| 00:00.28 | cr2 | druidu: is it better to have an ifdef there ? |
| 00:00.34 | Bally3 | aah.. no.. still working on polaris wifi as yet.. and reading up about setting up the sources |
| 00:00.47 | tmzt | so what branch are you working on now? |
| 00:01.08 | cr2 | druidu: the most smart was will be to trap the proc_comm, and do the right thing for wince (if amss=wince) |
| 00:02.32 | tmzt | there's a proc_comm api right? |
| 00:02.47 | tmzt | there should be an ops table |
| 00:03.00 | cr2 | int msm_proc_comm(unsigned cmd, unsigned *data1, unsigned *data2) |
| 00:03.32 | cr2 | and it hides gpio config, vreg & DEX, and clocks+rpc |
| 00:03.38 | tmzt | otherwise it will be hard to support trout as well |
| 00:05.40 | cr2 | tmzt: is the gpio alt config protected by MPU on g1/g2 ? |
| 00:05.56 | cr2 | tmzt: if not, then our code may be used on g1/g2 too |
| 00:06.17 | tmzt | I don't know, I guess we could enable it though |
| 00:06.33 | cr2 | you mean disable |
| 00:06.37 | tmzt | yeah |
| 00:06.39 | tmzt | right |
| 00:06.53 | tmzt | who was building openmoko-msm again? |
| 00:08.23 | cr2 | tmzt: there is a nice rpc call clk_regime_sec_msm_get_clk_freq_khz() |
| 00:08.47 | cr2 | but it has some clock handle as a parameter |
| 00:09.16 | tmzt | nice |
| 00:09.30 | tmzt | so there's a copy_smem_user or something? |
| 00:09.37 | tmzt | what kind of pointers are these? |
| 00:14.12 | cr2 | i'vedited clk_regime_msm AP here |
| 00:14.18 | cr2 | http://www.htc-linux.org/wiki/index.php?title=MSM_CLK |
| 00:14.28 | *** join/#htc-linux Shinto (n=John@f049070072.adsl.alicedsl.de) |
| 00:15.08 | druidu | wtf is clamshell anyway? |
| 00:17.04 | tmzt | he means the slide on the keyboard |
| 00:17.52 | druidu | microp-ksc: rewritten kovsky clamshell support -- this still breaks the build if compiling without CONFIG_MACH_HTCKOVSKY |
| 00:17.57 | cr2 | btw, why does clock-wince.c set pmdh clock to 0xa00 when it's 0xa19 for rap100 and 0xa41 for raph800 ? |
| 00:18.11 | cr2 | and g1 has a different value too |
| 00:18.33 | druidu | no idea :D |
| 00:18.39 | cr2 | CONFIG_MACH_HTCKOVSKY is evil |
| 00:19.05 | cr2 | druidu: sometimes i'm wondering why the LCD works at all :) |
| 00:19.19 | cr2 | most things there are setup differently in wince. |
| 00:19.30 | tmzt | does the clock really matter? |
| 00:24.28 | druidu | cr2: simple, we just write stuff in the buffer and do a dma transfer, everything else is just for transition to the android firmware :) |
| 00:25.09 | cr2 | tmzt: the mddi speed |
| 00:25.43 | tmzt | yeah, but if it being wrong doesn't affect the screen how could it be used? |
| 00:25.55 | tmzt | or it's just close enough? |
| 00:25.58 | druidu | anybody know address at which to push git changes? |
| 00:26.02 | cr2 | druidu: we also push mddi speed setting and toshiba config. even if the mddi client is epson. and it still works :) |
| 00:26.53 | druidu | hey, that's what I call robust design... works even if you push bullshit |
| 00:27.23 | cr2 | it may be interesting to cal clk_regime_sec_msm_get_clk_freq_khz(0x38) on wince |
| 00:27.33 | cr2 | s/cal /call / |
| 00:27.53 | cr2 | and (probably) get the mddi clock speed |
| 00:27.58 | tmzt | we could list the proper clock settings? |
| 00:29.03 | cr2 | (at least) only for the clocks with regime idx, and those with known clock algorithm (BT and SD/wifi) |
| 00:29.46 | cr2 | i2c is 0x2f , for example |
| 00:30.02 | tmzt | there's clk_regime_sec, is there a non-sec |
| 00:30.03 | cr2 | i'm wondering if it'll report the ~100 or ~400kHz clock |
| 00:30.09 | tmzt | A clocks |
| 00:30.11 | cr2 | most are sec |
| 00:30.46 | cr2 | it can be an interesting rpc test driver too |
| 00:30.56 | cr2 | or calling it inside haret. |
| 00:31.50 | cr2 | on linux it's an rpc call 0x3000000f ,0, 0x25 |
| 00:32.08 | cr2 | and for wince it's possible to just use the clock API |
| 00:36.04 | cr2 | ali1234: about mdoc_g34.c the g3/g4 should be a module parameter probably. and g4 may have a different register location |
| 00:36.13 | druidu | shit, I pushed the wrong changes |
| 00:36.18 | druidu | how can I undo a git push? :D |
| 00:36.33 | tmzt | where did you push it? |
| 00:36.35 | tmzt | how recently? |
| 00:41.27 | cr2 | ali1234: http://belgarath.falguerolles.org/download/mio_a701/docs/disassembly/RomBootStrap.S |
| 00:41.30 | druidu | right now |
| 00:41.37 | druidu | http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=shortlog;h=refs/heads/htc-msm-2.6.27 |
| 00:42.37 | druidu | well, I was going to push it anyway but didn't get to test it so there will be some more changes |
| 00:42.50 | druidu | any ideas? |
| 00:44.24 | druidu | so I want to reset HEAD to abfb31f3e1d5e8096eb6668401963d1ec5212889 |
| 00:48.17 | *** join/#htc-linux mrmoku|a` (n=mrmoku@ppp-93-104-61-155.dynamic.mnet-online.de) |
| 00:49.01 | cr2 | Sapphire (MACH_SAPPHIRE) [Y/n] (NEW) n |
| 00:49.22 | cr2 | <PROTECTED> |
| 00:50.37 | cr2 | arch/arm/mach-msm/qdsp5/adsp.c: In function âmsm_adsp_probeâ: |
| 00:50.38 | cr2 | arch/arm/mach-msm/qdsp5/adsp.c:805: warning: assignment makes integer from pointer without a cast |
| 00:50.40 | cr2 | arch/arm/mach-msm/qdsp5/adsp.c:806: warning: assignment makes integer from pointer without a cast |
| 00:50.41 | cr2 | arch/arm/mach-msm/qdsp5/adsp.c:807: warning: assignment makes integer from pointer without a cast |
| 00:50.47 | marex | cr2: Support for OpenBSD (OPENBSD) [simply Y] ;-) |
| 00:51.43 | cr2 | marex: linux is the one true operating system :-) |
| 00:52.00 | marex | cr2: try telling that to the mighty OpenBSD $DEITYs ;-) |
| 00:52.25 | marex | now that'd be one hell of a rough flamewar ;-) |
| 00:52.28 | druidu | wohooo, managed to undo the commit :P |
| 00:52.33 | druidu | -f (force) rulez |
| 00:53.24 | marex | cr2: it's an university project ... and it allows me to build my own little empire (without RMS anywhere close) in the OpenBSD kernel ;-) |
| 00:53.59 | tmzt | marex: what device? |
| 00:54.14 | marex | tmzt: palms ... |
| 00:54.24 | marex | tx t5 ld z72 |
| 00:54.39 | cr2 | marex: lol. i don't mind switching to a gplv3 licensed os |
| 00:54.57 | cr2 | drivers/usb/function/msm_hsusb.c: In function âusb_find_descriptorâ: |
| 00:54.59 | cr2 | drivers/usb/function/msm_hsusb.c:1755: warning: large integer implicitly truncated to unsigned type |
| 00:55.05 | marex | cr2: seen windows7sins already ? |
| 00:55.12 | marex | cr2: Im starting to be suspicious of RMS |
| 00:55.24 | marex | or rather of his overall mental health |
| 00:55.40 | druidu | RMS? |
| 00:55.43 | ali1234 | anybody know about qualcomm pistachio? |
| 00:55.45 | cr2 | $windows is filtered out by my brain automatically |
| 00:56.07 | cr2 | ali1234: it's an L4 based microkernel |
| 00:56.11 | ali1234 | druidu: RMS is overall arm linux maintainer |
| 00:56.21 | marex | ali1234: bullcrap ... RMK is |
| 00:56.29 | ali1234 | oops |
| 00:56.30 | marex | RMS is the GPL fatman ;-) |
| 00:56.31 | ali1234 | yeah |
| 00:56.58 | ali1234 | damn TLAs |
| 00:56.59 | marex | I dont mind RMK ... even though in the OpenBSD kernel Im the RMK to some extent ;-) |
| 00:57.06 | marex | (it really does feel great :-) ) |
| 00:57.23 | ali1234 | well RMS, hasn't everyone know he's crazy for years? |
| 00:57.44 | marex | ali1234: yeah, but they are afraid to put him under lock into some madhouse ;-) |
| 00:58.06 | marex | possibly the psychiatrists are scared they'd get infected or something ... ;-) |
| 00:58.16 | druidu | cr2: any descriptions for your patches? |
| 00:59.17 | marex | cr2: btw. it's good to know more OS internals besides linux ;-) |
| 00:59.36 | cr2 | adsp irq and parameters, bt clocks and halt implementation |
| 00:59.49 | druidu | 2nd one is power off support right? |
| 00:59.51 | cr2 | marex: make n560 resume, then i'll listen to you :) |
| 00:59.57 | druidu | ah, halt |
| 01:00.05 | cr2 | halt=poweroff |
| 01:00.18 | marex | cr2: you're getting me heated up ... I might end up porting openbsd to it too ;-) |
| 01:01.17 | cr2 | BT clocks are derived from PLL0, while SD clocks are from PLL1 and TCX0 |
| 01:01.31 | cr2 | marex: make resume work in any os. |
| 01:01.40 | cr2 | wince != os |
| 01:01.54 | marex | cr2: hey that's cheating |
| 01:02.35 | marex | cr2: Ballmer calls it an OS and this time I'll take his word for it ... so go ahead, finish sound on N560 |
| 01:03.22 | cr2 | marex: it's of little use without resume. talk to kiozen |
| 01:04.09 | cr2 | druidu: but i need to write some demonstrating code and comments for clock-wince.c about it |
| 01:04.18 | marex | cr2: it's certainly very useful ... dont try to talk your way out of it ;-) |
| 01:04.56 | cr2 | marex: ok, i promise you to write the sound driver, once the resume will work. |
| 01:05.30 | marex | cr2: it works in wince |
| 01:06.55 | marex | cr2: I'll look into the resume once I have some free time ... but fttb it's x86 assembler till 8th sept ... then I have to design some serial-GPIB bus converter ... then work on the OpenBSD some more (yay!) ... so possibly sometimes after those are sorted out |
| 01:07.02 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
| 01:07.27 | marex | btw the asoc audio is a piece of cake and you know it :/ |
| 01:07.39 | tmzt | what are those for? |
| 01:09.19 | cr2 | GPIB is not difficult |
| 01:09.54 | cr2 | druidu: how do i roll back your commit ? |
| 01:11.42 | *** join/#htc-linux rafyvitto (n=rafyvitt@24.54.253.34) |
| 01:11.58 | marex | tmzt: exam from x86 asm (including the lovely FPU, yummy MMX and tasty SSE in all it's beauty) I have on 8th -- and then some voluntary work at uni (GPIB) -- and the OpenBSD port is an university project (and I like it anyway) |
| 01:12.13 | marex | tmzt: that's just cr2 who doesnt understand what a real UNIX should be like |
| 01:13.20 | cr2 | marex: x86 is the most evil cpu i know |
| 01:14.05 | *** join/#htc-linux marajin_ (n=marajin@87-194-102-189.bethere.co.uk) |
| 01:14.20 | rafyvitto | hey guys |
| 01:14.57 | marex | cr2: I can agree with that more than fully ... the x86 assembler makes me want to cry and vomit |
| 01:15.13 | marex | rafyvitto: hello |
| 01:15.32 | rafyvitto | did you guys ever read about this http://forum.xda-developers.com/showthread.php?t=388273&highlight=linux |
| 01:15.59 | rafyvitto | its a dual boot linux/windows mobile 6.1 for the typhoon |
| 01:16.42 | rafyvitto | some chinese guys made it happen |
| 01:19.09 | rafyvitto | is this usefull in anyway for the raph android port? |
| 01:20.01 | tmzt | 3.SD card (linux mirror and wince XIP boot from SD card) |
| 01:20.03 | tmzt | nice |
| 01:20.35 | rafyvitto | is this method better than haret? |
| 01:21.11 | tmzt | yeah, if it works |
| 01:21.18 | tmzt | the console sounds interesting, what is it? |
| 01:21.34 | rafyvitto | i dont have a clue |
| 01:21.53 | rafyvitto | i have always wonder how does chinese dudes made it happen |
| 01:22.23 | tmzt | we could use source |
| 01:22.28 | rafyvitto | http://forum.xda-developers.com/showthread.php?t=388231 here are some screenshots |
| 01:22.30 | tmzt | for the bootloader |
| 01:23.05 | rafyvitto | someone would have to contact the project leader or main dev |
| 01:23.13 | rafyvitto | and see if he could share the source |
| 01:23.32 | rafyvitto | problem is, hes chinese >< |
| 01:26.22 | druidu | cr2: pushed your patches: http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=shortlog;h=refs/heads/htc-msm-2.6.27 |
| 01:29.45 | tmzt | rafyvitto: you in #xda-devs? |
| 01:40.14 | *** join/#htc-linux stickboy (n=anonymou@128.153.211.117) |
| 01:48.30 | rafyvitto | yes |
| 01:53.15 | rafyvitto | wait if your asking if im that channel no xD, i though you meant the forum it self |
| 01:53.22 | rafyvitto | in* |
| 01:53.25 | *** join/#htc-linux miknix_ghost (n=miknix@bl4-81-27.dsl.telepac.pt) |
| 01:55.48 | rafyvitto | hey tmzt |
| 02:05.17 | *** join/#htc-linux Rafyvitto (n=rafymerc@24.54.253.34) |
| 02:06.56 | *** join/#htc-linux Rafyvitto (n=rafymerc@24.54.253.34) |
| 02:18.51 | *** part/#htc-linux Rafyvitto (n=rafymerc@24.54.253.34) |
| 02:59.02 | *** join/#htc-linux AFBCamaro (n=Criton@fl-71-52-102-205.dhcp.embarqhsd.net) |
| 03:00.00 | AFBCamaro | . |
| 03:07.14 | *** join/#htc-linux surge (i=surge@pool-98-118-154-21.bflony.fios.verizon.net) |
| 03:17.44 | *** join/#htc-linux jlcox (n=root@5.30.233.220.static.exetel.com.au) |
| 03:50.08 | *** join/#htc-linux Amaranth (n=travis@ubuntu/member/Amaranth) |
| 03:59.41 | *** join/#htc-linux BHSPitLappy (n=BHSPitLa@unaffiliated/bhspitmonkey) |
| 04:00.37 | *** join/#htc-linux MrPippy (n=pip@adsl-75-37-163-177.dsl.sndg02.sbcglobal.net) |
| 04:08.48 | *** join/#htc-linux Moku (n=John@f054169173.adsl.alicedsl.de) |
| 04:25.04 | *** join/#htc-linux droid001 (n=g1@p4FDCEB43.dip.t-dialin.net) |
| 07:24.27 | *** join/#htc-linux ArteK (n=Artur@81.15.241.96) |
| 07:25.43 | *** join/#htc-linux kiozen (n=kiozen@rgnb-5d87c9a0.pool.mediaWays.net) |
| 07:27.28 | *** join/#htc-linux stickboy (n=anonymou@128.153.211.117) |
| 07:28.42 | *** join/#htc-linux niccoswe (n=niccoswe@78-70-202-58-no156.tbcn.telia.com) |
| 07:31.11 | *** join/#htc-linux swc|666 (n=infidel2@unaffiliated/swc666/x-4934821) |
| 07:39.37 | *** join/#htc-linux niccoswe (n=niccoswe@78.70.202.58) |
| 08:02.06 | *** join/#htc-linux niccoswe (n=niccoswe@78-70-202-58-no156.tbcn.telia.com) |
| 08:05.14 | *** part/#htc-linux niccoswe (n=niccoswe@78-70-202-58-no156.tbcn.telia.com) |
| 08:13.08 | *** join/#htc-linux cr2 (n=cr2@ip-77-25-155-176.web.vodafone.de) |
| 08:13.28 | cr2 | Application stopped. Please contact your administrator |
| 08:13.33 | cr2 | eh? |
| 08:28.03 | Perkka | hey guys... |
| 08:28.10 | Perkka | this it totally off topic... |
| 08:28.35 | Perkka | but a nice gift for my gf who goes 22 in a few days... |
| 08:28.45 | Perkka | ideas? :P |
| 08:37.34 | *** join/#htc-linux Echo31 (n=olivier@mir31-4-82-240-194-54.fbx.proxad.net) |
| 08:57.06 | Echo31 | cr2: pH5: hi |
| 09:08.50 | *** join/#htc-linux xperia (n=chatzill@80-218-229-113.dclient.hispeed.ch) |
| 09:09.00 | xperia | hello to all. |
| 09:10.31 | xperia | cr2: i am analysing the microp wiki changes. a lot of the new stuff makes sense for me now. i think there are one or tewo things that are missing. |
| 09:16.23 | xperia | as a example the lines cc,14,<0a,x pointsequence cmd+buf; wr mul and cc,41,3c,x -> LED TBL (hardcoded) ; wr mul |
| 09:18.12 | xperia | P.S. can somebody look why the htc-linux logger from netripper dont work since yesterday ? http://irclog.netripper.com/?chan=htc-linux wihout a log the people will miss a lot of conversation |
| 09:18.28 | *** join/#htc-linux Squarc (n=Squarc@82.217.32.29) |
| 09:18.39 | dzo | hi cr2, i got myself a diamond. does audio work on raph yet? |
| 09:19.14 | druidu | not yet dzo |
| 09:19.31 | dzo | on diamond it doesn't register the m2a rpc and calls the wince ept. |
| 09:19.38 | *** join/#htc-linux StarLite (n=nnscript@s55916ca6.adsl.wanadoo.nl) |
| 09:19.39 | dzo | for the adsp. |
| 09:20.50 | dzo | perhaps something wrong with rpc startup. since wince has already registered the rtosadsp rpc we cant take it over. |
| 09:21.20 | dzo | i might try smsm change state to reinit rpc on the a9. |
| 09:22.03 | dzo | i've got some questions too do you have some time? |
| 09:22.34 | druidu | me? not sure if I can help you, but you can ask :D |
| 09:22.50 | dzo | is there no cpld on these devices? |
| 09:25.40 | druidu | don't know... |
| 09:26.12 | xperia | cr2: in my i2c traces i have this here. |
| 09:26.14 | xperia | cc,14,<0a,x pointsequence cmd+buf; wr mul |
| 09:26.15 | xperia | 003451: mmutrace 7805fd34: e5863000(str) b2300000 000001cc (00000000) |
| 09:26.17 | xperia | 003451: mmutrace 7805fd74: e5863000(str) b2300000 00000014 (00000000) |
| 09:26.18 | xperia | 003451: mmutrace 7805fddc: e5863000(str) b2300000 00000094 (00000000) |
| 09:26.20 | dzo | never mind, i'll come back later and see if cr2 is around. |
| 09:26.20 | xperia | 003523: mmutrace 7805fddc: e5863000(str) b2300000 00000228 (00000000) |
| 09:26.21 | xperia | and this here |
| 09:26.23 | xperia | cc,41,3c,x -> LED TBL (hardcoded) ; wr mul |
| 09:26.24 | xperia | 003751: mmutrace 7805fd34: e5863000(str) b2300000 000001cc (00000000) |
| 09:26.26 | xperia | 003751: mmutrace 7805fd74: e5863000(str) b2300000 00000041 (00000000) |
| 09:26.28 | xperia | 003751: mmutrace 7805fddc: e5863000(str) b2300000 00000000 (00000000) |
| 09:26.29 | xperia | 003808: mmutrace 7805fddc: e5863000(str) b2300000 00000000 (00000000) |
| 09:26.31 | xperia | My first question to you is the 0x3c right for the hardcoded LED TBL ? |
| 09:28.24 | xperia | P.S. the two lines after the 41 lines is not a clearing. i have just posted the first two lines after 0x41 from nearly 40 (str) lines. |
| 09:34.36 | cr2 | dzo: good news :D the sound does not work, but the latest 2.6.27 git should be a good start. |
| 09:35.43 | dzo | yes, i saw that, how far does it get for you, does it start to write to the adsp? |
| 09:36.14 | cr2 | dzo: there is a 3000ffff amss server, which sends data to arm11 even while in linux. we may reregister some dummy client for it. |
| 09:36.52 | dzo | ok, i'll look at that, thanks, got to go now, bye. |
| 09:36.57 | cr2 | it sends the audiomgr rpcs without a problem, but then hangs the machine |
| 09:38.35 | cr2 | druidu: do you have BT in .config ? |
| 09:39.10 | cr2 | xperia: 0x3c is the size |
| 09:40.07 | cr2 | dzo: 7x01A phones have enough gpios, so they don't use cpld |
| 09:40.51 | dzo | except dream and saphire, they do. |
| 09:41.07 | cr2 | dzo: no known wince phones |
| 09:41.25 | dzo | ok, that makes things simpler anyway. |
| 09:41.28 | cr2 | dzo: this is the latest doc collection http://www.htc-linux.org/wiki/index.php?title=Raphael |
| 09:42.03 | dzo | yes, i saw that. has anybody looked at power collapse? |
| 09:42.58 | dzo | and data is on smd7, yes? |
| 09:43.40 | dzo | is the smd_private table correct? |
| 09:47.29 | xperia | cr2: ahh great. the size of the buffer. what i dont understand now is how this Hardcoded Led Table is created. when i compare the source table with the i2c log they are two differnt things. |
| 09:47.29 | xperia | here a litlle example |
| 09:47.29 | xperia | that is the i2clog here http://pastebin.com/d6ffcb25b |
| 09:47.29 | xperia | and that here should be the hardcoded table. http://pastebin.com/m48c21e8b |
| 09:51.06 | druidu | cr2: actually, i'm trying to get android to work again... don't know why it locks up at boot |
| 09:51.13 | druidu | anybody have a working diam100 image? |
| 09:53.14 | druidu | none of the download links on connect-utb work |
| 09:55.45 | druidu | hmm, why do I keep getting "msm_i2c: msm_i2c.o: Warning bus was busy (xx) |
| 09:56.36 | druidu | while booting android, it keeps showing this error on the console while the android logo flashes |
| 10:00.00 | xperia | cr2: are you sure about the size? becouse i have as a size value zero. |
| 10:00.02 | xperia | the anoying thing is that we have two blocks that repeat in my i2c logs. |
| 10:00.03 | xperia | first three lines with writes. 100 mili seconds delay |
| 10:00.05 | xperia | four lines writes 60 miliseconds delay |
| 10:00.06 | xperia | repeat the whole again. |
| 10:00.08 | xperia | Do you have also such a thing or is this only on my device becouse I am having on the htc-kovsky device a spectral colour effect when the leds turn on. |
| 10:00.09 | xperia | for me my i2c log show the change of the spectral colour. |
| 10:00.11 | xperia | but i cant find the sources or the buffer in the sources with this numbers that are sended to be stored. |
| 10:00.12 | xperia | http://www.colour.org.uk/spectrum.gif |
| 10:03.21 | *** part/#htc-linux druidu (n=druidu@78.97.155.94) |
| 10:07.36 | *** join/#htc-linux ogr3_ (n=ogr3@adsl-99-3-81-175.dsl.lsan03.sbcglobal.net) |
| 10:07.49 | xperia | cr2: here is a full i2c log for blinking of a led http://pastebin.com/m6eb4b420 |
| 10:09.28 | Echo31 | cr2 : Hi |
| 10:09.44 | cr2 | dzo: you have a gsm diam ? |
| 10:09.53 | cr2 | Echo31: hi |
| 10:10.53 | cr2 | Echo31: what is the full url to pH5 git ? let's try to compile the kernel |
| 10:11.06 | cr2 | xperia: looking at your data |
| 10:15.03 | Echo31 | cr2: PH5 want to add the patches only the next week. |
| 10:17.12 | cr2 | Echo31: i'll try if it works for me |
| 10:17.43 | Echo31 | cr2: ok |
| 10:18.37 | Echo31 | cr2: I tries to implement ATA ide on athena CS3. What is EGPIO1_C4_HTCATHENA_EPSON_POWER_SLEEP meaning (cpld1) ? |
| 10:20.00 | cr2 | Echo31: i'll check it with the .dll . is it traceable ? |
| 10:24.04 | Echo31 | cr2: The ide driver launches the hda: unfortunely, the error message displays with "hda: timeout in ide_busy_sleep" . I ll try to set this egpio to 1 |
| 10:24.32 | cr2 | check with 'pd' if it's enabled in wince. i think so |
| 10:27.33 | Echo31 | cr2: I go to check it, now |
| 10:30.17 | Echo31 | cr2: HaRET(1)# pd 0x08000000 16 |
| 10:30.17 | Echo31 | 08000000 | 000e0004 00d00013 007e0000 000a0021 | ..........~.!... |
| 10:31.28 | *** join/#htc-linux cbandarrinha (i=55f5b9ae@gateway/web/freenode/x-tdyvqazhhhzbrgso) |
| 10:32.27 | *** join/#htc-linux mallochigh (n=Administ@cpe-173-169-235-155.tampabay.res.rr.com) |
| 10:33.46 | xperia | cr2: the led spectral settings has a time frame from 3.8 second till 5.2 seconds |
| 10:33.48 | xperia | 003808: mmutrace 7805fddc: e5863000(str) b2300000 00000000 (00000000) |
| 10:33.49 | xperia | 005235: mmutrace 7805fddc: e5863000(str) b2300000 00000208 (00000000) |
| 10:33.51 | xperia | the setting of the led colour must be done in a routine. |
| 10:34.21 | *** join/#htc-linux pH5 (n=ph5@g229065192.adsl.alicedsl.de) |
| 10:34.53 | cr2 | Echo31: how is the epson dll called ?? |
| 10:35.20 | Echo31 | cr2: I don't know |
| 10:36.58 | mallochigh | Howto htc raphael / touch pro android? |
| 10:37.10 | mallochigh | Seems like a reading in the forum so I thought I would ask |
| 10:37.13 | mallochigh | Its for my gf |
| 10:38.26 | cr2 | Echo31:hmm. v03usbfn |
| 10:38.47 | Echo31 | cr2 ATA_WRITEBACK.dll with folder Epson-ata and FSD=FATFSD.dll |
| 10:39.27 | cr2 | mallochigh: not there yet, but i guess with dzo joining in we will finish it soon. |
| 10:39.38 | cr2 | Echo31: ok, sounds good. |
| 10:39.59 | *** join/#htc-linux SOG (n=SOG@n058152150059.netvigator.com) |
| 10:40.02 | cr2 | Echo31: don't have ATA* |
| 10:40.13 | cr2 | heh. it's lowercase |
| 10:41.00 | cr2 | looking |
| 10:42.15 | cr2 | aee* is pxa gpio |
| 10:46.31 | *** join/#htc-linux druidu (n=druidu@78.97.155.94) |
| 10:46.33 | cr2 | Echo31: checks the board version (upper CPLD2 bits). do you print it out ? |
| 10:47.15 | cr2 | orr_cpld1(3,0x10) |
| 10:48.26 | cr2 | then something with the pxa gpio 0x10 |
| 10:48.35 | cr2 | msleep 0x4b0 |
| 10:50.59 | Echo31 | cr2: HaRET(10)# pd 0x0a000000 16 |
| 10:50.59 | Echo31 | 0a000000 | 00000000 00000000 00000000 00000000 | ................ |
| 10:51.06 | cr2 | GPIOC4 0x0010 epson power sleep 4b0 |
| 10:51.32 | cr2 | ok, that means 'on=1, msleep 0x4b0) |
| 10:51.50 | cr2 | editing the wiki |
| 10:52.24 | cr2 | GPIO 68 00000010 EPSON power? O |
| 10:52.30 | cr2 | need to check this one too |
| 10:53.08 | cr2 | +0x20 bit |
| 10:54.23 | cr2 | #define GPSR2 __REG(0x40E00020) |
| 10:54.38 | cr2 | ok, so pxa gpio68=1 |
| 10:56.03 | cr2 | #define GPCR2 __REG(0x40E0002C) |
| 10:56.07 | cr2 | on off=0 |
| 10:59.37 | cr2 | Echo31: is the gpio68 changing ? |
| 11:00.49 | Echo31 | cr2: I set the GPIO68 to 1 but egpio not yet |
| 11:01.32 | cr2 | ok |
| 11:01.56 | Echo31 | cr2 : sleep 4b0 ? |
| 11:02.20 | cr2 | msleep 4b0 ? |
| 11:02.56 | cr2 | mdelay(1200) |
| 11:04.00 | Echo31 | cr2: I must expect a mdelay |
| 11:04.27 | cr2 | after setting 2 gpios, delay for 1.2s |
| 11:05.44 | Echo31 | cr2: But I set the egpio during the initlialisation of cpld1 |
| 11:06.54 | cr2 | ? |
| 11:07.31 | *** join/#htc-linux phh (n=quassel@lns-bzn-50f-81-56-224-142.adsl.proxad.net) |
| 11:07.44 | cr2 | it's not for the initial value |
| 11:08.26 | Echo31 | cr2: I set 0x10 for the inital value |
| 11:08.41 | cr2 | it's wrong |
| 11:09.25 | *** join/#htc-linux x29a_ (n=x29a@unaffiliated/x29a) |
| 11:12.54 | *** part/#htc-linux ArteK (n=Artur@81.15.241.96) |
| 11:13.09 | Echo31 | cr2: I set EGPIO register C of CPLD1 with 0x10 value ? |
| 11:15.16 | cr2 | GPIOC4=1 |
| 11:16.04 | cr2 | GPIO68=1 |
| 11:16.10 | cr2 | mdelay(1200); |
| 11:16.24 | cr2 | to powerup epson |
| 11:16.39 | *** join/#htc-linux ArteK (n=Artur@81.15.241.96) |
| 11:21.10 | cr2 | Echo31: btw, why are you using joystick for accelerometer ? afair the raph accelerometer driver is not using it. |
| 11:23.15 | Echo31 | cr2: GPIOC4= 1 -> initiale value=0x10 and I cannot wait 1200 because the pata_platfor and cpld1 are initilised at the same time in platform_add_devices |
| 11:25.15 | cr2 | why can't you just set gpio ? |
| 11:25.34 | cr2 | .init is a completely wrong place for setting this gpio |
| 11:25.52 | cr2 | eventually we will remove the LCD .init bits too. |
| 11:26.05 | cr2 | because they do not belong there |
| 11:27.08 | Echo31 | cr2: I set GPIO068 before the function of platform_add_devices |
| 11:29.06 | Echo31 | cr2: the Gsensor for athena is the raph acelerometer but on I2C |
| 11:29.43 | cr2 | ok, then you can only hope that it will work |
| 11:30.03 | druidu | cr2: i get a `ARM9 has CRASHED` `smem: DIAG ''` when booting android with latest kernel |
| 11:30.05 | druidu | any tips? |
| 11:30.12 | cr2 | raph accelerometer is also i2c, just a different chip |
| 11:30.37 | druidu | btw, your halt works... before disabling panic on oops it shut down completely |
| 11:30.52 | cr2 | druidu: boot with angstrom initrd. if it will work (works for me) then it's an android fault |
| 11:31.12 | druidu | it books ok into console mode |
| 11:31.17 | druidu | I can telnet to it even |
| 11:31.28 | druidu | then while booting android, the flashing logo, I get that ARM9 has crashed |
| 11:31.38 | druidu | just before that is a request suspend state: wakeup |
| 11:31.44 | cr2 | disable power off, then you'll see the crash reason |
| 11:31.59 | cr2 | does the android ram console work ? |
| 11:32.27 | cr2 | nobody looked into suspend/resume yet |
| 11:32.31 | druidu | I disabled panic on ops and this is all I get |
| 11:32.42 | druidu | thing is, latest update from upstream needs CONFIG_HAS_EARLYSUSPEND=y |
| 11:32.50 | druidu | so I activated that |
| 11:33.08 | cr2 | disable it |
| 11:33.56 | druidu | won't work with the latest upstream ver, it needs that to compile htc_pwrsink.c |
| 11:34.11 | druidu | I'll try again without that patch, see if it works |
| 11:35.05 | cr2 | pwrsink is not needed |
| 11:37.56 | druidu | btw |
| 11:37.57 | druidu | <PROTECTED> |
| 11:37.57 | druidu | <PROTECTED> |
| 11:37.57 | druidu | <PROTECTED> |
| 11:38.11 | druidu | so that's what happened, handle_modem_crash was called |
| 11:38.45 | cr2 | we can reboot only with a DEX call |
| 11:39.08 | cr2 | the purpose of gpio25 is not 100% clear |
| 11:40.41 | druidu | <PROTECTED> |
| 11:40.46 | druidu | this is just hilarious :) |
| 11:46.19 | Echo31 | cr2: sorry, the athena Gsensor chip used from htc-shift chip. The i2c between LIS3LV02DL and device is ok but the polling function don't yet work . I don't know why. |
| 11:48.43 | Echo31 | cr2: Where can i find the raph accelero ? name and url ? |
| 11:49.55 | druidu | cr2: your power_down is the weirdest thing ever |
| 11:50.26 | druidu | it doesn't actually shut down (leds are still glowing), soft reset button doesn't do anything and the power button resets the device :) |
| 11:50.57 | druidu | it's a kind of deep sleep |
| 11:51.07 | cr2 | Echo31: drivers/i2c/chips/kionix-kxsd9.c |
| 11:51.39 | Echo31 | cr2: thanks |
| 11:52.36 | cr2 | druidu: the leds are controlled by a microcontroller |
| 11:53.00 | cr2 | druidu: the power_off is only for the pmic |
| 11:53.35 | cr2 | at least it is more useful than RESET_ARM9 :) |
| 11:54.14 | cr2 | druidu: is 'soft reset' a gpio ? |
| 11:55.56 | druidu | don't now, doubt it |
| 11:56.03 | druidu | works right after I boot the kernel |
| 11:56.08 | druidu | so no drivers, no nothing |
| 11:56.53 | druidu | you might be able to disable it though, maybe |
| 11:59.32 | *** join/#htc-linux phh (n=quassel@lns-bzn-50f-81-56-224-142.adsl.proxad.net) |
| 12:00.32 | Echo31 | cr2: the raph accelero is based on hrtimer for polling : the htc-shift uses input-polldev. I ll try with hrtimer |
| 12:10.15 | Echo31 | cr2: for epson, i must to set GPIOC4=1 ( 0x10 to cpld1 register C) and GPIO68=1. However, I cannot wait mdelay(1200) |
| 12:11.01 | cr2 | Echo31: the device driver should be able to set power. also suspend/resume |
| 12:19.01 | druidu | cr2: ok, this is weird, while booting android I just get a "Power down." message then msm_pm_power_off is called |
| 12:19.13 | druidu | no other message printed on the console |
| 12:24.10 | phh | print the backtrace at the start of msm_pm_power_off |
| 12:29.20 | druidu | good idea |
| 12:29.37 | druidu | lol, I can play mp3s in android, even if I can't hear them :) |
| 12:35.10 | *** join/#htc-linux Bally3 (n=Bally3@cpc1-blac6-0-0-cust618.manc.cable.ntl.com) |
| 12:42.29 | druidu | LOL |
| 12:42.44 | druidu | looks like the batter driver doesn't work and because of that android thinks it's best to shutdown |
| 12:42.47 | druidu | so calls sys_reboot |
| 12:46.40 | *** join/#htc-linux x_coRe_x (i=coRe@95-37-233-84.dynamic.mts-nn.ru) |
| 12:49.04 | parmaster_ | wow cr2 is back! we're saved! |
| 13:22.44 | *** join/#htc-linux leaigor (n=laigor@188.134.36.14) |
| 13:23.02 | *** join/#htc-linux dream_kill (n=nospam@89.131.127.37) |
| 13:23.58 | phh | druidu: "doesn't work" means report 0 ? |
| 13:24.38 | druidu | damn, just noticed |
| 13:24.46 | druidu | board diamond doesn't register the battery driver |
| 13:24.48 | druidu | so it's not there |
| 13:25.05 | druidu | gonna port some stuff from board-htcraphael to diamond |
| 13:26.50 | *** join/#htc-linux Echo31 (n=olivier@mir31-4-82-240-194-54.fbx.proxad.net) |
| 13:27.01 | druidu | kdiff3 rulez again :) |
| 13:29.03 | phh | oh you're doing diamond's port ? great. if i can help with my diam100+ a diam 160 battery (not the standard one but a 2Ah, don't know if it changes anything) |
| 13:29.15 | phh | (but way too lazy to setup a crosscompiler.) |
| 13:29.23 | *** join/#htc-linux Squarc (n=Squarc@82-217-32-29.cable.quicknet.nl) |
| 13:29.57 | *** join/#htc-linux marajin (n=marajin@87-194-102-189.bethere.co.uk) |
| 13:32.58 | Echo31 | cr2: I tried some ways, but platform_add_devices function initialises in the same time the pata_platform ( ata ide) and cpld1. I cannot insert a mdelay. |
| 13:39.07 | Echo31 | cr2: Your Idea is it to suspend the pata_platform during 1200 for the initiatlisation of devices, and do after a resume ? |
| 13:40.44 | cr2 | Echo31: i'm surprised. how does this driver look like ? |
| 13:43.25 | Echo31 | cr2: what ? |
| 14:18.08 | Echo31 | cr2: I do |
| 14:18.09 | Echo31 | init GPIO68, |
| 14:18.09 | Echo31 | platform_add_device, |
| 14:18.09 | Echo31 | mdelay(1200); |
| 14:18.09 | Echo31 | platform_device_register(&athena_device_ide0); |
| 14:26.28 | *** join/#htc-linux rafyvitto (n=rafyvitt@24.54.253.34) |
| 14:27.45 | *** join/#htc-linux marex (n=marex@thor.hackndev.com) |
| 14:28.21 | cr2 | Echo31: does athena_device_ide0 support .probe, .init, .exit, .suspend, .resume functions ? |
| 14:28.52 | cr2 | why should you put the power init before the platform_device_ register ? |
| 14:29.27 | cr2 | registering device just creates some structures and calls already predefined functions. |
| 14:30.31 | cr2 | druidu: still here ? |
| 14:31.27 | Echo31 | cr2: I think athena_device_ide0 supports .probe, .init, .exit, .suspend, .resume functions |
| 14:34.23 | druidu | yes |
| 14:34.25 | druidu | finally, it works |
| 14:34.50 | druidu | I ported all the raphael stuff to diamond (the lack of a battery driver was causing the shutdown anyway) |
| 14:35.11 | druidu | what is *really weird* is that I distinctly remember booting a certain android image yesterday |
| 14:35.12 | rafyvitto | hey cr2 |
| 14:35.26 | druidu | and that image had a touchscreen driver which reported the size of the touched area!!! |
| 14:35.35 | druidu | and with latest kernel this doesn't happen |
| 14:35.43 | *** join/#htc-linux mrmoku|away (n=mrmoku@ppp-82-135-3-67.dynamic.mnet-online.de) |
| 14:35.43 | rafyvitto | did you guys ever heard about this http://forum.xda-developers.com/showthread.php?t=388273&highlight=linux |
| 14:36.03 | rafyvitto | dual os on the htc typhoon linux/windows mobile |
| 14:36.23 | druidu | yeah, somebody mentioned it here yesterday :) |
| 14:36.32 | rafyvitto | me |
| 14:36.34 | rafyvitto | xD |
| 14:36.38 | druidu | hehe |
| 14:36.57 | rafyvitto | anyways im out guys, gl to all and keep up the good work cheers to all of ya |
| 14:37.01 | druidu | it's not a priority anyway, it won't make phone or gprs magically work :) |
| 14:37.07 | druidu | bye |
| 14:37.23 | rafyvitto | yeh but its an alternative to booting linux xD |
| 14:37.27 | rafyvitto | instad of using haret |
| 14:37.36 | rafyvitto | instead* |
| 14:37.59 | druidu | cr2: ok, so the pointer location app from dev tools clearly and correctly showed touch area size |
| 14:38.00 | *** join/#htc-linux Gnutoo (n=gnutoo@ABordeaux-152-1-40-189.w83-193.abo.wanadoo.fr) |
| 14:39.06 | *** join/#htc-linux dcordes (n=luke-g@unaffiliated/dcordes) |
| 14:40.10 | Gnutoo | hi cr2 |
| 14:40.20 | Gnutoo | cr2, did you get my mail? |
| 14:43.44 | cr2 | Gnutoo: yes. |
| 14:44.03 | cr2 | Gnutoo: fyi, i don't write non-free software. |
| 14:44.22 | Gnutoo | cr2, ok so what's the license |
| 14:44.23 | Gnutoo | ? |
| 14:44.30 | Gnutoo | btw thatnks a lot and it's great |
| 14:44.46 | cr2 | but since most of the code is taken from android source, it's apache license i guess. |
| 14:44.54 | Gnutoo | ok |
| 14:44.59 | Gnutoo | apache is like BSD |
| 14:45.10 | Gnutoo | + a close against patents agression |
| 14:45.17 | cr2 | ok |
| 14:45.36 | Gnutoo | so if it's apache you are free to make it proprietary that's why I mailed you |
| 14:45.45 | Gnutoo | the original is apache 2.0 |
| 14:45.52 | cr2 | lol |
| 14:46.26 | xperia | cr2: new stuff for the raphael microp wiki page |
| 14:46.27 | Gnutoo | and unfortunately a source file without a license is proprietary by default |
| 14:46.28 | xperia | http://pastebin.com/m3299b929 |
| 14:46.32 | Gnutoo | so I needed your permission |
| 14:47.15 | cr2 | gplv2+ is ok in my eyes, but as i've said, this code is mostly cut'n'paste |
| 14:47.36 | Gnutoo | cr2, ok could you send it or pastebin to me a version with the apache header? |
| 14:47.57 | Gnutoo | me too I prefer GPLv3 |
| 14:48.04 | Gnutoo | but it realy depend on the goal |
| 14:48.13 | Gnutoo | or Agplv3 |
| 14:48.34 | cr2 | ok. leave it apache2.0 and paste the license from android |
| 14:48.39 | Gnutoo | ok |
| 14:48.41 | Gnutoo | thanks |
| 14:57.54 | druidu | cr2: should the accelerometer work? |
| 14:58.18 | druidu | <PROTECTED> |
| 14:58.36 | druidu | I get some debug output but sensors don't update in android |
| 15:04.22 | cr2 | druidu: it works as input device |
| 15:04.32 | cr2 | druidu: od -x /dev/input/event* |
| 15:13.48 | druidu | what's the stuff about memory banks in board fixup? |
| 15:13.55 | druidu | diamond only has one memory bank? |
| 15:15.50 | cr2 | 128MB |
| 15:16.08 | cr2 | there may be a 256MB diamond, i don't know |
| 15:16.31 | cr2 | the sram parts may be reused too |
| 15:16.35 | druidu | RAM-capacity: 192 MiB |
| 15:16.38 | druidu | according to specs |
| 15:17.00 | cr2 | but the g1 kernel prefers to use static allocation |
| 15:17.12 | cr2 | 192=64+128 ? |
| 15:17.29 | cr2 | you have 64MB sram ? |
| 15:17.34 | druidu | ah, I see touch pro has RAM:capacity: 288 MiB (accessible: 201 MiB) |
| 15:17.47 | cr2 | raph has 32+256=288 |
| 15:18.02 | cr2 | "accessible" is useless |
| 15:18.28 | cr2 | but parts of sram and main sdram are used by arm9, and are protected by MPU |
| 15:18.45 | cr2 | maybe we can disable this protection, i don't know |
| 15:18.56 | druidu | well, maybe arm9 really needs`em |
| 15:18.57 | cr2 | the spl in sram is protected, for example |
| 15:19.00 | druidu | it does run the AMSS |
| 15:19.01 | cr2 | lol |
| 15:19.11 | cr2 | yes, it's the amss ram |
| 15:19.24 | druidu | how can I find out how big the ram module are? |
| 15:19.28 | cr2 | are you sure that you have 64MB sram ? |
| 15:19.28 | druidu | WM doesn't show |
| 15:19.31 | druidu | only that I have 192 |
| 15:19.35 | druidu | total |
| 15:19.47 | druidu | so it must be 2 modules, there's no 192 ram module |
| 15:19.51 | cr2 | this info should be in the MFG section in nand |
| 15:20.02 | druidu | MFG? |
| 15:20.13 | druidu | how can I access it? |
| 15:20.20 | cr2 | 192 means 64MB sram and 128MB sdram |
| 15:20.34 | cr2 | in linux ? |
| 15:20.40 | druidu | dunno, wherever |
| 15:20.44 | cr2 | MFG is the manufacturer area |
| 15:20.57 | druidu | so which one is first? |
| 15:21.01 | druidu | 64 or 128? |
| 15:21.10 | cr2 | in linux you need a patched haret, to provide the partition table as on g1 |
| 15:21.20 | cr2 | sram is at 0x0 |
| 15:21.32 | cr2 | the sdram is at 0x10000000 |
| 15:21.49 | cr2 | if you have a second bank (like raph), then it's at 0x20000000 |
| 15:22.24 | cr2 | look at http://www.htc-linux.org/wiki/index.php?title=RaphaelMemoryMap |
| 15:23.26 | druidu | what I still don't get it is how it comes up with 89 for raphael |
| 15:23.34 | druidu | <PROTECTED> |
| 15:26.25 | cr2 | See board-htcraphael.h |
| 15:26.46 | *** join/#htc-linux miknix (n=miknix@gentoo/developer/miknix) |
| 15:27.03 | cr2 | some ram is reserved by amss, and there is android static allocation |
| 15:27.05 | druidu | aaah, #define MSM_LINUX_SIZE 0x5900000 /* 89mb */ |
| 15:27.36 | cr2 | i guess we can get rid of the static allocation, at least for the non-android use |
| 15:30.00 | cr2 | druidu: can you test BT ? |
| 15:31.45 | druidu | hmm, driver's active |
| 15:31.47 | druidu | tell me how |
| 15:34.24 | druidu | dmesg shows: HCI UART, HCI H4, HCI BCSP, HCILL, L2CAP, SCO, RFCOMM, BNEP |
| 15:34.40 | druidu | but it doesn't show that any device was detected |
| 15:35.46 | cr2 | did you enable msm uart hs in .config ? |
| 15:36.06 | *** join/#htc-linux slawson (n=kw@adsl-178-60-36.gnv.bellsouth.net) |
| 15:36.22 | cr2 | the HCI* et al. is a higher level protocol |
| 15:39.33 | druidu | yeah, noticed |
| 15:39.47 | druidu | I added all the drivers to the board- file, gonna check config now |
| 15:39.57 | druidu | but don't know what happened that all my buttons are messed up |
| 15:40.33 | druidu | back and home don't work anymore :-s |
| 15:41.21 | druidu | (I configured the end call button to be the home button) |
| 15:42.56 | slawson | my back and home never worked :o |
| 15:44.25 | cr2 | i'm not familiar with diamond |
| 15:44.48 | cr2 | the only problem i have is the wrong keyboard leds |
| 15:45.53 | slawson | Your radio works etc? |
| 15:49.51 | druidu | ok, accelerometer works |
| 15:50.16 | druidu | home worked (after configuring it in spare parts app) |
| 15:50.22 | druidu | and back too |
| 15:50.32 | slawson | Amazing. |
| 15:50.45 | *** join/#htc-linux MethoS- (n=clemens@dyndsl-085-016-167-109.ewe-ip-backbone.de) |
| 15:50.46 | cr2 | slawson: gsm and umts work for me. |
| 15:50.56 | slawson | ah i am diam500 |
| 15:50.56 | druidu | btw, when you mean gsm |
| 15:51.05 | druidu | you can actually call or receive calls? |
| 15:51.22 | cr2 | druidu: AT works (i.e. SMS too) |
| 15:51.36 | druidu | because when I receive a call, it is automatically rejected (and the caller hears a busy signal, like when you push the hang up button) |
| 15:51.39 | cr2 | druidu: but sice the ADSP does not work as expected, you don't have sound |
| 15:51.47 | druidu | and it shows up a missed call |
| 15:52.04 | cr2 | android does not send ata |
| 15:52.13 | druidu | ata? |
| 15:52.15 | druidu | sms works too for me |
| 15:52.17 | cr2 | you need some ril library |
| 15:52.37 | cr2 | ata is ATA |
| 15:52.41 | druidu | [ 6.150367] warning: `rild' uses 32-bit capabilities (legacy support in use) |
| 15:52.42 | cr2 | answer the call |
| 15:52.56 | cr2 | 64bit cpu ? |
| 15:53.12 | druidu | dunno, this shows up right after booting android |
| 15:53.24 | druidu | in dmesg |
| 15:55.32 | druidu | btw, od /dev/input/event0 (navipad) shows total garbage... I keep pressing the upper left button and except one value that goes to 1 while it is pressed and 0 when it's not, all is garbage |
| 15:57.22 | druidu | ah... I think I understand what it does |
| 15:59.26 | *** part/#htc-linux ArteK (n=Artur@81.15.241.96) |
| 16:01.01 | druidu | SERIAL_MSM_HS |
| 16:01.03 | druidu | it was enabled |
| 16:01.04 | druidu | cr2: |
| 16:01.26 | cr2 | ttyHS1 in the dmesg ? |
| 16:02.03 | druidu | hmmm |
| 16:02.07 | druidu | you disabled the clock |
| 16:02.12 | druidu | in clock-7x01a.c |
| 16:02.16 | druidu | clock-7x01a.c: CLOCK("uartdm_clk", UART2DM_CLK, &msm_device_uart_dm2.dev, OFF), |
| 16:02.38 | cr2 | it's ok |
| 16:02.40 | druidu | Aug 30 18:33:05 (none) user.info kernel: [ 1.686317] msm_serial_hs.1: ttyHS1 at MMIO 0xa0300000 (irq = 53) is a MSM HS UART |
| 16:02.53 | cr2 | yes |
| 16:02.57 | druidu | Aug 30 18:33:05 (none) user.info kernel: [ 1.788062] msm_hsusb: functions bound. starting. |
| 16:02.57 | druidu | Aug 30 18:33:05 (none) user.info kernel: [ 1.796577] hsusb: IDLE -> ONLINE |
| 16:02.57 | druidu | Aug 30 18:33:05 (none) user.info kernel: [ 1.808295] hsusb: reset controller |
| 16:03.13 | cr2 | it's usb |
| 16:03.26 | cr2 | what does stty -F /dev/ttyHS1 say ? |
| 16:03.45 | druidu | # stty -F /dev/ttyHS1 |
| 16:03.45 | druidu | speed 9600 baud; |
| 16:03.45 | druidu | intr = ^C; quit = ^\; erase = ^?; kill = ^U; eof = ^D; eol = <undef>; eol2 = <undef>; start = ^Q; stop = ^S; susp = ^Z; rprnt = ^R; werase = ^W; lnext = ^V; flush = ^O; |
| 16:03.45 | druidu | min = 1; time = 0; |
| 16:03.45 | druidu | -brkint -imaxbel |
| 16:04.00 | cr2 | ok |
| 16:04.04 | cr2 | enable power |
| 16:04.21 | cr2 | in /sys/class/rfkill/* |
| 16:04.21 | druidu | don't know BT commands :D |
| 16:04.26 | cr2 | not bt |
| 16:04.42 | druidu | ok |
| 16:04.42 | druidu | wait |
| 16:04.45 | cr2 | it's just a serial port |
| 16:04.57 | cr2 | no BT commands |
| 16:05.03 | druidu | echo 1 > /sys/class/rfkill/rfkill0/power/wakeup ? |
| 16:05.15 | druidu | no, wait |
| 16:05.31 | cr2 | <PROTECTED> |
| 16:05.38 | druidu | yeah, rfkill0 is brf6300 which sounds like BT |
| 16:05.51 | druidu | [ 1954.767158] bluetooth rfkill state ON |
| 16:06.00 | druidu | and some clock messages |
| 16:06.12 | cr2 | call stty again. and get the hard lock ;) |
| 16:06.36 | druidu | LOL |
| 16:06.43 | druidu | now I see what you mean by hard lock |
| 16:06.44 | druidu | =) |
| 16:06.53 | cr2 | yes, it's my status for today |
| 16:06.55 | druidu | good sign right? :) |
| 16:06.58 | cr2 | must be some DMA problem. |
| 16:07.37 | cr2 | g1 kernel put uart2dm tx on channel 10 |
| 16:07.42 | druidu | problem is we get absolutely no console feedback |
| 16:07.49 | cr2 | which is already used by usb = braindead |
| 16:07.51 | druidu | I could try to do that while having the tty win open |
| 16:07.55 | druidu | maybe something shows up? |
| 16:08.04 | cr2 | but i've changed it to 9 |
| 16:08.16 | cr2 | wince uses channel 8 |
| 16:08.24 | cr2 | but we use 8 for SD dma |
| 16:08.33 | cr2 | wince does not support SD dma |
| 16:08.35 | cr2 | btw |
| 16:08.36 | druidu | test-root/init.rc:service hfag /system/bin/sdptool add --channel=10 HFAG |
| 16:08.36 | druidu | test-root/init.rc:service hsag /system/bin/sdptool add --channel=11 HSAG |
| 16:08.42 | druidu | in init.rc |
| 16:08.44 | druidu | some daemon |
| 16:09.26 | druidu | not sure what's with that... |
| 16:09.26 | cr2 | we need to make uart rx/tx work first |
| 16:09.26 | cr2 | before talking about any BT-related things |
| 16:09.30 | cr2 | another difference to g1 is the clock setup |
| 16:09.37 | druidu | so bluetooth is basicly standard +AT commands? :) |
| 16:09.58 | cr2 | androids use 4X clock speed. probably to avoid the bps < 1500 |
| 16:10.35 | cr2 | BT is a protocol stack used over (RF) uart (usually) |
| 16:11.14 | cr2 | btw, can you run with SD dma on channel 9 ? |
| 16:11.33 | druidu | hmm, dunno lemme see |
| 16:11.39 | cr2 | we don't know the DMA constoller capabilities, because there are no CPU docs |
| 16:12.21 | druidu | #define DMOV_USB_CHAN 11 |
| 16:12.30 | druidu | we have cpu docs |
| 16:12.31 | cr2 | #define DMOV_SDC1_CHAN 8 |
| 16:12.37 | druidu | #define DMOV_SDC1_CHAN 8 |
| 16:12.48 | druidu | #define DMOV_HSUART2_TX_CHAN 4 // =1 in wince ?? |
| 16:12.50 | cr2 | #define DMOV_HSUART2_RX_CHAN 9 // wince value, does not clash with USB |
| 16:12.51 | druidu | #define DMOV_HSUART2_RX_CHAN 9 // wince value, does not clash with USB |
| 16:12.56 | cr2 | yes, i've mixed the numbers |
| 16:13.20 | cr2 | can you exchange DMOV_SDC1_CHAN and DMOV_HSUART2_RX_CHAN ? |
| 16:13.23 | druidu | did you get any console after crash? any msg at all? |
| 16:13.37 | druidu | I can try to see if there's any message with the console active |
| 16:13.45 | cr2 | nothing. the machine is locked, no irq processing anymore |
| 16:14.10 | druidu | I was hoping that a message would show before that |
| 16:15.07 | druidu | what is _CRCI? |
| 16:15.29 | cr2 | the easiest thing to test is SD on DMA9 |
| 16:15.40 | cr2 | then we can put UART2DM on DMA8 and will have more or less the wince setup |
| 16:16.03 | cr2 | it's some throttling mechanism for DMA |
| 16:16.16 | druidu | ah, so we don't need to mess with it |
| 16:16.16 | cr2 | forget about it for now. |
| 16:17.24 | cr2 | echange 8 and 9 and try to boot |
| 16:18.09 | cr2 | another thing to try is reducing the uart2dm clock by 4 times, so we will have wince setup there too. |
| 16:21.10 | *** join/#htc-linux Dunedan (n=dunedan@phoenitydawn.de) |
| 16:30.11 | druidu | btw, except audio routing for gsm, would playing sound work? |
| 16:31.48 | *** join/#htc-linux BHSPitLappy (n=BHSPitLa@unaffiliated/bhspitmonkey) |
| 16:32.14 | cr2 | no |
| 16:32.22 | cr2 | at least not for me |
| 16:33.00 | cr2 | try it. i got some rpc messages with playwav2 and then it locked too |
| 16:37.44 | druidu | what's the audio device? |
| 16:37.58 | druidu | cat /dev/random > /dev/sound or whatever might also be good to test :)) |
| 16:42.36 | *** join/#htc-linux phh (n=quassel@lns-bzn-50f-81-56-224-142.adsl.proxad.net) |
| 16:44.06 | druidu | what's the major/minor for ttyHS1 ? |
| 16:50.33 | cr2 | <PROTECTED> |
| 16:50.50 | cr2 | drivers/serial/msm*hs.c |
| 16:53.41 | druidu | same |
| 16:54.01 | druidu | (neah, I was trying without android booted up and had to look up the device in /sys/devices) |
| 16:54.48 | cr2 | same ? |
| 16:55.05 | cr2 | does the SD work with dma9 ? can you post the dmesg ? |
| 17:02.41 | druidu | LOL |
| 17:02.45 | druidu | I have debug output!!! |
| 17:02.49 | druidu | from the console |
| 17:03.10 | druidu | pc_clk_set_rate: id=35 rate=7372800 |
| 17:03.44 | druidu | set mdns: 35, 7372800; bitidx=26 offset=dc, ns=0 |
| 17:03.51 | druidu | LOOOL |
| 17:03.56 | druidu | just got a stacktrace |
| 17:04.49 | druidu | uart_open => uart_startup => msm_hs_startup => request_irq => setup_irq + 0x0/0x2ac |
| 17:05.07 | druidu | request_irq+0x94/0xbc |
| 17:06.30 | druidu | hmm, there's also some backtrace before that, looks like inside the hrtimer_interrupt |
| 17:13.33 | cr2 | yes, it's a problem in irq processing |
| 17:13.45 | cr2 | i think it all starts with some DMA timeout |
| 17:14.41 | druidu | weird thing is that it was hung but only apparently |
| 17:14.52 | druidu | after like 1 minute, the screen refreshed and I could see the backtrace |
| 17:15.08 | druidu | they request 2 irqs, one if wakeup is enabled |
| 17:17.18 | druidu | also, having BT enabled before booting kernel doesn't help |
| 17:19.42 | cr2 | i don't really know how to debug the dma |
| 17:22.50 | marex | cr2: use a scope ... there's no other way |
| 17:23.07 | marex | cr2: or possibly jtag and trace memory accesses ... but that's all you can do about dma |
| 17:27.57 | cr2 | ok, i must leave now. |
| 17:27.58 | cr2 | bbl |
| 17:28.57 | druidu | tried outputting stuff directly to /dev/msm_pcm_out but get somethink like [ 99.780616] audio_disable() in dmesg and no sound |
| 17:28.59 | druidu | ok, later |
| 17:30.50 | druidu | btw, just enabling the device causes msm_hs_startup to request two irqs: 53 and 85, and no crash yet |
| 17:32.10 | druidu | btw, back and home buttons magically work again... |
| 17:33.11 | *** join/#htc-linux luc_ (n=luc@89-115-128-35.cl.ipv4ilink.net) |
| 17:44.38 | *** join/#htc-linux slawson (n=nnscript@adsl-178-60-36.gnv.bellsouth.net) |
| 17:48.00 | slawson | Does the cdma radio work yet? |
| 17:51.32 | *** join/#htc-linux mblancom (n=mblancom@193.Red-79-149-207.staticIP.rima-tde.net) |
| 17:52.36 | druidu | dunno, I have a gsm one |
| 17:53.14 | druidu | btw, back and home now magically work again |
| 17:53.44 | druidu | (home as in the end call button, configured in spare parts app) |
| 17:54.13 | slawson | Hmm. |
| 17:54.22 | slawson | Are you the only one working on this project? |
| 17:55.55 | *** join/#htc-linux luc_ (n=luc@89-115-128-35.cl.ipv4ilink.net) |
| 17:56.40 | slawson | yeah cdma radio doesn't work |
| 17:57.59 | druidu | for diamond I could be the only one, but code is 99% identical to raphael |
| 17:59.38 | slawson | You said that accelerometer works? |
| 17:59.40 | slawson | what build |
| 18:01.04 | druidu | works but not in android |
| 18:01.27 | druidu | I'm compiling from sources, latest git version with a few mods |
| 18:01.41 | druidu | mostly porting all the raphael board code to the diamond board |
| 18:01.47 | druidu | I'll post some patches after I test some more |
| 18:01.56 | slawson | Okay awesome |
| 18:02.00 | slawson | Greatly appreciate it |
| 18:02.08 | slawson | if i had a linux box i would be compiling |
| 18:04.04 | druidu | anyway, you can test accelerometer by doing (after booting the kernel from cmd line or via telnet): od /dev/input/event3 |
| 18:06.35 | *** join/#htc-linux ArteK (n=Artur@81.15.241.96) |
| 18:13.29 | *** join/#htc-linux MrPippy (n=pip@75.37.163.177) |
| 18:38.12 | *** part/#htc-linux mblancom (n=mblancom@193.Red-79-149-207.staticIP.rima-tde.net) |
| 18:42.19 | *** join/#htc-linux phh (n=quassel@2a01:e35:2e4b:b2b0:250:8dff:fee1:c793) |
| 18:43.05 | *** join/#htc-linux Amaranth (n=travis@ubuntu/member/Amaranth) |
| 18:54.55 | *** join/#htc-linux Bally3_ (n=Bally3@cpc1-blac6-0-0-cust618.manc.cable.ntl.com) |
| 18:55.08 | Bally3_ | http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=shortlog;h=refs/heads/htc-vogue |
| 18:55.17 | Bally3_ | you mean this? |
| 18:56.16 | *** join/#htc-linux Bally3__ (n=Bally3@cpc1-blac6-0-0-cust618.manc.cable.ntl.com) |
| 18:58.18 | xperia | Bally3_ the source that druidu is working are here |
| 18:58.20 | xperia | http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=tree;h=refs/heads/htc-msm-2.6.27;hb=refs/heads/htc-msm-2.6.27 |
| 18:58.21 | xperia | the board files you can find in arch/arm/mach-msm |
| 18:58.23 | xperia | He use the raph boardfile with new changes |
| 18:58.54 | druidu | yep |
| 18:59.26 | druidu | except the keypad there are very few differences between raphael and diamond |
| 19:00.14 | druidu | lol, actually he just left |
| 19:00.34 | *** join/#htc-linux fon (n=fon@5ac08d96.bb.sky.com) |
| 19:00.47 | druidu | hm, or not |
| 19:03.44 | Bally3 | thanks xperia :) |
| 19:06.07 | Bally3 | sorry.. was tryng to use linux in a vm.. :S |
| 19:08.58 | Bally3 | I take it this is for the hermes port? |
| 19:09.52 | druidu | not really, diamond/raphael :) |
| 19:10.41 | Bally3 | oh |
| 19:10.47 | Bally3 | :S |
| 19:11.08 | Bally3 | I donthave a damond or a raphael tho |
| 19:12.05 | Bally3 | actually snce I got you two here with raoh and diamoind.. can you tell me if these listed are software? |
| 19:12.41 | Bally3 | http://www.htc-linux.org/wiki/index.php?title=Raphael_IRQ |
| 19:13.42 | Bally3 | I'm trying to find the wifi irq for polaris.. we know the kaiser one is "93" ...used haretconsole to watch gpios.. now lost |
| 19:14.30 | *** join/#htc-linux darkstone (n=LOL@78-86-59-180.zone2.bethere.co.uk) |
| 19:16.44 | *** join/#htc-linux pH5 (n=ph5@g229065192.adsl.alicedsl.de) |
| 19:18.30 | Bally3 | xperia: sorry for the confusion mate..I pasted that link in the wrong room.. now I understand what you were talking about lol |
| 19:30.34 | *** join/#htc-linux luc_ (n=luc@89-115-128-35.cl.ipv4ilink.net) |
| 19:32.34 | *** join/#htc-linux luc_ (n=luc@89-115-128-35.cl.ipv4ilink.net) |
| 19:33.43 | *** join/#htc-linux luc_ (n=luc@89.115.128.35) |
| 19:35.51 | *** join/#htc-linux Bally3 (n=Bally3@cpc1-blac6-0-0-cust618.manc.cable.ntl.com) |
| 19:43.26 | *** join/#htc-linux toi (n=toi@d54C2AAB7.access.telenet.be) |
| 19:44.22 | *** join/#htc-linux ogr3_ (n=ogr3@adsl-99-62-2-195.dsl.lsan03.sbcglobal.net) |
| 19:49.09 | *** join/#htc-linux dzo (n=dzo@mail.marginz.co.nz) |
| 19:53.44 | *** join/#htc-linux pH5 (n=ph5@g229239070.adsl.alicedsl.de) |
| 20:00.11 | *** join/#htc-linux glyph` (n=glyph@c-76-24-223-175.hsd1.ma.comcast.net) |
| 20:00.12 | *** join/#htc-linux marajin (n=marajin@87-194-102-189.bethere.co.uk) [NETSPLIT VICTIM] |
| 20:00.15 | *** join/#htc-linux mrb__ (n=mrb@160.89-10-30.nextgentel.com) [NETSPLIT VICTIM] |
| 20:00.15 | *** join/#htc-linux maejrep (n=madCoder@68.64.144.211) [NETSPLIT VICTIM] |
| 20:00.15 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) [NETSPLIT VICTIM] |
| 20:10.51 | *** join/#htc-linux leaigor (n=laigor@188.134.36.14) |
| 20:14.39 | *** join/#htc-linux x_coRe_x (i=coRe@95-37-233-84.dynamic.mts-nn.ru) |
| 20:28.08 | *** join/#htc-linux mallochigh1 (n=Administ@cpe-173-169-235-155.tampabay.res.rr.com) |
| 20:38.37 | *** join/#htc-linux mblancom (n=mblancom@193.Red-79-149-207.staticIP.rima-tde.net) |
| 20:42.12 | *** join/#htc-linux x29a (n=x29a@unaffiliated/x29a) |
| 21:01.22 | *** part/#htc-linux ArteK (n=Artur@81.15.241.96) |
| 21:02.13 | *** join/#htc-linux Echo31 (n=olivier@mir31-4-82-240-194-54.fbx.proxad.net) |
| 21:13.26 | *** part/#htc-linux fon (n=fon@5ac08d96.bb.sky.com) |
| 21:40.32 | *** join/#htc-linux cr2 (n=cr2@ip-77-25-251-112.web.vodafone.de) |
| 21:42.43 | *** join/#htc-linux cr2 (n=cr2@ip-77-25-251-112.web.vodafone.de) |
| 21:43.08 | *** join/#htc-linux eway (n=ellis@78-86-162-166.zone2.bethere.co.uk) |
| 21:45.32 | druidu | cr2: |
| 21:45.34 | druidu | I have news :) |
| 21:45.44 | druidu | I know what's happening to the bt driver |
| 21:45.53 | cr2 | ok |
| 21:46.10 | cr2 | the log is not available again |
| 21:46.32 | cr2 | tell me |
| 21:46.43 | druidu | I didn't say anything on the channel anyway |
| 21:46.43 | druidu | so |
| 21:46.54 | druidu | in the moment the irq is requested |
| 21:46.57 | druidu | 58 |
| 21:47.02 | druidu | I think |
| 21:47.16 | druidu | it starts to fire constantly, like a few times / second |
| 21:47.37 | cr2 | it's not that mich |
| 21:47.50 | druidu | and msm_hs_isr is called all the time |
| 21:47.50 | cr2 | s/mi/mu/ |
| 21:47.53 | druidu | <PROTECTED> |
| 21:47.57 | druidu | this returns 0 all the time |
| 21:48.31 | druidu | well bottom line is that the request_irq doesn't return |
| 21:48.40 | druidu | it hasn't got time... the irq keeps firing |
| 21:48.55 | druidu | I used printk before and after it |
| 21:49.00 | cr2 | hmm. but g1 must have the same problem ? |
| 21:49.03 | druidu | and a printk in the _isr |
| 21:49.17 | druidu | it's true that I bias the results by adding a printk in the _isr, but the isr_status is always 0 |
| 21:49.24 | druidu | so something's wrong |
| 21:49.30 | cr2 | are there any 2.6.29 changes ? |
| 21:49.32 | druidu | the isr is firing and we don't clear it |
| 21:49.40 | druidu | I'm gonna port the changes now |
| 21:49.46 | cr2 | UARTDM_MISR_ADDR is ok. uport ? |
| 21:49.53 | cr2 | great. |
| 21:53.14 | druidu | dunno |
| 21:53.22 | druidu | uport is the _isr handler param |
| 21:53.35 | druidu | should be ok, it's a ptr to a structure |
| 21:54.09 | druidu | uport->irq is 53 |
| 21:54.59 | druidu | #define INT_ARM11_DMA (32 + 11) |
| 21:55.01 | druidu | hmmm |
| 21:55.21 | druidu | it is the dma :)) |
| 21:55.39 | druidu | and because reading that port returns 0, the _isr doesn't do anything so the irq keeps firing? |
| 21:56.07 | *** join/#htc-linux cr2_ (n=cr2@ip-77-25-43-175.web.vodafone.de) |
| 21:56.51 | *** join/#htc-linux pH5 (n=ph5@g229239070.adsl.alicedsl.de) |
| 21:57.39 | druidu | actually, I just realized... if irq 53 is the dma one, it keeps firing because MDP used DMA right? |
| 21:57.44 | druidu | *uses |
| 22:00.56 | druidu | ok, forget what I just said, it seems I have trouble adding numbers |
| 22:00.56 | druidu | #define INT_UART2DM_RX (32 + 20) |
| 22:00.57 | druidu | #define INT_UART2DM_IRQ (32 + 21) |
| 22:01.03 | druidu | these are the uart2dm irqs |
| 22:01.11 | druidu | to irq 53 keeps firing |
| 22:01.53 | *** join/#htc-linux stickboy (n=anonymou@128.153.211.112) |
| 22:02.59 | Bally3 | druidu: can you tell me where to find the irq for wifi in the kernel source please so i can change it? |
| 22:03.13 | cr2_ | it#s ok, in haret you can see a lot of them |
| 22:03.24 | druidu | all IRQs should be in arch/arm/mach-msm/include/mach/irqs.h |
| 22:03.26 | cr2_ | but only because the .bts is loaded. |
| 22:03.37 | Bally3 | thanks mate :) |
| 22:03.44 | druidu | np |
| 22:03.45 | druidu | hmm... but still, request_irq does not return |
| 22:03.54 | druidu | so the device doesn't actually lock |
| 22:04.01 | druidu | it's just overloaded |
| 22:04.16 | druidu | irq processing eats up all the cpu |
| 22:05.45 | druidu | Bally3: actually, I'm not sure which one of those is the wifi irq |
| 22:07.10 | Bally3 | I got a long print out from haret console ..after turning off and on wifi in winmo.. and saw where it spiked.. but dont know how to get the irq from that |
| 22:07.48 | druidu | but which one is the wifi driver anyway? |
| 22:07.58 | druidu | I don't see it anywhere around there |
| 22:07.58 | Bally3 | wlan.ko |
| 22:08.09 | Bally3 | oh |
| 22:08.25 | druidu | btw http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=commit;h=d7fd4ed7de626c07d253d28ee194c89f3a593133 |
| 22:08.31 | druidu | aaa, it's a separate module |
| 22:08.34 | druidu | right |
| 22:09.34 | *** join/#htc-linux x29a (n=x29a@unaffiliated/x29a) |
| 22:09.39 | druidu | CSR unifi-105x WiFi device |
| 22:09.45 | druidu | or CSR unifi-60xx WiFi device |
| 22:09.45 | druidu | ? |
| 22:09.58 | Bally3 | you see that? |
| 22:10.30 | druidu | that what? |
| 22:10.30 | druidu | http://android.git.kernel.org/?p=platform/vendor/csr/unifi-105x.git;a=tree;f=unifi_hostsw_linux_147;h=feba699ef19feafea13319b582c032ead2816f9c;hb=HEAD |
| 22:11.05 | Bally3 | sorry where did you see the above? |
| 22:15.47 | druidu | dunno, found it in the android projects |
| 22:15.52 | druidu | but I don't think it's that one... |
| 22:16.04 | druidu | might be the tools & stuff, but not the driver itself |
| 22:17.48 | cr2_ | /* Read Only register */ |
| 22:17.49 | cr2_ | #define UARTDM_MISR_ADDR 0x10 |
| 22:18.10 | Bally3 | dont think it is either druidu |
| 22:18.14 | Bally3 | never heard of it :S |
| 22:18.51 | Bally3 | the kaiser and the polaris share the same firmware and driver except irq's are different as the kaiser has a keyboard.. the kaisers irq for wifi is "93" |
| 22:20.13 | cr2_ | #define UARTDM_ISR_ADDR 0x14 |
| 22:20.20 | Bally3 | I now have the enviable task of trying to figure out and try he Polaris irq :S |
| 22:25.24 | cr2_ | druidu: i don't understand how it can't return |
| 22:26.12 | cr2_ | if isr_status == 0 it will just fall through |
| 22:26.37 | druidu | yeah |
| 22:26.43 | druidu | but then it fires again |
| 22:26.46 | druidu | and again |
| 22:26.55 | druidu | it's eating up all the cpu |
| 22:27.02 | cr2_ | ok |
| 22:27.06 | druidu | and the _isr just returns _HANDLED |
| 22:27.25 | druidu | but it doesn't handle the condition that generated the irq in the first place |
| 22:27.52 | druidu | if it's an irq that's saying you can send data now, it's gonna keep firing (since we didn't disable it nor told it not to notify us of that) |
| 22:28.26 | druidu | one thing I'm not 100% sure if for first irq request it returns 0 too |
| 22:28.30 | druidu | it was too fast to see |
| 22:29.14 | cr2_ | but why is the _isr=0 in the first place ?? |
| 22:29.27 | druidu | isr_status |
| 22:29.32 | cr2_ | i think at the beginning there was some dma error messages |
| 22:29.36 | cr2_ | yes |
| 22:29.49 | cr2_ | i need to grep through my logs |
| 22:30.43 | cr2_ | <PROTECTED> |
| 22:30.56 | cr2_ | is it trigger_high in wince ? |
| 22:31.04 | druidu | don't know... |
| 22:31.18 | druidu | I'm testing with latest upstream 2.6.27 now |
| 22:31.36 | cr2_ | haret does not parse the gpio irq config i think |
| 22:31.51 | cr2_ | you may change HIGH to re/fe maybe |
| 22:32.39 | druidu | it's irrelevant on the trigger I guess... the fact is something does change |
| 22:32.44 | druidu | and we don't handle it |
| 22:32.53 | druidu | let's try to make it work with not irq |
| 22:32.55 | druidu | busy waiting :) |
| 22:33.14 | druidu | just let it read and write from the port, raw mode |
| 22:34.43 | tmzt | sl: what device? diam cdma? |
| 22:35.00 | druidu | wohooo, bootez with latest changes |
| 22:35.06 | cr2_ | the trigger HIGH is strange anyway. |
| 22:35.09 | druidu | diam gsm |
| 22:44.03 | tmzt | 13:37 < druidu> [ 6.150367] warning: `rild' uses 32-bit capabilities (legacy support in use) |
| 22:44.38 | tmzt | no, linux capabilities |
| 22:44.44 | tmzt | securoty thing |
| 22:53.10 | druidu | I didn't understand that warning |
| 22:53.17 | druidu | it *is* a 32 bit system |
| 22:53.20 | druidu | why legacy support? |
| 22:53.36 | druidu | 32bits capabilities vs what |
| 22:53.37 | druidu | ? |
| 23:08.44 | *** join/#htc-linux BHSPitMonkey (n=stephen@unaffiliated/bhspitmonkey) |
| 23:11.18 | *** join/#htc-linux tsdogs (n=tsdogs@net203-187-146.mclink.it) |
| 23:13.39 | cr2_ | druidu: i still don't understand why is it triggered high |
| 23:13.52 | cr2_ | maybe try "rising" ? |
| 23:14.32 | druidu | let's look at it this way... g1 uses same chip and the driver works, so well maybe our problem is with something else |
| 23:14.43 | druidu | hmmm |
| 23:14.55 | druidu | actually I see what you mean |
| 23:14.58 | cr2_ | this is something to be compared with wince too |
| 23:15.27 | tmzt | on the caps tjing, it's just a syscall with a 32bit field |
| 23:15.37 | cr2_ | ok |
| 23:15.51 | tmzt | which has changed to a 64bit fields to support more caps |
| 23:16.22 | tmzt | 20:15 < tmzt> sl: what device? diam cdma? |
| 23:16.30 | tmzt | I meant slawson: |
| 23:17.29 | cr2_ | i think wince irqs are configured with a kioctl. and the parameter meaning is not known to me. |
| 23:17.42 | cr2_ | but we can still compare with the other irqs |
| 23:17.59 | *** join/#htc-linux MethoS- (n=clemens@dyndsl-085-016-167-109.ewe-ip-backbone.de) |
| 23:18.41 | tmzt | 20:57 < cr2_> i think wince irqs are configured with a kioctl. and the parameter meaning is not known to me. |
| 23:18.42 | *** join/#htc-linux marajin_ (n=marajin@87-194-102-189.bethere.co.uk) |
| 23:19.26 | cr2_ | #define VIC_INT_TYPE1 VIC_REG(0x0044) /* 1: EDGE, 0: LEVEL */ |
| 23:19.31 | cr2_ | #define VIC_INT_POLARITY1 VIC_REG(0x0054) /* 1: NEG, 0: POS */ |
| 23:19.51 | cr2_ | druidu: maybe dump these regs with bt enabled ? |
| 23:20.17 | *** join/#htc-linux Tinyboom_ (n=nahh@41.80-202-154.nextgentel.com) |
| 23:20.20 | *** join/#htc-linux mrb (n=mrb@160.89-10-30.nextgentel.com) |
| 23:22.08 | cr2_ | #define MSM_VIC_PHYS 0xC0000000 |
| 23:22.27 | cr2_ | pd 0xc0000044 4 |
| 23:22.37 | cr2_ | <PROTECTED> |
| 23:22.50 | *** join/#htc-linux rafyvitto (n=rafyvitt@24.54.253.34) |
| 23:23.10 | cr2_ | #define INT_UART2DM_RX (32 + 20) |
| 23:23.11 | cr2_ | #define INT_UART2DM_IRQ (32 + 21) |
| 23:23.22 | druidu | yeah, I know |
| 23:23.25 | druidu | 2nd one is 53 |
| 23:23.57 | cr2_ | 20 and 21 bit |
| 23:24.39 | druidu | what does VIC stand for? |
| 23:25.23 | ali1234 | something interrupt controller |
| 23:25.30 | cr2_ | vector |
| 23:25.36 | druidu | ah |
| 23:25.41 | druidu | I knew that :D |
| 23:26.09 | cr2_ | druidu: dump the regs |
| 23:26.49 | tmzt | vectored |
| 23:27.03 | tmzt | same as we would on pxa, cool |
| 23:27.12 | tmzt | no tracing ce calls |
| 23:35.51 | druidu | just a moment |
| 23:37.34 | druidu | with BT off |
| 23:37.35 | druidu | HaRET(2)# pd 0xc0000044 4 |
| 23:37.35 | druidu | c0000044 | 0010003c | <... |
| 23:37.35 | druidu | HaRET(3)# pd 0xc0000054 4 |
| 23:37.35 | druidu | c0000054 | 00000000 | .... |
| 23:38.07 | cr2_ | polarity is pos |
| 23:38.28 | cr2_ | 0,4,8,12,16,20,24,28 |
| 23:38.30 | druidu | HaRET(9)# pd 0xc0000044 4 |
| 23:38.30 | druidu | c0000044 | 0010003c | <... |
| 23:38.30 | druidu | HaRET(10)# pd 0xc0000054 4 |
| 23:38.30 | druidu | c0000054 | 00000000 | .... |
| 23:38.33 | druidu | BT on -- same |
| 23:38.45 | tmzt | what happens if you change it in ce? |
| 23:38.57 | tmzt | shouldn't it lockup? |
| 23:39.03 | druidu | dunno |
| 23:39.06 | cr2_ | 20 is EDGE, 21 is level |
| 23:39.33 | cr2_ | #define INT_UART2DM_RX (32 + 20) <- edge |
| 23:39.49 | cr2_ | #define INT_UART2DM_IRQ (32 + 21) <- level |
| 23:39.56 | cr2_ | which one do we have ? |
| 23:40.30 | cr2_ | 20 is rising, 21 is high |
| 23:41.32 | druidu | bit 20 is 1, 21 is 0 |
| 23:41.53 | druidu | true |
| 23:47.57 | druidu | weeeeee, wince dmesg really rulez |
| 23:48.50 | druidu | I just can't find a logical explanation for the reason why microsoft made firmware with all that debug info, and shipped dlls compiled with debug info |
| 23:49.47 | tmzt | they didn't? |
| 23:49.52 | cr2_ | it's htc, not m$ |
| 23:51.29 | druidu | well, don't know about this abut mine, but xperia sent me some dlls and they had debug info in them |
| 23:52.06 | druidu | s/this abut //g |
| 23:52.26 | druidu | yeah, true |
| 23:52.33 | druidu | htc actually released some sources |
| 23:53.26 | druidu | developer.htc.com - android kernel sources for dream and magic |
