00:00.16 | cr2_ | NetRipper: 3 funcs, one after another |
00:00.20 | NetRipper | ok |
00:02.52 | cr2_ | NetRipper: http://xda.privatepaste.com/f6fcf77b80 |
00:03.05 | *** join/#htc-linux ToAsT_ (~quassel@235-41.200-68.tampabay.res.rr.com) |
00:04.44 | ToAsT_ | phh: woot.. quassel is pretty nice:P |
00:04.46 | NetRipper | that first one reads cpu id, and identifies the processor |
00:06.16 | NetRipper | and recursively clears cache i think |
00:06.32 | NetRipper | it's a whole lot shorter than the hierarchial cache flush in linux |
00:06.35 | cr2_ | the last is called with R1 param, but it's only a loop counter |
00:06.46 | cr2_ | ok |
00:07.30 | NetRipper | that c7, c14, 2 clears cache by set/way, but there are multiple caches so you need to set a cache, clear it, move on to next |
00:07.40 | cr2_ | ok |
00:07.56 | NetRipper | c7, c14, 0 clears all caches when they're ARMv7 compatible, but doesnt work on this cpu |
00:08.12 | cr2_ | not used in wince too |
00:08.44 | cr2_ | buggy hw ? :) |
00:08.47 | NetRipper | so they just clear hierarchially and simply don't finish with c7, c14, 0 (which they do in linux) |
00:08.57 | NetRipper | different hw |
00:09.02 | cr2_ | ok |
00:09.03 | NetRipper | maybe buggy rev yes |
00:09.06 | cr2_ | nexus ? |
00:09.24 | cr2_ | and debug boards. |
00:09.31 | cr2_ | they have the same cpu |
00:09.44 | NetRipper | well i don't know if they use zImage or Image for nexus |
00:09.49 | cr2_ | hm. different rev. maybe. |
00:09.57 | NetRipper | if they use zImage there, that means they don't have this problem |
00:10.23 | cr2_ | can we compare your dmesg with nexus ? |
00:10.30 | NetRipper | (cos the problems is in the decompression code) |
00:10.35 | cr2_ | or is the cpuid not so accurate ? |
00:10.52 | cr2_ | decompression is written in asm ? |
00:10.53 | *** join/#htc-linux M1DLGpc_ (~M1DLGpc@bb-87-81-252-83.ukonline.co.uk) |
00:11.24 | NetRipper | cr2_, no, but the c-stack setup is, and for the decompression it enables mmu temporarily |
00:11.55 | cr2_ | insane :) |
00:12.15 | NetRipper | and when they're done, they disable mmu again.. and to get it back in proper state for loading the uncompressed linux kernel.. as it expects D-cache to be clear, they need toc lear it |
00:12.56 | cr2_ | ok |
00:13.37 | cr2_ | googling for nexus dmesg |
00:13.52 | NetRipper | cr2_, btw do you want msmclkkhz as a seperate command as well or just move it all into dump? |
00:14.12 | cr2_ | NetRipper: dump what ? |
00:14.31 | NetRipper | that msmclkkhz |
00:14.47 | NetRipper | i'm moving it into a dump msmclkkhz <start> <end> command |
00:15.05 | cr2_ | as you like |
00:15.07 | NetRipper | btw is it 0-255 or 1-256? |
00:15.14 | cr2_ | 0 works |
00:15.21 | NetRipper | yea but 1000000 works too |
00:15.25 | cr2_ | maybe you can provide a full int32 |
00:15.37 | cr2_ | the rpc param is int32 |
00:15.41 | NetRipper | i think it does a modulus 256 |
00:15.44 | tmzt | NetRipper: does nexus use zImage? |
00:15.52 | NetRipper | tmzt, no that's what i'd like to know |
00:15.57 | NetRipper | tmzt, so i dont know |
00:16.00 | tmzt | there's 2.6.29 and 2.6.32 |
00:16.01 | cr2_ | tmzt: do you know somebody with nexus ? |
00:16.15 | tmzt | maybe you can ask San since this is a cpu thing and not firmware |
00:16.18 | NetRipper | tmzt, i'm aware of that, but that doesnt tell us if they use Image or zImage |
00:16.18 | cr2_ | tmzt: nexus and devmem2 :) |
00:16.35 | NetRipper | san? |
00:16.43 | cr2_ | NetRipper: from android |
00:18.01 | tmzt | Makefile.boot is part of the zImage generation |
00:18.18 | NetRipper | tmzt, ah, it's not used in Image? |
00:18.26 | tmzt | the inner image/vmlinux is compressed as piggy.gz and attached to the decompressor |
00:18.35 | tmzt | this happends in boot/compressed |
00:18.53 | cr2_ | NetRipper: i think they should use zimage |
00:19.50 | tmzt | cr2_: I have a nice 4gb notebook |
00:20.04 | cr2_ | tmzt: >4G |
00:20.16 | tmzt | I don't see any reason more ram couldn't be installed, if greater than 2gb sodimms exist |
00:20.47 | cr2_ | NetRipper: it's zimage |
00:20.56 | NetRipper | cr2_, why "should"? |
00:21.05 | cr2_ | NetRipper: the gzip text strings at top |
00:21.10 | NetRipper | ok |
00:21.23 | *** join/#htc-linux phh_ (~quassel@2a01:e35:2e4b:b2b0:250:8dff:fee1:c793) |
00:21.44 | cr2_ | NetRipper: can i decode .img ? |
00:21.58 | NetRipper | .img? |
00:22.02 | tmzt | what's .img? |
00:22.14 | tmzt | boot.img? |
00:22.27 | tmzt | there should be a reverse mkbootimg |
00:22.59 | cr2_ | yes |
00:23.20 | cr2_ | hehe. microp firmware |
00:23.25 | tmzt | where? |
00:23.34 | tmzt | I really needed that |
00:23.45 | tmzt | I want to be able to flash my old one |
00:23.52 | tmzt | wanted |
00:23.59 | cr2_ | what is yamato ? |
00:24.07 | cr2_ | tmzt: on nexus |
00:24.33 | cr2_ | why is -france.acdb is a different file ?!? |
00:27.15 | *** join/#htc-linux jumoit (~samuel@121.77.87.145) |
00:27.48 | NetRipper | arf my reply on the mailing list is being held for approval |
00:28.01 | NetRipper | cos i used a mail address i didn't reg with |
00:29.29 | NetRipper | it's late.. bedtime, nn :) |
00:29.57 | tmzt | yamato? I don't know |
00:30.11 | cr2_ | tmzt: cant find nexus dmesg |
00:33.22 | *** join/#htc-linux mickey|sofa (~M@e180129213.adsl.alicedsl.de) |
00:33.41 | *** join/#htc-linux mickey_away (~mickey@80.81.242.146) |
00:37.37 | cr2_ | [ 0.000000] Linux version 2.6.29-g4c17402 (android-build@apa26.mtv.corp.google.com) (gcc version 4.4.0 (GCC) ) #1 PREEMPT Fri Dec 11 12:17:58 PST 2009 |
00:37.38 | cr2_ | [ 0.000000] CPU: ARMv7 Processor [510f00f2] revision 2 (ARMv7), cr=10c5387f |
00:38.20 | cr2_ | now, where is NetRipper's dmesg ? |
00:39.48 | cr2_ | mahimahi-microp 0-0066: microp version [0B][15] |
00:40.43 | cr2_ | nandid: 5500bcec maker ec device bc |
00:41.14 | *** join/#htc-linux DarkMasterHalo (~sam@modemcable187.118-80-70.mc.videotron.ca) |
00:41.38 | cr2_ | mmc0: MMC clock 144000 -> 50000000 Hz, PCLK 64000000 Hz |
00:42.04 | cr2_ | hm. how do they get this PCLK value ?? |
00:43.01 | cr2_ | ulpi is different, but that's ok |
00:44.27 | tmzt | they went to 50mhz from 49 |
00:44.37 | tmzt | oh |
00:45.04 | cr2_ | http://forum.xda-developers.com/showthread.php?p=5477033 |
00:45.48 | cr2_ | misc kgsl: firmware: requesting yamato_pm4.fw |
00:46.01 | cr2_ | kgsl is xserver ? |
00:46.17 | cr2_ | misc kgsl: firmware: requesting yamato_pfp.fw |
00:46.37 | cr2_ | hm. the kernel has acdb parser ? |
00:46.51 | cr2_ | acdb: 588 bytes for device 607, rate 48000. |
00:47.26 | *** join/#htc-linux Neo31 (~Neo31@unaffiliated/neo31) |
00:48.29 | tmzt | kgsl is the gl kernel driver |
00:48.29 | *** join/#htc-linux balsat (~balsat@87.72.13.34) |
00:48.48 | cr2_ | so what is yamato ? |
00:48.48 | tmzt | is this on leo? |
00:48.50 | tmzt | ah |
00:48.54 | tmzt | gpu I think |
00:48.54 | cr2_ | no, on nexus |
00:48.59 | tmzt | I had seen it |
00:49.01 | tmzt | same thing |
00:49.15 | tmzt | I just wondered if wm used the term kgsl |
00:50.02 | tmzt | never saw a gpu needing firmware though |
00:50.14 | tmzt | there's an xda thread about the gpu core |
00:50.58 | cr2_ | w100 ? |
00:51.11 | tmzt | don't think so |
00:51.17 | tmzt | z430 or something |
00:51.20 | tmzt | ati chip |
00:51.25 | cr2_ | uploaded microcode |
00:51.35 | Neo31 | Hello, does the MicroSD write/read speed (mSD class 2, 4 or 6) inflect starting android from the memory card on an HTC TP ? |
00:51.52 | cr2_ | ok. tts |
00:53.23 | NetRipper | ~leo-ramconsole |
00:53.24 | apt | hmm... leo-ramconsole is pwf myramconsole 0x2FFC0000 0x40000 |
00:57.03 | NetRipper | cr2_, that line is exactly same in my dmesg: [ 0.000000] CPU: ARMv7 Processor [510f00f2] revision 2 (ARMv7), cr=10c5387f |
00:57.09 | NetRipper | *gone* |
01:18.10 | *** join/#htc-linux M1DLGpc (~M1DLGpc@bb-87-81-252-83.ukonline.co.uk) |
01:28.55 | *** join/#htc-linux jeremychang (~jeremycha@61-30-10-70.static.tfn.net.tw) |
02:04.38 | *** join/#htc-linux makkonen (~makkonen@cpe-66-69-229-9.austin.res.rr.com) |
02:10.00 | *** join/#htc-linux ToAsTcfh (~toast@235-41.200-68.tampabay.res.rr.com) |
02:15.26 | *** join/#htc-linux mcdull (~mcdull@110.4.31.122) |
02:16.43 | *** join/#htc-linux ToAsTcfh (~toast@235-41.200-68.tampabay.res.rr.com) |
02:16.59 | *** join/#htc-linux ToAsTcfh|lappy (~quassel@235-41.200-68.tampabay.res.rr.com) |
02:21.17 | *** part/#htc-linux LargePrime (~chatzilla@99-148-91-126.lightspeed.dttnmi.sbcglobal.net) |
02:29.01 | *** join/#htc-linux enjalot (~enjalot@wg-d232033.dsl.fsu.edu) |
02:32.32 | *** join/#htc-linux MrPippy (~pip@75.42.179.196) |
02:37.51 | *** join/#htc-linux nacho-g (~nacho-g@190.189.189.218) |
02:41.00 | *** join/#htc-linux mastermerlin_ (~merlin@pD957B51C.dip.t-dialin.net) |
03:19.02 | *** join/#htc-linux ToAsTcfh|lappy_ (~quassel@235-41.200-68.tampabay.res.rr.com) |
03:34.23 | *** join/#htc-linux BHSPitMonkey (~stephen@unaffiliated/bhspitmonkey) |
03:35.36 | *** join/#htc-linux randomblame (~randombla@c-71-59-184-96.hsd1.wa.comcast.net) |
03:52.33 | *** join/#htc-linux rzk (~rzk@daemonet.ru) |
03:53.43 | xJDx | EVERYONE SHUT UP NOW! |
03:53.59 | ToAsTcfh | lol |
03:54.03 | randomblame | you shut up |
03:54.03 | xJDx | :p |
03:54.07 | randomblame | :D |
03:54.07 | ToAsTcfh | u get it to work? |
03:54.11 | xJDx | fine :( |
03:54.15 | xJDx | ToAsTcfh: nah :'( |
03:54.27 | xJDx | It worked, but was rebooting |
03:54.29 | ToAsTcfh | dude its right there |
03:54.42 | xJDx | I know! |
03:54.51 | ToAsTcfh | u just gotta get it in the right hands |
03:54.59 | xJDx | your hands? |
03:55.00 | xJDx | lol |
03:55.12 | ToAsTcfh | u cant let the heroc out do the g1 |
03:55.17 | xJDx | fuck this hurts |
03:55.37 | xJDx | (eating) |
03:55.41 | xJDx | I know I cant |
03:55.48 | xJDx | The G1 is like the pwnage tool. |
03:56.11 | randomblame | msm_fb is complicated, far too many things interacting with each other |
03:57.04 | ToAsTcfh | bah buy a g1 fuck it |
03:57.07 | ToAsTcfh | :P |
03:57.14 | ToAsTcfh | n1 i mean |
03:57.18 | xJDx | lol |
03:57.27 | xJDx | I have 2 Magics though |
03:57.37 | ToAsTcfh | then everyone will want the desire |
03:58.13 | ToAsTcfh | anyone get the desire build to boot yet on the g1 or sapphire? |
03:58.25 | xJDx | No idea |
03:58.56 | xJDx | I came out with a release for 2.0 because everyone was asking how |
03:59.14 | xJDx | then I went back in my hole |
03:59.53 | randomblame | if I wanted to make a dummy board panel.c all I'd need to include was an init mddi_dummy_client for the most part right |
04:00.50 | tmzt | just nop the init |
04:00.57 | tmzt | like every other version |
04:01.01 | tmzt | dummy is broken |
04:01.40 | randomblame | the problem is it's not starting at all with my mach type, with diamond mtype mddi.c causes kernel panic |
04:01.59 | randomblame | with my mtype I don't get any messages from mddi or msmfb |
04:02.24 | randomblame | so I want to make it start how do I do that with the init nopd |
04:03.13 | tmzt | look at qualcomm-on-msm gito |
04:03.28 | tmzt | just #ifdef 0 all the actuall init |
04:03.48 | tmzt | #if 0 |
04:16.56 | xJDx | ToAsTcfh: I can't work out if it's the kernel being broken, or if it's how i'm adding it back ;\ |
04:19.06 | xJDx | Im guessing it's the kernel though, since different kernels give different results |
04:23.01 | *** join/#htc-linux reefermattness (~mjg@2002:4c12:4faa:1234:222:3fff:fef1:efa1) |
04:26.41 | randomblame | gitorious is sooooo slow |
04:35.57 | xJDx | Actually |
04:36.05 | xJDx | I think I might know why this one didnt |
04:36.06 | xJDx | work |
04:36.31 | xJDx | @ ToAsTcfh: I built it for a 32A board instead of a 32B |
04:55.41 | *** join/#htc-linux jooprzol_ (~quassel@ip70-189-79-143.ok.ok.cox.net) |
05:20.46 | xJDx | D/AndroidRuntime( 150): >>>>>>>>>>>>>> AndroidRuntime START <<<<<<<<<<<<<< |
05:20.49 | xJDx | FUCKING WOOT. |
05:29.25 | ToAsTcfh | xJDx u got it?! |
05:29.31 | xJDx | Yes, and no |
05:29.36 | xJDx | Yes, it was all perfect |
05:29.41 | xJDx | But then I had a boot loop |
05:29.45 | xJDx | Assumed kernel was bad |
05:29.54 | ToAsTcfh | :( |
05:29.54 | xJDx | rebooted with default kernel |
05:30.06 | xJDx | realised that webkit from system.img is causing the boot loop |
05:30.33 | ToAsTcfh | bah |
05:30.34 | xJDx | so the kernel appears to be correct ;\ |
05:30.51 | xJDx | Im just gonna run a sync, might force webkit from eclair branch |
05:30.55 | ToAsTcfh | dude if u get it right it will work. no doubt about it |
05:31.18 | xJDx | Im not actually even using the same thing anymore |
05:31.27 | ToAsTcfh | :/ |
05:31.28 | xJDx | I have bits everywhere now, I'm in desperation, lol |
05:31.57 | ToAsTcfh | dude phh and mrpippy patches are the way to go. they work |
05:32.15 | xJDx | That's pretty much what I did before |
05:32.17 | ToAsTcfh | my device is not much different then urs |
05:32.44 | ToAsTcfh | if u get the patches ported correctly it will work |
05:33.28 | ToAsTcfh | http://forum.xda-developers.com/showthread.php?t=593015 |
05:33.42 | ToAsTcfh | use the libs out of my rom for opengl |
05:33.52 | ToAsTcfh | egl file |
05:33.59 | ToAsTcfh | in lib |
05:34.16 | ToAsTcfh | anyhow those libs work with 2.x builds |
05:34.56 | ToAsTcfh | once u get the patches ported to ur kernel and they compile it should work without a hitch |
05:35.10 | ToAsTcfh | mrpippys patch alone will work |
05:35.12 | xJDx | Err |
05:35.16 | xJDx | hold up |
05:36.02 | ToAsTcfh | phhs patch only inccreases the framerates |
05:36.40 | ToAsTcfh | phh: patch actually works well too. we get better then stock frame rates |
05:37.07 | ToAsTcfh | 30.8 fps is the best so far |
05:37.58 | ToAsTcfh | with just mrpippys patch we only got 19.8 fps so its a big jump |
05:38.13 | xJDx | thats pretty impressive |
05:38.34 | ToAsTcfh | yeah they both did great work |
05:38.51 | xJDx | http://jd.pastebin.com/d43204b0f |
05:38.54 | xJDx | Check that out |
05:38.56 | xJDx | that's my diff |
05:39.04 | ToAsTcfh | but yeah just use the libs from my rom and use mrpippys patch |
05:40.41 | ToAsTcfh | but yeah u should just be able to apply mrpippys patch then go back and put the rejects in and make the proper changes |
05:41.15 | ToAsTcfh | if ur having trouble with all the merging |
05:41.18 | xJDx | But what are you applying the patch to? Surely it's not the default Android Kernel? |
05:41.31 | *** join/#htc-linux Bry8Star_ (~Bry8Star@cpe-76-170-247-232.socal.res.rr.com) |
05:41.36 | ToAsTcfh | the default htc kernel |
05:41.51 | ToAsTcfh | straight to the tar bal |
05:42.18 | xJDx | oh, the one HTC supplies? |
05:42.25 | ToAsTcfh | yes |
05:42.38 | xJDx | from |
05:42.38 | xJDx | http://developer.htc.com/ |
05:42.39 | xJDx | ? |
05:42.42 | ToAsTcfh | yes |
05:42.58 | xJDx | wow |
05:43.04 | xJDx | I've been using the Android one |
05:43.09 | ToAsTcfh | i g2g man but this is ur best bet |
05:43.17 | ToAsTcfh | trust me |
05:43.53 | ToAsTcfh | or u can backport the whole thing urself. but ill bet a hundred bucks this will take alot less time |
05:44.11 | xJDx | I have the source |
05:44.19 | xJDx | but the merge will f*** me up |
05:44.59 | ToAsTcfh | just apply the patch and go through and put the reject in where they go |
05:45.16 | xJDx | Aight |
05:46.03 | ToAsTcfh | trust me i worked hard on this and then phh: come along and i had it donre in 5 minutes |
05:46.42 | xJDx | Heh |
05:46.42 | ToAsTcfh | it took me longer to compile then apply the patch i think |
05:47.04 | xJDx | If it's not from GIT, can I use git apply? |
05:47.15 | ToAsTcfh | idk |
05:47.48 | ToAsTcfh | patch p0> ur.patch i think |
05:48.40 | xJDx | ;\ |
05:48.47 | xJDx | That wouldnt work for me before |
05:48.50 | xJDx | ill try it again lol |
05:49.38 | ToAsTcfh | patch -p0 < patch-file-name-here |
05:49.45 | xJDx | Yeh |
05:49.49 | xJDx | That's the normal way |
05:50.02 | xJDx | But I couldnt get it to work with the gitorious patches |
05:50.04 | xJDx | ;\ |
05:50.06 | ToAsTcfh | thats a ggod way to start |
05:51.32 | ToAsTcfh | http://www.pastebin.com/m37db1d8c |
05:52.04 | xJDx | Heh |
05:52.05 | xJDx | http://gitorious.org/linux-on-qualcomm-s-msm/linux-msm/commit/a95f49b0c480da1a9d9cdbe0cf1aee6de3828eb5.patch |
05:52.34 | ToAsTcfh | http://www.pastebin.com/m7f611186 |
05:53.28 | ToAsTcfh | yep thats it. put it in the root of ur kernel dir. and call it mrpippy.patch |
05:53.49 | ToAsTcfh | patch -p0 < mrpippy.patch |
05:54.05 | xJDx | Yeh I know how to patch |
05:54.11 | xJDx | but it was going nuts the other day |
05:54.16 | xJDx | with Gitorious output ;\ |
05:54.17 | ToAsTcfh | lol then it should work |
05:54.34 | xJDx | I'm not new to linux lol |
05:54.38 | xJDx | It's my primary OS |
05:54.54 | ToAsTcfh | me too. but this should work |
05:54.55 | xJDx | However, I'm new to kernel shit and GIT (I hate GIT, sure i've mentioned that before though :P) |
05:55.07 | ToAsTcfh | yes u have |
05:55.09 | xJDx | I hope so ToAsTcfh :P |
05:55.28 | xJDx | I might try a make before I patch |
05:55.34 | xJDx | that way I know everything is working |
05:55.40 | ToAsTcfh | anyhow g2g. good luck man and let me know how it goes |
05:55.47 | xJDx | Thanks bro |
05:55.52 | xJDx | will do :) |
05:56.07 | xJDx | 1am? |
05:56.13 | xJDx | Pfft its 5pm :P |
05:56.33 | ToAsTcfh | 1am here yes |
05:56.43 | xJDx | Sleep :) |
05:56.44 | xJDx | Peace |
05:57.14 | ToAsTcfh | peace |
05:57.31 | xJDx | Dont worry |
05:57.38 | xJDx | It wouldn't even build for me |
05:57.38 | xJDx | ahahahaha |
06:03.38 | *** join/#htc-linux f22 (~pcm_2@c-68-82-52-241.hsd1.nj.comcast.net) |
06:03.58 | *** join/#htc-linux nacho-g (~nacho-g@190.189.189.218) |
06:04.53 | ToAsTcfh | htc source doesnt come with a .config fyi |
06:10.32 | xJDx | Yeh I know |
06:10.56 | xJDx | just using the one that came with my vf ROM |
06:11.13 | xJDx | had to choose a few more with make oldconfig |
06:23.52 | tmzt | defconfig |
06:40.43 | Paul_away | do you sleep sometimes? |
06:44.45 | *** join/#htc-linux gigitt (~cad64c2a@gateway/web/freenode/x-tsujnxdrrdmyllxa) |
06:52.57 | MrPippy | oO got it loaded on rhod eth0: Broadcom Dongle Host Driver |
06:54.55 | xJDx | who Paul_away? |
06:55.20 | *** join/#htc-linux M1DLGpc_ (~M1DLGpc@bb-87-81-252-83.ukonline.co.uk) |
06:57.23 | *** join/#htc-linux zachy (~george@ip-85-160-50-50.eurotel.cz) |
07:12.24 | *** join/#htc-linux mcdull (~blackston@n11648167249.netvigator.com) |
07:14.14 | *** join/#htc-linux mcdull_ (~blackston@n11648167249.netvigator.com) |
07:15.42 | mcdull | is the newest kernel fixed DIAM100 3D? |
07:25.29 | *** join/#htc-linux irlolcopter (~irlolcopt@host81-152-193-147.range81-152.btcentralplus.com) |
07:34.44 | mcdull | I got no GUI after setting fb to 3MB. |
07:39.57 | *** join/#htc-linux kiozen (~oeichler@p549211DB.dip0.t-ipconnect.de) |
07:40.13 | *** join/#htc-linux droid001 (~g1@p4FDCBED7.dip.t-dialin.net) |
07:40.21 | *** join/#htc-linux BabelO (~fcr@unaffiliated/babelo) |
07:47.47 | *** join/#htc-linux Speedy2 (~mike@bzq-79-180-18-85.red.bezeqint.net) |
07:48.33 | Speedy2 | www.search2.net |
07:48.57 | *** part/#htc-linux Speedy2 (~mike@bzq-79-180-18-85.red.bezeqint.net) |
07:55.16 | *** join/#htc-linux luminoso (~lumos@av-217-129-141-118.netvisao.pt) |
07:56.44 | *** join/#htc-linux mcdull (~blackston@n11648167249.netvigator.com) |
08:02.22 | randomblame | why do I never see this? console: switching to colour frame buffer device |
08:04.00 | randomblame | I wonder if my config is screwey |
08:37.33 | randomblame | hmm should be console=ttyMSM2 not tty0 |
08:43.52 | *** join/#htc-linux kvaster (~kvaster@93.84.112.80) |
08:57.46 | mcdull | randomblame, what are you exploring recently? |
08:58.18 | randomblame | trying to get a frame buffer working on an unsupported device |
08:58.34 | randomblame | same as usual |
08:58.36 | randomblame | :D |
09:01.57 | *** join/#htc-linux mickey|office (~Mickey@dialbs-092-079-168-007.static.arcor-ip.net) |
09:02.20 | randomblame | I don't understand what calls the panel files under mach-msm |
09:03.36 | mcdull | I see.. how big is the flamebuffer should be? how to calculate? |
09:04.08 | randomblame | my understanding is that it doesn't matter the size or even the memory address |
09:04.32 | randomblame | my problem is that with my mach type msm_fb doesn't start at all |
09:04.34 | mcdull | you mean it doesnt matter of the size or the kernel / driver will re-assign the size? |
09:05.13 | randomblame | it's automagic as far as phh has lead me to believe |
09:05.33 | randomblame | the size doesn't matter, it's 2mb which is plenty big for my resolution |
09:05.56 | randomblame | and the address doesn't matter, even though it's physical not virtual - which puzzles me deeply |
09:06.20 | mcdull | umm.. so as the gpu0. If the calculation in the description is correct, it may requires > 2MB for a WVGA device |
09:06.42 | randomblame | this device is 240x400 |
09:06.50 | randomblame | not really that thirsty for video memory |
09:06.57 | mcdull | me 2.. I have no idea on what the pre-setting of pmem do. where 1MB gpu0 will do the job nicely. |
09:07.28 | randomblame | yeah 1mb would be fine for this device, later that might have to be lowered for extra application memory |
09:07.44 | randomblame | but right now I can't get msm fb to even start |
09:08.04 | mcdull | 1mb is fine on my WVGA device which is hard to believe. Even it is 16bit color. |
09:08.41 | randomblame | I swear to god grep is broken in ubuntu |
09:08.50 | mcdull | well.. 1 MByte seems to be enough on 800x480x16 |
09:09.00 | mcdull | ha.. any proof? |
09:09.09 | randomblame | I've been waiting 10 minutes for it to finish this query |
09:09.43 | mcdull | it takes forever if you are running on virtual machine. |
09:10.30 | randomblame | trying to figure out what file printk's the message "console: switching to colour frame buffer device" |
09:10.40 | randomblame | because I never get that |
09:10.55 | randomblame | yeah virtual machines suck |
09:11.21 | randomblame | I tried installing ubuntu natively but it hates my fake raid array and it killed my windows partition *sigh* |
09:11.52 | mcdull | let me check if I got the message .. |
09:11.56 | mcdull | in dmesg? |
09:12.25 | mcdull | [ 1.863189] Console: switching to colour frame buffer device 60x100 |
09:12.41 | mcdull | cant do a global string for string? |
09:12.46 | randomblame | what messages are before and after that |
09:14.03 | mcdull | [ 1.831298] mddi_epson_probe |
09:14.04 | mcdull | [ 1.831542] setup_vsync |
09:14.04 | mcdull | [ 1.831787] vsync on gpio 97 now 1 |
09:14.04 | mcdull | [ 1.833099] msmfb_probe() installing 480 x 800 panel |
09:14.04 | mcdull | [ 1.863189] Console: switching to colour frame buffer device 60x100 |
09:14.04 | mcdull | [ 1.922973] msm_serial: driver initialized |
09:14.06 | mcdull | [ 1.923767] msm_serial_hs.1: ttyHS1 at MMIO 0xa0300000 (irq = 52) is a MSM HS UART |
09:14.08 | mcdull | [ 1.924957] msm_serial_hs module loaded |
09:14.10 | mcdull | [ 1.929077] loop: module loaded |
09:14.12 | mcdull | [ 1.929687] pmem: 1 init |
09:14.26 | mcdull | should be the msmfb_probe() matter |
09:14.49 | randomblame | ahh so it's dependent on the board-x-panel.c |
09:15.29 | randomblame | I wonder if it matters what console you call in your default.txt |
09:15.39 | randomblame | shouldn't |
09:15.47 | mcdull | but I cannot find the string in the kernel source. |
09:16.04 | randomblame | you're on a new device too right? |
09:19.35 | mcdull | blac100 |
09:20.50 | randomblame | oh |
09:23.37 | *** join/#htc-linux aavan (~aavan@91.183.57.198) |
09:23.41 | *** part/#htc-linux aavan (~aavan@91.183.57.198) |
09:26.34 | MrPippy | arrrgh i can load dhd (compiled from the devour source) (but i have to use polling instead of interrupts), start wpa_supplicant, scan successfully with wpa_cli, but it won't associate to a network, wpa_supplicant spits out ioctl[SIOCSIWESSID]: Invalid argument |
09:34.29 | MrPippy | not sure where to go from there...i guess tomorrow i'll try getting it to work with our in-tree bcm4329 |
09:35.32 | MrPippy | the devour source is actually a newer drop from broadcom, but the nexus one driver is getting bug fixes |
09:37.01 | MrPippy | someone can think about how we can possibly make this work with android userland....afaik the wlan.ko path and name are hardcoded into the source |
09:37.11 | Wout | MrPippy: which board are you working on, rhod? |
09:37.17 | MrPippy | everything else (eth0 instead of tiwlan0, wpa_supplicant args) can be taken care of through shell somehow |
09:37.18 | MrPippy | y |
09:37.20 | MrPippy | yeah |
09:37.54 | randomblame | mr pippy any idea what calls the board-x-panel.c files in mach msm? |
09:38.19 | Wout | cool :) so with my limited knowledge I assume that what you just said means that there is progress on wifi? |
09:38.57 | MrPippy | yeah real progress on wifi, it scanned for networks successfully |
09:39.26 | *** join/#htc-linux playdo (~88020165@gateway/web/freenode/x-fpfgbkcytveecwhr) |
09:39.29 | Wout | wow :) great news |
09:42.58 | MrPippy | randomblame: they get compiled in and run themselves, the device_initcall() at the bottom tells which function runs when it loads |
09:43.43 | MrPippy | and then in the init functions, have to make sure that only one panel init runs for each device |
09:44.30 | randomblame | bah |
09:44.41 | randomblame | so I could just make my own and it would work fine |
09:44.48 | randomblame | it would load itself? |
09:44.50 | *** join/#htc-linux M1DLGpc__ (~M1DLGpc@bb-87-81-252-83.ukonline.co.uk) |
09:46.23 | *** join/#htc-linux Wout (~55927d21@gateway/web/freenode/x-plfewnwaxhfkbegl) |
09:46.51 | MrPippy | yeah as long as its being compiled in |
09:46.59 | randomblame | with make file |
09:47.02 | randomblame | nice |
09:47.21 | randomblame | now I have to figure out the bare minimum to get output |
09:54.26 | *** join/#htc-linux mcdull (~mcdull@203.210.9.159) |
10:06.57 | Wout | Guys, is there any way to make the androd clock sync with winmo? Right now, it's always an hour ahead |
10:08.03 | *** join/#htc-linux goxboxlive (~jrs@mail2.hjellnesconsult.no) |
10:08.38 | Wout | and can someone please explain this to me? |
10:08.41 | Wout | [root@localhost Desktop]# killall -9 adb adb: no process found [root@localhost Desktop]# ./adb shell * daemon not running. starting it now * * daemon started successfully * error: device not found |
10:09.06 | Wout | phone is in debug mode, I'm in the terminal as root, not processes running |
10:09.12 | Wout | (previous adb processes) |
10:10.03 | mcdull | winmo with rtc set to local clock |
10:10.11 | *** join/#htc-linux leobaillard (~leobailla@leobaillard.org) |
10:10.30 | mcdull | android with rtc set to gmt |
10:10.36 | mcdull | for me 8hr diff |
10:10.48 | randomblame | my clock is 72 years off |
10:10.51 | randomblame | :D |
10:11.46 | Wout | linux is hard... :/ |
10:15.41 | *** join/#htc-linux kri5 (~kri5@cowdy.vlmc.org) |
10:40.14 | *** join/#htc-linux kvaster (~kvaster@93.84.112.80) |
10:42.54 | *** join/#htc-linux sxe (~quassel@ip-62-143-102-161.unitymediagroup.de) |
10:43.00 | phh | [10:30:28] <MrPippy> arrrgh i can load dhd (compiled from the devour source) (but i have to use polling instead of interrupts), start wpa_supplicant, scan successfully with wpa_cli, but it won't associate to a network, wpa_supplicant spits out ioctl[SIOCSIWESSID]: Invalid argument <----- nnnooooo :'( I wanted to make it working ... anyway I'll have to take care of the android part I guess. |
10:44.03 | randomblame | yay I'm making progress |
10:50.23 | phh | randomblame: such as ? |
10:50.45 | *** join/#htc-linux MethoS- (~clemens@134.102.106.250) |
10:52.33 | *** join/#htc-linux jazka (jaska@b0xi.net) |
10:59.02 | *** join/#htc-linux kri5 (~kri5@cowdy.vlmc.org) |
10:59.48 | *** join/#htc-linux Wout (~wout85@s55927d21.adsl.wanadoo.nl) |
11:02.09 | *** join/#htc-linux dzo__ (~dzo@mail.marginz.co.nz) |
11:05.52 | *** join/#htc-linux leaigor (~laigor@188.134.16.241) |
11:09.06 | phh | gsensor works \o/ |
11:09.10 | phh | @1sample/second. |
11:09.36 | Paul_Ishenin | it did not work before? |
11:09.37 | Wout | !! cool |
11:09.51 | phh | Paul_Ishenin: not on rhod |
11:09.56 | Paul_Ishenin | ah... |
11:10.05 | Wout | you know what that means...... |
11:10.13 | Wout | papijump! :D |
11:10.14 | Paul_Ishenin | it works not reliavble on diamond too |
11:10.19 | phh | Wout: papijump sucks |
11:10.24 | phh | Paul_Ishenin: Oo |
11:10.28 | phh | it works perfectly on my dia |
11:10.28 | phh | m |
11:10.32 | phh | Wout: abduction is better than that |
11:10.37 | phh | Wout: but only 1sample/second |
11:10.40 | phh | won't be playable |
11:10.45 | Paul_Ishenin | in games - yes. in android interface - no |
11:10.48 | Wout | whats the normal sample refresh rate? |
11:10.54 | randomblame | phh: mddi inits |
11:10.59 | phh | Wout: ~ 50/s I think |
11:11.04 | randomblame | then crashes |
11:11.04 | Paul_Ishenin | sometimes it rotates the applications, sometimes not |
11:11.21 | phh | randomblame: you can see it inits ? |
11:11.27 | phh | I mean on the screen |
11:11.49 | randomblame | no in ramconsole i see its messages |
11:12.08 | randomblame | it crashes before anything good happens |
11:12.19 | randomblame | [ 9.327850] mddi: init() base=0xc500e000 irq=16 |
11:12.19 | randomblame | [ 9.292083] htcraphael_mddi_power_client(1) |
11:12.23 | randomblame | [ 9.292602] mddi cmd send rtd: int 23a000, stat 808063, rtd val e |
11:12.23 | randomblame | [ 9.292816] Unable to handle kernel NULL pointer dereference at virtual address 00000000 |
11:12.23 | Paul_Ishenin | calculator, browser and other application does not rotate often. while the game with ball works fine |
11:12.30 | phh | randomblame: so maybe it doesn't work :p |
11:12.40 | phh | Paul_Ishenin: android bug afaik |
11:12.57 | randomblame | lol at least it starts now |
11:13.19 | Paul_Ishenin | phh, I saw no reports for android phones |
11:13.52 | Paul_Ishenin | reports that it works bad. vice versa I often read how nice it works in the android interface |
11:14.05 | phh | really ? |
11:14.06 | phh | hum. |
11:14.13 | phh | I'll have to check I guess. |
11:14.24 | phh | Paul_Ishenin: just in case, you're not running a january build ? |
11:14.39 | phh | (a major bug has been fixed meanwhile) |
11:14.47 | Paul_Ishenin | no, I don't |
11:15.24 | Paul_Ishenin | I'm runing a build from 8 Feb + own built kernel |
11:15.27 | phh | though, I don't have major troubles with it, just need to hit it a bit in the direction I want it to go to (some sort of hysteresis I guess) |
11:17.28 | Paul_Ishenin | phh. maybe we can include more application with the next release? like calculator, gallery and camera? |
11:18.32 | phh | i have no real opinion about that |
11:18.41 | Paul_Ishenin | I can give you a link for them |
11:19.17 | Paul_Ishenin | they are standard android application but as you know they are hard to find |
11:19.26 | phh | yes |
11:19.39 | phh | which gallery btw ? the awesome looking one in 3D or the other one ? |
11:20.11 | Paul_Ishenin | standard gallery application from android virtual image |
11:20.29 | Paul_Ishenin | do you need a link? |
11:20.34 | phh | yes |
11:28.35 | Paul_Ishenin | imo it will simplify things for others if we include them |
11:40.58 | *** join/#htc-linux Vito89 (~quassel@195.113.242.139) |
11:44.52 | randomblame | wtf, anyone see the date of the latest commit on msm |
11:45.27 | Paul_Ishenin | feb 22 |
11:45.44 | randomblame | Friday April 05 2019 |
11:46.17 | Paul_Ishenin | ah, that one if there for a long time already |
11:46.24 | randomblame | lol |
11:52.35 | *** join/#htc-linux kvaster (~kvaster@93.84.112.80) |
11:56.31 | *** join/#htc-linux GNUtoo (~GNUtoo@host49-13-dynamic.54-79-r.retail.telecomitalia.it) |
12:13.29 | phh | wow. bma150 mostly reports only 0s |
12:14.31 | phh | well, not only. |
12:14.59 | phh | also some garbage |
12:16.54 | phh | why is there no haret for linux :'( |
12:20.21 | Paul_Ishenin | what haret for linux do you need? |
12:20.45 | Paul_Ishenin | to start new android kernel from android? |
12:20.46 | phh | mmutrace |
12:20.56 | phh | to know what my driver actually does |
12:21.58 | Paul_Ishenin | phh, before the release I suppose you need to fix hw3d for diamond? |
12:22.22 | phh | Paul_Ishenin: stop speaking about bad things :( |
12:22.34 | Paul_Ishenin | :)) |
12:22.54 | phh | yes I'll fix it... but I want to fix it for my rhod first ! :'( |
12:23.04 | phh | (but for diam the fix is really easy. for rhod it's totally unknown.) |
12:25.05 | Paul_Ishenin | maybe move the gpu0 block to the case? |
12:25.32 | phh | the problem is not how the code handle it, the problem is to which address set it for WVGA devices |
12:25.34 | Paul_Ishenin | then you can fix rhoduim without breaking the diamond? |
12:25.50 | phh | ah that. yes right. |
12:26.28 | *** join/#htc-linux GeekLad (~GeekLad@adsl-178-230-51.jax.bellsouth.net) |
12:43.17 | *** join/#htc-linux DaveWM (~dave@173-19-136-12.client.mchsi.com) |
13:01.08 | *** join/#htc-linux jazka (jaska@b0xi.net) |
13:11.32 | phh | ah, gsensor definitly works, I only have to calibrate the axis and I can play abduction on this awesome device |
13:12.00 | makkonen | good that you're getting down to the really important stuff. :-) |
13:12.06 | phh | :p |
13:12.23 | phh | but something is really weird ... |
13:13.01 | phh | I'm currently playing with memory allocation (for 3D) too, so I have only 100MB of RAM (less than diamond) |
13:13.09 | phh | and abduction is way smoother than on diamond ... |
13:13.26 | phh | I wonder if SMI is really that fast. |
13:14.23 | *** part/#htc-linux jumoit (~samuel@121.77.87.145) |
13:14.57 | *** join/#htc-linux GNUtoo (~GNUtoo@host49-13-dynamic.54-79-r.retail.telecomitalia.it) |
13:15.07 | phh | I remember I benchmarked it, but I don't remember of the results -_-' |
13:19.08 | *** join/#htc-linux Wr4i7h (~abc@89.180.41.114) |
13:19.23 | Paul_Ishenin | phh, you init change does not work for some reason |
13:19.31 | Paul_Ishenin | I mean camera related |
13:19.39 | phh | Paul_Ishenin: really ? |
13:19.40 | phh | hum |
13:19.51 | Paul_Ishenin | I had to create that devices manyally again |
13:20.02 | makkonen | abduction drags like crazy on raph. gets faster when you get near the top and there's less things to render. |
13:20.31 | phh | makkonen: with default graphic settings, it's just unplayable on diamond |
13:20.38 | phh | on my rhod it's just perfect |
13:22.24 | phh | might be related to the "boost HW" stuff |
13:22.28 | phh | or to pmem in SMI VS EBI |
13:22.40 | phh | I need a raph to test :p |
13:24.17 | phh | Paul_Ishenin: try adding mount -o remount,rw / before the mkdir ? |
13:34.02 | *** join/#htc-linux luminoso (~lumos@av-217-129-141-118.netvisao.pt) |
13:35.06 | *** join/#htc-linux luminoso_ (~lumos@av-217-129-141-118.netvisao.pt) |
13:35.48 | *** join/#htc-linux luminoso__ (~lumos@av-217-129-141-118.netvisao.pt) |
13:37.31 | phh | haha |
13:37.37 | phh | I commited RAM deletion of rhod -_-' |
13:40.08 | Paul_Ishenin | phh, remount does not help |
13:41.41 | Wout | phh, did you push the gsensor fix already? :P want to try me some abduction |
13:41.48 | phh | Wout: yup |
13:41.50 | phh | Paul_Ishenin: hum |
13:41.52 | phh | I'll have to try |
13:42.04 | phh | Wout: should work as is on markinus' donut |
13:42.37 | phh | Wout: you have to change the gsensor_axis in startup |
13:42.41 | phh | gsensor_axis=2 |
13:42.54 | phh | gsensor_axis=2,1,3 |
13:43.06 | phh | this gets abduction working, I'm not sure it's the correct settings though |
13:43.16 | Wout | ok :) |
13:44.02 | makkonen | is there an app that lets you check all axes? I still think mine is wrong. |
13:44.15 | phh | makkonen: motion observer |
13:44.28 | phh | but it doesn't explain which axis is what :p |
13:44.40 | *** join/#htc-linux luminoso (~lumos@av-217-129-141-118.netvisao.pt) |
13:44.40 | makkonen | mm. |
13:44.59 | makkonen | better than nothing, I guess. |
13:58.00 | phh | Wout: works ? |
13:58.34 | Wout | will check in a minute, in the middle of flashing a new winmo rom |
13:58.42 | phh | pfff |
13:58.53 | Wout | hey, this is still my primary phone :P |
13:59.12 | phh | I still have the ROM that ReeferMattness putted on it :p |
13:59.34 | *** join/#htc-linux irlolcopter (~irlolcopt@host81-152-193-147.range81-152.btcentralplus.com) |
14:01.27 | *** join/#htc-linux god206 (~infidel20@unaffiliated/swc666/x-4934821) |
14:01.49 | Wout | I still have to use winmo on a daily basis, so a rom with 6.5.5 and Sense 2.5 makes life a bit easier :) |
14:02.46 | *** join/#htc-linux fleixius (~fleixius@unaffiliated/fleixius) |
14:03.15 | fleixius | mooornnin |
14:03.30 | Wout | ok, booting haret, gsensor here I come |
14:05.29 | *** join/#htc-linux reefermattness (~mattgnm@c-76-18-79-170.hsd1.nm.comcast.net) |
14:05.48 | Wout | phh: would it be difficult to have xdandroid boot up in portrait mode? |
14:05.56 | phh | don't know |
14:05.57 | reefermattness | sup guys? :) |
14:06.07 | phh | reefermattness: gsensor :p |
14:06.28 | reefermattness | aye, am installing abduction as we speak :p |
14:06.35 | phh | :) |
14:06.45 | phh | works only on markinus' donut |
14:06.51 | phh | (and on the eclair that is on my phone) |
14:07.07 | reefermattness | awwww |
14:08.01 | Wout | hmm, doesn't work on glemsom buidl |
14:08.21 | phh | Wout: with latest markinus' donut ? |
14:08.44 | Wout | what's that exactly? |
14:09.03 | phh | not xdandroid |
14:09.27 | Wout | i've only used glemsom's autobuilds so far |
14:09.34 | phh | I mean the *android* build |
14:09.39 | phh | not kernel |
14:10.10 | Wout | ok, this is where my n00bness is on full display |
14:10.26 | Wout | zimage/modules = kernel? so what is the android build? |
14:10.39 | reefermattness | LOL |
14:10.44 | phh | Wout: all that is not kernel :p |
14:10.49 | Wout | right |
14:10.52 | Wout | good to know |
14:10.55 | Wout | :/ |
14:11.44 | phh | AMSS, I HATE YOU |
14:12.01 | Wout | so could someone give me a quick explanation on how the builds are constructed, what's in which files? |
14:12.14 | phh | awww |
14:12.22 | phh | not me. |
14:13.03 | Wout | or perhaps just tell me where I can find markinu's donut android build? |
14:15.57 | *** join/#htc-linux YellowGTO|Nix (~quassel@c-71-226-239-246.hsd1.pa.comcast.net) |
14:16.02 | reefermattness | http://wiki.xda-developers.com/index.php?pagename=TOPAZ_ANDROID |
14:16.07 | reefermattness | I think its that donut Wout |
14:16.17 | reefermattness | feb 13th update |
14:17.15 | reefermattness | I hope ;) |
14:18.01 | phh | 13 feb ? |
14:18.05 | phh | hum not sure it's recent enough |
14:18.43 | reefermattness | damn, thats the newest I could find easily. |
14:20.15 | Wout | hmm, I think it's best if I just wait for the next cab by reef :P |
14:20.31 | reefermattness | LOL shhhh I dont know what Im doing either. ;) |
14:21.03 | phh | Wout: basically, reefermattness will wait for my android release |
14:21.07 | phh | and then he'll do a cab |
14:21.09 | reefermattness | yup |
14:21.14 | Wout | ok :) |
14:21.16 | reefermattness | I do the easy part |
14:21.24 | makkonen | it's trickle-down development. |
14:21.39 | reefermattness | I tell all the ladies im a hacker :p |
14:21.41 | reefermattness | LMAO |
14:21.50 | makkonen | ...you should maybe rethink that. |
14:22.04 | makkonen | not because it's dishonest... just because there are better things to tell the ladies. |
14:22.10 | reefermattness | hahaha |
14:22.27 | reefermattness | Is that why it never works? |
14:23.51 | *** join/#htc-linux jeremychang (~jeremycha@61.57.131.211) |
14:26.56 | phh | could be :p |
14:27.00 | phh | or it's) just you |
14:27.02 | phh | -) |
14:27.09 | reefermattness | :'( |
14:27.15 | reefermattness | sad face |
14:28.49 | reefermattness | Maybe it would help if I was as smooth as a frenchman. :p |
14:29.01 | phh | lol |
14:29.27 | reefermattness | gotta run guys, ttyl! |
14:29.32 | Wout | later |
14:29.52 | *** part/#htc-linux reefermattness (~mattgnm@c-76-18-79-170.hsd1.nm.comcast.net) |
14:30.52 | phh | Paul_Ishenin: please check if I correctly fixed diam's 3D |
14:31.32 | Paul_Ishenin | yes, correctly |
14:31.37 | Paul_Ishenin | I have the same fix locally |
14:31.44 | phh | lol ok |
14:32.00 | phh | Paul_Ishenin: can you test this: set the size to 6MB instead of 7 |
14:32.04 | phh | first with the same bas address |
14:32.09 | phh | and then with base_address+=1MB |
14:32.23 | Paul_Ishenin | I will test |
14:32.26 | phh | thanks |
14:33.04 | phh | Wout: in 6minutes, download http://htcandroid.xland.cz/system.sqsh and use it |
14:33.13 | phh | well, 7 minutes |
14:33.22 | phh | 4 now ? stupid wput. |
14:33.27 | Paul_Ishenin | phh, could that be init.android who removes our msm_camera? |
14:33.39 | phh | Paul_Ishenin: yes |
14:33.42 | phh | aaahhhhhhh |
14:33.43 | phh | right. |
14:33.46 | phh | it mount /dev as tmpfs |
14:33.58 | Wout | k, thx :) |
14:34.34 | *** join/#htc-linux AstainHellbring (~AstainHel@unaffiliated/astainhellbring) |
14:34.46 | phh | groumpf, what's the haret command to get MMU mapping again |
14:35.48 | phh | dump mmu... ok. |
14:40.06 | phh | Wout: uploaded |
14:40.08 | Paul_Ishenin | phh, why you don't want to move pmp_gpu0 config inside the case block? |
14:40.15 | AstainHellbring | morning |
14:40.20 | phh | Paul_Ishenin: I think gpu0 is the same for everyone |
14:40.43 | Paul_Ishenin | but memory location could be different |
14:42.07 | phh | Paul_Ishenin: there must be a SMI location for 3D |
14:42.10 | phh | and there isn't much place in ther |
14:46.06 | Paul_Ishenin | phh, it works with 6MB as well |
14:46.12 | phh | ok |
14:46.26 | Paul_Ishenin | I have 3D in games and boot shows me animation |
14:47.43 | phh | and with offset ? |
14:48.21 | Paul_Ishenin | offset have not tested yet |
14:48.29 | phh | ok |
14:56.41 | *** join/#htc-linux Markinus (~Miranda@gtng-4db04372.pool.mediaWays.net) |
15:01.40 | *** join/#htc-linux Ekips (spike@unaffiliated/ekips) |
15:10.12 | phh | any android msm7k device user around ? |
15:11.39 | *** join/#htc-linux enjalot (~enjalot@2002:80ba:e894:b:223:12ff:fe03:9239) |
15:11.40 | makkonen | you mean g1/magic/etc? |
15:12.17 | *** join/#htc-linux Neo31 (~Neo31@unaffiliated/neo31) |
15:12.27 | phh | not necessarly htc |
15:12.32 | phh | but I asked on #android, it's more logical :p |
15:12.41 | makkonen | :-) |
15:14.04 | phh | ok no answer. I think I'll just try what I want to try. |
15:14.36 | phh | I'd like to know what AXI register phys 0xa8210800 controls :/ |
15:19.00 | Wout | phh: is there any specific reason why xdandroid is 1.6 instead of 2.1, does upgrading the android build require a lot of tweaking/coding/adusting? |
15:19.23 | phh | ?!? |
15:19.30 | phh | xdandroid is 2.{0,1} |
15:19.58 | Wout | oh, then I'm mistaken, sorry :P |
15:20.04 | Wout | nexus one is 2.1 right? |
15:20.17 | phh | yes |
15:24.19 | Wout | was confused about the version numbers because you mentioned the Donut build, which I thought is 1.6 |
15:24.28 | phh | yes. |
15:24.33 | phh | the donut build is no xdandroid build. |
15:24.54 | Wout | ah, ok |
15:25.06 | phh | Wout: so gsensor works ? |
15:30.36 | Paul_Ishenin | phh, it works with offset too |
15:30.44 | phh | ok thanks |
15:30.59 | phh | so I could try MB by MB to see where the memory is supposed to be |
15:30.59 | Paul_Ishenin | I used MSM_SMI_BASE + 2MB |
15:31.38 | phh | does anyone know which DLL handles 3D in windows mobile ? |
15:32.12 | phh | or maybe the name of MS' 3D api |
15:32.32 | Wout | phh: will try in a minute, did a clean android install on sd so need to setup first |
15:32.44 | phh | Wout: ok |
15:34.14 | Paul_Ishenin | phh, could the location in memory also affect camera? |
15:34.21 | phh | yes |
15:35.10 | *** part/#htc-linux zachy (~george@ip-85-160-50-50.eurotel.cz) |
15:36.55 | Wout | phh: gsensor on rhod100 confirmed! |
15:41.52 | *** join/#htc-linux balsat (~balsat@87.72.13.34) |
15:42.31 | *** join/#htc-linux Captnoord (~Captnoord@dc5147a47b.adsl.wanadoo.nl) |
15:42.41 | Captnoord | got my laptop whoooo |
15:42.44 | Captnoord | cheers |
15:42.50 | phh | Captnoord: 3D, NOW |
15:42.55 | Captnoord | nope |
15:42.59 | phh | pff :p |
15:43.08 | Captnoord | you sarcastic piece of coder |
15:43.20 | phh | stop repeating yourself old man. |
15:43.37 | Captnoord | hehe.... same applies to you... old man |
15:43.49 | phh | :p |
15:43.54 | phh | Captnoord: what are you doing with it then ? |
15:44.02 | Captnoord | installing |
15:44.11 | Captnoord | prepairing |
15:44.14 | phh | ah. right. |
15:44.21 | Captnoord | so I can do my normal goto school routine again |
15:44.26 | Captnoord | and finish what I started |
15:44.46 | Captnoord | I will do 3d |
15:44.54 | Captnoord | but first I need to prep |
15:45.41 | Wout | abduction has never been so much fun :P |
15:45.51 | phh | Wout: I prefer it on my diamond |
15:45.55 | phh | the rhod is too big :/ |
15:46.50 | adamw | phh: ooh, you got the gsensor working? neat. anything else? |
15:47.01 | *** join/#htc-linux enjalot (~enjalot@wg-d232033.dsl.fsu.edu) |
15:47.06 | phh | adamw: still fighting with 3D. |
15:47.28 | phh | I'll give wifi a try |
15:47.30 | adamw | who's winning? |
15:47.44 | Captnoord | 3d |
15:47.50 | phh | :'( |
15:48.27 | Wout | phh: you've talked to mrpippy(?) yet? he was making progress on wifi last night |
15:48.33 | phh | Wout: no |
15:48.49 | phh | I'm still waiting for him to come back :p |
15:48.52 | Wout | he got it to init and was able to scan and detect networks |
15:49.00 | Wout | no connection though |
15:49.01 | phh | anyway, the hard part has still be done |
15:49.02 | phh | (ie the android part.) |
15:49.58 | adamw | why's the android part hard? |
15:50.17 | phh | no proper wifi api |
15:50.24 | Wout | :/ |
15:50.27 | adamw | seriously? jeez, what a bunch of jackasses |
15:51.35 | adamw | there's a reason the kernel devs go to the trouble of providing a standard interface across all network interface drivers. le sigh. |
15:52.04 | makkonen | so tiwlan is good just because there happens to be an android-specific... driver? is that the level at which the changes have been made? |
15:52.13 | adamw | i was assuming as soon as we got the interface actually working via iwconfig etc it'd start working in android automatically. |
15:53.07 | makkonen | but doesn't the N1 use a very similar chip? shouldn't there be a good deal of crossover there for android functionality? |
15:53.10 | Captnoord | nothing is easy |
15:53.12 | Captnoord | it never is |
15:53.26 | phh | makkonen: yes, but not binary time compatibility AFAIK |
15:53.32 | phh | should be doable still |
16:12.38 | *** join/#htc-linux mcdull (~mcdull@203.210.9.151) |
16:13.18 | Wout | phh: Realized I hadn't 'thanked' you yet for all the good work |
16:13.23 | Wout | a little something is on its way |
16:13.39 | phh | thanks :D |
16:14.12 | Wout | np, it's cheaper than buying an android phone ;) |
16:15.41 | phh | with rhod it's more like "it's the only way to get a high end android phone" |
16:16.03 | Wout | and that too, dont understand why they havent released a proper qwerty phone yet |
16:16.33 | makkonen | droid/milestone |
16:16.55 | Wout | tp2's keyboard is so much better though |
16:16.56 | AstainHellbring | makkonen I think with the proper comment he wants a decent keyboard too |
16:17.17 | makkonen | ah. I haven't actually spent any time with one. is it that bad? |
16:17.37 | makkonen | oh. no separate line for #s |
16:17.40 | AstainHellbring | dunno but thats the word I've heard |
16:18.27 | makkonen | tp2's keyboard does seem pretty nice. I saw the pictures and thought 'eh, doesn't really look any better than the raphael's'... then I saw it in person. |
16:18.42 | Neo31 | Hello, what is the internal memory available on htc touch pro? can we store files on the 512 rom memory? |
16:18.52 | phh | Neo31: yes but no. |
16:18.59 | phh | but yes. |
16:19.02 | Neo31 | ^^ |
16:19.07 | Neo31 | explain phh |
16:19.17 | Wout | makkonen: they keys themselves are much better, but it's the layout that makes is so good |
16:19.19 | phh | we can write to the nand |
16:19.29 | phh | but when windows mobile detects such writes |
16:19.39 | phh | it erases the nand, ie do a hard reset |
16:19.45 | phh | so you lose everything |
16:19.55 | Neo31 | well the information i need, is if i dont have a memory card, how many pictures i can take with it phh ? |
16:19.57 | phh | but, there is a ROM somewhere, that disables this feature. |
16:20.09 | phh | wait, you're not speaking about linux ? |
16:20.16 | mickey|office | tp2's keyboard seems worse to me than tp1; it's missing the tab key |
16:20.23 | Neo31 | TP WM6 |
16:20.43 | phh | Neo31: .... here it's #htc-*linux* -_-' |
16:20.48 | Wout | mickey_away: you get used tot that, but granted...that was a bad choice by htc |
16:20.59 | phh | Neo31: well, it depends on the ROM you use |
16:21.00 | makkonen | neo31: why are you asking that here? anyway, it's got about 250mb free, so... 100 pictures? |
16:21.24 | phh | mickey|office: ATT added a "ok" key, which can be used as a tab key :p |
16:21.26 | makkonen | or maybe it's more like 150mb free. |
16:21.26 | Neo31 | thanks makkonen , phh that's the answer i was looking for :) |
16:21.26 | mcdull | neo31. do a simple maths pls |
16:21.44 | phh | makkonen: depends on the rom you use :p |
16:21.52 | phh | I guess some can make ~ 350MB free |
16:21.54 | Neo31 | i needed to know the internal memory available on that device |
16:22.07 | mcdull | it really depens on rom use. true |
16:22.07 | Neo31 | that's great |
16:22.28 | phh | but a uSD is really cheap -_-' |
16:22.52 | Neo31 | 350 is great but even a 256Mb is fine |
16:23.00 | mickey|office | heh, fair enough |
16:23.03 | phh | damn, why does radiant uses OpenGL :'( |
16:23.54 | mcdull | even the 2d ball games uses opengl. |
16:24.17 | phh | mcdull: radiant is an asteroid clone -_-' |
16:24.28 | Neo31 | well that means i can take about 50 pics with no memory card (just 50 to leave space for some software and other stuff) |
16:24.29 | Neo31 | :) |
16:24.39 | phh | you know, the thing that runs on 80s consoles. |
16:25.16 | *** join/#htc-linux Forsaken|Desk (~Forsaken@dslb-084-059-000-131.pools.arcor-ip.net) |
16:26.36 | mcdull | what i means is, everything that doesnt really need 3g goes for opengl. |
16:26.48 | phh | 3d* |
16:27.05 | mcdull | 3d shold be. damn soft keyboard. |
16:27.06 | *** join/#htc-linux lucx (~luc@89-115-128-35.cl.ipv4ilink.net) |
16:27.19 | mcdull | omg. typo again. |
16:27.48 | phh | :p |
16:29.18 | mcdull | i have tested different pmem settings. still no luck on 3d. |
16:30.08 | mcdull | i even put 40mb gpu0 and 3mb fb. |
16:30.20 | Captnoord | mcdull: its majorly related to both pmem, imem, grp clock settings, axi* stuff.... |
16:30.30 | Captnoord | pmem imem and grp clocks we can control |
16:30.36 | Captnoord | but the rest |
16:30.37 | Captnoord | is unknown |
16:30.41 | Captnoord | and propbebly the reason |
16:31.12 | phh | Captnoord: if we let wince init the clocks |
16:31.15 | phh | baaaah |
16:31.21 | mcdull | much beyond my knowledge |
16:31.52 | phh | mcdull: much beyond the knowledge of everyone here :/ |
16:31.59 | phh | it has to be reversed |
16:32.01 | *** join/#htc-linux Forsaken|Desk1 (~Forsaken@dslb-084-058-226-169.pools.arcor-ip.net) |
16:33.08 | *** join/#htc-linux Captnoord (~Captnoord@dc5147a47b.adsl.wanadoo.nl) |
16:33.15 | phh | Captnoord: if we let wince init the clocks |
16:33.18 | phh | and we use the same pmem |
16:33.22 | phh | it should work |
16:33.25 | phh | unless clever driver. |
16:33.34 | Captnoord | yup |
16:33.58 | phh | I tried to dump mmu and guess where is their 3D stuff, but no luck :/ |
16:34.09 | Captnoord | hmmm.... |
16:34.20 | Captnoord | I think its still in the 2d ati driver |
16:34.21 | phh | I mean there is no logical scheme |
16:34.27 | phh | 2D ATI ? |
16:34.30 | Captnoord | that is where I didn't really look |
16:34.34 | Captnoord | at there is to much stuff there |
16:35.28 | Captnoord | ahi2dati.dll |
16:35.40 | Captnoord | ahi 2d ati |
16:37.15 | phh | hum |
16:37.23 | phh | I can't see any link between ati and 2D |
16:37.38 | Captnoord | I don't know |
16:37.41 | Captnoord | I gues |
16:37.49 | Captnoord | unless I say that I actualy know it as a fact |
16:37.57 | Captnoord | like how the battery shit works |
16:38.00 | Captnoord | or should work |
16:38.11 | Captnoord | amss describes the method how todo batt stuff |
16:38.15 | Captnoord | like I sayd |
16:38.24 | Captnoord | so best thing is to use rpc for battery |
16:38.25 | phh | there is batt stuff in amss ?!? |
16:38.32 | phh | yes... but no. |
16:38.34 | Captnoord | in the dumps I made yesterday |
16:38.36 | phh | it's not declared by amss |
16:38.44 | Captnoord | k |
16:38.47 | phh | *unless* that's what the TASK_REGISTER dex does |
16:39.11 | Captnoord | lets put it this way.... I don't know what I dumped yesterday ( amss + something extra ) but it wasn't nk.exe |
16:39.24 | Captnoord | and it contained loads and loads of references on functions |
16:39.27 | Captnoord | log messages |
16:39.29 | Captnoord | debug stuff |
16:39.31 | phh | the addresses I gave you has no wince stuff in it |
16:39.37 | Captnoord | k |
16:48.49 | Captnoord | what does it contain then? |
16:49.14 | phh | just amss |
16:49.35 | Captnoord | k |
16:49.44 | Captnoord | then it contains loads of stuff I didn't find in the dlls |
16:51.07 | Paul_Ishenin | phh, http://pastebin.com/d6fc3529b |
16:51.37 | phh | I/mm-camera( 1290): Open Device node : /dev/msm_camera/frame0 |
16:51.37 | phh | W/WindowManager( 1307): App freeze timeout expired. |
16:51.39 | phh | same as me then |
16:52.17 | Paul_Ishenin | I have this long logcat for the first time |
16:52.39 | Wout | apps on winmo are soooooo slow |
16:52.45 | phh | Wout: lol |
16:52.51 | Markinus | phh: hi, is your error on 3d that the screen is still black and no errors in logcat? But can allocate memory etc. ? |
16:52.52 | phh | Wout: and so ugly :p |
16:53.00 | phh | Markinus: yes |
16:53.16 | Wout | phh: TouchTwit looks and feels pretty good for winmo, but starting Twidroid is about 5x faster |
16:53.44 | phh | Markinus: the same thing happens on raph/diam when pmem addresses are wrong |
16:53.50 | phh | so we "just" have to figure out the correct addresses. |
16:53.59 | phh | (or disable AXI VG ?) |
16:55.05 | Captnoord | which we haven't found yet |
16:55.46 | Markinus | phh: hmmmm |
16:59.17 | Paul_Ishenin | phh, this is what in my dmesg http://pastebin.com/d6de718b7 |
17:00.13 | phh | Paul_Ishenin: same as me |
17:00.16 | phh | I think |
17:00.28 | *** join/#htc-linux mcdull_ (~blackston@061093132071.ctinets.com) |
17:01.18 | *** join/#htc-linux [1]Captnoord (~Captnoord@dc5147a47b.adsl.wanadoo.nl) |
17:01.19 | mcdull | phh, any idea why 8MB gpu0 doesn't work on dima100? |
17:01.59 | phh | mcdull: some memory protection I guess |
17:03.51 | phh | hum, I might compare diamond's mmu dump and rhod's eone |
17:04.11 | Neo31 | what does it mean "dump an htc tp rom"??? |
17:05.01 | Captnoord | tp would be touch pro |
17:05.19 | Neo31 | i am interrested inthe word DUMP |
17:05.57 | phh | it's... dump... dumping... dumped... |
17:06.03 | Neo31 | lool |
17:06.05 | Captnoord | not dumb |
17:06.27 | mcdull | you mean the mmu value now using is a diamond one? |
17:06.28 | Wout | extracting the software from the TP and saving it to a computer |
17:06.32 | Neo31 | ok, phh do you have something useful about it? |
17:06.35 | phh | mcdull: no. |
17:06.48 | phh | mcdull: but I know the GPU area of the diamond |
17:06.57 | phh | I can see in the MMU dump what's the link |
17:07.20 | mcdull | GPU area refers to memory address? |
17:07.28 | phh | if you want |
17:07.32 | Neo31 | well here is what I need to know, how can i make a full backup of my htc tp so in case of any problem i can restore everything as it was? |
17:08.13 | Wout | SPB backup |
17:08.39 | mcdull | Neo, you don't really need to back up your rom. Android wont destroy your rom. |
17:09.08 | mcdull | Neo31, if you refers to the data, yes spb or spite backup will do. |
17:09.09 | Neo31 | it seams to be a great software Wout, thanks :) |
17:09.13 | *** join/#htc-linux [1]Captnoord (~Captnoord@dc5147a47b.adsl.wanadoo.nl) |
17:09.39 | Neo31 | ok thanks mcdull |
17:09.54 | Neo31 | i need it not only to use android |
17:10.02 | Neo31 | but you know windows really suck |
17:10.36 | Neo31 | and i expect WM not to be far of desktop windows |
17:10.59 | Neo31 | so a backup would be a great thing to do when i get the phone |
17:11.01 | mcdull | Neo31, but you should really ask those in xda, you will get lots of info there, |
17:11.28 | mcdull | Neo31, if you have no phone yet, why you need a backup? |
17:11.37 | Neo31 | I am Googling for now to know about mobiles phones, WM and android |
17:12.08 | Neo31 | and i'll ask on xda in case i need more details informations before i do something on my tp |
17:12.57 | Neo31 | i asked a friend to buy it for me, and i will get it when i see him in 2 or 3 weeks |
17:13.57 | Neo31 | will be back |
17:14.18 | Neo31 | thanks mcdull, Wout|AFK Captnoord phh |
17:17.30 | *** join/#htc-linux emwe (~michael@86.56.50.180) |
17:26.17 | *** join/#htc-linux kvaster (~kvaster@live.bn.by) |
17:42.26 | *** join/#htc-linux reefermattness (~mattgnm@c-76-18-79-170.hsd1.nm.comcast.net) |
17:42.48 | phh | reefermattness: I'm just thinking... you already put a screen protection on my phone, haven't you ? |
17:43.00 | reefermattness | Indeed. |
17:43.14 | reefermattness | Screen should look brand new underneath that, was on there from the moment I got it. |
17:43.25 | phh | it's already brand new with it :p |
17:43.38 | reefermattness | lol yeah |
17:50.37 | *** join/#htc-linux MrPippy (~pip@75.42.179.196) |
17:51.04 | Neo31 | spb backup just let us backup the data, programs, registry... but not the system itself, in case the system gets infected and slower, what can i do to make a copy of all the rom and resotre it later? |
17:51.59 | reefermattness | Hard Reset Neo, thats what we do. ;) |
17:52.22 | reefermattness | No backup, just get very good at knowing your settings menu |
17:52.26 | reefermattness | :p |
17:52.30 | phh | :p |
17:52.50 | Neo31 | so hard reset is an option is WM reefermattness ? |
17:53.25 | xJDx | How are you guys measuring OpenGLs FPS? |
17:53.32 | phh | xJDx: neocore |
17:53.35 | xJDx | ta |
17:53.38 | Neo31 | I mean in case the phone doesn't start due to a problem in the rom, does that work? the hard reset? |
17:54.04 | phh | Neo31: na, you must use bootloader mode to flash it back |
17:54.05 | reefermattness | Thats what your Hard SPL is for, PM me and ill explain, this really isnt the right chat for this. :) |
17:54.17 | phh | reefermattness: do you know a right chat for that ? |
17:54.26 | phh | I'd be interested too, I'm still a noob in wince :p |
17:54.40 | reefermattness | LOL you can PM me too noob. :p |
17:59.20 | *** join/#htc-linux DarkMasterHalo (~ssoulard@modemcable187.118-80-70.mc.videotron.ca) |
17:59.23 | Neo31 | ok phh :) |
18:00.02 | phh | STUPID PM |
18:00.15 | adamw | phh: http://forum.xda-developers.com/showthread.php?t=550540 , http://wiki.xda-developers.com/index.php?pagename=HTC_Rhodium%2FResets |
18:00.30 | phh | adamw: I know how to flash a rom |
18:00.42 | phh | but that's quite all I know with wince :p |
18:00.42 | *** join/#htc-linux Zoolooc (~fredsibar@p54955975.dip.t-dialin.net) |
18:00.59 | adamw | phh: there's more info in the post if you read it all, and the second is about doing hard resets and stuff. |
18:02.21 | phh | yipi |
18:02.24 | phh | gsensor is killing sleep. |
18:02.53 | phh | (I mean, it's even worst than before :p) |
18:03.23 | adamw | is that possible? :P |
18:04.11 | *** join/#htc-linux m3dlg (~m3dlg@212.183.140.51) |
18:04.21 | phh | yup :p |
18:05.10 | Paul_Ishenin | ToAsTchfh, according to the log you alreay had the next problem "adsp: module VFETASK not enabled before write" |
18:05.16 | Paul_Ishenin | did you fixed that? |
18:05.42 | *** join/#htc-linux M1DLGpc (~M1DLGpc@bb-87-81-252-83.ukonline.co.uk) |
18:08.24 | emwe | phh: hi. as you dare to mention the g-sensor ... it's working now on topa100 + eclair2.0.1 (rootfs and system.sqsh from 2010-02-22). just to let you know. thx! |
18:08.50 | phh | emwe: you mean the one I updated on htcandroid.xland.cz ? |
18:09.12 | emwe | phh: yap. don't tell now that it shouldn't be used ;) just saw it. |
18:09.28 | phh | I already knew it works :p |
18:10.04 | emwe | sorry for the noise :P |
18:10.24 | emwe | got a new phone? |
18:11.45 | phh | no |
18:11.52 | phh | well |
18:11.55 | phh | depend on how new you mean :p |
18:12.02 | phh | I got a rhod three days ago :p |
18:12.22 | phh | well 4 |
18:12.33 | phh | adamw: [ 242.900146] SMEM_SMSM_INT_INFO 0 0 2 |
18:12.36 | phh | what do you think about it ? |
18:12.50 | emwe | congrats to the "new" 4 day old phone ;) |
18:12.56 | *** join/#htc-linux kiozen (~oeichler@rgnb-5d878d2d.pool.mediaWays.net) |
18:12.59 | adamw | phh: i think you are overestimating me :) |
18:13.51 | phh | adamw: I think nothing about it either :'( |
18:15.45 | *** part/#htc-linux DarkMasterHalo (~ssoulard@modemcable187.118-80-70.mc.videotron.ca) |
18:20.45 | *** join/#htc-linux zachy (~george@85.193.4.18) |
18:21.24 | Paul_Ishenin | why do we have 2 msm_camera drivers? one of them compilable and the second one not |
18:22.03 | *** join/#htc-linux Captnoord (~Captnoord@dc5147a47b.adsl.wanadoo.nl) |
18:22.07 | Captnoord | ame |
18:22.08 | Captnoord | lolz |
18:22.15 | Captnoord | re |
18:26.46 | *** join/#htc-linux TimY (~timy@rue92-7-88-164-189-67.fbx.proxad.net) |
18:33.46 | mcdull | phh, blac100 wont boot with the new most kernel. And I have skipped all the bma150 commits. dunno which change broke it. |
18:34.52 | mcdull | will try boot once again. or will do a git bisect tomorrow. |
18:36.22 | phh | mcdull: I guess it's the pmem commit. |
18:36.26 | phh | Paul_away: two ? |
18:36.32 | phh | Paul_away: which ones ? |
18:36.43 | mcdull | phh, you mean the one fixed diam100 ? |
18:36.48 | phh | mcdull: yes |
18:37.05 | mcdull | phh, well, we tried 7MB before, right? and I'd tried it myself. |
18:37.21 | phh | mcdull: I mean it can make the phone crash |
18:37.48 | *** join/#htc-linux JairunCaloth (~m00@c-24-98-17-158.hsd1.ga.comcast.net) |
18:38.26 | mcdull | according to the user feedback, the version before pmem commit doesnt boot as well. |
18:38.33 | mcdull | but I will test it myself. |
18:38.40 | phh | really ?!? |
18:39.03 | phh | and bma150 can't cause any problem |
18:39.56 | mcdull | phh, umm.. is it possible to add the commit description next to the download package as well? it is quite difficult to match them up. |
18:40.17 | phh | mcdull: like moving your cursor on the "build log" logo ? :p |
18:40.25 | mcdull | phh, could be some false alarm. I am testing the version before. |
18:40.52 | mcdull | phh, opps.. sorry, didnt notice that, used to search by date. |
18:42.43 | mcdull | What does the " Merge branch .... " actually do? |
18:43.36 | Captnoord | mcdull: its simple... |
18:43.37 | Captnoord | on git |
18:43.40 | Captnoord | everything is a branch |
18:43.52 | Captnoord | it makes it easier to find stuff if you have fixed something or so |
18:43.54 | Captnoord | nah... |
18:44.01 | Captnoord | when you have a branch on your local machine |
18:44.04 | Captnoord | and you commit it |
18:44.13 | Captnoord | you can merge that branch with the main one |
18:44.16 | Captnoord | that is what you see |
18:44.53 | mcdull | Captnoord, so there are lots of changes expected if I see a merge branch commit? |
18:45.01 | phh | no |
18:45.07 | phh | it just means someone commited, and only then updated his tree |
18:45.30 | mcdull | So like this "Merge branch 'htc-msm-2.6.27' of gitorious.org:linux-on-qualcomm-s-msm/linux-msm into htc-msm-2.6.2" |
18:45.48 | phh | sleep stupid phone |
18:45.49 | phh | sleep. |
18:45.52 | mcdull | does it affect the source tree of gitorious.org:linux-on-qualcomm-s-msm/linux-msm that we are using? |
18:45.58 | phh | I'm your mom, I tell you to sleep. |
18:46.14 | Captnoord | check nk.exe for fake sleep phh |
18:46.17 | mcdull | phh, ok, false alarm on the user feedback. blac100 boots. |
18:46.19 | Captnoord | got some nice info |
18:46.22 | phh | Captnoord: hhuuuuuu ? |
18:46.28 | phh | mcdull: ok |
18:46.45 | Captnoord | fake sleep function |
18:47.13 | phh | what's that ? |
18:47.15 | phh | mmm maybe idle |
18:47.32 | Captnoord | yup |
18:47.41 | Captnoord | possible |
18:49.38 | mcdull | good night, I will try to figure out a pmem setting that will at least boot blac100 with 7MB gpu0 tomorrow. |
18:54.42 | *** join/#htc-linux GregLem (~greg@lal69-5-88-171-98-250.fbx.proxad.net) |
18:56.42 | Captnoord | goes hmmm..... |
18:57.44 | Captnoord | if only I could find where the structure is inited |
18:57.57 | phh | which structure ? |
18:58.34 | Captnoord | there are some functions in the dll talking about pmem |
18:58.35 | Captnoord | but |
18:58.43 | Captnoord | there are also some export functions |
18:58.45 | Captnoord | talking about exi2 |
18:59.04 | *** join/#htc-linux Wr4i7h (~abc@89.180.41.114) |
19:00.29 | Captnoord | also a function called |
19:00.29 | Captnoord | gles_IMEM_alloc |
19:00.43 | phh | alloc ? |
19:00.43 | phh | weird |
19:00.47 | *** join/#htc-linux mickey_away (~mickey@openmoko/coreteam/mickey) |
19:00.47 | Captnoord | which does a pmem_alloc |
19:00.51 | *** join/#htc-linux lucx (~luc@89-115-128-35.cl.ipv4ilink.net) |
19:00.58 | Captnoord | checks if the mem is already there |
19:01.01 | Captnoord | and allocates if not.. |
19:01.54 | Captnoord | MOV R0, R5 |
19:01.54 | Captnoord | BL pmem_malloc |
19:01.54 | Captnoord | CMP R0, #0 |
19:01.54 | Captnoord | STRNE R5, [R4,#0x70] |
19:01.54 | Captnoord | STRNE R0, [R4,#0x74] |
19:02.20 | Captnoord | and r5 is what I want to know where it gets allocated |
19:02.39 | Captnoord | s/allocated/inited |
19:21.09 | *** join/#htc-linux Vito89 (~quassel@gw.loccal.net) |
19:22.21 | *** join/#htc-linux Zoolooc_ (~fredsibar@p54954E84.dip.t-dialin.net) |
19:28.30 | *** join/#htc-linux toi (~toi@d54C2A96D.access.telenet.be) |
19:36.52 | phh | awesome, wakeup on vsync. |
19:37.05 | phh | that's why rhod can't sleep for that long |
19:37.15 | phh | or maybe not vsync should be disabled |
19:37.35 | Captnoord | or the wakeup flag should be removed |
19:37.41 | phh | that works too |
19:37.42 | Captnoord | as it would limit irq's |
19:37.51 | phh | uh ? |
19:37.57 | phh | we should limit irqs yes |
19:38.03 | Captnoord | k |
19:38.05 | phh | but removing the wakeup flag won't limit it |
19:38.15 | Captnoord | k |
19:38.21 | phh | weird, noone sets the wake up flag ?!? |
19:38.32 | phh | maybe wince's setting mmm |
19:42.22 | *** join/#htc-linux GlemSom (~glemsom@0x5da34bca.cpe.ge-1-1-0-1105.sdnqu1.customer.tele.dk) |
19:46.44 | phh | hum for toshiba I had to completly remove the irq ... |
19:46.45 | phh | bah why not. |
19:48.59 | *** join/#htc-linux GregLem (~greg@lal69-5-88-171-98-250.fbx.proxad.net) |
19:52.43 | phh | chamonix: I remember you took care of the 3G problem for some mobile ISP, what's the current status of this bug ? |
19:54.13 | phh | hum I haven't read the bugtracker for a while. |
19:57.58 | *** join/#htc-linux obrienmd (~obrienmd@173.160.148.166) |
20:03.02 | NetRipper | phh, you use github or gitorious? |
20:03.09 | phh | gitorious |
20:03.36 | NetRipper | phh, is it possible to keep new projects "hidden" or "not published" for a while, whilst you work on it.. and then publish it at a later date? or is everything you do publically accessible righta way? |
20:04.08 | phh | I don't think you can do hidden projects on gitorious |
20:04.12 | phh | and on github, you have to pay for hit |
20:04.13 | phh | it* |
20:04.24 | NetRipper | yea |
20:04.34 | NetRipper | was just wondering |
20:04.44 | NetRipper | cant find easily how much github costs though |
20:04.53 | phh | really ? |
20:05.02 | phh | last time I seeked, I found it easily |
20:05.10 | phh | http://github.com/plans |
20:05.17 | phh | pricing link, on the front page. |
20:05.25 | stinebd | heh |
20:05.42 | phh | I think you'll need at least the medium account for a kernel fork :p |
20:05.46 | phh | hi stinebd |
20:05.59 | stinebd | hi phh |
20:11.04 | *** join/#htc-linux cr2_ (~cr2@ip-109-85-91-198.web.vodafone.de) |
20:12.30 | phh | cr2_: new gps idea ? |
20:13.22 | cr2_ | are you sure that the arm9 is reset ? |
20:13.50 | cr2_ | can you do TRACES on the status ? |
20:14.20 | phh | the arm9 reset is what the status gives |
20:14.37 | phh | the bit SMSM_RESET is written (by who ?) in status[1] |
20:14.37 | cr2_ | are you sure that the arm9 is reset ?in wince ? |
20:14.42 | phh | ah |
20:14.43 | phh | no |
20:14.53 | cr2_ | then we do something wrong |
20:15.02 | phh | I'm not sure, doesn't mean it doesn't reset |
20:15.11 | cr2_ | what is this whole CB business ? |
20:15.21 | phh | CB business ? |
20:15.29 | cr2_ | is it parsing the CB response packets ? |
20:15.44 | phh | there is no response packets |
20:15.45 | cr2_ | all magic callback names |
20:16.18 | phh | all rpc stuff about GPS we get are only acks |
20:16.50 | *** join/#htc-linux m3dlg (~m3dlg@212.183.140.36) |
20:17.39 | cr2_ | PDSM_PD_EVENT_END |
20:17.41 | cr2_ | PDSM_PD_EVENT_UPDATE_FAILURE |
20:17.42 | cr2_ | PDSM_PD_EVENT_COMM_BEGIN |
20:17.52 | cr2_ | CB business == EVENT handling |
20:18.26 | cr2_ | where are these events coming from ? |
20:18.38 | cr2_ | from the 31* CB packets ? |
20:19.03 | phh | I guess so |
20:19.42 | cr2_ | maybe be some smsm ? |
20:19.51 | *** join/#htc-linux DaveWM (~dave@173-19-136-12.client.mchsi.com) |
20:20.08 | cr2_ | haret has some uart code for writing to serial console. |
20:20.29 | cr2_ | maybe it makes sense to write a command for reading from the serial port |
20:21.03 | cr2_ | then we can trigger the gps start/stop events from haret |
20:21.08 | phh | right |
20:21.39 | cr2_ | i think it can be ripped out from roadmap |
20:24.08 | cr2_ | RoadMapSerial roadmap_serial_open(const char *name, const char *mode, |
20:24.10 | cr2_ | <PROTECTED> |
20:25.24 | Markinus | phh: do you know where the call stuff is handled? The Topaz isn't waking up on call on pm=0 but I've no Idea in the Moment where I can find this problem . . |
20:25.45 | Markinus | phh: is thgis kernel stuff? |
20:25.48 | phh | Markinus: can you check which IRQs are triggered on call ? |
20:25.58 | Markinus | phh: none |
20:26.02 | phh | ?!, |
20:26.09 | phh | I mean from wince |
20:26.14 | Markinus | yes . .I wundered too . .but there is NOTHING happend |
20:26.22 | Markinus | I know . . |
20:26.47 | phh | see http://htc-linux.org/wiki/index.php?title=Raphael_IRQ |
20:27.04 | phh | my idea, which might be wrong, is that we need to set wake on IRQS 1..4 too |
20:27.15 | phh | but if you say there is no IRQ at all ... |
20:27.39 | Markinus | phh: hmm, I test it one time more . . |
20:27.45 | cr2_ | phh: cdma ? |
20:28.09 | Markinus | cr2_: on my gsm . . |
20:28.11 | phh | cr2_: no |
20:28.17 | phh | cr2_: but 6125 |
20:28.43 | cr2_ | 1-4 are cdma irqs |
20:28.46 | phh | bah |
20:28.47 | phh | ok. |
20:29.23 | cr2_ | NetRipper: ping |
20:30.15 | cr2_ | phh: back to adsp |
20:31.04 | phh | Markinus: at least I get irq 0 |
20:31.07 | phh | but right, no 1-5 |
20:31.11 | phh | cr2_: yup ? |
20:31.23 | cr2_ | phh: let me compare your dump with my .dll |
20:31.27 | Markinus | phh: IRQ0 is comming the hole whole time . . |
20:31.47 | phh | Markinus: it's because it's the AT (among other) irq :p |
20:32.06 | Markinus | phh: ahhh |
20:32.26 | phh | Markinus: so you get all signal level updates on it |
20:33.04 | Markinus | phh: whitch updates?!? |
20:33.11 | phh | Markinus: signal level. |
20:33.23 | *** join/#htc-linux onen|openBmap (~quassel@vbo91-1-89-87-201-85.dsl.club-internet.fr) |
20:33.43 | cr2_ | phh: your dump looks strange. tell me again how you have done it |
20:34.11 | Markinus | phh: sry, I don't know what you mean. I'Ve the last kernel with last xandroid 2.01 . . |
20:34.48 | phh | Markinus: I mean the AT channel, and so the IRQ 0 gets all signal level updates |
20:35.06 | phh | cr2_: lsmod, then pdump output base size |
20:35.12 | phh | s/pdump/vwf/ |
20:35.46 | cr2_ | phh: can you find the physical base addr ? |
20:35.51 | cr2_ | and use pd ? |
20:36.09 | phh | ok |
20:36.11 | Markinus | phh: ahhhhh, ok |
20:36.11 | cr2_ | v2p() should work |
20:37.14 | Markinus | phh: the first IRQ on call what is comming is MDDI, but I think this is for the display update. . . |
20:37.34 | phh | yes it's for display update |
20:37.38 | Markinus | then MDP and ADSP |
20:37.41 | cr2_ | Markinus: in wince ? |
20:37.46 | Markinus | cr2_: yes |
20:38.03 | cr2_ | ok. |
20:38.09 | cr2_ | strange |
20:38.21 | cr2_ | Markinus: maybe gpio ? |
20:38.27 | phh | od -x adspdll.2 |
20:38.27 | phh | 0000000 ffff ffff ffff ffff ffff ffff ffff ffff |
20:38.27 | phh | * |
20:38.27 | phh | 0011000 |
20:38.29 | Markinus | cr2_: I try to find the problems with wakup on call on pm=0 |
20:38.41 | phh | HaRET(9)# pwf adspdll.2 p2v(0x8f917708) 011000 |
20:38.41 | phh | HaRET(10)# |
20:38.45 | phh | ah v2p not p2v -_-' |
20:39.18 | phh | cr2_: http://husson.hd.free.fr/adspdll.2 |
20:39.19 | *** join/#htc-linux lucx (~luc@89-115-128-35.cl.ipv4ilink.net) |
20:39.48 | cr2_ | phh: 403 Forbidden |
20:39.52 | phh | bah |
20:40.05 | phh | fixed |
20:40.55 | cr2_ | phh: 1 page ? |
20:41.09 | phh | cr2_: the size that lsmod showed |
20:41.38 | cr2_ | must be wrong |
20:42.03 | phh | which size do you need ? |
20:42.22 | Markinus | cr2_: GPIOS only I2C |
20:42.29 | cr2_ | phh: let me check on raph |
20:43.13 | cr2_ | Markinus: it seems you miss some irqs, or it is arm9 controlled |
20:43.38 | phh | call answering, arm9 controlled ? that would hurt me a lot :p |
20:43.46 | cr2_ | but then arm9 should provide some notification |
20:44.13 | Markinus | cr2_: I'm checking it the third time, I can't see anythink . . |
20:44.19 | Markinus | hmmmmm |
20:44.34 | Markinus | cr2_: But on haret I see all IRQs or? |
20:44.52 | cr2_ | Markinus: during sleep ? |
20:45.04 | cr2_ | Markinus: on pxa it worked |
20:45.11 | cr2_ | don't know about msm |
20:45.46 | Markinus | cr2_: the devices isn't sleeping, do you mean this is different if device sleeping or not? Activating the IRQs only if the device sleep? |
20:46.07 | phh | that'd be weird |
20:46.28 | cr2_ | it dpeends on the wakeup irq setup |
20:46.42 | cr2_ | i didn't dig too deep here |
20:46.58 | cr2_ | a good chunk of nk.exe is dealing with it |
20:47.06 | phh | Markinus: you can see irq.c code to know what to mmutrace |
20:49.25 | cr2_ | phh: the whole irq setup is a bit of a problem now |
20:49.43 | phh | cr2_: uh ? |
20:50.21 | cr2_ | DGB timer irq, i2c irq, and other things |
20:50.28 | *** join/#htc-linux Captnoord (~Captnoord@dc5147a47b.adsl.wanadoo.nl) |
20:50.28 | cr2_ | ADSP too |
20:50.38 | Captnoord | cheers |
20:50.38 | Captnoord | http://tweakers.net/nieuws/65737/knoppen-hd2-kunnen-door-tweak-dienen-als-notificatielampje.html |
20:50.50 | Captnoord | NetRipper: for president |
20:50.51 | Captnoord | :P |
20:50.53 | phh | Captnoord: dutch sucks. |
20:50.57 | Captnoord | mayor |
20:50.59 | phh | what does that mean ? :p |
20:51.11 | Captnoord | nah tweakers.net is one of the leading tech sites in holland... |
20:51.22 | Captnoord | its about the button app of NetRipper |
20:51.29 | phh | ah. |
20:51.39 | cr2_ | button app ? |
20:51.42 | phh | I've never been in the "press" yet :/ |
20:51.44 | phh | only side effects |
20:51.59 | stinebd | be glad. all it does is lead to spam at your email ;) |
20:52.14 | phh | stinebd: more or less... :p |
20:55.14 | cr2_ | 0x78030000 | 011000 | 0x8f58517c | 5 | adsprtossvccore.dll |
20:55.58 | cr2_ | phh: 0x011000 is 68K |
20:56.08 | phh | ah it's 0x011000 ? |
20:56.14 | phh | ok that's the problem then. |
20:56.19 | cr2_ | submit the patch |
20:56.36 | cr2_ | i took the output from itsutils and it was long |
20:56.51 | Markinus | phh: maybe I register all missing IRQs, go sleep, and look into /proc/interrupts what happends . . |
20:56.53 | cr2_ | but somebody has modified the output |
20:57.11 | phh | Markinus: maybe |
20:58.01 | cr2_ | 8f5851cc | 78030000 00000000 00000000 78032074 |
20:58.21 | phh | cr2_: http://husson.hd.free.fr/adspdll.3 |
20:59.08 | cr2_ | HaRET(5)# print "0x%x" v2p(0x78032074) |
20:59.10 | cr2_ | 0x27246074 |
20:59.23 | cr2_ | hm. a big high ? |
20:59.50 | phh | it's still in memory |
21:00.08 | cr2_ | HaRET(6)# dump mmu 2 0x27246074 0x100 |
21:00.19 | cr2_ | <PROTECTED> |
21:00.21 | cr2_ | 8f100000 | 27200000 | 1MB section | CB AP=1 T=7 |
21:00.22 | cr2_ | af100000 | 27200000 | 1MB section | AP=1 T=7 |
21:00.25 | phh | <PROTECTED> |
21:00.25 | phh | 8f200000 | 27200000 | 1MB section | CB AP=1 T=7 |
21:00.25 | phh | af200000 | 27200000 | 1MB section | AP=1 T=7 |
21:00.39 | cr2_ | ah, i forgot that raph has a hole |
21:00.51 | cr2_ | thoght about 512MB |
21:01.24 | cr2_ | HaRET(7)# pd 0x27246074 0x200 |
21:01.26 | cr2_ | 27246074 | e52de004 e3510001 0b0005d1 e3a00001 | ..-...Q......... |
21:01.28 | cr2_ | 27246084 | e49de004 e12fff1e e92d4030 e24ddf85 | ....../.0@-...M. |
21:01.29 | cr2_ | 27246094 | e1a03002 e1a02001 e59f108c e5911000 | .0... .......... |
21:01.41 | cr2_ | with e4* and e5* looks like a reasonable .text segment |
21:02.36 | cr2_ | yeah. here : |
21:02.40 | cr2_ | 27248984 | 775c3a63 3231366d 3939315f 705c3536 | c:\wm612_19965\p |
21:02.42 | cr2_ | 27248994 | 6674616c 5c6d726f 68706172 5c6c6561 | latform\raphael\ |
21:02.43 | cr2_ | 272489a4 | 5c707363 73736d61 7265735c 65636976 | csp\amss\service |
21:02.45 | cr2_ | 272489b4 | 69645c73 6d5c6761 682e6773 00000000 | s\diag\msg.h.... |
21:02.46 | cr2_ | 272489c4 | 00000000 00000000 00000001 0000002a | ............*... |
21:02.48 | cr2_ | 272489d4 | 0000002a 00000005 0000002a 0000002a | *.......*...*... |
21:02.49 | cr2_ | 272489e4 | 0000002a 0000002a 00000004 00000002 | *...*........... |
21:03.08 | cr2_ | the numbers below the text are module numbers |
21:03.59 | cr2_ | do oyu see it ? |
21:04.27 | cr2_ | and here they end |
21:04.31 | cr2_ | 27248c14 | 00000670 0000066c 00000000 00000000 | p...l........... |
21:04.32 | cr2_ | 27248c24 | 00004033 00444101 09401050 00004005 | 3@...AD.P.@..@.. |
21:04.34 | cr2_ | 27248c34 | 00534c01 09401050 00002078 00534c01 | .LS.P.@.x ...LS. |
21:04.42 | cr2_ | 0x670, 0x66c |
21:04.49 | phh | I'm lost. |
21:05.16 | cr2_ | so you want to trace the 0x272489c4-27248c1c area |
21:05.22 | cr2_ | physical |
21:05.44 | phh | in your case or in mine ? |
21:06.08 | cr2_ | in mine |
21:06.14 | phh | ok I have to redo the math |
21:06.27 | cr2_ | it is 600 bytes starting at 0x272489c4 |
21:06.36 | cr2_ | need to dump mmu first |
21:07.10 | cr2_ | aRET(1)# dump mmu 2 0x272489c4 600 |
21:07.22 | cr2_ | <PROTECTED> |
21:07.24 | cr2_ | 8f100000 | 27200000 | 1MB section | CB AP=1 T=7 |
21:07.25 | cr2_ | af100000 | 27200000 | 1MB section | AP=1 T=7 |
21:07.50 | cr2_ | add mmutrace, and play sound :) |
21:08.19 | phh | ok |
21:09.27 | cr2_ | phh: you have dumped some strange data |
21:09.51 | *** join/#htc-linux jamenlang (~jamenlang@209-193-77-154.mammothnetworks.com) |
21:10.01 | cr2_ | can you pastebin the vd 0xhandle 0x100 ? |
21:10.56 | phh | http://pastebin.com/m324f9b7f |
21:11.35 | cr2_ | ok, what is the physical address for 0x022b55a8 ? |
21:11.55 | phh | 0xbd2b55a8 |
21:12.05 | cr2_ | bd ?? |
21:12.27 | cr2_ | looks like some virtual address to me |
21:12.40 | phh | baah |
21:12.43 | phh | my bad |
21:12.47 | phh | 0x273215a8 |
21:12.55 | cr2_ | this looks much better |
21:13.12 | *** join/#htc-linux Reefermattness (~mjg@2002:4c12:4faa:1234:222:3fff:fef1:efa1) |
21:13.13 | cr2_ | now pd from this addr + 68K |
21:15.03 | phh | pd 0x273325A8 0x20 |
21:15.03 | phh | 273325a8 | 00000000 00000000 00000000 00000000 | ................ |
21:15.03 | phh | 273325b8 | 00000000 00000000 00000000 00000000 | ................ |
21:15.10 | phh | no text nearby |
21:15.57 | cr2_ | strange. works as expected on raph |
21:16.17 | cr2_ | can you run playsound before dumping ? |
21:16.43 | phh | still 0s |
21:17.14 | cr2_ | 0x273215a8 is not 0x273325A8 ? |
21:18.02 | cr2_ | pd 0x273215a8 0x100 |
21:18.14 | phh | uh ? |
21:18.17 | phh | some problems with my calc. |
21:18.23 | cr2_ | [22:12] <phh> 0x273215a8 |
21:18.32 | phh | 273215a8 | e52de004 e3510001 0b00060d e3a00001 | ..-...Q......... |
21:18.32 | phh | 273215b8 | e49de004 e12fff1e e92d4030 e24ddf85 | ....../.0@-...M. |
21:18.32 | phh | 273215c8 | e1a03002 e1a02001 e59f1090 e5911000 | .0... .......... |
21:18.32 | phh | 273215d8 | e58d1210 e6000010 e59f107c e28d0010 | ........|....... |
21:18.33 | phh | 273215e8 | eb0005cf e59f006c e3a05000 e3a03000 | ....l....P...0.. |
21:18.42 | cr2_ | yes, it's ok |
21:19.00 | cr2_ | scroll down for 1-byte in 4byte table |
21:19.18 | phh | uh ? |
21:19.34 | cr2_ | dump 16K |
21:19.46 | cr2_ | pd 0x273215a8 0x4000 |
21:20.18 | cr2_ | then you will see something like that |
21:20.19 | cr2_ | [22:02] <cr2_> 27248984 | 775c3a63 3231366d 3939315f 705c3536 | c:\wm612_19965\p |
21:20.21 | cr2_ | [22:02] <cr2_> 27248994 | 6674616c 5c6d726f 68706172 5c6c6561 | latform\raphael\ |
21:20.22 | cr2_ | [22:02] <cr2_> 272489a4 | 5c707363 73736d61 7265735c 65636976 | csp\amss\service |
21:20.24 | cr2_ | [22:02] <cr2_> 272489b4 | 69645c73 6d5c6761 682e6773 00000000 | s\diag\msg.h.... |
21:20.25 | cr2_ | [22:02] <cr2_> 272489c4 | 00000000 00000000 00000001 0000002a | ............*... |
21:20.27 | cr2_ | [22:02] <cr2_> 272489d4 | 0000002a 00000005 0000002a 0000002a | *.......*...*... |
21:20.30 | cr2_ | note the 2 lower lines |
21:20.51 | cr2_ | you will have 0x33 instead of 2a fair |
21:21.29 | cr2_ | yes. QDSP_MODULE_MAX=0x33 |
21:21.52 | Spirits-Sight | anyone can help with Android install to NAND? |
21:22.28 | phh | got it. I think. |
21:22.40 | phh | cr2_: http://husson.hd.free.fr/d2 |
21:22.47 | phh | @0x27322d48 I guess. |
21:23.19 | phh | so mmutace this area ? |
21:23.39 | cr2_ | phh: these are interesting too |
21:23.43 | cr2_ | 27322a98 | 00000000 00000000 ffffffff 0000057a | ............z... |
21:23.45 | cr2_ | 27322aa8 | ffffffff ffffffff ffffffff ffffffff | ................ |
21:23.46 | cr2_ | 27322ab8 | 00000582 00000586 0000058a 0000058e | ................ |
21:23.48 | cr2_ | 27322ac8 | 00000592 00000596 ffffffff ffffffff | ................ |
21:24.15 | cr2_ | 0xffffffff is QDSP_RTOS_NO_QUEUE |
21:24.15 | Spirits-Sight | first questin, I am trying to update the nbh for the installer/kernel/etc first for Android on Kaiser. which buttons do I push to do a rom update from SD card? |
21:24.34 | cr2_ | phh: the rest are table entries. compare with adsp_6120.c |
21:25.00 | cr2_ | phh: needs to be renamed adsp_6125.c ? like adsp_5200.c is adsp_5225.c ? |
21:25.05 | phh | yes |
21:26.10 | jamenlang | spirits-sight: you should google. |
21:26.14 | jamenlang | Enter Bootloader mode: Hold Camera + Power buttons together and reset with stylus |
21:26.33 | cr2_ | [22:22] <phh> @0x27322d48 I guess. |
21:26.41 | cr2_ | phh: i'd say d3c |
21:26.55 | cr2_ | phh: so you get the idea now :) |
21:27.04 | phh | kind of. |
21:27.09 | Spirits-Sight | jamenlang: I did google and I tryed that however nothing changed on the screen after doing so beside the tri-color screen |
21:27.12 | cr2_ | phh: and you can see how i have filled these tables |
21:27.15 | phh | yes |
21:28.19 | cr2_ | phh: recheck the tables, maybe i've done something wrong |
21:28.35 | phh | I'm sure I'll do more mistakes when trying to check. |
21:28.39 | cr2_ | the module->name of ifdef is pure guess |
21:28.48 | *** join/#htc-linux bzo (~chatzilla@netblock-68-183-234-14.dslextreme.com) |
21:29.05 | cr2_ | but the tables should be as they are in your dll |
21:29.25 | cr2_ | phh: btw, you may try to mmutrace this area, and playsound |
21:29.42 | phh | I checked AUDPPTASK with wince dmesg :/ |
21:29.59 | cr2_ | =4 |
21:30.03 | cr2_ | this is easy |
21:30.06 | phh | ?!? |
21:30.13 | cr2_ | but the cam related stuff is tough |
21:30.15 | phh | it's 3, and you setted it to 3 |
21:30.20 | phh | well for 6125 |
21:30.44 | jamenlang | did you rename the nbh to KAISIMG.nbh |
21:30.52 | cr2_ | eh ? |
21:31.01 | cr2_ | do you see 3 in your table ? |
21:31.17 | Spirits-Sight | jamenlang: Yes, I just reformated/part the card to try again |
21:31.20 | Markinus | cr2_: yes, QDSP_MODULE_AUDPPTASK=0x3 |
21:31.39 | cr2_ | Markinus: i mean in the dump |
21:31.46 | phh | 27322d58 | 0000000b 00000033 00000033 00000003 | ....3...3....... |
21:31.48 | phh | I see a 3 there |
21:31.53 | cr2_ | ah, ok |
21:31.59 | jamenlang | i'm out of ideas then, i've never flashed from sd. |
21:32.07 | phh | not sure if it's this one or not :p |
21:32.15 | cr2_ | would be strange if your dump differed from my dump, and wince dmesg ;) |
21:32.38 | cr2_ | it was 4 on raph afair. |
21:32.49 | phh | yes |
21:32.49 | cr2_ | sorry for the confusion |
21:33.11 | phh | I still don't understand this table :/ |
21:33.21 | cr2_ | Markinus: you see the same numbers ? |
21:33.53 | phh | the table is 0x01, 0x09, 0x33, 0x33, 0x03, ... ? |
21:34.02 | cr2_ | yes |
21:34.05 | Markinus | cr2_: I need some more time for this . . but I'M on it . . |
21:34.08 | cr2_ | the array length is 32 ? |
21:34.31 | cr2_ | the module array |
21:34.44 | phh | don't know. |
21:36.10 | cr2_ | phh: anyway, the mmutrace will be the most interesting |
21:36.27 | phh | <PROTECTED> |
21:36.27 | phh | 8f300000 | 27300000 | 1MB section | CB AP=1 T=7 |
21:36.27 | phh | af300000 | 27300000 | 1MB section | AP=1 T=7 |
21:36.29 | phh | which one ? |
21:36.32 | cr2_ | to see module=3 and the respective queue |
21:36.38 | Spirits-Sight | jamenlang: for the kaiser 100 what panel type should I use? I ask this as when I boot the phone it has some of the first splash showing right at the bottom |
21:37.13 | cr2_ | phh: dump mmu 2 0x27322a88 0x1000 |
21:37.35 | phh | <PROTECTED> |
21:37.35 | phh | <PROTECTED> |
21:37.35 | phh | 8f300000 | 27300000 | 1MB section | CB AP=1 T=7 |
21:37.35 | phh | af300000 | 27300000 | 1MB section | AP=1 T=7 |
21:37.38 | Spirits-Sight | jamenlang: when I was using haret I had it set at the default 0 and it was good but now that its a rom I don't know what the standard/default would be? |
21:37.50 | cr2_ | -0x27322a88+0x27322ffc |
21:38.18 | phh | <PROTECTED> |
21:38.18 | phh | <PROTECTED> |
21:38.18 | phh | 8f300000 | 27300000 | 1MB section | CB AP=1 T=7 |
21:38.18 | phh | af300000 | 27300000 | 1MB section | AP=1 T=7 |
21:38.29 | cr2_ | 1396 bytes |
21:38.42 | cr2_ | phh: ok, you need to mmutrace these |
21:38.51 | phh | all four ? |
21:38.58 | phh | one memory page ? |
21:39.17 | jamenlang | because the kaiser is booting from nand, i can only assume that whoever made that nbh is in control of the settings now. |
21:39.35 | cr2_ | phh: the page with these settings |
21:40.02 | cr2_ | phh: you would like to set the mask exactly to that region |
21:40.20 | cr2_ | otherwise you may be really spammed |
21:41.38 | cr2_ | in the ideal world you should see ldr for 3, and for some other number |
21:42.04 | phh | ok my brain is totally out... which mmutrace address should I use ? |
21:42.28 | cr2_ | try each of them one by one |
21:43.25 | cr2_ | phh: addlist mmutrace virt+0xa88 1400 |
21:43.41 | cr2_ | phh: addlist mmutrace virt+0xa8c 1400 |
21:43.50 | cr2_ | sorry. c is better |
21:44.12 | phh | Warning! Tracing non-section mapping (02200000) not well supported |
21:44.16 | phh | one day I'll understand what this means |
21:44.17 | cr2_ | i'll try 4k pages first |
21:44.33 | cr2_ | yeah, dzo wanted to fix it. |
21:44.40 | phh | 000049: Can't emulate insn access at 0223dec4 |
21:45.06 | cr2_ | try with 76 |
21:45.23 | phh | 76 what ? size ? |
21:45.33 | Spirits-Sight | jamenlang: or anyone that knows more: see I copyed KAISIMG.NBH to the mSD and then did reset with power+cam and it shows on the screen the tri-colors on top it says sercurity unlock, then under that KAIS1*0, SPL-3.56.Hard, CPLD-8 and the bottom it shows serial. Is this what it should say? |
21:45.45 | cr2_ | addlist mmutrace 0x7644fa8c 1400 |
21:46.09 | phh | ah this 76. |
21:46.40 | phh | cr2_: frozen \o/ |
21:47.01 | cr2_ | phh: i wish we could use axi traffic sniffer, like leo nk.exe |
21:47.10 | phh | oh you can. |
21:47.13 | phh | but I think we have no doc about it |
21:47.14 | cr2_ | phh: yeah. Warning! Tracing non-section mapping (02200000) not well supported |
21:47.25 | phh | cr2_: I cleared it |
21:47.40 | phh | the only mmutrace is the last one you gave me 0x76..... |
21:47.42 | cr2_ | phh: we can see what leo does |
21:47.55 | cr2_ | and it froze the device ? |
21:48.00 | phh | yes |
21:48.04 | cr2_ | ok |
21:48.14 | cr2_ | bt was not traceable on raph too |
21:48.27 | phh | bt is a hardware register |
21:48.34 | phh | this memory location is plain memory |
21:49.17 | cr2_ | well, try to compare the raw numbers then. |
21:49.23 | cr2_ | from the dump |
21:49.32 | phh | are they supposed to be in the same order ? |
21:50.00 | cr2_ | compare them with my numbers in the adsp_ driver |
21:50.21 | cr2_ | i think they are from image 5 to 0 |
21:50.44 | cr2_ | inside the image the order is from top to bottom |
21:50.58 | cr2_ | but maybe i got the 5 to 0 order wrong |
21:51.05 | cr2_ | though i doubt it |
21:51.16 | cr2_ | wince dmesg says image=1 is used |
21:51.17 | jamenlang | spirits-sight: fat32? |
21:51.33 | Spirits-Sight | jamenlang: yes |
21:51.43 | Spirits-Sight | 1GiB |
21:51.47 | cr2_ | phh: and image=1 in my list is pretty complete, so it was a good guess on my side |
21:52.26 | cr2_ | phh: if you are crazy, you can patch this table in ram :) and see if playsound will still work :) |
21:52.40 | jamenlang | i guess you're supposed to hit the power button to start flashing. |
21:52.43 | phh | if you call that crazy, I won't try. |
21:53.14 | cr2_ | phh: linux didn't crash in such situation |
21:53.20 | phh | I know |
21:53.25 | cr2_ | phh: there was just no sound |
21:53.31 | MrPippy | wow i missed this, android git has a kernel tree for nvidia tegra and google is writing the code |
21:53.31 | phh | but I don't really understand this table |
21:53.39 | phh | MrPippy: oO |
21:53.41 | phh | cool |
21:53.58 | phh | MrPippy: I didn't see you joined ... so about wifi |
21:54.08 | Spirits-Sight | jamenlang: nothing? |
21:54.27 | jamenlang | is it a 1GB kingston? |
21:54.29 | phh | MrPippy: if you think the driver I putted is wrong, feel free to change it |
21:54.36 | Spirits-Sight | should the ext be cap letters .NBH or .nbh |
21:54.54 | jamenlang | good question, all the tutorials i've seen have lowercase |
21:55.01 | MrPippy | nah i'm about to try it, see if it works better with wpa_supplicant than the other one |
21:55.15 | phh | Spirits-Sight: fat doesn't care about case. |
21:56.47 | Spirits-Sight | phh: I didn't think so but you never know :( any idea, should I redo the hardSPL? |
21:56.58 | phh | no |
21:57.52 | bzo | woo hoo. I'm running my diamond overclocked at 710mhz! |
21:58.35 | bzo | android is much more usable at this speed |
21:58.56 | phh | bzo: hum, how ? |
21:59.03 | phh | I can't see how you can do such a clock |
21:59.07 | phh | change pll ? |
21:59.29 | bzo | changed pll2 to a 1x multiplier |
21:59.43 | jamenlang | can i underclock mine to preserve battery life ;p |
21:59.50 | bzo | I originally oc at the default 2x, the the max speed is 1.2g/2 |
21:59.52 | phh | jamenlang: it's dynamically underclocked |
22:00.03 | bzo | yep, most of the time it's running at 128mhz |
22:00.13 | jamenlang | good to know |
22:00.26 | phh | jamenlang: we go further than wimo in low clocks |
22:00.31 | phh | iirc |
22:00.41 | jamenlang | that's what i like to hear |
22:00.42 | phh | not sure if it's useful. |
22:00.58 | cr2_ | phh: is it possible to run at 19.2 MHz ? |
22:01.02 | phh | cr2_: yes |
22:01.18 | jamenlang | yeeeeeaah |
22:01.20 | cr2_ | nice. i wonder how long is the battery life then |
22:01.31 | bzo | afaik, it runs at 19.2 during suspend |
22:01.32 | cr2_ | with lcd off, and gps on. |
22:01.36 | phh | jamenlang: most of the time it's) @ 32kHz you know. |
22:01.46 | phh | bzo: hum I have to check |
22:01.58 | phh | bzo: but cr2_ speaks about on :p |
22:02.00 | phh | not sleep |
22:02.15 | phh | wait for interrupt is done at 19.2 |
22:02.39 | Markinus | phh: is there a vreg for the CPU voltage too? A bit lower . . . where nice |
22:02.47 | phh | Markinus: see acpuclock.c |
22:02.56 | phh | but you can't set it clearly |
22:03.01 | bzo | phh: i will have a patch for you soon. I've added the oc code to acpuclock.c that integrates cleanly with the dynamic clocks |
22:03.01 | phh | you just have 7 values |
22:03.04 | phh | you don't know what they are |
22:03.10 | cr2_ | phh: 3 i think |
22:03.10 | jamenlang | 32K love it |
22:03.11 | phh | bzo: cool |
22:03.24 | phh | cr2_: well, theorically 7, but only 3 are used in our code |
22:03.43 | cr2_ | phh: androids say there are 3 in real life |
22:03.45 | phh | jamenlang: when the system is in sleep, the consumption goes almost everywhere but in the CPU :p |
22:03.48 | bzo | yeah in acpuclock the acpu_freq_table shows all the pll/clock/multiplier/voltage etc settings we are running |
22:03.51 | phh | cr2_: some devices has 7 |
22:04.12 | phh | bzo: but rhodium has "turbo HW", I want to play with that :p |
22:04.19 | cr2_ | phh: 3 bits, but some bits may mean the same voltage |
22:04.34 | phh | cr2_: maybe |
22:04.44 | cr2_ | phh: what is turbo HW ? |
22:04.46 | emwe | phh: turbo hw as in CONFIG_TURBO_MODE? |
22:04.47 | bzo | phh: i was looking at the heroc turbo code, all it seems to do is increase some intermediate clocks and the axi bus |
22:04.51 | phh | emwe: yes |
22:04.56 | cr2_ | phh: your pll1 is already at a higher clockrate |
22:05.00 | cr2_ | 960 vs. 768 |
22:05.11 | phh | cr2_: ebi @ 200MHz instead of 166 |
22:05.12 | phh | I think |
22:05.19 | emwe | phh: backported that to linux-msm-topaz/htc-msm-2.6.27-exp. |
22:05.31 | phh | emwe: oh really ? |
22:05.32 | cr2_ | phh: visible with msmclkkhz ? |
22:05.42 | phh | cr2_: I don't know which one is ebi |
22:05.48 | emwe | phh: http://gitorious.org/~mweirauch/linux-on-qualcomm-s-msm/linux-msm-topaz/commits/htc-msm-2.6.27-exp |
22:05.54 | cr2_ | but you can see the difference in output |
22:06.05 | phh | emwe: thanks, I'm a bit lost in all the forks. |
22:06.15 | cr2_ | hehe |
22:06.24 | emwe | phh: give all commit access, let them screw the tree and be gone with loosing track of all the forks ;) |
22:06.27 | phh | http://gitorious.org/linux-on-qualcomm-s-msm ----> 11 forks. youhou. |
22:07.08 | cr2_ | emwe: lol. then the tree will look even workse that it is already :) |
22:07.16 | phh | emwe: hum your commits seems cleaner now. you might get access too. |
22:07.54 | emwe | phh: oh, they've been looking dirty at some point?! :P |
22:08.00 | phh | emwe: iirc yes |
22:08.34 | emwe | phh: might be. might also have been the diff marathon in the beginnings of the topaz-get-going-madness |
22:08.41 | phh | yup. |
22:09.07 | cr2_ | emwe: out tree is really dirty. but the g1 tree is not much better |
22:09.32 | emwe | phh: i wouldn't mind getting access, but i'd rather do some basics. rather cleanup/consolidation. |
22:09.44 | phh | emwe: oh really ? |
22:09.47 | phh | you get instant commit access. |
22:09.50 | emwe | lol |
22:09.52 | phh | well |
22:09.58 | phh | instant as in wait for me to find the option in gitorious |
22:10.13 | emwe | phh: well, that could wait till tomorrow i guess ;) |
22:10.27 | emwe | getting late and work is calling at 7am |
22:10.35 | phh | added |
22:10.39 | emwe | oh, thank you! |
22:11.37 | emwe | will get back to you about what i would imagine... i still have common memory bank setup and pwrsink in mind. |
22:11.58 | phh | I can't see how you can do that, but would be cool |
22:12.21 | phh | well for pwrsink yes you can. |
22:12.21 | cr2_ | pwrsink is some androidism ? |
22:12.26 | phh | cr2_: yes |
22:12.29 | cr2_ | ok |
22:12.35 | phh | cr2_: estimation of the power consumption using drivers' reports |
22:12.41 | phh | sounds totally stupid to me. |
22:12.41 | Spirits-Sight | jamenlang: its press the cam and reset that worked |
22:13.01 | phh | ok, time to have fun |
22:13.05 | phh | let's review all of forks ! :p |
22:14.09 | emwe | phh: linux-msm-topaz/htc-msm-2.6.27 is clean, except for the topaz pwrsink-table. the *-exp-tree has only the last three commits of intresset. just so you know. |
22:14.16 | NetRipper | hmm |
22:14.30 | phh | emwe: you have commit access, you merge :p |
22:14.36 | NetRipper | new leo OS roms don't boot if you don't update the radio rom first |
22:14.45 | NetRipper | i wonder what changed.. incompatible ril? |
22:14.56 | cr2_ | NetRipper: if you have the same cpu as nexus, why can't you boot zimage ? |
22:14.56 | emwe | phh: ok. will do. tomorrow then. thx! |
22:14.59 | emwe | gnight everyone. |
22:15.06 | cr2_ | NetRipper: security settings ? |
22:15.11 | phh | cr2_: can nexus boot zImage ? |
22:15.11 | NetRipper | cr2_, maybe it's somehow protected |
22:15.23 | Markinus | emwe: I didn'T commit the pwrsink table because I wasn't sure that this it correct for Topaz . . . a bit different hardware |
22:15.23 | NetRipper | phh, cr2 seems to have verified the nexusone boots zImage |
22:15.27 | phh | ok |
22:15.28 | NetRipper | :) |
22:15.30 | cr2_ | NetRipper: there is some region id gpio |
22:15.56 | cr2_ | phh: if you provide me unimg program |
22:15.57 | emwe | Markinus, it was in IIRC. i just changed to heroc/desirec/tattoo. |
22:16.14 | cr2_ | phh: then i'll send you zimage :) |
22:16.23 | phh | cr2_: I already searched, no clue :/ |
22:16.24 | emwe | anyways, gnight. |
22:16.39 | Spirits-Sight | OK, now for my next question is there suppose to be any white on the splash screen (kaiser/Tilt), and or screen that looks lines that fly by real fast before it shows the pingun <-spelt wrong ? |
22:16.44 | cr2_ | phh: should be open source ? |
22:16.51 | phh | cr2_: yes |
22:17.31 | Markinus | emwe: night |
22:18.53 | bzo | phh: here is the overclocking patch - http://pastebin.com/m121af19e |
22:18.57 | phh | thanks |
22:19.27 | bzo | in the config, you have to increase CONFIG_MSM_CPU_FREQ_ONDEMAND_MAX |
22:19.34 | bzo | otherwise, it will not exceed 528mhz |
22:19.42 | NetRipper | cr2_, hm ok maybe the radio changes region? |
22:19.46 | bzo | I set mine to 999999 |
22:19.51 | phh | bzo: noone uses PLL2 ? |
22:20.05 | cr2_ | NetRipper: no idea |
22:20.07 | bzo | afaik, it is only used for 528mhz devices for the cpu clock |
22:20.16 | phh | bzo: ok |
22:20.23 | phh | cr2_: you confirm ? |
22:20.26 | cr2_ | NetRipper: but if arm9 reports wrong region, you get a reboot dex call from wince kernel ;) |
22:20.42 | cr2_ | phh: pll2 is used only by arm11 |
22:20.48 | phh | cool then |
22:21.02 | phh | bzo: I think ToAsTcfh might be interested in your patch too :p |
22:21.08 | ToAsTcfh | yes |
22:21.12 | ToAsTcfh | :D |
22:21.22 | phh | and it's pretty easy to merge |
22:21.24 | cr2_ | phh: at least all known peripheral devices use pll0 and pll1 |
22:21.25 | bzo | seems like it should work on any 528mhz device that has the same CLK_CTL registers |
22:21.32 | ToAsTcfh | where too? |
22:21.39 | NetRipper | cr2_, ah |
22:21.42 | phh | ToAsTcfh: http://pastebin.com/m121af19e |
22:21.47 | ToAsTcfh | i have the turbo |
22:21.55 | phh | ToAsTcfh: not related |
22:21.57 | phh | turbo is PLL1 |
22:22.02 | ToAsTcfh | sweet |
22:22.07 | phh | or ... not ? |
22:22.09 | NetRipper | does anyone here know how cpu load is calculated on arm processors? |
22:22.13 | NetRipper | i heard the cpu doesn't natively support it |
22:22.14 | phh | damn, don't remember I have to check |
22:22.20 | ToAsTcfh | so what doees it do to oc |
22:22.21 | phh | ToAsTcfh: well pll2 isn't the one for turbo. |
22:22.25 | phh | ToAsTcfh: changes pll2. |
22:22.29 | bzo | phh: all the turbo settings for the heroc seemed to be only on pll1 |
22:22.33 | ToAsTcfh | too? |
22:22.36 | ToAsTcfh | 2x |
22:22.38 | phh | bzo: really ? |
22:22.44 | Markinus | phh: I reduced the VDDs for all steps and the device is working well . . |
22:22.56 | cr2_ | NetRipper: load average is number of running processes. how it is related to hw ? |
22:23.03 | phh | Markinus: as cr2_ said, we have no clue of what these values are actually doing |
22:23.05 | bzo | aside from changing the axi clk also on 528 on pll2 |
22:23.21 | phh | bzo: axi clk is on pll2 ? |
22:23.36 | bzo | phh: i mean on the 528 mhz pll2 setting |
22:23.52 | NetRipper | cr2_, you know the cpu usage gauge in windows or in linux, is it "virtualized" for arm processors? |
22:23.56 | ToAsTcfh | bzo: whats this increase the clock too? |
22:24.08 | NetRipper | x86 processors have regs that tell you the cpu load... arm processors dont (as i heard) |
22:24.10 | ToAsTcfh | ans or up too? |
22:24.17 | bzo | ToAsTcfh: whatever you want :) |
22:24.20 | cr2_ | phh: clock distribution is nonobvious. i still don't get hot they obtain 64MHz for SD pclk |
22:24.24 | bzo | it's a module param |
22:24.27 | phh | cr2_: ok |
22:24.28 | ToAsTcfh | whats the best u guys get? |
22:24.47 | stinebd | goes for 7GHz |
22:24.51 | bzo | hehe, I think I'm the only one that has ever tried so far, just created the ptach |
22:24.57 | ToAsTcfh | phh: no luck on the camera yet? |
22:25.01 | phh | ToAsTcfh: no |
22:25.02 | cr2_ | NetRipper: can you boot on leo now ? |
22:25.03 | bzo | 710mhz ended being a little unstable for me |
22:25.07 | bzo | trying 650mhz |
22:25.15 | ToAsTcfh | sweet man ill try |
22:25.16 | phh | bzo: and BIG question |
22:25.22 | phh | does it warm your hands enough ? |
22:25.27 | ToAsTcfh | notice a difference? |
22:25.28 | phh | it's cold outside. |
22:25.31 | NetRipper | cr2_, no.. this is just an unrelated question so far :) something to do at work ;) |
22:25.41 | cr2_ | NetRipper: what do you think about removing swordfish and wrong drivers from defconfig ? |
22:25.50 | bzo | phh: doesn't seem to make a big difference in heat, but I haven't run that long |
22:25.55 | phh | bzo: ok |
22:26.12 | NetRipper | cr2_, i've removed swordfish, but doesnt help... i'm also not sure what drivers to remove.. removing some drivers breaks the build |
22:26.20 | phh | bzo: with the same voltage, I don't think it heats a lot more |
22:26.25 | ToAsTcfh | does it bence any better? |
22:26.26 | phh | so if it's stable, it's all good for us :D |
22:26.39 | ToAsTcfh | bench |
22:26.46 | cr2_ | NetRipper: on linux it's just scheduler statistics |
22:26.59 | bzo | phh: yeah, worst case is probably linear relationship 650/528 = 23% more power consumption |
22:27.04 | phh | bzo: benchs ! benchs ! benchs ! benchs ! benchs ! benchs ! |
22:27.18 | phh | bzo: neocore results would be interesting |
22:27.23 | *** join/#htc-linux cr2_ (~cr2@ip-109-85-91-198.web.vodafone.de) |
22:27.26 | ToAsTcfh | yet neoccore |
22:27.30 | bzo | phh: I've been using "CPU benchmark" to double check it is not my imagination |
22:27.37 | phh | bzo: ok |
22:27.43 | bzo | at 528mhz, it takes about 5500ms a run |
22:27.47 | NetRipper | cr2_, hm ok |
22:27.49 | bzo | at 650mhz it takes about 4500ms |
22:27.54 | ToAsTcfh | woot |
22:28.04 | bzo | but the results are very inconsistent run to trun |
22:28.06 | phh | bzo: I asked how you do that, because ToAsTcfh already tried an "overclock" which was totally ridiculous :p |
22:28.30 | Spirits-Sight | OK, does any one have a picture of the booting of a kaiser phone with NAND? |
22:28.32 | bzo | phh: you mean didn't work? |
22:28.39 | ToAsTcfh | yeah it was the dream devs shit so i want no credit for it :P |
22:28.50 | phh | ToAsTcfh: yeah sure :p |
22:28.51 | Spirits-Sight | the splash screen is what I would like to see if it is right? |
22:28.53 | phh | bzo: { 528000, ACPU_PLL_2, 2, 1, 132000, 3, VDD_7, 128000, 0, 11, -1 }, |
22:28.54 | phh | see this ? |
22:28.59 | phh | change the 528000 to 768000 :p |
22:29.04 | ToAsTcfh | lol |
22:29.04 | stinebd | hah |
22:29.05 | phh | and here's there patch |
22:29.06 | cr2_ | NetRipper: if the hardware can do it, then the cpu has a little less work |
22:29.10 | ToAsTcfh | renamed the shit |
22:29.13 | bzo | phh: lol |
22:29.26 | NetRipper | cr2_, well it seems ARM doesnt have support for this in the hardware |
22:29.28 | ToAsTcfh | bzo: are u using setcpu? |
22:29.37 | ToAsTcfh | or just echoing it? |
22:29.46 | bzo | ToAsTcfh: no, why? |
22:29.59 | phh | Markinus: I can't see anything in the docs about the A11 VDD registers :/ |
22:30.03 | ToAsTcfh | just wondering how ur settinfg the clock |
22:30.26 | phh | ToAsTcfh: it's using classic cpufreq stuff |
22:30.27 | bzo | I'm reconfiguring the low level MSM clock PLL through the CLK_CTL registers |
22:30.53 | Markinus | phh: Hmm, I have running my device now, and It feels colder . . hmmm Imagination or reality . . |
22:31.04 | phh | Markinus: I'd vote imagination :p |
22:31.12 | phh | Markinus: ok, msm's doc has no doc about it |
22:31.14 | Markinus | phh :) |
22:31.18 | stinebd | if you're outside, reality. inside, imagination |
22:31.23 | phh | next power (voltage) level |
22:31.23 | phh | encoding is TBD |
22:31.37 | phh | thanks qualcomm |
22:31.56 | bzo | Markinus: you're already running the overclock patch? |
22:32.04 | phh | bzo: no |
22:32.08 | phh | bzo: he is trying to reduce the vdd |
22:32.12 | Markinus | bzo: no, I trie to reduce the voltage of the CPU |
22:32.15 | Markinus | try |
22:32.30 | bzo | by changing the acpu_freq_table? |
22:32.38 | phh | Markinus: I think I've seen in some source code of acpuclock.c the match between vdds and actual voltage mmm |
22:32.54 | phh | <PROTECTED> |
22:32.58 | phh | in hero's acpuclock.h |
22:32.58 | phh | in hero's acpuclock.c |
22:33.14 | Markinus | phh: this is the only what I seen |
22:33.23 | phh | but I don't believe it, in another code they say they can't have precise voltage |
22:33.31 | Markinus | I reduced all 3 in 2 and the 7 in 6 |
22:33.40 | Markinus | and 2 in 1 |
22:33.50 | phh | Markinus: try setting everything to 3 |
22:33.52 | Markinus | Ni Idea what <3 is |
22:33.55 | phh | and see if 528MHz is stable |
22:34.19 | phh | <3 is used for <100MHz clock |
22:34.28 | phh | we might try 128MHz on VDD0 |
22:34.40 | phh | could save a lot of idle power |
22:34.46 | phh | (but we'd better fix idle sleep ;) |
22:35.03 | Markinus | phh :) yeah |
22:35.16 | NetRipper | cr2_, i've been trying to completely remove that DBG timer from timer.c but even after removing everything, the kernel simply doesn't boot.. i can't find any references to it anymore.. it's weird |
22:35.17 | ToAsTcfh | bzo: so ur saying once i apply the patch i can set the clock to any value? with an app or how do i set the clock? |
22:35.28 | phh | ToAsTcfh: with kernel parameter |
22:35.30 | bzo | phh: does your latest vsync suspend commit affect the diamond? |
22:35.39 | phh | bzo: no, diamond already had this one |
22:36.07 | phh | I totally forgot I had to do that on diamond |
22:36.07 | cr2_ | NetRipper: we need to dig deeper in the general irq setup |
22:36.12 | phh | until I went into its file. |
22:36.12 | NetRipper | cr2_, as if somewhere there's an array that contains all timers and there's a direct references to the 2nd timer, which isn't there.. no idea :x |
22:36.13 | cr2_ | NetRipper: in wince |
22:36.21 | phh | and so I lost twice the same amount of time for the same thing -_-' |
22:36.27 | cr2_ | NetRipper: hmm. |
22:36.51 | cr2_ | NetRipper: btw, i suggest to add the amss version detection patch to the tree. |
22:36.54 | Markinus | phh: ok, all 3 and I set the ONDEMAND_MIN to 528 |
22:37.01 | NetRipper | cr2_, and if i just leave in that 2nd clock, and make it fail.. it does work.. even though it fails and bails out |
22:37.03 | phh | Markinus: right. |
22:37.11 | phh | Markinus: good way to test |
22:37.16 | cr2_ | NetRipper: maybe after cleaning it up, anddoing some failsafe |
22:37.23 | phh | Markinus: then try max to 128, and vdd=0 |
22:37.28 | bzo | ToAsTcfh: also, don't forget to change your ONDEMAND_MAX in the config |
22:37.47 | NetRipper | cr2_, i'm curious what wince does with that DBG_TIMER interrupt, probably ignores it totally |
22:38.05 | ToAsTcfh | ok so i just apply the patch and set the clock in the ondemand in the connfig? |
22:38.06 | NetRipper | cr2_, do you have a link to what patch you mean exactly? |
22:38.15 | cr2_ | NetRipper: and then add 2 switches into proc_comm. the first on amss version, and the second one on PCOM_enum. |
22:38.40 | bzo | ToAsTcfh: yep, and add acpuclock.oc_freq_khz=710000 (or whatever you want to try) |
22:38.43 | cr2_ | NetRipper: it should be in the gitoriuos tree |
22:39.00 | bzo | phh: why is the ONDEMAND MIN/MAX hardcoded in config anyway? |
22:39.07 | bzo | shouldn't it be derived from the table |
22:39.07 | cr2_ | NetRipper: it's like awk -F. '{print $1$3}' |
22:39.12 | bzo | in acpuclock.c? |
22:39.16 | tmzt | that populates the table? |
22:39.25 | tmzt | and the ondemand governor I think |
22:39.51 | bzo | tmzt: cpufreq just uses the config values, it could easily get them from the acpuclock table |
22:40.00 | cr2_ | NetRipper: but the DBG_TIMER is there afair from haret ? |
22:40.21 | NetRipper | cr2_, yes but it may be enabled in SPL on purpose to annoy us in linux |
22:40.22 | NetRipper | :P |
22:40.39 | NetRipper | (i love conspiracy theories) |
22:41.05 | NetRipper | maybe it's some sort of debug mode, or debug protection for that matter |
22:41.09 | tmzt | umm |
22:41.10 | NetRipper | anti-debug |
22:41.19 | tmzt | is it used for fiq debugger? |
22:41.47 | NetRipper | tmzt, on nexusone they use DBG_TIMER as high-precision timer on 32mhz |
22:42.08 | cr2_ | NetRipper: i can check oemsbl, though it's not cleanly disassembled (because of the evil extra bytes) |
22:42.10 | NetRipper | the normal GP_TIMER is like 24khz or so |
22:42.18 | phh | bzo: don't want 19.2MHz as min value. |
22:42.28 | phh | bzo: system starts at slower clock. |
22:42.33 | cr2_ | NetRipper: maybe it's easier to check it for raph |
22:42.44 | bzo | phh: good point |
22:42.44 | NetRipper | cr2_, yea it was same there |
22:42.46 | cr2_ | where i have a clean smi dump |
22:43.26 | cr2_ | NetRipper: and we need the serial cable |
22:43.42 | NetRipper | cr2_, well at the very least, ramconsole should init a lot sooner than it does now :p |
22:43.51 | cr2_ | tmzt: is there somebody with a nexus serial cable, who can share the experience ? |
22:44.26 | cr2_ | NetRipper: i still don't get why nexus can use zimage and you can't |
22:44.43 | cr2_ | NetRipper: let's assume that i booted zimage by some black magic :) |
22:44.49 | NetRipper | cr2_, i'd love to know |
22:44.59 | NetRipper | cr2_, i asked cmonex about it either but she doesn't know unfortunately |
22:45.06 | phh | cr2_: uncompressor write to forbidde address ? |
22:45.08 | NetRipper | s/either/as well/ |
22:45.09 | cr2_ | Markinus: where is your amss detection patch ? |
22:45.25 | cr2_ | ok |
22:45.38 | NetRipper | she's quite into the asm due to her hardspl work |
22:45.51 | Markinus | cr2_: amss_para.c |
22:46.08 | cr2_ | NetRipper: check amss_para.c |
22:46.15 | cr2_ | Markinus: thanks |
22:46.23 | Spirits-Sight | OK here is a little more maybe someone would be able to answer with my informaiton. I was wondering how the splash screen should look on a kaiser using the native res of 240x320 the screen that shows when you first turn the phone on? my has a about 1inch white on the bottom, I have tryed all three panel types also right after it I see lots of distion of the screen then it comes clear with the linux logo and text flying by |
22:49.21 | cr2_ | Markinus: where is it located ? i'm interested only in the amss strsep() parser |
22:49.34 | MrPippy | so weird, poweroff worked with 95civic's rhod kernel (from end of january), but now poweroff and even reboot don't work |
22:49.39 | phh | cr2_: mach-msm/qdsp5 |
22:49.57 | phh | MrPippy: need to trace "power reason" smem I think |
22:49.58 | cr2_ | phh: hm. funny place. |
22:50.06 | phh | cr2_: I know. |
22:51.03 | phh | bzo: I hate pastebin. |
22:51.03 | Markinus | cr2_: http://pastebin.com/m2869723f |
22:51.03 | bzo | phh: want me to email you the patch? |
22:51.04 | Markinus | cr2_ under qdsp5 |
22:51.04 | phh | bzo: no it's fine |
22:51.33 | NetRipper | hmm interesting |
22:51.34 | NetRipper | :) |
22:51.37 | cr2_ | NetRipper: i suggest to create the toplevel amss_version.c file |
22:51.46 | phh | bzo: oc_freq_khz = (oc_freq_khz / 19200) * 19200; |
22:51.50 | phh | this is totally useless you know ? :p |
22:52.00 | cr2_ | NetRipper: with a global amss_version varialble/symbol |
22:52.02 | phh | oh right not totally |
22:52.06 | bzo | phh: freq has to be rounded to 19.2mhz |
22:52.47 | bzo | phh: neocore @ 633mhz = 20.5fps - not much improvement here |
22:52.47 | phh | wow. that's weird. |
22:52.59 | phh | bzo: please send me the patch at phhusson@free.fr |
22:53.03 | phh | I really hate pastebin. |
22:53.20 | phh | bzo: I think your editor forgot some part in your copy paste :p |
22:53.28 | phh | see line 83 for example, it's not ended |
22:53.34 | NetRipper | cr2_, why would we re-create an amss implementation? isn't it easier to continue working their way? |
22:53.44 | bzo | phh: k, sending email in a moment |
22:54.17 | NetRipper | or am i misunderstanding you? |
22:54.44 | cr2_ | NetRipper: we need to know the amss version in proc_comm |
22:55.02 | cr2_ | NetRipper: so we can use the right api for doing PCOM_ ops |
22:55.21 | cr2_ | NetRipper: and it can be parsed from smem once at boot |
22:55.30 | ToAsTcfh | bah this acually isnt merging easy. or at least as easy as i though |
22:56.37 | bzo | phh: email sent. By the way, the patch also has a couple of fixes from the heroc code |
22:56.45 | phh | bzo: I know |
22:56.51 | phh | well, one fix. |
22:56.54 | phh | ToAsTcfh: yes the patch is wrong |
22:56.55 | NetRipper | cr2_, ok so there'll be like a switch() statement in proc_comm_wince to make things specific for leoo? |
22:56.58 | ToAsTcfh | :/ is this the same patch? |
22:57.13 | cr2_ | NetRipper: there will be no proc_comm_wince at all |
22:57.25 | ToAsTcfh | phh: oh sorry |
22:57.32 | bzo | ToAsTcfh: looks like there was a copy paste error in pastebin |
22:57.41 | cr2_ | NetRipper: there will be 2 switches in "standard" proc_comm |
22:57.58 | ToAsTcfh | yeah it looked really different from my source |
22:58.01 | phh | ToAsTcfh: maybe try this one: |
22:58.04 | phh | http://en.pastebin.ca/1806157 |
22:58.24 | cr2_ | "standard" proc_comm function will be called if amss=6210 || = 6220 || =6225 |
22:58.43 | bzo | phh: hmm, looks like I accidentally undid one on the heroc fixes |
22:58.53 | cr2_ | NetRipper: we need to rename it too, but it's a very small function , so it's not a problem |
22:59.33 | cr2_ | NetRipper: so the first switch is on board->amss_version, and the second one on PCOM_ enum |
23:00.07 | cr2_ | g1 et al. proc_comm then becomes proc_comm_googel() |
23:00.37 | cr2_ | and we can call it for all PCOM if amss=6220 |
23:01.10 | phh | bzo: pushed |
23:01.36 | cr2_ | otherwise we need to call our dex driver (aka old proc_comm_wince+vreg), or our clock implementation of our gpio implementation. |
23:01.37 | NetRipper | keeping proc_comm as is, and just re-implementing it for a wince-kernel prevents a lot of merging troubles |
23:01.56 | cr2_ | yes. the drivers and driver api remains as is |
23:02.05 | bzo | phh: cool :) |
23:02.25 | cr2_ | NetRipper: we could not afford doing it from the very beginning, but now we more than 99% to do it right |
23:03.08 | bzo | phh: not sure if we care, but the other heroc fix was in line 373: if (cur_s->pll != ACPU_PLL_TCXO |
23:03.11 | NetRipper | i'm still not getting the full picture |
23:03.23 | bzo | their comment for the fix was: /* Power collapse should also request pll.(19.2->528) */ |
23:03.39 | cr2_ | NetRipper: btw, our gpio and clk driver can be tested on g1 too, instead of the goggle api |
23:03.40 | phh | bzo: we don't do pll poweroff afaik, so no problem here |
23:03.41 | NetRipper | also probably because i'm not really sure how PCOM/DEX/VREG is done in nexus compared to wince |
23:04.08 | cr2_ | NetRipper: the bigger picture is actually quite simple |
23:04.28 | cr2_ | NetRipper: now there is a proc_comm api with PCOM_ ops |
23:04.35 | bzo | phh: acpuclock turns plls on/off as needed - though I imagine this only applies to pll2 |
23:04.42 | cr2_ | NetRipper: used by goggle |
23:04.53 | NetRipper | ok |
23:04.56 | phh | bzo: really ? |
23:04.57 | phh | huim |
23:05.01 | NetRipper | and what about DEX and VREG now in nexus by google? |
23:05.08 | cr2_ | NetRipper: but it is limited to "google" amss only |
23:05.29 | cr2_ | NetRipper: because all these PCOM_ ops are rerouted to "google" amss |
23:05.34 | phh | bzo: where do you see that ? |
23:05.48 | bzo | phh: one moment, let me see where I saw it |
23:05.53 | phh | bzo: you mean the pc_pll_request part ? |
23:05.56 | phh | that is commented out ? :p |
23:06.12 | cr2_ | NetRipper: we need to reimplement the PCOM_ api calls that "our wince" amss can't perform |
23:06.36 | cr2_ | NetRipper: so it depends on knowing the amss version ( |
23:06.39 | NetRipper | such as? what is done by google amss what isn't done by wince amss? |
23:06.45 | cr2_ | NetRipper: google vs. wince |
23:06.48 | NetRipper | (or other way around) |
23:06.49 | phh | NetRipper: clocks |
23:06.52 | phh | *but* |
23:06.55 | phh | there is rpc for that |
23:06.56 | bzo | phh: look at acpuclk_set_rate, and go to the comment /* Disable PLLs we are not using anymore. */ |
23:06.58 | phh | for both |
23:07.03 | cr2_ | NetRipper: and them a PCOM_ switch for our vs google ops |
23:07.10 | phh | bzo: yup, pc_pll_request |
23:07.14 | NetRipper | phh, but, rpc = amss.. so it's done by amss after all |
23:07.18 | bzo | phh: oh, that is disabled? |
23:07.21 | cr2_ | NetRipper: setting clock speed for example |
23:07.25 | phh | NetRipper: yes, but it has to be coded. |
23:07.28 | phh | bzo: yes |
23:07.48 | phh | cr2_: which sources have you seen that uses clock regim rpc ? |
23:07.52 | cr2_ | NetRipper: or everything DEX. |
23:07.53 | NetRipper | so, google amss doesn't use rpc design at all? |
23:08.01 | cr2_ | NetRipper: or setting gpio alt functions |
23:08.07 | NetRipper | dex = via rpc, right? |
23:08.12 | cr2_ | no |
23:08.13 | bzo | phh: hehe, now ic the #if 0 |
23:08.23 | cr2_ | NetRipper: what is rpc to you ? |
23:08.28 | NetRipper | remote procedure call |
23:08.43 | NetRipper | setting data in smem, calling soft irq |
23:08.59 | cr2_ | NetRipper: there is direct hw write, DEX aka remote pmic control and ONC rpc on wince |
23:09.24 | NetRipper | am i right about rpc? if so, how does dex differ from that |
23:09.26 | cr2_ | NetRipper: and on google it's all hidden behind the proc_comm(PCOM_*) |
23:09.33 | ToAsTcfh | phh that one was a good to go. |
23:09.43 | cr2_ | yes. ONC RPC has nothing to do with DEX |
23:09.45 | Markinus | phh: running now ca 15 min with reduced voltage, no problems . . |
23:09.56 | phh | Markinus: 528 @ 3 ? |
23:10.01 | ToAsTcfh | although i dont have a ondeman_max in my current .config |
23:10.05 | *** join/#htc-linux DarkMasterHalo (~sam@modemcable187.118-80-70.mc.videotron.ca) |
23:10.07 | NetRipper | cr2_, what is the ONC? |
23:10.07 | Markinus | phh: yes |
23:10.16 | NetRipper | the abbreviation |
23:10.22 | phh | Markinus: we should add this setting though debugfs I guess |
23:10.50 | cr2_ | ONC RPC (Open Network Computing Remote Procedure Call |
23:10.56 | Markinus | phh: yes, I do it, I will test the VDD_0 stuff for 128 too .. |
23:10.58 | cr2_ | i think it's sun invention ? |
23:10.59 | NetRipper | ok |
23:11.07 | phh | Markinus: cool |
23:11.17 | phh | Markinus: something else we'd have that android devices won't have :) |
23:11.32 | phh | I love that. |
23:11.35 | Markinus | phh: yeah ! ;) |
23:11.51 | cr2_ | NetRipper: http://hissa.nist.gov/rbac/5277/node14.html |
23:12.27 | *** join/#htc-linux Wr4i7h_ (~abc@89.180.41.114) |
23:12.44 | NetRipper | cr2_, so basically.. we get a huge switch in proc_comm.. and i.e. when PM_VREG_PDOWN_USB_ID cmd is called, then we internally call a DEX (or RPC) driver that performs the command for us? |
23:13.05 | ToAsTcfh | bzo: can u post the ondemand part of ur config? so i can make sure i got it right? |
23:13.32 | bzo | CONFIG_MSM_CPU_FREQ_ONDEMAND_MAX=999999 |
23:13.34 | bzo | CONFIG_MSM_CPU_FREQ_ONDEMAND_MIN=128000 |
23:13.48 | ToAsTcfh | and the other? |
23:13.51 | cr2_ | NetRipper: yes. or we write directly into the register |
23:14.06 | cr2_ | NetRipper: mimicking what wince does in this situation |
23:14.13 | ToAsTcfh | acpuclock.oc_freq_khz=710000? |
23:14.33 | bzo | ToAsTcfh: no other changes in config, that goes in as a kernel param (i.e. default.txt) |
23:14.43 | chamonix | phh: I have to check on the 3G status for our dutch users who used to have probs with vodafone.nl |
23:14.50 | cr2_ | bzo: can you describe the clock relations somewhere in wiki ? |
23:14.52 | phh | chamonix: ok |
23:14.52 | ToAsTcfh | eh no defaut text here |
23:14.58 | phh | chamonix: I have several requests for you :p |
23:15.02 | chamonix | shoot |
23:15.08 | chamonix | I'm bored |
23:15.10 | phh | chamonix: little app to create a new bigger data.img |
23:15.10 | chamonix | :) |
23:15.13 | cr2_ | bzo: afair there was some bug in the (full) acpuclk.c table |
23:15.33 | NetRipper | cr2_, ok.. and for msm7k we already have dex and rpc drivers, so we can re-use them i suppose |
23:15.34 | phh | chamonix: I can give you the shell lines if you can't see how to do |
23:15.34 | bzo | cr2_: sure, what would you like me to add? |
23:15.46 | phh | chamonix: and an app for the soon to be patch from Markinus to set VDDs |
23:15.53 | cr2_ | NetRipper: the rpc is a separate subsystem |
23:16.17 | *** join/#htc-linux stinebd (~bryan@gentoo/developer/battousai) |
23:16.22 | cr2_ | NetRipper: but it has its own problems, because hte VERS and FUNC numbers depends on the amss version too |
23:16.33 | bzo | cr2_: even the abbreviated table used seems a little odd, |
23:16.36 | chamonix | phh: I guess both could be done using my generic app issuing whatever su command |
23:16.43 | phh | chamonix: guess so |
23:16.45 | chamonix | I just need to publish it |
23:16.48 | chamonix | :) |
23:16.48 | phh | :p |
23:16.54 | NetRipper | cr2_, hm ok |
23:17.08 | phh | cr2_: we can bypass VERS I think |
23:17.11 | cr2_ | bzo: for example how axi and ahb0/1 clocks relate to the plls |
23:17.19 | chamonix | ok, I'll do that.... did u have the opportunity to review the wiki page about ehco to dbgfs phh ? |
23:17.23 | ToAsTcfh | arch/arm/mach-msm/acpuclock.c: In function 'acpuclk_set_rate': |
23:17.23 | ToAsTcfh | arch/arm/mach-msm/acpuclock.c:378: error: 'else' without a previous 'if' |
23:17.25 | chamonix | echo* |
23:17.31 | phh | chamonix: iirc it was just right |
23:17.37 | phh | ToAsTcfh: you applied the patch b hand ? |
23:17.38 | phh | by* |
23:17.44 | ToAsTcfh | yeah |
23:17.47 | phh | ToAsTcfh: tss .. |
23:17.47 | chamonix | ok, then I'll include those as default |
23:17.48 | cr2_ | phh: it can be solved independently from proc_comm |
23:18.00 | phh | ToAsTcfh: somewhere in your code you have: |
23:18.01 | phh | } |
23:18.03 | phh | } else { |
23:18.07 | phh | remove the first } . |
23:18.12 | cr2_ | phh: so we can develop a completely new proc_comm implementation now |
23:18.44 | chamonix | phh: an I the only one having had real trouble with wifi on kernel 19.02? |
23:18.45 | cr2_ | phh: doing a clean dex.c driver and sane clock control directly in the PCOM_ switch |
23:18.46 | phh | cr2_: for clocks I'd prefer relying on AMSS... |
23:18.47 | NetRipper | cr2_, downside is that every wince driver will need to use the google api to do its thing? |
23:18.49 | bzo | cr2_: ok, I'll look at the MSM_CLK page and see if there is anything new to add there. Is the wiki public editable? |
23:19.01 | phh | chamonix: not much people uses wifi |
23:19.08 | phh | bzo: yes it is |
23:19.09 | cr2_ | phh: you can't. unless it's done as rpc clock in wince |
23:19.11 | phh | just not front page |
23:19.27 | chamonix | awwww, I forgot..... do not read build comments :-D |
23:19.45 | cr2_ | phh: and even then it is possible to do the right thing |
23:19.49 | ToAsTcfh | phh: what to do if i dont have a default.txt to add acpuclock.oc_freq_khz=710000? |
23:20.05 | phh | cr2_: it's not done as rpc clock in wince, but if amss can have get clk khz, I'd guess it can set freq too |
23:20.05 | cr2_ | phh: just cam/vfe is not such a big priority for me :) |
23:20.13 | ToAsTcfh | just add it to acpuclock.c? |
23:20.23 | phh | ToAsTcfh: that's a way |
23:20.29 | phh | don't know how you can pass kernel parameters |
23:20.30 | ToAsTcfh | or? |
23:20.36 | cr2_ | phh: if wince does not use it, it does not exist |
23:20.38 | phh | bzo: you should enable on-the-fly parameter editing |
23:20.42 | NetRipper | thanks for explanation cr2_ |
23:20.45 | NetRipper | gtg off to bed now :) |
23:20.53 | phh | cr2_: wince uses msmcklkhz ? |
23:20.56 | cr2_ | NetRipper: good night |
23:21.00 | phh | I mean it declares it, but that's it. |
23:21.04 | cr2_ | phh: for setting clocks |
23:21.08 | phh | really ? |
23:21.09 | phh | ah. |
23:21.21 | phh | cr2_: and in wince, the rpc API exists |
23:21.29 | phh | (see what captnoord dumped) |
23:21.39 | cr2_ | phh: wince either writes intop MSM_CLK directly, or uses rpc (only for cam,btw) |
23:21.41 | bzo | phh: yes I wanted to. But I have to figure out how to reconfigure cpufreq tables so it doesn't get confused |
23:22.05 | cr2_ | phh: we have no idea how to use it anyway |
23:22.14 | cr2_ | phh: and htc could know better |
23:22.28 | phh | cr2_: have you said one board (acer ?) uses the rpc api ? |
23:22.39 | cr2_ | different amm |
23:22.45 | cr2_ | s/amm/amss/ |
23:22.46 | phh | it gives the idea. |
23:22.51 | phh | and we have the rpc ids |
23:23.05 | chamonix | phh: overclocking mean underclocking as well right? |
23:23.09 | cr2_ | it's pure guesswork |
23:23.14 | phh | chamonix: no |
23:23.17 | phh | cr2_: yup. |
23:23.17 | chamonix | :-/ |
23:23.30 | phh | chamonix: you can already underclock to 19MHz /o\ |
23:23.33 | phh | isn't that enough ? |
23:23.34 | cr2_ | the register accesses (at least for MD/NS and i2c) are perfectly known and work |
23:23.49 | phh | cr2_: yes but it's heavy work to set all clocks correctly ...; |
23:23.55 | phh | especially the "ns only" ones |
23:24.01 | cr2_ | not really |
23:24.23 | chamonix | dunno, just thought about testing effect on battery life for non heavy use |
23:24.28 | cr2_ | just the current clock-wince.c driver needs to be replaced by a sane implementation |
23:24.42 | cr2_ | phh: the vogue clock driver is even more strange :) |
23:24.56 | phh | cr2_: it's not a reason ... |
23:25.04 | tmzt | why are timers late? |
23:25.10 | tmzt | do we know why yet? |
23:25.16 | phh | tmzt: no :'( |
23:27.07 | cr2_ | phh: we will do it for leo, in the most recent tree |
23:27.37 | phh | cr2_: pfff you said you'd do one for msm7k at mid january |
23:27.42 | Markinus | phh: no problems with VDD_0 on 128 :) |
23:27.45 | cr2_ | phh: because there is no reason to pull all the legacy crap lke clock-wince and proc_comm_wince there. |
23:27.45 | bzo | chamonix: the new oc patch will let you set any freq you want, including lower than 528 |
23:28.04 | phh | bzo: but multiple of 19.2MHz. |
23:28.05 | cr2_ | phh: yes, it's the same driver |
23:28.10 | phh | bzo: and we already have a 19.2 clock :p |
23:28.28 | cr2_ | phh: the 7200/7201A/8250 differences are all doucmented in MSM_CLK |
23:28.29 | *** join/#htc-linux Zoolooc__ (~fredsibar@p54955DBB.dip.t-dialin.net) |
23:28.36 | bzo | phh: right, but I was thinking in the case of if you wanted to test a top 400mhz speed or something |
23:28.50 | phh | bzo: right. |
23:29.01 | phh | cr2_: true |
23:29.13 | cr2_ | phh: 8250 is actually more close to 7200 for many common regs |
23:29.44 | cr2_ | phh: and the latest nexus tree works also on 7201A |
23:30.02 | chamonix | I'll give it a try.... I use my phone always in the same way and I'm curious if it reflects on battery in some significant way |
23:30.29 | bzo | I don't know underclocking will save much power though. It is probably better to finish stuff sooner, and drop down to 128mhz as soon as possible |
23:30.32 | phh | chamonix: I'm curious too |
23:30.48 | phh | bzo: even better. idle sleep. |
23:31.17 | chamonix | bzo: that's the assumption yes but I want to confirm that |
23:31.27 | cr2_ | phh: lol. aka "boot CE, arm9 in idle sleep" |
23:31.28 | bzo | I'm so happy that suspend is working again. I can actually use my phone like a regular phone :) |
23:31.41 | ToAsTcfh | phh: ive been through it ... yeah its not making http://en.pastebin.ca/1806187 |
23:31.42 | phh | cr2_: uh ? |
23:31.55 | cr2_ | phh: the oemsbl option yesterday |
23:31.56 | ToAsTcfh | i even applied the patch |
23:31.56 | phh | ToAsTcfh: can't you just use patch ? |
23:32.05 | ToAsTcfh | i tried it |
23:32.07 | phh | cr2_: it was idle sleep ? |
23:32.09 | *** join/#htc-linux stinebd (~bryan@gentoo/developer/battousai) |
23:32.12 | phh | I thaught it was infinite loop mmm |
23:32.12 | ToAsTcfh | it does the same thing |
23:32.16 | chamonix | I mean what does my phone do when it wakes up? most alarms are for xmpp and other stuff that may go well with low freq |
23:32.27 | phh | ToAsTcfh: what's the error ? |
23:32.41 | phh | chamonix: mmm |
23:32.54 | cr2_ | phh: 3 : Boot CE, ARM9 enter endless loop |
23:32.59 | chamonix | I know, we had that discussion already about freq with screen off :P |
23:33.03 | ToAsTcfh | arch/arm/mach-msm/acpuclock.c: In function 'acpuclk_set_rate': |
23:33.03 | ToAsTcfh | arch/arm/mach-msm/acpuclock.c:378: error: expected '}' before 'else' |
23:33.11 | ToAsTcfh | fixed that then got |
23:33.17 | phh | ToAsTcfh: .... |
23:33.27 | phh | remove from line 378 to 385 |
23:33.37 | phh | you applied fixes from hero's tree :p |
23:33.41 | phh | (which you already has.) |
23:33.51 | phh | chamonix: it's not exactly the same discussion thouogh |
23:34.07 | phh | chamonix: btw, you may want to try low clock @ 19.2MHz |
23:34.14 | phh | but set it once kernel is booted, not before. |
23:34.43 | phh | chamonix: echo 19200 > /sys/devices/platform/cpu/cpufreq0/scaling_min_speed or something like that |
23:35.54 | chamonix | that looks like something for an applet to control freq = f(time of days) |
23:36.04 | cr2_ | good night |
23:36.19 | ToAsTcfh | phh: arch/arm/mach-msm/acpuclock.c:603: error: invalid storage class for function 'lpj_init' |
23:36.20 | ToAsTcfh | arch/arm/mach-msm/acpuclock.c:632: error: expected declaration or statement at end of input |
23:36.36 | phh | ToAsTcfh: how do you do that ... |
23:36.47 | phh | seriously, applying patches is easy ... even with conflicts |
23:36.57 | ToAsTcfh | <phh> remove from line 378 to 385 |
23:37.02 | stinebd | haha |
23:37.32 | *** join/#htc-linux Zoolooc_ (~fredsibar@p54954F34.dip.t-dialin.net) |
23:37.35 | ToAsTcfh | bah the patch applied clean then wont make |
23:37.35 | chamonix | oh btw I saw u guys added dropbear to rootfs.... that's cool! |
23:37.48 | stinebd | makkonen did it! |
23:38.08 | chamonix | did u build it? |
23:38.08 | stinebd | all bug reports go to him |
23:38.12 | chamonix | lol |
23:38.17 | chamonix | kk, I'll test |
23:39.07 | phh | stinebd: tsss |
23:39.17 | phh | you hate bugreports ? :p |
23:39.20 | stinebd | whistles nonchalantly |
23:39.27 | *** join/#htc-linux myn\ (cereal@cpe-71-64-8-167.insight.res.rr.com) |
23:39.28 | stinebd | i had to file a bug report to myself earlier |
23:39.45 | stinebd | for the rootfs build service |
23:40.10 | chamonix | I had a feature request but I forgot :-X |
23:40.28 | phh | chamonix: ah |
23:40.44 | chamonix | yea, got it.... it was some housekeeping / cleanup |
23:40.48 | stinebd | echo 19200 >/sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq |
23:40.49 | stinebd | btw |
23:40.56 | chamonix | ty |
23:41.18 | chamonix | I'll add to the wiki and to my defaults |
23:41.21 | phh | stinebd: I wasn't that far. |
23:41.40 | chamonix | u guys have something against slideME? |
23:41.46 | phh | no |
23:41.51 | phh | why ? |
23:41.56 | stinebd | what's that? |
23:41.59 | phh | you're putting your apps on it ? |
23:42.01 | phh | stinebd: market |
23:42.03 | chamonix | because I was thinking of publishing there |
23:42.06 | stinebd | ah |
23:42.20 | chamonix | it's better than market because fee free for devs |
23:42.41 | stinebd | not quite as work safe from the looks of it ;) |
23:42.42 | phh | chamonix: can you add a machine filter ? |
23:42.50 | phh | ie free app for wince devices, paying for android devices :p |
23:42.51 | chamonix | huh? |
23:43.14 | chamonix | lol |
23:43.32 | chamonix | who would like to run ledeffect on a N1? :-D |
23:43.57 | phh | chamonix: just implement "color leds" stuff |
23:43.58 | *** join/#htc-linux Zoolooc__ (~fredsibar@p549543D8.dip.t-dialin.net) |
23:44.02 | phh | you'll get many users |
23:44.10 | phh | chamonix: and code the "low voltage" thing |
23:44.14 | phh | you'll get even more. |
23:44.24 | phh | implement jogdial color stuff |
23:44.26 | chamonix | what low voltage thing? |
23:44.28 | phh | and you'll be the king of the wolrd. |
23:44.39 | chamonix | ok, who sends me a N1? |
23:44.42 | phh | chamonix: the thing that Markinus is working on, set the CPU To a lower voltage |
23:44.43 | phh | lol |
23:45.19 | chamonix | have u added dbgfs for color leds? |
23:45.41 | phh | still the old thing |
23:45.49 | phh | the parameters are still not fully known |
23:45.54 | phh | (like blinking) |
23:46.09 | phh | I think microp is able to do advanced blink stuff, that wince doesn't use |
23:46.31 | phh | (no I don't want to do better than wince.) |
23:46.33 | chamonix | I would be more than happy to add more support/settings |
23:46.34 | phh | (ok yes I do.) |
23:47.05 | Markinus | phh: I set some parameters in the wiki for the Topaz led, blinking is there too (only red is missing .. ) |
23:47.10 | ToAsTcfh | phh: got it :P |
23:47.17 | phh | Markinus: ah you know how to blink ? |
23:47.17 | phh | cool |
23:47.22 | phh | Markinus: and there is no red I think |
23:47.27 | chamonix | Markinus: what wiki? |
23:47.28 | ToAsTcfh | user error :P |
23:47.33 | ToAsTcfh | again |
23:47.35 | phh | ToAsTcfh: NO WAY |
23:47.38 | phh | chamonix: htc-linux.org |
23:47.39 | ToAsTcfh | lol |
23:47.44 | chamonix | kk |
23:47.46 | phh | chamonix: but don't use that for your userland app |
23:47.52 | phh | we will provide you with a proper app |
23:48.01 | ToAsTcfh | one to many } |
23:48.05 | chamonix | ok |
23:48.07 | ToAsTcfh | again |
23:48.09 | ToAsTcfh | ? |
23:48.19 | phh | Markinus: cc, 51, 0, 3, ff, ff -> LED Green flashing slow ( ca. 6 sek ). |
23:48.20 | phh | really ? |
23:48.21 | phh | damn |
23:48.24 | chamonix | u add dbgfs, I take care of prefs and settings |
23:48.32 | Markinus | http://www.htc-linux.org/wiki/index.php?title=TopazColorLED |
23:48.41 | phh | Markinus: that's not how it works on rhod |
23:48.45 | chamonix | I have to get in touch with ppl from ebuddy too to add broadcast support |
23:48.50 | Markinus | phh: :( |
23:48.50 | phh | Markinus: I'd thaught that the second half of the message are timers |
23:49.17 | phh | I have to try. |
23:49.21 | phh | anyway, now is sleeping |
23:49.25 | phh | MrPippy: any news on wifi ? |
23:49.26 | Markinus | phh: on Topaz not . . . |
23:50.12 | ToAsTcfh | still gotta figure a way around a default.txt |
23:50.14 | phh | ok MrPippy is sleeping, so I do too |
23:50.15 | Markinus | (argh . . .every device different . . . did they didn't know platform strategies . . .. . . .) |
23:50.18 | phh | ToAsTcfh: just hardcode your setting |
23:50.27 | chamonix | does any1 know a good xmpp client with source? |
23:50.47 | phh | Markinus: they first did topaz, then they wanted to do nexus one, so they started TP2 |
23:50.57 | phh | they realized nexus one as TP2 without QSD would be bad |
23:51.05 | phh | so they released TP2 as a wince device |
23:51.09 | phh | and did a bad N1. |
23:51.15 | ToAsTcfh | phh: static int oc_freq_khz = 0; ? or no im gonna have to write in a big deal? |
23:51.15 | phh | (no I'm not fantasing) |
23:51.17 | phh | (ok just a bit) |
23:51.30 | phh | ToAsTcfh: yes set it to 610000 |
23:51.46 | ToAsTcfh | ok great |
23:51.58 | Markinus | phh: yes . . . . . . . argh |
23:52.03 | stinebd | is it stable at 610? |
23:52.15 | ToAsTcfh | i wonder if i can use an app like setcpu to control it in userland |
23:52.23 | bzo | I've been running at 633mhz for a little while now, works great |
23:52.31 | bzo | unstable at 710mhz though |
23:52.55 | bzo | ToAsTcfh: not as it is currently implemented, only at startup |
23:53.09 | stinebd | command line arg works? |
23:53.29 | ToAsTcfh | bzo good work man |
23:53.38 | bzo | yes, if you have a kernel that includes the patch |
23:53.45 | bzo | ToAsTcfh: thx |
23:53.52 | ToAsTcfh | although im still compiling with my socket 478 |
23:53.58 | ToAsTcfh | lol |
23:54.01 | stinebd | oh i do |
23:54.40 | ToAsTcfh | bzo u got some where i can lead people too if they wanna donate or thank u? |
23:55.25 | bzo | ToAsTcfh: no donations necessary |
23:55.29 | chamonix | gnite guys |
23:55.53 | ToAsTcfh | well ill just link u on xda |
23:56.25 | ToAsTcfh | although if this works itll be awesome |
23:56.40 | bzo | cool. I should let no2chem know about my patch, maybe it will help him figure out the winmo oc tool |
23:59.26 | ToAsTcfh | yeah we were trying his but i wouldnt work for us |
23:59.55 | ToAsTcfh | his was way more do able then the dream devs |