lot* less than we do over here
19:23:47prpplaguei pay about 35%
19:24:21erikmseletz_: but OTOH, everyday is twice as expensive, and you don't have the social security we have over here in europe
19:24:22seletz_erikm: hmmm. 36%, well, that comes close to my almost 50% :/
19:25:12seletz_erikm: i always thought its cheaper to life in USA. Food, accomodation and such.
19:25:22Russ|werkjust add 7% state sales tax and it does
19:25:28prpplagueonly in certain areas
19:25:44erikmseletz_: certainly not. what does 1 liter of milk cost in .de?
19:25:45prpplaguecheaper to live that is
19:25:45Russ|werkits pretty cheap living in phoenix
19:26:03prpplaguei'm in east texas and its dirt cheap
19:26:15Russ|werkapartment for $800, condo for $1200, good home for 120k
19:26:19seletz_well, we pay 16% sales tax, and milk is DM1 to DM 2
19:26:39Russ|werkand gas for $1.30/gal
19:26:42erikmseletz_: ok, double that and you have the US price
19:26:52Russ|werkhow much is a DM?
19:26:54prpplagueRuss: i bought a brand new 4bedroom/2bath brick on a 1 acre of land for $62k
19:27:07seletz_erikm: ugh. DM rougly 1/2 $
19:27:26Russ|werksounds about right for a liter of milk
19:27:59seletz_hmmmmm. $62k for a home with 4 bedrooms? How far away to the next town?
19:28:36Russ|werksan jose though....
19:29:12prpplagueseletz_: about 15 miles
19:29:19erikmRuss|werk: or NYC area...
19:30:53seletz_well, well. as an indep software developer i earn rougly the same, but pay much more tax, and pay about $1000 for a flat with 120m2 in a decent little town. 54% tax is the maximum in de. (insurance, sales tax etc. not included)
19:32:44seletz_what about the NYC area? I suppose thats a different story altogether?
19:33:34erikmit's quite expensive as well
19:34:44seletz_oh my, one "there is no such thing as a free lunch" i keep saying to myself .... :^)
19:34:58seletz_s/one//g
19:36:14seletz_well, I've to get some sleep now (and a quick game of GT3 :)
19:36:25seletz_bye all!
19:36:31erikmbye seletz
19:36:37prpplaguelater
19:42:51erikmeeeek, somebody got so insane as to put both SDRAM and EDORAM in the same system
19:43:16Russ|werkhow's that working for them?
19:43:27erikmit isn't :)
19:43:43erikmthe SA1110 supports it in theory
19:44:30ed__you know as shitty as rambus is...having an interface would make building these boards alot easier
19:44:57erikmbut it's a lot of juggling to get the memory timings right
19:46:11erikmwonders what the correct register jargon was for rambus
19:46:33ed__register?
19:46:35ed__theregister?
19:46:40erikmthat one, yes
19:46:43ed__ahh gotcha
19:46:47ed__ummm i don't remember
19:46:51ed__i'm sure they called them lots of things
19:46:54erikmit wasn't dramurai
19:47:09ed__dramuarai is everyone else (asian chip makers)
19:47:28erikm*nod*
19:48:26ed__it'd be cool to have the main bus be firewire
19:53:41erikmfound it: rambust
19:58:38ed__hahah
19:59:08erikmhttp://www.theregister.co.uk/content/28/16586.html
19:59:55ed__awesome
20:29:04BZFlagerikm: looks like I'm getting a paying contract I'll use blob for. less that a month or so till it starts
20:29:23BZFlaga sa based handheld with integrated fold out keyboard.
20:29:36erikmah, cool!
20:29:49prpplaguethe sharp?
20:30:00BZFlagthey want tftp updates, so I'll start hacking on that. no, not sharp.
20:30:37BZFlagwe hope to show/announce it at CES in January.
20:30:44prpplagueahh
20:30:48erikmBZFlag: I think we first want parameter block support
20:30:55Russ|werkBZFlag: I would start from scratch on tftp
20:31:29Russ|werkerikm: the base is there
20:32:30BZFlagI think I agree. I'm just excited to actually have a paying customer. ;-)
20:32:47Russ|werkpolicy just needs to be hammered out
20:32:58erikmRuss|werk: policy like what?
20:33:33erikmRuss|werk: I'm trying to clean up as much as possible, and this weekend's hack session did quite some good
20:34:52BZFlagcool. well I'll keep you posted. should be signed next week.
20:34:57erikmok
20:35:09erikmbtw, I put some features on the blob sourceforge page
20:35:26erikmfeature requests, that is
20:54:01erikmgoes zzz
20:54:02erikmbye
20:57:50prpplagueRuss|werk: looks like the problem was in my uclibc build
20:58:39Russ|werkhuh
20:59:17prpplagueRuss|werk: i guess steped on something while i was playing around
20:59:22Russ|werkin what way?
21:02:26prpplaguei've got the libs from the tuxscreen rootfs and the libs i built from source, i can only image i got something overwriten
21:05:57prpplagueRuss|werk: now that i've got the file over on the unit it won't execute
21:06:04prpplaguehmm
21:06:32Russ|werkcould be is packed full of x86'isms
21:06:55Russ|werkalignment issues, etc
21:07:41prpplagueRuss|werk: i'm sure i've done some wrote, this is hodge-podge code anyway
21:08:49Russ|werkwhat does it fil with?
21:08:53Russ|werker, fail with
21:10:24prpplaguedump_cis: file or directory not found
21:11:16prpplaguedump_cis: no such file or directory    
21:11:40prpplagueits executable and i'm starting with #./dump_cis
21:12:04Russ|werkoh, it can't find ld
21:12:12Russ|werkstrings dump_cis | head
21:14:59prpplagueRuss|werk: did that, what am i looking for?
21:18:03Russ|werkthe ld-uclibc.so or whatever
21:20:59prpplaguenone listed
22:05:15prpplagueRuss|werk: seems i can compile this program against a fresh build of uclibc but not the libs on the rootfs of tuxscreen
22:05:47Russ|werkwhat do you mean by against the libs on the rootfs of the tuxscreen
22:06:04Russ|werkand do you have the string /lib/ld-uClibc.so.0 is the compiled file
22:10:02prpplagueyes, the /lib/ld-uClibc.so.0 is in the string, i downloaded the rootfs from tuxscreen.net and tried to use the libs from there so that i'd have common libs
22:11:04Russ|werkwhat do you mean by "use"
22:12:41prpplagueRuss|werk: i place the tuxscreen rootfs libs in the dir where the uClibc would link to instead of the one i built from source
22:13:13Russ|werkprobably not too wise...
22:13:31Russ|werkdoes the /lib on your tuxscreen contain /lib/ld-uClibc.so.0?
22:13:32prpplaguedidn't think so, i was just experimenting
22:13:45prpplagueyes
22:25:53Russ|werkhmmm
22:26:04Russ|werkI would do it the other way around
22:27:20prpplagueuse the fresh build on the rootfs?
22:28:22Russ|werkget the libs you are building with into the tuxscreen
22:29:38prpplaguethat means i'll have to re-compile busybox,tinylogin,etc.

Generated by irclog2html.pl by Jeff Waugh - find it at freshmeat.net! Modified by Tim Riker to work with infobot logs, split per channel and by date, etc.