#qi-hardware IRC log for Friday, 2012-09-14

DocScrutinizer05so, how does "qualcomm"  (http://maps.google.de/maps?oe=UTF-8&um=1&ie=UTF-8&q=qualcomm&fb=1&gl=de&hq=qualcomm&hnear=0x479f57aeb5b61cd3:0xdd5daf85a98c21b7,N%C3%BCrnberg&cid=0,0,7383782245350231276&sa=X&ei=tI9SUIrZMMel4gSvtIE4&ved=0CJ8CEPwSMAA) sound to you, for my new job?02:01
DocScrutinizer05wpwrak: how's your hayfever?02:04
wolfspraulDocScrutinizer05: you work for qualcomm now?02:22
DocScrutinizer05maybe next one, yes02:22
wolfspraulhope this works out for you02:22
DocScrutinizer05that's why I ask ;-)02:23
DocScrutinizer05I'm not sure I like qualcomm02:23
DocScrutinizer05otoh who says you need to like your job, or even your employer**02:24
DocScrutinizer05(** double deref)02:24
DocScrutinizer05funny detail: it's so close to STE http://maps.google.de/maps?num=30&hl=de&newwindow=1&safe=off&um=1&ie=UTF-8&q=st-ericsson&fb=1&gl=de&hq=st-ericsson&hnear=0x479f57aeb5b61cd3:0xdd5daf85a98c21b7,N%C3%BCrnberg&cid=0,0,7050262622794191287&sa=X&ei=uZVSUK-aHqvZ4QSynICgAQ&ved=0CLkCEPwSMAM that I probably will meet my ex-colleagues in restaurant during lunch02:26
wolfspraulI agree that you don't need to like your job or employer, but... at the very least there does need to be cultural 'compatibility', whatever that means :-)02:28
wolfspraulif the company culture is a bunch of assholes raping the world, then you better have that same culture, and you will do well. equally for any other culture.02:28
DocScrutinizer05actually that's another reason i'm not entirely happy with the perspective: there are realy few shops and restaurants in this whole area02:28
wolfspraulI don't know the qualcomm company culture, but culture compatibility is something you should check, for everybodys happiness...02:29
wolfspraulthe few qc people I met were all very professional and serious, I liked them02:29
DocScrutinizer05aiui my potentially new job would be more hw-centric :-D02:30
wolfspraulqc has also recently been restructuring their open source efforts02:30
wolfsprauldon't know the thinking behind that though02:31
DocScrutinizer05ooh, they have a thing like opens source efforts?02:31
wolfsprauldefinitely, and they are restructuring it02:31
wolfspraulI *guess* they want to separate a little better from their IP businesses so that they can protect that IP better, and at the same time can invest in anything 'open' better02:32
wolfspraulbut you need to talk with them, maybe you are soon one of them :-)02:32
wolfspraulthen you can explain to us!02:32
DocScrutinizer05thanks for the tips!02:35
wolfspraulgotta run too02:37
wolfspraulhow can you be up at 4.30 btw? :-)02:37
wolfspraulk l8...02:37
DocScrutinizer05I'm still vacant02:37
DocScrutinizer05my pimp threated me to call yesterday, to tell me about possible meeting at qualcomm on monday, but he didn't02:38
DocScrutinizer05so otoh he "forced" me to take my remaining holiday for this year, which ends today02:40
DocScrutinizer05rather, yeaterday02:40
wpwrakqualcomm ? heart of evil ? :)03:28
wpwrak(at the company level no idea how the people are)03:29
wpwrak(open effort) now that sounds like a departure from the past. if it's true, maybe they won't be that bad.03:30
wpwrakgrmbl. go out and protest for a few hours in spring, sneeze bucketfulls the rest of the night :-( at least this time we were a lot. no complaint there :)08:33
whitequarkwpwrak: sounds like the atheros acquisition wasn't that bad10:00
mthlarsc: the current battery driver will use a negative (error code) voltage value in some cases14:59
mththis diff is for the 4770 driver, but that's almost an exact copy of the 4740 driver14:59
mthcurrently anyway: it doesn't work yet, so I'll probably have to add some more differences14:59
larscmth: my plan is to add a IIO ADC driver for the jz4740 ADC and then eventually switch to the generic IIO battery driver (which currently is in development).15:28
larsca sane strategy would probably to do the same for the jz477015:29
kyakthe most dangerous command is "uptime". as soon as i run it, it doesn't take long till something happens to my server15:31
larscwhy does svn suck so much? Just found out that you can't set global ignores, but rather have to specify them for each directory...16:09
viriclarsc: svn works by directory :)16:58
larschm, that might be a point17:09
larscbut I still don't like it 17:09
Aylalarsc: ping17:12
Ayladid you remember about the INI library? :)17:13
larscbut I'm still at work17:13
viric[<bf2004f4>] (btrfs_num_copies+0xe8/0xf4 [btrfs]) from [<bf1bd89c>] 17:14
viricwhat is 0xe8 and 0xf4?17:14
viricone should be the offset since the function entry point17:15
larscthe other is the size of the function17:18
larscAyla: So far I could find out that the header file is called ini.h17:19
viricah 0xf4 is the size? nice.17:19
viricthank you17:19
larscAyla: but that's all I can find for the moment17:22
larscmore on monday. the person that's using it isn't in the office right now17:27
AylaI won't wait that long ;)17:56
larscmth: Can I mention to sjhill that you are working on jz4770 support?18:04
larschi sjhill 18:13
sjhilllarsc -> Lars-Peter Clausen18:13
sjhillmth -> Maarten ter Huurne18:13
sjhilli have not idea where Thierry and Antony are18:14
larscthierry acutally has nothing to do with jz47xx, he just brought over the pwm driver to the generic pwm framework since he is the pwm maintainer18:14
sjhillso i won't include him on future emails18:14
larscand to be honest i don't know who Antony is18:14
Aylayou're doing a PWM driver for jz4770?18:15
larscsjhill: is working on jz4760 support18:16
sjhilllarsc: how did you get a lepus board?18:18
larscwas a bit of a coincidence, I did a bit of contract work for somebody who wanted to sell a jz4760 based board18:20
sjhilli'll have to talk to my colleagues at MIPS and email my one contact at Ingenic18:21
sjhillmaybe i'll get lucky18:21
viricmh what's that kernel config option, to have a populated /dev mounted before running init?20:22
viric(when rootfs doesn't have it)20:22
sjhillviric: what version kernel?20:57
viricI looked for it in 3.5.320:58
viricI recall it's about DEVFS something20:58
viricbut F8 in nconfig does not show me any DEVFS thing20:58
viricdamn. It's inside device drivers, not filesystems.21:01
viricthx larsc 21:01
rjeffriesSince several people use KiCad -- http://dangerousprototypes.com/2012/09/14/dangerous-prototypes-kicad-eda-experience/21:04
viricI hate when I fail to get 'init' running.21:12
viricVFS: Mounted root (btrfs filesystem) readonly on device 0:11.21:12
viricdevtmpfs: mounted21:12
viricKernel panic - not syncing: Attempted to kill init! exitcode=0x0000000421:12
viric(it's a statically linked 'sh', init)21:13
viricsignal 4... SIGILL... illegal instruction. damn.21:13
mthhi sjhill22:38
mthsjhill: this is the JZ4770 device that Ayla and I are working on: http://game-consoles-worldwide.com/?page_id=1522:40
mthlarsc: Antony is doing an OpenDingux port for the Ritmix (Russian) game console (JZ4755)22:42
mthsjhill: Ayla has done a fair bit of kernel for for the 4740 as well, for example the watchdog driver was written by him22:44
mths/kernel for/kernel work/22:44
sjhillmth: cool22:58
sjhillSo, who's going to send me one of those game consoles? :)22:59
sjhillI'll buy one, if there are any left.23:01
Aylathere are23:01
Aylaif you're up to do some kernel work, you could ask for a prototype23:01
sjhillWho do I ask for a prototype?23:05
wpwrakhmm. i have small and fixed number of variables. from these, i want to determine the median value. i could sort the variables and pick the value in the middle, but that's quite inefficient. a more efficient approach is to directly search for the median value.23:45
wpwraki've solved this manually for three variables. doing it for five smells like messy work. i wonder if there's an algorithm for this somewhere ?23:46
wpwrakthis is a solution for three variables: http://pastebin.ca/220435523:47
mthis five enough or is that just the next step up from three?23:51
wpwrakit's fun to google for this. there's a thread in a vb basic forum. proves that vb basic is for people who still have a lot to learn ;-)23:51
wpwrakwhat the fortran folks do looks more interesting ... http://compgroups.net/comp.lang.fortran/find-the-median-value-of-an-array/32696223:52
wpwraknot sure yet. i very much hope it'll be enough. it has to run in a poor little microcontroller crawling at less than one MIPS23:53
mththe number of values is smaller than the range of the values?23:53
wpwrakyes, the range is much larger23:55
mthwhat you can do, is a limited type of quicksort23:57
mthevery quicksort iteration separates the array into a segment where all elements are smaller than X, a segment where all elements are equal to X and a segment where all elements are greater than X23:58
wpwrakhmm, this seems to be the place for the algorithms: http://en.wikipedia.org/wiki/Selection_algorithm23:58
wpwrakquicksort is way too expensive23:58
mthif the "equals" segment is in the middle, you're done, otherwise continue to work on the largest remaining segment (the one that contains the middle element)23:59
--- Sat Sep 15 201200:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!