00:00.23 | cr2 | i'll commit the latest one to the hh.org CVS |
00:01.53 | cr2 | 1213 HTC Hermes |
00:03.40 | Kevin2 | cr2: I'd like to pull in the new mach-types.h file - where is the source? |
00:03.41 | *** join/#htc-linux Ralith (n=ralith@soggy202.drizzle.com) |
00:05.09 | pof | cr2: do u have any idea about why it hangs when we copy SPL into memory ? |
00:05.18 | *** join/#htc-linux ljp (n=lpotter@203.94.178.46) |
00:05.41 | cr2 | arch/arm/tools/mach-types |
00:05.46 | cr2 | update from hh.org |
00:06.20 | cr2 | pof: you allocate a buffer in RAM, put the SPL there. |
00:06.40 | cr2 | it is relocated to RAMSTART+0x0 |
00:06.47 | pof | we copy it into 0x30000000 |
00:06.55 | cr2 | and jump ther. |
00:07.12 | cr2 | do you deal with the tags ? |
00:07.17 | pof | yes this is the addres of RAMSTART AFAIK |
00:07.21 | cr2 | i.e. have you removed that code. |
00:07.24 | pof | we removed the 'tags' part |
00:07.33 | pof | and the initrd part too |
00:07.37 | cr2 | ok. where does it hang ? |
00:07.42 | cr2 | ok. |
00:07.50 | cr2 | do you know the vibra gpio ? |
00:07.54 | pof | see here http://pof.eslack.org/hermes-unlocker/haret/bootimage.tar.gz |
00:08.01 | pof | no |
00:08.13 | pof | do you mean, to debug ? |
00:09.00 | cr2 | yes. |
00:09.14 | cr2 | it's just 2 asm commands. |
00:09.15 | pof | we made it boot to 0x0 and if it booted then we knew it did not hang, and otherwise we know it hungs |
00:09.46 | pof | which commands ? |
00:10.06 | cr2 | look at the eten-g500 patch. |
00:10.10 | pof | this was only way we could think of to debug :) |
00:10.29 | cr2 | lkcl has put some gpios into wiki |
00:10.30 | pof | yes... we "ported" this patch to old haret code we used to start :) |
00:10.47 | pof | yes, I saw it... but we did this before :( |
00:10.59 | cr2 | http://wiki.xda-developers.com/index.php?pagename=HermesGPIO |
00:11.05 | cr2 | but no vibra there. |
00:11.18 | cr2 | i have tested the "nled" command, it works. |
00:11.38 | cr2 | so we need to convince Kevin to add gpio trace now :) |
00:11.56 | lkcl | for hermes? |
00:12.00 | cr2 | yes. |
00:12.13 | lkcl | yep. only a couple. |
00:12.31 | cr2 | and dump gpio, but i need to read the manual first. afaik it is not implemented. |
00:12.50 | cr2 | lkcl: i have registers htchermes |
00:12.56 | lkcl | excellent |
00:13.02 | cr2 | s/ers/ered/ |
00:13.19 | lkcl | i understood :) |
00:13.28 | cr2 | what else do we have ? greenphone :) |
00:13.45 | lkcl | yep - greenphone. |
00:14.09 | lkcl | i think i need to get an htc magician, and help asylumed with the sound, check the uda1380 on that. |
00:14.23 | cr2 | magician ? |
00:14.35 | cr2 | pH5 has 2. |
00:14.47 | cr2 | but the second is dead. |
00:15.17 | lkcl | oops |
00:15.29 | lkcl | oh. yes. it was pH5 not asylumed. |
00:16.13 | Kevin2 | cr2 / pof: New version of haret at http://handhelds.org/~koconnor/haret/ |
00:16.27 | Kevin2 | This version will just detect the wizard / hermes machines. |
00:16.47 | Kevin2 | Note, we need to register a mach type for wizard. |
00:17.36 | pof | thanks Kevin2, will try right now :) |
00:18.00 | cr2 | Kevin2: i'll check http://linwizard.sf.net |
00:18.12 | cr2 | eten g500 is registered |
00:18.16 | Kevin2 | There is a way to watch gpios on hermes - use the "addwatch P2V(0xaddr)" on all the registers. |
00:18.40 | cr2 | yes, but it's not user-friendly :) |
00:18.43 | Kevin2 | It's a little low-level, but should have the same effect as the code in s3 haret. |
00:19.31 | cr2 | Kevin2: are there ALT gpio states. imho should be. then "dump gpio" is welcomed. in wiki format :) |
00:19.59 | cr2 | Kevin2: currently "dump gpio" does not see that it's not on a non-pxa cpu. |
00:20.33 | cr2 | i vote for a wiki-formatted output. saves a lot of hassle and manual editing. |
00:20.39 | Kevin2 | cr2: Yes - the command should check if it is PXA during registration. |
00:21.06 | Kevin2 | cr2: Let me know when the patch is ready. :-) |
00:21.13 | cr2 | ok. :) |
00:21.35 | pof | i love u guys :D |
00:21.37 | cr2 | it should be easier than the current 4 column hack anyway. |
00:22.51 | cr2 | ak4641 is a cool chip. with the right driver you can record the phone. |
00:22.59 | cr2 | pof: is it possible on wince ? |
00:23.00 | lkcl | cool! |
00:23.36 | cr2 | lkcl: even the BT headset. |
00:23.43 | lkcl | ooooo |
00:24.18 | cr2 | you can also play mono sound to the BT |
00:24.20 | pof | cr2: i don't know :? |
00:24.34 | pof | new haret version detects hermes correctly, log here: |
00:24.36 | pof | http://pof.eslack.org/tmp/haretlog.txt |
00:25.24 | cr2 | pof: check with the wince guyz. the hardware allows that. whether lame duck wince lets you do that is unknown ;) |
00:26.04 | pof | i know there is soft to record conversations, but i don't know if it is implemented directly using the ak4641 chip or by any other API... |
00:26.22 | pof | let me search in xda-dev, i remember reading an article about it |
00:26.45 | cr2 | api is ok too. it can be traced too. |
00:27.42 | cr2 | Kevin2: you have hardcoded \Storage Card ? |
00:28.51 | Kevin2 | cr2: No. Haret finds which directory it starts in and uses that as a "current directory". |
00:29.50 | pof | cr2: http://forum.xda-developers.com/showthread.php?t=284953 |
00:30.09 | cr2 | Kevin2: ok. |
00:30.37 | pof | http://forum.xda-developers.com/showthread.php?t=276351 --> check this one too |
00:31.03 | cr2 | pof: it's on hermes. i meant the universal. on hermes we need to trace the i2c first. |
00:31.45 | cr2 | LOL. |
00:31.48 | cr2 | It's available here and costs $9.95. |
00:32.46 | cr2 | <PROTECTED> |
00:32.58 | pof | :P |
00:33.54 | pof | http://www.softtrends.com/Downloads/LIVEPVR_WM5_PPC_TRIAL.zip |
00:34.05 | pof | there's a free trial :) |
00:34.22 | cr2 | i'm too disconnected from the wince world :) |
00:34.52 | Hymie | cr2: GOOD |
00:34.52 | pof | you still have an hermes in your pocket... which doesn't run linux (yet!) ;) |
00:35.10 | NK` | cr2: , |
00:35.14 | NK` | ? |
00:36.08 | Hymie | lkcl: DON'T DO IT MAN! |
00:36.29 | cr2 | i prefer "free software" to "free trial" ;) |
00:36.56 | pof | so do we |
00:37.30 | Hymie | for me, the big thing is access to the code.. I prefer GPL over closed copyright.. but really? I don't mind a closed copyright, if I have access to the sources |
00:37.42 | Hymie | it's closed source that burns by britches |
00:37.45 | *** join/#htc-linux teslamike (n=teslamik@CPE0011d864891e-CM0011e6c7aaa5.cpe.net.cable.rogers.com) |
00:37.54 | lkcl | mwahaaha |
00:38.59 | teslamike | Hello. |
00:39.36 | teslamike | I have an HTC BlueAngel and would _really_ like to get linux running on it. I'm willing to help out with debugging and code, but I have no idea where to start. |
00:40.14 | pof | teslamike: see the url's in channel topic :) |
00:40.21 | teslamike | As it is, I know nothing about writing code for the device (not even which compiler to use), but I'm willing to put in the time to learn if anyone can help point me in the right direction. |
00:40.40 | teslamike | pof: Been reading the last 2, just getting to the first now. |
00:41.09 | lkcl | teslamike: cool |
00:41.35 | teslamike | The first link just looks like IRC logs? |
00:42.29 | pof | yes, check also http://www.gnulinux.biz/ |
00:43.31 | pof | http://www.gnulinux.biz/files/blueangel/sd/ --> here you've initrd, kernel... for blueangel |
00:43.57 | pof | 1 year old, but can serve you as a start :) |
00:44.18 | Hymie | TOMATO AND MOZZARELLA FLAVOURED PRINGLES !!! :( They don't have that here :((( |
00:44.20 | Hymie | damn |
00:44.29 | pof | cr2: did u see my query ? |
00:44.47 | pof | Hymie: LOL :D |
00:46.46 | teslamike | pof: You'd suggest I start off just trying to run it eh? |
00:47.07 | pof | yes, setup the partitions in SD card and run haret |
00:47.15 | teslamike | pof: From what I've read haret and WM5 don't get along well. Is that still true? |
00:47.33 | pof | teslamike: for blueangel WM5 yes |
00:48.06 | pof | if you have WM5 rom on your blueangel you'll need to go back to wm2003SE again if you want to run linux on it |
00:48.25 | pof | on other devices there's no problem with WM5... it'Âs just blueangel :( |
00:48.30 | teslamike | really? |
00:48.32 | teslamike | any idea why? |
00:48.48 | pof | no idea... i wonder the same :? |
00:48.58 | teslamike | lol |
00:49.09 | teslamike | traditional microsoft wizardry it is then! |
00:49.19 | Kevin2 | I think BabelOued has got BA working with wm5. |
00:49.31 | pof | and he can run linux on it ?? |
00:50.03 | pof | i got mine with helmi's WM5 aku 3.2 rom... last time i tried it was not possible |
00:51.39 | *** join/#htc-linux Ralith (n=ralith@soggy202.drizzle.com) |
00:52.57 | teslamike | has anyone tried completely switching to linux? |
00:53.04 | teslamike | no haret at all? |
00:53.20 | teslamike | or is that just not possible at this point? |
00:53.51 | lkcl | teslamike: what's required to do that is to have access to the flash memory chip - read and write. |
00:54.14 | lkcl | the HTC devices use proprietary m-systems devices (blueangel has the 3rd generation device, called duh g3) |
00:54.43 | teslamike | wouldn't you also need this same access to update windows? |
00:54.46 | lkcl | the 2nd revision - g2 - aka millenium plus - was successfully reverse-engineered a few months ago. nearly three years after it first came out. |
00:54.58 | teslamike | Ouch. |
00:55.11 | lkcl | well actually the driver had been working some time ago |
00:55.22 | lkcl | but that was an older driver |
00:55.46 | lkcl | and the 'newer' driver - as usual - didn't take into account the 'interleave' of two 16mb chips. |
00:56.55 | lkcl | hmm my uda1380 isn't detected by i2cdetect any more. |
00:56.56 | teslamike | so you're saying we had this read/write access on blueangel until the driver was replaced by one that doesn't work? |
00:57.15 | lkcl | correct! but it was like _really_ early 2.6 kernels. |
00:57.56 | teslamike | that sounds a little shitty if you ask me. |
00:59.27 | teslamike | so, i've never written code for an embedded device before, and i've never worked with device drivers before. my only low-level experience is in MIPS, and only dealing with a cpu and some ram. any good places to start reading up on how to write drivers, and how to interact with other hardware? |
01:01.29 | teslamike | sorry for all the questioning. i just have no idea where to get started and i'd like to help with development |
01:08.53 | teslamike | alright, i'll come back when i'm ready to get linux running and i've figured out more about hardware basics... |
01:10.14 | pof | teslamike: tomorrow mornint this will be more active, more people will be able to help you out |
01:10.24 | pof | i guess everyone is sleeping now :P |
01:10.37 | pof | check handhelds.org, it has great documentation |
01:12.57 | teslamike | pof: thanks! |
01:26.21 | lkcl | ok i have a clue... |
01:26.29 | lkcl | i loaded i2c-pxa first. |
01:26.32 | lkcl | waited for a bit |
01:26.39 | lkcl | then i loaded blueangel_audio... |
01:32.48 | lkcl | oh jesus - asic3 has alternate functions??? |
01:32.50 | lkcl | oh maaaan. |
01:36.27 | lkcl | huh. loading the blueangel_audio module screws with the i2c bus, ha ha. |
02:05.57 | *** join/#htc-linux LunohoD (n=alex@e180064065.adsl.alicedsl.de) |
04:36.52 | *** join/#htc-linux RoEn_PC (n=roen@p54A66BD1.dip.t-dialin.net) |
04:41.09 | *** join/#htc-linux Ro_En_PC (n=roen@p54A67259.dip.t-dialin.net) |
05:32.33 | *** join/#htc-linux psokolovsky_ (n=psokolov@237.usernat.ip.net.ua) |
06:48.51 | *** join/#htc-linux goxboxlive (n=jrs@gw.hjelcowi.no) |
07:44.57 | *** join/#htc-linux rob_w (n=rob_w@p549B9EDC.dip0.t-ipconnect.de) |
08:07.47 | *** join/#htc-linux Fossi (n=cmewes@195.177.48.115) |
08:27.15 | *** join/#htc-linux rmoravcik (n=rmoravci@pc-3s0zt5w2e4y0vzmhnrzq3a21zqajzfw.users.student.utc.sk) |
09:59.21 | *** join/#htc-linux babeloued (n=Fabrice@faim1.itesoft.com) |
10:02.18 | *** join/#htc-linux bd2 (n=ufo@89.113.84.151) |
10:45.19 | *** join/#htc-linux booba (n=booba@AAnnecy-256-1-5-119.w90-10.abo.wanadoo.fr) |
10:55.12 | lkcl | ok - but i can do an i2cdump 0 0x12 and that shows data, so it's there. |
10:55.50 | babeloued | hi lkcl :) |
10:55.55 | lkcl | allo babeloued |
10:56.11 | babeloued | what do you have at 0x12 ? |
10:59.44 | lkcl | o yeh. sorry. 0x12 is the ak4641. 0x18 is the uda1380 |
11:00.15 | lkcl | hm battery charging on the blueangel with linux doesn't work unless you load the modules, does it? |
11:00.26 | lkcl | i just totally flattened the battery overnight |
11:00.37 | lkcl | so i got to wait 20 mins for it to charge up a bit... |
11:00.55 | babeloued | yes i have same problem. |
11:01.29 | babeloued | but it is better to have a working battery monitor module before ? |
11:03.38 | babeloued | i don t know if there charger security is software or harware |
11:27.18 | lkcl | ok - damnit. sound actually works after a wince boot. |
11:27.21 | lkcl | just like last year. |
11:27.52 | lkcl | ok - so something in the configuration is missing. |
11:27.58 | lkcl | what the xxxx is it??? :) |
11:28.16 | lkcl | that survives a haret boot??? |
11:28.20 | *** join/#htc-linux lamikr (n=chatzill@aragorn.kortex.jyu.fi) |
11:29.24 | lkcl | where's that memory-dump program? |
11:29.27 | lkcl | for linux? |
11:30.33 | lkcl | x.c or something silly |
11:31.07 | lkcl | ok i found it. |
11:38.35 | *** join/#htc-linux cr2 (n=konversa@crpl22.physik.uni-wuppertal.de) |
11:42.05 | lkcl | ok with that program x i've done a memory dump of gpio, asic3, ssp and i2c. |
11:42.31 | lkcl | now i'm going to reboot, _not_ play a sound when running haret, and then compare. |
11:44.26 | lkcl | blueangel has four serial ports, one at 0x41600000??? |
11:44.37 | lkcl | oops damnit. battery's not charged up enough yet. |
11:44.51 | lkcl | nuts. |
11:45.31 | lkcl | ok am gonna get some breakfast. |
12:02.09 | *** join/#htc-linux asbjorn (n=asbjorn@2.81-167-1.customer.lyse.net) |
12:52.50 | *** part/#htc-linux goxboxlive (n=jrs@gw.hjelcowi.no) |
13:15.10 | *** join/#htc-linux epinephrine (i=geek@82-37-152-150.cable.ubr01.wals.blueyonder.co.uk) |
13:18.46 | *** join/#htc-linux psokolovsky_ (n=psokolov@237.usernat.ip.net.ua) |
14:28.07 | *** part/#htc-linux babeloued (n=Fabrice@faim1.itesoft.com) |
14:50.00 | lkcl | ok i found some uda1380 i2c registers that might not be being set. |
14:53.37 | lkcl | where the _heck_ is the volume set???? |
14:54.38 | lkcl | ok i got it... |
14:55.38 | *** join/#htc-linux babeloued (n=Fabrice@faim1.itesoft.com) |
15:11.41 | lkcl | ah fricking xxxxing hell. |
15:11.52 | lkcl | i2c 'master mute' bit is set on the uda1380. |
15:18.35 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
15:21.55 | lkcl | huh. uda1380 isn't responding to mute 'off' command. odd. |
15:23.13 | lkcl | that's really weird |
15:34.52 | *** join/#htc-linux WizMaui (n=WizMaui@62.112.90.231) |
15:55.39 | *** join/#htc-linux Ralith (n=ralith@66.150.12.86) |
16:01.09 | lkcl | argh ok i need to do an i2c register dump (wi) from wince - anyone got any clues on what to do? |
16:01.30 | lkcl | do we have any i2c data decoding scripts? didn't cr2 write one last month? |
16:21.40 | lkcl | ok i found this |
16:21.42 | lkcl | http://pastebin.ca/159143 |
16:27.44 | *** join/#htc-linux pH5 (n=ph5@e178231142.adsl.alicedsl.de) |
16:32.43 | *** join/#htc-linux JTRipper86 (n=jtripper@p54B1D177.dip.t-dialin.net) |
16:38.15 | lkcl | okayyyy.... this is really weird. the uda1380 flatly refuses to accept changes to master-mute. |
16:38.30 | lkcl | from linux. |
16:38.34 | pH5 | lkcl: please explain |
16:38.53 | pH5 | I have a similar problem - it seems to ignore writes to every register but 0x00 to 0x04 |
16:38.57 | lkcl | ok - as usual, i'm booting into linux whilst playing a sound. |
16:39.22 | lkcl | ok well that could be due to the way that the i2c writes are done |
16:39.26 | lkcl | this is register 0x13 |
16:39.41 | lkcl | wouldn't that be ironic... |
16:40.18 | *** part/#htc-linux epinephrine (i=geek@82-37-152-150.cable.ubr01.wals.blueyonder.co.uk) |
16:40.22 | pH5 | lkcl: can you write other registers with the i2c tools from the lmsensors package? |
16:40.27 | lkcl | no. |
16:40.43 | lkcl | i2cset 0x0 0x18 0x08 w |
16:40.43 | lkcl | fails. |
16:41.00 | lkcl | it double-checks the response and finds that it's the same. |
16:41.13 | pH5 | that is exactly my problem |
16:41.22 | lkcl | looking at the data sheet it says you're supposed to wait for an acknowledgement |
16:41.34 | lkcl | but setting volume, which is a 0x20 register, is ok. |
16:41.44 | lkcl | have you been looking at this code for a long time? |
16:42.22 | pH5 | define long time |
16:42.27 | lkcl | weeks? |
16:42.30 | pH5 | no |
16:42.32 | lkcl | ah. |
16:42.50 | pH5 | it's more like "week" |
16:42.58 | lkcl | basically i was wondering if you knew if there was an acknowledgement from the i2c bus devices - every device |
16:44.15 | lkcl | the uda1380 data sheet mentions 'acknowledgement' - it'd be good to know that that was part of the hardware spec, rather than something that just the uda1380 does. |
16:45.01 | pH5 | lkcl: I thought this is handled in the i2c layer - it "asserts an acknowledge (A)" after every byte sent, whatever that means |
16:45.06 | pH5 | I should learn how i2c works |
16:45.08 | lkcl | ok. good. |
16:48.16 | lkcl | huh. don't tell me... uhn??? wtf??? the uda1380 has two modes, or something, with two addresses. |
16:49.03 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
16:49.39 | lkcl | which explains the 0x18 / 0x1e funny addressing thing |
16:50.06 | pH5 | it is common to have two different addresses available |
16:50.15 | pH5 | this has nothing to do with the two control modes (L3-bus or i2c) |
16:51.10 | lkcl | huh. |
16:51.36 | pH5 | I think you can select the address by driving some pin high or low |
16:52.02 | lkcl | why isn't the read-write mode bit set to 0 on write, in i2c_pxa_do_xfer |
16:53.29 | *** part/#htc-linux babeloued (n=Fabrice@faim1.itesoft.com) |
16:54.20 | pH5 | lkcl: I think this is handled in i2c_pxa_addr_byte |
16:54.51 | pH5 | or are you talking about the TB bit? |
16:55.14 | lkcl | #define ISR_RWM (1 << 0) /* read/write mode */ |
16:55.20 | lkcl | it's only read - it's never set. |
16:55.51 | pH5 | isr is a read-only register, right |
16:55.58 | pH5 | this must be for slave mode or something |
16:57.55 | lkcl | hm. ok. ok, you're right. |
16:58.00 | lkcl | there's a DOM - data operation mode. |
16:59.20 | lkcl | hmm. |
17:02.14 | lkcl | ok apparently you're supposed to set an extra bit - bit 1 - when doing a write. |
17:02.43 | *** join/#htc-linux psokolovsky_ (n=psokolov@237.usernat.ip.net.ua) |
17:02.45 | lkcl | i'll give it a shot... |
17:02.49 | lkcl | allo psokolovky |
17:02.57 | lkcl | am having fun with i2c on uda1380 |
17:03.32 | pH5 | hey psokolovsky_ |
17:03.45 | psokolovsky_ | Hi! |
17:05.29 | lkcl | we got a weird one for you, psoko. uda1380 i2c registers are unreliable in being written to (!) |
17:05.34 | lkcl | some work. some don't |
17:12.38 | lkcl | ok. what else can i try... |
17:13.06 | lkcl | i can try looking at other memory locations on the i2c bus registers... |
17:13.18 | pH5 | lkcl: can you soft-reset the uda1380? |
17:13.28 | lkcl | eep! yes? you know how? |
17:13.36 | lkcl | there's a configure command isn't there... |
17:13.46 | pH5 | send 0x00 to 0x7f, look at uda1380.c |
17:13.52 | lkcl | i've seen some code to do that... |
17:13.53 | lkcl | oh ok. |
17:14.08 | lkcl | arse i gotta reboot into linux to do it :) |
17:14.37 | pH5 | and if that doesn't work, perhaps hard-reset. I know how to do a hard reset, but I'm sure I#m not doing it correctly - the specs say somehting about the clock that must be running during the reset |
17:15.59 | lkcl | ahh that would explain why ssp CKEN is set in between. |
17:16.04 | lkcl | ok. great. |
17:16.18 | *** join/#htc-linux BabelOued (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
17:16.23 | lkcl | that's good to hear, pH5 |
17:17.30 | lkcl | it explains the asic3gpio traces i've seen |
17:23.02 | *** join/#htc-linux g3gg0_ (n=g3gg0@ppp-88-217-6-241.dynamic.mnet-online.de) |
17:25.54 | lkcl | huh??? i'm looking at a trace of a large file being played, and on the face of it, it appears that the uda1380 is being hard-reset several times a second! |
17:25.56 | psokolovsky_ | lkcl, sorry, I'm not really a sound stuff expert. are you bother yourself with asoc? |
17:27.02 | pH5 | lkcl: wha..? |
17:27.03 | lkcl | psokolovsky: i've gone back to 'known state', to eliminate some of the unknowns. |
17:27.03 | lkcl | asoc is unknown. |
17:27.03 | lkcl | 2.6.16 and above is unknown (messed up). |
17:27.08 | pH5 | that can't be true. the power-on reset takes at least one µs. |
17:27.11 | lkcl | i'm back with the code from over a year ago - 2.6.12 xanadux sf.net - which we know how to at least get sound playing |
17:27.36 | psokolovsky_ | lkcl, as you probably no, pxa-i2sound works perfectly for uda1380 on all versions. |
17:27.49 | lkcl | once i know what's going on, then i can report on how it can be fixed |
17:28.09 | lkcl | psokolovsky: the uda1380 on the blueangel is locking 'mute' on. |
17:28.20 | lkcl | we never managed to get to use pxa-i2sound for the blueangel. |
17:28.33 | psokolovsky_ | lkcl, damn, I did everything to stop hearing that .12 whine, but guess it's just unhelpable ;-))) |
17:28.35 | pH5 | psokolovsky_: blueangel and magician use ssp for sound output, not i2s |
17:28.48 | psokolovsky_ | ah, that sucks |
17:28.54 | pH5 | indeed :) |
17:29.05 | lkcl | oh. that'd explain why i couldn't get it to work, ha ha |
17:29.23 | lkcl | oh. and that would... ahh... |
17:29.35 | lkcl | let me just go try something. |
17:29.42 | pH5 | i2s is used for input only |
17:29.52 | lkcl | oh. |
17:29.52 | lkcl | no. |
17:29.58 | lkcl | it isn't. |
17:30.00 | lkcl | really? |
17:30.06 | pH5 | really |
17:30.11 | pH5 | watch the clocks with haret |
17:30.16 | lkcl | *sigh*. ok. |
17:30.26 | pH5 | playing --> cken ssp(1), recording --> cken i2s |
17:30.45 | lkcl | o good grief :) |
17:30.54 | lkcl | *sigh* ok let me try something... |
17:32.41 | lkcl | ok i should do a memory-dump of the clocks as well. |
17:32.49 | lkcl | in linux. when sound is working. |
17:33.33 | lkcl | i wonder if the uda1380 simply doesn't respond if there's no clock (from cken ssp(1)) |
17:33.40 | lkcl | to that unmute... |
17:34.46 | lkcl | ok - so, let's guess here: what do you reckon, pH5, that they simply use the SSP clock wire... because they can? |
17:35.39 | lamikr | has anyone of you tried http://www.minisip.org/ with HTC phones? |
17:37.10 | pH5 | lkcl: you're genius, look at page 29 of the specs |
17:37.18 | lkcl | page 29, page 29.... |
17:37.27 | pH5 | 11, register mapping |
17:37.42 | pH5 | "Interpolation filter" (registers 10h .. 14h) |
17:37.48 | pH5 | "running at 128fs interpolator clock |
17:37.49 | pH5 | " |
17:37.50 | lkcl | err.... |
17:37.53 | pH5 | at guess what, |
17:37.59 | pH5 | interpolator clock does *not* come from i2c |
17:38.04 | lkcl | err? |
17:38.17 | pH5 | this is driven by the sysclk |
17:38.20 | goxboxlive | lamikr: no i havent . I am buidling it now for testing it. |
17:38.33 | *** join/#htc-linux JTRipper (n=jtripper@p54B1D177.dip.t-dialin.net) |
17:38.40 | lkcl | ohhh, that would explain why the 0x18 register status was wrong. |
17:38.45 | lkcl | oooops. |
17:38.47 | lamikr | goxboxlive: Cool, let me know how it goes. |
17:38.50 | pH5 | ergo, having ssp1 running writing to those registers looks like a sane idea |
17:39.01 | goxboxlive | yes |
17:52.11 | lkcl | status of register 0x18 is 7c00 |
17:53.47 | *** join/#htc-linux kops (n=kops@87.135.103-84.rev.gaoland.net) |
17:56.23 | lkcl | ARGH :) |
17:59.31 | kops | hi |
18:12.50 | pH5 | lkcl: there's nothing wrong with that 0x18 value. why argh? |
18:15.03 | *** join/#htc-linux g3gg0__ (n=g3gg0@host-82-135-95-8.customer.m-online.net) |
18:16.46 | *** join/#htc-linux ellisway (n=ellis@host-87-74-241-174.bulldogdsl.com) |
18:20.40 | lkcl | argh because i want this to damn well work, that's why! |
18:21.09 | lkcl | ok. dump of ssp needed when running sound. |
18:22.30 | pH5 | well, "argh" |
18:22.41 | lkcl | :) |
18:24.07 | lkcl | ok - SSCR0_P1 changes from 0x75f to 0x7df when sound is playing. |
18:24.25 | lkcl | bit 7 |
18:24.49 | lkcl | #define SSCR0_SSE (1 << 7) /* Synchronous Serial Port Enable */ |
18:35.23 | pH5 | wince definitely doesn't use i2s emulation on ssp. I wonder how to make asoc uda1380.c accept anything but i2s daifmt |
18:37.27 | *** join/#htc-linux JTRipper (n=jtripper@p54B1D177.dip.t-dialin.net) |
18:47.24 | *** join/#htc-linux FossiFoo (n=Fossi@e176118052.adsl.alicedsl.de) |
19:13.05 | pH5 | lkcl: I got back to the bad-quality overdrive sound, and indeed when changing values with alsamixer *while playing* at least *something* happens (sudden silence), so we are on the right track |
19:14.34 | pH5 | and oops when I rmmod the sound module with alsamixer opened... |
19:14.40 | pH5 | time to reboot |
19:33.05 | lkcl | ah ha |
19:43.30 | pH5 | unfortunately, I'm now back to no sound at all(tm) after the reboot. confusing. |
20:31.11 | goxboxlive | lamikr: GPE-ERROR: Cant get mixer element 'Phone' What do i have to change source? |
20:37.54 | lamikr | goxboxlive: The errors comes from the init_mixer() method in mixer.c. I have added there a check that the mixer is tried to use only phoneConfig->initMixer == TRUE |
20:38.26 | lamikr | goxboxlive: so you can disable the mixer opening to your device from the configmodel.c |
20:40.56 | lamikr | goxboxlive: The real fix would be to add either "Phone" or "Target" mixer elemenet to your devices alsa-mixer. |
20:41.22 | goxboxlive | I know, but i have enabled it. And it seems like it is activating the speakers, but i dont have any wav file yet, have to add it.. |
20:42.10 | goxboxlive | It also seems like it get info about a incomming sms, but it doesnt show up in gui. "AT+WMUSS=1:+CMGR=3' |
20:42.57 | goxboxlive | Maybe the message storage should only be twice as h6300 and not three times if u know what i mean. |
20:47.14 | *** join/#htc-linux rmoravcik (n=rmoravci@pc-3s0zt5w2e4y0vzmhnrzq3a21zqajzfw.users.student.utc.sk) |
20:53.39 | lamikr | goxboxlive: Yeah, I know... I can try to check your sms at cmd a little bit later. I am just a little bit in hurry. |
20:54.06 | goxboxlive | that sounds great lamikr. |
21:04.06 | lkcl | pH5: have you been loading/unloading modules? |
21:04.44 | lkcl | because if so, it could be that earlier experiments reset the codec and then you of course were able to play sound with a subsequent module... |
21:06.16 | pH5 | oh dear, this is a whole lot more complicated than I like it to be. |
21:07.10 | pH5 | I did change modules. But I didn't touch the reset code. I only played around with the DAI matching, I2S emulation mode and clock settings. |
21:07.56 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
21:10.21 | *** join/#htc-linux BoSJo (n=BoSJo@105.84-48-112.nextgentel.com) |
21:15.19 | *** join/#htc-linux asylumed (n=insanity@196.211.28.91) |
21:53.28 | lkcl | allo asylumed |
21:54.07 | lkcl | well it involves i2s, ssp, unknown gpio start/reset, unknown speaker/mic gpio switching.... are you surprised? |
21:58.40 | lkcl | ok - when sound is played, only bit 7 is set in the SSCR0_P1. |
21:59.08 | lkcl | pH5: you might want to play a long sound in wince, and do a dump pxa27xssp |
21:59.13 | lkcl | then compare it to afterwards. |
21:59.57 | lkcl | when the sound isn't playing |
22:00.11 | lkcl | what else can i look at while i'm here... |
22:01.43 | pH5 | lkcl: been there, done that |
22:01.46 | pH5 | SSCR0 4000009f SSCR1 00201dc0 SSTO 0 SSPSP 0 SSSR 0000f0d8 SSACD 00000011 |
22:02.10 | lkcl | ok so that's when sound is on, right? |
22:03.04 | pH5 | yup |
22:03.23 | pH5 | 0x4000001f otherwise |
22:04.17 | pH5 | if only haret could dump the uda1380 i2c regs.. |
22:06.38 | lkcl | well... that shouldn't be too difficult, directly accessing the registers |
22:06.53 | lkcl | integrating i2cdump |
22:07.21 | pH5 | right, but it's a whole little project on its own.. |
22:09.24 | lkcl | ok another boot.... :) |
22:09.40 | pH5 | lkcl: earlier you said it looks like it is resetting several times a second - are you sure you were watching the correct gpios? |
22:10.14 | lkcl | yeh... A15. |
22:11.03 | lkcl | oops. the i2c bus is _completely_ shagged :) |
22:11.39 | pH5 | lkcl: got liam's mail? |
22:11.48 | lkcl | oops. i think i managed to switch the device off. finally. |
22:12.04 | lkcl | liam, liam... mmmm no? |
22:12.31 | lkcl | 00: XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XXXXXXXXXXXXXXXX |
22:12.31 | lkcl | 10: XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XXXXXXXXXXXXXXXX |
22:12.37 | lkcl | oops! |
22:13.07 | lkcl | ok so i know it doesn't look like it but i think i am getting somewhere. |
22:13.16 | pH5 | bye bye uda1380 :) |
22:13.45 | lkcl | oops - damnit. |
22:14.00 | lkcl | i was checking device 0x12 instead of 0x18 |
22:14.04 | lkcl | o well... |
22:32.10 | goxboxlive | lamikr: I figure out one thing.Universal sends : AT+CNMI=1,2,2,2,0 And for sms storage it sends like h6300: AT+CPMS="SM","SM". |
22:32.20 | pH5 | good night |
22:32.22 | *** join/#htc-linux ljp_ (n=lpotter@203.94.178.46) |
22:32.35 | *** part/#htc-linux pH5 (n=ph5@e178231142.adsl.alicedsl.de) |
22:37.07 | goxboxlive | lamikr: When i try to send a sms it seems to work as it should, but it stops at "sendCommand('AT+CMGS=18') The "18" noumber seems to differ for everytime. |
22:37.09 | goxboxlive | lamikr: When i recive a sms, the screen flashes, but it is register as a call in the gui. When i hit answer it goes straight to the "missing calls" windows and shows a missed call with during 1 sec and with no phonenumber. |
22:37.54 | *** join/#htc-linux iggy (n=iggy@12.45.184.235) [NETSPLIT VICTIM] |
22:46.36 | *** join/#htc-linux iggy (n=iggy@12.45.184.235) |
22:46.36 | *** join/#htc-linux frankie (n=nnnnnnfr@lugbari/people/frankie) [NETSPLIT VICTIM] |
22:46.43 | *** join/#htc-linux iggy_ (n=iggy@gentoo/developer/iggy) |
22:49.38 | *** join/#htc-linux frankie (n=nnnnnnnf@klecker.ba.issia.cnr.it) |
23:11.04 | goxboxlive | lamikr: I also know that they now are stored on the sim card, because when i reboot to wince i recive the sms i have sendt to the Universal when running linux |
23:15.07 | *** join/#htc-linux FossiDDR (n=Fossi@e176118052.adsl.alicedsl.de) |
23:33.33 | *** join/#htc-linux Ralith (n=ralith@soggy202.drizzle.com) |
23:33.43 | *** join/#htc-linux cr2 (n=konversa@crpl22.physik.uni-wuppertal.de) |
23:33.47 | lkcl | allo cr2 |
23:33.52 | cr2 | hi lkcl |
23:34.07 | lkcl | how are ya? |
23:34.18 | cr2 | i must give a talk tomorrow, so i'm perparing the slides. not time for hacking :) |
23:34.45 | cr2 | well, maybe a bit. |
23:35.04 | lkcl | what kind of talk? |
23:36.51 | cr2 | satellite observation of the temperature in the lower thermosphere. |
23:37.23 | cr2 | 120-150km |
23:37.49 | cr2 | maybe 120-180km is a better description. |
23:40.09 | lkcl | bizarre! |
23:40.45 | goxboxlive | hi cr2: You are really late on this days :-) |
23:40.45 | lkcl | nice easy stuff, then |
23:41.09 | cr2 | i do some rocket science is real life, but will probably leave sometime next year, because our group is dissolving and the future financing is unclear. |
23:41.45 | cr2 | goxboxlive: i need to work sometimes too, you know :) |
23:41.50 | goxboxlive | :-) |
23:42.36 | cr2 | the bt headset discovery was very nice. |
23:43.20 | goxboxlive | Well i am trying out gomunicator. I am getting further. Still not able to send sms, but there are something happening when i recive a sms, but it is detected as a call, and not sms. |
23:43.33 | cr2 | what we need now is the real mixer. |
23:43.44 | goxboxlive | Ok, did you get it working? |
23:44.03 | cr2 | i've seen in the log, you are using some strange h6300 SMS commands. |
23:44.21 | cr2 | the i2c/ak4641.c needs some major rewrite. |
23:44.49 | goxboxlive | Well i debugged in wince, and that's what it is saying. |
23:44.56 | cr2 | how good is the alsa mixer code in the gomunicator ? |
23:45.40 | goxboxlive | Well it's not workimg. Lamikr said someting about adding a "Phone" or "Target" mixer elemenet to ourdevice alsa-mixer |
23:46.13 | cr2 | it's too simple. what is "phone" ? |
23:46.18 | cr2 | AT+WMUSS=1 |
23:46.26 | cr2 | it's h6300-specific. |
23:46.39 | goxboxlive | But when i call my self (of course from another phone) i can hear that the headphones is being activated, but no sound. Also dmesg shows active speakers |
23:47.16 | cr2 | the speakers are switched to nowhere. |
23:47.17 | goxboxlive | Phone is declared in the gomunicator source. |
23:48.08 | goxboxlive | You mean AT+WMUX right? I am not using that. |
23:48.09 | cr2 | ak4641 has 2 mics (int ext), BT, I2S, stereo line out, mono in, mono out, mono out2 |
23:48.27 | goxboxlive | wow |
23:48.47 | cr2 | and we need a real mixer support for them. |
23:49.02 | goxboxlive | What do you mean by external? The one on headset right? |
23:49.03 | cr2 | and the phone "AT@AUDIOSET" |
23:49.13 | cr2 | yes. |
23:49.18 | cr2 | 21:42 goxboxlive It also seems like it get info about a incomming sms, but it doesnt show up in gui. "AT+WMUSS=1:+CMGR=3' |
23:49.28 | cr2 | AT+WMUSS=1 ? |
23:50.07 | goxboxlive | hmm, man i am getting senil |
23:51.05 | goxboxlive | ok, i'll try sending me a sms again |
23:53.33 | goxboxlive | well i didnt get any WMUSS now |
23:53.52 | goxboxlive | Maybe beacuse i have added mixer, i'll try to disable it again |
23:56.42 | cr2 | i don't yet understand who sends the external notification to the asic3 that is linked to the mic. |
23:56.53 | cr2 | both in phone and non-phone mode. |
23:57.21 | cr2 | you press the record button, and somebody from outside pulls the external irq. |
23:57.56 | cr2 | how did he know is a puzzle for me now. |
23:59.32 | cr2 | openoffice.org finally has 64bit support :) |