| 00:12.38 | *** join/#htc-linux BHSPitMonkey (n=stephen@unaffiliated/bhspitmonkey) |
| 00:14.21 | *** join/#htc-linux AstainHellbring (n=AstainHe@unaffiliated/astainhellbring) |
| 00:30.42 | *** join/#htc-linux StarLite (n=nnscript@s55916ca6.adsl.wanadoo.nl) |
| 01:01.05 | *** join/#htc-linux x29a_ (n=x29a@unaffiliated/x29a) |
| 01:12.28 | *** join/#htc-linux AstainZZZZZZ (n=AstainHe@unaffiliated/astainhellbring) |
| 01:19.41 | *** join/#htc-linux mdrobnak_ (n=mdrobnak@ool-457e706e.dyn.optonline.net) |
| 01:20.01 | mdrobnak_ | tmzt: You here? |
| 01:27.11 | mdrobnak_ | ok, gotta run. I am going to try the latest dzo rootfs.img...Word of warning to those using mine - dont hit a key to backup. it hosed my data.img :-( |
| 01:27.18 | mdrobnak_ | Gotta remove that code. |
| 01:27.24 | mdrobnak_ | Off to the movies :-) |
| 01:58.56 | tmzt | mdrobnak: yeah |
| 02:03.15 | *** join/#htc-linux BHSPitLappy (n=BHSPitLa@unaffiliated/bhspitmonkey) |
| 02:10.52 | *** join/#htc-linux ppman (n=mkern@pool-71-250-13-156.nwrknj.east.verizon.net) |
| 02:10.59 | ppman | just got my raphael today |
| 02:11.10 | ppman | that is one hell of a phone! |
| 02:19.48 | *** join/#htc-linux Abracadabra (n=Abracada@unaffiliated/abracadabra) |
| 02:26.37 | *** join/#htc-linux Perkka (n=perkka@ua-83-227-207-92.cust.bredbandsbolaget.se) |
| 02:38.34 | *** join/#htc-linux dzo_ (n=dzo@mail.marginz.co.nz) |
| 02:43.24 | *** join/#htc-linux ppman_ (n=mkern@pool-70-111-244-91.nwrk.east.verizon.net) |
| 02:45.20 | *** join/#htc-linux stickboy (n=anonymou@128.153.211.150) |
| 02:49.24 | *** join/#htc-linux g55 (n=g55@93.135.219.215) |
| 02:53.42 | *** join/#htc-linux alexandernst (n=alexande@212.183.198.145) |
| 03:01.01 | tmzt | ppman_: hey, which one? |
| 03:01.42 | *** join/#htc-linux dzo (n=dzo@mail.marginz.co.nz) |
| 03:12.00 | tmzt | AstainHellbring: ping |
| 03:12.11 | tmzt | <PROTECTED> |
| 03:12.11 | tmzt | <PROTECTED> |
| 03:12.11 | tmzt | - max_irq = (machine_is_htcraphael_cdma() || machine_is_htcdiamond_cdma()) |
| 03:12.15 | tmzt | + max_irq = (machine_is_htcraphael_cdma() || machine_is_htcraphael_cdma500 |
| 03:12.18 | tmzt | please add that to smd.c |
| 03:13.50 | ppman_ | tmzt: raph110... htc FUZE |
| 03:13.55 | tmzt | cool |
| 03:14.00 | tmzt | what did you pay? |
| 03:14.04 | ppman_ | $200 |
| 03:14.09 | tmzt | ok |
| 03:14.17 | tmzt | not from at&t? |
| 03:14.21 | ppman_ | used |
| 03:14.26 | tmzt | right |
| 03:14.31 | tmzt | still good for vga though |
| 03:14.37 | ppman_ | the last guy scratched up the screen a little, |
| 03:14.43 | AstainHellbring | ahh tmzt I dont have a build env anymore |
| 03:14.47 | ppman_ | but at least he left it with a ton of cool software |
| 03:15.16 | ppman_ | oh yeah, and gps won't even work in winmo |
| 03:15.25 | ppman_ | but apparently that's common |
| 03:15.27 | tmzt | no, it won't |
| 03:15.36 | tmzt | I think I found the instructions to fix it though, on the pre site |
| 03:15.43 | tmzt | it should be the same |
| 03:15.51 | tmzt | I'm working on some that don't require QPST |
| 03:16.01 | tmzt | not quite the same |
| 03:16.14 | tmzt | and you would have to have diag access which might be missing on fuze |
| 03:16.58 | ppman_ | QPST? |
| 03:17.55 | tmzt | qualcomm tool |
| 03:20.07 | ppman | I really want to get started with linux on this thing.. |
| 03:20.32 | ppman | I'm thinking of using the sd card I had on my kaiser for now and buying a replacement for kaiser. |
| 03:20.44 | tmzt | ppman: are you in #ofono? |
| 03:20.51 | tmzt | sure |
| 03:21.04 | ppman | no... anything good happening there? |
| 03:21.07 | tmzt | android or something else? I though you mentioned you wanted to work on a driver for raph100 |
| 03:21.19 | tmzt | join, denkinz is there now |
| 03:21.24 | ppman | I want to get mer up to coolness |
| 03:21.44 | tmzt | well, we need a gsm stack and it looks like that's the one maemo will be using at some point |
| 03:21.55 | tmzt | there's still the option of using fso an fsogsmd |
| 03:22.37 | ppman | well, we want to follow the mer guys... make it easy for ourselves. |
| 03:26.31 | *** join/#htc-linux dilinger (n=dilinger@wireless.queued.net) |
| 03:33.50 | ppman | um... am I supposed to get 17.3MB/s from a class 2 microsd card? |
| 03:38.09 | *** join/#htc-linux dzo (n=dzo@mail.marginz.co.nz) |
| 04:00.10 | *** join/#htc-linux dzo (n=dzo@mail.marginz.co.nz) |
| 04:03.27 | *** join/#htc-linux NexVision (n=a@c-76-109-33-88.hsd1.fl.comcast.net) |
| 04:09.51 | *** join/#htc-linux stickboy (n=anonymou@128.153.211.150) |
| 04:28.51 | *** join/#htc-linux stickboy (n=anonymou@128.153.211.150) |
| 04:37.55 | *** join/#htc-linux dzo (n=dzo@121.98.128.127) |
| 05:00.11 | *** join/#htc-linux dzo (n=dzo@mail.marginz.co.nz) |
| 05:01.09 | *** join/#htc-linux droid001 (n=g1@p4FDCF3ED.dip.t-dialin.net) |
| 05:22.08 | *** join/#htc-linux dzo (n=dzo@mail.marginz.co.nz) |
| 05:25.37 | *** join/#htc-linux evildarknight (n=charles@41.211.74.250) |
| 05:43.38 | *** join/#htc-linux dzo (n=dzo@mail.marginz.co.nz) |
| 06:23.55 | *** join/#htc-linux dzo (n=dzo@mail.marginz.co.nz) |
| 06:37.05 | *** join/#htc-linux CZero (n=CZero@87-194-204-58.bethere.co.uk) |
| 06:49.39 | *** join/#htc-linux dzo (n=dzo@mail.marginz.co.nz) |
| 07:06.02 | tmzt | welcome, goodnight |
| 07:06.37 | CZero | :) |
| 07:19.17 | *** join/#htc-linux x29a (n=x29a@unaffiliated/x29a) |
| 07:35.05 | *** join/#htc-linux alexandernst (n=alexande@212.183.198.145.dyn.user.ono.com) |
| 07:36.39 | *** join/#htc-linux luc (n=luc@89-115-128-35.cl.ipv4ilink.net) |
| 07:45.22 | *** join/#htc-linux DasFx (n=John@535482D5.cable.casema.nl) |
| 08:23.38 | kri5 | Hi there |
| 08:23.45 | tmzt | hello |
| 08:53.33 | *** join/#htc-linux Gnutoo (n=gnutoo@host93-155-dynamic.51-79-r.retail.telecomitalia.it) |
| 09:18.30 | *** join/#htc-linux dcordes (n=dcordes_@unaffiliated/dcordes) |
| 09:27.40 | *** join/#htc-linux Zinbolic (n=zinbolic@84.238.80.215) |
| 09:33.13 | tmzt | mdrobnak: do you have raph110 keymap somewhere? |
| 09:35.24 | swc|666 | tmzt, thx btw from last night |
| 09:35.32 | tmzt | ? |
| 09:35.43 | swc|666 | meant to say thx to u last night |
| 09:35.49 | swc|666 | then ctrl+d |
| 09:36.08 | swc|666 | i asked a question about something and saw u answered 1 min b4 |
| 09:36.11 | swc|666 | lol |
| 09:36.38 | swc|666 | tmzt, |
| 09:36.38 | tmzt | ah |
| 09:36.38 | tmzt | just forget what it was |
| 09:36.49 | tmzt | mdrobnak: please forward to Jason8 or |Jason8| |
| 09:36.49 | swc|666 | does the new wlxx driver work for ti acx100 ? |
| 09:36.56 | tmzt | it should |
| 09:37.03 | tmzt | but we don't know how yet |
| 09:37.08 | tmzt | and it might need some work |
| 09:37.08 | swc|666 | on what stack is that btw... mac80211? |
| 09:37.11 | tmzt | yes |
| 09:37.14 | swc|666 | nice |
| 09:37.22 | swc|666 | tmzt, hmm |
| 09:37.26 | swc|666 | how about for the uni |
| 09:37.34 | tmzt | the pcmcia one? |
| 09:37.41 | tmzt | you would have to write a new if_ at least |
| 09:37.45 | swc|666 | uni has tiacx, 16 bit slave mem tho |
| 09:37.50 | swc|666 | not sdiio |
| 09:37.53 | swc|666 | -i |
| 09:38.05 | *** join/#htc-linux FR^2 (i=frr@frquadrat.de) |
| 09:38.08 | swc|666 | yea |
| 09:38.12 | swc|666 | hmm |
| 09:39.39 | swc|666 | i have some patches from a while ago, from sameo @ ohand iirc |
| 09:40.23 | swc|666 | its been a while since i've done real c, but i might try and adapt the wl driver for the uni and the ba |
| 09:41.05 | swc|666 | tmzt, does par still have all of those ba's btw? |
| 09:41.32 | tmzt | I think |
| 10:08.03 | *** join/#htc-linux kvaster (n=kvaster@live.bn.by) |
| 10:16.36 | *** join/#htc-linux thedicemaster (n=thedicem@j89051.upc-j.chello.nl) |
| 10:17.10 | *** join/#htc-linux Captnoord (n=Captnoor@dc5147a47b.adsl.wanadoo.nl) |
| 10:17.42 | *** join/#htc-linux BabelO (n=fcr@unaffiliated/babelo) |
| 10:19.28 | *** join/#htc-linux Squarc (n=Squarc@82-217-32-29.cable.quicknet.nl) |
| 10:27.35 | *** join/#htc-linux pleemans (n=toi@d54C2A96D.access.telenet.be) |
| 10:30.38 | DJWillis | swc|666: WL12xx will need some work for acx100 but it is VERY possible, that said, acx100 got some love recently I notice. For the mac80211 vairent anyway. There was a post to the Angstrom mailing list about fixing up the acx100 driver for the hp hx4700 (so not far off the BA). |
| 10:32.32 | DJWillis | swc|666: wl12xx has layers for SPI and SDIO so that should not give you any issues (the n900 uses SPI for it's WL1251 so nokia wrote the driver to support that ;-)) |
| 10:34.32 | *** join/#htc-linux NexVision (n=a@c-76-109-33-88.hsd1.fl.comcast.net) |
| 10:39.23 | DJWillis | swc|666: http://lists.linuxtogo.org/pipermail/angstrom-distro-devel/2009-October/003494.html for info on the acx100 love. |
| 10:40.26 | *** join/#htc-linux sd00 (n=smac@82-39-194-231.cable.ubr02.jarr.blueyonder.co.uk) |
| 10:42.52 | *** join/#htc-linux luc_ (n=luc@89.115.128.35) |
| 10:48.01 | *** join/#htc-linux GlemSom (n=glemsom@0x5da34bca.cpe.ge-1-1-0-1105.sdnqu1.customer.tele.dk) |
| 11:07.54 | *** join/#htc-linux leaigor (n=laigor@188.134.16.241) |
| 11:26.55 | *** join/#htc-linux DarkMasterHalo (n=DarkMast@modemcable070.63-56-74.mc.videotron.ca) |
| 11:39.28 | *** join/#htc-linux luc_ (n=luc@89-115-128-35.cl.ipv4ilink.net) |
| 11:42.40 | *** join/#htc-linux droid001 (n=g1@p4FDCC737.dip.t-dialin.net) |
| 11:44.39 | *** join/#htc-linux MethoS- (n=clemens@134.102.106.250) |
| 12:43.03 | *** join/#htc-linux dream_kill (n=nospam@92.56.53.50) |
| 12:48.51 | *** join/#htc-linux marex (n=marex@vasut.kolej.mff.cuni.cz) |
| 12:58.02 | *** part/#htc-linux sd00 (n=smac@82-39-194-231.cable.ubr02.jarr.blueyonder.co.uk) |
| 13:03.23 | *** join/#htc-linux Xyis (i=5772181c@gateway/web/freenode/x-tiqcjgqwtiujjdrm) |
| 13:03.48 | Xyis | wow, lots more people here than i expected. How its everyone? |
| 13:04.39 | *** join/#htc-linux Xyis (i=5772181c@gateway/web/freenode/x-xxpknmhjodtigxyz) |
| 13:05.40 | Xyis | Out of interest, anyone on Google Wave here? |
| 13:06.47 | phh | not really related to the chan ? |
| 13:07.25 | Xyis | Well i ask as i was wondering if someone wanted to help me make a htc-linux wave. |
| 13:10.58 | Xyis | Oh well, Just a quick question, i am going to attempt to put android on my Raph, will i need a linux machine to compile anything? |
| 13:12.48 | phh | only if you want to debug |
| 13:16.18 | Xyis | okay, thanks, i'll look into it. I read the wiki briefly it looks complicated, if i go over it enough is everyhing i need to know to get running on there? |
| 13:16.41 | phh | there is nothing complicated |
| 13:17.07 | phh | just grab an android build, check the xda's forum to find one, extract it on a sdcard |
| 13:17.10 | phh | and that's it |
| 13:17.58 | Xyis | Dont i have to get haret installed? |
| 13:19.44 | phh | most of the time it's in the build |
| 13:20.24 | Xyis | Brilliant, sounds easier than flashing a rom :) Thanks for the infor, i know n00bs like me can be annoying. So much appreciated. |
| 13:22.13 | phh | it's easier to debug :p |
| 13:39.36 | alexandernst | What are the bad things about the HTC Touch Pro 2 Vario V (if there are bad things)? I red that this is a T-Mobile phone version. |
| 13:42.57 | *** join/#htc-linux Squarc (n=Squarc@82-217-32-29.cable.quicknet.nl) |
| 14:04.46 | *** join/#htc-linux tsdogs (n=tsdogs@net203-187-146.mclink.it) |
| 14:15.25 | *** join/#htc-linux [1]Captnoord (n=Captnoor@81.71.164.123) |
| 14:44.50 | *** join/#htc-linux tuxhero (n=tuxhero@122.169.157.97) |
| 14:59.25 | *** join/#htc-linux jack_baumer (n=jack_bau@dslb-088-065-048-172.pools.arcor-ip.net) |
| 15:00.10 | *** part/#htc-linux jack_baumer (n=jack_bau@dslb-088-065-048-172.pools.arcor-ip.net) |
| 15:11.07 | *** join/#htc-linux jack_baumer (n=jack_bau@dslb-088-065-048-172.pools.arcor-ip.net) |
| 15:13.37 | *** join/#htc-linux tuxhero_ (n=tuxhero@122.169.169.172) |
| 15:30.47 | *** join/#htc-linux dcordes (n=dcordes_@unaffiliated/dcordes) |
| 15:32.12 | *** join/#htc-linux balsat (n=kll@87.72.13.87) |
| 15:41.51 | *** join/#htc-linux Hellgringo (n=holger@p3EE01AD9.dip0.t-ipconnect.de) |
| 15:43.03 | Hellgringo | hello, i tried a android build on my HTC Diamond |
| 15:43.54 | Hellgringo | the device booted up the linux kernel .... and so on |
| 15:44.29 | Hellgringo | but now i can't execute the haret.exe |
| 15:44.57 | Hellgringo | windows says that it is broken or the certificate isn't correct |
| 15:45.08 | Hellgringo | reinstalling harte.exe doesn't change anything |
| 15:45.25 | Hellgringo | has somone an idea, what the problem might be? |
| 15:50.45 | *** join/#htc-linux Bry8Star_ (n=Bry8Star@adsl-99-32-0-250.dsl.lsan03.sbcglobal.net) |
| 15:58.28 | Hellgringo | no idea? |
| 16:03.34 | *** join/#htc-linux Marajin (n=marajin@87-194-102-189.bethere.co.uk) |
| 16:12.18 | *** join/#htc-linux DarkMasterHalo (n=DarkMast@modemcable070.63-56-74.mc.videotron.ca) |
| 16:24.08 | *** join/#htc-linux Gnutoo (n=gnutoo@host93-155-dynamic.51-79-r.retail.telecomitalia.it) |
| 16:46.57 | *** join/#htc-linux BHSPitLappy (n=BHSPitLa@unaffiliated/bhspitmonkey) |
| 16:48.52 | *** join/#htc-linux balans (n=balans@212-123-149-239.ip.telfort.nl) |
| 17:24.31 | *** join/#htc-linux pleemans (n=toi@d54C2A96D.access.telenet.be) |
| 17:34.29 | *** join/#htc-linux pH5 (n=user@g229228210.adsl.alicedsl.de) |
| 17:36.19 | *** join/#htc-linux Epsylon3 (i=kvirc@85-170-22-251.rev.numericable.fr) |
| 17:36.30 | Epsylon3 | hi |
| 17:37.23 | Epsylon3 | i have a battery problem with my kaiser, even fully charged (and tested with new one) Windows tell me Battery is low, then shutdown... |
| 17:37.54 | Jason8 | ...this is #htc-linux |
| 17:38.03 | Epsylon3 | i had this problem since a few months, and worked again last week..... i installed android on |
| 17:38.10 | Epsylon3 | i come... |
| 17:38.25 | Epsylon3 | so i was able to boot android with haret |
| 17:38.35 | Jason8 | yep. |
| 17:38.50 | Epsylon3 | the problem... i dont have time to launch it |
| 17:39.02 | Epsylon3 | and dont have time to change registry too... |
| 17:39.10 | Epsylon3 | to place it at startup |
| 17:39.10 | Jason8 | Try a hard-reset? |
| 17:39.34 | Epsylon3 | it doesnt work... tried the 2 "menus" buttons method |
| 17:39.54 | Epsylon3 | and i fear to flash with the battery problem |
| 17:40.00 | Epsylon3 | a big image |
| 17:40.20 | dcordes | Epsylon3, #xda-devs might be better help for hard resetting kaiser questions |
| 17:40.25 | Jason8 | hold those two softkeys, hit the soft-reset button with the stylus, and keep holding them until it asks you to push a button to continue |
| 17:40.26 | Epsylon3 | so... is there a way to load kernel via sdcard at bootloader ? |
| 17:40.48 | Epsylon3 | the zImage.... |
| 17:40.58 | Epsylon3 | i've seen something with IMAGE.BIN |
| 17:41.12 | Epsylon3 | and INITRD.BIN |
| 17:41.22 | dcordes | Epsylon3, not yet |
| 17:42.26 | Epsylon3 | ya ok, so the only way i have is to flash again the WM rom |
| 17:42.40 | Epsylon3 | i tried the hard reset... |
| 17:42.51 | Epsylon3 | maybe my spl is not compatible |
| 17:43.00 | Epsylon3 | i've the 3.14 |
| 17:43.37 | dcordes | please discuss rom flashing in #xda-devs . |
| 17:43.51 | Epsylon3 | yup... ok |
| 17:43.53 | dcordes | or even better use xda wiki |
| 17:43.58 | Epsylon3 | i know how to do that |
| 17:44.25 | Epsylon3 | kaisimg.nbh on sdcard and lets go |
| 17:45.08 | Epsylon3 | is there a linux nbh ? |
| 17:45.19 | Epsylon3 | os.nb |
| 17:45.31 | dcordes | Epsylon3, again: not yet. |
| 17:48.32 | *** join/#htc-linux dream_kill (n=nospam@92.56.53.50) |
| 17:48.53 | Epsylon3 | ok ok... will try understand haret source |
| 17:49.58 | dcordes | Epsylon3, tmzt was looking into this |
| 17:50.19 | Epsylon3 | and ? |
| 17:51.08 | Epsylon3 | yea, will ask him thanks |
| 17:51.10 | dcordes | Epsylon3, you might talk to him if you want to help developing an msm bootloader |
| 17:51.35 | Epsylon3 | ok, yep |
| 17:53.07 | Epsylon3 | do you know if android htc devices also uses nbh system for updates ? |
| 17:55.31 | dcordes | https://www.codeaurora.org/patches/quic/le/ |
| 18:03.24 | phh | dcordes: do you know anything about htc_battery ? |
| 18:04.33 | Epsylon3 | ok i found the problem : "It's not just a question of the bootloader. There's some hardware initialization being done by WM that is still needed." |
| 18:05.24 | Epsylon3 | ergh, i really need to set haret as default shell (like gps devices) |
| 18:14.55 | Epsylon3 | erfg, even when i connect a stabilised 4.3VDC source in parallel to the battery, i've same problem... booting then shutdown on PIN code |
| 18:18.01 | dcordes | phh, no |
| 18:19.55 | Epsylon3 | ahhh i'll maybe do it... i have some more seconds at 4,8VDC and usb powered |
| 18:22.57 | *** join/#htc-linux audrysu (n=audrius@rrcs-71-43-112-220.se.biz.rr.com) |
| 18:25.44 | *** join/#htc-linux pleemans (n=toi@d54C2A96D.access.telenet.be) |
| 18:47.59 | *** join/#htc-linux thedicemaster (n=thedicem@24.132.89.51) |
| 18:49.07 | *** join/#htc-linux stickboy (n=anonymou@128.153.211.150) |
| 18:49.30 | Epsylon3 | made the hard reset, no luck.... same battery problem |
| 18:50.36 | phh | anyway to cycle the battery ? |
| 18:50.46 | phh | (meaning discharge it and charge it totally) |
| 18:51.06 | Epsylon3 | i tested it yes, with a 0.5A 3.6V pocket lamp |
| 18:51.37 | Epsylon3 | that works... 45mn later i disconnected it |
| 18:52.10 | Gnutoo | dcordes, I default the kernel to 2.6.29 in oe for htcdream right? other branch are less advanced for GNU/Linux(not android)? |
| 18:52.18 | Epsylon3 | my device had a water problem some month ago |
| 18:52.48 | Epsylon3 | i think there is something bad with the battery sensor... |
| 18:52.53 | dcordes | Gnutoo, htc-msm-2.6.29 branch yes |
| 18:52.53 | Epsylon3 | or something like that |
| 18:52.57 | phh | maybe try a new battery |
| 18:53.14 | dcordes | Gnutoo, see how it's done for the other machines with the branch selection etc? |
| 18:53.17 | Epsylon3 | i have a big one too, same problem |
| 18:53.30 | phh | Epsylon3: isn't your windows rom the problem ? |
| 18:53.33 | Gnutoo | ok because someone told me that everything worked in an older branch but I don't know who it was and if he was talking about android or GNU/Linux |
| 18:53.44 | Epsylon3 | hmm maybe something like that yes |
| 18:53.53 | Epsylon3 | i dont had this problem with Android |
| 18:53.53 | Gnutoo | dcordes, I know how to do that |
| 18:54.03 | dcordes | ok |
| 18:54.09 | phh | Epsylon3: then flash the stock one |
| 18:54.21 | Epsylon3 | yeah, device is 2 years old :p |
| 18:54.31 | phh | ... and ? |
| 18:54.56 | Epsylon3 | dont know if i can find the old one... will try to flash another one yes |
| 18:55.15 | Epsylon3 | or maybe the htc one |
| 18:56.08 | Epsylon3 | i think its a problem with the battery driver (or sensor) |
| 18:56.23 | phh | that's why you should try another ROM ... |
| 18:56.27 | Epsylon3 | worked again a full week |
| 18:56.34 | Epsylon3 | last week |
| 18:57.55 | Epsylon3 | now all is erased i can do that sure.... |
| 18:58.33 | tmzt | cdma? |
| 18:59.05 | tmzt | try cleaning the contacts |
| 18:59.32 | dcordes | tmzt, I think there's only gsm kaiser |
| 18:59.41 | tmzt | yeah, they use nbh |
| 18:59.43 | tmzt | why? |
| 19:00.05 | Epsylon3 | to get the os.nb ;) |
| 19:00.32 | tmzt | it was recommended we start with kaisdiag.nbh |
| 19:00.40 | tmzt | I'm still looking in to that |
| 19:01.00 | tmzt | if you know arm asm and what the "periphrial port remap register" is you could really help |
| 19:01.02 | Epsylon3 | yea, like the diagnostic SD card |
| 19:01.29 | Epsylon3 | i worked on mips router WRT54G some years ago |
| 19:01.54 | Epsylon3 | and i'm coding Ollydbg Script plugin ;p |
| 19:02.18 | Epsylon3 | and also made PIC assembler |
| 19:02.22 | tmzt | Epsylon3: you could also use the early registry stuff (rgu) and #xda-devs should be able to help with that |
| 19:02.31 | tmzt | if you do that use startup.txt instead of default.txt |
| 19:02.46 | Epsylon3 | will autostart right ? |
| 19:03.40 | Epsylon3 | rgu is a shell commands to write registry via usb ? |
| 19:04.06 | tmzt | not via usb, you put it in xip |
| 19:04.37 | tmzt | cool, can you look at this, people.openezx.org/tmzt/green7.S |
| 19:04.41 | tmzt | it's a very early start |
| 19:04.59 | tmzt | other wise we can adapt Qi or uboot from qualcomm |
| 19:05.23 | Epsylon3 | =) |
| 19:05.55 | Epsylon3 | happy to know you, it's hard to find infos on xda forum |
| 19:06.08 | Epsylon3 | for that |
| 19:06.30 | tmzt | I don't know much about ce though, the people in #xda-devs, no2chem, cmonex, Olipro, TFGDB do |
| 19:06.46 | tmzt | TFGBD |
| 19:06.47 | Epsylon3 | the asm code directly access to GPIOs ? |
| 19:06.55 | tmzt | sure, no reason why not |
| 19:06.58 | Epsylon3 | write to vibra register |
| 19:07.01 | tmzt | but I'm using proc comm |
| 19:07.08 | tmzt | that's in the shared memory area |
| 19:07.15 | tmzt | the problem is none of it works booted from haret |
| 19:07.31 | tmzt | and I don't understand how the haret trampoline works |
| 19:07.51 | tmzt | since mmu has to be off to boot linux |
| 19:07.51 | Epsylon3 | was reading that in sources :p |
| 19:07.51 | Epsylon3 | some minuts ago |
| 19:08.35 | Epsylon3 | in linboot.cpp |
| 19:08.43 | tmzt | yes |
| 19:10.40 | Epsylon3 | i think its possible to force to free some app memory blocks... |
| 19:10.56 | *** join/#htc-linux cr2 (n=cr2@ip-109-84-56-184.web.vodafone.de) |
| 19:11.06 | Epsylon3 | but for a bootloader, i think we have almost all memory |
| 19:11.41 | tmzt | we do have all memory? what do you mean? |
| 19:12.45 | tmzt | that's just to ensure there's enough memory with physically contiguous mappings to boot |
| 19:12.46 | Epsylon3 | under windows, there is many things loaded yea |
| 19:13.13 | Epsylon3 | but on startup (before windows) all is free |
| 19:13.13 | tmzt | right, but after linboot |
| 19:13.25 | cr2 | hi |
| 19:13.29 | tmzt | not sure what you mean |
| 19:13.42 | tmzt | it's just to prevent it writing over windows before Go Go Go |
| 19:13.48 | tmzt | at which point ce is goine |
| 19:13.49 | tmzt | gone |
| 19:13.51 | *** join/#htc-linux Alex[sp3dev] (n=alex_dfr@ip-95-221-54-65.bb.netbynet.ru) |
| 19:13.58 | cr2 | tmzt: it's not so easy to disable pmem. the adsp code needs to be fixed, and i fear that we may loose the sound ;) |
| 19:14.18 | tmzt | disable pmem? |
| 19:14.20 | tmzt | ah |
| 19:14.24 | tmzt | why? |
| 19:14.33 | cr2 | !android |
| 19:14.55 | tmzt | I just mean we don't need the driver for userspace |
| 19:14.55 | cr2 | btw, where is the framebuffer patch of Gnutoo ? |
| 19:15.04 | tmzt | we should be able to replace it with sg |
| 19:15.06 | Gnutoo | cr2, in the git |
| 19:15.07 | tmzt | Gnutoo: ping |
| 19:15.11 | tmzt | hey |
| 19:15.13 | Gnutoo | tmzt, hi |
| 19:15.27 | Epsylon3 | i think firmwares are directly sent to other chips |
| 19:15.38 | cr2 | Gnutoo: ok, then we need to update the defconfig, and do it the default |
| 19:15.41 | tmzt | I don't see it |
| 19:15.44 | Gnutoo | ok |
| 19:15.53 | tmzt | what patch? |
| 19:15.58 | Gnutoo | I don't know what you do with defconfigs...if you do android or GNU/Linux |
| 19:16.05 | Gnutoo | so I didn't touched it |
| 19:16.13 | cr2 | if somebody wants android, they should use android_defconfig |
| 19:16.13 | Gnutoo | anyway I'm doing a defconfig for oe |
| 19:16.17 | Gnutoo | ah ok |
| 19:16.47 | cr2 | tmzt: kernel doublebuffering |
| 19:16.48 | tmzt | cr2: looks like android images will be built by somebody now |
| 19:16.58 | tmzt | if we can get at least raph800 included |
| 19:17.06 | tmzt | to disable it? we actually need it in X |
| 19:17.13 | cr2 | tmzt: ? |
| 19:17.20 | tmzt | I'm trying to adapt the s3c ddx to use it properly |
| 19:17.23 | cr2 | why do you need pmem ? |
| 19:17.32 | tmzt | we don't need pmem |
| 19:17.35 | tmzt | we can mmap fb |
| 19:17.59 | cr2 | yes. if we drop android from defconfig, we can purge a lot of crappy code |
| 19:18.28 | tmzt | I mean there's someone who is here occassionally who is building zImage's for android |
| 19:18.47 | tmzt | if Untouchable was using them, since people seem to like his userlands |
| 19:18.53 | tmzt | we could still support them |
| 19:18.58 | Epsylon3 | (ok, understund the rgu way... something like the OEM stuff in tuned XP distros) |
| 19:18.59 | *** join/#htc-linux Dayd (n=ddaydd@AToulouse-159-1-118-121.w92-129.abo.wanadoo.fr) |
| 19:19.05 | tmzt | Epsylon3: yes |
| 19:19.22 | Epsylon3 | yea, very nice way |
| 19:19.34 | Epsylon3 | thanks for help, will try that |
| 19:19.37 | cr2 | i don't really understand why there are multiple android images for each device. |
| 19:19.37 | tmzt | cr2: what do you think about splitting out gpio defines from the board headers |
| 19:19.46 | tmzt | and making RAPH100 the basic set of them |
| 19:19.50 | Gnutoo | btw I apoligize for the orthography |
| 19:19.57 | tmzt | and then having BLAK100, etc. only were they differ |
| 19:20.08 | tmzt | Gnutoo: can I see this patch? |
| 19:20.13 | Gnutoo | tmzt, cr2 if you need explanations on the options just ask |
| 19:20.21 | Gnutoo | tmzt, which one? |
| 19:20.37 | Alex[sp3dev] | tmzt: a great idea.. so as not to duplicate code, right? |
| 19:20.46 | tmzt | yes |
| 19:20.47 | Gnutoo | refresh(that is not from me),double-buffering,tslib ? |
| 19:20.54 | tmzt | Gnutoo: right |
| 19:20.59 | Gnutoo | which one of the 3? |
| 19:21.09 | tmzt | probably double-buffering |
| 19:21.17 | tmzt | if you have a clean msmts_raw though I would like that |
| 19:21.39 | cr2 | tmzt: we can create some more generic names for gpios |
| 19:21.43 | tmzt | by the way, I'm not using tslib anymore but evdev |
| 19:22.03 | tmzt | cr2: I'm fine with RAPH100 like it is now, I just want the additional devics in the same header |
| 19:22.15 | tmzt | so I can split the smi parts into smi32 and sim64 |
| 19:22.19 | tmzt | like on sapp kernel |
| 19:22.21 | cr2 | tmzt: as long as you don't peddle calibration into the kernel ;) it's all ok |
| 19:22.22 | tmzt | and fix my board |
| 19:22.27 | Alex[sp3dev] | sorry, i haven't followed the discussion for about two weeks.. has anyone looked into fixind the ram on kovsky and others yet? |
| 19:22.34 | Gnutoo | tmzt, ok I'll send you the patch's url |
| 19:22.42 | tmzt | cr2: X wants the min/max in kernel, but we can work around it using xinput |
| 19:22.50 | cr2 | tmzt: i'd like to get rid of fixed allocations there. |
| 19:22.57 | cr2 | as much as it is possible |
| 19:22.57 | tmzt | I want the calibration routine out of the kernel |
| 19:23.03 | cr2 | and move fbram into SMI |
| 19:23.08 | Gnutoo | tmzt, you don't want/need it for xorg/kdrive: http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=commit;h=d61ae9e6fd1616391f696cb662c388433891c426 |
| 19:23.19 | tmzt | it might make sense to have a separate tree for android with shared commits |
| 19:23.26 | tmzt | something git or at least topgit can do |
| 19:23.27 | Gnutoo | the help says: |
| 19:23.38 | Gnutoo | Android relies on a double buffer to smoothly render page flips.Select this options if you use Android |
| 19:23.45 | tmzt | cr2: I also want to use samsung version of mdp driver |
| 19:23.49 | tmzt | because it makes more sense |
| 19:23.57 | tmzt | that means fixing the panel/client drivers as well |
| 19:24.11 | cr2 | tmzt: codeaurora ? i want tvout |
| 19:24.18 | cr2 | as fb1 |
| 19:24.18 | tmzt | either way |
| 19:24.28 | tmzt | do we have a 2.6.27 version of it? |
| 19:24.37 | cr2 | no |
| 19:24.41 | tmzt | well, I need to disable tv completely |
| 19:24.53 | tmzt | make it an option and adjust smi accordingly |
| 19:25.02 | tmzt | because I only have 192mb total |
| 19:25.08 | tmzt | minus 40+ for amss |
| 19:25.22 | cr2 | option ? |
| 19:25.31 | tmzt | we can use the samsung version then |
| 19:25.34 | cr2 | can't we found its location in nand ? |
| 19:25.37 | tmzt | it has tvout and is based on ca |
| 19:25.41 | tmzt | nand? |
| 19:26.06 | cr2 | the "engineering" data is on nand |
| 19:26.15 | tmzt | sure |
| 19:26.21 | tmzt | cmonex knows where that is |
| 19:26.27 | *** join/#htc-linux Abracadabr4 (n=Abracada@92.24.47.47) |
| 19:27.13 | Gnutoo | btw there is that in oe in ./conf/machine/include/htc-msm7.inc : |
| 19:27.17 | Gnutoo | PREFERRED_PROVIDER_virtual/xserver = "xserver-kdrive" |
| 19:27.17 | Gnutoo | XSERVER = "xserver-kdrive-fbdev" |
| 19:27.21 | Gnutoo | what about xorg |
| 19:27.34 | Gnutoo | I had to comment them for shr |
| 19:27.39 | Gnutoo | because they use xorg |
| 19:27.45 | tmzt | talk to Weiss and Jama in #openmoko-cdevel |
| 19:27.53 | tmzt | yes |
| 19:28.04 | Gnutoo | me? |
| 19:28.07 | tmzt | if ANDROID is going into the kernel as one option, which I would prefer not |
| 19:28.16 | tmzt | please keep it as a config ANDROID select ... |
| 19:28.20 | tmzt | not depends |
| 19:28.24 | Gnutoo | tmzt, talk to Weiss and Jama in #openmoko-cdevel was for me? |
| 19:28.26 | cr2 | tmzt: ANDROID_PMEM is enough |
| 19:28.35 | tmzt | also, double-buffer should be a seperate issue from android |
| 19:28.46 | tmzt | I need to be able to enable disable it completely seperately |
| 19:28.48 | tmzt | Gnutoo: yes |
| 19:28.51 | Gnutoo | tmzt, it's and android specific bit according to: |
| 19:28.59 | Gnutoo | http://www.kandroid.org/android_pdk/display_drivers.html |
| 19:29.01 | tmzt | it historical |
| 19:29.09 | tmzt | from the emulator |
| 19:29.16 | tmzt | it's wasted a year of our time :) |
| 19:29.17 | Gnutoo | I heard that double buffering can be done differently for kdrive/xorg |
| 19:29.36 | Gnutoo | tmzt, ok thanks |
| 19:29.36 | tmzt | Gnutoo: I would suggest looking at the new s3c6410 ddx |
| 19:29.55 | Gnutoo | tmzt, ok what's that? |
| 19:30.01 | tmzt | g1 doesn't even use that method, it uses a gles driver for swap buffers |
| 19:30.06 | Gnutoo | apart that s3c is a soc |
| 19:30.08 | tmzt | driver for Xorg |
| 19:30.10 | cr2 | tmzt: i have seen a recent rpc bugfix in codeaurora |
| 19:30.19 | Gnutoo | ok so userspace xorg driver |
| 19:30.25 | tmzt | right |
| 19:30.35 | tmzt | using vyres like the "android" one |
| 19:30.37 | Gnutoo | ok |
| 19:30.39 | tmzt | it can be adapted |
| 19:30.49 | Gnutoo | what should I do with this ddx driver? |
| 19:30.50 | Gnutoo | ok |
| 19:31.11 | Gnutoo | I'll go in #openmoko-cdevel to ask for the ./conf/machine/include/htc-msm7.inc |
| 19:35.42 | tmzt | Gnutoo: sorry, I mean I thought we had to disable the double buffering for X to work, but we should really fix instead. the method google used is pretty much standard for embedded framebuffers |
| 19:36.07 | *** join/#htc-linux stickboy (n=anonymou@128.153.211.150) |
| 19:36.16 | Gnutoo | tmzt, ah ok sorry then...btw Xorg can work with it...just that it's not great because it breaks autodetection |
| 19:36.27 | tmzt | with? |
| 19:36.38 | Gnutoo | with the double buffering patch |
| 19:36.41 | tmzt | the solution is to put the driver in udev |
| 19:36.46 | tmzt | until there's a better way |
| 19:36.51 | tmzt | that's why we need a custom ddx |
| 19:37.09 | tmzt | I like the Mer way, keep everything generic but have per-device packages or sets of packages |
| 19:37.15 | tmzt | that customize certain things |
| 19:37.20 | cr2 | hehe. ppp over adb. insane :) |
| 19:37.25 | tmzt | I'm sure oe/shr is similar |
| 19:37.44 | tmzt | cr2: can we use gadget instead of function in our new kernel? |
| 19:38.02 | tmzt | cr2: I've tried it but I don't have phy setup right so detect no longer works |
| 19:38.23 | Gnutoo | tmzt, I use gadget |
| 19:38.28 | Gnutoo | but on GNU/Linux |
| 19:38.29 | tmzt | you do? |
| 19:38.31 | tmzt | on 2.6.27? |
| 19:38.35 | Gnutoo | 2.6.29 |
| 19:38.42 | tmzt | how is that working for you? |
| 19:38.53 | tmzt | should we switch to it for X and leave 2.6.27 for android? |
| 19:39.06 | tmzt | do you have any of that wifi stuff patching core left in there? |
| 19:39.10 | Gnutoo | fine |
| 19:39.15 | tmzt | or is this a clean 2.6.29 from ca again? |
| 19:39.32 | tmzt | okay, you can support raw msmts? |
| 19:39.39 | tmzt | no calibration/vkeyb? |
| 19:39.40 | Gnutoo | if you're talking to me for wifi I've not added wii |
| 19:39.55 | tmzt | ok, openmoko-msm is not you? |
| 19:40.01 | Gnutoo | no |
| 19:40.03 | tmzt | ok |
| 19:40.03 | Gnutoo | it's leviathan |
| 19:40.04 | Gnutoo | 's |
| 19:40.06 | tmzt | right |
| 19:40.20 | tmzt | wl12xx does work with compat though? |
| 19:40.21 | Gnutoo | but replicant's kernel patches which just get merged was me |
| 19:40.24 | Gnutoo | it works |
| 19:40.36 | Gnutoo | just that I was hacking on alsa...so I didn't have the time to try |
| 19:40.37 | tmzt | nice, so I port raph500 to it it should work |
| 19:40.45 | tmzt | when I get calibration |
| 19:41.00 | tmzt | where is your tree ltg? |
| 19:41.31 | Gnutoo | I used tslib |
| 19:41.31 | tmzt | whose patches? |
| 19:41.38 | Gnutoo | so calibration is in userspace |
| 19:41.41 | tmzt | wifi calibration |
| 19:41.42 | tmzt | I mean |
| 19:41.44 | tmzt | right |
| 19:41.58 | Gnutoo | ah ok I thought you were talking about tslib |
| 19:42.05 | Gnutoo | I didn't try wifi at all so I don't know |
| 19:42.13 | Gnutoo | I didn't have time to try it |
| 19:42.15 | tmzt | aer you on Xfbdev? |
| 19:42.19 | tmzt | if not, try evdev |
| 19:42.42 | tmzt | wait, this is on g1? |
| 19:42.46 | tmzt | we don't have rmi driver |
| 19:43.02 | Gnutoo | let's restart the conversation from scratch...I'm lost |
| 19:43.09 | Gnutoo | so I've an htcdream |
| 19:43.41 | tmzt | okay, I'm using a msm-single directory in drivers/video |
| 19:43.41 | Gnutoo | I've made 2 patch and 1 is not mine but was commited as beeing made by me(error from the commiter) |
| 19:44.15 | Gnutoo | one patch is for having tslib for the synaptics_rmi driver |
| 19:44.20 | Gnutoo | which the htcdream uses |
| 19:44.34 | Gnutoo | the second patch is for better xorg auto-detection |
| 19:44.37 | tmzt | the correct approach on double buffer seems to be to divide the fbram value by size of the screen in bytes |
| 19:44.44 | tmzt | and determine the number of screens from there |
| 19:44.53 | tmzt | look at s3c6410 fb driver |
| 19:44.58 | Gnutoo | ok |
| 19:45.17 | Gnutoo | thanks |
| 19:45.24 | tmzt | it's not "auto detection" fbdev ddx will fail if the fbset fails to succeed by X's terms |
| 19:45.39 | tmzt | okay, any reason we can't support wm device with this kernel? |
| 19:46.01 | tmzt | we should probably document are changes and create a patchset |
| 19:46.07 | tmzt | but that requires amss_ops |
| 19:46.15 | Epsylon3 | android 2.0 : http://www.boygeniusreport.com/2009/10/16/android-2-0-screenshot-walkthrough/ |
| 19:46.16 | tmzt | cr2: have you though about what that might look like? |
| 19:46.23 | tmzt | Epsylon3: it's not 2.0 |
| 19:46.32 | Epsylon3 | oh ? |
| 19:46.32 | *** join/#htc-linux luc_ (n=luc@89-115-128-35.cl.ipv4ilink.net) |
| 19:46.49 | tmzt | ctate in #android says there is no offical release number for eclair |
| 19:47.17 | tmzt | cr2: amss_ops I mean |
| 19:47.39 | tmzt | we should drop all ifdef's on machine where possible |
| 19:47.40 | Epsylon3 | ok, could be a 1.8 or something |
| 19:47.51 | tmzt | Epsylon3: could be, but as I said, there's no official number |
| 19:48.19 | balsat | 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 |
| 19:48.20 | cr2 | tmzt: what about rpc numbers and versions and parameters ? |
| 19:48.42 | tmzt | keep the separate .o's for amss |
| 19:49.03 | tmzt | but have a per board declration of which is used |
| 19:49.17 | cr2 | it's per-amss |
| 19:49.18 | tmzt | borrow the infrastructure from pxa that google refused to use |
| 19:49.21 | cr2 | not per-board |
| 19:49.28 | tmzt | set_whatever_pdata |
| 19:49.37 | tmzt | yes |
| 19:49.42 | tmzt | sorry |
| 19:49.55 | tmzt | I mean you need amss_xxxx.o in qdsp5/ |
| 19:50.02 | tmzt | as well as board-amss-xxxx.o in msm/ |
| 19:50.13 | cr2 | not really |
| 19:50.24 | cr2 | amss version is shared by many boards |
| 19:50.26 | tmzt | it will setup amss_ops for things like gpios and proc_comm |
| 19:50.40 | tmzt | yes, the file name is board-amss-$(VERSION) |
| 19:50.42 | cr2 | gpios are per-board |
| 19:50.48 | tmzt | not htcraphael-amss-$(VERSION) |
| 19:51.02 | tmzt | but the method of using gpios is per amss, isn't it? |
| 19:51.07 | cr2 | no |
| 19:51.16 | tmzt | amss->set_gpio_value(nr, value) |
| 19:51.25 | cr2 | why amss ? |
| 19:51.27 | tmzt | amss->msm_set_gpio_function(nr, func) |
| 19:51.36 | tmzt | because that's where ce and g1 differ |
| 19:51.44 | cr2 | blac100 and raph100 has the same amss, but the gpios are differemnt |
| 19:51.52 | tmzt | so you have |
| 19:52.05 | tmzt | config CONFIG_MACH_HTCDIAMOND |
| 19:52.16 | tmzt | ' select CONFIG_AMSS_6250 or whatever |
| 19:52.21 | *** join/#htc-linux rayman18 (i=opera@36.pool85-49-144.dynamic.orange.es) |
| 19:52.26 | tmzt | the way google did it is wrong |
| 19:52.30 | tmzt | that should have been a select |
| 19:52.34 | tmzt | not a depends |
| 19:52.35 | cr2 | diamond does not support multiple amss |
| 19:52.47 | cr2 | unless you 'd like to merge cdma |
| 19:52.51 | tmzt | select keyword selects one of them |
| 19:53.10 | tmzt | we can only merge cdma when smi infrastructure is in place |
| 19:53.29 | tmzt | I mean the Kconfig select keyword selects a particular config option |
| 19:53.30 | cr2 | smi has nothing to do with cdma |
| 19:53.38 | mdrobnak | Hey everyone. |
| 19:53.39 | cr2 | in principle |
| 19:53.41 | tmzt | it does for raph500, I need smi64 |
| 19:53.58 | cr2 | it's just board-raph500 specific setting |
| 19:54.01 | tmzt | I mean we can't merge the boards until the defines are renamed SMI32_whatever |
| 19:54.02 | *** join/#htc-linux Abracadabr4 (n=Abracada@92.24.47.47) |
| 19:54.12 | tmzt | not if there is a conflict in the pre processor |
| 19:54.16 | cr2 | they could have created a device with 256+smi64 |
| 19:54.27 | tmzt | sure |
| 19:54.41 | tmzt | anyway |
| 19:54.42 | mdrobnak | tmzt: You wanted Raph110 keyboard map for console? |
| 19:54.47 | tmzt | this fixes the proc_comm_wince |
| 19:54.56 | tmzt | we do use the correct structure now right? |
| 19:55.02 | tmzt | mdrobnak: yes, can we please collect those |
| 19:55.14 | mdrobnak | tmzt: http://www.drobnak.com/raph110.kmap.gz |
| 19:55.16 | tmzt | give it dcordes so he can put on ~lgorris ? |
| 19:55.23 | tmzt | ok |
| 19:55.50 | tmzt | dcordes: can you put a copy on ~lgorris |
| 19:56.00 | tmzt | dcordes: I will help collect the linux and android versions |
| 19:56.01 | cr2 | tmzt: how much ram does ADSP and MDP need ? |
| 19:56.13 | tmzt | minimum? I think mdp can use two screens |
| 19:56.25 | tmzt | there's no way adsp needs 8mb |
| 19:56.40 | cr2 | tmzt: they use 10MB now ;) |
| 19:56.45 | tmzt | I can't tell you what amss wants though |
| 19:56.51 | cr2 | for jpeg encoding or so |
| 19:56.55 | tmzt | yeah |
| 19:57.08 | cr2 | #define MSM_FB_SIZE 0x200000 /* 2mb */ |
| 19:57.08 | tmzt | cr2: I don't think I'm explaining myself well |
| 19:57.15 | cr2 | this is just plain stupid. |
| 19:57.20 | cr2 | ok |
| 19:57.20 | tmzt | fine, but then they add gpu to that |
| 19:57.45 | tmzt | cr2: by amss_ops I mean the things that we need to be able to define for g1 and ce seperately |
| 19:57.54 | tmzt | for most things the ops will be shared between amss versions |
| 19:58.13 | tmzt | the only thing the CONFIG_AMSS_xxxx should do is determine whether a particular versio is built |
| 19:58.17 | tmzt | not which is used |
| 19:58.24 | tmzt | DEFAULT should go away |
| 19:58.26 | tmzt | it's just broken |
| 19:58.36 | cr2 | ok |
| 19:58.45 | tmzt | there's no reason a mach-msm zImage should not work on g1, sapp, raph100, raph800, raph500, etc. |
| 19:58.54 | tmzt | including blak, kovs and whatever else |
| 19:58.58 | cr2 | tmzt: can you explain me what is GPU0 and GPU1 ? |
| 19:59.08 | tmzt | offscreen stuff I don't understand |
| 19:59.17 | tmzt | but gpu also uses the fb define |
| 19:59.19 | mdrobnak | That's android related stuff, no? |
| 19:59.22 | tmzt | for the front and back buffer |
| 19:59.25 | cr2 | can they go away for non-android ? |
| 19:59.28 | tmzt | for now |
| 19:59.30 | mdrobnak | I haven't dug into the Android stuff enough. |
| 19:59.46 | tmzt | until we can use HW3D or a better replacement |
| 20:00.07 | cr2 | yes. we are happily using pxafb |
| 20:00.17 | cr2 | without any HW3D |
| 20:00.20 | tmzt | cr2: anyway, amss_ops then would have proc_comm and set_gpio and set_gpio_function and some clocks stuff |
| 20:00.32 | tmzt | so we can do them manually, g1 can use amss, etc. |
| 20:00.48 | tmzt | this will also determine the method used to access the battery smem data |
| 20:00.53 | tmzt | or rpc, etc. |
| 20:01.08 | tmzt | this will not include qdsp functions, those will be changed to something similar later |
| 20:01.17 | tmzt | for now we build per machine for that |
| 20:01.39 | tmzt | each arch/arm/mach-msm/amss_xxxx.o defines a global symbol |
| 20:01.50 | tmzt | which is the struct amss_ops amss_xxxx_ops |
| 20:01.55 | tmzt | static struct |
| 20:02.03 | tmzt | including the function pointers for that version of amss |
| 20:02.26 | tmzt | okay, I did mess up early |
| 20:02.38 | tmzt | the board's machine_init function will call a global function |
| 20:02.47 | tmzt | msm_set_board_amss_ops |
| 20:02.54 | tmzt | to set the pointer to that struct |
| 20:03.10 | tmzt | then all of proc_comm_wince, etc. will be a stub using the ops functions |
| 20:03.18 | cr2 | you can pick the amss version in board_init |
| 20:03.26 | tmzt | yes, or per board setup or whatever |
| 20:03.27 | cr2 | it's in smem already |
| 20:03.39 | tmzt | fine, if that can be trusted |
| 20:03.55 | cr2 | yes |
| 20:04.02 | tmzt | but this should fix rpc and let us use a reference of g1 |
| 20:04.10 | cr2 | we rely on the SMD layout and other things done by amss |
| 20:05.12 | cr2 | how do we know about the MPU locked ram for amss ? |
| 20:05.30 | tmzt | I guess there is one issue, the raph500 smd layout is different but uses the same amss version as raph800? |
| 20:05.34 | tmzt | what does that mean? |
| 20:05.39 | cr2 | without extracting amss from nand, and parsing the elf |
| 20:05.41 | tmzt | yeah, I think xda people know the mpu |
| 20:06.33 | cr2 | saving us the static allocation mess in board-*.h will be great |
| 20:06.45 | tmzt | how will that work? |
| 20:06.51 | cr2 | i thing topa100 does not work because of that |
| 20:07.37 | cr2 | you enable all available ram in _fixup() |
| 20:07.42 | tmzt | me? |
| 20:07.44 | cr2 | look what amss needs, and so on |
| 20:07.46 | tmzt | mine is still broken |
| 20:07.56 | cr2 | not you, the kernel code. |
| 20:08.04 | cr2 | we check for the amss version already |
| 20:08.07 | tmzt | I don't understand |
| 20:08.14 | tmzt | how do you read this layout? |
| 20:08.19 | mdrobnak | So we are going to try and do this re-organization of the current tree? |
| 20:08.20 | cr2 | just don't bring it to the end solution |
| 20:08.41 | tmzt | mdrobnak: I would say pull a new tree from google or better ca |
| 20:08.50 | cr2 | mdrobnak: switching to codeaurora 2.6.29 ? |
| 20:09.16 | tmzt | this is why we need a document outline the changes need for htc ce A devices |
| 20:09.23 | cr2 | but it's easier to fix the mess in the current tree first |
| 20:09.25 | tmzt | so we can make them as a clean set of abstractions |
| 20:09.33 | cr2 | instead of trying to merge it into 2.6.29 |
| 20:09.37 | tmzt | yeah |
| 20:09.41 | mdrobnak | Well, Here's my thing... The Samsung stuff is based on .27 .. .and that has some of what we are talking about, right? |
| 20:09.41 | tmzt | we know 2.6.27 better |
| 20:09.50 | tmzt | yeah |
| 20:10.08 | tmzt | the fact is we should be able to go to any kernel we want, including upstream, once we know what's happening |
| 20:10.16 | mdrobnak | So maybe we should clean up as much as possible, then move to .29...then latest.. |
| 20:10.51 | cr2 | mdrobnak: yes. i'd like to separate android as the first step. and cleanup the ram allocation |
| 20:11.18 | mdrobnak | cr2: Well, here's my issue with that. All the 'reference' code we have...is for Android. |
| 20:11.37 | *** part/#htc-linux rayman18 (i=opera@36.pool85-49-144.dynamic.orange.es) |
| 20:11.48 | cr2 | but the root problem now is to determine the MPU'd areas dynamically for $board != raph100 |
| 20:11.49 | mdrobnak | that being said..I see no reason to seperate out Android code. |
| 20:12.04 | mdrobnak | How can we do that? |
| 20:12.08 | tmzt | no2chem: sorry, here is better |
| 20:12.09 | mdrobnak | Or we don't know? |
| 20:12.42 | cr2 | mdrobnak: it's easier t do certain things without thinking about android. |
| 20:12.42 | no2chem | oh |
| 20:12.56 | no2chem | mpu is controlled by a9, so |
| 20:13.04 | no2chem | why do you need to control it? |
| 20:13.29 | tmzt | cr2: what is the goal with dumping elf? |
| 20:13.32 | cr2 | no2chem: not to control it, but to look which areas are requested by amss |
| 20:13.42 | no2chem | you mean locked? |
| 20:13.59 | cr2 | tmzt: the elf has all areas listed. |
| 20:14.03 | cr2 | no2chem: yes |
| 20:14.19 | mdrobnak | cr2: But can we dump the ELF now? |
| 20:14.23 | cr2 | no2chem: so we don't try to use it for linucx |
| 20:14.45 | tmzt | I'm lost, you mean the areas of memory it has in use and we can't use? |
| 20:14.46 | cr2 | mdrobnak: yes, you can disable the NAND MPU and dump the amss from nand |
| 20:14.55 | tmzt | not the ones it allocates for us, I see |
| 20:15.15 | tmzt | no2chem: you were able to gain additional ram with custom nk right? |
| 20:15.16 | mdrobnak | cr2: That sounds like a great idea. I imagine the parsing problem is much harder. |
| 20:15.19 | cr2 | tmzt: the areas it allocates for itself. |
| 20:15.23 | cr2 | oemsbl and such |
| 20:15.38 | tmzt | oemsbl should not be in use when amss is running |
| 20:16.09 | no2chem | eh |
| 20:16.09 | cr2 | tmzt: but this ram location is mpu'd |
| 20:16.13 | cr2 | so you can't reuse it |
| 20:16.47 | cr2 | no2chem: dzo disabled the mpu for spl@0x0 |
| 20:16.56 | no2chem | well, you've read the msm7200 interface manual |
| 20:17.04 | no2chem | its there.. in chapter 2 |
| 20:17.05 | cr2 | so we can actually overwrite and use this memory for our needs |
| 20:17.10 | no2chem | ah |
| 20:17.11 | no2chem | nice |
| 20:17.32 | tmzt | no2chem: where you able to get more with with nk changes? |
| 20:17.41 | no2chem | too bad i don't know where MPU_DEF_PART_PERM1 is |
| 20:17.53 | no2chem | tmzt: i dunno, most of the work ive been doing is bsp |
| 20:19.09 | cr2 | no2chem: btw, do you know the rpc clock id -> clock name mapping ? |
| 20:19.44 | cr2 | 0x33, 0x28 and 0x29 are mdc_clk, vfe_clk and mdc_vfe_clk. but i don't know which one is what. |
| 20:21.08 | tmzt | what they do? |
| 20:21.08 | tmzt | or which number? |
| 20:21.09 | cr2 | camera |
| 20:21.47 | cr2 | tmzt: i have this amss allocation problem with topaz now |
| 20:22.03 | cr2 | wince uses only 0x51=81MB |
| 20:22.07 | tmzt | you have topaz? |
| 20:22.20 | tmzt | mdc is mobile display client I think |
| 20:22.25 | cr2 | no, but i have written some code to support it. amss6120 |
| 20:22.33 | tmzt | since the msm is the client in the case of camera that makes sense |
| 20:22.46 | tmzt | vfe is video front end which is hardware as well as dsp |
| 20:22.47 | cr2 | tmzt: these clocks are controleed over rpc |
| 20:22.55 | cr2 | in wince |
| 20:23.02 | tmzt | mdc_vfe_clk most be the dma engine between those two parts |
| 20:23.34 | cr2 | #define MSM_EBI_LOCKED_BASE MSM_FB_BASE + MSM_FB_SIZE |
| 20:23.35 | cr2 | #define MSM_EBI_LOCKED_SIZE 0xD00000 /* 13mb */ |
| 20:23.47 | cr2 | these values are for raph100 |
| 20:24.05 | cr2 | nobody checked them on other devices. |
| 20:24.20 | cr2 | maybe they are ok for amss52xx |
| 20:24.30 | cr2 | but who knows about amss612x |
| 20:25.38 | tmzt | well, why do you need to know this at runtime? |
| 20:25.45 | tmzt | isn't it in radio.img/nb? |
| 20:28.11 | mdrobnak | cr2: The problem with that locked base thing -- if you change the layout...that changes. *Can* it change? Ie, did I move something I shouldn't have in mine? |
| 20:28.52 | tmzt | can we probe this with haret if we have to? |
| 20:28.55 | mdrobnak | Ie, I changed: #define MSM_PMEM_MDP_BASE0x23D00000 |
| 20:28.57 | tmzt | just try writes and see what happens? |
| 20:29.02 | *** join/#htc-linux dcordes_ (n=dcordes_@unaffiliated/dcordes) |
| 20:29.30 | mdrobnak | Which moved all the pmem stuff into the 2nd bank..giving me 107MB of RAM (cr2 said up to 115MB) |
| 20:30.21 | tmzt | how much do you normally have? |
| 20:30.34 | mdrobnak | 89MB in the first bank.. |
| 20:30.37 | tmzt | when I inscreased bank1 above 109 I got fb corruption |
| 20:30.52 | tmzt | below that it seems to work |
| 20:30.53 | cr2 | mdrobnak: 640*480 18bit LCD. how much ram does it need ? |
| 20:30.59 | cr2 | 640*480*4 ? |
| 20:31.27 | tmzt | yeah, think so |
| 20:31.29 | cr2 | mdrobnak: the raph100 memory layout is documented. |
| 20:31.31 | tmzt | since it's not packed |
| 20:31.42 | tmzt | can you use 16bit with your client? |
| 20:31.46 | mdrobnak | cr2: About 1MB? 3bytes / pixel? |
| 20:31.56 | tmzt | there's no point in using 18 because you lose so much ram |
| 20:32.01 | tmzt | mdrobnak: very slow on arm |
| 20:32.04 | cr2 | 1228800 |
| 20:32.06 | cr2 | bytes |
| 20:32.23 | cr2 | tmzt: we don't know how to change the panel to 16bit |
| 20:32.45 | tmzt | we have the registers don't we? |
| 20:32.50 | tmzt | which client? |
| 20:32.52 | tmzt | oh, panel |
| 20:33.03 | tmzt | can't the mddi scan out 18 with 16 buffer? |
| 20:33.03 | cr2 | 1536000 for 800*480*4 |
| 20:33.26 | mdrobnak | cr2: So I'm not sure if you answered my question - is the FB movable, or must it be at a certain address? |
| 20:33.40 | cr2 | mdrobnak: it's movable |
| 20:33.53 | mdrobnak | Ok. Thanks. |
| 20:33.56 | cr2 | mdrobnak: you may move it to SMI, because this one is faster |
| 20:34.00 | cr2 | afair |
| 20:34.18 | cr2 | but then you need to get rid of this GPU0 android blob |
| 20:34.19 | mdrobnak | It seemed Ok to me. |
| 20:34.20 | mdrobnak | :-) |
| 20:34.42 | mdrobnak | My coworker with a non-S iPhone is jealous of my Fuze in Android hehehe |
| 20:34.51 | tmzt | gpu0 is probably not being used, but unless you aren't doing dma from it to gpu it will break things |
| 20:34.54 | cr2 | dzo said somehting like that android does want it to be in SMI |
| 20:34.57 | tmzt | is my understanding |
| 20:35.10 | tmzt | mdrobnak: android works with full phone on raph100/110 now? |
| 20:35.16 | leviathan | Gnutoo: did you push now? |
| 20:35.17 | mdrobnak | tmzt: Using it everyday. |
| 20:35.27 | Gnutoo | one sec |
| 20:35.32 | leviathan | ok |
| 20:35.32 | Epsylon3 | there is a project htc-tools/htc-boot to flash a zImage and boot from usb bootloader : http://git.linuxtogo.org/?p=ph5/htc-tools.git;a=summary |
| 20:35.34 | Gnutoo | leviathan, push what? alsa or in oe? |
| 20:35.40 | Gnutoo | I'm about to push in oe |
| 20:35.41 | cr2 | mdrobnak: that's why if i had a non-android setup, i would not wait a second to remove the GPU0 :) |
| 20:36.01 | Gnutoo | leviathan, btw : |
| 20:36.07 | Gnutoo | *you said oe was broken |
| 20:36.14 | cr2 | Epsylon3: it's unfinished. and it's for older devices. |
| 20:36.14 | Gnutoo | *you wanted oe commit access |
| 20:36.23 | cr2 | Epsylon3: for newer devices there is htcflasher |
| 20:36.24 | Gnutoo | so maybe send a lot of patches in the oe ml |
| 20:36.26 | mdrobnak | cr2: Here's my thought. If you can start ifdefing items...and not break android..go right ahead :-) |
| 20:36.36 | Epsylon3 | yea compiled it but support magician hx4700 and another one |
| 20:36.39 | Gnutoo | because if you said it was broken you may have fixes |
| 20:36.49 | Epsylon3 | blue angel |
| 20:36.49 | leviathan | Gnutoo: I've now built a running mokotouch image |
| 20:37.04 | Gnutoo | leviathan, ok do they use oe? |
| 20:37.15 | leviathan | I've now fixed and merged the fixes into current org.oe.dev |
| 20:37.18 | Gnutoo | ok |
| 20:37.19 | cr2 | mdrobnak: why 107MB ? |
| 20:37.24 | Gnutoo | did you get commit access btw? |
| 20:37.32 | leviathan | Gnutoo: hmm, my mokotouch-oe-package does ;) |
| 20:37.35 | leviathan | Gnutoo: nope |
| 20:37.38 | Gnutoo | ok |
| 20:37.40 | leviathan | they refused |
| 20:37.47 | Gnutoo | maybe send more patches |
| 20:37.47 | leviathan | I stopped trying for now |
| 20:37.48 | Gnutoo | then |
| 20:37.48 | tmzt | they don't want it in core |
| 20:37.53 | tmzt | try shr or whatever |
| 20:37.55 | tmzt | or angstrom |
| 20:38.09 | Gnutoo | I saw only the htcdream support message |
| 20:38.17 | Gnutoo | I'll read my mail later |
| 20:38.20 | leviathan | yeah, or upload it to gitorious, then they can merge it theire own |
| 20:38.21 | *** join/#htc-linux cr2 (n=cr2@ip-109-84-56-184.web.vodafone.de) |
| 20:38.23 | leviathan | =D |
| 20:38.29 | cr2 | hm |
| 20:39.27 | cr2 | mdrobnak: the amss area starts at 0x17300000 |
| 20:39.38 | mdrobnak | hmm |
| 20:39.45 | cr2 | these are 155MB |
| 20:39.52 | cr2 | s/155/115/ |
| 20:40.07 | cr2 | the 114'th MB has wince dmesg |
| 20:40.15 | mdrobnak | My Linux area should end at 0x16000000 |
| 20:40.19 | cr2 | so it's clearly writable from arm11 |
| 20:40.30 | cr2 | mdrobnak: why ? |
| 20:40.33 | tmzt | ah |
| 20:40.44 | mdrobnak | That's just the way I had it set up. |
| 20:40.54 | mdrobnak | The diamond board had 107 as well |
| 20:40.58 | mdrobnak | I figured it was well tested ;-) |
| 20:41.06 | cr2 | wince fb starts here 0x1666a000 |
| 20:41.16 | mdrobnak | fwiw, the Samsung kernel uses 109MB. |
| 20:41.32 | cr2 | it is all amss -dpeendent |
| 20:41.48 | cr2 | i'm talking about raph100 with 52xx amss now. |
| 20:41.51 | mdrobnak | ...which is why if we could parse it, we'd have a better idea. |
| 20:41.59 | mdrobnak | *better idea of where we can use or not use. |
| 20:41.59 | cr2 | other devices are not that interesting |
| 20:42.16 | mdrobnak | True, cause I don't have one :-D |
| 20:42.17 | leviathan | Gnutoo: BTW: I ment the alsa driver |
| 20:42.38 | cr2 | mdrobnak: well. do 'pwf mywincedmesg 0x17200000 0x100000' |
| 20:42.48 | mdrobnak | But for now, I wanted something I could use, until we could get back to the RAM issue. |
| 20:42.52 | Gnutoo | leviathan, ok I'll push in replicant's repo after pushing into oe |
| 20:43.07 | Gnutoo | pusing into oe means not for the alsa driver |
| 20:43.11 | cr2 | mdrobnak: i hit this issue for topa100 |
| 20:43.14 | Gnutoo | I don't push alsa to oe |
| 20:43.15 | leviathan | Gnutoo: Your pushing it into OE? |
| 20:43.15 | Gnutoo | yet |
| 20:43.18 | leviathan | ahh, ok |
| 20:43.24 | Gnutoo | leviathan, yes kernel and machine.conf |
| 20:43.27 | Gnutoo | basic things only |
| 20:43.33 | Gnutoo | dcordes, wanted me to push things |
| 20:43.36 | cr2 | because it's different amss, and the memory layout seems to be different. and i don't have amss dump for it |
| 20:43.37 | mdrobnak | cr2: Uptime 23:47:30 on Android :-) |
| 20:43.40 | mdrobnak | I'll reboot. |
| 20:43.40 | Gnutoo | so I push |
| 20:43.48 | cr2 | mdrobnak: :D |
| 20:43.56 | leviathan | okee, good that you're telling me that, because I was going to do the same ^^" |
| 20:44.24 | leviathan | so, is it now in gitorious? |
| 20:44.36 | tmzt | cr2: wince fb is not used after htcfb finishes? |
| 20:44.45 | tmzt | can that me iounmapped? |
| 20:44.57 | mdrobnak | tmzt: Btw, still not sure how I want to do the android keymap. |
| 20:45.02 | cr2 | tmzt: only by druidu-fb |
| 20:45.26 | cr2 | tmzt: we should ioremap it in the first place |
| 20:45.55 | cr2 | tmzt: using SMI :) |
| 20:46.05 | leviathan | Gnutoo: in OE, ok |
| 20:46.07 | leviathan | :) |
| 20:46.13 | mdrobnak | cr2: Btw, tmzt told me someone tried to flash a g1 AMSS to a raph or diam...do you know if it worked? |
| 20:46.18 | cr2 | tmzt: if you thing GPU0 is unused by android, we can drop it |
| 20:46.34 | cr2 | mdrobnak: how can it work ? |
| 20:46.42 | tmzt | mdrobnak: dream_kill? |
| 20:46.50 | cr2 | mdrobnak: no cpld, and all the gpios are different. |
| 20:46.59 | mdrobnak | I figured as much. |
| 20:47.03 | tmzt | I don't know, but I was told yesterday that pretty much everything is recoverable with serial cable, even on g1 |
| 20:47.06 | mdrobnak | tmzt: Yeah I tink that's it. |
| 20:47.08 | cr2 | so you can't just use the plain g1 kernel |
| 20:47.10 | mdrobnak | *that's who |
| 20:47.33 | cr2 | tmzt: unless you blow away the oemsbl |
| 20:47.35 | mdrobnak | getting wince message now |
| 20:48.06 | mdrobnak | Is it just me, or is HaRET slower in WinMo 6.5? |
| 20:48.32 | cr2 | mdrobnak: let's try to squeeze the last byte from the first EBI1 bank :) |
| 20:48.50 | tmzt | cr2: true |
| 20:48.57 | leviathan | Gnutoo: did you push now? |
| 20:49.02 | mdrobnak | It's a shame the second one is a pain in the butt with Android. |
| 20:49.03 | Gnutoo | in oe yes |
| 20:49.04 | tmzt | cr2: but we can do that in ram at runtime without issue I would think |
| 20:49.09 | mdrobnak | cr2: I'll try a build with 114MB. |
| 20:49.15 | Gnutoo | I'll push in replicant now |
| 20:49.15 | cr2 | mdrobnak: android problem. |
| 20:49.16 | leviathan | hmm, did a pull, nothing there |
| 20:49.26 | leviathan | which branch? in OE? |
| 20:49.26 | cr2 | mdrobnak: 115 |
| 20:49.37 | cr2 | mdrobnak: but you need to put fbram somewhere else |
| 20:49.41 | leviathan | oh |
| 20:49.42 | leviathan | now |
| 20:49.43 | leviathan | delay |
| 20:49.59 | mdrobnak | cr2: Fbram is already moved.. |
| 20:50.05 | mdrobnak | cr2: Wow that's a long dmesg |
| 20:50.09 | cr2 | mdrobnak: where ? |
| 20:50.31 | mdrobnak | (The one I just dumped from WinCE) |
| 20:50.35 | mdrobnak | err |
| 20:50.42 | cr2 | mdrobnak: i mean the fbram |
| 20:51.06 | mdrobnak | yeah |
| 20:51.08 | mdrobnak | I realized |
| 20:51.09 | mdrobnak | hehe |
| 20:51.10 | mdrobnak | one sec |
| 20:51.36 | cr2 | #define MSM_EBI_SIZE 0x8000000 /* 128mb */ |
| 20:51.37 | mdrobnak | Ok, so original is: |
| 20:51.45 | cr2 | this is a very misleading #define |
| 20:51.56 | mdrobnak | #define MSM_FB_BASE MSM_PMEM_GPU1_BASE + MSM_PMEM_GPU1_SIZE |
| 20:51.56 | mdrobnak | #define MSM_FB_SIZE 0x200000 /* 2mb */ |
| 20:52.22 | mdrobnak | Which is what is is currently.. |
| 20:52.41 | cr2 | mdrobnak: where is GPU1 ? |
| 20:52.45 | mdrobnak | I changed #define MSM_PMEM_MDP_BASE MSM_LINUX_BASE + MSM_LINUX_SIZE to 0x23D00000 |
| 20:53.01 | tmzt | what is LINUX size? it's not the kernel |
| 20:53.24 | tmzt | actually, it might be helpful if we used a macro here |
| 20:53.25 | cr2 | tmzt: is it used somewhere ? |
| 20:53.27 | tmzt | and used ranges |
| 20:53.44 | tmzt | I think it's just to set up the offset correctly for drivers |
| 20:53.52 | tmzt | so they can find ranges in smem |
| 20:53.52 | cr2 | tmzt: it all comes from g1 code ;) |
| 20:53.58 | mdrobnak | ADSP = 0x23D00000 + 0x800000, GPU1 = + 0x800000, and FB is +0x800000 |
| 20:54.16 | cr2 | 2d+8= ? |
| 20:54.19 | cr2 | 3d+8= ? |
| 20:55.00 | tmzt | see, I could read that :) |
| 20:55.08 | mdrobnak | ADSP = 0x24500000 |
| 20:55.22 | mdrobnak | GPU1 = 24D00000 |
| 20:55.38 | cr2 | awk '{print 0x3d+8+8}' |
| 20:55.39 | cr2 | 77 |
| 20:55.50 | cr2 | why in the middle ? |
| 20:56.01 | mdrobnak | FB = 0x25500000 |
| 20:56.17 | mdrobnak | in the middle of? |
| 20:56.33 | mdrobnak | I had a solid reason for doing it. Can't remember now lol |
| 20:56.43 | mdrobnak | I didn't want to pick the absolute top of the 2nd bank, if I remember. |
| 20:56.44 | cr2 | 0x20000000 - 0x28000000 |
| 20:57.07 | cr2 | mdrobnak: why not start at 0x20000000 ? |
| 20:57.21 | mdrobnak | I wanted to start from the top and work backward. |
| 20:57.34 | mdrobnak | So that I could try to add more memory to Linux if we figured out what was not right. |
| 20:57.54 | cr2 | the second bank is 100% ok for linux |
| 20:58.07 | cr2 | !android though |
| 20:58.15 | leviathan | Gnutoo: alsa in linuxtogo-kernel would be nice |
| 20:58.17 | mdrobnak | True. |
| 20:58.26 | Gnutoo | but it's buggy |
| 20:58.33 | cr2 | mdrobnak: what about GPU0 ? |
| 20:58.42 | mdrobnak | GPU0 is where it was. Didn't move that. |
| 20:58.48 | cr2 | ok |
| 20:58.50 | Gnutoo | I'll do a clean thing for replicant |
| 20:58.58 | Gnutoo | it'll take a min |
| 20:58.59 | leviathan | ok |
| 20:59.00 | mdrobnak | I only moved anything PMEM related. |
| 20:59.14 | leviathan | just tell me, when you have finished |
| 20:59.17 | mdrobnak | (and the other stuff as a side effect) |
| 21:00.04 | mdrobnak | cr2: Oh, right, what did you want to see from the WinCE dmesg? |
| 21:00.52 | leviathan | I've created an compat-wireless-msmwifi_2.6.bb |
| 21:00.55 | cr2 | mdrobnak: nothing. i only wanted to tell you that the 114th MB is perfectly usable |
| 21:01.20 | leviathan | and added msm_wifi.patch to URI for htcdream |
| 21:01.21 | cr2 | so you should be able to boot with 115MB in the first EBI1 bank |
| 21:01.42 | leviathan | I'll send the patch later to the ML |
| 21:01.59 | mdrobnak | [K] MainMemoryEndAddress = 0x85A00000 |
| 21:02.00 | mdrobnak | [K] PmemHeapPhyAddr = 0x00200000 |
| 21:02.00 | mdrobnak | [K] dwPmemHeapSize = 0x00700000 |
| 21:02.01 | mdrobnak | Interesting. |
| 21:02.34 | cr2 | 5a=90MB |
| 21:02.46 | cr2 | but it's only for wince |
| 21:02.47 | tmzt | oh |
| 21:02.57 | cr2 | on topa100 they use only 81MB for wince. |
| 21:03.38 | cr2 | pmem -> 2+7=9MB |
| 21:04.23 | cr2 | in SMI |
| 21:04.45 | cr2 | 512K is locked for appsbl aka spl |
| 21:04.52 | cr2 | can be unlocked though |
| 21:05.25 | cr2 | but 512K -> 2MB is not used on wince |
| 21:06.00 | cr2 | dzo put ramconsole at 0xe0000, but afair he moved it to 8MB now (on diamond) |
| 21:07.24 | cr2 | mdrobnak: does linux boot with 115MB in the first EBI1 bank ? |
| 21:08.06 | dzo | hi cr2 and mdrobnak |
| 21:08.53 | cr2 | hi dzo |
| 21:15.29 | *** join/#htc-linux tuxhero (n=tuxhero@122.169.169.172) |
| 21:15.47 | tuxhero | cr2 hi |
| 21:19.03 | *** join/#htc-linux SOG (n=SOG@n058152129154.netvigator.com) |
| 21:21.34 | *** join/#htc-linux ptl_ (n=patola@201-68-170-252.dsl.telesp.net.br) |
| 21:21.58 | *** join/#htc-linux FR^2 (i=frr@frquadrat.de) |
| 21:25.44 | mdrobnak | cr2: Sorry, was afk |
| 21:26.19 | mdrobnak | Hi dzo. |
| 21:27.43 | dzo | mdrobnak: hi, does data stay connected for you, it drops sometimes and won't return (on edge). |
| 21:27.58 | mdrobnak | dzo: Correct. It's a DNS thing. |
| 21:28.12 | mdrobnak | I was going to try your rootfs to fix it hehehe |
| 21:28.31 | mdrobnak | You have ip-up working, right? |
| 21:28.41 | dzo | ok, i thought so, the arrows should flash with traffic too, probably the same problem. |
| 21:28.51 | dzo | yes, ip-up works. |
| 21:29.40 | mdrobnak | Well, here's the thing |
| 21:29.43 | mdrobnak | It *gets* a DNS value. |
| 21:29.47 | mdrobnak | of 10.11.12.13 |
| 21:29.47 | mdrobnak | :-( |
| 21:30.03 | mdrobnak | Thats what happens to me anyway |
| 21:30.06 | tmzt | you could use rewrite |
| 21:30.11 | mdrobnak | Exactly. |
| 21:30.11 | dzo | yes, mine too. initially it's correct then gets it wrong. |
| 21:30.14 | tmzt | but fixing it properly is much better |
| 21:30.25 | mdrobnak | Not sure how to. |
| 21:30.42 | mdrobnak | But we could check to see if it's 10.11.12.23 and replace it with opendns in that case |
| 21:30.53 | tmzt | where? |
| 21:30.59 | mdrobnak | in /etc/ppp/ip-up |
| 21:31.09 | mdrobnak | cr2: O |
| 21:31.14 | mdrobnak | cr2: I'll try a build in 2 minutes. |
| 21:31.42 | dzo | also, does your usb reconnect after power collapse? |
| 21:32.41 | mdrobnak | no idea. I don't think I have full power collapse. |
| 21:34.06 | dzo | my remote-ip is 10.64.64.64, i think that's wrong too. |
| 21:34.27 | mdrobnak | That's fine. |
| 21:34.38 | mdrobnak | It just stuffs it along the interface |
| 21:34.41 | mdrobnak | it works :-) |
| 21:35.31 | mdrobnak | Lots of warnings doing this compile... |
| 21:35.46 | mdrobnak | I'm at Oct 12 GIT |
| 21:35.49 | dzo | do you know what the net.dnschange property does? |
| 21:35.51 | mdrobnak | Should I update to latest? |
| 21:35.59 | mdrobnak | dzo: Makes Android re-read the server list. |
| 21:38.07 | mdrobnak | Ok, trying to boot with 115MB |
| 21:38.46 | mdrobnak | Looks good. |
| 21:39.02 | tmzt | on raph110? |
| 21:39.13 | mdrobnak | Yeah |
| 21:39.17 | mdrobnak | I've got a cool Boot logo too hehe |
| 21:39.43 | mdrobnak | And we're at the home screen. |
| 21:40.00 | mdrobnak | cr2: Thanks for 8MBs more ;-) |
| 21:40.40 | mdrobnak | 0-17 17:40:17.756 W/pppd ( 372): Could not determine remote IP address: defaulting to I |
| 21:40.41 | mdrobnak | 10-17 17:40:17.756 I/pppd ( 372): local IP address I |
| 21:40.41 | mdrobnak | 10-17 17:40:17.766 I/pppd ( 372): remote IP address I |
| 21:40.41 | mdrobnak | 10-17 17:40:17.766 I/pppd ( 372): primary DNS address I |
| 21:40.41 | mdrobnak | 10-17 17:40:17.766 I/pppd ( 372): secondary DNS address I |
| 21:40.41 | mdrobnak | 10-17 17:40:17.906 D/MobileDataStateTracker( 120): CONNECTED event did not supply interface name. |
| 21:40.43 | mdrobnak | 10-17 17:40:17.906 D/MobileDataStateTracker( 120): DNS server addresses are not known. |
| 21:40.47 | mdrobnak | Stupid stuff happens on 3G |
| 21:41.07 | mdrobnak | root@mdrobnak-desktop:~/android-repo/android-sdk-linux_x86-1.5_r3/tools# ./adb shell |
| 21:41.07 | mdrobnak | # ping www.google.com |
| 21:41.07 | mdrobnak | PING google.navigation.opendns.com (208.69.36.230) 56(84) bytes of data. |
| 21:41.08 | mdrobnak | 64 bytes from 208.69.36.230: icmp_seq=1 ttl=49 time=145 ms |
| 21:41.10 | mdrobnak | But it works fine :-) |
| 21:41.29 | cr2 | mdrobnak: lol. the nadroid ril code is so primitive ;) |
| 21:41.40 | cr2 | mdrobnak: i suggest to create a wiki table: |
| 21:42.20 | cr2 | AT, wince RIL, htc android RIL, android RIL |
| 21:42.48 | Gnutoo | leviathan, http://gitorious.org/replicant/gnulinuxkernel/commit/92945ccd0921e940fc2675dc394141db0f7c1386 |
| 21:43.04 | mdrobnak | I don't understand. How can we know what's in the G1 one without disassembling it? |
| 21:43.08 | mdrobnak | It does NOT echo AT commands. |
| 21:43.10 | tmzt | use ppp0 for interface |
| 21:43.13 | tmzt | this makes sense |
| 21:43.30 | tmzt | mdrobnak: add somethign to kernel to dump read/write on smd0 |
| 21:43.39 | mdrobnak | That'll work. |
| 21:43.49 | tmzt | you need to know it's rpc format though |
| 21:43.54 | mdrobnak | Oh, right. |
| 21:44.00 | tmzt | since rild communicates with the rest of android over a custom rpc |
| 21:44.01 | cr2 | mdrobnak: it's called "strings" and objdump |
| 21:44.01 | mdrobnak | That's pretty useless to us then? |
| 21:44.02 | *** join/#htc-linux stickboy (n=anonymou@128.153.211.150) |
| 21:44.10 | mdrobnak | cr2: -P |
| 21:44.14 | mdrobnak | :-P |
| 21:44.24 | tmzt | the interface it brings up must be encoded in an event some how |
| 21:44.25 | tmzt | well |
| 21:44.31 | tmzt | you could proxy the socket as well |
| 21:44.50 | tmzt | you also have rild source |
| 21:44.55 | cr2 | tmzt: lol, gsmd :) |
| 21:44.57 | mdrobnak | tmzt: But if you're spying on smd0, who cares. You can get a timed log from Android. Correlate the events. |
| 21:45.29 | cr2 | wtf is 76XXD-6125-DSDCAAPZM |
| 21:46.00 | tmzt | cr2: ? |
| 21:46.10 | mdrobnak | 6125 = ADSP version? |
| 21:46.12 | tmzt | that looks like a amss version string |
| 21:46.18 | cr2 | mdrobnak: amss |
| 21:46.19 | tmzt | oh |
| 21:46.20 | tmzt | from topa? |
| 21:46.25 | cr2 | tmzt: yes |
| 21:46.30 | mdrobnak | That's what I meant |
| 21:46.30 | tmzt | so it's not 6150 |
| 21:46.36 | cr2 | no |
| 21:46.41 | cr2 | it's gsm |
| 21:46.43 | tmzt | but it's msm7600 |
| 21:46.50 | tmzt | no it's not apparently |
| 21:46.56 | cr2 | 7201A |
| 21:47.02 | tmzt | Jason8: did you say they all use 7600? |
| 21:47.02 | tmzt | oh |
| 21:47.06 | mdrobnak | Wait, which one is the Topaz? |
| 21:47.11 | tmzt | what is the 76XXD for then? |
| 21:47.16 | cr2 | no idea |
| 21:47.26 | Jason8 | It wasn't me, but someone else said it. |
| 21:47.31 | cr2 | it's some SMD item |
| 21:47.41 | mdrobnak | Because I know for Sprint, the Touch Pro 2 is both CDMA and Quad Band GSM |
| 21:48.12 | Epsylon3 | hmmm found something.... with boot loader terminal, when i type "Task 38" (that show something like an Windows CE init log) |
| 21:48.17 | Epsylon3 | B4WaitARM9DexReady PV_DRIVERGLOBAL->fixAddress.bIsBatteryReplaceB4=0xffffffff |
| 21:48.25 | tmzt | Jason8: he said for at&t also |
| 21:48.30 | Jason8 | yep |
| 21:48.32 | Epsylon3 | should be 0 |
| 21:48.42 | Jason8 | I think it was trip that said the pure was msm7600 |
| 21:48.47 | tmzt | cr2: that string is also in the device info |
| 21:49.08 | tmzt | cr2: how do we know it's msm7201a? |
| 21:49.10 | tmzt | actually, it's not |
| 21:49.14 | tmzt | it's msm7200a |
| 21:49.22 | cr2 | tmzt: it's the same. |
| 21:49.25 | Jason8 | GSMArena says 7200a |
| 21:49.39 | cr2 | 7201A is a politically correct name |
| 21:49.55 | cr2 | they are identical i think |
| 21:50.26 | tmzt | not usb at least |
| 21:50.34 | tmzt | well, that was 7200 non-A |
| 21:50.52 | tmzt | we can always look at trout/sapp kernel differences |
| 21:50.53 | cr2 | 7200A is 7201A |
| 21:50.55 | tmzt | which are minimal |
| 21:51.40 | *** join/#htc-linux Abracadabr4 (n=Abracada@92.24.47.47) |
| 21:54.35 | phh | has anyone tried running Xegl against android's libgl ? |
| 21:54.52 | tmzt | not that I know of |
| 21:55.03 | cr2 | leviathan: can we fix /dev/mem on g1 ? i'd like to see a full smem dump from g1 |
| 21:55.21 | leviathan | cr2: sure we can |
| 21:55.38 | leviathan | but first I have to finish hacking around in userspace |
| 21:55.52 | cr2 | ok |
| 21:56.03 | leviathan | atm: Im trying to get compat-wireless building as bitbake-package |
| 21:56.18 | *** join/#htc-linux dcordes__ (n=dcordes_@unaffiliated/dcordes) |
| 21:56.41 | leviathan | And I'm adding gnutoos-patch to the depencies for the linux-msm7xxxx-kernel |
| 21:56.42 | Gnutoo | leviathan, how hard is /dev/mem fix? |
| 21:56.45 | leviathan | in OE |
| 21:56.50 | leviathan | hmm |
| 21:57.14 | Gnutoo | cr2, is it because of /dev/mem problem that gdb bt makes the phone reboot? |
| 21:58.13 | leviathan | hmm |
| 21:58.29 | leviathan | /dev/mem is optimized to be used by the android tools |
| 21:58.47 | leviathan | to inject graphics and initialize sound, etc. |
| 21:59.09 | Gnutoo | ok |
| 21:59.13 | tmzt | pmem is |
| 21:59.18 | leviathan | so a lot of things you should not be allowed to do on an ordinary /dev/mem interface |
| 21:59.40 | leviathan | pmem? |
| 21:59.47 | leviathan | mom |
| 21:59.54 | cr2 | leviathan: lol, "optimized" |
| 22:00.17 | cr2 | leviathan: embraced and extended |
| 22:00.30 | mdrobnak | cr2: I will agree, they did some stupid stuff. |
| 22:00.31 | Gnutoo | lol |
| 22:00.32 | leviathan | cr2: yes, that would be the correct term |
| 22:00.34 | leviathan | ^^ |
| 22:00.56 | mdrobnak | The key map is ... annoying. |
| 22:00.56 | leviathan | yeahh |
| 22:00.58 | Gnutoo | mdrobnak, can the keyboard be remapped? |
| 22:01.25 | leviathan | mdrobnak: would be nice, if you could do this |
| 22:01.31 | leviathan | I tried but gave up |
| 22:01.35 | mdrobnak | Yeah, trying. |
| 22:01.36 | leviathan | its tooo ennoying for me |
| 22:01.37 | mdrobnak | it's a pain |
| 22:01.46 | mdrobnak | I got it in console/X |
| 22:01.46 | leviathan | full agreement |
| 22:01.48 | mdrobnak | but not And |
| 22:01.51 | leviathan | its pita |
| 22:01.51 | Gnutoo | because android remaps it |
| 22:01.56 | leviathan | pain in the a* |
| 22:02.05 | Gnutoo | in the init.something scripts that are in / |
| 22:02.06 | leviathan | PITA |
| 22:02.08 | leviathan | :) |
| 22:02.12 | Gnutoo | if I remember well |
| 22:02.17 | leviathan | yeahh |
| 22:02.31 | leviathan | its binary f* s* tinkering |
| 22:02.45 | leviathan | but not beautiful hacking tinkering |
| 22:02.51 | leviathan | more the ugly and dirty kind |
| 22:03.03 | mdrobnak | Actually |
| 22:03.05 | mdrobnak | it's quite easy |
| 22:03.07 | leviathan | make cat and you can type reset in the keymap |
| 22:03.09 | mdrobnak | you set a property value |
| 22:03.16 | mdrobnak | to the bin file |
| 22:03.17 | mdrobnak | but |
| 22:03.24 | mdrobnak | getting that bin file right is a pain |
| 22:03.25 | mdrobnak | brb |
| 22:03.31 | leviathan | ok |
| 22:04.21 | Gnutoo | mdrobnak, how did you get it in X? |
| 22:04.23 | Gnutoo | xmodmap? |
| 22:05.35 | cr2 | 9*4=36 |
| 22:05.39 | cr2 | 8*4=32 |
| 22:06.49 | tmzt | I will be back, I guess your going to bed soon cr2? |
| 22:07.43 | cr2 | tmzt: yes |
| 22:08.48 | cr2 | 8 vs. 9 |
| 22:12.39 | cr2 | <PROTECTED> |
| 22:12.40 | cr2 | <PROTECTED> |
| 22:12.41 | cr2 | <PROTECTED> |
| 22:12.43 | cr2 | <PROTECTED> |
| 22:12.47 | cr2 | 20+3*4=32 |
| 22:13.47 | tmzt | what's this? |
| 22:15.48 | cr2 | struct smd_alloc_elm { |
| 22:16.16 | cr2 | seems to have one more unt32_t on amss612x ;) |
| 22:16.46 | tmzt | see why we need ops? |
| 22:16.48 | cr2 | between cid and ref_count |
| 22:17.03 | cr2 | #ifdef for now :) |
| 22:17.24 | tmzt | what do you think is there? |
| 22:18.02 | cr2 | unused |
| 22:18.10 | cr2 | it's not on samsung |
| 22:18.37 | tmzt | what is ctype? |
| 22:18.48 | cr2 | not on sapp |
| 22:19.05 | cr2 | should be stream vs. packet ? |
| 22:19.15 | cr2 | but not used. |
| 22:19.15 | tmzt | what values |
| 22:19.27 | cr2 | 0 |
| 22:19.38 | cr2 | i've seen in in some code. |
| 22:20.12 | tmzt | ok |
| 22:20.20 | tmzt | so just a place holder for now then? |
| 22:20.38 | cr2 | tmzt: there was a seriuos (i think) rpc bug fixed two days ago in codeaurora code. |
| 22:20.40 | cr2 | tmzt: yes |
| 22:20.52 | tmzt | ok |
| 22:21.05 | cr2 | * 'type' field of smd_alloc_elm structure |
| 22:21.07 | cr2 | <PROTECTED> |
| 22:21.08 | cr2 | <PROTECTED> |
| 22:21.10 | cr2 | <PROTECTED> |
| 22:21.11 | cr2 | <PROTECTED> |
| 22:21.27 | tmzt | right |
| 22:21.30 | tmzt | so we never use packet? |
| 22:22.08 | cr2 | no |
| 22:22.15 | tmzt | need to go |
| 22:22.20 | tmzt | goodnight |
| 22:22.37 | cr2 | #define SMD_CHANNEL_TYPE(x) ((x) & 0x000000FF) |
| 22:22.50 | cr2 | #define SMD_XFER_TYPE(x) (((x) & 0x00000F00) >> 8) |
| 22:22.50 | cr2 | good night |
| 22:24.57 | mdrobnak | Gnutoo: Set the default keymap |
| 22:25.36 | Gnutoo | mdrobnak, ok but with xmodmap? |
| 22:25.43 | Gnutoo | with which application |
| 22:25.56 | mdrobnak | General X should use the default console stuff, right? |
| 22:26.11 | mdrobnak | Otherwise that's highly annoying |
| 22:26.22 | mdrobnak | I was testing Mer |
| 22:26.28 | mdrobnak | but I dont remember if I tried it in X |
| 22:26.43 | Gnutoo | mdrobnak, ah ok with console...I'll try...what was the command to set the console keymap? |
| 22:26.43 | mdrobnak | Oh. |
| 22:26.49 | mdrobnak | loadkey? |
| 22:26.51 | Gnutoo | thanks |
| 22:27.04 | mdrobnak | /etc/console-setup/boottime.gz or something |
| 22:28.11 | Gnutoo | ok |
| 22:29.08 | Gnutoo | mdrobnak, but did you get android keys mapping under mer...that is the alt key thing |
| 22:29.13 | Gnutoo | like the - |
| 22:29.23 | *** join/#htc-linux BabelO (n=fcr@lun34-2-82-238-28-28.fbx.proxad.net) |
| 22:29.25 | Gnutoo | the / |
| 22:29.26 | Gnutoo | etc... |
| 22:29.41 | Gnutoo | because I've qwerty under X |
| 22:29.46 | Gnutoo | s/X/kdrive |
| 22:29.55 | Gnutoo | but...that's not great without the - |
| 22:31.13 | mdrobnak | I mapped Fn to Alt |
| 22:31.19 | mdrobnak | and so that worked for me |
| 22:31.52 | mdrobnak | the problem is Android's keys are wrong... |
| 22:31.52 | leviathan | goes sleeping now |
| 22:32.09 | mdrobnak | Good night leviathan |
| 22:32.16 | mdrobnak | I love Android messages: 10-17 18:31:40.396 I/NotificationService( 120): enqueueToast pkg=com.dataviz.stargate callback=android.app.ITransientNotification$Stub$Proxy@43430f58 duration=0 |
| 22:32.21 | leviathan | mdrobnak: thanks, same to you =) |
| 22:32.28 | mdrobnak | enqueueToast? lol |
| 22:32.49 | Gnutoo | ok |
| 22:32.53 | Gnutoo | leviathan, good night |
| 22:32.55 | mdrobnak | hmm.. |
| 22:33.02 | mdrobnak | Can't get it to load my stupid map |
| 22:33.03 | mdrobnak | argh |
| 22:36.21 | cr2 | at@agpsfeature=0 |
| 22:36.30 | mdrobnak | yeah I saw that cr2 |
| 22:36.33 | mdrobnak | does that help you? |
| 22:36.41 | mdrobnak | how is gps coming? |
| 22:38.08 | cr2 | not yet |
| 22:39.57 | mdrobnak | location via GSM isn't fun, as I discovered |
| 22:39.58 | mdrobnak | hehe |
| 22:40.15 | mdrobnak | Ie, the RIL code needs quite a bit of work there. It wasn't really even done. |
| 22:40.40 | cr2 | mdrobnak: do you have fieldtest.exe ? |
| 22:40.56 | mdrobnak | in windows? Not sure. |
| 22:43.23 | cr2 | good night |
| 22:43.56 | mdrobnak | good night |
| 22:45.19 | mdrobnak | needs to rework system.img to load keymaps from rootfs.img |
| 22:45.26 | mdrobnak | dinner time |
| 22:58.01 | dcordes__ | Gnutoo, rats.. I have no luck with OE these days |
| 22:58.20 | dcordes__ | Gnutoo, do you have dream rootfs somewhere? |
| 22:58.21 | dcordes__ | shr |
| 23:48.39 | *** join/#htc-linux Abracadabr4 (n=Abracada@92.24.47.47) |