| 00:01.27 | *** join/#htc-linux rafyvitto (n=rafyvitt@adsl-72-50-48-43.prtc.net) |
| 00:13.27 | *** join/#htc-linux swc|666 (n=infidel2@unaffiliated/swc666/x-4934821) |
| 00:17.02 | *** join/#htc-linux [1]Captnoord (n=Captnoor@dc5147a47b.adsl.wanadoo.nl) |
| 00:26.30 | *** join/#htc-linux maejrep[w] (n=madCoder@smtp-n.myyearbook.com) |
| 00:28.48 | *** join/#htc-linux fnord_ (n=fnord@24-151-90-116.static.nwtn.ct.charter.com) |
| 00:49.37 | *** join/#htc-linux TommyBres (n=TommyBre@unaffiliated/tommybres) |
| 00:50.23 | tmzt | tb |
| 00:50.30 | TommyBres | yessir |
| 00:50.32 | tmzt | what's the problem? |
| 00:50.59 | TommyBres | i try to fastboot, "fastboot flash userdata c:\nandroid\data.img" |
| 00:51.54 | TommyBres | and it gives errors like, "Failed (data transfer failure (Too many links))" |
| 00:52.08 | TommyBres | or "FAILED (command write failed (Unknown error))" |
| 00:53.43 | TommyBres | but now i gotta go for a bit... i'll be back later :\ |
| 01:45.23 | *** join/#htc-linux LTxda (n=anon@unaffiliated/ltxda) |
| 01:55.53 | *** join/#htc-linux BabelO (n=fcr@unaffiliated/babelo) |
| 02:10.00 | *** join/#htc-linux replican (n=replican@adsl-074-183-167-249.sip.bhm.bellsouth.net) |
| 02:49.36 | *** join/#htc-linux surge (n=surge@pool-98-118-158-217.bflony.fios.verizon.net) |
| 03:08.57 | *** join/#htc-linux mrmoku|a` (n=mrmoku@ppp-93-104-33-177.dynamic.mnet-online.de) |
| 03:57.47 | *** part/#htc-linux TommyBres (n=TommyBre@unaffiliated/tommybres) |
| 03:59.05 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
| 04:24.42 | *** join/#htc-linux Shinto (n=John@f048048066.adsl.alicedsl.de) |
| 04:57.27 | *** join/#htc-linux droid0011 (n=mc@p4FDCD713.dip.t-dialin.net) |
| 05:45.40 | *** join/#htc-linux Echo31 (n=olivier@mir31-4-82-240-194-54.fbx.proxad.net) |
| 06:06.22 | Echo31 | good morning |
| 06:07.08 | xsacha | hi |
| 06:10.34 | cr2 | morning |
| 06:10.57 | cr2 | Echo31: the ad7877 driver was written by Kevin2 |
| 06:12.29 | Echo31 | cr2: yes, but i used only in the specific boards arch/blackfin/mach-bf548/boards/ |
| 06:13.13 | Echo31 | cr2: I has to implement SPI ? |
| 06:13.23 | cr2 | Echo31: not this one |
| 06:13.49 | cr2 | the pxa2xx_spi driver already exists |
| 06:14.26 | cr2 | Echo31: can you post the full link to blackfin driver ? |
| 06:14.48 | cr2 | eh ? android hardcodes device names ? http://www.netmite.com/android/mydroid/cupcake/system/core/init/devices.c |
| 06:17.29 | xsacha | cr2: im having a problem with my battery. android finds the battery status (eg. charging) but doesnt find capacity :( im just registering all my PROPs as a normal power supply. does it look elsewhere? |
| 06:18.08 | cr2 | xsacha: i'm not an android expert |
| 06:18.12 | xsacha | :( k |
| 06:18.27 | cr2 | xsacha: did you actually trace the raw i2c data ? |
| 06:18.35 | xsacha | huh? |
| 06:18.49 | xsacha | i looked at what winmo sent/received |
| 06:19.01 | cr2 | mmutrace or trace for pxa |
| 06:19.03 | cr2 | yes |
| 06:19.10 | xsacha | mmutrace? no idea |
| 06:19.19 | xsacha | i read zylonite_battery.dll |
| 06:19.30 | cr2 | how did you "looked at what winmo sent/received" |
| 06:19.38 | cr2 | ok |
| 06:19.40 | xsacha | went to the call to xllipi2cwrite/read |
| 06:19.57 | xsacha | and looked at parameters, slave address is third parameter (0x34) |
| 06:20.05 | cr2 | you can also watch what is sent over the wire |
| 06:20.09 | xsacha | and it send 0xD221 for temperature |
| 06:20.19 | xsacha | oh ok, how do i do that? |
| 06:20.33 | cr2 | is i2c the same location as on pxa270 ? |
| 06:20.51 | xsacha | not sure, im using i2c_pxa controller.. is it based on 270? |
| 06:22.36 | xsacha | a lot of things i need use i2c so watching wire would be useful |
| 06:29.30 | Echo31 | cr2: for example the cm_bf548 used AD7877 but it is not arm. The magician uses the pxa2xx_spi for the touscscreen tsc2046 |
| 06:32.36 | xsacha | for some reason when i type 'reboot', the device shuts down and reboots in to windows. other times it only shuts down |
| 06:32.52 | xsacha | sometimes* |
| 06:34.21 | tmzt | hey |
| 06:34.39 | tmzt | did you talk to WyrM? |
| 06:35.43 | xsacha | <4>BATT: LE output: 0x0506 |
| 06:35.44 | xsacha | <7>power_supply sgh_batt-0: prop CAPACITY=80 |
| 06:35.52 | xsacha | where does WyrM hang out? |
| 06:36.17 | tmzt | Echo31: does implement the spi bus/host, the chip driver should have no dependen.cy on the bus/host driver |
| 06:36.22 | tmzt | or the cpu |
| 06:36.28 | tmzt | here recently |
| 06:36.33 | tmzt | and #openezx |
| 06:36.36 | Echo31 | cr2: ok . I will take pxa2xx_spi. |
| 06:38.14 | cr2 | tmzt: i'm writing playwav2.c |
| 06:38.18 | xsacha | this is my dmesg: http://pastebin.com/d67fe3587 |
| 06:38.44 | tmzt | cool |
| 06:39.06 | tmzt | did you get the message about nuvi? |
| 06:39.11 | xsacha | no.. |
| 06:39.17 | tmzt | cr2 |
| 06:39.30 | cr2 | tmzt: yes. looks like some intermediate version |
| 06:39.44 | cr2 | tmzt: and also uses dejavu fonts :) |
| 06:40.03 | xsacha | does it have similar hardware? |
| 06:40.17 | cr2 | tmzt: is pxa3xx using the same pxa27x i2c driver ? |
| 06:40.59 | xsacha | all it tells me is: <6>I2C: i2c-0: PXA I2C adapter |
| 06:41.02 | cr2 | tmzt: the linux kernel code for nüvi6xx and 7xx is available |
| 06:41.09 | cr2 | xsacha: ok |
| 06:42.10 | cr2 | <PROTECTED> |
| 06:42.23 | cr2 | xsacha: look in haret where it is mapped |
| 06:42.29 | Echo31 | cr2: the AD7877 is it connected to ssp2 ? |
| 06:42.42 | cr2 | xsacha: dump mmu 2 0x40301688 4 |
| 06:42.48 | cr2 | Echo31: yes |
| 06:42.56 | cr2 | Echo31: it's in the wiki |
| 06:43.35 | Echo31 | cr2: it is similar with magician. |
| 06:43.59 | xsacha | cr2: 88300000 | 40300000 | 1MB section | CB AP=1 |
| 06:43.59 | xsacha | a8300000 | 40300000 | 1MB section | AP=1 |
| 06:44.14 | xsacha | and.. cp15: r1=00003a7f r2=bc390000 r3=00000001 r13=1c000000 |
| 06:45.36 | xsacha | http://pastebin.com/m6c124c5e |
| 06:45.48 | xsacha | is that what it is meant to be like? |
| 06:47.29 | tmzt | I think 3xx is the same i2c block |
| 06:47.43 | cr2 | xsacha: addlist mmutrace 0x88301688 4 |
| 06:47.45 | cr2 | xsacha: addlist mmutrace 0xa8301688 4 |
| 06:47.49 | tmzt | but marvell has only just started really supporting the chip |
| 06:47.53 | tmzt | in linux |
| 06:48.02 | cr2 | and then 'wi 2' |
| 06:48.08 | cr2 | mask all spam |
| 06:48.23 | cr2 | then you can really trace what goes over i2c |
| 06:48.59 | cr2 | Echo31: we may do something like that, but for the spi. |
| 06:49.47 | xsacha | 26 was spamming so ignored that.. |
| 06:49.53 | xsacha | now it's boring :P empty |
| 06:50.02 | cr2 | tmzt: why does android hardcode device names ? |
| 06:50.14 | xsacha | 87 when i touch screen, getting 134 occasionly |
| 06:50.42 | xsacha | only getting IRQs though |
| 06:52.09 | tmzt | where? |
| 06:52.28 | tmzt | I don't know |
| 06:52.30 | xsacha | cr2: with `wi` im not getting anything in 0x883.. or 0xa83.. |
| 06:52.52 | xsacha | just IRQs in 0xa8e and 0xa8d |
| 06:54.33 | cr2 | xsacha: you can try the pxa-specific trace |
| 06:54.46 | xsacha | where's that? |
| 06:54.50 | cr2 | with 'set trace' and 'set tracemask' |
| 06:54.54 | xsacha | k |
| 06:55.00 | cr2 | type 'help' |
| 06:55.17 | xsacha | set variable trace? |
| 06:55.31 | xsacha | doesnt have much help in there :P |
| 06:56.50 | xsacha | so i set trace to a number like that 0x88 one.. and then? :\ |
| 06:58.02 | cr2 | xsacha: read this first http://www.handhelds.org/moin/moin.cgi/ApachePhoneTrace |
| 06:59.30 | cr2 | xsacha: and here at the bottom http://wiki.xda-developers.com/index.php?pagename=MagicianCPLD |
| 07:00.57 | xsacha | these addresses must be wrong |
| 07:01.51 | xsacha | am i meant to just get IRQs? |
| 07:02.59 | xsacha | not working here, took a photo, called someone, nothing happening |
| 07:04.38 | cr2 | xsacha: it is strange, but may happen |
| 07:04.54 | cr2 | xsacha: how does the .dll access i2c ? |
| 07:05.34 | xsacha | it uses a simple i2c_transfer (equivalent on linux) and sends a command to 0x34 on the bus |
| 07:05.52 | xsacha | and the bus is external, defined somewhere else.. didnt look at that |
| 07:06.04 | Echo31 | cr2: i wish your help for spi |
| 07:06.46 | xsacha | cr2: the XllpI2CWrite/Read for windows ce is opensource |
| 07:06.58 | xsacha | well, the source is available* |
| 07:07.28 | xsacha | can see it here: http://read.pudn.com/downloads110/sourcecode/embed/457638/WinceKeypadCode/init2c.c__.htm |
| 07:07.43 | Echo31 | cr2: I don't find the topic dedicated for spi |
| 07:07.44 | xsacha | line 236 |
| 07:11.54 | xsacha | by the way, it turns out the i780 board is.. literally a zylonite? i was looking through marvell's zylonite kit and my device has the same chips except for the ones that are labelled 'optional' |
| 07:12.05 | cr2 | I2C_regs->XLLP_IDBR = (XLLP_UINT32_T) (*bytesBuf++); |
| 07:12.07 | xsacha | here's my USB3318: http://img29.imageshack.us/img29/3434/screenshotoyr.png |
| 07:12.13 | cr2 | it writes into IDBR |
| 07:12.23 | xsacha | IDBR? |
| 07:12.40 | cr2 | [08:41] <cr2> #define IDBR __REG(0x40301688) /* I2C Data Buffer Register - IDBR */ |
| 07:12.46 | xsacha | ah |
| 07:12.54 | cr2 | so IDBR must have a virtual mapping |
| 07:13.41 | xsacha | but defined at that address |
| 07:13.58 | cr2 | tmzt: handset, speaker, headset,bt. what is "handset" ? |
| 07:14.17 | cr2 | 0x40301688 is a physical address |
| 07:14.56 | cr2 | Echo31: use the driver of Kevin2 a in 2.6.21-hh20 |
| 07:14.59 | xsacha | addlist mmutrace p2v(0x40301688)? |
| 07:15.16 | cr2 | xsacha: ok |
| 07:15.38 | cr2 | xsacha: addlist mmutrace p2v(0x40301688) 4 |
| 07:15.47 | tmzt | the phone itself |
| 07:15.50 | cr2 | otherwise you will trace the whole megabyte |
| 07:16.03 | cr2 | tmzt: these are sound devices |
| 07:16.25 | cr2 | handset=whatever it is now=0 ? |
| 07:16.25 | xsacha | the speaker next to your ear? |
| 07:16.35 | tmzt | handset is the equivalent of Handy |
| 07:17.00 | cr2 | tmzt: i'm just curious about the gogle code |
| 07:17.11 | tmzt | it's the mobile industry's term for the phone |
| 07:17.19 | tmzt | for cellphones |
| 07:17.22 | cr2 | SND_DEVICE_BT |
| 07:17.29 | cr2 | SND_DEVICE_BT_EC_OFF |
| 07:17.46 | tmzt | echo cancelation |
| 07:17.51 | tmzt | is my guess |
| 07:18.20 | xsacha | still just getting irqs :( |
| 07:18.28 | cr2 | the id is inherited from somewhere else in an intransparent way ;) |
| 07:18.58 | tmzt | ? |
| 07:19.30 | *** join/#htc-linux Zoolooc (n=fredsiba@p54951DA3.dip0.t-ipconnect.de) |
| 07:20.04 | cr2 | ok, compiled. |
| 07:20.38 | tmzt | what git tree? |
| 07:20.40 | Echo31 | cr2: why cannot i use the driver AD7877 and spi from 2.6.31 from dev/pH5 code |
| 07:20.50 | *** join/#htc-linux kiozen (n=oeichler@rgnb-5d87de91.pool.einsundeins.de) |
| 07:21.15 | cr2 | Echo31: you can |
| 07:23.20 | xsacha | hmm the value at 0x40301688 keeps changing (if i dump it) |
| 07:23.35 | xsacha | 2d, 2f,35,31,etc.. |
| 07:24.02 | xsacha | are those changes meant to be seen by wi ? |
| 07:24.55 | cr2 | xsacha: yes |
| 07:25.02 | cr2 | root@htcraphael:~# ./playwav2 |
| 07:25.03 | cr2 | Segmentation fault |
| 07:26.35 | Echo31 | cr2: magician uses ssp2 but for athena, i think using ssp1 ( because only one ssp ?) |
| 07:28.17 | cr2 | GPIO 23 00800000 SSP CLK(bat+ts) O 2 |
| 07:28.19 | cr2 | GPIO 24 01000000 SSP FRM(bat+ts) O 2 1 |
| 07:28.20 | cr2 | GPIO 25 02000000 SSP TXD(bat+ts) O 2 |
| 07:28.22 | cr2 | GPIO 26 04000000 SSP RXD(bat+ts) I 1 |
| 07:29.33 | cr2 | Echo31: yes, i think it's SSP1 |
| 07:29.51 | xsacha | well no idea,.. this thing isnt working |
| 07:30.17 | xsacha | i tried addlist 0x40301688 4 (physical instead of virtual) and it said cant map unless i set permissivemmutrace or something |
| 07:30.27 | cr2 | ioctl(3, 0xc0047305, 0xbeffcaa8) = 0 |
| 07:30.28 | cr2 | --- SIGSEGV (Segmentation fault) @ 0 (0) --- |
| 07:30.49 | cr2 | set permissivemmutrace |
| 07:30.53 | cr2 | 1 |
| 07:32.06 | *** join/#htc-linux Zoolooc (n=fredsiba@p54951DA3.dip0.t-ipconnect.de) |
| 07:32.43 | xsacha | yeah i tried that and then it didnt watch anything :( |
| 07:32.54 | xsacha | and now i have to restart my phone :P screen frozen and red light is stuck on |
| 07:35.18 | cr2 | found 39 snd endpoints |
| 07:35.39 | cr2 | root@htcraphael:~# gdb |
| 07:35.40 | cr2 | -sh: gdb: not found |
| 07:41.07 | cr2 | [ 1085.401171] snd_ioctl set volume: invalid method. |
| 07:42.39 | cr2 | hmm |
| 07:44.54 | cr2 | tmzt: hehe. it works, but i don't hear sound ? |
| 07:46.59 | tmzt | did you add the paras? |
| 07:48.15 | cr2 | for the preprocessor ? |
| 07:48.34 | cr2 | hanged on playing a big .wav file |
| 07:49.13 | cr2 | i need to check the "invalid method" |
| 07:49.51 | tmzt | audioparas |
| 07:50.01 | cr2 | playwav: 2 ch, 22050 hz, 16 bit, PCM |
| 07:50.06 | xsacha | think sound will be easy on i780? because we have a really big opensource driver.. |
| 07:50.06 | tmzt | people.openezx.org/tmzt |
| 07:50.21 | xsacha | it only complains of wrong reg's |
| 07:50.32 | cr2 | http://people.openezx.org/tmzt |
| 07:50.36 | tmzt | depends on codec as well as pxa support |
| 07:51.00 | xsacha | codec has a big opensource team from the manufacturer themselves and pxa seems to support |
| 07:51.00 | tmzt | cr2: yeah, it's there somewhere |
| 07:51.21 | xsacha | driver loads, it just complains of wrong reg's |
| 07:51.46 | cr2 | tmzt: the params should be already in sram |
| 07:52.25 | tmzt | I'm quite sure it changed things n my phoneo |
| 07:52.35 | xsacha | they make my touchscreen + sound codec: http://opensource.wolfsonmicro.com/ |
| 07:52.40 | tmzt | just that led to static on phone call |
| 07:52.46 | tmzt | yeah, ok |
| 07:53.05 | cr2 | tmzt: i want to play .wav |
| 07:53.45 | tmzt | again you want to WyrM on this |
| 07:53.53 | tmzt | talk to |
| 07:54.00 | cr2 | <PROTECTED> |
| 07:54.02 | cr2 | <PROTECTED> |
| 07:54.03 | cr2 | <PROTECTED> |
| 07:54.05 | cr2 | <PROTECTED> |
| 07:54.06 | cr2 | <PROTECTED> |
| 07:56.16 | *** join/#htc-linux Zoolooc_ (n=fredsiba@p54952D5D.dip0.t-ipconnect.de) |
| 08:00.20 | xsacha | i got android working with my battery haha! |
| 08:00.27 | xsacha | i had to change battery name to "battery" |
| 08:01.48 | *** join/#htc-linux Squarc (n=Squarc@82-217-32-29.cable.quicknet.nl) |
| 08:02.54 | tmzt | hah |
| 08:03.14 | tmzt | cr2: where is that code? |
| 08:03.22 | *** join/#htc-linux onen|openBmap (n=quassel@mry91-1-89-87-198-158.dsl.club-internet.fr) |
| 08:05.31 | cr2 | tmzt: in the kernel |
| 08:07.27 | *** join/#htc-linux ptitjes (n=didier@80.125.172.60) |
| 08:09.38 | *** join/#htc-linux onen|openBmap (n=quassel@mry91-1-89-87-198-158.dsl.club-internet.fr) |
| 08:11.39 | tmzt | cr2: I see, it has to be SND METHOD VOICE |
| 08:12.13 | cr2 | 2,5 snd_set_volume_modem_reset (dev=d,method=1,vol=5,cb,data) |
| 08:14.43 | Echo31 | cr2: what is the difference between GPIO 114 Toucscreen DAV ant GPIO 27 TS_PENDOWN ? |
| 08:15.36 | tmzt | DAV? |
| 08:16.52 | xsacha | pen down is only when you are on the screen? |
| 08:17.31 | cr2 | Echo31: it's all in the Kevin2's driver |
| 08:17.55 | cr2 | Echo31: this code is used on htc apache too. |
| 08:18.14 | cr2 | actually it comes from apache |
| 08:19.00 | cr2 | Echo31: the pen down irq is GPIO 27 TS_PENDOWN |
| 08:21.56 | cr2 | tmzt: i have verified that all adsp modules for 5200 make sense. you need to recheck your camera modules |
| 08:23.04 | tmzt | how do I do that? |
| 08:23.35 | tmzt | how is sound working now? does it appear to work but no sound? |
| 08:24.42 | cr2 | only for small files |
| 08:24.59 | cr2 | for the big files like kde startup it hangs as before |
| 08:25.24 | cr2 | the rpc calls tracked @wince all go through. |
| 08:25.37 | cr2 | i'll try to recheck the sequence. |
| 08:25.50 | cr2 | because of the 'mute' settings |
| 08:27.35 | tmzt | ok |
| 08:27.49 | tmzt | can you dump the aucoustic settings? |
| 08:28.05 | tmzt | from linux I mean |
| 08:28.18 | cr2 | how ? |
| 08:28.56 | cr2 | from /dev/mem ? they should have stayed as in wince |
| 08:29.29 | cr2 | i guess if you will modify them, you need to call the 'audio update' DEX |
| 08:29.35 | tmzt | in the driver |
| 08:29.51 | tmzt | ah |
| 08:30.03 | tmzt | maybe the get written to the dsp then |
| 08:30.53 | tmzt | not stay in smem |
| 08:33.50 | cr2 | probably |
| 08:34.27 | cr2 | the only problem i have with adsp, is how to verify the translation tables hardcoded there, and the queue numbers. |
| 08:35.14 | cr2 | <PROTECTED> |
| 08:35.29 | cr2 | <PROTECTED> |
| 08:35.52 | tmzt | but we are getting silence |
| 08:36.11 | tmzt | at least noise would mean it's working |
| 08:37.00 | cr2 | let me check how much modules i've seen from this list: |
| 08:37.05 | cr2 | 252 static struct adsp_module_info module_info[] = { |
| 08:37.06 | cr2 | 253 QDSP_MODULE(AUDPPTASK), |
| 08:37.08 | cr2 | 254 QDSP_MODULE(AUDRECTASK), |
| 08:37.09 | cr2 | 255 QDSP_MODULE(AUDPREPROCTASK), |
| 08:37.11 | cr2 | 256 QDSP_MODULE(VFETASK), |
| 08:37.12 | cr2 | 257 QDSP_MODULE(QCAMTASK), |
| 08:37.14 | cr2 | 258 QDSP_MODULE(LPMTASK), |
| 08:37.15 | cr2 | 259 QDSP_MODULE(JPEGTASK), |
| 08:37.17 | cr2 | 260 QDSP_MODULE(VIDEOTASK), |
| 08:37.18 | cr2 | 261 QDSP_MODULE(VDEC_LP_MODE), |
| 08:38.44 | cr2 | <PROTECTED> |
| 08:38.58 | cr2 | <PROTECTED> |
| 08:39.13 | cr2 | <PROTECTED> |
| 08:39.14 | cr2 | <PROTECTED> |
| 08:39.31 | cr2 | <PROTECTED> |
| 08:39.43 | cr2 | <PROTECTED> |
| 08:39.45 | cr2 | <PROTECTED> |
| 08:39.46 | cr2 | <PROTECTED> |
| 08:40.20 | cr2 | tmzt: you should recheck the last 4 |
| 08:40.46 | cr2 | don't know how to test VIDEOTASK and VDEC_LP_MODE |
| 08:40.46 | tmzt | how do I check them? |
| 08:40.52 | cr2 | wtf is LP_MODE ? |
| 08:41.29 | tmzt | low power? |
| 08:41.32 | cr2 | enable the camera, dump rpccall, save jpeg, dump rpccall, etc. |
| 08:41.39 | tmzt | ah |
| 08:41.53 | tmzt | dump the smem like before? |
| 08:41.59 | cr2 | yes |
| 08:42.06 | tmzt | ok |
| 08:42.22 | tmzt | do you have the command? |
| 08:42.35 | tmzt | well, I can't upload very well right now |
| 08:42.45 | cr2 | QTV_LP image |
| 08:43.07 | cr2 | no, must grep through the logs |
| 08:43.42 | tmzt | limited profile? |
| 08:46.11 | *** join/#htc-linux Squarc (n=Squarc@82-217-32-29.cable.quicknet.nl) |
| 08:46.54 | cr2 | msm_dmov_enqueue_cmd(11) error datamover stalled, status 0 |
| 09:07.23 | *** join/#htc-linux Zoolooc_ (n=fredsiba@p54952DF4.dip0.t-ipconnect.de) |
| 09:15.27 | *** join/#htc-linux ecze (n=ecze@eczema.ecze.com) |
| 09:27.54 | *** join/#htc-linux KindofBlue (n=KindofBl@rgnb-5d87d216.pool.einsundeins.de) |
| 09:27.58 | cr2 | root@htcraphael:~# stty -F /dev/ttyHS1 speed 921600 |
| 09:28.00 | cr2 | 115200 |
| 09:28.01 | cr2 | root@htcraphael:~# stty -F /dev/ttyHS1 |
| 09:28.03 | cr2 | speed 921600 baud; |
| 09:28.27 | cr2 | looks like a bug in the g1 code. |
| 09:28.49 | cr2 | haha |
| 09:28.54 | cr2 | root@htcraphael:~# stty -F /dev/ttyHS1 speed 4000000 |
| 09:28.55 | cr2 | stty: invalid argument '4000000' |
| 09:29.53 | cr2 | [ 204.300301] clock-wince: set_mdns_host_clock: UART2DM_CLK, freq=7372800 |
| 09:29.53 | *** join/#htc-linux g55 (n=g55@rgnb-5d87d216.pool.einsundeins.de) |
| 09:29.54 | cr2 | [ 207.705910] clock-wince: pc_clk_set_rate: id=UART2DM_CLK rate=7372800 |
| 09:29.56 | cr2 | [ 207.723946] clock-wince: set mdns: UART2DM_CLK, 7372800; bitidx=26, offset=dc, ns=0 |
| 09:30.03 | cr2 | but the real problem is dma stall |
| 09:32.40 | tmzt | oh, that error was about bt?? |
| 09:34.15 | cr2 | yes, i have many patches in the kernel |
| 09:40.19 | cr2 | #define DMOV_HSUART2_TX_CHAN 4 |
| 09:40.29 | cr2 | why do i get 11 ? |
| 09:40.44 | cr2 | hmm |
| 09:40.47 | cr2 | #define DMOV_HSUART2_RX_CHAN 1 |
| 09:40.55 | cr2 | #define DMOV_HSUART2_RX_CHAN 11 |
| 09:41.04 | xsacha | tmzt, any ideas bout my battery registers? i dumped all the registers here: http://andromnia.pastebin.com/d5735922b |
| 09:42.30 | xsacha | every fourth register gives '0x0000' |
| 09:42.53 | Echo31 | cr2: how can i check the initialisation of spi and ad7877. I don't see nothing from dmesg ? |
| 09:42.54 | xsacha | i thought i had formula for voltage/charge right but it doesnt seem to be right at all :\ |
| 09:45.25 | Echo31 | cr2: i find pxa2xx-spi.1 in /proc/interrupts but i don't find ad7877 |
| 09:45.32 | xsacha | i dont think those 'E's are relevant but if i dont use them i'm left with a simple byte |
| 09:57.06 | cr2 | Echo31: then you must be doing something wrong |
| 09:57.47 | Echo31 | cr2: i think so. |
| 10:00.29 | Echo31 | cr2: find the piece of code http://fr.pastebin.ca/1483731 |
| 10:03.38 | cr2 | tmzt: ok, now it's gone. |
| 10:04.06 | cr2 | tmzt: tell San that he is a google genius for sharing uart2dm with usb ;-) |
| 10:07.04 | tmzt | hah |
| 10:07.31 | cr2 | Echo31: don't see anything wrong |
| 10:08.27 | cr2 | tmzt: the driver is still buggy, because it produces oops in _hs_shutdown() on a dma error. |
| 10:09.29 | tmzt | but it works? |
| 10:09.40 | cr2 | wait_event(msm_uport->rx.wait, msm_uport->rx.flush == FLUSH_SHUTDOWN); |
| 10:09.40 | tmzt | if you don't close the device? |
| 10:09.47 | tmzt | does cat work? |
| 10:10.00 | cr2 | this call causes an oops in case of a dma error |
| 10:10.15 | cr2 | i need to powerup the bt first |
| 10:10.33 | cr2 | how can i kick this rfkill from userspace ? |
| 10:11.36 | tmzt | in sysfs? |
| 10:12.02 | cr2 | which one ? |
| 10:13.21 | tmzt | I really don't know |
| 10:13.26 | tmzt | is it standard? |
| 10:14.29 | tmzt | '/sys/class/misc/rfkill |
| 10:14.40 | tmzt | '/sys/class/rfkill |
| 10:15.56 | cr2 | ok, i'll revert some debugging patches first |
| 10:16.35 | xsacha | this battery driver is weird. all the registers increase about 0x40 when i put in charger |
| 10:16.52 | xsacha | but there's a register that is on for when i have a charger and 0 for when i dont and there's also a GPIO for it |
| 10:17.44 | cr2 | wee. stty is fixed now. |
| 10:18.07 | *** join/#htc-linux vts (n=vts@62-47-219-175.adsl.highway.telekom.at) |
| 10:18.10 | cr2 | tmzt: where do we post the g1 code bugs ? |
| 10:18.30 | tmzt | b.android.com |
| 10:18.34 | xsacha | http://source.android.com/submit-patches |
| 10:18.45 | xsacha | oh bugs |
| 10:18.54 | tmzt | it's not a hardware thing? |
| 10:19.01 | cr2 | it's not a patch |
| 10:19.33 | cr2 | tmzt: i don't know, but an oops in the kernel is not acceptable |
| 10:20.06 | cr2 | dma failure is not a legitimate hardware condition ? |
| 10:20.08 | *** part/#htc-linux vts (n=vts@62-47-219-175.adsl.highway.telekom.at) |
| 10:21.44 | cr2 | tmzt: how do i enable power in rfkill ? |
| 10:21.48 | *** join/#htc-linux onen|openBmap (n=quassel@mry91-1-89-87-198-158.dsl.club-internet.fr) |
| 10:22.06 | cr2 | <PROTECTED> |
| 10:22.18 | tmzt | echo 0 |
| 10:22.22 | cr2 | root@htcraphael:/sys/class/rfkill/rfkill0# cat state |
| 10:22.24 | cr2 | 0 |
| 10:22.34 | tmzt | 1 then 0 |
| 10:22.35 | cr2 | to which node ? |
| 10:22.45 | tmzt | probably state |
| 10:22.53 | cr2 | [ 390.182137] bluetooth rfkill state ON |
| 10:23.03 | tmzt | the otheres will say invalid anyway |
| 10:23.21 | cr2 | hm. but no debug powerup message ?? |
| 10:23.22 | tmzt | I mean the uart/usb dm overlap was not hardware? |
| 10:23.35 | cr2 | no |
| 10:24.28 | cr2 | the hsuarts share the rx dma channel already (in the g1 source), so you can't use them simultaneously |
| 10:24.55 | cr2 | but the uart2dm channel was shared with usb ;) |
| 10:25.59 | cr2 | hciattach /dev/ttyHS1 texas |
| 10:26.07 | cr2 | lockup |
| 10:26.21 | tmzt | and just cat? |
| 10:26.35 | cr2 | i'll reboot first |
| 10:26.38 | tmzt | or stty |
| 10:26.44 | cr2 | stty is ok |
| 10:26.57 | tmzt | that's the first thing hciattach does |
| 10:27.09 | cr2 | i'll use strace now |
| 10:27.14 | tmzt | but dm is only enabled above a certain speed I think |
| 10:27.34 | tmzt | funny, strace fixed the bt on my q |
| 10:27.42 | tmzt | hciattach failed otherwise |
| 10:27.51 | tmzt | it slowed it down |
| 10:27.58 | tmzt | no idea why it worked |
| 10:28.03 | cr2 | hehe |
| 10:28.50 | cr2 | actually, i need to check which dma channel is used by uart2dm in wince |
| 10:29.07 | cr2 | the dma channel usage is something for wiki |
| 10:30.17 | cr2 | hm. i need to check the powerup messages first. before going to hciattach |
| 10:30.48 | *** join/#htc-linux vts (n=vts@88-117-92-104.adsl.highway.telekom.at) |
| 10:30.51 | NAiL | cr2: Did you have any insight on my kaiser usb-issue on XP? :) |
| 10:31.11 | *** join/#htc-linux leaigor (n=laigor@188.134.36.14) |
| 10:31.26 | cr2 | NAiL: it's not a linux problem :) |
| 10:31.45 | cr2 | <PROTECTED> |
| 10:31.47 | cr2 | ok |
| 10:32.06 | NAiL | cr2: In android it doesn't show up at all :-P |
| 10:32.49 | tmzt | it's probably your kernel |
| 10:33.03 | tmzt | on the deviceon the device |
| 10:34.47 | tmzt | recent dzo tree has usb support |
| 10:34.53 | tmzt | older ones don't |
| 10:35.45 | *** join/#htc-linux BabelO (n=fcr@unaffiliated/babelo) |
| 10:36.51 | cr2 | tmzt: in case of the device lockup i can ping it, but other interrupts do not work |
| 10:37.43 | tmzt | interesting |
| 10:38.00 | tmzt | ping? over usb? |
| 10:38.24 | tmzt | maybe only a process locked up? |
| 10:38.33 | tmzt | do you have multiple getty's |
| 10:38.44 | cr2 | yes, over usb |
| 10:38.48 | tmzt | but I guess you can't switch between them |
| 10:38.50 | cr2 | got an oops on the screen |
| 10:39.18 | tmzt | what interrupts don't work? |
| 10:40.12 | cr2 | _hs_startup, request_irq, setup_irq, handle_edge_irq |
| 10:40.30 | cr2 | keyboard/i2c |
| 10:40.55 | cr2 | usb/inout works, lcd is refreshed |
| 10:41.28 | cr2 | but can't ssh |
| 10:41.47 | tmzt | lsusb works? |
| 10:42.05 | cr2 | ping works |
| 10:42.06 | *** join/#htc-linux pH5 (n=ph5@p5485C2FC.dip.t-dialin.net) |
| 10:42.08 | *** join/#htc-linux lama (n=lama@netbsd.pl) |
| 10:42.24 | *** join/#htc-linux Red_Lion (n=lion@77.41.93.140) |
| 10:42.28 | cr2 | even ping -f |
| 10:42.46 | cr2 | 0% packet loss, |
| 10:43.47 | tmzt | so it's probably gethostbyname or something else the sshd process does after forking |
| 10:44.12 | cr2 | i2c irq ? |
| 10:44.42 | tmzt | not sure |
| 10:44.50 | tmzt | but oops breaks things randomly |
| 10:44.57 | tmzt | random things |
| 10:45.03 | cr2 | ok |
| 10:45.16 | tmzt | I had an oops cause me not to be able to sudo |
| 10:45.30 | tmzt | because of gethostbyname |
| 10:45.33 | cr2 | yes, the ram corruption |
| 10:46.37 | cr2 | i need to check my wince dmesg for the dma usage |
| 10:47.44 | *** part/#htc-linux vts (n=vts@88-117-92-104.adsl.highway.telekom.at) |
| 10:48.40 | cr2 | root@htcraphael:~# stty -F /dev/ttyHS1 speed 115200 |
| 10:48.42 | cr2 | 9600 |
| 10:49.20 | tmzt | 9600? |
| 10:50.39 | cr2 | yes, set in the kernel as default speed |
| 10:50.47 | cr2 | strace: |
| 10:50.51 | cr2 | ioctl(3, SNDCTL_TMR_START or TCSETS, {B115200 -opost -isig -icanon -echo ...}) = 0 |
| 10:50.52 | cr2 | ioctl(3, TCFLSH, 0x2) = 0 |
| 10:50.54 | cr2 | write(3, "\1\1\20\0", 4) = 4 |
| 10:50.55 | cr2 | read(3, 0xbe963ae8, 1) = ? ERESTARTSYS (To be restarted) |
| 10:50.57 | cr2 | --- SIGALRM (Alarm clock) @ 0 (0) --- |
| 10:50.58 | cr2 | write(2, "Initialization timed out.\n", 26Initialization timed out. |
| 10:51.00 | cr2 | ) = 26 |
| 10:51.01 | cr2 | io_submit(0x1, 0x1, 0xfbad2088 |
| 10:51.22 | cr2 | write was ok, or seemed to be ok |
| 10:51.40 | cr2 | read(3, 0xbe963ae8, 1) = ? ERESTARTSYS (To be restarted) |
| 10:51.50 | cr2 | this one was not very successful |
| 10:52.02 | tmzt | did it return 9600 though? what did you mean |
| 10:52.16 | cr2 | may be the wrong clock speed in wince-clock.c |
| 10:52.27 | cr2 | stty returns the old speed |
| 10:52.59 | tmzt | ok |
| 10:53.00 | cr2 | [ 219.808724] clock-wince: pc_clk_set_rate: id=UART2DM_CLK rate=7372800 |
| 10:53.02 | cr2 | [ 219.823555] clock-wince: set mdns: UART2DM_CLK, 7372800; bitidx=26, offset=dc, ns=0 |
| 10:53.03 | cr2 | [ 219.838082] clock-wince: set_mdns_host_clock: UART2DM_CLK, freq=7372800 |
| 10:53.10 | cr2 | this time without oops |
| 10:53.22 | cr2 | strace -ff hciattach -t 5 /dev/ttyHS1 texas |
| 10:53.35 | cr2 | need to powerup :) |
| 10:54.05 | cr2 | echo 1 > /sys/class/rfkill/rfkill0/state |
| 10:54.11 | tmzt | no |
| 10:54.12 | tmzt | 0 |
| 10:54.17 | cr2 | [ 460.745949] bluetooth rfkill state ON |
| 10:54.42 | cr2 | <PROTECTED> |
| 10:54.49 | cr2 | bluetooth rfkill state OFF |
| 10:55.17 | tmzt | so it should be powered |
| 10:56.00 | cr2 | mprotect(0x2a000000, 24576, PROT_READ|PROT_EXEC) = 0 |
| 10:56.02 | cr2 | rt_sigaction(SIGALRM, {0x2a001c50, [], SA_NOCLDSTOP|0x4000000}, NULL, 8) = 0 |
| 10:56.03 | cr2 | setitimer(ITIMER_REAL, {it_interval={0, 0}, it_value={5, 0}}, {it_interval={0, 0}, it_value={0, 0}}) = 0 |
| 10:56.09 | cr2 | hangs hee |
| 10:56.19 | cr2 | the oops is on the lcd |
| 10:56.24 | tmzt | really? |
| 10:56.42 | cr2 | yes |
| 10:56.47 | tmzt | nothing to do with hsuart/uartDM? |
| 10:57.00 | cr2 | at least not directly |
| 10:57.22 | cr2 | ping works, and the irqs are screwed |
| 10:57.30 | tmzt | real time priority? |
| 10:58.07 | cr2 | usb works, ping works |
| 10:58.35 | cr2 | may be a dma error, but usb also uses dma |
| 10:59.02 | cr2 | rebooting |
| 10:59.13 | tmzt | what is itimer? |
| 10:59.22 | cr2 | interval timer |
| 10:59.32 | tmzt | yeah |
| 10:59.42 | cr2 | it's hciattach -t 5 |
| 10:59.43 | tmzt | have never seen it |
| 10:59.47 | tmzt | ah |
| 10:59.51 | cr2 | timeout for this call |
| 10:59.58 | tmzt | what does that do? |
| 11:00.00 | tmzt | ok |
| 11:00.46 | tmzt | then sends sigalrm? |
| 11:02.18 | cr2 | <PROTECTED> |
| 11:02.19 | cr2 | <PROTECTED> |
| 11:02.24 | cr2 | need to check this one |
| 11:02.37 | cr2 | <PROTECTED> |
| 11:02.39 | cr2 | <PROTECTED> |
| 11:02.40 | cr2 | <PROTECTED> |
| 11:02.42 | cr2 | <PROTECTED> |
| 11:02.45 | cr2 | and this google hack too |
| 11:04.55 | cr2 | <PROTECTED> |
| 11:04.59 | tmzt | no non low power version? |
| 11:05.47 | *** part/#htc-linux Echo31 (n=olivier@mir31-4-82-240-194-54.fbx.proxad.net) |
| 11:05.49 | cr2 | it's wakup from suspend on bt headset, for example |
| 11:06.40 | *** join/#htc-linux Echo31 (n=olivier@mir31-4-82-240-194-54.fbx.proxad.net) |
| 11:08.03 | tmzt | rts? |
| 11:08.24 | tmzt | from bt chip? |
| 11:08.33 | cr2 | yes |
| 11:08.35 | tmzt | how can we trace that gpio? |
| 11:08.52 | cr2 | you can try |
| 11:08.54 | tmzt | when hci packet is ready? |
| 11:09.05 | cr2 | but you need a bt headset for it |
| 11:09.13 | cr2 | ? |
| 11:09.17 | tmzt | yeah |
| 11:09.38 | tmzt | hci packet from the bluetooth chip |
| 11:10.27 | tmzt | bt headset is a AT command over rfcomm channel |
| 11:10.50 | tmzt | linux rfcomm should simulate that |
| 11:11.01 | tmzt | just not sco channel |
| 11:11.17 | cr2 | root@htcraphael:~# cat /proc/interrupts |
| 11:11.32 | cr2 | <PROTECTED> |
| 11:11.38 | cr2 | 223 ?? |
| 11:11.50 | Echo31 | cr2: what is the better distribution for the athena porting ? Angstrom ? |
| 11:12.13 | cr2 | <PROTECTED> |
| 11:12.30 | tmzt | it's dm? |
| 11:12.31 | cr2 | Echo31: non-gui |
| 11:12.42 | cr2 | tmzt: it's the dma controller |
| 11:12.54 | cr2 | tmzt: it seems that the mdp uses a different dma |
| 11:13.08 | cr2 | <PROTECTED> |
| 11:13.37 | Echo31 | cr2: how can i check the touchscreen ? |
| 11:14.18 | cr2 | Echo31: od -x /dev/input/event* |
| 11:14.42 | cr2 | tmzt: don't see a hsuart irq ? |
| 11:15.22 | cr2 | <PROTECTED> |
| 11:15.52 | cr2 | a0300000-a0300fff : msm_serial_hs.1 |
| 11:15.54 | tmzt | do you see a uart irq? |
| 11:16.13 | cr2 | maybe it registers it only on open() |
| 11:17.04 | cr2 | <PROTECTED> |
| 11:17.05 | cr2 | <PROTECTED> |
| 11:17.20 | cr2 | yeah |
| 11:17.23 | cr2 | /* Called when port is opened */ |
| 11:17.25 | cr2 | static int msm_hs_startup(struct uart_port *uport) |
| 11:17.50 | tmzt | ok |
| 11:18.27 | tmzt | but with PM not enabled the low power thing might be false |
| 11:19.16 | cr2 | it's another irq |
| 11:20.32 | cr2 | i'm using this as a "base" clock now |
| 11:20.35 | cr2 | <PROTECTED> |
| 11:20.59 | cr2 | the spl uses // MSM_CLOCK_REG( 7372800, 2, 0xc8, 0x64, 3, 2, 1, 1), /* unknown */ |
| 11:21.35 | cr2 | i think for this one i can calculate the frequency, because it's the same class as SD |
| 11:22.09 | cr2 | but then i get some strange frequency |
| 11:22.38 | cr2 | 460800*16=7372800 btw |
| 11:23.23 | cr2 | the direct dump of the lcokc registers gives |
| 11:23.26 | cr2 | MSM_CLOCK_REG( 460800*16, 3, 0x64, 0x32, 0, 2, 4, 1) |
| 11:24.25 | cr2 | it's a scale difference P=4 |
| 11:25.01 | cr2 | it's difficult to say what is right. without disassembling the amss ;) |
| 11:26.31 | *** join/#htc-linux Guest87676 (n=fnord@24-151-90-116.static.nwtn.ct.charter.com) |
| 11:26.47 | *** join/#htc-linux kiozen (n=oeichler@rgnb-5d87de91.pool.einsundeins.de) |
| 11:30.34 | cr2 | tmzt: does the codeaurora use the same driver ? |
| 11:30.45 | tmzt | haven't looked |
| 11:30.57 | tmzt | 2.6.29 might be better |
| 11:31.24 | tmzt | tuxhero gave me a patch with the 27 code ported to 29 |
| 11:31.30 | Echo31 | cr2: it is not easy to see . However, if I want to include the w228x, how I do ? |
| 11:31.34 | tmzt | but the Kconfig has some issues |
| 11:32.05 | tmzt | hold on |
| 11:32.59 | cr2 | +CONFIG_USB_MSM_OTG=y |
| 11:34.00 | tmzt | host? |
| 11:34.07 | tmzt | get pm? |
| 11:35.40 | cr2 | +#define PM_APP_OTG_PROG 0x30000080 |
| 11:35.47 | cr2 | too high for us |
| 11:37.18 | *** join/#htc-linux kiozen (n=oeichler@rgnb-5d87de91.pool.einsundeins.de) |
| 11:41.47 | *** join/#htc-linux pleemans (n=toi@d54C2AAB7.access.telenet.be) |
| 11:48.06 | cr2 | codeaurora uses the gogle driver |
| 11:50.01 | cr2 | <PROTECTED> |
| 11:57.11 | jaSOnGg | :o |
| 11:57.58 | tmzt | what's too high? |
| 11:59.40 | cr2 | we don't have such rpc numbers |
| 12:00.26 | tmzt | can't see why otg needs rpc |
| 12:10.05 | *** join/#htc-linux kiozen (n=oeichler@rgnb-5d87de91.pool.einsundeins.de) |
| 12:12.27 | cr2 | tmzt: can we talk to San about the dma channel ? |
| 12:12.38 | *** join/#htc-linux MethoS (n=clemens@host-091-097-243-127.ewe-ip-backbone.de) |
| 12:12.40 | cr2 | the right clock seems to be a difficult task. |
| 12:13.02 | cr2 | maybe i'd create my own divisors using the SD formula ? |
| 12:13.19 | cr2 | and not using the wince setup at all. |
| 12:13.54 | cr2 | and revisit the spl settings again |
| 12:18.58 | cr2 | so the problem is how to generate the 7372800 Hz clock out of TCXO=19.2 MHz or the PLL1=768 MHz clock, using M/N/P |
| 12:20.36 | cr2 | PLL1=40*TCX0 |
| 12:22.38 | tmzt | he won't be in till later if at all today |
| 12:23.07 | *** join/#htc-linux miknix (n=miknix@gentoo/developer/miknix) |
| 12:28.14 | cr2 | 768000000*M=P*N*7372800, P=1,2,4, M<N (i think) |
| 12:28.44 | cr2 | tmzt: can this equation be solved in for integers M and N ? |
| 12:30.33 | cr2 | P*N is N, the prescaler is probably used to avoid feeding M/N with too high frequency. |
| 12:31.30 | cr2 | so it's 7680000*M=7372800*N |
| 12:32.12 | tmzt | integers? |
| 12:32.13 | cr2 | need to factor 7680000 and 73728 into prime numbers :) |
| 12:32.19 | cr2 | err |
| 12:32.26 | cr2 | 7680000*M=73728*N |
| 12:32.49 | cr2 | never thought that number theory is useful in real life :) |
| 12:33.35 | cr2 | both sides can be divided by 4 |
| 12:33.53 | tmzt | so p is 4? |
| 12:33.58 | ali1234 | the other day i was looking at the source code for a xorg driver which calculated the pixel clock coefficients by having a for loop over the valid ranges of each one |
| 12:34.38 | cr2 | ali1234: that's an engineer's solution :) |
| 12:34.50 | ali1234 | i think it was intel driver |
| 12:35.20 | cr2 | it's easier to divide 7680000 by 73728 and see what comes out |
| 12:35.23 | tmzt | kernel? |
| 12:35.40 | tmzt | division is hard in the kernel I think |
| 12:35.54 | tmzt | npot division |
| 12:35.57 | cr2 | awk '{printf("%.6f\n",7680000/73728)}' |
| 12:35.59 | cr2 | 104.166667 |
| 12:36.01 | ali1234 | N = 625M/6 |
| 12:36.20 | tmzt | rational! |
| 12:36.20 | cr2 | 0.166667 |
| 12:36.32 | cr2 | multily by 6 |
| 12:36.51 | cr2 | ali1234: 625 ? |
| 12:37.13 | ali1234 | 104.166667 = 625/6 |
| 12:37.27 | tmzt | this is why I want gen clk tree support |
| 12:37.38 | cr2 | :) |
| 12:37.44 | cr2 | ali1234: yes |
| 12:37.49 | tmzt | teach the kernel what the tree looks like and let it figure it out |
| 12:38.37 | cr2 | now we need to split 625 into the prescaler and N |
| 12:38.37 | cr2 | the known values are 1-4 and < 0xff |
| 12:38.42 | cr2 | seems to be impossible |
| 12:39.22 | ali1234 | 625 = 5^4 |
| 12:39.23 | cr2 | 125 * 5 would have been nice. |
| 12:40.13 | cr2 | ali1234: i think the prescaler uses 2 bits |
| 12:40.33 | ali1234 | that's too convenient... |
| 12:41.16 | ali1234 | where did 7680000 and 73728 come from? |
| 12:41.20 | cr2 | tmzt: for gen clk you need to know how all clocks are generated and distributed |
| 12:41.27 | tmzt | yes |
| 12:41.36 | tmzt | that's what I mean |
| 12:42.05 | cr2 | 7372800 is hardcoded into the g1 driver, and 768 MHz is known from the SD clocks and uboot source |
| 12:42.41 | tmzt | maybe uart is different? |
| 12:42.49 | tmzt | not the same formula |
| 12:43.08 | cr2 | awk '{print 7372800/4}' |
| 12:43.09 | cr2 | 1843200 |
| 12:43.10 | tmzt | what does uboot say for our soc? |
| 12:43.17 | cr2 | ali1234: does it tell you something ? |
| 12:43.36 | ali1234 | no, just wondered. how is it "hardcoded" into the driver? |
| 12:44.02 | ali1234 | btw 19.2M/7372800 = 125/48 |
| 12:44.06 | cr2 | ali1234: the 18.432MHz is something very familiar |
| 12:44.21 | tmzt | crystal? |
| 12:44.39 | cr2 | it's a standard crystal for the uart clocks. |
| 12:44.53 | tmzt | look at c driver |
| 12:44.55 | tmzt | pc |
| 12:45.08 | tmzt | 16550 or whatever |
| 12:45.13 | cr2 | ali1234: 125/48 sounds good |
| 12:45.27 | cr2 | ali1234: the 768=19.2*40 |
| 12:45.53 | tmzt | is it right? |
| 12:46.09 | cr2 | so i'm a bit surprised. |
| 12:46.14 | tmzt | how can it be 19.2M? |
| 12:46.42 | cr2 | tmzt: the base clock on msm is TCX0=19.2MHz |
| 12:46.53 | cr2 | the PLL1 is TCX0*40 |
| 12:46.58 | xsacha | im really interested in seeing what goes over the i2c wire but i cant work out this wi thing. isn't wi short for watch irqs? |
| 12:47.03 | tmzt | ok |
| 12:47.11 | cr2 | xsacha: yes |
| 12:47.23 | xsacha | what am i meant to see when data is written to i2c? |
| 12:47.32 | tmzt | where did you get 18.432? |
| 12:47.38 | tmzt | then |
| 12:47.58 | tmzt | xsacha: addr=value |
| 12:48.03 | tmzt | use haretconsole |
| 12:48.11 | xsacha | all i see is irqs :\ |
| 12:48.39 | tmzt | the you probably have the wront virt addr |
| 12:48.40 | cr2 | ali1234: you are right. i'm just surprised how it happens. |
| 12:48.42 | xsacha | like IRQS a8....(26) = 0x4000 |
| 12:48.47 | tmzt | dump mmu 2 phys |
| 12:49.01 | ali1234 | cr2: i am? |
| 12:49.20 | cr2 | awk '{print 19200000*48,7372800*125}' |
| 12:49.22 | cr2 | 921600000 92160000 |
| 12:49.38 | cr2 | [14:43] <ali1234> btw 19.2M/7372800 = 125/48 |
| 12:49.56 | ali1234 | yeah, i just used wolfram alpha, it does all this math stuff real nice :) |
| 12:50.14 | ali1234 | telling you factors and stuff. much better than google |
| 12:51.10 | cr2 | factor is included in baseutils |
| 12:51.36 | cr2 | factor 7372800 |
| 12:51.37 | cr2 | 7372800: 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 5 5 |
| 12:51.40 | cr2 | yeah |
| 12:51.59 | cr2 | so |
| 12:52.21 | cr2 | P=1, M=0x30, N=0x7d src=TCX0 |
| 12:53.11 | cr2 | <PROTECTED> |
| 12:53.36 | xsacha | tmzt, it's humungous |
| 12:53.47 | tmzt | ? |
| 12:55.33 | xsacha | made me dump of the virtual address space? dump mmu 2 phys |
| 12:55.50 | cr2 | <PROTECTED> |
| 12:56.07 | cr2 | the 125 can't be divided by 2. |
| 12:56.53 | cr2 | what was the spl using... |
| 12:57.28 | tmzt | is it the 4k or 1mb sections we want? |
| 12:57.33 | tmzt | cr2 |
| 12:57.35 | cr2 | <PROTECTED> |
| 12:57.39 | cr2 | tmzt: both |
| 12:58.01 | xsacha | i see a tonne of 4k and some 1MB |
| 12:58.15 | xsacha | it took 10 seconds to print them all |
| 12:59.08 | cr2 | xsacha: yes. now look for you physical address page |
| 12:59.16 | xsacha | :O |
| 12:59.40 | xsacha | but the first message i can see on console starts with a 'c' cause it flooded console |
| 12:59.45 | cr2 | xsacha: dump mmu 2 should have done it for you |
| 13:00.12 | cr2 | tmzt: weird |
| 13:00.16 | cr2 | awk '{printf("%d\n",768000000*2/0xc8/4)}' |
| 13:00.17 | cr2 | 1920000 |
| 13:00.44 | cr2 | and not 1843200 |
| 13:01.33 | cr2 | there is 3 vs 2 difference though ;) |
| 13:01.59 | xsacha | this is dump mmu 2: http://pastebin.com/m18d17a16 |
| 13:01.59 | xsacha | list is much bigger |
| 13:03.14 | cr2 | xsacha: dump mmu 2 0x40301688 0x1000 ? |
| 13:03.50 | xsacha | http://pastebin.com/d9e8f5de |
| 13:04.35 | xsacha | just has the block it comes from |
| 13:04.37 | cr2 | tmzt: and it's again ~4 times less than the required clock |
| 13:05.11 | xsacha | im missing something |
| 13:05.29 | cr2 | xsacha: then it's some haret trick. you should talk to Kevin2 |
| 13:05.37 | xsacha | k |
| 13:05.43 | xsacha | Kevin2, hi |
| 13:07.07 | *** join/#htc-linux ccube (n=ccube@ssh.ccube.de) |
| 13:12.29 | *** join/#htc-linux ptitjes (n=didier@80.125.173.60) |
| 13:18.52 | *** join/#htc-linux IceBone (n=IceBone@89-212-106-54.dynamic.dsl.t-2.net) |
| 13:20.31 | cr2 | hmm. then it should be bumped by factor 4 |
| 13:20.45 | cr2 | MSM_CLOCK_REG( 7372800, 8, 0xc8, 0x64, 3, 2, 1, 1) |
| 13:23.43 | *** join/#htc-linux ImCoKeMaN (n=imcokema@pool-173-67-180-143.hrbgpa.fios.verizon.net) |
| 13:23.53 | *** join/#htc-linux fnord_ (n=fnord@24-151-90-116.static.nwtn.ct.charter.com) |
| 13:26.38 | xsacha | are there any msm power management drivers in kernel? |
| 13:39.27 | xsacha | would think there would be one for pm6550 since almost all quallcom's use it |
| 13:48.52 | xsacha | Palm Pre: A Murata WLAN/Bluetooth module, which is ostensibly the same as the one found in the iPhone 3G. This Murata module features a Marvell W8686B12 and a CSR BlueCore 6 solution. |
| 13:49.30 | xsacha | same as my phone |
| 13:52.23 | *** join/#htc-linux pleemans (n=toi@d54C2AAB7.access.telenet.be) |
| 13:53.03 | *** join/#htc-linux nebi (n=nebi@217.142.147.19) |
| 13:57.18 | *** join/#htc-linux vts (n=vts@88-117-92-104.adsl.highway.telekom.at) |
| 13:57.25 | *** part/#htc-linux vts (n=vts@88-117-92-104.adsl.highway.telekom.at) |
| 14:01.13 | *** join/#htc-linux NetRipper (n=netrippe@netripper.nl) |
| 14:09.02 | *** join/#htc-linux vts (n=vts@88-117-92-104.adsl.highway.telekom.at) |
| 14:09.13 | *** part/#htc-linux vts (n=vts@88-117-92-104.adsl.highway.telekom.at) |
| 14:21.17 | Kevin2 | xsacha: Hi |
| 14:21.31 | xsacha | hi |
| 14:21.50 | xsacha | do you know how to get read the i2c bus from haret on a pxa3xx? |
| 14:23.22 | Kevin2 | xsacha: Huh? |
| 14:23.41 | *** join/#htc-linux CoKeSero (n=imcokema@pool-173-67-156-146.hrbgpa.fios.verizon.net) |
| 14:23.49 | xsacha | i want to see data transferred on the i2c line of my pxa3xx using haret |
| 14:24.18 | xsacha | i found the address that i2c commands run on from winmo.. |
| 14:24.32 | pH5 | xsacha: mmutrace the IDBR register |
| 14:24.52 | xsacha | 0x40301688 |
| 14:25.03 | xsacha | thats my IDBR register |
| 14:25.47 | xsacha | can you explain the mmutracing? |
| 14:25.53 | xsacha | i did a dump mmu 2 on that address.. |
| 14:26.13 | Kevin2 | xsacha: Are you using haretconsole? |
| 14:26.16 | xsacha | yes |
| 14:26.23 | xsacha | http://pastebin.com/d9e8f5de |
| 14:27.05 | Kevin2 | Can you post the commands you ran and the output of "wi"? |
| 14:27.13 | xsacha | it's just irqs |
| 14:27.50 | Kevin2 | It would help to see the commands you ran and the output. |
| 14:28.02 | pH5 | I use "addlist mmutrace p2v(0x40301688)" + "wirq" on pxa27x |
| 14:28.28 | xsacha | yeah i did that |
| 14:29.23 | xsacha | http://pastebin.com/m3471381e |
| 14:31.55 | Kevin2 | xsacha: Our you sure the i2c bus was active during that capture? If so, the most likely case is wince is using DMA. See: http://www.handhelds.org/moin/moin.cgi/HaRET_20Tracing_20Details |
| 14:31.57 | xsacha | "MMU table merging disabled" |
| 14:32.08 | xsacha | yeah was using i2c |
| 14:33.23 | xsacha | tried making phone calls, using camera, etc.. had it on for 3 minutes and still only normal irqs |
| 14:36.52 | xsacha | does that MMU table merging disabled mean anything? |
| 14:37.52 | Echo31 | cr2: there are still some error on egpio (cpld), how can i fix it ? |
| 14:40.51 | xsacha | Kevin2, so i should try this loading VirtualCopy method? do i have to watch something when i `set insn` the address? |
| 14:41.29 | Kevin2 | xsacha: I've never seen that work. I'd try watching the DMA registers to see if they're being set for i2c. |
| 14:42.10 | xsacha | oh ok, if i find they are being set... ? i can read off them? |
| 14:42.11 | Kevin2 | xsacha: The MMU merging is an enhancement to try and find changed addresses - it being off wont adversely impact anything - I'll see if I can find the documentation for that. |
| 14:43.49 | xsacha | i have a battery driver that gives me weird values like 0xF510 for temperature, i know to ignore the first 4 bits but no idea how to decode the rest of it. was hoping i could check the temperature in winmo and compare to the word received through i2c bus |
| 14:44.14 | xsacha | same with capacity and voltage. absolutely no documentation on this chip |
| 14:44.34 | xsacha | all i know is the registers that winmo checks, i guess i could check how it decodes them |
| 14:54.26 | *** join/#htc-linux Captnoord (n=Captnoor@dc5147a47b.adsl.wanadoo.nl) |
| 14:57.58 | Captnoord | I am wondering.... do I need to unlock my phone to flash a diff radio in it? I do find some references on the forum that I have to, but also a lot that itsn not required. Can anyone verify this for me? |
| 15:02.18 | Echo31 | hi pH5 |
| 15:02.53 | pH5 | hej Echo31 |
| 15:03.06 | Echo31 | pH5: there are still some error on Athena egpio (cpld 1), how can i fix it ? |
| 15:03.06 | Echo31 | dmesg:Platform driver 'htc-egpio' needs updating - please use dev_ |
| 15:03.06 | Echo31 | EGPIO failed to setup |
| 15:03.06 | Echo31 | htc-egpio: probe of htc-egpio.0 failed with error -22 |
| 15:05.11 | pH5 | Echo31 dev_pm_ops? that's unrelated to the error (just a transition warning) |
| 15:08.04 | Echo31 | pH5: this warning is displayed for cpld2: so, cpld2 works correctly. The problem is cpld1 |
| 15:10.09 | *** join/#htc-linux Red_Lion (n=lion@77.41.93.140) |
| 15:25.52 | pH5 | Echo31: yes, there's something wrong with your cpld1 platform data, thus -22 (-EINVAL) |
| 15:25.55 | *** join/#htc-linux miknix (n=miknix@gentoo/developer/miknix) |
| 15:29.12 | *** join/#htc-linux madCoder- (n=madcoder@c-71-225-60-178.hsd1.pa.comcast.net) |
| 15:32.14 | Echo31 | pH5: How to analyze where is the problem? htc_egpio in debug mode ? |
| 15:38.15 | *** join/#htc-linux mickey|ICE924 (n=M@92.116.182.56) |
| 15:40.54 | Echo31 | pH5: htc-egpio returns failed because it don't compliant with |
| 15:40.54 | Echo31 | .reg_width = 8, |
| 15:40.54 | Echo31 | .bus_width = 8, |
| 15:47.30 | pH5 | yeah, because .bus_width=8 is wrong, I told you already |
| 15:48.04 | xsacha | isn't it 2 for a bus width of 4 and 3 for a bus width of 8? |
| 15:49.23 | *** join/#htc-linux BabelO (n=fcr@unaffiliated/babelo) |
| 15:50.06 | pH5 | .bus_width is in bits, either 16 or 32. |
| 16:00.19 | Captnoord | xsacha: you mean 1 << 3 and 1 << 2 |
| 16:01.30 | xsacha | hmm that'd just be 8 and 4 |
| 16:02.11 | Captnoord | yup |
| 16:03.35 | xsacha | no... |
| 16:04.00 | mickey|ICE924 | morning |
| 16:04.06 | mickey|ICE924 | cr2: got anywhere yet wrt. audio? |
| 16:11.51 | *** join/#htc-linux Echo31 (n=olivier@mir31-4-82-240-194-54.fbx.proxad.net) |
| 16:20.07 | cr2 | mickey|ICE924: i have written a playwav2.c program, the adsp modules are ok, but it still does not work. |
| 16:20.55 | mickey|ICE924 | who can we hang for the audio subsystem in g1? :) |
| 16:24.46 | cr2 | mickey|ICE924: g1 has the code for 3 different amss versions. raph100 has another one, and raph800 is also different from raph100 |
| 16:25.56 | cr2 | mickey|ICE924: but a big part of raph100 -speciifc adsp code is the same as on 6210. |
| 16:26.45 | mickey|ICE924 | ok |
| 16:29.06 | *** join/#htc-linux BabelO (n=fcr@unaffiliated/babelo) |
| 16:44.11 | *** join/#htc-linux fnord_ (n=fnord@24-151-90-116.static.nwtn.ct.charter.com) |
| 16:50.58 | Echo31 | pH5: for .reg_width =16, the egpio_write_cache function fails |
| 16:52.12 | *** join/#htc-linux StarLite (n=nnscript@s55916ca6.adsl.wanadoo.nl) |
| 16:53.31 | Echo31 | pH5: How can i debug htc-egpio ? with pr_debug |
| 16:57.04 | *** join/#htc-linux mickey|train (n=M@92.116.172.137) |
| 17:03.32 | pH5 | Echo31: yes, make sure to #define DEBUG and add debug to the kernel cmdline to see it on the console. |
| 17:03.59 | Echo31 | pH5: ok |
| 17:04.35 | Echo31 | pH5: the crash is located to egpio_writew |
| 17:10.29 | *** join/#htc-linux mickey|ICE924 (n=M@92.116.160.104) |
| 17:16.04 | *** join/#htc-linux stickboy (n=anonymou@ool-457e4101.dyn.optonline.net) |
| 17:19.01 | *** join/#htc-linux solerant-dc (n=kamikaze@adsl-76-202-7-26.dsl.hstntx.sbcglobal.net) |
| 17:27.04 | cr2 | Echo31: most htc cplds have .reg_width=16 |
| 17:30.57 | Captnoord | cr2 the strugle for sound continue's? |
| 17:31.58 | *** join/#htc-linux Echo31 (n=olivier@mir31-4-82-240-194-54.fbx.proxad.net) |
| 17:33.20 | *** join/#htc-linux surge (i=surge@pool-98-118-158-217.bflony.fios.verizon.net) |
| 18:11.59 | *** join/#htc-linux nebi (n=nebi@217.142.147.19) |
| 18:18.28 | Echo31 | pH5: with #define DEBUG in the code and debug in the Haret CMDLINE, i don't yet find the debug messages. |
| 18:21.43 | cr2 | Echo31: can you post the link to htc-egpio.c you are using ? |
| 18:27.20 | Echo31 | cr2: fin the link http://fr.pastebin.ca/1484030. The comments 'echo31' in the code for displaying the traces |
| 18:29.20 | pH5 | Echo31: "#define DEBUG", not "debug", it's case sensitive. |
| 18:30.15 | Echo31 | pH5: thanks |
| 18:30.20 | cr2 | writew(value, ei->base_addr + (reg << ei->bus_shift)); |
| 18:30.27 | cr2 | readw(ei->base_addr + (reg << ei->bus_shift)); |
| 18:30.40 | cr2 | i really don't understand what can go wrong here |
| 18:30.54 | *** join/#htc-linux kiozen_ (n=oeichler@rgnb-5d87de91.pool.einsundeins.de) |
| 18:33.32 | pH5 | Echo31: let's see what the debug output shows. when you have it, also consider enabling the egpio_writew again, to see which write breaks. |
| 18:34.36 | cr2 | Echo31: printk all the parameters to egpio_writew before calling it |
| 18:35.17 | *** part/#htc-linux Echo31 (n=olivier@mir31-4-82-240-194-54.fbx.proxad.net) |
| 18:35.25 | cr2 | egpio->cached_values, shift, ei->reg_mask |
| 18:35.31 | cr2 | too late :) |
| 18:35.55 | *** join/#htc-linux Echo31 (n=olivier@mir31-4-82-240-194-54.fbx.proxad.net) |
| 18:37.54 | cr2 | pH5: i'll pull my hair out with such arithmetic :) ei->bus_shift = fls(pdata->bus_width - 1) - 3; |
| 18:38.11 | Echo31 | pH5: i don't watch the debug messages |
| 18:38.24 | cr2 | Echo31: use printk |
| 18:38.33 | cr2 | <PROTECTED> |
| 18:39.34 | Echo31 | cr2: for the first call of writw, the function crashs |
| 18:40.25 | cr2 | fls - find last bit set |
| 18:40.43 | cr2 | Echo31: return before writing, but after printk |
| 18:41.11 | cr2 | or put at endless loop, whatever |
| 18:42.58 | *** join/#htc-linux fnord_ (n=fnord@24-151-90-116.static.nwtn.ct.charter.com) |
| 18:44.18 | *** join/#htc-linux kiozen (n=oeichler@rgnb-5d87de91.pool.einsundeins.de) |
| 18:45.08 | Echo31 | cr2: printk is ok. i will send the messages |
| 18:51.30 | Echo31 | cr2: pH5: find the messages EGPIO: setting 0 to 0, was 14 |
| 18:51.31 | Echo31 | EGPIO: setting 1 to 0, was 20 |
| 18:51.31 | Echo31 | EGPIO: setting 2 to 12, was 13 |
| 18:51.31 | Echo31 | EGPIO: setting 3 to d, was d0 |
| 18:51.31 | Echo31 | EGPIO: setting 7 to 0, was a |
| 18:59.46 | *** join/#htc-linux kiozen (n=oeichler@rgnb-5d87de91.pool.einsundeins.de) |
| 19:00.24 | Echo31 | the messages written by printk("EGPIO: setting %x to %x, was %x\n", reg, |
| 19:00.24 | Echo31 | (egpio->cached_values >> shift) & ei->reg_mask, |
| 19:00.24 | Echo31 | egpio_readw(ei, reg)); |
| 19:02.04 | pH5 | cr2: maybe I should change that to a switch statement |
| 19:04.44 | pH5 | Echo31: which of the writes fails, and how? |
| 19:06.16 | Echo31 | pH5: I create a counter. For the first value, the system crashes |
| 19:07.37 | *** join/#htc-linux ImCoKeMaN (n=imcokema@pool-71-173-203-229.hrbgpa.fios.verizon.net) |
| 19:08.19 | *** join/#htc-linux nebi (n=nebi@217.142.147.19) |
| 19:09.43 | Echo31 | pH5: the principle: outside the function static int counter ;in the function if counter < 1 Writew and after counter ++ |
| 19:14.26 | *** join/#htc-linux Kevin2 (n=Kevin2@207-237-194-161.c3-0.avec-ubr2.nyr-avec.ny.cable.rcn.com) |
| 19:18.13 | Echo31 | pH5: cr2: I try with if ( egpio_readw(ei, reg) == 0x14 ) |
| 19:18.13 | Echo31 | <PROTECTED> |
| 19:18.13 | Echo31 | & ei->reg_mask, ei, reg); |
| 19:28.55 | Echo31 | pH5: cr2; the first values 'EGPIO: setting 0 to 0, was 14' fails |
| 19:49.44 | *** part/#htc-linux hachiya (n=flash@mail.filmpr.com) |
| 19:54.33 | cr2 | Echo31: you should post the full patch to pH5 tree, so i can apply and test it |
| 19:56.49 | Echo31 | cr2: ok, few minutes |
| 20:00.01 | xsacha | cr2: why does ARM use X * CCCCCCD >> 35 = X/10 ? |
| 20:00.28 | xsacha | wouldnt X*CCD >> 15 = X / 10 would just as well? |
| 20:00.49 | xsacha | or is it to work for all size numbers? |
| 20:14.07 | Echo31 | vr2: pH5: find the patch http://fr.pastebin.ca/1484125 with htc-egpio.c changed |
| 20:17.07 | *** join/#htc-linux pleemans (n=toi@d54C2AAB7.access.telenet.be) |
| 20:20.32 | *** join/#htc-linux Captnoord (n=Captnoor@ip4da104ed.direct-adsl.nl) |
| 20:32.14 | cr2 | Echo31: ok |
| 20:52.54 | xsacha | battery driver works! yipee |
| 21:00.15 | *** join/#htc-linux BabelO (n=fcr@unaffiliated/babelo) |
| 21:00.34 | *** join/#htc-linux darkstar62 (n=darkstar@97-126-107-190.tukw.qwest.net) |
| 21:11.58 | pH5 | xsacha: congrats! |
| 21:17.27 | *** join/#htc-linux darkstar62_ (n=darkstar@m0b0e36d0.tmodns.net) |
| 21:26.57 | xsacha | pH5: is there a better way to add batt_vol and batt_temp props (android looks for) than adding it in to power_supply.h? |
| 21:34.15 | xsacha | i mean i did that and it works but seems bad |
| 21:36.14 | *** join/#htc-linux pr0p (n=pr0p@84.228.139.160) |
| 21:39.46 | kiozen | cr2: around? |
| 21:52.48 | pH5 | xsacha: are those just nonstandard names for the voltage_now and temp properties? |
| 21:54.43 | xsacha | yes |
| 21:55.19 | xsacha | i dont know why they dont use the standard names |
| 21:55.34 | pH5 | urghs, no way to fix userspace? |
| 21:55.53 | pH5 | is it possible to create links in sysfs by now? |
| 21:55.54 | tmzt | custom libhardware |
| 21:56.08 | tmzt | maybe we should go that direction |
| 21:56.33 | xsacha | what does everyone else use? HTC-linux for eg. |
| 21:56.45 | tmzt | make on for standard 2.6 with full enumeration and contribute it for the upstream work |
| 21:57.02 | tmzt | was debian, now custom gentoo |
| 21:57.16 | xsacha | k |
| 21:57.31 | tmzt | we'll need it for asoc support later anyway |
| 22:02.38 | cr2 | har |
| 22:03.03 | tmzt | hello, cr2 |
| 22:03.18 | tmzt | can you look at something quickly? |
| 22:03.29 | cr2 | what ? |
| 22:03.40 | tmzt | this bootloader thing |
| 22:03.42 | cr2 | i'm looking for dcordes |
| 22:03.59 | cr2 | bootloader ? |
| 22:04.55 | *** join/#htc-linux dcordes (n=dcordes@unaffiliated/dcordes) |
| 22:05.20 | dcordes | hi tmzt, cr2 others |
| 22:06.11 | cr2 | hi dcordes |
| 22:10.01 | dcordes | cr2: you need help with anything? |
| 22:10.44 | cr2 | dcordes: will you be here tomorrow , aka today ? |
| 22:11.11 | cr2 | i'd like to commit some patches |
| 22:11.56 | cr2 | and i've written a playwav2.c program, that talks to /dev/msm_snd , and is more or less a standalone player/recorder |
| 22:12.05 | tmzt | cr2: pH5 sets up ltg accounts? |
| 22:12.06 | cr2 | very nice for initramfs. |
| 22:12.14 | cr2 | i have no time for that |
| 22:12.41 | tmzt | I mean for a 2.6.29 tree |
| 22:12.46 | cr2 | tmzt: will leave on tuesday for a vacation, i've told you |
| 22:12.52 | tmzt | yeah |
| 22:14.37 | *** join/#htc-linux dcordes_ (n=dcordes@unaffiliated/dcordes) |
| 22:14.51 | *** join/#htc-linux nebi (n=nebi@217.142.147.19) |
| 22:14.52 | cr2 | dcordes_: we need a qdsp_6150.c file too. |
| 22:15.12 | cr2 | dcordes_: this will create an even bigger mess in the common code. |
| 22:15.31 | cr2 | <PROTECTED> |
| 22:15.33 | cr2 | <PROTECTED> |
| 22:15.39 | dcordes_ | cr2: I have to work from 2pm til 10pm tomorrow |
| 22:15.51 | dcordes_ | CET |
| 22:15.53 | cr2 | ok |
| 22:16.12 | cr2 | i'll wake up early |
| 22:16.19 | dcordes_ | if you are here in the morning we can handle it before I leave |
| 22:16.23 | cr2 | was hacking since 8:00 today. |
| 22:16.38 | cr2 | never happened to me on saturday before ;) |
| 22:16.48 | dcordes_ | was sleeping at that time. had same late shift yesterday |
| 22:16.57 | cr2 | i see |
| 22:17.26 | dcordes_ | wednesday is last day of service |
| 22:17.48 | cr2 | tmzt: you should verify other entries for 6150 |
| 22:17.57 | cr2 | dcordes_: congratulations :) |
| 22:18.02 | tmzt | ok, if I can haret to work |
| 22:18.59 | dcordes_ | cr2: thanks. actually I'm still public servant until end of july but I have the full vacation |
| 22:21.23 | cr2 | ok |
| 22:22.11 | cr2 | tmzt: actually, adapting the adsp_xxxx.c is the main task for making sound work. i have some progress here. |
| 22:22.39 | tmzt | ok |
| 22:22.53 | tmzt | so I need the module numers |
| 22:23.27 | cr2 | i'll pastebin them |
| 22:23.51 | cr2 | and i was looking for these numbers in adsp*.dll |
| 22:24.24 | cr2 | need to cross-compare the tables to other g1 values |
| 22:25.33 | cr2 | tmzt: can you create a wiki table with name, raph100_5200 number, raph800_6150 number, g1_6210, g1_6225, ... columns ? |
| 22:26.03 | cr2 | tmzt: yes, and vogue_4xxx too |
| 22:26.32 | cr2 | so we can compare them like MSM_CLK values for 7200, 7201A and 8250 |
| 22:26.39 | tmzt | not on here |
| 22:26.51 | tmzt | even iris makes it unusable |
| 22:27.57 | tmzt | I can make a csv if you have the data |
| 22:28.32 | cr2 | there are 7 tables in adsp_5200 |
| 22:28.48 | cr2 | i've edited the module table, the first one. |
| 22:28.54 | cr2 | pasting it now |
| 22:31.59 | *** join/#htc-linux Aiv (n=Aiv@office.mdsoftware.pl) |
| 22:32.02 | Aiv | hi |
| 22:32.17 | cr2 | tmzt: https://privatepaste.com/e6su4ozqAR |
| 22:34.47 | Aiv | i trying to get gpios from x1, but haret don't recognize this device so "watch gpios x" give me max 10 gpios |
| 22:35.04 | cr2 | x1 ? |
| 22:35.05 | Aiv | is there any option to configure haret for x1? |
| 22:35.08 | Aiv | xperia |
| 22:35.20 | Aiv | aka htc-kovsky/venus |
| 22:35.37 | cr2 | can you recompile haret ? |
| 22:35.45 | Aiv | why not |
| 22:36.16 | cr2 | is it a registered mtype ? |
| 22:36.27 | tmzt | ask dcordes |
| 22:36.31 | Aiv | yes, AFAIR 2006 |
| 22:36.37 | tmzt | when he gets back |
| 22:36.47 | cr2 | dcordes_: here ? |
| 22:37.08 | Aiv | i found some gpios list of couse but i need to check this manualy at my phone |
| 22:37.10 | tmzt | cr2: you want those in csv? |
| 22:37.11 | cr2 | Aiv: is this mtype already in the haret CVS/git ? |
| 22:37.16 | tmzt | or html table |
| 22:37.24 | cr2 | tmzt: mediawiki |
| 22:37.34 | Aiv | cr2 i dont think so, w8 i will check |
| 22:37.53 | cr2 | tmzt: you can add it to htc-linux.org wiki |
| 22:38.09 | tmzt | not on my phone |
| 22:38.14 | tmzt | I mean |
| 22:38.18 | cr2 | Aiv: you can hijack some existing number, but it's not nice. |
| 22:38.29 | cr2 | tmzt: ah, ok |
| 22:39.36 | dcordes_ | yea semi present |
| 22:39.49 | dcordes_ | Aiv: any x1 success? did the initrd usage help? |
| 22:39.56 | cr2 | dcordes_: is kovky registered, and in haret ? |
| 22:40.06 | Aiv | dcordes_ yep, thx |
| 22:40.11 | dcordes_ | well it's in the arm registry |
| 22:40.16 | Aiv | dcordes_ i just run git and make patches |
| 22:40.33 | Aiv | dcordes_ now i trying to got gpios my self |
| 22:41.07 | dcordes_ | cr2: shouldn't kovsky gpio work in the same way as raphael gpio?? |
| 22:41.16 | dcordes_ | I mean we have no cplds and it's the same chip |
| 22:42.24 | cr2 | dcordes_: haret needs to detect it as 7201A |
| 22:42.39 | cr2 | which is hardcoded into machtypes.txt |
| 22:43.17 | Aiv | Welcome, this is HaRET pre-0.5.3-20081105_214354 running on WindowsCE v5.2 |
| 22:43.17 | Aiv | Minimal virtual address: 00010000, maximal virtual address: 7FFFFFFF |
| 22:43.17 | Aiv | Detected machine Generic MSM7xxx/MSM7xxx (Plat='PocketPC' OEM='X1i') |
| 22:43.18 | Aiv | CPU is ARM ARM arch 6 stepping 2 running in system mode |
| 22:43.38 | dcordes_ | 7xxx no good |
| 22:43.44 | *** join/#htc-linux nebi (n=nebi@217.142.147.19) |
| 22:44.05 | dcordes_ | Kevin2: are you there? |
| 22:44.16 | Aiv | i dont see anything about x1/kovsky in http://linuxtogo.org/cgi-bin/gitweb.cgi?p=groups/haret/haret.git;a=blob;f=src/mach/machlist.txt;h=2eea55ad4cd23dbac2a9abcf53f8604886aa5071;hb=577253518ce0dcb3ff754cc294a03e40d94f15f6 |
| 22:44.39 | cr2 | tmzt: i think i have found these enums in the .dll |
| 22:44.53 | cr2 | tmzt: but i need to compare them value for value |
| 22:45.02 | cr2 | tmzt: and it's for my raph100 only |
| 22:45.18 | tmzt | haret? |
| 22:45.29 | tmzt | you have the raph800 dlls? |
| 22:48.20 | cr2 | no |
| 22:48.34 | cr2 | let me check it for 100 first |
| 22:49.36 | dcordes_ | Aiv: so we need to add it along with blackstone |
| 22:50.49 | Aiv | dcordes_: if i set mtype in haret to blackstone (1951) i will get baclstone specyfic gpios? |
| 22:51.45 | dcordes_ | Aiv: no there's no blackstone in the kenrel |
| 22:51.49 | cr2 | Aiv: no, but you can compare them to the wiki values |
| 22:51.56 | dcordes_ | Aiv: did you try what I proposed and open keyboard to see if it triggers the sd detection? |
| 22:52.08 | cr2 | dcordes_: should i pastebin playwav2.c ? |
| 22:52.25 | dcordes_ | cr2: yes. I can prepare new initrd with it tomorrow |
| 22:52.29 | Aiv | dcordes_ yes, but my kernel dont recognized it, but now i will try to focus on keyboard, navipanel etc |
| 22:52.39 | cr2 | ok |
| 22:52.43 | dcordes_ | cr2: have it stay a bit longer in the pastebin |
| 22:52.46 | Aiv | dcordes_: with initrd and ssh i dont need sd now |
| 22:52.48 | Kevin2 | dcordes_: Hi |
| 22:53.19 | cr2 | Kevin2: can you merge some new devices to haret machtypes.txt ? |
| 22:53.49 | Kevin2 | cr2: Sure - just send a patch. |
| 22:53.51 | dcordes_ | we could also put the other new msm. topaz etc |
| 22:53.53 | cr2 | Kevin2: maybe it's better to register them dynamically by reading and parsing machtypes.txt ? |
| 22:54.06 | tmzt | cmonex should have the Plat names |
| 22:54.22 | Kevin2 | cr2: That would be nice, but it's not setup that way currently. |
| 22:54.26 | *** join/#htc-linux nebi (n=nebi@217.142.147.19) |
| 22:54.33 | cr2 | ok |
| 22:54.38 | cmonex | tmzt? |
| 22:54.50 | tmzt | Plat/Oem for topaz, etc. |
| 22:55.01 | tmzt | for haret |
| 22:55.12 | Aiv | "ibit gpios X" it's nor working when haret dont recognise my phone? |
| 22:55.31 | tmzt | it things you have an older cpu |
| 22:56.40 | Aiv | I try to ignore flooding gpio (9). After ibit gpios 9 i have gpio 9 in watch gpios command |
| 22:57.07 | tmzt | it wonqt work until it's detected properly |
| 22:57.14 | Aiv | k |
| 22:57.41 | tmzt | cmonex: I mean we need the Plat and Oem names that ce reports on diamond2 and tp2 |
| 22:57.49 | tmzt | to add them to haret |
| 22:58.00 | Aiv | so what i can do to correct detect it ? ;] |
| 22:58.26 | cmonex | tmzt : |
| 22:58.31 | cmonex | not sure what you mean, |
| 22:58.35 | cmonex | can you give me what they are for raphael? |
| 22:58.39 | cmonex | so i can see what you mean |
| 22:59.45 | cr2 | tmzt: the module tables seem to be the same as on 6210 |
| 22:59.50 | tmzt | I would but I just realized mines not dxetected either |
| 22:59.53 | tmzt | good |
| 22:59.57 | tmzt | hold on |
| 23:00.12 | cr2 | tmzt: but the queue offsets seem to be different |
| 23:00.38 | tmzt | ok |
| 23:00.44 | tmzt | same as smd |
| 23:00.48 | tmzt | is different |
| 23:01.57 | Aiv | guys, can i help with something with adding support for kovsky to haret? |
| 23:01.57 | tmzt | cmonex: Plat='PocketPC' OEM='RAPH500' |
| 23:02.15 | tmzt | Aiv: get cegcc toolchain |
| 23:02.55 | tmzt | hah, I have Detected Generic MSM7xxx/MSM7xxx as well |
| 23:04.10 | Aiv | tmzt: and compile haret from git? Do i dont need to add somethinf to machtypes? |
| 23:05.20 | tmzt | yes, that's easy |
| 23:05.27 | cmonex | ah |
| 23:05.32 | cmonex | well they're all PocketPC then :) |
| 23:05.35 | cmonex | the OEM will be the modelid |
| 23:05.40 | tmzt | ok |
| 23:05.42 | cmonex | TOPAxxx and RHODxxx |
| 23:05.47 | cmonex | xxx is for gsm anyway.. |
| 23:05.54 | cmonex | TOPA100, 210, etc-... |
| 23:06.04 | cmonex | several modelids exist, rebranded devices have diff one |
| 23:06.10 | cmonex | pointless usually :) |
| 23:06.30 | tmzt | if we had proper cpu detection we wouldn't need it |
| 23:06.36 | tmzt | but that's how haret works |
| 23:06.54 | Aiv | tmzt: compiling is easy, but that's all? after i compile from git, haret detects my phone better than now? |
| 23:07.16 | tmzt | no, add a line to a .txt file first |
| 23:07.24 | tmzt | then compile |
| 23:09.59 | Aiv | tmzt: ok, i will try |
| 23:12.03 | cr2 | dcordes_: how many days ? |
| 23:18.56 | cr2 | tmzt: https://privatepaste.com/e81A1bINab |
| 23:21.23 | tmzt | and the patches for kernel? |
| 23:22.51 | cr2 | i need to sync to the current git first |
| 23:23.27 | tmzt | ok |
| 23:23.40 | tmzt | how do I find the mailboxes? |
| 23:23.48 | cr2 | i've left msm72xx_enable_audpp() for you :) |
| 23:23.52 | tmzt | if they're different |
| 23:23.56 | tmzt | ok |
| 23:23.56 | Aiv | hmm |
| 23:23.57 | Aiv | make |
| 23:23.57 | Aiv | mkdir out |
| 23:23.57 | Aiv | <PROTECTED> |
| 23:23.58 | Aiv | make: *** [out/machines.o] Floating point exception |
| 23:24.08 | tmzt | wow |
| 23:24.27 | tmzt | did it compile before your changes? |
| 23:24.56 | Aiv | i dont know, w8 i will check |
| 23:25.53 | Aiv | some results |
| 23:26.32 | Aiv | i only add line "X1i,MSM7200A,X1i,2006" to machlist.txt |
| 23:27.30 | tmzt | try without it |
| 23:27.43 | tmzt | X1i? |
| 23:28.01 | Aiv | i tried |
| 23:28.16 | Aiv | name from current haret: (Plat='PocketPC' OEM='X1i') |
| 23:28.47 | Aiv | this is htc venus/kovsky |
| 23:29.36 | Aiv | strange |
| 23:29.37 | Aiv | $ /opt/mingw32ce/bin/arm-mingw32ce-cpp |
| 23:29.37 | Aiv | Floating point exception |
| 23:29.42 | Aiv | this is problem with toolchain |
| 23:30.03 | tmzt | yeah, thought so |
| 23:30.22 | tmzt | dcordes: do you have an X1i? |
| 23:37.38 | Aiv | ok, on another server toolchains is working fine |
| 23:41.12 | Aiv | out/mach-autogen.cpp:608: error: expected class-name before '{' token |
| 23:41.12 | Aiv | out/mach-autogen.cpp: In constructor 'MachX1i::MachX1i()': |
| 23:41.12 | Aiv | out/mach-autogen.cpp:611: error: 'name' was not declared in this scope |
| 23:41.12 | Aiv | out/mach-autogen.cpp:612: error: 'PlatformType' was not declared in this scope |
| 23:41.12 | Aiv | out/mach-autogen.cpp:613: error: 'OEMInfo' was not declared in this scope |
| 23:41.13 | Aiv | out/mach-autogen.cpp:614: error: 'machType' was not declared in this scope |
| 23:41.15 | Aiv | out/mach-autogen.cpp: At global scope: |
| 23:41.17 | Aiv | out/mach-autogen.cpp:617: error: cannot convert 'MachX1i*' to 'Machine*' in initialization |
| 23:41.19 | Aiv | make: *** [out/mach-autogen.o] Error 1 |
| 23:41.21 | Aiv | ;[ |
| 23:41.35 | Aiv | something is wrong with code generation |
| 23:41.38 | Aiv | i add line: X1i, MSM7200A,X1i, HTCVENUS |
| 23:41.47 | cr2 | Aiv: it should match the mtype name |
| 23:42.39 | Aiv | root@fortuna:~/haret/haret# cat include/mach-types.h | grep HTCVENUS |
| 23:42.39 | Aiv | #define MACH_TYPE_HTCVENUS 2006 |
| 23:44.36 | Aiv | k, there is no definition for MachineMSM7200A |
| 23:46.13 | tmzt | 7201A |
| 23:47.01 | Aiv | i just create alias in mach-msm.h |
| 23:47.12 | Aiv | downloading exe now |
| 23:48.32 | *** join/#htc-linux fnord_ (n=fnord@24-151-90-116.static.nwtn.ct.charter.com) |
| 23:49.04 | cr2 | tmzt: 2a not 2b |
| 23:49.28 | tmzt | ioctl? |
| 23:49.46 | cr2 | <PROTECTED> |
| 23:49.53 | cr2 | for 6210 |
| 23:49.59 | tmzt | ok |
| 23:50.08 | cr2 | but it should be 2a on 5200 |
| 23:50.09 | IceBone | 1701-D |
| 23:50.13 | tmzt | thatqs for openmax? |
| 23:51.38 | *** join/#htc-linux nebi (n=nebi@217.142.147.19) |
| 23:54.06 | Aiv | Detected machine X1i/MSM7xxxA (Plat='PocketPC' OEM='X1i') |
| 23:54.23 | cr2 | tmzt: QDSP_RTOS_NO_QUEUE distribution in enums is very helpful |
| 23:54.42 | cr2 | QDSP_RTOS_NO_QUEUE = -1 = 0xffffffff |
| 23:55.24 | Aiv | but i still done have correctl list of gpios ;[ |
| 23:55.34 | cr2 | and QDSP_MODULE_MAX |
| 23:55.43 | Aiv | and ibit gpios 9 not ignore gpio 9 at watching |
| 23:55.54 | Guest94912 | lost his Tilt |
| 23:56.30 | Raketemensch | such a bummer, it's just gone. |
| 23:56.46 | cr2 | the offset tables are completely different. not even a single match ;) |
| 23:57.19 | Aiv | oo wirq not show something |
| 23:57.25 | Aiv | but watch gpios still not ;[ |
| 23:57.35 | *** join/#htc-linux ccube (n=ccube@ssh.ccube.de) |
