01:02.13 | *** join/#htc-linux XirXes (~XirXes@67-2-31-70.slkc.qwest.net) |
01:09.39 | *** join/#htc-linux sleepee (~sleepee@wifirev.centennialpr.net) |
02:26.26 | *** join/#htc-linux AstainHellbring (AstainHell@unaffiliated/astainhellbring) |
03:48.21 | *** join/#htc-linux MacDrunk (~marper@201.165.163.183) |
03:50.48 | *** part/#htc-linux MacDrunk (~marper@201.165.163.183) |
05:30.29 | *** join/#htc-linux eR^zeRa` (~zzeratul@ip-84-42-202-42.net.upcbroadband.cz) |
05:50.01 | *** join/#htc-linux sdub (~sdub@Aircrack-NG/Friend) |
06:52.42 | *** join/#htc-linux kiozen (~kiozen@p578a42db.dip0.t-ipconnect.de) |
07:29.34 | *** join/#htc-linux lamikr (lamikr@nat/nokia/x-jfziiivsrzamlwqm) |
08:47.10 | *** join/#htc-linux DuperMa (~Duper@93-172-145-8.bb.netvision.net.il) |
08:56.54 | *** join/#htc-linux marc1706 (~Marc@phpbb/modifications/marc1706) |
10:18.27 | *** join/#htc-linux marmotta (~skodde@unaffiliated/skodde) |
10:31.50 | *** join/#htc-linux LargePrime (~LargePrim@173-120-174-120.pools.spcsdns.net) |
11:04.49 | *** join/#htc-linux LargePrime (~LargePrim@173-120-174-120.pools.spcsdns.net) |
11:10.15 | *** join/#htc-linux LargePrime_ (~LargePrim@173-120-174-120.pools.spcsdns.net) |
11:54.21 | *** join/#htc-linux zeusk_ (77eb360e@gateway/web/freenode/ip.119.235.54.14) |
12:13.22 | *** join/#htc-linux helicopter88 (~helicopte@host216-7-dynamic.30-79-r.retail.telecomitalia.it) |
12:14.23 | *** join/#htc-linux XirXes (~XirXes@67-2-31-70.slkc.qwest.net) |
12:15.04 | *** join/#htc-linux gauner1986 (~Adium@87.253.171.205) |
13:02.34 | *** join/#htc-linux conantroutman (~conantrou@cpc5-pert5-2-0-cust768.16-4.cable.virginmedia.com) |
13:03.11 | *** join/#htc-linux Cotulla (~myfakemai@109.205.253.11) |
13:11.27 | gauner1986 | ah |
13:11.30 | gauner1986 | the great cotulla |
13:13.30 | marc1706 | perfect timing :D |
13:13.33 | marc1706 | hi Cotulla |
13:13.58 | Cotulla | hi |
13:14.31 | marc1706 | quick question: why is the baudrate of the smd tty driver set to 115200 while the newest code at CAF is using 38400? |
13:14.55 | Cotulla | because it's ignored??? |
13:15.09 | Cotulla | there is no data rate limits |
13:15.14 | Cotulla | it's always works with maximum speed |
13:15.20 | marc1706 | ah ok, so the set baudrate is just bogus? |
13:15.23 | Cotulla | yes |
13:15.26 | marc1706 | ok |
13:15.28 | Cotulla | comes from real COM ports |
13:15.49 | marc1706 | I thought you might know :P |
13:16.11 | zeusk_ | that rate is there probably just to satisfy linux tty driver specs |
13:16.19 | Cotulla | yes for that too |
13:17.16 | marc1706 | btw, kexec told me bye and then nothing happened xD |
13:17.27 | marc1706 | probably loading it to the wrong address |
13:17.43 | marc1706 | the documentation is really crappy |
13:17.50 | zeusk_ | Yesterday, I was writing timer code and i looked at kernel/timer.c and i got cancer... They call ticks, "jiffies", have some wierd time vector buckets for marking ticks with a registered timer and what not.. complicated something so easy.. |
13:18.43 | Cotulla | shit do u have enough money for good hearse? |
13:19.17 | Cotulla | :D yes it sucks I believe looked to it some time ago |
13:20.20 | zeusk_ | I just resorted to using unsorted list timers, register:O(1) unregister:O(1) execute:O(N) |
13:20.44 | zeusk_ | and it's less than 100 lines of code.. |
13:22.07 | marc1706 | wasn't there something about kexec zimage offset being too small? |
13:24.41 | zeusk_ | marc, any specific reason for kexec ? |
13:25.01 | zeusk_ | usually it's only for those poor guys with secure boot |
13:25.19 | zeusk_ | or the 1337 ninja kernel hackers |
13:25.29 | marc1706 | well I was trying if we maybe can get the .38 kernel to boot with kexec |
13:26.05 | marc1706 | stops after handing off to the new kernel though ... |
13:26.06 | zeusk_ | Hmm |
13:26.16 | marc1706 | I was just making sure that the kexec is not the cause |
13:26.30 | zeusk_ | lol, cotulla, we need you |
13:27.02 | Cotulla | I am not "1337 ninja kernel hackers" |
13:27.24 | Cotulla | I am number 1338 I think |
13:27.40 | marc1706 | you can try to get it to run the way you want |
13:27.41 | Cotulla | cuz I joined party too late :( |
13:27.46 | marc1706 | I'll grant you that :P |
13:27.47 | Cotulla | run what? |
13:28.03 | zeusk_ | linux 2.6.38+ on leo |
13:28.10 | zeusk_ | doesn't run and we can't find why |
13:28.35 | zeusk_ | maybe, just maybe, another stupid thing like _irqsave ? LOL |
13:30.16 | Cotulla | hm |
13:30.26 | Cotulla | maybe |
13:30.56 | Cotulla | Markinus ported some new kernel afaik . he started work at least |
13:31.09 | marc1706 | that was 2.6.35 |
13:31.13 | marc1706 | and it works |
13:31.16 | marc1706 | at least on magldr |
13:31.26 | Cotulla | so why not take 3.1.X |
13:31.35 | marc1706 | for some not so obvious reasons it just randomly reboots on cLK :D |
13:32.02 | marc1706 | don't have a problem with 3.1.x but I think porting 2.6.38 first would be a bit easier |
13:32.16 | Cotulla | no |
13:32.19 | Cotulla | u are wrong |
13:32.25 | marc1706 | I am? hm |
13:32.44 | Cotulla | porting any linux kernel version to another == AAA FUCK IT'S EATING MY LEEDD |
13:32.48 | zeusk_ | 3.0.x should be better otherwise 3.4, 3.1 was never adopted by android mainline or caf |
13:33.04 | marc1706 | well 3.4 lacks a lot of stuff on caf |
13:33.09 | marc1706 | doesn't even contain msm_rmnet O.o |
13:33.12 | Cotulla | htc released any kernel sources with 3.XX? |
13:33.19 | zeusk_ | Yes, many |
13:33.24 | zeusk_ | even for 7227A |
13:33.26 | Cotulla | so better take them |
13:35.57 | marc1706 | htc kernels are all at 3.0.16 as far as I can see |
13:36.17 | Cotulla | so take that port and should work |
13:36.32 | Cotulla | maybe copy old video |
13:37.01 | marc1706 | I think we need to copy video drivers from our currenty kernels, at least in parts |
13:37.07 | marc1706 | htc has hacked up gpu drivers |
13:37.08 | marc1706 | for sense |
13:37.13 | zeusk_ | there was a wierd mention of kernel domain bug in 8x50 inside some linux mmu pte file, something about domain 0 being faulty in older revision |
13:37.14 | Cotulla | how |
13:37.16 | Cotulla | they can't |
13:37.30 | Cotulla | 3D is prebuilt from QCT |
13:37.39 | Cotulla | zeusk just a hardware bug |
13:37.43 | Cotulla | usual thing |
13:37.46 | marc1706 | Cotulla: wait I think it's video drivers |
13:37.46 | marc1706 | sry |
13:37.48 | Cotulla | needs workaround |
13:37.59 | marc1706 | mdp is slightly different |
13:38.09 | marc1706 | for some special htc reasons |
13:38.36 | marc1706 | one v or sensation kernel? |
13:38.46 | zeusk_ | sense crap in kernel is unlikely though not impossible, reminds how bravo port of 2.6.35 suddenly fixed the glitches in sense 2.1 |
13:38.55 | zeusk_ | one V should be better |
13:39.03 | zeusk_ | sensation is another generation ahead |
13:39.05 | marc1706 | yes, but it doesn't work on cm7 ... |
13:39.22 | marc1706 | that's what I meant when I talked about hacked up video code |
13:39.22 | Cotulla | u need just track things |
13:39.30 | Cotulla | not look from "blackbox" position |
13:39.33 | Cotulla | :) |
13:39.41 | zeusk_ | marc, then take a CM kernel for one V ? |
13:39.42 | marc1706 | there is also desire c kernel code |
13:39.58 | marc1706 | zeusk_: I think the video drivers are the last of our problems ;) |
13:40.20 | zeusk_ | I wonder why QCT offers android 4+ support for 7227A but not 8x50 |
13:40.50 | marc1706 | they provided gpu libs for 8x50? |
13:40.50 | zeusk_ | only reason i know is higher AXI clock on 7227 with smaller die size |
13:40.57 | marc1706 | or was that general arm7? |
13:41.04 | marc1706 | adreno200? |
13:41.06 | zeusk_ | Yeah, just the adreno libs for a2xx |
13:41.10 | Cotulla | 8x50 is too old for them |
13:41.19 | Cotulla | 7227A is rather new |
13:41.24 | marc1706 | they still sell it though :P |
13:41.34 | Cotulla | u think so? |
13:41.46 | marc1706 | at least they still advertise it ... |
13:41.49 | Cotulla | it's EOL already long time |
13:42.18 | marc1706 | sry, confused that with snapdragon s1 |
13:42.27 | Cotulla | but driver should be suitable from 7227A |
13:42.29 | Cotulla | in general |
13:42.51 | marc1706 | desire c is msm7225A and snapdragon s1, also with 3.0.16 kernel |
13:42.59 | marc1706 | and I think adreno 200 |
13:43.25 | Cotulla | Yamato |
13:43.25 | marc1706 | one v has adreno 205 |
13:43.50 | marc1706 | yamata is the firmware name? |
13:43.54 | marc1706 | yamato* |
13:44.13 | marc1706 | yes, desire c has adreno 200 |
13:44.17 | zeusk_ | codename :P |
13:44.21 | marc1706 | ah ^^ |
13:44.31 | *** join/#htc-linux LargePrime (~LargePrim@173-120-174-120.pools.spcsdns.net) |
13:44.38 | marc1706 | desire c might be better suited? what do you think? |
13:44.59 | Cotulla | for what? |
13:45.05 | Cotulla | for GLES? |
13:45.07 | marc1706 | for porting to leo |
13:45.10 | marc1706 | yes |
13:45.16 | Cotulla | they should be same |
13:45.26 | Cotulla | cuz there is no working 8x50 code |
13:45.36 | Cotulla | and it's rather different from other directions |
13:45.52 | marc1706 | well one v code is newer |
13:45.54 | zeusk_ | only problem will be forward porting qdsp_1550 |
13:46.00 | marc1706 | yup |
13:46.21 | zeusk_ | linux will break your code like shit |
13:46.52 | zeusk_ | they even introduced wierd stuff like clk_prepare and crap in 3.x series afaik |
13:47.46 | marc1706 | qdsp_1550 seemed to compile on 3.0.8 IIRC |
13:48.02 | marc1706 | only needs changing from .ioctl to .unlocked_ioctl |
13:48.38 | marc1706 | doesn't mean it works but still |
13:49.17 | Cotulla | yeah |
13:49.22 | Cotulla | whu unlockrf? |
13:49.28 | Cotulla | why unlocked? |
13:49.38 | marc1706 | .ioctl seems to be deprecated |
13:49.53 | Cotulla | so u need mutexes AROUND? |
13:50.08 | marc1706 | not as far as I know |
13:50.20 | marc1706 | unlocked_ioctl also existed in 2.6.35 |
13:50.28 | marc1706 | I changed it in one or 2 files |
13:50.32 | marc1706 | no change in behavior O.o |
13:50.38 | zeusk_ | marc, i think he is asking what that "unlocked_" means, does it mean you need locks to protect ioctls or what |
13:50.48 | marc1706 | http://lwn.net/Articles/119652/ |
13:50.55 | Cotulla | yes |
13:50.58 | Cotulla | it is |
13:51.09 | zeusk_ | Ahh, linux.. the new way of everything.. |
13:51.14 | Cotulla | :D |
13:52.24 | marc1706 | funny that it exists since 2005 xD |
13:53.16 | marc1706 | wow, htc now provides an awesome readme on how to compile the kernel source |
13:53.32 | marc1706 | it's like 10 lines ... |
13:53.37 | marc1706 | thank you thc |
13:53.39 | marc1706 | lol |
13:53.41 | marc1706 | I fail |
13:54.01 | zeusk_ | what is it ? gunzip tar, cd, make config, make zimage ? |
13:54.20 | marc1706 | no, basically get the toolchain from google |
13:54.27 | marc1706 | edit a file and add the toolchain path |
13:54.37 | marc1706 | make primou_defconfig |
13:54.38 | marc1706 | and make |
13:54.42 | marc1706 | and then where the files are |
13:55.22 | marc1706 | --Files path |
13:55.22 | marc1706 | arch/arm/boot/zImage |
13:55.22 | marc1706 | drivers/*/*.ko |
13:55.56 | Cotulla | I will need to dedicate a Linux day for porting :) |
13:56.01 | marc1706 | utf8 fail: NOTE: the tool ¡¥git¡¦ [...] |
13:56.09 | marc1706 | what is a linux day? |
13:57.42 | marc1706 | wow, why would you put the file primou-panel.c in a seperate folder in mach-msm? |
13:58.07 | Cotulla | ? |
13:58.38 | marc1706 | they put the file primou-panel.c in arch/arm/mach-msm/display/ |
13:59.03 | zeusk_ | behold my 80 line timer code, http://pastebin.com/3iELzr7D |
13:59.55 | marc1706 | what's the diff between list_for_each_entry_safe() and list_for_every_entry_safe()? |
13:59.57 | Cotulla | overflow is handled? |
14:00.17 | zeusk_ | cotulla, no, not yet |
14:00.24 | Cotulla | :P |
14:00.57 | zeusk_ | marc, inside safe i can delete the node while inside normal function if i delete list node inside the loop, it should cause null dereference on next cycle |
14:01.01 | zeusk_ | it is double linked list |
14:01.42 | marc1706 | I get the reason for save, but what's the diff between for_every and for_each? |
14:01.48 | marc1706 | sounds very similar to me ^^ |
14:01.58 | zeusk_ | nothing... :P |
14:02.08 | zeusk_ | it just iterates over timer objects in my list |
14:02.08 | marc1706 | aha ^^ |
14:20.17 | zeusk_ | cotulla, at 100Hz, it should take 497 days to overflow, i think that should be fine :) |
14:20.26 | Cotulla | lol |
14:20.30 | marc1706 | no :P |
14:20.36 | Cotulla | take 64 bit maybe |
14:20.57 | zeusk_ | since timers are managed at internal tick count of 100Hz, I can safely reset HW timer without caring about stupid mdelay like in lk |
14:21.18 | Cotulla | 100 hz? |
14:22.59 | zeusk_ | Board should call sys_tick at 100Hz, not at hw timer freq |
14:24.10 | zeusk_ | then of course it can implement high resolution timers using HW timer count |
14:24.41 | zeusk_ | which is very unlikely to overflow in the small timegap of a high resolution timer |
14:31.49 | Cotulla | but 100 hz can be impossible to reach |
14:31.53 | Cotulla | so how it's RTOS |
14:34.32 | zeusk_ | impossible ? |
15:13.58 | *** join/#htc-linux conantroutman (~conantrou@cpc5-pert5-2-0-cust768.16-4.cable.virginmedia.com) |
15:15.02 | *** join/#htc-linux LargePrime (~LargePrim@173-120-174-120.pools.spcsdns.net) |
15:45.24 | *** join/#htc-linux LargePrime (~LargePrim@173-120-174-120.pools.spcsdns.net) |
15:51.56 | *** join/#htc-linux leviathan (~quassel@2001:470:26:484:6ef0:49ff:fee6:8dca) |
15:52.56 | *** part/#htc-linux Cotulla (~myfakemai@109.205.253.11) |
16:14.52 | *** join/#htc-linux daviathan (~quassel@2001:470:26:484:6ef0:49ff:fee6:8dca) |
16:18.16 | *** join/#htc-linux LargePrime (~LargePrim@173-120-174-120.pools.spcsdns.net) |
16:29.25 | *** join/#htc-linux LargePrime (~LargePrim@173-120-174-120.pools.spcsdns.net) |
16:29.52 | *** join/#htc-linux rob_w (~rob_w@unaffiliated/rob-w/x-1112029) |
16:36.18 | zeusk_ | marc, you working on htc's 3.x kernel then ? I'm interested in it too |
16:36.24 | zeusk_ | marc1706 |
16:51.06 | marc1706 | zeusk_: yes I'll start working on it, probably end of the week |
16:53.55 | *** join/#htc-linux mes (~mes@sentry.lazo.ca) |
17:12.04 | *** join/#htc-linux rossimo (~Ross@75-138-55-162.static.gnvl.sc.charter.com) |
17:12.15 | rossimo | hey all, do htc android phones use ALSA? |
17:14.49 | helicopter88 | no |
17:15.59 | rossimo | so they've written thier own audio interface software? |
17:21.41 | *** join/#htc-linux gauner1986 (~Adium@hmbg-4d06c223.pool.mediaWays.net) |
17:36.30 | *** join/#htc-linux paulk-desktop (~paulk@lib33-1-82-233-88-171.fbx.proxad.net) |
17:43.25 | *** join/#htc-linux detule (~detule@unaffiliated/d3tul3) |
17:45.18 | helicopter88 | is away: Busy as fuck |
18:06.50 | *** join/#htc-linux LargePrime_ (~LargePrim@173-148-126-5.pools.spcsdns.net) |
18:39.17 | *** join/#htc-linux gauner1986 (~Adium@hmbg-4d06c223.pool.mediaWays.net) |
18:47.31 | *** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info) |
18:54.05 | *** join/#htc-linux kiozen (~kiozen@46.244.203.191) |
18:58.35 | *** join/#htc-linux BabelO (~fcr@unaffiliated/babelo) |
19:48.19 | *** join/#htc-linux DuperMan (~Duper@93-172-145-8.bb.netvision.net.il) |
19:56.32 | *** join/#htc-linux LargePrime (~LargePrim@173-111-209-237.pools.spcsdns.net) |
20:21.45 | *** join/#htc-linux gauner1986 (~Adium@hmbg-4d06c223.pool.mediaWays.net) |
20:34.02 | *** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info) |
20:57.08 | *** join/#htc-linux LargePrime_ (~LargePrim@173-99-210-73.pools.spcsdns.net) |
21:08.09 | *** join/#htc-linux LolekTD (b22a9c07@gateway/web/freenode/ip.178.42.156.7) |
21:59.03 | *** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info) |
22:38.18 | *** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info) |
22:41.39 | *** join/#htc-linux Bry8Star (~Bry8Star@gateway/tor-sasl/bry8star) |