00:09.36 | *** join/#htc-linux ali1234 (~ajbuxton@robotfuzz.co.uk) |
00:15.06 | *** join/#htc-linux mitsutaka (~mitsutaka@rt.miraclelinux.com) |
00:36.06 | *** join/#htc-linux phh (~quassel@quassel.rezel.enst.fr) |
00:52.30 | *** join/#htc-linux AstainHellbring (AstainHell@unaffiliated/astainhellbring) |
01:02.48 | *** join/#htc-linux furtardo (~mks@nat/yahoo/x-gtwtxbjkhdcyklsb) |
01:05.13 | *** join/#htc-linux BabelO (~fcr@AMontpellier-553-1-104-14.w86-209.abo.wanadoo.fr) |
01:05.13 | *** join/#htc-linux BabelO (~fcr@unaffiliated/babelo) |
01:21.32 | *** join/#htc-linux d3tul3 (~detule@pool-96-234-141-27.bltmmd.east.verizon.net) |
01:30.24 | *** join/#htc-linux hardwalker (~hardwalke@122-117-115-146.HINET-IP.hinet.net) |
01:31.52 | *** join/#htc-linux surge (surge@pool-72-88-82-28.bflony.fios.verizon.net) |
01:57.00 | *** join/#htc-linux mitsutaka (~mitsutaka@rt.miraclelinux.com) |
02:17.45 | *** join/#htc-linux surge (surge@pool-72-88-82-28.bflony.fios.verizon.net) |
02:24.57 | *** join/#htc-linux arif-ali_ (~arif-ali@94-192-24-56.zone6.bethere.co.uk) |
02:27.49 | *** join/#htc-linux WisTilt2 (~wisgreg@wireless248.wirelesstcp.net) |
02:44.37 | *** join/#htc-linux raymonddull (~raymonddu@c-98-209-145-76.hsd1.mi.comcast.net) |
03:01.58 | *** join/#htc-linux NeoMatrixJR (~chatzilla@173-20-63-62.client.mchsi.com) |
03:02.00 | *** join/#htc-linux toastcfh (~toastcfh@41-32.200-68.tampabay.res.rr.com) |
03:02.21 | *** join/#htc-linux toastcfh (~toastcfh@unaffiliated/toastcfh) |
03:28.08 | *** join/#htc-linux toastcfh (~toastcfh@41-32.200-68.tampabay.res.rr.com) |
03:28.49 | *** join/#htc-linux toastcfh (~toastcfh@unaffiliated/toastcfh) |
03:51.11 | *** join/#htc-linux mitsutak_ (~mitsutaka@rt.miraclelinux.com) |
03:56.02 | *** join/#htc-linux NeoMatrixJR (~chatzilla@173-20-63-62.client.mchsi.com) |
03:59.10 | *** join/#htc-linux AstainHellbring (AstainHell@unaffiliated/astainhellbring) |
04:03.56 | *** join/#htc-linux Rob2222 (~Miranda@p4FFF10B0.dip.t-dialin.net) |
04:08.54 | *** join/#htc-linux Bry8Star (~Bry8Star@gateway/tor-sasl/bry8star) |
04:32.58 | *** join/#htc-linux Entropy512 (~quassel@cpe-69-205-145-233.stny.res.rr.com) |
04:37.23 | *** join/#htc-linux TheXev (~user@177.sub-174-252-212.myvzw.com) |
05:08.56 | jonpry | dur |
05:47.38 | *** join/#htc-linux kiozen (~kiozen@ppp-93-104-92-93.dynamic.mnet-online.de) |
05:54.26 | WisTilt2 | jonpry ping ping |
05:56.38 | jonpry | hi WisTilt2 |
05:57.06 | WisTilt2 | hey want you to look at this and see if you notice anything different before i hit the sack -> |
05:57.08 | WisTilt2 | [ 0.000000] On node 0 totalpages: 51200 |
05:57.09 | WisTilt2 | [ 0.000000] free_area_init_node: node 0, pgdat c048e79c, node_mem_map c0549000 |
05:57.09 | WisTilt2 | [ 0.000000] Normal zone: 716 pages used for memmap |
05:57.09 | WisTilt2 | [ 0.000000] Normal zone: 0 pages reserved |
05:57.09 | WisTilt2 | [ 0.000000] Normal zone: 50484 pages, LIFO batch:15 |
05:57.09 | WisTilt2 | [ 0.000000] Memory: 193460k/193460k available, 11340k reserved, 0K highmem |
05:57.50 | jonpry | not really |
05:59.10 | jonpry | is it broken? |
05:59.18 | WisTilt2 | ? ok. unlocked the rest of smi and gained 32mb usable. not done with it but doing a major rewrite on pmem |
05:59.25 | WisTilt2 | working like a charm |
05:59.32 | jonpry | whoa 32mb? |
06:00.01 | WisTilt2 | yeah, if you compare that paste with normal you will see. my available memory after boot is 106.5mb |
06:00.08 | jonpry | impressive |
06:00.17 | jonpry | need that for gpu |
06:00.26 | WisTilt2 | fb in smi is by far faster too. |
06:00.55 | WisTilt2 | do you know if gpu0/1 need 8mb each? |
06:01.23 | WisTilt2 | gpu0=7 actually but wonder why they are set at 7 and 8 |
06:01.45 | jonpry | so modem just doesn't use smi2? |
06:02.11 | WisTilt2 | it uses the upper protected part but not much |
06:02.16 | jonpry | part of it is hard coded in the driver blob |
06:02.16 | jonpry | the other can probably be whatever you want |
06:02.26 | jonpry | cdma will probably blow up |
06:02.59 | WisTilt2 | yes cdma wont work with this im sure. i really need a 400 to work with |
06:03.38 | *** join/#htc-linux mitsutaka (~mitsutaka@rt.miraclelinux.com) |
06:05.26 | WisTilt2 | i need to hit the sack. ill be working on this tomorrow so ill catch you then. |
06:10.18 | *** join/#htc-linux swc|666 (~gecko@unaffiliated/swc666/x-4934821) |
06:37.29 | *** join/#htc-linux rpierce99 (~rpierce99@96-42-102-103.dhcp.stcd.mn.charter.com) |
06:41.14 | *** join/#htc-linux TheXev (~user@230.sub-174-252-200.myvzw.com) |
06:53.10 | *** join/#htc-linux kiozen (~kiozen@ppp-93-104-92-93.dynamic.mnet-online.de) |
06:58.00 | *** join/#htc-linux jonpry (~jon@unaffiliated/jonpry) |
06:59.32 | *** join/#htc-linux mitsutak_ (~mitsutaka@rt.miraclelinux.com) |
06:59.41 | *** join/#htc-linux YataroX (~user@230.sub-174-252-200.myvzw.com) |
07:02.33 | toastcfh | sup jonpry |
07:02.48 | jonpry | hey toastcfh |
07:03.23 | jonpry | tired, getting ready to sleep :( |
07:03.30 | toastcfh | u do winmo ports as well? |
07:03.59 | toastcfh | i havent fuct with winmo since the diam500 |
07:04.41 | jonpry | yeah i have a rhodium. thats all i ever did before touchpad |
07:04.52 | toastcfh | hehe |
07:04.58 | jonpry | i'm working on some ics stuff for msm7k |
07:05.32 | toastcfh | true i got it booting and gl seems to work for the most part |
07:05.55 | toastcfh | gralloc working |
07:06.10 | jonpry | on leo? |
07:06.27 | toastcfh | shooter and speedy 8660 and 7x30 |
07:06.46 | jonpry | mine doesn't have gles2 |
07:06.53 | toastcfh | right |
07:06.57 | toastcfh | ;/ |
07:07.14 | toastcfh | it run ok? |
07:07.15 | jonpry | trying to make a new libhwui |
07:07.28 | toastcfh | ahh im not even using it |
07:07.35 | jonpry | well no its bad. like 3-4fps |
07:08.32 | jonpry | now that i am into it. it seems like launcher2 is the only thing trying to use libhwui |
07:08.44 | jonpry | must be disabled unless composer is on or something |
07:08.53 | toastcfh | yeah ui rendering with hw |
07:09.36 | toastcfh | the rest ofits like rendering apps with hw |
07:10.07 | toastcfh | u seen the sf_bypass stuff i ported to cm? |
07:10.16 | toastcfh | pretty much the same shit |
07:10.30 | toastcfh | just not so device specific |
07:10.38 | jonpry | sf_bypass eh? |
07:10.45 | toastcfh | qcom only enabled it on 7x30 and 8660 tho |
07:11.05 | toastcfh | surfaceflinger bypass |
07:11.30 | jonpry | afaict there is really only a few lines of code that are gles2 specific in libhwui |
07:11.43 | toastcfh | basically use overlay in apps and with gl. u could also render the ui with overlay as well |
07:12.17 | jonpry | we don't really have overlay |
07:12.26 | toastcfh | ;/ |
07:12.35 | toastcfh | can try with gralloc ;x |
07:12.45 | toastcfh | >_< |
07:13.35 | jonpry | yeah like copybit or something? |
07:13.54 | toastcfh | actually yeah we used copybit to do it as well |
07:13.59 | toastcfh | c2d comp |
07:14.31 | toastcfh | but it looks to me qcom dumped it (copybit) |
07:14.42 | jonpry | i wonder if that would make it go faster |
07:14.52 | toastcfh | hehe idk |
07:15.20 | toastcfh | libc2d2.so and copybit_c2d are all u need |
07:15.30 | jonpry | for hw_composer? |
07:15.34 | toastcfh | libc2d2 is closed sourced tho |
07:15.37 | toastcfh | no |
07:15.45 | toastcfh | for copybit |
07:15.52 | toastcfh | c2d composition |
07:16.16 | toastcfh | which it boggles me why they are bailing on something they just got implemented |
07:23.23 | jonpry | g00g |
07:55.25 | *** join/#htc-linux mitsutaka (~mitsutaka@rt.miraclelinux.com) |
08:13.06 | *** join/#htc-linux BabelO (~fcr@AMontpellier-553-1-104-14.w86-209.abo.wanadoo.fr) |
08:13.06 | *** join/#htc-linux BabelO (~fcr@unaffiliated/babelo) |
08:18.15 | *** join/#htc-linux kiozen (~kiozen@p54BB60BA.dip.t-dialin.net) |
08:18.47 | *** join/#htc-linux kiozen (~kiozen@p54BB60BA.dip.t-dialin.net) |
08:19.52 | *** join/#htc-linux rzk (~rzk@95-27-68-96.broadband.corbina.ru) |
08:23.27 | *** join/#htc-linux paulk (~paulk@lib33-1-82-233-88-171.fbx.proxad.net) |
08:41.31 | *** join/#htc-linux balans2 (~user@82-170-217-205.ip.telfort.nl) |
08:48.23 | *** join/#htc-linux gauner1986 (~Miranda@87.253.171.219) |
09:52.43 | *** join/#htc-linux ychavan (ychavan@nat/redhat/x-pswgksmamrrqvjan) |
10:03.47 | *** join/#htc-linux rajkosto (~rajkosto@2001:470:d76b:da7a:8c21:9c22:f19b:688a) |
10:04.02 | *** join/#htc-linux dobrin (~dobrin@85.91.150.26) |
10:08.04 | *** join/#htc-linux lamikr (lamikr@nat/nokia/x-owaghfodtceubozn) |
10:16.32 | *** join/#htc-linux bartman (~bart@2607:f2c0:a000:175:2e0:81ff:fe47:3d01) |
10:24.44 | *** join/#htc-linux Cass (~Cass@nat/sun/x-qomdwntkccjztpum) |
10:25.42 | *** join/#htc-linux bukington (~bukington@fac34-2-82-228-151-145.fbx.proxad.net) |
10:30.00 | *** join/#htc-linux AstainHellbring (AstainHell@unaffiliated/astainhellbring) |
10:38.07 | *** join/#htc-linux dr1337 (78931828@gateway/web/freenode/ip.120.147.24.40) |
11:09.41 | *** join/#htc-linux stroughtonsmith (~steven@86-44-99-130-dynamic.b-ras2.bbh.dublin.eircom.net) |
11:23.07 | *** join/#htc-linux MethoS- (~clemens@134.102.106.250) |
11:48.25 | *** join/#htc-linux bartman (~bart@2607:f2c0:a000:175:2e0:81ff:fe47:3d01) |
11:53.21 | *** join/#htc-linux jonpry (~jon@unaffiliated/jonpry) |
12:27.11 | *** join/#htc-linux helicopter88 (~helicopte@95.239.9.67) |
12:32.49 | *** join/#htc-linux Rajko (~rajkosto@ds.wan.rajkonet.info) |
12:37.01 | *** join/#htc-linux helicopter88 (~helicopte@95.239.9.67) |
13:51.19 | *** join/#htc-linux mitsutaka (~mitsutaka@p2004-ipbf2206marunouchi.tokyo.ocn.ne.jp) |
15:23.16 | *** join/#htc-linux rpierce99 (~rpierce99@96-42-102-103.dhcp.stcd.mn.charter.com) |
15:25.54 | *** join/#htc-linux hammadi (c500863e@gateway/web/freenode/ip.197.0.134.62) |
15:29.00 | *** join/#htc-linux helicopter88 (~helicopte@95.239.9.67) |
15:40.53 | *** join/#htc-linux NeoMatrixJR (~chatzilla@173-20-63-62.client.mchsi.com) |
15:51.17 | *** join/#htc-linux rob_w (~bob@ppp-188-174-26-88.dynamic.mnet-online.de) |
15:58.47 | *** join/#htc-linux kiozen (~kiozen@p54BB60BA.dip.t-dialin.net) |
16:20.10 | *** join/#htc-linux emwe (~mweirauch@cable-86-56-10-252.cust.telecolumbus.net) |
16:33.44 | *** join/#htc-linux mgross029 (c0234f46@gateway/web/freenode/ip.192.35.79.70) |
17:07.37 | *** join/#htc-linux swc|666 (~gecko@unaffiliated/swc666/x-4934821) |
17:11.10 | *** join/#htc-linux AstainHellbring (~AstainHel@unaffiliated/astainhellbring) |
17:13.15 | *** join/#htc-linux |Jeroen| (~jeroen@d5152B25B.access.telenet.be) |
17:21.33 | *** join/#htc-linux balans2 (~user@82-170-217-205.ip.telfort.nl) |
17:25.30 | *** join/#htc-linux WisTilt2 (~wisgreg@wireless251.wirelesstcp.net) |
17:26.57 | WisTilt2 | arrrghhh or rpierce99: either of you guys available to test a .39 kernel with this new pmem to see if it works on cdma? |
17:27.13 | rpierce99 | yep |
17:27.23 | arrrghhh | i was going to say, so long as i don't have to test audio |
17:27.25 | arrrghhh | i should be good :D |
17:27.54 | WisTilt2 | no audio, just need to see if it blows up on non gsm:) should gain you 32mb free memory |
17:28.23 | phh | using smi2 ? |
17:28.45 | WisTilt2 | i'll never tell hehe |
17:29.03 | phh | ok |
17:29.07 | phh | and seriously ? |
17:29.25 | WisTilt2 | yes its all unlocked now |
17:29.50 | WisTilt2 | might be placebo but everything seems so much faster in smi memory |
17:30.11 | WisTilt2 | guys, dl the usual kernel pack name, its up there |
17:30.38 | arrrghhh | k |
17:31.20 | WisTilt2 | phh: any reason we can't just allocate 16mb for pmem that you know of? |
17:31.44 | phh | WisTilt2: you mean put 16MB of "generic" ram (ie non pmem) in smi ? |
17:32.38 | WisTilt2 | no, im talking about in mach-msm/pmem.c, pmem_size currently 32 |
17:32.40 | phh | at that time, hole in memory wasn't working really gerat |
17:33.17 | phh | mmmmm ? |
17:34.09 | WisTilt2 | pmem=32 now and its in ebin starting up at 78mb offset. just wondering why that needs to be 32mb and couldn't be smaller |
17:34.55 | rpierce99 | WisTilt2: sorry can you send me the link again i seem to have misplaced it |
17:35.11 | WisTilt2 | you should have that memorized by now:) |
17:35.23 | rpierce99 | this particular pack i've only downloaded once :) |
17:35.28 | phh | what |
17:35.34 | phh | in current code SMI is not used at all oO |
17:35.35 | arrrghhh | lol |
17:35.36 | rpierce99 | i have the full domain and path memorized, not the file name |
17:36.44 | WisTilt2 | phh: smi currently is only being used by gpu0 anc ram console, everything else in ebi and ebin |
17:36.52 | WisTilt2 | is current public code that is |
17:37.54 | jonpry | phh where did this 3d blob come from? |
17:37.58 | rpierce99 | WisTilt2: did you have a chance to look into the issue with mac + 39 |
17:38.08 | arrrghhh | yea, don't use mac |
17:38.09 | phh | jonpry: god knows |
17:38.12 | arrrghhh | problem solved |
17:38.22 | rpierce99 | problem avoided, not solved |
17:38.25 | jonpry | Q3Dimension MSM7500 01.02.08 0 4.0.0 |
17:38.28 | WisTilt2 | rpierce99 not yet. i'll try that today, actually forgot i was going to test that |
17:38.33 | arrrghhh | solved, avoided... it's no longer a problem. |
17:38.43 | jonpry | there should be a newer blob. this thing is only gles 1.0 |
17:38.47 | rpierce99 | WisTilt2: no biggy, it's low priority |
17:39.53 | WisTilt2 | riperce99: after you get this booted up and it settles down, if you could kill all or most of your apps then check available memory, should be up around 105mb or so free |
17:40.12 | jonpry | we should be able to use this right? http://www.xda-developers.com/android/working-opengles-1-1-on-htc-hero/ |
17:40.14 | phh | WisTilt2: anyway pmem in SMI should be a speed up by itself |
17:40.22 | rpierce99 | best way to read free mem? |
17:40.25 | rpierce99 | widget? |
17:40.57 | detule | cat /proc/meminfo is informative though not necessarily about what android sees |
17:41.14 | WisTilt2 | phh: it must have. i currently have everything running in smi except adsp, which doesn't seem to like it for whatever |
17:41.22 | arrrghhh | rpierce99, i think it's in about phoen |
17:41.23 | arrrghhh | phone* |
17:41.45 | WisTilt2 | rpierce99 if you have systempanel app it shows available memory accurately |
17:42.10 | phh | WisTilt2: SMI goes through AHB bus too ? |
17:42.13 | detule | WisTilt2, did you make the extra memory available to the kernel in board-rhod* |
17:42.27 | WisTilt2 | detule yep |
17:42.38 | detule | rpierce99, you should see it if you just type in "free" in connectbot |
17:42.43 | WisTilt2 | phh need to look at docs, dont know of top of my head about that |
17:49.10 | WisTilt2 | phh: looks like smi can be controlled via AHB or AXI. i have no idea how we have it setup in the current code to know which mode it's currently in. |
17:49.28 | phh | wat. there are both ? |
17:49.32 | phh | okay i'm lost -_-' |
17:50.01 | WisTilt2 | yes there is a mode bit that determines ahb or axi mode |
17:50.51 | WisTilt2 | in the smi controller config register |
17:51.21 | WisTilt2 | rpierce99 ? chomping at the bit here |
17:51.42 | rpierce99 | killing stuff with system panel atm |
17:51.51 | rpierce99 | i'm at 77m and some stuff won't close |
17:52.02 | WisTilt2 | you mean it didnt blow up on cdma wow |
17:52.09 | arrrghhh | lol |
17:52.23 | detule | the kernel should see the memory if it's allocated in the board file, you can just see it in a terminal |
17:52.36 | rpierce99 | too early to tell how stable it is, but no it is working fine for getting into market, installing apps, killing process, running apps, etc |
17:52.55 | phh | "The AMBA 3 AHB interface specification enables highly efficient interconnect between simpler peripherals in a single frequency subsystem where the performance of AMBA 3 AXI is not required." |
17:52.56 | WisTilt2 | systempanel wont let you kill everything? |
17:52.57 | phh | ok set it to AXI. |
17:53.13 | rpierce99 | no, email, facebook, gvoice, status bar all come back |
17:53.15 | jonpry | now i just need to know where acl patched the blob |
17:53.16 | rpierce99 | and connectbot is running |
17:53.25 | phh | jonpry: wasn't it cotulla ? |
17:53.29 | rpierce99 | each of those is 12-15M |
17:53.37 | jonpry | no don't think so. acl and bzo |
17:53.38 | rpierce99 | system panel is 17M itself |
17:53.40 | WisTilt2 | 77mb free with all that running, how much was free with the other kernel? |
17:53.48 | jonpry | if i had the old blob i could diff it |
17:53.52 | rpierce99 | dunno, just installed this app |
17:53.58 | rpierce99 | i can revert kernels |
17:54.11 | detule | rpierce99, just to humor me what's the output of "free" |
17:55.04 | rpierce99 | 197,624 -> 194,724 -> 2900 -> 0 -> 1388 |
17:55.12 | detule | cool i get 174 |
17:55.14 | jonpry | are the blobs checked into git somewhere? |
17:55.18 | *** join/#htc-linux kiozen (~kiozen@ppp-88-217-21-58.dynamic.mnet-online.de) |
17:55.22 | detule | so there's certainly +20MB of memory you are seeing |
17:56.21 | WisTilt2 | detule you see 174 total? |
17:56.23 | detule | more specifically "total = 174264" |
17:56.32 | detule | yeah so the kernel sees the extra memory fine |
17:56.44 | WisTilt2 | <PROTECTED> |
17:56.44 | WisTilt2 | <PROTECTED> |
17:56.49 | WisTilt2 | my free |
17:57.10 | detule | yeah i am at the above total and at 6568 free |
17:57.16 | detule | after being booted since this morning |
17:57.43 | WisTilt2 | rpierce99 does it seem snappier at all? |
17:57.57 | rpierce99 | well i've never killed all of my apps before :) |
17:58.08 | detule | rpierce99, try something 3d |
17:58.26 | WisTilt2 | here's where cdma will blow up lol |
17:59.09 | rpierce99 | i just somehow managed to lose my launcher bar |
18:00.50 | WisTilt2 | phh: changing the mode bit to AXI is easy enough to do but won't we have other things to change also? |
18:01.07 | phh | WisTilt2: no fucking clue :D |
18:01.12 | WisTilt2 | me neither:) |
18:01.25 | phh | i can barely see why the hell they would put axi *and* ahb support |
18:01.31 | *** join/#htc-linux paulk (~paulk@lib33-1-82-233-88-171.fbx.proxad.net) |
18:01.32 | jonpry | arrrghhh i see in the logs that bzo emailed you a 1.1 3d lib? |
18:01.40 | rpierce99 | 3d didn't kill it |
18:01.43 | arrrghhh | lol |
18:01.44 | rpierce99 | it doesn't seem super fast |
18:01.48 | rpierce99 | but it's a dumb game |
18:01.55 | arrrghhh | jonpry, hum. how long ago? |
18:02.04 | jonpry | jan 30 |
18:02.33 | phh | aka long long ago. |
18:02.38 | arrrghhh | oh of this year? |
18:02.42 | arrrghhh | hrm i should still have it lemme look |
18:02.50 | jonpry | thanks |
18:03.45 | rpierce99 | regular apps seem plenty fast, i mean it's not burning my hands or anything, more free ram doesn't mean more speed, except when avoiding IO by storing stuff in memory |
18:04.13 | phh | it means less GC ... |
18:04.22 | arrrghhh | hrm. i can't find it in my regular gmail |
18:04.29 | arrrghhh | i wonder if it was sent to my xdandroid acct... |
18:05.01 | jonpry | he may not have sent it. just said that he would shortly |
18:05.04 | detule | we'll need new oom_values :) we changed them because we couldn't envision operating with over 80MB free |
18:06.18 | arrrghhh | jonpry, did you check the ML? |
18:06.22 | arrrghhh | i am now, nothing so far |
18:07.58 | detule | jonpry, i think gles 1.1 is somehow in 7.1 |
18:08.28 | jonpry | frx7.1? |
18:08.58 | arrrghhh | yes |
18:09.09 | arrrghhh | where would it be? i can try to yank it for you |
18:09.13 | arrrghhh | it's also on my server :P |
18:09.22 | jonpry | system/lib/egl/libgles_qcom.so |
18:09.55 | mgross029 | Wasn't xofrats working on that iirc? |
18:10.25 | arrrghhh | that dude kinda fell off ther |
18:10.32 | arrrghhh | not sure what happened. |
18:10.50 | mgross029 | Yeah noticed that. |
18:10.51 | rpierce99 | i'd guess he got tired of hyc complaining about his ideas |
18:11.04 | detule | complaining is one way of saying it |
18:11.06 | rpierce99 | it got pretty brutal on the ML there for bit |
18:11.19 | arrrghhh | lol |
18:11.24 | arrrghhh | yea |
18:11.27 | arrrghhh | hawkward |
18:13.45 | jonpry | i have my doubt 1.1 is actually any different. its like they made made a bunch of optional extensions into mandates |
18:14.34 | detule | sometimes, i wish i knew what you are talking about.... |
18:14.39 | arrrghhh | lol |
18:16.04 | WisTilt2 | jonpry do you ever sleep? |
18:16.12 | jonpry | more than you know |
18:16.15 | arrrghhh | lol |
18:16.24 | arrrghhh | ok i have the libGLES_qcom.so |
18:16.29 | jonpry | email? |
18:16.31 | arrrghhh | where do you want it? /home/jon, or weblink? |
18:16.32 | arrrghhh | ok |
18:17.24 | arrrghhh | sent |
18:17.34 | jonpry | thanks |
18:17.39 | arrrghhh | np |
18:19.39 | WisTilt2 | btw, i setup one of our 300's for a 24hr test in sleep with scbs running. 19.2% battery drop over 24hrs. radio on, no syncing. winmo's got nothing on us now. |
18:20.11 | jonpry | probably uses like 10%/hour on my 210 |
18:20.29 | mgross029 | But what happens when a call comes in? Does it wake up? :p |
18:20.41 | rpierce99 | my phone died before I went to bed last night :( |
18:20.46 | WisTilt2 | of course it does |
18:20.59 | WisTilt2 | rpierce99 which kernel? |
18:21.08 | *** join/#htc-linux stroughtonsmith (~steven@86-44-87-122-dynamic.b-ras2.bbh.dublin.eircom.net) |
18:21.20 | mgross029 | Cool... Even better then. |
18:24.06 | mgross029 | jonpry: Came across this link. It include libGLES_qcom.so, gralloc.msm7k.so and gralloc.morrison.so for 1.1 you made need to replace all of them. http://www.simply-android.com/discussion/833/mod-opengl-es-1.1-for-motorola-cliq/p1 |
18:24.38 | detule | by the way, I too, like bzo am noticing an increase in power consumption |
18:24.45 | jonpry | ics required a new and improved gralloc |
18:25.11 | rpierce99 | WisTilt2: your power saver |
18:25.15 | detule | this not having booted the the kernel with the latest pm adjustments..... |
18:25.15 | WisTilt2 | detule on .39? |
18:25.32 | detule | yeah .39 and .30 |
18:25.56 | WisTilt2 | rpierce99 how long did you last and was this better or worse under the same conditions? |
18:26.12 | rpierce99 | with the new kernel, fresh boot without killing anything i have 47M Available in SPL |
18:26.40 | detule | that's ok, lowmemory killer doesn't kick in until the 48MB mark |
18:26.48 | WisTilt2 | detule are you running sensors or something that's causing that? no way should the kernel be causing more drain now unless something else is keeping things awake |
18:26.54 | rpierce99 | you don't know what my OOM settings are |
18:27.20 | mgross029 | jonpry: wouldn't they have to matchup with the libGLES? Just a thought. |
18:27.32 | detule | right you are rpierce99 i assumed you are using something out of emwe's lab, if you cat minfree what's the last value |
18:27.44 | detule | perhaps it's even less than 48MB :) |
18:27.53 | rpierce99 | what is cyanogenmod.android.fotakill and why do i have it |
18:27.58 | jonpry | it seems to be working, whatever that is worth |
18:28.34 | detule | WisTilt2, i have my eye on scbs as the culprit |
18:28.48 | jonpry | boo |
18:28.56 | arrrghhh | detule, do you have the demon enabled in rootfs? |
18:29.04 | arrrghhh | i thought it wouldn't do anything if the demon wasn't turned up |
18:29.05 | WisTilt2 | i dont think its scbs. i only see a couple ma difference running scbs |
18:29.16 | rpierce99 | yes WisTilt2 i know i am running sensors, i assume detule is since he gave them to me |
18:29.41 | detule | i have auto rotate turned off in settings -> i don't think it's polling the sensor then |
18:29.48 | rpierce99 | oh i dont |
18:29.56 | detule | at least it's not in froyo |
18:32.50 | detule | could be sensors don't know |
18:33.17 | rpierce99 | my last minfree value atm is 7680 |
18:33.29 | detule | much more conservative than emwe's rootfs |
18:33.39 | detule | it doesn't kick in till 30MB |
18:33.44 | detule | (no of pages x 4KB) |
18:34.19 | arrrghhh | well, booted .39 |
18:34.41 | arrrghhh | new data.img so no way to test 3d yet. gonna dl some poop |
18:35.42 | detule | i think my battery is also getting fried though.....took it off the charger this morning, already getting blinking red light |
18:36.26 | arrrghhh | detule, that's bad |
18:36.29 | arrrghhh | i thought mine was dying... |
18:40.01 | detule | i think emwe's hit the sweet spot with the oom values |
18:45.17 | *** join/#htc-linux LordDeath (~LordDeath@cable-81-173-166-124.netcologne.de) |
18:50.01 | rpierce99 | I don't think there is a sweet spot, depends on usage |
18:50.15 | rpierce99 | someone who uses the same three apps over and over should just keep it all in memory |
18:50.35 | rpierce99 | someone who downloads and tries the free app of the day every day should probably run with some free |
18:50.42 | mgross029 | jonpry: Just another thought what do you have in your build.prop for ics with this entry? ro.opengles.version |
18:52.51 | jonpry | nothing |
18:53.25 | mgross029 | So you don't have that set or is is =(blank) |
18:53.35 | jonpry | not set |
18:54.27 | mgross029 | According to what I've read that setting needs to be there and it should be ro.opengles.version=65537 for Opengles 1.1 |
18:55.03 | jonpry | i'm not 100% convinced 3d is operation |
18:55.24 | jonpry | new egl might just not set the prop though |
18:56.19 | *** join/#htc-linux kiozen (~kiozen@ppp-88-217-21-58.dynamic.mnet-online.de) |
18:56.25 | mgross029 | Well thought it would be worth a shot to get to see if that does anything for you. |
18:56.49 | mgross029 | http://translate.google.com/translate?hl=en&sl=ru&u=http://www.gizmonder.com/2011/01/huawei-u8230-opengl-11.html&ei=eO3LTo6lFoeNsQLIhNi2Dg&sa=X&oi=translate&ct=result&resnum=6&sqi=2&ved=0CE0Q7gEwBQ&prev=/search%3Fq%3Dicecream%2Bsandwich%2Blibgles_qcom.so%26hl%3Den%26biw%3D1191%26bih%3D556%26prmd%3Dimvnsfd |
18:57.02 | arrrghhh | lol |
18:57.06 | arrrghhh | that's quite a linkn |
18:57.09 | mgross029 | Sorry had to translate that page. |
18:57.15 | mgross029 | I don't read russian well. |
18:57.28 | jonpry | i am not totally sure what it is i am trying to get to work atm |
18:58.14 | arrrghhh | lol |
18:58.18 | mgross029 | lol Well then we'll just keep throwing answers your way until one sticks. |
18:59.53 | jonpry | it seems like launcher2 probably needs libhwui to be usable. but everything else sucks because of no hw_composer |
19:01.05 | arrrghhh | launcher2 sucks |
19:01.11 | arrrghhh | although i haven't seen the ICS version i guess |
19:01.12 | arrrghhh | seen/used |
19:04.46 | jonpry | having to swap out the launcher would be more than fine |
19:07.10 | detule | rpierce99, sure.... what i mean is that with emwe's rootfs i find that there is plenty of available memory for sudden load, while at the same time there is enough room for usually a few apps to persist.... regardless of who the user is, more often than not these will be the apps you use daily, email maps etc... |
19:07.53 | detule | at the same time, _some_ settings need to be loaded by default, and i feel like as such, these do the jobs....down the road if a user wants to, they can customize to their liking |
19:09.46 | detule | certainly better than whatever defaults are preloaded in the kernel driver |
19:10.47 | arrrghhh | ^^ |
19:10.56 | arrrghhh | and it's a pretty happy medium IMHO |
19:11.06 | arrrghhh | set it too aggressive and some apps won |
19:11.06 | mgross029 | WisTilt2: All of your latest commits are only in your local repo and not on gitorious right? |
19:11.08 | arrrghhh | t run |
19:12.12 | WisTilt2 | mgross029 which latest commits you referring to, there are a bunch |
19:12.29 | mgross029 | pmem and such |
19:13.06 | mgross029 | I found a way to make a1010 fire like mad in the logs to see if it is working or you get errors |
19:13.14 | WisTilt2 | PM stuff with idle suspend/collapse is already on git. |
19:13.45 | mgross029 | Ok. I was just going to compile a kernel and test it out in .39 to see what happens |
19:14.19 | WisTilt2 | so is noise canceling working now or is the a1010 just up and doing nothing? |
19:14.59 | mgross029 | As far as I can tell phone calls seem a little clearer. |
19:15.28 | WisTilt2 | that would only be in speakerphone mode too |
19:15.45 | mgross029 | Also I was streaming some music through bt to my car audio and it also did not skip like it did in the past |
19:16.04 | WisTilt2 | if you turn it up full volume you can really tell if canceling is working or not |
19:16.37 | *** join/#htc-linux mastermerlin (~Adium@pD9E2EC69.dip.t-dialin.net) |
19:17.35 | mgross029 | I did turn it up loud but not full. Plus I had my gps audio talking behind the music which was another big difference. It seemed to lower the gps speach while the music was playing |
19:18.08 | mgross029 | It was all clear and not garbled |
19:18.23 | *** join/#htc-linux detule (~detule@hw-ma-6l13f61.mat.jhu.edu) |
19:18.46 | WisTilt2 | sounds like its working. this was on .35 i assume? |
19:19.22 | mgross029 | Correct. |
19:20.03 | WisTilt2 | let me know if your method works on .39 and we can try it |
19:21.05 | arrrghhh | hrm. why the hell can't i install apps... i think i should give up on android, i can't do the basics anymore. |
19:21.26 | WisTilt2 | old age maybe |
19:21.31 | arrrghhh | lol |
19:21.54 | arrrghhh | i am trying to install an app for testing 3d on this new .39 kernel |
19:22.00 | arrrghhh | and it just endlessly says installing |
19:22.01 | mgross029 | I have noticed a lot of points where a1010 is getting called throughout the kernel and in libaudio_wince so I'm wondering without it enabled if it would cause things to break. Just my thoughts |
19:22.04 | WisTilt2 | cant install any apps? |
19:22.33 | mgross029 | FC the vending app and try again. |
19:25.11 | *** join/#htc-linux GNUtoo (~gnutoo@host249-67-dynamic.116-80-r.retail.telecomitalia.it) |
19:26.07 | *** join/#htc-linux Cotulla (~myfakemai@nat100-255-205-109.tvoe.tv) |
19:26.10 | arrrghhh | trying to install another app now |
19:26.39 | Cotulla | hi, the other part of earth |
19:26.56 | detule | mgross029, a lot of points in the kernel and libaudio? |
19:26.56 | WisTilt2 | hi Cotulla way over there in the dark |
19:27.16 | mgross029 | detule: yes |
19:27.20 | Cotulla | yeah |
19:27.24 | Cotulla | dark already here :( |
19:27.25 | detule | where at |
19:27.39 | Cotulla | no white nights :( |
19:29.17 | WisTilt2 | phh: from what i can tell with the current code we are already axi on smi. |
19:30.30 | mgross029 | detule: mainly libaudio_wince/AudioHardware.cpp and a1010.h which is also used by AudioHardware.cpp |
19:30.38 | detule | i see one one place in libaudio where it calls doAudience |
19:31.19 | *** join/#htc-linux rajkosto (~rajkosto@2001:470:d76b:da7a:8c21:9c22:f19b:688a) |
19:33.22 | detule | but that's all, and that after properly checking to see if audience is available or not...don't think that one call is breaking much |
19:37.19 | WisTilt2 | detule: in your .39 merge do you remember was clock-wince.c changed at all? |
19:37.25 | mgross029 | Ok. Just thought I would throw that out there. |
19:38.18 | detule | WisTilt2, changed plenty |
19:38.41 | detule | https://gitorious.org/~detule/linux-msm-rhod/detules-linux-msm-rhod/commit/c36233d312ba5e1180480bf2ee783f137b33292e |
19:39.13 | WisTilt2 | got a couple problems with the cam and vfe registers unless my eyes are playing tricks on me against the docs |
19:39.38 | detule | could be |
19:40.41 | detule | they were brought en masse from Alex via emwe's tree |
19:41.00 | WisTilt2 | hmm, and camera is working fine on their tree? |
19:41.58 | mgross029 | camera isn't enabled on 35 yet |
19:43.11 | detule | WisTilt2, which registers |
19:43.26 | detule | CAM_MD and VFE_NS? |
19:43.34 | WisTilt2 | MSM_CAM_MD_REG MSM_VFE_NS_REG |
19:43.58 | WisTilt2 | yep |
19:44.04 | detule | they haven't changed from what they were before |
19:44.36 | WisTilt2 | docs say different. all other defines are right though |
19:45.09 | detule | before we had them as : 53#define MSM_CAM_MD_REG ( MSM_CLK_CTL_BASE+0x40 )#define MSM_CAM_NS_REG ( MSM_CLK_CTL_BASE+0x44 ) |
19:45.10 | WisTilt2 | i thought i looked at this last year on .27 and came to same conclusion. |
19:45.25 | WisTilt2 | they should be 0x3c and 0x40 |
19:45.37 | WisTilt2 | 0x3c being CAM_MD |
19:45.49 | detule | ok, just checking it wasn't my error :) |
19:46.08 | detule | i guess they were wrong in JB's tree as well? |
19:46.09 | WisTilt2 | 0x44 as it is set is actually ecodec_md |
19:46.59 | WisTilt2 | they had to be right at some point since camera worked so i have no idea |
19:47.36 | detule | assuming that is the probelm with the camera |
19:47.50 | detule | git log doesn't lie either |
19:47.50 | jonpry | nothing has to be right with clk_regime |
19:47.57 | WisTilt2 | dont think this would be causing the memory clashes though, |
19:50.20 | *** join/#htc-linux _kiozen (~kiozen@ppp-88-217-6-109.dynamic.mnet-online.de) |
19:52.06 | detule | for the record i remember trying the camera with the autobuild kernel before any of the recent changes to the tree - it failed much in the same way as it does now |
19:53.36 | jonpry | 27 still uses clk_regime. so it might not be using the MD/NS that it appears to be |
19:54.42 | detule | yes the usage of those registers may have changed |
19:57.41 | jonpry | i think we need somebody with a panda board, or omap4 device |
19:57.56 | rpierce99 | buy me one |
19:58.10 | jonpry | well |
19:59.34 | jonpry | <PROTECTED> |
20:02.06 | mgross029 | WisTilt2: Just tried booting .39 with your latest git commits and a1010 enabled and I didn't even get bootani. :p heh |
20:02.31 | mgross029 | going to give it another try. |
20:02.34 | WisTilt2 | any build errors? |
20:02.53 | mgross029 | None that I saw. Let me check my compile log |
20:03.43 | WisTilt2 | if its off the clean tree it builds fine. enabling that i2c a1010 shouldnt cause it not to boot |
20:04.59 | WisTilt2 | jonpry any idea why adsp doesn't like to run in smi? |
20:05.04 | mgross029 | I agree. I've got bootani now and I did not see anything in the log |
20:05.37 | jonpry | some of the blobs seem to bound the memory regions |
20:05.53 | jonpry | maybe even the qdsp5 |
20:06.15 | jonpry | i think i can get gpu1 to work in ebin |
20:06.53 | *** join/#htc-linux GNUtoo (~gnutoo@host40-58-dynamic.116-80-r.retail.telecomitalia.it) |
20:14.06 | mgross029 | WisTilt2: well a1010 fired in 39 but then broke. It took out adc3001 too |
20:22.18 | WisTilt2 | mgross029: guess that's another project down the road then |
20:23.20 | mgross029 | http://dl.dropbox.com/u/45928762/dmsg39_1122.txt Here is the dmesg output just so you can see |
20:24.45 | WisTilt2 | jonpry: if we can get adsp into smi we're good to go. there's enough room and i changed pmem to 16mb, made no difference that i can see as far as anything negative so still dont know why it needs to be 32mb. |
20:25.36 | WisTilt2 | mgross you tried taking a picture it looks like, which break other stuff as it is right now |
20:27.11 | mgross029 | Hmm. I did not try to take a picture |
20:27.19 | WisTilt2 | and it looks like adc3001 got powered down for some reason |
20:27.29 | WisTilt2 | something used the camera app |
20:27.30 | mgross029 | zDevice Test probably initialized the camera |
20:29.31 | WisTilt2 | arrrghhh did you ever get .39 up? |
20:30.19 | arrrghhh | WisTilt2, yea. i think i need to reboot tho, no apps will install. |
20:31.16 | WisTilt2 | k just wondered if i missed your comment and left you hanging. im out of here for lunch shortly so ill get your results when i get back. |
20:31.27 | arrrghhh | np mate |
20:38.08 | *** join/#htc-linux kiozen (~kiozen@ppp-88-217-6-109.dynamic.mnet-online.de) |
20:44.55 | *** join/#htc-linux TheXev (user@pool-108-11-143-164.eriepa.dsl-w.verizon.net) |
20:58.48 | *** join/#htc-linux AstainHellbring (~AstainHel@unaffiliated/astainhellbring) |
21:09.24 | *** join/#htc-linux balans2 (~user@82-170-217-205.ip.telfort.nl) |
21:12.25 | *** join/#htc-linux Cass (~Cass@188-220-34-222.zone11.bethere.co.uk) |
21:19.36 | jonpry | no installz? |
21:20.03 | jonpry | WisTilt2, what good is adsp in smi? |
21:20.24 | jonpry | need to put stuff in there that sucks up bandwidth |
21:20.38 | jonpry | maybe like gpu0 = smi1, gpu1 = smi2 |
21:20.45 | phh | jonpry: already tried ;) |
21:20.59 | phh | and it didn't work |
21:21.06 | jonpry | there is a gralloc patch for such things |
21:21.20 | phh | mmm i'm pretty sure that wasn't the problem |
21:21.21 | jonpry | and i think a blob hack will be required |
21:21.49 | jonpry | the patches existence implies it is possible no? |
21:22.02 | phh | ah you've already seen such a config ? ok |
21:22.31 | *** join/#htc-linux MethoS- (~clemens@134.102.106.250) |
21:25.10 | jonpry | https://www.codeaurora.org/gitweb/quic/le/?p=platform/hardware/msm7k.git;a=commit;h=9c2781aab28f21e74b7a30bbaf50747d82743d54 |
21:26.20 | Cotulla | 33 C0 CD 10 C3 |
21:26.55 | Cotulla | 0F 00 A0 E1 <--------- who can? |
21:28.21 | mgross029 | I'm out guys. Laters |
21:28.31 | Cotulla | good luck |
21:40.42 | *** join/#htc-linux rob_w (~bob@ppp-188-174-26-88.dynamic.mnet-online.de) |
21:45.11 | *** join/#htc-linux dobrin (~dobrin@85.91.150.26) |
21:54.02 | WisTilt2 | jonpry: gpu0 is already in smi along with ram console. i only wanted adsp in smi to free up available memory for the rest of the system |
21:55.10 | WisTilt2 | so to move gpu1 into smi we need to do some hard address coding or what? |
21:55.16 | *** join/#htc-linux TheXev (user@pool-108-11-143-164.eriepa.dsl-w.verizon.net) |
22:00.26 | jonpry | WisTilt2, can't the rest of the system just use smi2? |
22:00.50 | jonpry | moving gpu1 might be as simple as changing the pmem stuff |
22:02.05 | WisTilt2 | modem uses the upper part of it along with parts of smi1 so we'd have to be damn sure about areas we use. i am in process of setting up something to do just that to see what changes. |
22:02.28 | WisTilt2 | i could try moving gpu1 right now though, that's a quick test |
22:02.41 | Willd | WisTilt2: Hit me up if you manage to get it working |
22:02.56 | WisTilt2 | gpu size again? why is it at 8mb |
22:03.14 | WisTilt2 | i just wonder what the minimum we ever need is |
22:03.34 | jonpry | or maximum? |
22:03.41 | jonpry | some games will need more |
22:03.44 | WisTilt2 | yeah that's what i meant in a backwards way |
22:04.05 | jonpry | not sure. but ics could need lots |
22:04.21 | WisTilt2 | also just to confirm, page align is ok on 2mb boundaries correct? |
22:04.50 | jonpry | yeah it seems good on 1mb |
22:05.03 | WisTilt2 | doesnt have to be multiple of 2 then? |
22:05.23 | jonpry | no |
22:05.56 | WisTilt2 | ok let me try gpu1 in smi2 first and we'll see if it crashes it |
22:06.42 | jonpry | it possibly can be fixed with a gralloc patch or patch to the blob |
22:07.15 | jonpry | btw. does smi have its own sdram pads or is it just a chip select? |
22:07.40 | WisTilt2 | smi is on the msm chip directly |
22:07.58 | jonpry | esdram? |
22:08.04 | jonpry | i got the impression it was stacked die |
22:10.57 | phh | jonpry: smi is on axi bus too |
22:11.02 | phh | so chip select alike |
22:12.38 | jonpry | not necessarily. axi is a multiple issue bus |
22:13.40 | *** join/#htc-linux [acl] (~abel@96.246.167.90) |
22:18.29 | WisTilt2 | i need to get kexec setup, this rebooting is a pita. got a dma error with that setup jonpry, booting again to be sure |
22:19.58 | jonpry | are you using the gles 1.1 blob? |
22:24.59 | WisTilt2 | never changed that. looks like gpu1 needs a change elsewhere. i moved fb and camera up to smi2 and works fine. gpu1 wants to be in ebi |
22:26.14 | WisTilt2 | dma errors were from datamover_irq |
22:32.13 | WisTilt2 | detule did you already push those changes you had with removing registering gpu0/1 since hw3d already did that? |
22:32.43 | WisTilt2 | i guess i could look myself but i have 15 things going on right now |
22:53.54 | *** part/#htc-linux Cotulla (~myfakemai@nat100-255-205-109.tvoe.tv) |
23:01.39 | detule | sorry i was afk, no haven't pushed anything like that out WisTilt2 |
23:02.01 | detule | jonpry, that CA commit isn't that already in our gralloc? |
23:02.23 | jonpry | i didn't check |
23:02.24 | WisTilt2 | was it just removing the register of both gpu0/1 you were talking about? i was going to do that on my end to test |
23:02.25 | jonpry | probably |
23:02.45 | detule | WisTilt2, right just removing the registering |
23:02.49 | WisTilt2 | k thanks |
23:03.05 | detule | in whatever limited testing it didn't have much of an effect on my device, positive or negative |
23:03.41 | detule | jonpry, i am seeing that FB_arena stuff here http://gitorious.org/xdandroid-eclair/xdandroid-hardware_msm7k/blobs/froyo/libgralloc/gralloc.cpp |
23:03.52 | WisTilt2 | jonpry, looks like modem is using part of upper smi2 and some bottom. we're only going to have around 16mb usable in that area with a bit more in smi1 also |
23:10.12 | jonpry | yeah gpu1 probably not worth while |
23:10.27 | jonpry | i figure just allocate the 16mb to linux |
23:11.17 | WisTilt2 | looks like 16mb in smi2 will be safe and 8-10mb in smi1, although the smi1 area will be wrong on cdma for sure, i remember trying this with arrrghhh long time ago. |
23:12.06 | jonpry | i am starting to think hardware renderer is not as important as this compositor thing |
23:12.44 | jonpry | need someone with working ics to confirm though |
23:22.40 | jonpry | WisTilt2, you put saphire radio on yet? |
23:22.57 | WisTilt2 | no one's got it to me to dl yet:) |
23:23.42 | jonpry | getting it is easy. its putting it some reasonable format that might be hard |