Monday, 2011-06-27

*** GPFerror has joined #xdandroid00:12
*** madplaya has joined #xdandroid00:19
madplayajuh-00:19
*** madplaya has quit IRC00:48
*** arrrghhh_ has joined #xdandroid01:04
arrrghhhhyc, ping01:04
smiley-hyc: the raph100 is now on about 70 hours uptime with the smd-patch :)01:07
arrrghhhnice01:08
arrrghhhwell hyc it was re: the whole rev.0 and rev.a thing01:10
arrrghhhi'll try to find where it is in winmo to see if i am truly in a rev.a or 0 area.  Not sure WinMo even shows that info...01:11
arrrghhhand in reality the issue is probably cosmetic01:11
arrrghhhi doubt it would affect xfer speeds drastically.01:11
arrrghhhif at all for that matter :P01:12
*** arrrghhh has quit IRC01:18
hyci doubt it makes any difference01:49
*** rpierce99 has joined #xdandroid03:26
*** rpierce99 has quit IRC03:31
*** diego_kynetics has joined #xdandroid03:34
*** emwe has joined #xdandroid04:29
emwehyc: heyho. does this look like a smd tty alloc fix candidate? http://dl.dropbox.com/u/12183048/htctopaz/dot35/logs/rhod_ramconsole_last_smdtty_alloc.txt04:29
emwenever seen this ever on .35.04:29
emwewas with rhod... SOD (but sleeping, not wake). didn't like to wakeup no matter how long i pressed the buttons. neither did wake on ac-plug.04:30
*** sigma6 has joined #xdandroid04:31
smiley-emwe: I have over 70 hours uptime on my RAPH100 now with latest kernel and hyc smd tty pppd blahblah patch :)04:35
*** emwe has quit IRC04:39
*** emwe has joined #xdandroid05:25
emwesmiley-: 70 hours with one charge? i assume no :)05:26
smiley-hehe..  no :(05:26
smiley-probably 7-8 charges ;)05:26
smiley-but moving from wifi->3G->E alot05:27
*** Miga has quit IRC05:57
*** Miga has joined #xdandroid05:59
hycemwe: page allocation failure? it sure does look like it06:06
emweyah, looks like it.06:07
hycso you haven't applied that fix to .35 yet?06:07
emwenope. never had it06:08
emwenever seen this and i am "running" .35 for quite some time.06:08
hycsounds like it's time to do so ;)06:08
emwei saw no need yet06:08
hyceither way the original code is terrible06:08
emweis that patch applicable right away to autobuild .27? i think so yes, or?06:08
hycpretty sure yes06:09
emwei mean, it can go out already i take it?06:09
emweok06:09
hycnobody has reported any problems on kernels with that patch06:09
hycit has no downside at all, all win.06:09
emweyah, rather the other way around. problems without it.06:09
emweiirc never had it though with .27 either.06:10
emwebtw, you ever looked into fn+caps controlled by userland? wasn't there some htc-extension to lights service exactly for that?06:10
hycyou were pretty lucky then. It wasn't frequent for me, but often enough that I recognizedi t whnever it hit.06:10
hycI don't know of any htc extensions06:10
emweok. is aosp capable of doing so? iirc the api to liblights got extended to do so. pushing some more events...06:11
hycwould be happy to rig something up though06:11
emweneed to look again.06:11
hycI've seen nothing in froyo source that could do it06:11
hycmebbe gingerbread06:11
emweperhaps cyano has something already.06:12
hycI looked extensively in the froyo code06:12
hychm, didn't look there06:12
emwei wouldn't even know where to start. frameworks/base?06:12
hycprobably06:12
hycpowermanager service lives there I think06:12
hycand that's what invokes liblights06:12
emweah, thanks.06:12
emweprivate LightsService.Light mCapsLight;     private LightsService.Light mFnLight;06:15
emwe^^ on cyanogen that is06:16
hycoh ho06:16
hycfroyo?06:16
emwelooking at their ginger. but we don't have it. in either branch06:18
emweso they likeley extended like htc used to06:18
hycwill have to see what kernel interface they used too06:20
*** hardwalker has quit IRC06:25
*** sigma6 has quit IRC06:25
emwehm, no fn/caps in their hw-msm7k liblights but i'll take a look at it then. "just" a matter of adding to liblights when PowerManagerService is updated.06:28
hychm if not in their msm7k, then what phone did they implement it for?06:44
emwedon't they do their dist for a shitload of devices? perhaps some non-htc keyboard phone.06:45
emweor for evo-shift e.g. some samsung are with keyb iirc too06:46
hycah06:46
*** Detule has joined #xdandroid07:03
*** sigma6 has joined #xdandroid07:09
emwehyc: just to make sure... you were seeing the same stack traces before the smd/tty patch fixup? or do "mine" rather look like it can't allocate the memory requested? (asking as jonpry has seen some memory allocation pressure on .39 initially)07:46
hycI never saw a trace like this, but it matches the description jonpry gave07:58
hycBut I used to see other smd_tty panic traces frequently on my phone07:58
hycof course, this raises the question of whether there's a memory leak in .35 and/or .3907:59
hycI mean, thus far, the only kernel crashes I've had with .27 on my rhod have been triggered by manually enabling 2G-only mode.08:00
hycthe modem command to change modes causes the radio to disconnect and reconnect, and sometimes (but not every time) this interacts badly with smd and pppd ...08:01
hycnote that simply roaming between 2G and 3G has no such problem08:01
smiley-after the patch my raph is very stable (if it's the same patch we talk about)08:02
hycyep, same patch08:02
hycnote that there are two aspects to the patch - jonpry's original patch which fixes the memory buffer stupidity08:02
hycand my patch which adds tty hangup support to smd_tty08:02
hycI'm pretty sure my tty hangup patch fixed my 2G/3G toggling problem.08:03
smiley-cpuspy says 4h on 691MHz, 43h on 128MHz and 31h deep sleep :)08:03
hycfor me jonpry's patch was simply an opportunity to optimize the driver and increase the efficiency08:03
smiley-the reason for high number on 128MHz is because I have left it in the charger over night08:04
hyccool08:04
hycsounds like it's all healthy08:04
smiley-very :)08:04
hycwhat kind of battery life do you get from winmo? comparable, or very different?08:04
smiley-very different :(08:05
smiley-8h-10h at most08:05
hycreally. so winmo is worse?08:05
smiley-oh08:05
smiley-I mean in android08:05
smiley-on winmo..    I can't remember..   it's 2 years ago since I used it ;)08:06
hycheh ok08:06
smiley-but more than 8h-10h08:06
hycI've never tested battery life on mine, just installed xdandroid as soon as the phone arrived.08:06
smiley-this one has been on the shelf for 1.5-2 years08:06
smiley-but I'm running gingerbread on it..  not sure if that drains more battery08:07
hycwell, LiIon batteries only have a useful life of ~3 years before they're all oxidized08:07
hycso you might have gotten more than 10 hours 2 years ago08:07
hycbut not any more08:07
smiley-yeah true08:08
hycI get 14-16 hours on my rhod210, no real trouble08:08
hycif I just leave it idling it will last a few days08:09
smiley-ok08:14
smiley-yeah that will not happen with this one :)08:14
smiley-the Kovsky has a bit better battery time..    currently I'm trying it out with the same kernel as on RAPH100 and same gingerbread build08:15
smiley-so far it behaves nice08:15
smiley-did port alex charger code so it's actually charging when USB is plugged in08:16
hycI'm sure that makes life better :P08:16
smiley-but still needs to be booted with wall charger plugged for activating the charge08:16
smiley-and apparently that also helps to prevent the panel from freezing08:16
hycweird. I guess device init in the kernel is still missing a lot08:17
smiley-yeah.. looking forward to the 2.6.35 he is working on :)08:17
smiley-have tried it and it works but no wifi so far08:18
smiley-the ti-driver compiles but it can't load it because of symbols missing..  probably lot of old code has been refactored and the wl12* is gonna be used instead08:19
hycwl1251 has been split out from wl12xx in latest linux-wireless source08:19
*** helicopter88 has joined #xdandroid08:39
*** diego_kynetics has quit IRC08:45
*** diego_kynetics has joined #xdandroid08:47
*** diego_kynetics has quit IRC08:49
*** arrrghhh has joined #xdandroid08:58
arrrghhhemwe, re: SoD on .35... unfortunately they were all with the old kernel, before you had hacked in the sleep LED...09:04
arrrghhhhyc, re: battery life on RAPH - they don't have panel collapse... :/09:06
arrrghhhi think every device supported in our port does, except DIAM/RAPH.  actually, i don't know about BLAC...09:06
hycah, that explains a lot09:08
smiley-collapse?09:08
hycLCD powerdown09:08
*** fishhead2567 has joined #xdandroid09:08
smiley-ah09:08
smiley-well it gets completly dark..  but maybe there is power still going there09:09
smiley-(any parameter in /sys /proc that shows it?)09:09
arrrghhhyes09:09
arrrghhhso backlight is shut off smiley-09:09
arrrghhhand brightness is set to 009:09
arrrghhhbut panel is still sucking powa09:09
smiley-ok09:09
arrrghhhi want to say wistilt2 figured out how to collapse the panel09:10
arrrghhhbut they had a ton of issues when he would try to bring it back to life :P09:10
smiley-hehe09:11
smiley-D/lights  (  211): Setting brightnessMode=0 brightness=009:11
smiley-I/power   (  211): *** set_screen_state 009:11
smiley-D/Sensors (  211): LightSensor::enable(4, 0)09:11
smiley-is whats visible when it's shutdown09:11
arrrghhhi wouldn't like smiley-09:11
arrrghhh;)09:11
arrrghhhder09:11
arrrghhhs/like/lie/09:11
smiley-the kovsky says something about epsonpanel also09:12
smiley-lie? :)09:12
arrrghhhyes :P09:12
*** mcm1900 has joined #xdandroid09:12
smiley-no.. I believe you ;)09:12
smiley-from time to time the backlight on the keyboard isn't shut off either..  that can drain battery very fast too09:16
arrrghhhnot sure that LED will pull that much power09:17
arrrghhhhonestly, the panel is probably the #1 drainor09:17
smiley-yeah09:17
smiley-I hope alex gets the 2.6.35 kernel working and I will use the kovsky instead :)09:17
arrrghhhi honestly thought kovs was farther than it is09:18
arrrghhhbut so few people talk about them here09:18
arrrghhhit's like they're not a part of the project, but are :P09:18
arrrghhhjust like blackstone users...09:18
smiley-hehe09:23
smiley-actually the gingerbread build works great on it after replacing the libsurfaceflinger libs09:23
smiley-the panel is mounted upside down on the kovsky..  so you get a 180° rotated image on the LCD ;)09:24
smiley-without the fixed lib09:24
hycFunny, i would have thought setting the screen calibration would tell it the screen orientation too09:25
smiley-all the cordinates gets right on the screen.. but not the image itself09:26
emwearrrghhh: thx for the info.09:41
arrrghhhnp mate09:41
emwebtw, my raph100 survives 24h with ease :P09:44
emweok, idling... but nevertheless. way bettern then my topa idling.09:44
arrrghhhreally!?!?09:44
emweso panel collapse seems - at least for my variant - quite good09:44
arrrghhhi thought TOPA had panel collapse09:44
emweyeah. 27 idle with wifi last great.09:44
emweyeah, but topa's been a bitch ever09:45
arrrghhhlol09:45
emweok, then 1350mAh battery in raph makes it go better too09:45
emwetopa has only 1100mAh.09:45
arrrghhhheh09:45
arrrghhhTOPA is a bastard child09:45
arrrghhhdoes it have a front-facing cam?09:45
emwealready exchanged the battery topa battery from the tattoo hand-me-down from my gf09:45
arrrghhhnice09:45
emweno change. both still in good shape imho09:45
smiley-hehe..  is there a way to check if it has collapsed or not?  in /proc /sys   whatever09:46
emweno sleep led? ;)09:46
smiley-hmm09:46
smiley-I never get that..  but I have ledeffects installed09:46
emwealternative: pm.debug_mask=2   and then dump dmesg after some sleep without cable.09:46
emweledeffects is useless09:46
emweespecially when it comes to sleep09:47
emwewith pm.debug_mask=2 you should get sth like: [ 2031.124725] msm_pm_collapse(): returned 109:47
smiley-pm.debug_mask=2 sounds like I have to reboot.. ;)09:47
emweyap.09:47
emweths is kovs?09:47
arrrghhhLEDEffects is only useful for DIAM/RAPH09:47
arrrghhhi wish it worked for other devices09:48
arrrghhhwas great software.09:48
smiley-yeah this is on the raph09:48
emwewell, but it will never informed about sleep the last time (about a year ago) i checked.09:48
smiley-the kovs is switched off right now..  making a noise in the speaker when LED is flashing09:48
smiley-LEDs09:48
arrrghhhlol09:48
arrrghhhthat's a feature09:48
emweaudiovisual sleep indicator.09:48
smiley-it sucks.. there are kernels that doesnt do that on the xdadev-forums..  but looks like they never sent any patches back for all the tweakings they have done09:49
smiley-duckly something..  familiar name?09:49
emwehyc: jonpry confirmed the stack trace to be his issue on .39 and is also positive your smd tty fixes are gonna fix it.09:49
emwe(or rather your combined effort to name it correctly)09:50
smiley-because if that kernels could be compiled with the smd tty fix that would be a quite stable 2.6.27 for kovs09:50
smiley-that/those09:50
*** helicopter88 has quit IRC09:54
smiley-uninstalled ledeffects and emwe seems to be right..  no sleep led10:09
arrrghhhheh10:12
arrrghhhi thought you'd still get sleep LED on RAPH w/o LEDEffects10:13
arrrghhhi just thought it helped customize it10:13
arrrghhhbut perhaps it is more than just customizing...10:13
smiley-yeah..  we do get that10:13
smiley-maybe the ledeffects settings is still active10:13
smiley-thats one reason I installed it..  i didn't want the flashing light ;)10:14
arrrghhhhehe10:14
arrrghhhi like having an indicator of sleep10:14
arrrghhhif all the sleep bugs get squashed10:14
arrrghhhi'm not sure how i'll deal :P10:14
smiley-then you will love the Macbook ;)10:14
arrrghhhat first i don't think i'll like it10:14
arrrghhhbwhahahahaha10:14
arrrghhhlike other laptops don't indicate they're sleeping :P10:14
arrrghhhoh and my new laptop should be here tomorrow.  it's at some sort facility, thinking about trying to pick it up today :D10:15
arrrghhhwork tiem10:21
*** arrrghhh has quit IRC10:21
*** Detule has quit IRC10:22
emwesleep on raph is "circling led" on that circle control thingy on navipad10:29
emwegtg, laters.10:29
*** emwe has quit IRC10:29
smiley-something seems to keep wifi alive11:00
smiley-connectbot offcourse11:06
*** programmer8922 has quit IRC11:11
*** GlemSom has joined #xdandroid11:32
*** GlemSom has joined #xdandroid11:34
*** helicopter88 has joined #xdandroid11:34
*** Yusuke has joined #xdandroid11:43
*** Yusuke14 has quit IRC11:43
*** Yusuke is now known as Yusuke1411:43
*** Detule has joined #xdandroid11:57
*** arrrghhh has joined #xdandroid12:22
*** |Jeroen| has joined #xdandroid12:32
*** NeoMatrixJR has joined #xdandroid12:44
*** NeoMatrixJR has quit IRC12:55
*** |Jeroen| has quit IRC13:03
*** Detule has quit IRC13:29
*** Detule has joined #xdandroid13:31
*** Detule has quit IRC13:46
*** Detule has joined #xdandroid13:48
*** helicopter88 has quit IRC14:01
*** NeoMatrixJR has joined #xdandroid14:04
*** NeoMatrixJR has quit IRC14:45
*** fishhead2567 has quit IRC14:47
*** NeoMatrixJR has joined #xdandroid15:01
*** emwe has joined #xdandroid15:09
*** fishhead2567 has joined #xdandroid15:19
arrrghhhemwe, howdy16:15
emwegetting nuts about rhod ;)16:17
arrrghhhlol16:20
arrrghhhdid you get my mail?16:22
arrrghhhi hope the log was helpful16:22
arrrghhhthe ends of those ramconsole logs are always full of such garbage.16:22
emwei've seen thos msm_72k errors before i think. at least used to. not recently. nothing like i had here though.16:24
arrrghhhhrm16:24
emwethat wake_lock: alarm_rtc looks familiar honestly.16:25
emweit *might* be insta-plug related.16:25
arrrghhhwell it happeend while on usb/adb16:25
emweoh usb? hm.16:25
arrrghhhyes16:26
arrrghhhi was using adb, checked something16:26
arrrghhhit fell asleep while on usb/adb (thought that was odd)16:26
arrrghhhgot the ding it d/c'd too16:26
arrrghhhand then i was unable to wake it16:26
arrrghhhsolid green light, no kbd backlight or anything would respond.16:26
emweit fell asleep on usb? uhoh16:26
arrrghhhoddly the PC recognized a USB device, but it didn't know what to do with it.16:26
arrrghhhso i rebooted, and pulled ramconsole.16:26
emwebut green=on and nothing would wake is what i had too this morning16:27
emwefirst time ever.16:27
arrrghhhhrm16:27
arrrghhhi had a few of those before, but no sleep indicator.16:27
arrrghhhjust wouldn't wake, no response.16:28
emweif nothing helps i'll have to disable that dex isr again like on .27 :/16:29
arrrghhhbummer16:30
arrrghhhi like that feature16:30
arrrghhhbut if it blows stuff up, not worth it16:30
emweit perhaps just needs to be finetuned. i have a feeling that battery-smem code might have an influence. not sure. always easy to blame something else ;)16:32
arrrghhhlol16:33
arrrghhhindeed16:33
smiley-my raph still up without reboots :)16:35
*** Detule has quit IRC16:35
*** tabunet_iptt has joined #xdandroid16:35
emweok, smd/tty fix pushed to .27 autobuild.16:37
smiley-yay :)16:40
emweyou wouldn't mind to test this one out, no? ;)16:40
*** GlemSom has quit IRC16:40
smiley-haha ;)  on next reboot ;)16:41
smiley-in 201316:41
smiley-;)16:41
emwearrrghhh: you had "random" data loss during sleep, recovering on resume?16:41
xdandroidNew kernel available (20110627_223755) at http://zimages.googlecode.com/files/htc-msm-linux-20110627_223755-package.tar.bz216:47
*** TheSeven has quit IRC17:08
*** TheSeven has joined #xdandroid17:09
arrrghhhemwe, sorry, didn't see your message17:10
arrrghhhi blame work17:10
arrrghhhby data loss, you mean connectivity?17:10
emwenp17:10
emweyap, like seen before the panel vreg-fix for rhodw on gsm17:11
emwei think i saw this after resume several times this afternoon17:11
emwebut it got never broken like before the vreg-fix17:11
arrrghhhhrm17:12
arrrghhhnothing like that, but our setups are different.17:12
*** NeoMatrixJR has quit IRC17:34
*** Orgg has quit IRC17:59
emwearrrghhh: 2 little commits to .35. autobuild should popup soon if you are still giving .35 a go.17:59
emweauto-bl and lcd-backlight.17:59
*** Orgg has joined #xdandroid17:59
arrrghhhcool18:00
emweand smd tty ported from htc ace rather identical to .27. strangely hycs tty hangup fix closes smd channel immedately. perhaps it's that flush of the workqueue in tty_close() which emulates the same. hyc, feel free to comment ;)18:00
arrrghhhemwe, did you post 20110627_223755 on .27?18:00
emwearrrghhh: well, it's just a copying of the .27 stuff18:00
arrrghhhcommit log is... non-existant.18:00
emweoh, that is broken then.18:01
emweGlemSom: if you read logs... commit-log fetching broken :)18:01
arrrghhhheh18:01
arrrghhhthe 6/22 kernel is fine18:01
emwe6/22 from autobuild .35?18:02
arrrghhh.2718:02
emweah heh :)18:02
arrrghhhthe production tree ;)18:02
emwehad you tried earphones with .35?18:02
emwesomehow my audiorouting is broken again18:02
arrrghhhw/extUSB?18:03
emweyap18:03
arrrghhhno... i can tho18:03
emweas if it starts after using a fresh data.img for some time...18:03
arrrghhhhrm18:03
emwebeen making jb crazy already with telling broken/fixed/broken/fixed18:03
arrrghhhlololol18:03
arrrghhhi haven't talked to that dude in ages18:03
arrrghhhhow is his stuff going?18:04
emwesaid he's going to do some forum testing before his patches to kernel should be applied18:04
emweor well.. he just wan'ts to be sure18:04
emweneed to get him in one room with stinebd.18:04
stinebdtell him to come here18:05
arrrghhhindeed18:05
arrrghhhi never see that cat on IRC18:05
emwewe mail quite regularly :)18:06
emwestinebd: heyho.18:06
emweso ok... earphone detection is rboken. doesn't recognize 35mm plug via extusb18:06
emwewtf...18:06
arrrghhhweird18:07
arrrghhhjust fix the actual 3.5mm jack18:07
arrrghhh/done18:07
arrrghhh:P18:07
emwestinebd: i'd like to break stuff again. one little pending change from alex to gb-gps freeing a stack allocated object. can i merge? so one less merge request.18:07
emwearrrghhh: well, that work branch got it going since saturday. but it's a bit of a mess.18:08
stinebdthe gb stuff is fluid so go ahead18:08
arrrghhhstinebd, how is TI wifi goin?18:08
stinebdarrrghhh: fighting with wpa_supplicant18:08
arrrghhhdamn18:08
emwestinebd: you got that power.c issue sorted, right? strangeness it didn't work for your rhod. raph was all i tested and fine...18:08
stinebdemwe: it's all good here, he was writing the wrong thing to sysfs (mem vs off or whatever)18:09
emwestinebd: if you need some help, how about pushing what you got and for rootfs to another prerelease branch so we can chime in testing?18:09
emwestinebd: strangeness though it worked on raph...18:09
stinebdwhat i'm fighting with is in libhardware_legacy18:09
stinebdemwe: it was broken on both my raph and rhod18:09
emwestinebd: really? man, i did a full rebuild of froyo just to make sure. then i am not trustworthy anymore and you should revoke my commit rights. damn build system driving me nuts18:10
arrrghhhlol18:10
emweofc this was a joke!18:10
arrrghhhhehehe18:10
emweno, that buildsystem is a pain18:10
stinebdit is18:10
emwechange on thing, think it's picked up and nada. or at leat it does not land in system.img reliably imho18:11
emweperhaps i am just too dumb...18:11
arrrghhhemwe, so far new .35 is great on batt life18:11
arrrghhhphone isn't getting hot either18:12
arrrghhhi think before something was keeping it awake, and i had no clue.18:12
arrrghhhbrb18:12
emweif you have watched it more closely, i am always eager about some subjective comparisons. as you know them all. ( the kernels i mean)18:12
emwestinebd: so if you need any help with some pre FRX07 testing, please don't hesitate.18:13
emweat least the raph with wifi and SIM is setup quickly. can do testing with topa and rhod then as well if required.18:13
emwealways a pita to backup data and system images, though.18:14
emwearrrghhh: if bored, try the .35 autobuild.18:16
emweso, gtg. gnight.18:16
hycemwe: re: smd_tty hangup - dunno what you're seeing, a dmesg might help18:17
*** xdandroid has joined #xdandroid18:24
arrrghhhwb18:24
*** fishhead2567 has quit IRC18:39
*** NeoMatrixJR has joined #xdandroid18:42
*** arrrghhh has quit IRC19:50
*** Fudge has quit IRC20:09
*** Fudge has joined #xdandroid20:13
*** jonpry has joined #xdandroid20:18
*** tabunet_iptt has quit IRC20:45
*** ORi| has quit IRC21:03
*** ORi|| has joined #xdandroid21:04
*** arrrghhh has joined #xdandroid21:05
*** tabunet has joined #xdandroid21:40
*** TheSeven has quit IRC22:06
*** [7] has joined #xdandroid22:06
*** hardwalker has joined #xdandroid22:19
*** Hackbat has joined #xdandroid22:28
*** SG has joined #xdandroid22:54
*** SG has left #xdandroid22:54
*** programmer8922 has joined #xdandroid23:48
*** Hackbat has quit IRC23:49
*** phh has quit IRC23:56
*** phh has joined #xdandroid23:56

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!