00:12.08 | *** join/#htc-linux snadge (~snadge@unaffiliated/snadge) |
00:13.03 | WisTilt2 | jonpry: while im freeing up all this memory and moving things around, any benefit to increase ram console from 256k up to say 1mb? |
00:13.37 | jonpry | probably not. never really run out of log |
00:17.03 | WisTilt2 | k. been tied up here on and off but have done some fancy things with memory but not letting the cat out of the bag until i have arrrghhh and rpierce99 test it on cdma:) |
00:17.28 | arrrghhh | heh |
00:17.34 | rpierce99 | ready and willing, I'm still running your "slow" kernel with the huge gobs of memory from last week |
00:18.00 | WisTilt2 | stable in that configuration? other than slow desktop |
00:20.21 | rpierce99 | haven't had any abnormal issues that i've noticed |
00:20.39 | rpierce99 | i don't really do much "weird" stuff though, i don't use the camera or bluetooth etc |
00:22.42 | d3tul3 | want to test out a possible camera fix while you are at it? |
00:22.46 | d3tul3 | not kernel related |
00:23.41 | rpierce99 | it's broken? |
00:25.09 | d3tul3 | don't know is it workiing for you right now |
00:25.21 | arrrghhh | lol |
00:27.41 | d3tul3 | it's broken on the autobuild kernels not sure if the kernel you are on is hacked to disregard pmem overlaps as reported by userland |
00:28.40 | WisTilt2 | the kernel of mine he is running has pmem check removed so camera should work. |
00:30.28 | jonpry | hrm. codeaurora hw3d.c doesn't have any PM related stuff |
00:35.17 | d3tul3 | so er these revoke hooks they are called on suspend |
00:35.30 | d3tul3 | and what force close the gpu? |
00:36.18 | *** join/#htc-linux raymonddull (~raymonddu@c-24-11-68-98.hsd1.mi.comcast.net) |
00:40.37 | d3tul3 | i think system_server is listed as holding hw3d open |
00:41.15 | d3tul3 | and when it does something like [ 1656.585357] hw3d: Requesting revoke for suspend |
00:41.16 | d3tul3 | [ 1658.585205] hw3d: force revoke from pid=2198 |
00:41.16 | d3tul3 | <PROTECTED> |
00:46.48 | jonpry | yeah that is done in hw3d.c itself |
00:47.12 | jonpry | we could stop it. but this could have negative impact on power consumption |
00:48.04 | jonpry | might be able to turn off some of the clocks. but who knows if collapse will clear registers in hw3d that the blob just can't deal with |
00:48.47 | d3tul3 | i'll try compiling a kernel without that early_suspend call |
00:52.05 | jonpry | got tracking on new touchpads |
00:52.19 | jonpry | rpierce99 is the man :p |
00:52.59 | rpierce99 | :) |
00:53.07 | rpierce99 | got my # as well, but not "shipped" yet |
00:53.33 | jonpry | looks pretty legit to me |
00:54.33 | d3tul3 | what's the plan for the new touchpads - presents or toys? |
00:56.18 | jonpry | mine has a crappy usb port and strange bright spot on the screen. so trade first |
00:56.24 | jonpry | then present |
00:57.51 | rpierce99 | i'm planning on installing an OS on it that is just unstable enough that it drives my wife batty so she uses the kids and lets me tinker with the TP |
00:57.51 | d3tul3 | jonpry, you should open an xda_frameworks_base repo |
00:58.09 | jonpry | for your mods? |
00:58.16 | d3tul3 | for your mods |
00:58.36 | jonpry | i'm already maxed out on github space |
00:59.07 | rpierce99 | i thought stinebd said that was a soft cap |
00:59.26 | jonpry | yeah so i have to email and beg for more |
01:03.04 | rpierce99 | just tell them "yeah, you know that touchpad thing and how the touchscreen didn't work? you're welcome. Now give me more space" |
01:03.28 | jonpry | lol |
01:04.27 | jonpry | i think we should just make a working cm7. why bother maintaining all this other stuff |
01:04.41 | *** join/#htc-linux asnos (~Tob@unaffiliated/asnos) |
01:07.39 | d3tul3 | because right now I am the only other person that can build it, and it makes me feel special |
01:08.11 | d3tul3 | oh wait.. |
01:10.58 | d3tul3 | on a more serious note, you're probably right... |
01:11.45 | jonpry | right is not my specialty |
01:12.01 | rpierce99 | right, fast, cheap, pick 2 |
01:12.10 | jonpry | i usually deal in "works" and open to interpretation |
01:13.04 | rpierce99 | sadly that's how my dev has always been, it's fun until it meets the spec, never much cared for niceness like UI or edge cases |
01:13.31 | rpierce99 | at that point it because tedious and I stop working on it... see: getlogs |
01:13.39 | rpierce99 | s/because/becomes/ |
01:14.12 | d3tul3 | not sure what the problem is with getlogs |
01:14.20 | rpierce99 | which one |
01:14.45 | *** join/#htc-linux AstainHellbring (~AstainHel@unaffiliated/astainhellbring) |
01:15.01 | *** join/#htc-linux mitsutaka (~mitsutaka@219.143.36.82) |
01:16.06 | jonpry | some times i get so discusted with the solution that i just refuse to implement it on principle |
01:17.35 | jonpry | getlogs looks like a nice tool |
01:18.05 | rpierce99 | it serves it's purpose, but it's very buggy and ugly |
01:18.33 | d3tul3 | i guess i've yet to hit a bug on it, or perhaps i just don't notice |
01:18.55 | rpierce99 | add a bunch of custom filtered words and then try to hit the pastebin button |
01:19.13 | rpierce99 | or flip it to landscape and add just like 1 or two custom words |
01:19.26 | d3tul3 | i've never used the filter :) |
01:19.31 | d3tul3 | too much work |
01:20.36 | jonpry | should remove that. whats a couple of esn's between you and the internet |
01:22.20 | *** join/#htc-linux furtardo (~mks@nat/yahoo/x-qjqrmlefsufbjtii) |
01:23.11 | *** join/#htc-linux mitsutaka (~mitsutaka@219.143.36.82) |
01:25.47 | WisTilt2 | arrrghhh/rpierce99: kernel is up. give it a run through and see if anything crashes it. should have 200+ MB free and curious if desktop is smooth and things open quick. Heading home so will check in later. |
01:26.52 | *** join/#htc-linux defendthecommons (~defendthe@50-48-20-8.drr05.mdtw.ny.frontiernet.net) |
01:31.01 | *** join/#htc-linux hardwalker (~hardwalke@122-117-115-146.HINET-IP.hinet.net) |
01:33.00 | *** join/#htc-linux programmer8922 (~Evan@c-174-56-72-235.hsd1.nm.comcast.net) |
01:37.00 | *** join/#htc-linux Cass (~Cass@188-220-34-222.zone11.bethere.co.uk) |
01:38.10 | d3tul3 | well commenting out do_force_revoke didn't do much |
01:39.25 | rpierce99 | free says mem total is 200696 |
01:39.39 | rpierce99 | i can never remember what baseline is for that |
01:40.36 | d3tul3 | try fp2sd |
01:41.26 | rpierce99 | hm, SPL says 96 + 15 for itself "free" after force killing everything |
01:41.56 | rpierce99 | haha fps2d is bouncing all over the place |
01:42.02 | jonpry | d3tul3, it stopped the crashing for me. just the screen is blank on wake |
01:42.19 | d3tul3 | wait i completely commented out do_force_wake |
01:42.23 | d3tul3 | and it still restarts everything |
01:42.33 | d3tul3 | do_force_revoke |
01:42.44 | jonpry | +//register_early_suspend(&info->early_suspend); |
01:42.50 | jonpry | ftw |
01:43.05 | rpierce99 | hm, much more stable this time |
01:43.24 | rpierce99 | 30, stdev 4.2 |
01:43.51 | d3tul3 | damnit i even have most of early_suspend commented out |
01:43.52 | jonpry | i can't figure out what this calls. mDisplayEventThread->acquireScreen(); |
01:44.37 | jonpry | i have usb plugged in to make sure it doesn't actually do any real suspending |
01:45.27 | rpierce99 | 467 score in quadrant |
01:45.44 | rpierce99 | 963/229/888/71/183 |
01:47.11 | rpierce99 | 23.4 FPS |
01:47.40 | rpierce99 | camera works |
01:50.34 | rpierce99 | maybe it's just because I was on the slow kernel for a few days but it feels very snappy navigating around the OS and in apps |
02:18.41 | snadge | quick question about a2dp which someone in here might be familiar with |
02:18.57 | snadge | in CM7.. its called liba2dp.. and confusingly, theres a proprietary lib with the same name.. that afaik, is not in use |
02:19.24 | snadge | in ICS, its called audio.a2dp.default |
02:20.08 | snadge | does anyone have a2dp working in an ICS build? |
02:32.30 | Cass | you compiled ics ? |
02:35.47 | snadge | yeah |
02:35.58 | snadge | a few of us are working on htc visionn |
02:36.06 | snadge | andromadus port on xda |
02:36.21 | snadge | its CM9 based.. public git |
02:36.41 | snadge | we're not trying to be show ponies.. and compete with other roms and devices etc |
02:37.17 | snadge | unfortunately.. some people are doing that :/ |
02:37.48 | snadge | its frustrating that we're having to replicate work thats already been done.. and put into other roms |
02:37.54 | snadge | simply because they wont share |
02:38.44 | snadge | and of course.. when we fix stuff.. they're more than happy to pinch those fixes from us etc |
02:38.53 | snadge | now i know how CM devs must feel |
02:43.05 | *** join/#htc-linux WisTilt2 (~wisgreg@wireless248.wirelesstcp.net) |
02:44.27 | WisTilt2 | rpierce99: im home. get a chance to run that kernel yet? i'll be around several hours so just ping me. |
02:44.58 | rpierce99 | pm, so you don't have to crawl logs |
02:45.32 | WisTilt2 | i dont do logs:) nice numbers. no crashes after running all that stuff? |
02:45.42 | rpierce99 | nope haven't seen any issues |
02:46.00 | rpierce99 | beyond a lockup on the first kexec boot, a full reboot didn't have any issues |
02:46.38 | WisTilt2 | should feel pretty quick. fps isnt going to be any different. this kernel basically recovered 24mb of memory we can use so should speed things up somewhat. |
02:46.53 | rpierce99 | my memory numbers don't seem that high |
02:46.54 | WisTilt2 | yeah i get that at times with kexec, freeze partially through boot |
02:47.06 | rpierce99 | free says mem total is 200696 |
02:47.08 | WisTilt2 | 230+ is good not being oc'd |
02:47.16 | rpierce99 | SPL says 96 + 15 for itself "free" after force killing everything |
02:48.18 | WisTilt2 | thats about right. normal kernel will be around 173-4 free. i have more to do, just wanted to test this in phases so we can catch anything breaking step by step |
02:49.00 | WisTilt2 | there is no memory or gpu oc'ing with this btw so you might try acpu oc to 672 again. should pick up a bit |
02:49.39 | rpierce99 | yeah not really worried about it unless you need that tested, not much of an ocer |
02:50.28 | WisTilt2 | have you had any incoming calls with this yet? im seeing things wake up much faster than normal. |
02:50.39 | rpierce99 | hm, no but i can test that quick |
02:50.51 | WisTilt2 | put it to sleep and see |
02:51.37 | WisTilt2 | screen comes up to lock screen pretty quick. might be a fluke, i only had 2 calls come in but both were fast wake |
02:52.27 | rpierce99 | 30 seconds from the time i hit send on gvoice to the time i could answer on the phone |
02:52.36 | rpierce99 | so no, not any better on that try |
02:53.01 | WisTilt2 | well, once the ringtone starts how long before screen comes up? |
02:53.04 | rpierce99 | vibrated a good 10 or more times, showed lock screen for like 5 seconds, before showing the swipe to answer |
02:54.05 | WisTilt2 | that seems like a long time. is that how long you normally have to go before you can answer? |
02:54.10 | rpierce99 | yeah |
02:54.15 | rpierce99 | it seemed about normal |
02:54.59 | WisTilt2 | wow, i usually get the swipe screen within a couple seconds. probably something different on the cdma side of the ril |
02:55.26 | snadge | im assuming none of you know why liba2dp.so has moved to hw/audio.a2dp.default ? |
02:55.37 | rpierce99 | yeah, i don't know about you gsm folk but pretty much all of the cdma folks miss calls entirely waiting for phone app to wake |
02:55.41 | WisTilt2 | so as far as moving between desktops and opening app menu does it seems snappier? |
02:55.53 | snadge | or more to the point.. how you're supposed to link libaudio with a2dp, now that its moved into the hw folder.. the linker cant find it |
02:55.55 | rpierce99 | yeah, the desktop feels really smooth, apps seem to open quicker |
02:57.03 | WisTilt2 | ill make up another kernel with memory oc'd and we can compare. im going to finish tying in the rest of the smi memory first so if you're going to be around a bit ill ping you with another kernel when ready. |
02:57.09 | rpierce99 | but again i was on the slow kernel for days, so it might be relative |
02:58.12 | WisTilt2 | actually, before i do that let me make a quick test for you to try. i need to see if camera will work on cdma devices like this. give me a few mins |
03:01.36 | snadge | ok, so you guys are busy working on kernel related stuff.. and im assuming a2dp is wildly offtopic atm.. or that you dont know |
03:02.12 | rpierce99 | snadge: yeah someone will answer you if they have a clule, it would be dumb if 69 people were all like "no idea dude" |
03:02.21 | rpierce99 | clue* |
03:03.58 | WisTilt2 | rpierce99: kernel is up |
03:04.15 | jonpry | we can't even get bt to work on gb |
03:12.27 | rpierce99 | WisTilt2: camera app opens, shows preview, hangs on shutter |
03:13.26 | WisTilt2 | ok guess that answers that. why cant gsm/cdma be the same. htc idiots |
03:13.46 | rpierce99 | want a log? |
03:14.01 | WisTilt2 | no, i know why it wont work this way, just irritating |
03:14.15 | WisTilt2 | you mind one more test on camera? |
03:14.36 | rpierce99 | sure |
03:16.28 | WisTilt2 | ok give this one a try |
03:23.12 | WisTilt2 | jonpry: can ashmem create regions that are not contiguous? |
03:23.30 | jonpry | it does |
03:23.55 | rpierce99 | WisTilt2: same issue |
03:24.28 | WisTilt2 | so in the virtual space it appears contiguous once it creates the region? |
03:24.31 | rpierce99 | #123 |
03:24.46 | WisTilt2 | ok let me try one more then we'll move on |
03:24.54 | jonpry | yes only contiguous in virt |
03:24.58 | jonpry | uses vmalloc |
03:26.11 | WisTilt2 | rpierce99 last camera test kernel up |
03:27.14 | jonpry | but we can make copybit work with ashmem |
03:27.26 | WisTilt2 | jonpry thanks. doing some brainstorming |
03:28.04 | WisTilt2 | we need to do that. i now have another 24mb available and if we can get copybit into ashmem that will free up that other 32mb pmem |
03:28.16 | *** join/#htc-linux d3tul3 (~detule@pool-96-234-141-27.bltmmd.east.verizon.net) |
03:29.01 | d3tul3 | how's the testing coming along |
03:29.23 | jonpry | blank is wrecking my world |
03:29.59 | d3tul3 | WisTilt2, perhaps on one of these kernels you can leave the camera business on status quo and we can have rpierce99 try out a userland lib |
03:30.31 | WisTilt2 | fixed userland for the overlap? |
03:30.45 | d3tul3 | i think so |
03:31.08 | WisTilt2 | after this test ill add pmem check back in and he can give it a whirl |
03:31.49 | d3tul3 | it would be nice if we can leave the pmem check, as tasks other than JPEG* register adsp memory...not that i expect anything else will blow up |
03:32.14 | WisTilt2 | i agree. hate hacking the kernel to fix userland |
03:32.38 | rpierce99 | hm, somehow i'm still on #123 |
03:33.56 | WisTilt2 | bfd33e5f4f0b51b46eaf770905b37b36 |
03:43.56 | rpierce99 | ok #124 is the same |
03:46.02 | WisTilt2 | ok. uploaded the working camera one but with pmem check back on to try d3tul3's userland fix |
03:46.16 | d3tul3 | http://dl.dropbox.com/u/38520332/libcamera.so |
03:47.00 | d3tul3 | you might even try just pure autobuild |
03:48.58 | rpierce99 | what's the path to libcamera normally |
03:49.50 | d3tul3 | system/lib |
03:51.51 | jonpry | i got something that appears to let the screen be off |
03:52.07 | d3tul3 | kernel or userland |
03:52.24 | jonpry | both. but i might be able to revert the userland stuff |
03:52.45 | jonpry | er kernel stuff |
03:53.53 | jonpry | will anyone really want to use this? giving up 3d apps? |
03:58.20 | rpierce99 | it's taking pictures, but there are some weird artifacts in the picture |
03:58.24 | rpierce99 | some pictures |
03:58.43 | d3tul3 | which kernel are you on |
03:58.44 | rpierce99 | and it showed up in the preview window at one point |
03:58.51 | rpierce99 | the one wis posted, i'll try ab |
04:00.34 | d3tul3 | well i certainly messed with the little thumbnail view |
04:08.12 | *** join/#htc-linux SG (~SG@cpe-76-186-49-202.tx.res.rr.com) |
04:09.26 | rpierce99 | took about 10 pictures on ab and couldn't get the issue to pop up |
04:10.21 | d3tul3 | cool, this is all somewhat academic as i doubt this patch will likely ever make it to userland |
04:10.23 | WisTilt2 | probably the memory cluster ive got going on and camera is within it |
04:10.29 | jonpry | d3tul3, only kernel change is that register_early_suspend |
04:12.20 | d3tul3 | alright i am with you |
04:12.33 | jonpry | sometimes when it wakes up. screen is blank and i have to do something that would cause a redraw |
04:12.38 | jonpry | might only be the first wake |
04:12.58 | jonpry | other code is pushed |
04:14.37 | d3tul3 | oh nice on current.xml this might actually apply clean |
04:15.05 | jonpry | i need to untrack that file |
04:16.08 | jonpry | rings unlocker is so hip |
04:16.56 | jonpry | we can turn on tablet tweaks to get a soft menu button |
04:19.32 | jonpry | looks like this thing is hitting some kind of low power state |
04:20.31 | *** join/#htc-linux [acl] (~abel@cpe-69-203-141-229.si.res.rr.com) |
04:22.59 | *** join/#htc-linux Rob2222 (~Miranda@p4FFF1B9F.dip.t-dialin.net) |
04:28.44 | jonpry | my build is apparently to jenky to allow installation of fps2d |
04:29.14 | d3tul3 | jonpry, well done |
04:29.21 | d3tul3 | survived first screen off |
04:29.33 | jonpry | cooel |
04:29.46 | d3tul3 | i got drawing stuff in logcat past screen off |
04:30.03 | jonpry | me too |
04:30.10 | d3tul3 | would this manage to actually suspend |
04:30.18 | jonpry | seems to for me |
04:30.30 | jonpry | i have no idea how thats possible |
04:31.22 | d3tul3 | right you are |
04:32.22 | jonpry | winmo performance? |
04:32.43 | d3tul3 | lol, you know i delete data.img and i am practically at first boot every time i try this |
04:33.39 | d3tul3 | but now that it survives suspend i might actually run with it for a day |
04:33.53 | jonpry | that would be interesting |
04:34.41 | jonpry | rendering quality is pretty good |
04:35.57 | d3tul3 | is it possible that lowmemorykiller would kill surfaceflinger |
04:36.35 | jonpry | i don't think so |
04:36.43 | d3tul3 | because every single time when i try to open market on first boot i get a blank screen and a reset |
04:37.11 | jonpry | bold |
04:38.12 | jonpry | i'm throwing in the towel for the day |
04:38.35 | d3tul3 | good night i am out as well |
05:38.46 | *** join/#htc-linux lardman|home (~lardman@138-38-179-9.docking.bath.ac.uk) |
05:39.08 | *** join/#htc-linux asnos (~Tob@unaffiliated/asnos) |
06:02.14 | *** join/#htc-linux XirXes (~XirXes@67-2-13-6.slkc.qwest.net) |
06:17.55 | *** join/#htc-linux sykosoft (~sykosoft@67-61-60-19.cpe.cableone.net) |
06:29.35 | *** join/#htc-linux programmer8922 (~Evan@c-174-56-72-235.hsd1.nm.comcast.net) |
06:46.47 | *** join/#htc-linux Mavy (mavfree@unaffiliated/mavy) |
07:05.15 | *** join/#htc-linux kiozen (~kiozen@p54BB6AFE.dip.t-dialin.net) |
07:17.03 | *** join/#htc-linux dobrin (~dobrin@85.91.150.26) |
07:32.18 | *** join/#htc-linux XeKToReX- (~xTx@pentac.lnk.telstra.net) |
07:38.40 | *** join/#htc-linux XeKToReX (~xTx@CPE-124-188-18-211.hdbq1.win.bigpond.net.au) |
07:41.40 | *** join/#htc-linux lamikr (lamikr@nat/nokia/x-awmapkfbvnplekvs) |
07:45.48 | *** join/#htc-linux kiozen (~kiozen@p54BB6AFE.dip.t-dialin.net) |
08:08.10 | *** join/#htc-linux FlawlesStyle (~LOL@c-98-253-109-196.hsd1.il.comcast.net) |
08:08.10 | *** join/#htc-linux FlawlesStyle (~LOL@unaffiliated/flawlesstyle) |
08:21.28 | *** join/#htc-linux ychavan (ychavan@nat/redhat/x-mnsuzzkvvhwcwfkb) |
08:22.46 | *** join/#htc-linux XirXes_ (~XirXes@67-2-21-203.slkc.qwest.net) |
09:53.21 | *** join/#htc-linux GNUtoo (~gnutoo@host1-132-dynamic.51-79-r.retail.telecomitalia.it) |
09:56.18 | *** join/#htc-linux rajkosto (~rajkosto@2001:470:d76b:da7a:197e:f5e2:6854:b1) |
10:17.38 | *** join/#htc-linux rajkosto (~rajkosto@2001:470:d76b:da7a:4d90:7425:96a5:3b58) |
10:20.28 | *** join/#htc-linux LjL (~ljl@unaffiliated/ljl) |
10:59.11 | *** join/#htc-linux bartman (~bart@2607:f2c0:a000:175:2e0:81ff:fe47:3d01) |
11:39.28 | *** join/#htc-linux LjL (~ljl@unaffiliated/ljl) |
12:56.36 | *** join/#htc-linux mitsutaka (~mitsutaka@123.112.97.210) |
13:47.37 | *** join/#htc-linux Andreyxxl1 (~Andreyxxl@94.52.236.48) |
14:32.25 | *** join/#htc-linux asnos (~Tob@unaffiliated/asnos) |
14:53.12 | *** join/#htc-linux gauner1986 (~Miranda@ip-178-201-214-71.unitymediagroup.de) |
15:58.01 | *** join/#htc-linux emwe__ (~mweirauch@cable-86-56-10-252.cust.telecolumbus.net) |
15:59.53 | *** join/#htc-linux rpierce99 (~rpierce99@96-42-102-103.dhcp.stcd.mn.charter.com) |
16:04.05 | *** join/#htc-linux arrrghhh (~arrrghhh@unaffiliated/arrrghhh) |
16:13.34 | *** join/#htc-linux WisTilt2 (~wisgreg@wireless251.wirelesstcp.net) |
16:48.20 | *** join/#htc-linux AstainHellbring (~AstainHel@206.71.78.158) |
16:48.23 | *** join/#htc-linux AstainHellbring (~AstainHel@unaffiliated/astainhellbring) |
16:51.26 | arrrghhh | WisTilt2, how goes it |
16:51.41 | arrrghhh | trolling the logs, interesting concept that the RIL is causing issues with waking the phone for inbound calls |
16:52.01 | arrrghhh | i always assumed it was the bag-o-ass Phone.apk... but since the RIL was rewritten for CDMA by someone without a CDMA device... |
16:52.05 | *** join/#htc-linux kiozen (~kiozen@ppp-93-104-73-46.dynamic.mnet-online.de) |
16:53.51 | WisTilt2 | hi arrrghhh, trying to get my coffee iv going. got a whole 2hrs sleep last night so dragging bad today |
16:53.58 | arrrghhh | damn |
16:54.08 | arrrghhh | you're the boss, can't you just go home and go back to sleep? :P |
16:54.27 | arrrghhh | this is probably one of the many reasons why i'm not the boss. |
16:54.49 | WisTilt2 | yeah thought about it but now im awake and going to just stay undisturbed in my office working on this stuff:) |
16:55.18 | WisTilt2 | everyone stays clear of me when they know i didnt get much sleep lol |
16:55.54 | *** join/#htc-linux skodde (~skodde@unaffiliated/skodde) |
16:57.22 | WisTilt2 | im not 100% sure on the ril being the prob but im sticking with that until proven otherwise. on gsm it wakes up quickly everytime. |
16:58.31 | rpierce99 | arrrghhh: I didn't mischaracterize your experiences with the phone waking for calls did I? |
17:00.29 | arrrghhh | nope |
17:00.38 | arrrghhh | exact same problem |
17:00.52 | arrrghhh | phone will usually ring/vibe for at least 10 secs before the screen wakes |
17:01.00 | arrrghhh | then typically i just see a lock screen, no answer slider |
17:01.24 | arrrghhh | and every once in a while it will come up and display the answer slider - but usually i miss the call entirely. |
17:01.43 | arrrghhh | WisTilt2, seems suspect considering it's fine on GSM |
17:03.13 | WisTilt2 | i get dozens of calls a day and longest i think ive ever had to wait was maybe 3sec before screen came on. |
17:03.55 | arrrghhh | yea, the only time i get to answer calls is when the screen is on. |
17:04.05 | arrrghhh | i pretty much miss every other call... |
17:04.17 | *** join/#htc-linux kiozen (~kiozen@ppp-93-104-73-46.dynamic.mnet-online.de) |
17:05.17 | WisTilt2 | some time if you could get me both dmesg, logcat, and logcat radio after an incoming call, i can take a look if it is indeed the ril |
17:05.30 | arrrghhh | sure |
17:05.37 | arrrghhh | any special setup ? |
17:05.48 | WisTilt2 | no hurry, im in the middle of trying to finish this ashmem for mdp this morning |
17:06.02 | arrrghhh | np |
17:06.03 | WisTilt2 | no, just normal with it in sleep before you call it |
17:06.12 | arrrghhh | sounds good. |
17:06.22 | arrrghhh | i'm on NAND right now, and in the middle of some production issues @ work.. |
17:06.27 | arrrghhh | so i won't be able to for a bit either ;) |
17:14.01 | detule | that's strange i don't usually see that issue |
17:15.06 | detule | then again i don't get many calls either... |
17:17.51 | jonpry | WisTilt2, ashmem for mdp? |
17:18.09 | WisTilt2 | mdp_ppp |
17:18.27 | jonpry | cool. what's the plan? |
17:18.59 | rpierce99 | i can get a set of logs for that via getlogs in the next hour or so |
17:22.09 | WisTilt2 | jonpry: i have it all coded except a couple things. dont know if this is going to work but it should. first attempt gave me interesting background colors from the 60's but icons overlaid looked fine. point problem obviously so thats what im working on now. |
17:23.25 | jonpry | sounds like a good stuff |
17:24.54 | WisTilt2 | looks like im not getting the correct pointer to the vm_file. i can see that this is going to be faster than pmem though when its done |
17:25.52 | jonpry | that sounds suspect. why faster? |
17:26.34 | WisTilt2 | maybe not, just seemed like switching desktops was faster |
17:26.48 | WisTilt2 | hard to tell with all the lsd colors on the screen |
17:26.56 | jonpry | heh |
17:27.26 | WisTilt2 | main goal of course is the gain that extra 32mb pmem |
17:27.50 | WisTilt2 | is the? jeeze, is to |
17:31.35 | jonpry | if you need a hand let me know |
17:45.42 | *** join/#htc-linux AstainHellbring (~AstainHel@unaffiliated/astainhellbring) |
17:47.03 | rpierce99 | WisTilt2: logs in PM |
17:47.19 | rpierce99 | i'm trying to track down the start of the call and the answer |
17:47.45 | rpierce99 | it started vibrating 11 seconds after I hit send on my phone, and i was able to answer it 12 seconds after that |
17:48.06 | WisTilt2 | rpierce99: thanks. i will take a look after i get this mdp working. |
17:48.11 | rpierce99 | np |
17:50.52 | arrrghhh | detule, you said you don't have the problem? |
17:51.05 | arrrghhh | i don't get many calls, but damn. every single call i do get is painful. |
17:51.14 | arrrghhh | my gf has just gotten in the habit of not calling me, lol. |
17:51.20 | detule | so wait what happens |
17:51.22 | arrrghhh | WisTilt2, perhaps you shouldn't fix it... :P |
17:51.23 | rpierce99 | I start to see acoustic stuff about 11:38.41, but the phone phone thing i get is 12-13 11:38:54.790 I/phone ( 1549): acceptCall: incoming... |
17:51.35 | arrrghhh | detule, just takes forever for the phone to wake/display answer swipe screen on incoming calls. |
17:51.46 | detule | this happens on the autobuild kernels? |
17:51.50 | arrrghhh | any kernel |
17:51.54 | rpierce99 | even .27 |
17:51.55 | arrrghhh | i haven't found one where it doesn't happen. |
17:52.28 | rpierce99 | 12-13 11:38:44.189 I/ActivityManager( 1464): Starting: Intent { act=android.intent.action.MAIN flg=0x10840000 cmp=com.android.phone/.InCallScreen } from pid -1 |
17:52.29 | detule | just woke up instantly over here |
17:52.37 | rpierce99 | detule: rhod400 right? |
17:52.40 | detule | yes |
17:52.45 | rpierce99 | hm, google voice setup? |
17:52.51 | detule | no |
17:52.55 | rpierce99 | hm, arrrghhh and i do |
17:53.20 | arrrghhh | bleh |
17:53.25 | arrrghhh | that would suck if this was all GV |
17:53.34 | arrrghhh | originally that was the reason i wanted Android. native GV app :P |
17:53.41 | detule | good troubleshooting rpierce99 |
17:54.33 | rpierce99 | 12-13 11:38:54.890 D/AudioHardwareInterface( 1437): setMode(RINGTONE) |
17:54.33 | rpierce99 | 12-13 11:38:54.910 D/InCallTouchUi( 1549): updateState: Too soon after last action; not drawing! |
17:54.51 | detule | well i usually get no ringer on my phone |
17:55.21 | detule | my phone is special |
17:56.45 | rpierce99 | i don't see any logcat references to google voice, so unless it is silent or affects another app just by being installed, i'm not sure how that would relate |
17:56.54 | detule | jonpry, i feel like sometimes certain parts of the ui lose context and i get snowed out lockscreen and snowed out taskbar pull-down....and they don't quite recover on screen off/on |
17:57.29 | rpierce99 | detule: I pmed you the log since you seem interested |
17:57.33 | jonpry | pretty sure thats caused by running out of memory |
18:01.09 | *** join/#htc-linux GNUtoo (~gnutoo@host1-132-dynamic.51-79-r.retail.telecomitalia.it) |
18:03.45 | jonpry | detule, need that 2d allocator to get more space :) |
18:07.55 | detule | are you doing anything with the pmem pool right now? |
18:08.01 | jonpry | no |
18:08.14 | detule | so it's just up for grabs :) |
18:08.32 | jonpry | i guess |
18:12.28 | *** join/#htc-linux |Jeroen| (~jeroen@d5152B25B.access.telenet.be) |
18:12.31 | detule | I/TelephonyRegistry( 1464): notifyDataConnection: state=3 isDataConnectivityPossible=true reason=2GVoiceCallStarted interfaceName=ppp0 networkType=7 |
18:12.54 | detule | GVoiceCall is certainly doing something during an incoming call |
18:13.30 | rpierce99 | what's the timestamp on that, rilj gets the call_ring at 11:38:39.495 |
18:13.59 | detule | 12-13 11:38:40.246 |
18:18.11 | rpierce99 | hm, then it takes 4 seconds after that for android to kick off the inCallScreen intent |
18:19.04 | rpierce99 | And then 10 seconds from inCallScreen to AudioHardwareInterface setMode(RINGTONE) |
18:19.30 | detule | not so worried about audiohardwareinterface |
18:19.35 | detule | as much as the incallscreen not showing up |
18:21.27 | detule | among other things this doesn't seem right 12-13 11:38:42.047 D/BT HS/HF( 1549): CDMA call state: IDLE prev state:IDLE |
18:23.09 | detule | is google voice implemented over gtalk? |
18:23.43 | rpierce99 | you can make a gvoice call from the gtalk window in gmail, but no the android apps are seperate |
18:24.21 | rpierce99 | i just uninstalled google voice, going to try again |
18:24.45 | detule | also see if you still get those xmpp messages in your log |
18:27.32 | rpierce99 | started ringing at :09, slide to answer at :19 |
18:27.44 | detule | what's 10 seconds here and there |
18:28.32 | detule | don't know try 3.1 from the autobuild that's what i am using right now, can't imagine it would make much of a difference |
18:29.11 | detule | wish i could get a log off my phone but adb is not cooperating |
18:29.27 | rpierce99 | this is .39 from autobuild, and this happens on .27 too, so i don't think it would be any diff |
18:30.12 | detule | good thing i have getlogs in my email |
18:30.22 | detule | i'll get some logs to compare |
18:30.55 | *** join/#htc-linux rob_w (~bob@unaffiliated/rob-w/x-1112029) |
18:32.22 | detule | i got the slide to unlock on the second ring as heard on the initiating end |
18:34.09 | *** join/#htc-linux mgross029 (c0234f46@gateway/web/freenode/ip.192.35.79.70) |
18:41.23 | detule | call came in 12-13 13:32:10.240 D/PHONE ( 1514): Event EVENT_CALL_RING Received state=IDLE |
18:41.33 | detule | by 32:16 i had already swiped to accept |
18:55.24 | *** join/#htc-linux rob_w_ (~bob@ppp-188-174-29-206.dynamic.mnet-online.de) |
19:23.02 | *** join/#htc-linux LordDeath (~LordDeath@cable-81-173-164-253.netcologne.de) |
19:50.22 | WisTilt2 | jonpry: incoming email |
19:50.58 | *** join/#htc-linux teemodk (~teemo@87-104-122-17-dynamic-customer.profibernet.dk) |
20:14.26 | *** join/#htc-linux kiozen (~kiozen@ppp-93-104-73-46.dynamic.mnet-online.de) |
20:15.40 | *** join/#htc-linux kiozen (~kiozen@ppp-93-104-73-46.dynamic.mnet-online.de) |
20:20.18 | WisTilt2 | Rhod400 just arrived woohoo |
20:20.22 | arrrghhh | nice |
20:20.55 | WisTilt2 | still going to need you guys with service to test though |
20:24.11 | arrrghhh | well you should be able to connect to towers |
20:24.14 | arrrghhh | and talk to the angry sprint guy |
20:24.19 | arrrghhh | but yea, you won't be able to use data. |
20:24.25 | arrrghhh | or make/receive phone calls... |
20:25.20 | *** join/#htc-linux d3tul3 (~detule@pool-96-234-141-27.bltmmd.east.verizon.net) |
20:27.19 | WisTilt2 | lunch has arrived, bbl |
20:30.16 | emwe | perhaps running the RHOD400 on gsm will show some behavioral differences to his RHOD300 aramada. |
20:30.16 | emwe | s/aramada/armada |
20:38.14 | *** join/#htc-linux programmer8922 (~Evan@esquel.math.unm.edu) |
20:39.26 | arrrghhh | armada lol |
20:39.30 | arrrghhh | that's a good way to put it. |
20:48.41 | jonpry | WisTilt2, you might be getting the right info |
20:48.50 | jonpry | small images work right? |
21:12.01 | WisTilt2 | jonpry: background image is all 60's style but icons and even fancy widget looks fine. so the way i did all this look ok to you or did i go about it the wrong way? |
21:12.27 | WisTilt2 | and i have pmem device unregistered btw |
21:13.42 | jonpry | whats done is fine. but i think your missing something |
21:14.14 | WisTilt2 | like sleep |
21:14.20 | WisTilt2 | lunch was not such a good idea |
21:14.29 | jonpry | like the mdp dma operation gets the a physical address that corresponds to the base of the first page |
21:14.45 | jonpry | but the second page is not contiguous. so its just drawing random memory to the screen |
21:15.22 | jonpry | smallish things taking up only 4kb will be fine |
21:15.49 | jonpry | but the background is like 100's of pages |
21:15.50 | WisTilt2 | yeah that could be a problem. thats pretty much what im seeing |
21:16.09 | WisTilt2 | ill just make this the 60's version of tp2 |
21:16.31 | jonpry | so the blit needs to be broken up |
21:16.55 | jonpry | cause there will always be subimages of the source that lie in contiguous memory |
21:17.46 | *** join/#htc-linux asnos (~Tob@unaffiliated/asnos) |
21:18.00 | WisTilt2 | worth pursuing or no? |
21:18.11 | jonpry | i think so |
21:18.18 | jonpry | get 32mb of ram |
21:19.32 | jonpry | why so little sleep? |
21:19.34 | WisTilt2 | probably going to need your help on this. you know that side of things far better than i do |
21:19.39 | jonpry | too much hacking? |
21:20.50 | WisTilt2 | lol, no, we had some problems with a break in out at one of our shop buildings at the ranch. |
21:21.18 | WisTilt2 | people going crazy stealing copper wire or anything that has copper in it |
21:22.09 | jonpry | ouch |
21:22.26 | jonpry | maybe they should start a copper mine |
21:23.48 | Willd | Hey, here in Sweden, they steal church roofs! |
21:23.52 | Willd | Blasphemy! |
21:23.54 | WisTilt2 | people are actually pulling out irrigation pumps and hauling them off for the copper wire in it |
21:24.16 | rpierce99 | people like to steal copper, I know because I saw it on The Wire |
21:24.33 | Willd | WisTilt2: When will they start tearing down phone lines? >_> |
21:24.35 | WisTilt2 | Willd thats sad |
21:24.46 | Willd | WisTilt2: Sure is.. |
21:25.04 | jonpry | vigilantes for fiber to the home |
21:25.16 | WisTilt2 | they actually did that. a month ago some guys in a truck hooked up some at&t underground copper lines and pulled out over a mile of it |
21:25.22 | WisTilt2 | they caught them though |
21:25.35 | Willd | Gee, handy! |
21:25.41 | Willd | jonpry: :D |
21:26.32 | toastcfh | jonpry, when u updated mdp for tp did u just grab it from caf or the msm folder |
21:26.43 | jonpry | CAF |
21:26.52 | toastcfh | as of that date? |
21:27.15 | jonpry | umm no. that commit may have been from msm non pe |
21:28.11 | toastcfh | how is the 7k progress going |
21:28.23 | toastcfh | with gl |
21:28.24 | jonpry | WisTilt2, in msm_mdp22.c there is code for doing tiled blits |
21:29.00 | jonpry | toastcfh, its working pretty well in gingerbread. haven't tried to port it to ics yet |
21:29.18 | jonpry | some crazed fellow has been running it for > 12 hours |
21:29.27 | toastcfh | heh |
21:29.35 | toastcfh | winmo ports that stable? |
21:30.11 | jonpry | regular compositor goes for weeks |
21:30.36 | jonpry | then i need new kernel |
21:32.32 | *** join/#htc-linux xirxes__ (~XirXes@67-2-14-9.slkc.qwest.net) |
21:34.25 | d3tul3 | xdandroid + gposit mhhh |
21:35.34 | WisTilt2 | jonpry i see the tiled code in mdp_ppp_do_blit but have no idea how to go about breaking the image up to call that at this point. |
21:36.47 | jonpry | just thinking aloud here. not sure of the best algorithm |
21:36.57 | jonpry | but say src image is like 1MB |
21:37.04 | jonpry | so 256 pages |
21:37.24 | jonpry | loop through from base to end, add 4k each time |
21:37.43 | jonpry | check to see if the current page is actually within the bounds of the blit |
21:38.32 | jonpry | if it is. then there is i think at most 3 rectangles that can represent the blit |
21:38.47 | jonpry | so figure out the coords and effect the operation |
21:39.16 | WisTilt2 | so this would be done in process_blit i take it? |
21:39.37 | jonpry | i think mdp_ppp_do_blit |
21:39.42 | jonpry | but it doesnt really matter |
21:42.41 | WisTilt2 | this looks like it kind of already does that in some sense, just needs some additional pre tests/changes before blit and wait |
21:44.06 | *** join/#htc-linux Rob2222 (~Miranda@p4FFF1B9F.dip.t-dialin.net) |
21:44.12 | WisTilt2 | so you think my vm pointer is getting the right info? i wont keep chasing that then. thats where i thought it was falling apart. |
21:44.26 | jonpry | yes |
21:44.45 | WisTilt2 | i guess none of the icons would be showing up if it was wrong |
21:44.58 | jonpry | but once you get the blits confined to a page. there is an another problem |
21:45.51 | jonpry | like process_blit and such configure the mdp registers using the base of the first page plus the offset. but the vm really needs to be queried for the base of the page in question |
21:46.52 | jonpry | i don't think the mdp_ppp_do_blit() code is quite right. but it could be adapted |
21:49.53 | WisTilt2 | i'll mess with it for a bit and see what kind of trouble i can get this thing into |
21:54.24 | *** join/#htc-linux asnos (~Tob@unaffiliated/asnos) |
22:15.17 | *** join/#htc-linux LordDeath (~LordDeath@cable-81-173-164-253.netcologne.de) |
22:20.00 | *** join/#htc-linux BabelO_ (~fcr@AMontpellier-553-1-150-152.w86-201.abo.wanadoo.fr) |
22:20.01 | *** join/#htc-linux BabelO_ (~fcr@unaffiliated/babelo) |
22:24.29 | *** join/#htc-linux LjL (~ljl@unaffiliated/ljl) |
22:30.23 | WisTilt2 | jonpry: thought you'd get a laugh from this 1st blit remodel attempt -> |
22:30.24 | WisTilt2 | [ 171.006408] mdp_ppp: final src or dst location is invalid, are you trying to make an image too large or to place it outside the screen? |
22:30.27 | WisTilt2 | [ 171.010467] [drivers/video/msm/mdp_ppp.c:blit_scale:337] crap, bad scale |
22:35.57 | *** join/#htc-linux Alex[sp3dev] (~alexander@178.176.20.181) |
22:40.16 | *** part/#htc-linux Alex[sp3dev] (~alexander@178.176.20.181) |
22:40.55 | *** join/#htc-linux vinceweis (4463ed13@gateway/web/freenode/ip.68.99.237.19) |
22:41.29 | *** join/#htc-linux NeoMatrixJR (~chatzilla@173-20-63-62.client.mchsi.com) |
22:43.47 | jonpry | WisTilt2, not too good :( |
22:44.10 | WisTilt2 | just testing the waters |
22:45.57 | arrrghhh | WisTilt2, lol |
22:46.06 | arrrghhh | i love those easter egg error messages :P |
22:46.45 | WisTilt2 | crap, bad scale lol. i need to stick some of those in my commits |
22:55.59 | *** join/#htc-linux d3tul3 (~detule@pool-96-234-141-27.bltmmd.east.verizon.net) |
23:05.08 | jonpry | when you hit one of those you have to be hot on the trail |
23:10.33 | WisTilt2 | jonpry, i have ignition. should ashmem be slower than pmem by a little bit though? |
23:11.22 | jonpry | maybe a little because of all this processing in mdp |
23:11.57 | WisTilt2 | ok thats probably it. its much faster than before i added the get ashmem file stuff but slower than pmem, not too much |
23:12.34 | WisTilt2 | everything looks good though but i do get an occasional blip of junk in random places for a split sec when i change desktops |
23:12.57 | jonpry | probably need to flush caches |
23:13.53 | WisTilt2 | yeah i didnt add the ashmem caches flush stuff yet. i am also getting invalid src/dst location in dmesg at times so still needs fine tuning |
23:14.21 | jonpry | impressive though |
23:14.23 | WisTilt2 | but its working and i have that extra 32mb available:) |
23:14.34 | WisTilt2 | no /dev/pmem anymore |
23:15.05 | jonpry | so like 232 total mem? |
23:15.29 | WisTilt2 | rebooting with another change, ill get mem shortly |
23:20.55 | *** join/#htc-linux programmer8922 (~Evan@esquel.math.unm.edu) |
23:25.04 | *** join/#htc-linux Rob2222 (~Miranda@p4FFF1B9F.dip.t-dialin.net) |
23:25.08 | *** join/#htc-linux NeoMatrixJR (~chatzilla@173-20-63-62.client.mchsi.com) |
23:28.04 | WisTilt2 | jonpry: im doing this on my other .39 tree that doesn't have that other memory magic stuff. total memory is 208824 and free after boot is 117mb |
23:28.41 | jonpry | <PROTECTED> |
23:29.40 | WisTilt2 | i need to track down where these occasional blips are coming from though |
23:29.41 | jonpry | i like this ashmem thing. it will be nice for fallbacks with compositor |
23:30.21 | jonpry | want to show me the algorithm? |
23:30.37 | WisTilt2 | im going to add the cache stuff first and see if that clears it up |
23:30.46 | WisTilt2 | while my brain is still somewhat working |
23:30.46 | jonpry | ok |
23:30.58 | WisTilt2 | i am really dragging |
23:31.21 | jonpry | <PROTECTED> |
23:32.02 | WisTilt2 | already had 1.5 pots of coffee and its starting to wear off |
23:32.57 | jonpry | <PROTECTED> |
23:34.15 | WisTilt2 | me too but that was decades ago when i was younger:) |
23:35.06 | arrrghhh | WisTilt2, that's awesome news. |
23:36.12 | WisTilt2 | coming down the home stretch anyway. |
23:36.24 | arrrghhh | indeed |
23:38.42 | WisTilt2 | jonpry: we need asm/cacheflush.h for this right? |
23:39.47 | jonpry | not so sure. i have the same problem. but there is some pmem ioctl |
23:40.32 | jonpry | dmac_flush_range(vaddr, vaddr + pmem_len(id, data)); |
23:40.52 | WisTilt2 | where is that? |
23:41.14 | jonpry | thats how its done in pmem. looks like as/cacheflush.h |
23:41.38 | WisTilt2 | yeah it looks like i need that header file |
23:49.07 | *** join/#htc-linux aeroevan (~aeroevan@spoon.lugatgt.org) |
23:58.31 | *** join/#htc-linux swc|666 (~gecko@unaffiliated/swc666/x-4934821) |