| *** rpierce99 has quit IRC | 00:26 | |
| d3tul3 | xdandroid, later tell GlemSom I think the .39 autobuild is down, thanks again for hosting it | 00:31 |
|---|---|---|
| xdandroid | d3tul3: The operation succeeded. | 00:31 |
| arrrghhh | d3tul3, hrm | 00:32 |
| arrrghhh | wistilt2's seems to not be fully up to date either. | 00:32 |
| d3tul3 | what was the address for that one i keep forgetting to bookmark it | 00:33 |
| arrrghhh | http://autobuild.wirelesstcp.net/ | 00:33 |
| d3tul3 | nice that one might actually be up to date | 00:34 |
| arrrghhh | orly | 00:34 |
| arrrghhh | there was a bunch of build "attempts" from 11/8 on glemsom's autobuild | 00:34 |
| d3tul3 | you are probably right | 00:35 |
| arrrghhh | i dunno dude, just staring & comparing :P | 00:35 |
| arrrghhh | i haven't spoken to wistilt2 in a few days | 00:36 |
| arrrghhh | or looked at his git | 00:36 |
| arrrghhh | or really done anything... | 00:36 |
| bzo | d3tul3: I think I fixed the problem | 00:37 |
| d3tul3 | the missing permissions? | 00:38 |
| d3tul3 | i am just booting with the patch | 00:38 |
| bzo | yeah | 00:38 |
| d3tul3 | so freaking stupid | 00:38 |
| bzo | yeah, how fricking retarded that the system server pid is hardcoded with perms | 00:38 |
| d3tul3 | i just learnt more about lmk than i ever want to know | 00:38 |
| arrrghhh | lol | 00:39 |
| bzo | no wonder this crap is not able to go back upstream | 00:39 |
| bzo | anyhow, ran a bunch of crap and still have plenty of memory | 00:39 |
| d3tul3 | i am about to throw the sink at it | 00:39 |
| bzo | arrrghhh: you back on haret right now? | 00:40 |
| arrrghhh | sorry not at the moment. | 00:40 |
| arrrghhh | sounds like you fixed the slowdown plague in .39? | 00:41 |
| bzo | hopefully, d3tul3 led the way | 00:41 |
| d3tul3 | i take no credit Mike Chan is apparently the patch author :) | 00:42 |
| arrrghhh | i don't see why this didn't effect wistilt2 ? | 00:42 |
| bzo | I forward ported my own patch :P | 00:42 |
| d3tul3 | yeah i had to add an extra NULL in generic_permissions:) | 00:42 |
| bzo | I think the fnct decls are different in 39, we may want to use mine | 00:43 |
| d3tul3 | i just added this http://git.ncl.cs.columbia.edu/git/?p=linux-android-common.git;a=commitdiff;h=004f0ae31748b0f2d77c3bf3bff927b6ce2780bc | 00:44 |
| bzo | I think it shoudl be tweaked: | 00:45 |
| bzo | static int oom_adjust_permission(struct inode *inode, int mask, unsigned int flags) | 00:45 |
| bzo | return generic_permission(inode, mask, flags, NULL); | 00:45 |
| bzo | arrrghhh: I suppose if you only use on the same few apps, you might have enough memory to make it work | 00:46 |
| bzo | s/on/only | 00:47 |
| arrrghhh | interesting. | 00:47 |
| d3tul3 | also if he is using a task killer it would alleviate the problems | 00:47 |
| arrrghhh | i couldn't make it work on the stock system | 00:47 |
| d3tul3 | (i was using one in 06+) | 00:47 |
| arrrghhh | ah, i know he uses system panel lite | 00:47 |
| bzo | yeah, if I kept killing apps, I could make mine last a bit longer | 00:47 |
| d3tul3 | bzo, i see how that would make sense, i just augmented it to return generic_permission(inode, mask, NULL, NULL) but yours seems more functional | 00:48 |
| bzo | probably works either way, but might as well do it they way it is supposed to be | 00:48 |
| d3tul3 | ok threw the sink i can't bring it down under 40MB of free usage as reported by whatever task manager i have in my launcher -> same as .35 | 00:49 |
| d3tul3 | and tasks actually disappear off the list....pretty remarkable | 00:49 |
| arrrghhh | heh | 00:49 |
| bzo | ok, I say it's fixed then :) | 00:50 |
| d3tul3 | ok i am off to bed, bzo you should format a patch and shoot it to WisTilt2....on that note perhaps it's time to expand the commit access to that tree | 00:51 |
| bzo | ok, will do | 00:51 |
| bzo | will also shoot a note to acl who is the owner of that tree | 00:51 |
| *** d3tul3 has quit IRC | 00:52 | |
| *** bzo has quit IRC | 01:13 | |
| *** raymonddull has joined #xdandroid | 01:18 | |
| *** [acl] has joined #xdandroid | 01:20 | |
| *** [acl] has quit IRC | 01:55 | |
| *** kalemas has joined #xdandroid | 02:24 | |
| *** raymonddull is now known as ray|actuallyslee | 03:17 | |
| *** ray|actuallyslee is now known as ray|sleeping | 03:17 | |
| *** corein has joined #xdandroid | 03:44 | |
| *** virsys has quit IRC | 03:45 | |
| *** virsys has joined #xdandroid | 04:03 | |
| *** Yusuke14 has quit IRC | 04:16 | |
| *** Tandyman100 has quit IRC | 04:16 | |
| *** Yusuke has joined #xdandroid | 04:16 | |
| *** Yusuke is now known as Yusuke14 | 04:16 | |
| *** ray|sleeping has quit IRC | 04:31 | |
| *** Yusuke14 has quit IRC | 05:14 | |
| *** Yusuke has joined #xdandroid | 05:14 | |
| *** Yusuke is now known as Yusuke14 | 05:14 | |
| *** hardwalker has quit IRC | 05:14 | |
| *** vinceweis has quit IRC | 05:23 | |
| *** panda84kde has joined #xdandroid | 05:32 | |
| *** mgross029 has joined #xdandroid | 08:12 | |
| *** Tandyman100 has joined #xdandroid | 08:42 | |
| *** helicopter88 has joined #xdandroid | 08:53 | |
| *** vinceweis_ has joined #xdandroid | 08:55 | |
| *** ryannathans has joined #xdandroid | 08:57 | |
| *** helicopter88 has quit IRC | 09:04 | |
| *** ryannathans is now known as ryannathansSLEEP | 09:24 | |
| *** helicopter88 has joined #xdandroid | 09:35 | |
| *** kalemas has left #xdandroid | 09:48 | |
| *** kalemas has joined #xdandroid | 09:50 | |
| *** helicopter88 has quit IRC | 11:03 | |
| *** helicopter88 has joined #xdandroid | 11:30 | |
| *** programmer8922 has joined #xdandroid | 11:30 | |
| *** programmer8922 has quit IRC | 11:31 | |
| detule | xdandroid, later tell bzo scratch camera off the list, userland issue, works fine with fast burst camera lite off market (from acl's nand ppcg thread) | 11:45 |
| xdandroid | detule: The operation succeeded. | 11:45 |
| *** bzo has joined #xdandroid | 11:49 | |
| *** emwe has joined #xdandroid | 12:02 | |
| *** emwe has joined #xdandroid | 12:03 | |
| *** emwe has quit IRC | 12:04 | |
| *** emwe has joined #xdandroid | 12:08 | |
| *** emwe has quit IRC | 12:15 | |
| *** e33 has joined #xdandroid | 12:35 | |
| e33 | Hi all, anyone download android reverse engineering? | 12:36 |
| *** emwe has joined #xdandroid | 12:41 | |
| *** e33 has quit IRC | 12:47 | |
| *** emwe has quit IRC | 12:49 | |
| *** emwe has joined #xdandroid | 12:51 | |
| *** SytheZN has joined #xdandroid | 12:55 | |
| SytheZN | green-tings etcetera | 12:55 |
| *** helicopter88 has quit IRC | 13:14 | |
| *** helicopter88 has joined #xdandroid | 13:34 | |
| *** tguy has joined #xdandroid | 13:38 | |
| *** |Jeroen| has joined #xdandroid | 13:39 | |
| *** tguy has joined #xdandroid | 13:39 | |
| *** tguy has left #xdandroid | 13:44 | |
| *** |Jeroen| has quit IRC | 13:59 | |
| *** panda84kde has quit IRC | 14:02 | |
| *** helicopter88 is now known as helicAWAY | 14:06 | |
| *** AndersG has joined #xdandroid | 14:11 | |
| *** programmer8922 has joined #xdandroid | 14:11 | |
| detule | emwe, not to sound like a broken record, but userland needs the caps patch :) | 14:33 |
| emwe | detule: i know. already stuffed in your ril patch. i will just throw it out. we don't have the time to discuss that "user error". | 14:34 |
| detule | ok i can kill the irq listener in microp-keypad | 14:35 |
| emwe | irq listener in microp-keypad? | 14:36 |
| detule | microp_keypad_led_event | 14:38 |
| detule | and the even listener that activates it | 14:39 |
| emwe | assume that got cleanuped up by alex on .35 | 14:40 |
| emwe | don't see it. | 14:40 |
| detule | it's not in .35 | 14:40 |
| emwe | ah i got it. | 14:41 |
| *** f00bar80 has joined #xdandroid | 14:41 | |
| f00bar80 | any idea why pptp vpn on froyo FRX07 ? | 14:42 |
| emwe | didn't even remember what was in it. ) | 14:42 |
| detule | emwe, your kernel has some sort of a swap you are keeping secret | 14:46 |
| emwe | ramzswap? that's just another name for compcache | 14:46 |
| emwe | it's enabled, but somehow never ever used. | 14:46 |
| emwe | i think some successor to ramzswap is kinda zcache now... | 14:47 |
| emwe | i dunno | 14:47 |
| emwe | dunno worth it. | 14:47 |
| detule | i see this is somehow compiled into the kernel, not as a module? | 14:47 |
| emwe | i think as module. loaded up by userinit.sh by default. | 14:47 |
| emwe | yes, module. | 14:48 |
| emwe | it's built separately in .27 | 14:48 |
| emwe | should be in staging. even looked at .39 and it was around iirc. | 14:49 |
| emwe | but i see no benefit/use of it. | 14:49 |
| detule | yeah i noticed on your kernel, with a functioning lmk, it hardly gets used | 14:51 |
| detule | my swap file prior to yesterday was getting a serious workout | 14:52 |
| emwe | i enabled swap just for a little amount of time. | 14:56 |
| emwe | actually have it disabled. | 14:56 |
| emwe | and compcache is not triggered at all, allthough it's known to the system. | 14:56 |
| emwe | but even with stock lmk-parameters, that swap never got hit. | 14:56 |
| mgross029 | should there be major differeces between .39 and .35 snd.c code? | 14:56 |
| emwe | mgross029: yes. alex modularized everything. | 14:57 |
| emwe | .39 has the .27 code more or less | 14:57 |
| emwe | so .35 is different in several places. | 14:57 |
| emwe | .39 and .27 is a bit more identical. | 14:57 |
| emwe | because most of the old code just got pulled over. | 14:57 |
| emwe | detule: damn, the fbase patch doesn't apply cleanly anymore. | 14:58 |
| mgross029 | Ok. Thanks. Makes sense | 14:58 |
| detule | mine? | 14:58 |
| detule | or cyanos? | 14:58 |
| emwe | well, yours. the reduced one from cyano without wimax and flashlight light stuff | 14:59 |
| detule | perhaps do cyanogen's then i think they are identical, mine was built against froyo, his perhaps against gb | 15:00 |
| emwe | we got two hyc patches in PowerManagerService.java afterwards which might interfere for finding the sweet spot for git apply/am | 15:00 |
| detule | oh | 15:00 |
| emwe | yeah, no biggi. | 15:02 |
| *** vinceweis_ has quit IRC | 15:04 | |
| *** kalemas has left #xdandroid | 15:05 | |
| detule | btw i should warn you | 15:11 |
| detule | the kernel side patch was only tested on rhod 400 | 15:11 |
| detule | :) | 15:12 |
| emwe | i am not gonna touch this, yet. | 15:21 |
| detule | looks complicated? | 15:21 |
| detule | for what is worth, the userland patch can be applied without the kernel side on .27, it won't break anything | 15:22 |
| detule | the kernel side patches can be applied only in .35 and .39 for starters, with the much more limited audience those kernels are geared towards | 15:22 |
| emwe | yeah, that sounds totally ok for me. | 15:23 |
| emwe | seems the force is going the .3x anyway. | 15:23 |
| emwe | dunno how long i'll stay here with .35 that alone... | 15:24 |
| emwe | damn git am. it will not allow me to resolve a merge/patch failure. | 15:24 |
| emwe | tree is clean after it bails out... | 15:25 |
| emwe | ok, with some trickery it seems in... | 15:28 |
| bzo | emwe - why not focus on 39, then bring in the 35 improvements? | 15:31 |
| bzo | detule has done that already with the alex clocks | 15:31 |
| bzo | and there are no showstopper bugs on 39 | 15:31 |
| detule | i've been thinking about this, .39 needs topa love | 15:32 |
| emwe | i am tired of doing it all again | 15:32 |
| bzo | detule, I think a kernel fix is needed for the camera | 15:32 |
| bzo | the jpeg encoder is broken, what you refer to is a workaround to not use the dsp encoder | 15:33 |
| emwe | who's been using his .39 full time without fuzz? | 15:33 |
| bzo | since the fix yesterday, it's been working better than 27 | 15:33 |
| bzo | albeit with a few things that need fixing (but equal to 35 in that regard I think) | 15:34 |
| bzo | I think wis has been running it for quite a long time | 15:34 |
| emwe | i think i am not ready with .35 somehow | 15:35 |
| emwe | and i am kinda afraid of switching ... not that i wouldn't like to | 15:36 |
| emwe | .35 has drifted off quite a bit... | 15:36 |
| emwe | from the "old" stuff. | 15:36 |
| emwe | detule: damnit, we need another fork.. hardware/libhardware just for the freaking lights.h defs | 15:37 |
| bzo | in any case, I think at least stopping 27 dev will get us more focused | 15:37 |
| bzo | and acl seemed amenable to switching over to .3x at some point | 15:37 |
| emwe | yeah, he said .35 several times :P | 15:38 |
| bzo | I'm not tied to either one, I'm just going with the flow :P | 15:39 |
| emwe | yeah :) | 15:39 |
| emwe | stinebd: you around my dear? | 15:39 |
| emwe | detule: which defconfig is .39? | 15:42 |
| emwe | msm_defconfig i assume? | 15:43 |
| bzo | yes | 15:43 |
| detule | bzo, ok about the camera, i know next to nothing about it | 15:46 |
| bzo | you don't wanna know ;) | 15:46 |
| emwe | bzo: as i just stumbled over your question again why not focus on .39... three people sound enough for that tree. i can try to get sth done for .35... if that improves it when my hands are in the game, dunno. | 15:47 |
| * emwe pings stinebd again | 15:48 | |
| bzo | sounds fine to me. I'm sure fixes are easily ported back and forth | 15:48 |
| bzo | and we probably have quite a few we want to pull from 35 | 15:48 |
| emwe | it might be not really fixes, but rather code refactorings, mainly done by alex. | 15:49 |
| bzo | yeah, I meant those as well when I say fixes | 15:50 |
| detule | i would like to get the kernels looking similarly enough, so when it comes to nand, they are interchangeable | 15:50 |
| emwe | yap. allthough .35 is LK-bootable according to alex. | 15:52 |
| emwe | iirc his tree is nand-ready. | 15:52 |
| emwe | i just neeed to find the time to rebase on his tree and replay all the non-kovsky stuff | 15:53 |
| emwe | i'll likely never get to it... | 15:53 |
| emwe | ugly... .39 has built through. | 15:54 |
| bzo | was there a fix for data not reconnecting? | 15:59 |
| * SytheZN yawns vehemently | 15:59 | |
| emwe | yes. but somehow it seems to work for cdma only | 15:59 |
| emwe | i do have reconnection issues at times. | 15:59 |
| emwe | sometimes this bitch also switches to cdma and i can't for the sake getting out of it | 16:00 |
| SytheZN | btw, what does iirc stand for? | 16:00 |
| emwe | If I Recall Correctly | 16:00 |
| SytheZN | ty :D | 16:00 |
| emwe | you're welcome. | 16:00 |
| bzo | was the fix pushed out? | 16:00 |
| emwe | of course :P | 16:00 |
| bzo | ok, must be hitting a different problem then | 16:01 |
| emwe | into frameworks_base | 16:01 |
| emwe | mgross029: reported it to be fine. i think arrrghhh also. | 16:01 |
| * SytheZN swears at his mobile provider | 16:01 | |
| SytheZN | gotta run... interesting convo tho *learning a lot | 16:02 |
| emwe | have fun. | 16:03 |
| *** SytheZN has quit IRC | 16:03 | |
| bzo | emwe, I suppose we should pull your last commit into 39 | 16:04 |
| emwe | yap. | 16:05 |
| emwe | initially wanted to go pdata way, but then i spotted there's machine checks in htc_acoustic_wince.c anyway... and did it like that. | 16:05 |
| emwe | not clean. | 16:05 |
| emwe | ideally on .35 this would have had to be pulled through two pdatas. one for amss_6125.c driver and then put into acoustic wince pdata. | 16:06 |
| mgross029 | emwe, maybe its the Global setting. :p Try it you know you want too. | 16:07 |
| *** helicAWAY is now known as helicopter88 | 16:08 | |
| bzo | emwe, since they are different repos, I guess I need to pull and apply the patch manually? | 16:08 |
| mgross029 | data reconnect working great here after that fix in the NetworkStateTracker.java | 16:09 |
| emwe | bzo: you can git remote add dot35 <git-url-to-linux-msm>; get fetch dot35; git cherry-pick <my-commits-commitish> | 16:12 |
| emwe | bzo: or git log dot35/htc-msm-2.6.35 to browse the logs | 16:13 |
| emwe | but i dunno the whole state of the .39 code and how clean - even this oneliner - will apply. | 16:14 |
| bzo | ok, thanks emwe | 16:14 |
| emwe | you need to keep the remote info up to date though with regular git fetch dot35 | 16:14 |
| bzo | ok | 16:15 |
| emwe | looks like nobody has dared to build tiwlan against .39 | 16:15 |
| *** Tandyman100 has quit IRC | 16:15 | |
| bzo | so are there any more issues with acoustic now? | 16:15 |
| emwe | switching to speaker during a call switches to max speaker loudness by default | 16:15 |
| mgross029 | emwe: pulling your latest autobuild. Thanks | 16:16 |
| emwe | you need to toggle vol for it to set the vol it was desired for. | 16:16 |
| emwe | mgross029: as you are getting more calls then I do, let me know how it does. | 16:16 |
| emwe | bzo: other then that... a little dialtone issue. switch to full vol and hit dial tones. their volume will increase till quite some too loud level. | 16:17 |
| mgross029 | Yep | 16:17 |
| mgross029 | The wife will appreciate a working phone. :p | 16:17 |
| emwe | bzo: then switch back to 3/4 and 3/4 will be louder as without going max before. | 16:17 |
| emwe | mgross029: lol. | 16:17 |
| bzo | weird | 16:17 |
| emwe | bzo: that might be a kernel thing, though. | 16:17 |
| emwe | haven't had the time. my RLs been kicking recently... | 16:17 |
| bzo | though I guess both those fall more into the annoying category as opposed to unusable | 16:17 |
| emwe | yap | 16:18 |
| emwe | incall vol is a major relief. | 16:18 |
| bzo | yeah, that one falls more into the unusable | 16:18 |
| emwe | detule: please refrain from sending frameworks/base patches again. these provoke near-full build times. | 16:22 |
| emwe | somebody wan't me to buy some core 2 duo t9900? | 16:23 |
| emwe | bzo: after doing nothing, my dataconn has fixed up itself. was gray before. now green again and has internets again. | 16:26 |
| bzo | i may just have a bad signal here then | 16:26 |
| emwe | in big cities this bitch totally freaks out here. | 16:27 |
| emwe | up till the point where it won't do anything and stay in "no-connection" mode... not airplane. | 16:27 |
| emwe | just that little x at the signal bars | 16:27 |
| emwe | likely it has fallen into cdma mode again not getting out of there | 16:27 |
| bzo | I found using a newer radio helped | 16:27 |
| emwe | how far will it get me with a new radio in DE here? | 16:28 |
| emwe | what have you flashed? | 16:28 |
| bzo | verizon 2.38 on my sprint 400 | 16:28 |
| bzo | winmo doesn't run too well with it for me though | 16:29 |
| emwe | hm. easy to flash? just nbh an dbe done? | 16:29 |
| bzo | yeah, just copy it to your sd as RHODIMG.nbh and power from bootloader | 16:29 |
| emwe | i assume i can flash back when that proves bad? | 16:30 |
| emwe | that's the last thing i need here. a broken RHOD400 in germany. | 16:30 |
| bzo | yah, I switched back and forth a couple of times | 16:30 |
| emwe | sorry to ask those question.. what was the key combo? pwr + vol up + vol down? | 16:31 |
| bzo | power off then, power on with vol - | 16:31 |
| emwe | k | 16:31 |
| bzo | hmm, 39 may have the same problem with usb as 35 | 16:31 |
| emwe | it's "fine" for me, kinda. sometimes looses connection after some time | 16:32 |
| emwe | but no more idle issues after letting it not fall into idle sleep | 16:32 |
| emwe | but mgross029 has quite some trouble with the same code | 16:32 |
| bzo | yeah, that problem, sod with usb | 16:33 |
| emwe | on RHOD500, though. | 16:33 |
| bzo | 400==500 afaict | 16:33 |
| emwe | i assume you are not suspending/resuming all over according to dmesg | 16:33 |
| emwe | ? | 16:33 |
| emwe | there seem to be subtle differences, or? | 16:33 |
| emwe | but now i got a wake SOD as well | 16:34 |
| bzo | I've never seen any differences, I think that you can interchange radios indicates they are the same | 16:34 |
| mgross029 | bzo: Just go to Applications -> Development and check Stay Awake. Good work around I found | 16:34 |
| emwe | panel still shown .. radiologs spits stuff out... but badly badly, i have logcat and dmesg piped with a grep for "snd"... damnit | 16:34 |
| bzo | yeah, I'll have to do that | 16:34 |
| emwe | can't open new adb conn | 16:34 |
| *** AndersG has quit IRC | 16:35 | |
| mgross029 | emwe: Just had a thought. last_kmsg should show what happened after the system drops into SOD I would think at least from a dmesg stand point | 16:36 |
| emwe | i hope it's filled. | 16:36 |
| mgross029 | That was were I pulled that kernel panic info from before so it is getting populated | 16:37 |
| emwe | ok. | 16:38 |
| emwe | bzo: you can point me to the 2.38 radio? | 16:39 |
| emwe | can't seem to find it on xda | 16:39 |
| mgross029 | So is anyone going to try the build I sent with Global setting? :p | 16:39 |
| emwe | mgross029: don't have the time, yet. | 16:40 |
| mgross029 | bzo, I'm only on 2.37 didn't hear of a 2.38 coming out. | 16:40 |
| bzo | err, I mean 2.37 | 16:40 |
| bzo | emwe http://forum.ppcgeeks.com/2009123-post498.html | 16:40 |
| f00bar80 | any idea why pptp vpn on froyo FRX07 ? | 16:41 |
| f00bar80 | is not working? | 16:42 |
| mgross029 | emwe, I know... Just throwing it out there. :p heh | 16:42 |
| emwe | f00bar80: might be some kernel modules are missing. | 16:42 |
| mgross029 | f00bar80: I wouldn't. Never used vpn over my phone. Sorry | 16:43 |
| f00bar80 | what i have to observe in dmesg or in any other logs ? | 16:44 |
| emwe | bzo: that "you must be security unlocked foobar" doesn't have to worry me in any way? | 16:44 |
| bzo | hmm, I don't believe you need that since it is an official signed package | 16:44 |
| emwe | cross your fingers. | 16:45 |
| *** helicopter88 has quit IRC | 16:45 | |
| mgross029 | f00bar80: dmesg and maybe logcat / radio | 16:45 |
| bzo | if you need to be security unlocked, it should just refuse to accept it | 16:45 |
| f00bar80 | mgross029, where can i find logcat / radio | 16:46 |
| mgross029 | f00bar80, adb logcat -v time -b radio | 16:47 |
| emwe | bzo: i need to reset after i see "Update Complete/UPDATE SUCCESS" or will it self reset? | 16:48 |
| mgross029 | If you aren't using adb you can just run logcat from your terminal on the phone (without the adb in front of course) | 16:48 |
| emwe | (asking in order to not break this thing) | 16:48 |
| bzo | I reset it myself last time, it didn't self reset for me | 16:48 |
| emwe | 2.37.00WV | 16:49 |
| bzo | success | 16:49 |
| mgross029 | Ok. guys outta here. Have a good one! | 16:50 |
| emwe | laters. | 16:50 |
| *** mgross029 has quit IRC | 16:50 | |
| emwe | oh that was a brilliant idea mgross029 (in case you read logs) | 16:54 |
| emwe | ah lol it was, but not useful for this case | 16:55 |
| emwe | because i pulled the sd card and that provoked a shitload of error messages. | 16:55 |
| *** turbolapin has joined #xdandroid | 17:05 | |
| *** turbolapin has quit IRC | 17:07 | |
| detule | emwe, 's last commit is already in .39 | 17:09 |
| bzo | not the one from today... | 17:10 |
| emwe | it's in the board file, yeah :) | 17:11 |
| emwe | i also just figued while standing on the balkony. | 17:11 |
| emwe | but it's safe to keep when we pull in alex' amss/qdsp refactorings | 17:11 |
| detule | bzo, the mic_bias_callback is nulled out in the rhodium board file | 17:11 |
| detule | (per JB's original code) | 17:11 |
| emwe | my brain hurts when we wan#t to pull in all those refactorings cleanly... | 17:12 |
| bzo | oh ok | 17:12 |
| emwe | took me a bit to keep things working somehow | 17:12 |
| emwe | gotta get some sleep | 17:13 |
| emwe | gnight peeps | 17:13 |
| detule | good night emwe | 17:13 |
| emwe | night detule and bzo. | 17:13 |
| bzo | might | 17:14 |
| bzo | err night | 17:14 |
| detule | bzo, sod with usb? | 17:14 |
| bzo | yep | 17:14 |
| detule | (sorry catching up with the log here) | 17:14 |
| detule | damn that's a first | 17:15 |
| detule | and i have the VZ radio as well | 17:15 |
| bzo | I think I confused some of the prior slowdown with this problem | 17:15 |
| bzo | because while on an adb shell, things start to crawl | 17:15 |
| detule | can you reproduce it consistently? | 17:15 |
| bzo | without the shell, you would assume a SOD | 17:15 |
| bzo | think so, not positive since we just fixed the other problem | 17:16 |
| *** emwe has quit IRC | 17:17 | |
| detule | bzo, btw, the memfree /adj values in emwe's rootfs need adjusting | 17:17 |
| detule | they override V6 | 17:18 |
| detule | and currently whenever you are under 80MB free but over 40MB it scans all /proc trying to find pid's with oom_adj > 14.....these are pretty hard to find | 17:18 |
| bzo | ic, I'm running a standard rootfs though | 17:18 |
| detule | so it keeps scanning and scanning, and returning empty handed | 17:18 |
| detule | what are the minfree and adj values in the standard rootfs anyway? | 17:21 |
| bzo | never looked into it, probably not setting it explicitly | 17:22 |
| detule | here they are write /sys/module/lowmemorykiller/parameters/minfree 1536,2048,4096,5120,5632,6144 | 17:23 |
| detule | so what this means that lowmemory killer doesn't kick in until less than 24MB free | 17:23 |
| detule | that seems .... suboptimal | 17:23 |
| bzo | these values probably came from the vogue, a 128mb device | 17:24 |
| *** turbolapin has joined #xdandroid | 17:40 | |
| detule | if V6 has not somehow overridden those values for you, it could account for some sluggishness | 17:43 |
| *** mes_ has joined #xdandroid | 17:43 | |
| turbolapin | is git clone git://gitorious.org/linux-on-qualcomm-s-msm/linux-msm.git the same as git clone git://gitorious.org/linux-on-qualcomm-s-msm/linux-msm.git -b linux-msm-2.6.27? | 17:45 |
| bzo | does git clone even take a -b arg? | 17:48 |
| detule | it might force to make the HEAD of the clone something different than the head of the remote | 17:48 |
| bzo | oh, it looks like it just sets the intial branch for you | 17:49 |
| detule | bzo, have you heard about this alignment trap business with the 2011q3 codesourcery toolchain | 17:50 |
| bzo | in passing, don't know much about it | 17:50 |
| detule | in theory it should improve memory allocation | 17:50 |
| turbolapin | @bzo i think i agree with you | 17:51 |
| detule | reduce some overhead from forcing software alignment | 17:51 |
| bzo | I suppose our ancient hardware could use all the optimization it can get | 17:51 |
| detule | i think so, emwe sent me a config/Kconfig patch needed to disable the software trap | 17:52 |
| detule | here it is http://dl.dropbox.com/u/12183048/htctopaz/dot35/patches/0001-defconfig-disable-sw-alignment-trap-with-codesourcer.patch | 17:53 |
| bzo | did he try it? | 17:54 |
| detule | i am trying to pawn this off on you because i built this once,and got some rather strange results | 17:54 |
| detule | not sure | 17:54 |
| bzo | mebbe later, in the middle of stuff rightnow | 17:54 |
| *** ImCoKeMaN has joined #xdandroid | 18:04 | |
| *** corein has quit IRC | 18:15 | |
| *** mes_ has quit IRC | 18:27 | |
| *** ryannathansSLEEP is now known as ryannathans | 18:48 | |
| *** virsys has quit IRC | 18:54 | |
| *** d3tul3 has joined #xdandroid | 18:55 | |
| *** virsys has joined #xdandroid | 19:06 | |
| *** turbolapin has quit IRC | 19:08 | |
| *** ryannathans is now known as ryannathansAFK | 19:18 | |
| *** raymonddull has joined #xdandroid | 19:25 | |
| *** raymonddull has quit IRC | 19:28 | |
| *** Tandyman100 has joined #xdandroid | 19:32 | |
| *** ryannathansAFK is now known as ryannathans | 19:34 | |
| *** ryannathans is now known as ryannathansAFK | 19:37 | |
| *** hardwalker has joined #xdandroid | 20:11 | |
| *** programmer8922 has quit IRC | 20:19 | |
| *** ringer has joined #xdandroid | 21:58 | |
| *** [7] has quit IRC | 22:58 | |
| *** TheSeven has joined #xdandroid | 22:58 | |
| *** raymonddull has joined #xdandroid | 23:05 | |
| *** virsys has quit IRC | 23:43 | |
| *** d3tul3 has quit IRC | 23:45 | |
| *** virsys has joined #xdandroid | 23:56 | |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!