| 00:04.55 | BabelO | why when doing 'cat /proc/interrupts' i didn t see acx irq ? |
| 00:18.23 | pH5 | .zZ |
| 00:35.01 | *** join/#htc-linux ellisway (n=ellis@host-87-74-241-174.bulldogdsl.com) |
| 01:04.54 | lkcl | pH5 - got it wrong: it's all GPIO 20. |
| 01:24.22 | lkcl | holy cow, this audio routing is just... horrendously complicated. |
| 02:03.17 | *** join/#htc-linux LunohoD (n=alex@e180083183.adsl.alicedsl.de) |
| 02:33.20 | *** join/#htc-linux jgw_ (n=jgw@c-24-22-48-124.hsd1.mn.comcast.net) |
| 05:06.29 | *** join/#htc-linux psokolovsky__ (n=psokolov@237.usernat.ip.net.ua) |
| 05:50.42 | *** join/#htc-linux LunohoD (n=alex@e180083183.adsl.alicedsl.de) |
| 06:46.47 | *** join/#htc-linux Kevin1 (n=Kevin@207-237-52-30.c3-0.avec-ubr12.nyr-avec.ny.cable.rcn.com) |
| 07:39.54 | *** join/#htc-linux rmoravcik (n=rmoravci@icm2-orange.orange.sk) |
| 08:54.20 | *** join/#htc-linux RoEn_PC (n=roen@p54A6497F.dip.t-dialin.net) |
| 08:59.18 | *** join/#htc-linux booba (n=booba@AAnnecy-256-1-94-95.w90-10.abo.wanadoo.fr) |
| 09:01.37 | *** join/#htc-linux pH5 (n=ph5@e178195199.adsl.alicedsl.de) |
| 10:01.05 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
| 10:04.54 | *** join/#htc-linux TripleDES (n=sergio@89.129.175.52) |
| 10:53.27 | *** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
| 11:18.55 | *** join/#htc-linux psokolovsky__ (n=psokolov@237.usernat.ip.net.ua) |
| 11:21.07 | goxboxlive | cr2: Are you around? |
| 11:49.20 | lkcl | allo darlins |
| 11:49.32 | pH5 | hey goxboxlive, lkcl |
| 11:49.36 | BabelO | hi lkcl |
| 11:49.42 | BabelO | hi pH5 |
| 11:49.50 | pH5 | hi BabelO |
| 11:49.51 | lkcl | moorning all |
| 11:50.18 | BabelO | how goes sound ? |
| 11:51.53 | lkcl | huh. still digging. |
| 12:02.21 | goxboxlive | hi |
| 12:10.12 | lkcl | allo goboxlive |
| 12:11.04 | lkcl | the sound initialisation is pretty horrendous - on a par with the init sequence for the gsm radio module - but it's buried in amongst some very comprehensive audio switching. |
| 12:11.20 | lkcl | i need to start identifying things to eliminate the other options. |
| 12:12.15 | lkcl | cr2 already identified what's involved with sound - _related_ to sound |
| 12:12.23 | lkcl | Babel0: do you have a bluetooth headset? |
| 12:13.14 | BabelO | lkcl: i use my computer with dongle to simulate headset :( |
| 12:13.17 | lkcl | if so, can you try doing a phone call whilst doing wasic3gpio? |
| 12:13.23 | lkcl | oh coo - that'll do |
| 12:13.27 | lkcl | that would do it. |
| 12:14.15 | lkcl | you can at least use the dongle to associate and confuse the blueangel into thinking it's attached to a headset, right? |
| 12:14.29 | BabelO | yes |
| 12:15.07 | BabelO | i ve tried same thing to simulate a2dp headset with no sucess |
| 12:15.19 | lkcl | ok - cool. would you mind trying a wasic3gpio and seeing if you can coax it into playing a sound? |
| 12:15.30 | lkcl | oh - hang on: i've got a motorola bluetooth headset.... |
| 12:15.38 | lkcl | <PROTECTED> |
| 12:18.09 | BabelO | ok |
| 12:23.56 | lkcl | arse - i just remembered where it is: on a shelf in holland. |
| 12:24.27 | lkcl | ok what else can i try.... i know: phone call conversation recording. |
| 12:24.39 | lkcl | sim card. ha ha. now i gotta find a sim card :) |
| 12:26.15 | BabelO | lkcl: i ll try bluetooth, need to buy some food before :) |
| 12:27.30 | pH5 | BabelO: is there a howto somewhere how to emulate a bt headset with an usb dongle? |
| 12:32.06 | BabelO | pH5: you have only to change you hci.conf class |
| 12:33.10 | BabelO | or try this howto :) |
| 12:33.13 | BabelO | http://www.soft.uni-linz.ac.at/_wiki/tiki-index.php?page=ProjectBluezHandsfree |
| 12:33.32 | goxboxlive | What is the easiest way to se the diff in lib/modules in two diffrent folders. SO it just list the modules that is missing or vica versa ( iam coamparing two diffnets images) |
| 12:33.34 | BabelO | it work for me :) |
| 12:33.59 | pH5 | BabelO: cool, thanks :) |
| 12:34.52 | *** join/#htc-linux FossiFoo (n=Fossi@e176119138.adsl.alicedsl.de) |
| 12:35.38 | pH5 | goxboxlive: I'd use a gui tool like meld or kdiff3 |
| 12:35.44 | lkcl | or dirdiff. |
| 12:36.08 | pH5 | hmm |
| 12:36.18 | *** join/#htc-linux WizMaui (n=WizMaui@62.112.90.231) |
| 12:36.42 | lkcl | i would use find . on the two locations, cat the output to files and then do a diff on the two files |
| 12:36.51 | lkcl | find /lib/modules.1 > f |
| 12:36.56 | lkcl | find /lib/modules.2 > g |
| 12:37.00 | lkcl | diff -u f g |
| 12:37.22 | goxboxlive | thx |
| 12:58.34 | lkcl | ok there's nothing obvious - no extra switching - when record is enabled on a phonecall. CKEN i2s goes on, and the i2s port goes mad gpio 28-32, but that's all. ok - gotta find my headset, next |
| 12:58.46 | lkcl | see if there's any gpio changed on recording, then... |
| 13:00.01 | BabelO | lkcl: how do you start reccording voice call on BA ? |
| 13:00.14 | lkcl | button in top right hand corner. press and hold. |
| 13:00.20 | lkcl | top left. sorry. |
| 13:02.03 | BabelO | lkcl: ok, i try again.few days ago cr2 tell me to test that, but i record only room noise, not phone call |
| 13:02.16 | lkcl | ahh. interesting. |
| 13:03.12 | lkcl | A10. hooray! |
| 13:03.26 | lkcl | A10 goes on instead of B15 when sound is played, through headphones. |
| 13:04.17 | *** join/#htc-linux asylumed (n=insanity@196.211.28.91) |
| 13:05.16 | BabelO | hi asylumed |
| 13:05.28 | asylumed | BabelO: heya |
| 13:05.53 | BabelO | lkcl yes, record phone call, but heard some room noise too |
| 13:06.00 | lkcl | ok. |
| 13:06.17 | lkcl | huh. another 3 - A6, A7 and A8 are all related to headphone record. |
| 13:07.18 | BabelO | ok going to talk to Ronald to have some food :) |
| 13:08.08 | lkcl | Babel!!! |
| 13:08.15 | lkcl | get yourself some better food, damnit! |
| 13:08.28 | lkcl | honestly... |
| 13:10.27 | *** join/#htc-linux JTRipper (n=jtripper@p54B1DB71.dip.t-dialin.net) |
| 13:14.31 | lkcl | ok. A8 and A9 flip from 0 to 1 and 1 to 0 when headphone record is active, and back again when main-mic is used. |
| 13:14.51 | lkcl | so, A8=0, A9=1 record from main-mic; A8=1, A9=0 record from headphone. |
| 13:14.53 | lkcl | jeez. |
| 13:14.59 | lkcl | ok. phone call. |
| 13:15.54 | lkcl | cool. the headset button works. |
| 13:27.23 | BabelO | cr2: i ve tested your code from htcuniversal_acx in xanadux cvs and got this http://pastebin.ca/raw/281060 |
| 13:49.17 | *** join/#htc-linux lkcl (n=lkcl@5ac267a5.bb.sky.com) |
| 13:51.11 | *** join/#htc-linux psokolovsky (n=psokolov@237.usernat.ip.net.ua) |
| 14:03.59 | BabelO | hi psokolovsky |
| 14:04.07 | psokolovsky | Hi! |
| 14:04.34 | pH5 | hey psokolovsky |
| 14:05.09 | psokolovsky | pH5, I'm going to make rts-sa1100 change soon. |
| 14:05.11 | psokolovsky | rtc |
| 14:05.31 | BabelO | psokolovsky: do you have a dump of /proc/drovers/acx_wlan0_diag ? |
| 14:05.37 | BabelO | after a ifconfig wlan0 up |
| 14:06.06 | psokolovsky | BabelO, nope ;-) |
| 14:06.59 | BabelO | i want to see if we have same problem :) |
| 14:08.11 | psokolovsky | I so far at the same stage - fw load ;-) |
| 14:10.24 | BabelO | ok.i see tx some tx buffer overflow on mine |
| 14:13.30 | rmoravcik | BabelO: http://pastebin.ca/281090 |
| 14:14.38 | pH5 | http://en.pastebin.ca/281092 on hx4700 |
| 14:15.08 | BabelO | thanks rmoravcik |
| 14:15.13 | BabelO | thanks pH5 |
| 14:15.25 | psokolovsky | pH5, that overlay thing, it's pxa2*7*x-specific? |
| 14:15.36 | pH5 | yes |
| 14:16.28 | rmoravcik | psokolovsky: i have a two question |
| 14:16.44 | pH5 | http://www.intel.com/design/pca/applicationsprocessors/faq/pxa27x_faq.htm#lcd |
| 14:16.50 | psokolovsky | rmoravcik, sure, if I'll be able to answer |
| 14:17.03 | rmoravcik | psokolovsky: h4000 has brf6101 or brf6150? |
| 14:17.37 | psokolovsky | pH5, good link, thanks |
| 14:17.38 | lkcl | http://wiki.xda-developers.com/index.php?pagename=BlueangelSound |
| 14:18.13 | lkcl | i managed to once get acx to work, on the hw6915. |
| 14:18.17 | psokolovsky | rmoravcik, we haev it as "brf6100" in our list. I myself didn't disable or peered too careful into photos ;-) |
| 14:18.50 | lkcl | i then went 'hooray' and then it never worked after that. |
| 14:19.14 | lkcl | that was about... three weeks ago, now.... |
| 14:19.27 | rmoravcik | psokolovsky: ok, do you have working it with 115200 or 921600 too? |
| 14:19.59 | lkcl | Babel0: i've created a sound page for the blueangel, because of the amount of stuff involved. |
| 14:20.23 | lkcl | hopefully by a process of elimination it will be possible to identify all of the IO pins used in wavedev.dll |
| 14:20.40 | psokolovsky | rmoravcik, nope, 115200 only. I heard 921600 needs firmware or something... |
| 14:21.12 | BabelO | lkcl: ok |
| 14:21.15 | lkcl | oo. |
| 14:21.19 | lkcl | i just spotted something. |
| 14:21.27 | BabelO | rmoravcik: , psokolovsky firmware is bts file in your wiondows directory |
| 14:21.44 | BabelO | load it usinf hciattach |
| 14:21.57 | psokolovsky | BabelO, h4000 doesn't have such actually ;-I |
| 14:22.41 | rmoravcik | psokolovsky: ok, and the second, did you remember about my problem with ds1wm and asic3 interrupt processing overrun? |
| 14:22.48 | lkcl | GPIO89 is switched _very_ fast. |
| 14:22.52 | cr2 | pH5: the overlay patch is too intransparent, maybe we can split it pieces. |
| 14:22.54 | BabelO | psokolovsky: i think it can work |
| 14:23.08 | cr2 | ooh. so much to read :) |
| 14:23.21 | psokolovsky | rmoravcik, yep, you're not the one ;-). Was able to solve? |
| 14:23.30 | lkcl | GPIO89 is set to on then off, within a few assembly instructions. that could mean that it's uda1380 reset. |
| 14:23.44 | lkcl | allo cr2. you poor thing. |
| 14:23.56 | cr2 | psokolovsky: 921600 does not need firmware. what is uploaded by BTS, is not really the firmware. |
| 14:24.00 | pH5 | lkcl: sounds like it. is there a delay inbetween, by chance? |
| 14:24.03 | psokolovsky | cr2, we really shouldn't apply it at all, unless we have good reasons for (and teh reason is obviously too good userspace support) |
| 14:24.09 | pH5 | hey cr2 |
| 14:24.09 | lkcl | no - no delay, whatsoever. |
| 14:24.14 | pH5 | what are your plans regarding 23c3? |
| 14:24.18 | psokolovsky | cr2, who care how you call it ;-) |
| 14:24.33 | lkcl | it's like only 2 instructions between the set and clear |
| 14:24.48 | rmoravcik | psokolovsky: no, but problem is with bit9 in asic3_intr register (owm) |
| 14:24.49 | psokolovsky | cr2, does it need that param file or no? whata re the diff brf6100 vs brf6150? |
| 14:24.54 | cr2 | psokolovsky: there are some good parts in this patch that can be merged without any question. |
| 14:25.07 | cr2 | not so much at once :D |
| 14:25.20 | pH5 | :) |
| 14:25.28 | psokolovsky | cr2, need reason for ;-) "good" is not enough ;-) otherwise, we'll get it via mainline eventually... |
| 14:25.34 | lkcl | hey cr2 you need a spare brain, or something? i got one here i'm not using... |
| 14:25.58 | cr2 | psokolovsky: the switch to 921600 can be done with 'hcitool', it's only once command, and does not depend on anything else. |
| 14:26.01 | psokolovsky | rmoravcik, that's "known issue". it happens wgen corresponding interrupt is not processed correctly. |
| 14:27.28 | rmoravcik | psokolovsky: and do you have a idea why it happens? |
| 14:27.30 | cr2 | psokolovsky: BabelO uses a BRF6150 firmware on his BA, but the command parameters differ a bit from "normal" BRF6150 |
| 14:28.18 | psokolovsky | rmoravcik, you don't have proper handler for w1 irq. likely, ds1wm doesn't do as it should. |
| 14:28.45 | cr2 | the BTS script does not upload firmware, it modifies onboard RAM and patches GPIO & so on. |
| 14:29.35 | psokolovsky | cr2, well, if you could capture this wisdom somewhere ;-). I personally again back form adv. things like BT to base stuff like SOC ;-\ |
| 14:29.52 | lkcl | ... blueangel bluetooth works?? |
| 14:29.56 | psokolovsky | But I guess that answers rmoravcik's q, of course |
| 14:30.14 | BabelO | lkcl: yes it works :) |
| 14:30.17 | lkcl | cool. |
| 14:31.00 | cr2 | lkcl: did you find out what is gpio89 ? |
| 14:31.28 | rmoravcik | psokolovsky: i will look to ds1wm |
| 14:31.54 | lkcl | cr2: i have some guesses - i think it might be uda1380 reset. it's switched to 1 and then to 0 _very_ quickly, within 2 instruction cycles. |
| 14:32.13 | BabelO | lkcl, cr2: gpio 85 is strange too |
| 14:32.27 | pH5 | lkcl: uda1380 power-on reset wants at least 1µs! |
| 14:32.35 | psokolovsky | rmoravcik, please keep k-d posted on adapting asic3 to ds1wm (or vice-versa ;-)), I may start to look into that too, so we don't duplicate... |
| 14:33.09 | lkcl | pH5: well, the thing can be 'blipped' and then you'd have to wait a bit - that i wouldn't.... 1 second??? |
| 14:33.21 | cr2 | pH5: i was there last year for a day, but it was a bit stressful. left at 7:00 (it was snowing on that day, so changed the tires ;), was in Berlin at 14:00, anf left at 2:00 on the next day. |
| 14:33.27 | pH5 | lkcl: 1 microsecond |
| 14:33.32 | lkcl | cr2: .text:015B9D58 loc_15B9D58 |
| 14:33.36 | rmoravcik | psokolovsky: i agree |
| 14:34.03 | goxboxlive | cr2: Do we need bfudb or can we remove it from our modules? |
| 14:34.10 | lkcl | yes. 1us. i've seen that in wavedev, pH5. when the clock is running. |
| 14:34.20 | goxboxlive | s/bfudb/bfusb/ |
| 14:34.37 | cr2 | goxboxlive: what is bfusb ? |
| 14:34.49 | goxboxlive | HCI BlueFRITZ! USB driver (BT_HCIBFUSB) |
| 14:34.51 | cr2 | lkcl: need to log in to your server. |
| 14:34.58 | cr2 | no. |
| 14:35.03 | lkcl | que? oh, ok. |
| 14:35.05 | BabelO | rmoravcik: pH5 , i got tx_desc_overfl 35 on my acx http://pastebin.ca/raw/281105 |
| 14:35.08 | cr2 | we don't have BT usb. |
| 14:35.12 | goxboxlive | ok, the remove it form defconfig |
| 14:35.19 | lkcl | .text:015B9D58 LDR R0, =PXAGPIO |
| 14:35.19 | lkcl | .text:015B9D5C CMP R1, #0 |
| 14:35.19 | lkcl | .text:015B9D60 LDR R1, [R0] |
| 14:35.19 | lkcl | .text:015B9D64 MOV R0, #0x2000000 |
| 14:35.19 | lkcl | .text:015B9D68 ADDNE R2, R1, #0x20 |
| 14:35.20 | lkcl | .text:015B9D6C STRNE R0, [R2] ; lkcl: GPIO89 on |
| 14:35.22 | lkcl | .text:015B9D70 ADDEQ R2, R1, #0x2C |
| 14:35.24 | lkcl | .text:015B9D74 STREQ R0, [R2] |
| 14:35.42 | lkcl | ; lkcl: GPIO89 off |
| 14:35.53 | goxboxlive | also wich one of the rtc drivers should we use. SA1100 or rtc-pcf8563 or de we need both? |
| 14:36.04 | goxboxlive | I amcleaning up the machine file in OE |
| 14:36.33 | cr2 | lkcl: looks like a set_gpio(on/off) function. who calls it ? |
| 14:37.21 | cr2 | goxboxlive: wince copies the data from rtc-pcf8563 to pxa27x-rtc and then uses only pxa27x-rtc |
| 14:37.27 | cr2 | at boot. |
| 14:37.44 | lkcl | cr2: it's in amongst a whole stack of bit-tests. if 0x1 do this (set some asic3 gpio) if 0x2 do this (set the ssp) if 0x4 do this (blip GPIO89) if 0x8 do this (set the i2s) etc. |
| 14:38.12 | cr2 | lkcl: is it in the ioctl section ? |
| 14:38.31 | lkcl | i wouldn't know the ioctl section if it bit me on the nose :) |
| 14:38.43 | pH5 | :D |
| 14:38.45 | goxboxlive | cr2: Does that mean we can remove SA1100 from the defconfig. |
| 14:39.11 | pH5 | goxboxlive: no, sa1100 is the driver used for the pxa27x-rtc, currently |
| 14:39.28 | pH5 | it is also needed for alarm wake-up |
| 14:39.31 | goxboxlive | hmm it is not enabled in the defconfig |
| 14:39.45 | cr2 | pH5: is it possible to have multiple rtc ? |
| 14:39.53 | goxboxlive | hmm sorry, it was enabled |
| 14:40.01 | cr2 | goxboxlive: not the char/rtc. |
| 14:40.16 | pH5 | cr2: sure |
| 14:40.54 | goxboxlive | This one is enabled: SA11x0/PXA2xx (RTC_DRV_SA1100) and the module is called rtc-sa1100. |
| 14:42.01 | goxboxlive | and this one: rtc-pcf8563 |
| 14:42.24 | cr2 | yes. leave them both enabled. |
| 14:42.27 | goxboxlive | Ok |
| 14:43.15 | cr2 | i'm thinking about porting the pxa_camera driver from a780 kernel. not the best code, still better than nothing. |
| 14:43.44 | cr2 | are there any other pxa CIF drivers around ? |
| 14:44.33 | lkcl | cr2: cool. |
| 14:45.10 | lkcl | didn't you find one somewhere? someone did... i'm sure it was you who was asking me to contact the author to see if they would add in support for the device on the sable |
| 14:45.11 | pH5 | cr2: I saw some in the e680 2.4.17 kernel, but never had a closer look |
| 14:45.21 | cr2 | but we may need the pxa27x overlay for it. |
| 14:45.31 | cr2 | pH5: a780==e680 |
| 14:45.36 | pH5 | aha :) |
| 14:45.54 | cr2 | it's the same kernel tree. |
| 14:46.28 | lkcl | oh - my a780 came back from the repair centre by mistake. i had to send it back to them again :) |
| 14:46.41 | BabelO | cr2: this one ? http://pastebin.ca/281116 |
| 14:46.43 | cr2 | i was looking for its firmware and probably have found a warez cite on sf.net :) |
| 14:46.47 | lkcl | they meant to send it to the manufacturer for them to replace it |
| 14:46.51 | BabelO | and i ve same for adcm2650 |
| 14:47.10 | cr2 | BabelO: you don't have pxa CIF. |
| 14:47.23 | lkcl | cr2: oops! well, probably a good idea to copy what's there and then notify the sf.net staff to have it removed |
| 14:47.29 | cr2 | you need to find the location of capture registers on AT3200. |
| 14:47.45 | lkcl | ohhh thoose.... ahh joy. |
| 14:48.07 | lkcl | Babel0: the camera on the BA and Him is via the ati. |
| 14:48.31 | lkcl | which means we're going to have fun with the hermes, cr2. the ati chip on that controls quite a lot of stuff... |
| 14:48.58 | cr2 | lkcl: there is also a closed forum there, i need to look at it. |
| 14:49.15 | cr2 | lkcl: vsfb first :) |
| 14:49.16 | lkcl | but then again, the wince kernel is 6mb on the hermes and is, i believe, involved in both the bootload process and the os, and has lots of debug strings in it. |
| 14:49.21 | lkcl | cr2: ack :) |
| 14:50.00 | cr2 | is vsfb a platform_device driver ? |
| 14:50.17 | cr2 | i need to look how to transfer the fb address to it. |
| 14:51.35 | lkcl | i think it's a parameter that you either set at boot-time or at compile-time |
| 14:51.56 | lkcl | in the end i just cut/paste the entire lot, for the early w100fb |
| 14:52.14 | cr2 | ok. |
| 14:53.02 | cr2 | i have learned a difference between the headphone and headset from the AK4641 datasheet. |
| 14:53.22 | cr2 | headphone does not have a mic, while headset has. |
| 14:53.51 | cr2 | says this manual, and even has a pin drawing. |
| 14:55.17 | cr2 | lkcl: some sound gpios are related to the "car kit", and you'll never trigger it without some soldering. |
| 14:55.42 | cr2 | afaik, documented for hima and should be the same for ba. |
| 14:57.17 | cr2 | lkcl: the CONTROL_CX may be the 32768 clock for the BT PCM codec, it needs such clock. |
| 14:59.43 | cr2 | BT PCM is not controlled by uda1380, unlike the ak4641. |
| 15:00.54 | *** join/#htc-linux g3gg0_ (n=g3gg0@ppp-82-135-86-240.dynamic.mnet-online.de) |
| 15:01.58 | cr2 | .text:015B9C60 CMP R0, #0x200 |
| 15:02.02 | cr2 | .text:015B9C64 BEQ loc_15B9D7C |
| 15:02.06 | cr2 | .text:015B9C68 CMP R0, #0x400 |
| 15:02.10 | cr2 | .text:015B9C6C BEQ loc_15B9D58 |
| 15:02.14 | cr2 | .text:015B9C70 CMP R0, #0x1000 |
| 15:02.18 | cr2 | .text:015B9C74 BEQ loc_15B9D28 |
| 15:02.22 | cr2 | .text:015B9C78 CMP R0, #0x2000 |
| 15:02.26 | cr2 | .text:015B9C7C BEQ loc_15B9CF8 |
| 15:02.26 | cr2 | .text:015B9C80 CMP R0, #0x4000 |
| 15:02.26 | cr2 | .text:015B9C84 BEQ loc_15B9CC4 |
| 15:02.27 | cr2 | .text:015B9C88 CMP R0, #0x8000 |
| 15:02.27 | cr2 | the numbers are familiar to be. |
| 15:02.33 | cr2 | some of them are dummies on universal. |
| 15:03.20 | cr2 | 0x200 is this (i2c-related) gpioa0 |
| 15:03.33 | cr2 | 0x400 is the headset power |
| 15:03.53 | cr2 | 0x800 is codec_reset |
| 15:04.04 | cr2 | 0x1000 is speaker_power |
| 15:04.32 | *** join/#htc-linux rmoravci1 (n=rmoravci@icm2-orange.orange.sk) |
| 15:04.36 | cr2 | the rest are mic power related and so on (dummies on universal, because the mic power is under i2c control) |
| 15:04.53 | *** join/#htc-linux rmoravci1 (n=rmoravci@icm2-orange.orange.sk) |
| 15:07.50 | asylumed | anyone know about the owm on the asic? |
| 15:07.58 | asylumed | how does it compare to the ds1wm? |
| 15:09.49 | psokolovsky | it is ds1wm |
| 15:10.09 | cr2 | asylumed: http://wiki.xda-developers.com/index.php?pagename=MagicianAIC3 |
| 15:10.18 | cr2 | .bus_shift is 1 |
| 15:12.01 | asylumed | cr2: are you sure? been trying to get that working for a while... |
| 15:12.12 | cr2 | yes. |
| 15:13.03 | asylumed | cr2: so it's definitely not like the hx4700_power? |
| 15:15.06 | cr2 | asylumed: the asic3 owm is not used on ba. |
| 15:15.38 | asylumed | cr2: thx... |
| 15:15.38 | pH5 | asylumed: and still, the asic3 owm is a ds1wm, too. |
| 15:50.03 | psokolovsky | going to tag -hh6. any concerns? |
| 15:51.21 | pH5 | psokolovsky: after you tag hh6, can I commit the sa/irqf and irqchip/irqdesc renaming changes? |
| 15:51.36 | pH5 | psokolovsky: and I just noticed that h2200 battery still is not platform_deviceified |
| 15:52.27 | psokolovsky | pH5, thanks for them. I didn't review them yet, but guess they're ok ;-). well, it's up to you ;-) |
| 15:53.07 | psokolovsky | pH5, we need to go over those upgrade tickets and fix them consistently ;-) |
| 15:55.02 | BabelO | lkcl: do you have comited latest change to blueangel.c audio ? |
| 15:55.03 | psokolovsky | pH5, you applied those changes even to code not maintained by us? |
| 15:56.25 | psokolovsky | pH5, is SA_INTERRUPT really maps to IRQF_DISABLED? 8-) |
| 15:56.44 | TripleDES | hmm anybody with wifi problems on hh4 (universal)? |
| 15:57.36 | pH5 | psokolovsky: probably. I didn't really discriminate. maybe we should pick out changes that concern our machines or wait with the IRQF update until january altogether |
| 15:57.59 | pH5 | psokolovsky: yes, SA_INTERRUPT always meant "this is a short handler, leave interrupts disabled while handling this" |
| 15:59.16 | psokolovsky | pH5, I guess applying such stuff asap is good way, but only to our codebase, because otherwise it creates spurious diffs with mainline tarball (of course, that's just thought for future, no need to bother now) |
| 16:00.10 | psokolovsky | pH5, heh, who'd new IRQF_DISABLED would mean that too. some stupid folks like would think "start with interrupt disabled" %-) |
| 16:00.13 | psokolovsky | ~lart mainline |
| 16:00.19 | pH5 | I think we should critically review all request_irq calls anyway. There are probably quite a lot that could have IRQF_SAMPLE_RANDOM added, for example |
| 16:01.54 | psokolovsky | ok |
| 16:02.36 | pH5 | psokolovsky: about IRQF_DISABLED, I think it is much better than SA_INTERRUPT, which could mean anything |
| 16:02.48 | pH5 | from interrupt.h: " * IRQF_DISABLED - keep irqs disabled when calling the action handler" |
| 16:02.53 | pH5 | I think this is quite clear. |
| 16:03.51 | psokolovsky | pH5, ok, just need to know that, thanks for hint |
| 16:10.23 | psokolovsky | goxboxlive, Koen's right. please don't add random modules to machine config. ;-I |
| 16:10.47 | goxboxlive | right, does that goes for fuse to? |
| 16:11.49 | psokolovsky | goxboxlive, you really should start with assumption "no modules need to be added on machine at all", and carefully add only those without which *machine* really won't work |
| 16:12.23 | psokolovsky | of ones in your patch, only two have 100% ground to be added: |
| 16:12.23 | psokolovsky | + kernel-module-rtc-pcf8563 \ |
| 16:12.24 | psokolovsky | + kernel-module-rtc-sa1100 \ |
| 16:12.47 | psokolovsky | the rest are not machine specific to start with |
| 16:13.11 | goxboxlive | psokolovsky: I have done that today. I have buildx199 time and checked wich modules is installed and wich not |
| 16:13.38 | goxboxlive | cr2: Told me to use both of them |
| 16:14.03 | psokolovsky | goxboxlive, "installed" or "not installed" is not criterion, unfortunately. What is installed depends on what you have in defconfig, not on what is really needed. |
| 16:14.38 | psokolovsky | two above are *needed*. the rest - either questionable, or not at all. |
| 16:15.12 | goxboxlive | psokolovsky: I have just edited it so it intalls all the modules choosen in htcuniversal_defconfig. And with the patch i aplied it installs everything. So... |
| 16:15.34 | goxboxlive | where should smbs and fuse be added it it is not in the machine file? |
| 16:15.39 | psokolovsky | goxboxlive, again, that's not right |
| 16:16.16 | psokolovsky | goxboxlive, smbfs and fuse and everything else are not machine specific, so has nothing to do with machine config file |
| 16:16.48 | goxboxlive | I know that. But where should they be added then? |
| 16:17.50 | psokolovsky | goxboxlive, if they are not machine, then they are apparently distro features, right? That's where they should be added, and implemented in task-base. |
| 16:18.33 | psokolovsky | -hh6 is tagged |
| 16:28.16 | goxboxlive | psokolovsky: I agree with you about with the smbfs and fuse. I'll ask koen. |
| 16:29.41 | *** join/#htc-linux lkcl_ (n=lkcl@5ac5d83e.bb.sky.com) |
| 16:33.09 | *** join/#htc-linux JTRipper86 (n=jtripper@p54B1CA1C.dip.t-dialin.net) |
| 16:45.10 | Kevin2 | Hi |
| 16:45.20 | pH5 | hey lkcl, can the blueangel osrom be downloaded somewhere? |
| 16:45.21 | pH5 | hi Kevin2 |
| 16:46.05 | lkcl | Babel0: no not yet. it's a bit of a mess. |
| 16:49.44 | Kevin2 | The HTC Apache has a keyboard pullout detect gpio pin. Any idea what the "best" way to map this in Linux is? Should I call it a key and map it through the input layer? |
| 16:50.05 | Kevin2 | Also, any thoughts on the keyboard backlight? Is that considered an LED? |
| 16:50.38 | pH5 | Kevin2: I have it as a led (only the phone keys on ma though), as well as vibra. |
| 16:51.32 | lkcl | pH5: you mean the bootloader? |
| 16:51.44 | lkcl | cr2: do you mean 0x8000 is codec reset? |
| 16:51.55 | pH5 | the keyboard pullout definitely needs to go to userspace. a key would be the easiest way, right? |
| 16:52.35 | pH5 | lkcl: I'm interested in the wince kernel and driver dlls mainly. |
| 16:54.12 | Kevin2 | pH5: Okay, thanks. |
| 16:55.13 | pH5 | Kevin2: there is SW_LID and SW_TABLET_MODE from zaurus already, which are similar, but not quite the same |
| 17:01.35 | lkcl | pH5: no - i have the himalaya one. |
| 17:04.06 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
| 17:04.47 | pH5 | lkcl: ah ok |
| 17:05.49 | psokolovsky | pH5, ok, feel free to commit your patches ;-). let's pile all changes now and test them yet over weekend |
| 17:16.53 | pH5 | psokolovsky: what will be tested? I guess pxa, sa1100 and s3c2410 at most. |
| 17:19.20 | psokolovsky | pH5, you tested your patches, right? and there's always what to test even with trivial changes ;-) |
| 17:19.37 | Kevin2 | cr2: I added some more info to http://www.handhelds.org/moin/moin.cgi/ApachePhoneTrace -- it now has some examples on the memory polling features of mainline haret. |
| 17:19.47 | Kevin2 | (At the end of the page.) |
| 17:20.43 | pH5 | psokolovsky: I only tested them for one architecture, both my devices are pxa27x.. |
| 17:21.26 | psokolovsky | pH5, sure. that's really just formal measure ;-) |
| 17:22.10 | pH5 | Kevin2: sweet |
| 17:27.43 | lkcl | guys, what's the status on pierrox's eten g500 (s3c2442) patches? |
| 17:28.03 | lkcl | should i look at integrating the majority of his stuff where it doesn't impact common code? |
| 17:28.20 | lkcl | and then leave the rest for further investigation and review? |
| 17:28.30 | lkcl | i should be asking this in #handhelds.org shouldn't i... |
| 17:28.45 | pH5 | lkcl: you should, and my answer would be yes. |
| 17:28.50 | lkcl | ack |
| 17:30.31 | cr2 | pH5: there are a lot of roms at xda-dev |
| 17:30.46 | cr2 | Kevin2: thanks for the docs :) |
| 17:30.58 | psokolovsky | lkcl, where's his patches? |
| 17:31.06 | lkcl | cr2: you sure about 0x800 being codec_reset? |
| 17:31.14 | pH5 | cr2: hm.. maybe I should read the wiki sections outside *Research for a change.. thanks. |
| 17:31.15 | cr2 | lkcl: on universal -yes. |
| 17:31.16 | lkcl | cos there ain't one listed in that code-fragment you noticed |
| 17:31.19 | lkcl | ok. |
| 17:31.27 | psokolovsky | pH5, cr2: do you have an idea how we should handle RTC timebase? |
| 17:32.08 | pH5 | I don't, because I don't know what you are talking about. Can you explain? |
| 17:32.14 | cr2 | psokolovsky: universal( and others) have battery-powered RTC like a desktop machine. |
| 17:32.24 | cr2 | i don't know what wince stores there. |
| 17:32.48 | cr2 | there are wince machines. where the base is stored in RAM (!) |
| 17:33.08 | psokolovsky | pH5, cr2: each wince machine potentially has its own time base. for example, for h4000, 0 == 2003-08-01 |
| 17:33.09 | pH5 | cr2: so timebase is the time that is loaded into the rtc during boot? |
| 17:33.30 | cr2 | pH5: i'll define it as OSMR offset |
| 17:33.31 | lkcl | hey i wonder if the docking thing sends sound out to a different port as well? |
| 17:33.36 | cr2 | OSCR, sorry. |
| 17:34.01 | cr2 | lkcl: docking with the carkit pins enabled. |
| 17:34.13 | lkcl | GPIO 10, right? |
| 17:34.35 | cr2 | lkcl: look in the himalaya wiki at xda-dev |
| 17:34.38 | lkcl | ack. |
| 17:37.06 | psokolovsky | cr2, pH5: so, do you guys care about resuing wince time? |
| 17:37.08 | cr2 | psokolovsky: we are very good at managing mach-pxa, but who is the authoritative source for mach-s3c2410 ? |
| 17:37.28 | cr2 | psokolovsky: me ? not :) |
| 17:37.56 | cr2 | wince users are used to hard resets. |
| 17:37.58 | psokolovsky | cr2, we too (s3) ;-) |
| 17:38.22 | cr2 | s3 is not in such a good shape as pxa. |
| 17:38.59 | pH5 | psokolovsky: I'd rather get my time via ntp than from wince. |
| 17:39.06 | cr2 | lol. |
| 17:39.50 | cr2 | the hw rtc with its own battery is not bad. i guess wince stores local time there, but need to check. |
| 17:40.09 | pH5 | neither hx4700 nor magician have one, unfortunately :( |
| 17:40.26 | cr2 | goxboxlive: can you look at the rtc-pcf8563 date/time data ? |
| 17:40.33 | psokolovsky | pH5, so, you don't care too? You know, I agree, but it still sux to have: "sa1100-rtc: setting the system clock to 2023-08-01 00:00:41" |
| 17:41.01 | *** join/#htc-linux WizMaui (n=WizMaui@62.112.90.231) |
| 17:41.39 | cr2 | psokolovsky: add the rtc_oscr_offset= kernel parameter, or something like that. |
| 17:42.06 | cr2 | sa1100_oscr_offset= may be better. |
| 17:42.51 | psokolovsky | cr2, I want to consider a way to specify that offset as part of machine definition. Do you think it's good idea? |
| 17:43.35 | pH5 | psokolovsky: not really. I bet hx4700/wm2003 has a different offset than hx4700/wm2005. |
| 17:44.15 | psokolovsky | pH5, so, it will have it not static, but will need to calc. we alreday do that for resume w/native bootloader. |
| 17:44.33 | Kevin2 | This is odd - when I start a phone call I get an FFUART iinterrupt before I see any traffic from the PXA to the phone. I must be missing some communication channel between the two components. |
| 17:46.07 | pH5 | psokolovsky: because there is no other way. but I'd rather have that wince dependent code in HaRET than in the kernel, if possible. |
| 17:46.45 | psokolovsky | pH5, so, you vote for the same timebase for Linux across all handhelds? |
| 17:47.31 | cr2 | psokolovsky: 1.01.1970 :) |
| 17:48.11 | cr2 | the rest we will put in haret. |
| 17:48.16 | pH5 | psokolovsky: cr2 proposed to add a kernel parameter. we could then have a list of offsets in haret |
| 17:48.22 | psokolovsky | Kevin2, are you listening to this RTC talk? what do you think regarding haret resetting RTC base? |
| 17:48.41 | cr2 | Kevin2: you also have the hw rtc ? |
| 17:49.03 | psokolovsky | pH5, yes, we could. it will then depend on haret. do we need that? |
| 17:49.19 | pH5 | psokolovsky: how else would you start linux from wince? |
| 17:49.30 | cr2 | :) |
| 17:49.31 | pH5 | if you flash linux, you don't care about the wince timebase |
| 17:50.16 | cr2 | pH5: there were some other bootloaders, but they are toast now. omho. |
| 17:50.54 | pH5 | hehe, haret world domination progressing as planned. |
| 17:50.55 | cr2 | with haret compilable with gcc its a non-issue. |
| 17:50.57 | Kevin2 | psokolovsky: I'd vote for having the offset as a cmdline parameter. That way we can default it to 19700101 for users that aren't inflicted with wince. For those users needing to dual boot, it can be a setting in their haret config. |
| 17:52.22 | psokolovsky | Kevin2, no, haret config is not what I mean. I you ok to hardcode machine-specific offsets into haret? |
| 17:53.08 | cr2 | psokolovsky: we will hardcode a lot machine-dependent stuff in haret anyway. |
| 17:53.28 | Kevin2 | psokolovsky: Sure. But how would we pass it to the kernel? (Just about everyone overrides cmdline.) |
| 17:53.45 | cr2 | because on some machines there is simply no other way to boot. |
| 17:54.46 | psokolovsky | Kevin2, as long as you're ok to handle that in haret, we'll find way. easiest is still to readjust CPU RTC regs. or that won't work for external RTC? |
| 17:56.03 | Kevin2 | One could modify the internal PXA RTC, but that'd be dangerous - booting is already a difficult procedure, and I'd be reluctant to have it do more tasks. Setting RTC via I2C is way too difficult. |
| 17:56.19 | Kevin2 | As for having machine specific info in haret -- we have lots of it and I'm not against adding more. |
| 17:56.32 | psokolovsky | Kevin2, that's why I was thiking to handle it in kernel ;-) |
| 17:56.40 | psokolovsky | Kevin2, ok! |
| 18:00.07 | psokolovsky | Ok, so does anyone knows how to set RTC alarm by using only /proc or /sys? ;-) |
| 18:02.02 | cr2 | is it really _that_ portable ? |
| 18:04.35 | psokolovsky | cr2, if you ask me, I want to set alarm w/o 3rd party utils, only with kernel and busybox |
| 18:05.02 | cr2 | ok. |
| 18:05.08 | Kevin2 | I was wondering if the RTC I2C has the alarm line hooked up to one of the gpios. |
| 18:07.41 | pH5 | psokolovsky: I think you'd have to add a sysfs wrapper for rtc_set_alarm to rtc-sysfs.c for that. |
| 18:07.53 | cr2 | Kevin2: how to check it? |
| 18:08.01 | psokolovsky | pH5, I'd do that, if it was accepted ;-) |
| 18:08.12 | cr2 | write an alarm time, and watch the gpio ? |
| 18:08.13 | pH5 | to my limited knowledge, there only is the RTC_SET_ALM ioctl in rtc-dev.c currently |
| 18:08.41 | cr2 | pH5: ioctl |
| 18:09.03 | Kevin2 | cr2: Yeah. I'd try it that way. |
| 18:09.39 | Kevin2 | There is an extra pin on my ad7877 ADC chip. I programmed the chip to set the alarm, and I couldn't find any gpio changes. |
| 18:10.39 | cr2 | cpld ? |
| 18:11.46 | Kevin2 | I think I checked the egpio lines too. I don't think wince uses the "alert" pin on my ADC chip. |
| 18:13.08 | Kevin2 | Hrmm. Looks like memory polling is paying off during phone traces. The output is very cryptic but it looks like good info. |
| 18:14.54 | cr2 | Kevin2: addirqwatch P2V(0x0a000000) 0x20ff 16 0 |
| 18:15.06 | cr2 | what are the 2 3 and 4 params ? |
| 18:15.52 | Kevin2 | mask (bits to ignore), read size (ie, 8/16/32), and compare value. |
| 18:16.27 | cr2 | compare is also 16 bit value ? |
| 18:16.39 | Kevin2 | You don't normally set a "compare value" - it will report the value if it changes. However, when checking for interrupts, sometimes it is useful to check if the value is not zero. |
| 18:16.51 | cr2 | ok. |
| 18:16.58 | Kevin2 | cr2: Yes, you'd only want to compare against a 16bit value if you use a 16 bit read. |
| 18:17.05 | cr2 | :) |
| 18:18.30 | cr2 | the number of 'addtracewatch' is not limited ? |
| 18:22.06 | Kevin2 | There is a max of 32 - but it would just be a recompile to increase. |
| 18:22.32 | cr2 | ok. |
| 18:22.40 | pof | so, this is how we should go to debug gpio on hermes, righ? |
| 18:23.56 | Kevin2 | pof: It is possible - use "ADDWATCH P2V(0xaddr) 0xmask" for each of the gpio registers, and then call "watch <seconds>". |
| 18:24.37 | *** join/#htc-linux WizMaui_ (n=WizMaui@62.112.90.231) |
| 18:24.39 | Kevin2 | Unfortunately, you can't use addtracewatch - that depends on pxa. You can't use addirqwatch, because that also uses pxa, but it should be possible to make addirqwatch not pxa specific. |
| 18:25.25 | pof | ok, only addwatch... thanks Kevin2, will try that when i have some time, now i'm damn busy with university stuff :'( |
| 18:50.43 | *** join/#htc-linux rmoravcik (n=rmoravci@icm2-orange.orange.sk) |
| 18:58.13 | cr2 | the amount of FUD spread about SD on sl-5500 is amazing |
| 18:59.54 | cr2 | i have opened the case and made the PCB scans. The SDIO irq is wired, at least on the daugthercard. |
| 19:06.33 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
| 19:07.05 | *** join/#htc-linux rmoravcik (n=rmoravci@icm2-orange.orange.sk) |
| 19:10.38 | goxboxlive | cr2: are you around. I need some help about the rtc stuff. |
| 19:14.40 | cr2 | is it compiled as a module? |
| 19:15.05 | goxboxlive | yes |
| 19:15.06 | cr2 | there is also another driver in i2c/chips |
| 19:15.18 | goxboxlive | but oe guys complaines about it |
| 19:15.31 | cr2 | about which ? |
| 19:15.52 | goxboxlive | I guess that we have to choose about sa1100 and pcf8563 |
| 19:16.32 | cr2 | choose sa1100 |
| 19:16.38 | goxboxlive | ok, thx |
| 19:16.47 | cr2 | but i'd like to check the epson too. |
| 19:17.49 | goxboxlive | ok, but do it localy. I am gonna commit a new defconfig file, and it's gonna be without the epson module. |
| 19:17.59 | cr2 | ok. |
| 19:18.49 | cr2 | angstrom is confused that there are 2 rtcs ? :) |
| 19:19.29 | goxboxlive | not angstrom but oe chiefs |
| 19:19.32 | goxboxlive | :-) |
| 19:19.51 | cr2 | ok. |
| 20:03.47 | goxboxlive | cr2: Well have you any news about universal? Figured out something about resuming usbnet and keyboard? |
| 20:29.19 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
| 20:45.49 | cr2 | goxboxlive: the external mic detection code looks good, why it can't detect the mic ? we may simply enforce the mic switch her.e |
| 20:49.09 | cr2 | e680-a780/e680-linux/drivers/media/video/pxa_camera_ov.c and e680-a780/e680-linux/drivers/media/video/pxa_camera.c |
| 20:49.25 | cr2 | which one to use as a base ?? |
| 21:02.40 | goxboxlive | cr2: Hmm dont ask me. Take the one with wich is cleanest |
| 21:04.08 | cr2 | both look like quick-and-dirty hacks ;) |
| 21:08.18 | goxboxlive | :) |
| 21:22.39 | goxboxlive | cr2: Is the greenphone having the same camera as we have? |
| 21:25.45 | cr2 | don't know. |
| 21:25.54 | goxboxlive | ok |
| 21:25.58 | cr2 | but it's 2.4 kernel too |
| 21:26.03 | goxboxlive | i c |
| 21:26.12 | cr2 | we can write something that will work, sure. |
| 21:26.37 | *** join/#htc-linux WizMaui (n=WizMaui@62.112.90.231) |
| 21:28.22 | *** join/#htc-linux lkcl (n=lkcl@5ac5d83e.bb.sky.com) |
| 21:28.30 | cr2 | the camera support is only a question of the kernel infrastructure |
| 21:28.31 | goxboxlive | ok good. cr2, take alook here, and notice the date: http://www.nabble.com/forum/ViewPost.jtp?post=7844095&framed=y |
| 21:28.45 | lkcl | bollocks - forgot to plug power in to my laptop. |
| 21:28.45 | goxboxlive | Seems like opieII is comming before xmas |
| 21:28.47 | cr2 | we are the first who need something like that in 2.6 kernel. |
| 21:28.51 | lkcl | *sigh* i'm back... |
| 21:28.59 | goxboxlive | i c |
| 21:31.42 | cr2 | lkcl: what are our priorities ? :) |
| 21:32.40 | cr2 | goxboxlive: it's stupid not to wait for the openmoko/qtopia now, before making any substantial decisions. |
| 21:33.13 | cr2 | our device is the best by its hardware merits, and will stay for a long time imho. |
| 21:36.32 | goxboxlive | Yes i think we should wait for Qtopia/OpenMoko and just leave opie behind us. I am working in GPE a little, because of Gomunicator, but GPE is slow compared to Opie. |
| 21:37.37 | lkcl | ha ha, you're asking me? |
| 21:38.12 | lkcl | cr2: i agree about waiting for openmoko. it's why i'm happy to focus on the device-drivers. |
| 21:38.22 | lkcl | goboxlive: ack. |
| 21:38.40 | lkcl | although, gomunicator is a good place to experiment with phone stuff |
| 21:44.17 | cr2 | lkcl: what is the current eten g500 status ? |
| 21:44.45 | lkcl | pierrox has quite a lot working, relatively speaking |
| 21:44.54 | lkcl | last month he got sd card. |
| 21:44.56 | cr2 | .19 ? |
| 21:45.03 | *** join/#htc-linux RoEn_PC (n=roen@p54A6497F.dip.t-dialin.net) |
| 21:45.08 | lkcl | http://www.handhelds.org/moin/moin.cgi/EtenG500Status |
| 21:45.14 | lkcl | yes, .19 |
| 21:45.47 | cr2 | i don't see any port that cleanly compiles from CVS. |
| 21:46.03 | cr2 | ok, what do we need to merge it to hh.org CVS, without breaking it for others ? |
| 21:46.14 | lkcl | eek. |
| 21:46.33 | cr2 | the patches are big, and touch some core sybsystems. |
| 21:46.56 | lkcl | pierrox had to send me a patch against linux-2.6.19 because when he did a cvs update he got masses of conflicts, and hasn't enough experience or contacts to sort them out. |
| 21:47.14 | cr2 | psokolovsky has said, hh.org is the central samsung repo. |
| 21:47.32 | lkcl | yes, but it was the only way he could send me a clean patch file. |
| 21:47.33 | cr2 | yeah, that's exactly this mess that scares me. |
| 21:48.12 | cr2 | are there any samsung devices in OE ? |
| 21:48.50 | cr2 | i can guess that the openmoko will be a patch relative to the mainline kernel. |
| 21:49.07 | cr2 | judging from openezx stuff. |
| 21:49.36 | lkcl | i'm in the middle of trying to do sound for BA - you reckon i should change focus onto the s3c stuff? |
| 21:49.49 | lkcl | push pierrox's stuff through, for example? |
| 21:50.29 | cr2 | imho, if we will get barebones samsung kernel for hermes, the people will research gpio. |
| 21:51.49 | lkcl | ok. |
| 21:51.57 | lkcl | wanna tackle that now? |
| 21:52.01 | lkcl | just because we can? |
| 21:52.25 | lkcl | i'm quite happy to bash through that for a couple of hours, like we did for sable. |
| 21:52.46 | lkcl | which didn't take very long. |
| 21:53.03 | lkcl | btw - A2 is for bluetooth audio switching on BA :) |
| 21:54.17 | lkcl | oh - and GPIO 22 |
| 21:54.25 | cr2 | good :) |
| 21:54.27 | cr2 | btw |
| 21:54.30 | cr2 | ./mach-pxa/eseries/e330_lcd.c |
| 21:54.35 | cr2 | this is the vsfb setup. |
| 21:54.45 | pH5 | is the asic3 OWM clock at a fixed frequency of 4 MHz or are there other possibilities how the AIC2 DS1WM could be clocked on the ba? |
| 21:55.06 | cr2 | pH5: the cdiv is 2 on hx4700 and universal. |
| 21:55.54 | cr2 | where can i measure this frequency ? |
| 21:56.01 | cr2 | on the battery pin ? |
| 21:56.20 | pH5 | cr2: I know. this is for a clock frequency of 4 to 5 MHz. the hx4700 enables the OWM_INT. I just wondered if there might be some clock setting register in the asic3 to change that frequency |
| 21:56.28 | pH5 | cr2: unfortunately not. |
| 21:56.45 | pH5 | the battery pin has the ~1MHz OWM frequency that the ds1wm makes out of its input frequency via the divider |
| 21:57.22 | pH5 | wait, actually yes. if you measure the battery pin and have cdiv=0x2, the input frequency should be 5 times what you measure. |
| 21:57.59 | cr2 | pH5: hmm. how can i measure the frequency on ba... |
| 21:58.09 | cr2 | the battery must be closed. |
| 21:58.31 | lkcl | i got an oscilloscope somewhere and i also got a hammer |
| 21:59.02 | lkcl | i was only partially joking |
| 21:59.12 | pH5 | lkcl: the oscilloscope part or the hammer part? |
| 21:59.19 | lkcl | the hammer part. |
| 21:59.37 | cr2 | lkcl: we do experimental physics here :) i have a 16channel logic analyzer. can capture stuff up to 0.5GHz . |
| 21:59.47 | lkcl | cooool |
| 21:59.49 | lkcl | mwahaahha |
| 22:00.14 | lkcl | that's more than adequate for all sorts of fun and games. |
| 22:00.28 | cr2 | the difficult part is to connect the wires ;) |
| 22:00.53 | lkcl | well... my friend richard lightman manages it ok |
| 22:01.06 | lkcl | so - you wanna tackle hermes, then? |
| 22:01.22 | cr2 | to measure 1 Mhz... |
| 22:01.54 | cr2 | that's something the most primitive frequency counter can do these days. |
| 22:02.15 | lkcl | the only gpio on BA sound i'm not sure about now is GPIO20 and the exact purpose of GPIO89 other than it's being flipped really quickly |
| 22:02.22 | cr2 | just 2 files for hermes. minimal setup and vsfb. |
| 22:02.59 | cr2 | is there any context when gpio89 is flipping ? |
| 22:03.04 | lkcl | gonna go get me hermess... |
| 22:03.48 | cr2 | i have created a hermes config from rx3715, but the damn kernel does not link because of ohter kernel bugs. |
| 22:04.14 | cr2 | then i have created a htchermes directory in mach-s3c2410 |
| 22:04.26 | cr2 | and botched all the configs completely. |
| 22:04.37 | cr2 | now need to cleanup all the mess in my tree. |
| 22:05.13 | cr2 | pH5: i have cracked open the BA battery. |
| 22:05.34 | cr2 | there are 4 pins. |
| 22:05.49 | cr2 | +3.7V, ground. |
| 22:06.11 | cr2 | where should i measure on the other pins ? |
| 22:06.36 | lkcl | cr2: well, let me have a go at it. commit what you've got and i'll make _another_ linux kernel tree... :) |
| 22:07.23 | lkcl | btw you mind if i cvs commit some gpio comments in asic3.cpp? |
| 22:07.33 | lkcl | i added CDEX watching. |
| 22:07.33 | pH5 | cr2: I have no idea, never seen a ba battery in real life |
| 22:07.47 | cr2 | in haret ? |
| 22:07.54 | lkcl | haret-gnu |
| 22:08.01 | cr2 | pH5: i can put it on the scanner |
| 22:08.06 | cr2 | lkcl: no prob. |
| 22:08.09 | lkcl | k. |
| 22:08.43 | pH5 | lkcl: thanks :) |
| 22:08.48 | lkcl | :) |
| 22:09.51 | pH5 | cr2: though I wouldn't be able to identify the correct owm pin from that |
| 22:09.51 | cr2 | lkcl: afaik i have replaced EGPIO for CDEX now |
| 22:09.51 | cr2 | on BA |
| 22:09.51 | lkcl | ah. yes. |
| 22:09.56 | cr2 | pH5: i'll try to look with the oscilloscope. |
| 22:10.14 | lkcl | yes you did. i disabled EGPIO on BA and added some comments for CDEX. so it actually prints out 'EX1' or whatever. |
| 22:10.35 | lkcl | not that it actually changes cos damn wince *grumble* |
| 22:11.11 | cr2 | add is as a "%s" |
| 22:11.20 | cr2 | the EGPIO vs. something else. |
| 22:11.37 | cr2 | so it stays EGPIO on universal. |
| 22:11.55 | pH5 | cr2: thanks |
| 22:12.24 | *** join/#htc-linux RoEn_PC (n=roen@p54A6497F.dip.t-dialin.net) |
| 22:14.16 | cr2 | hmm. where did i put the battery... |
| 22:14.55 | lpotter | http://opie.home.linuxtogo.org/ |
| 22:15.01 | lkcl | it's run away, cr2, cos it saw you with the hammer :) |
| 22:15.15 | lkcl | allo lpotter |
| 22:15.30 | lkcl | ooo, pretty! |
| 22:16.38 | cr2 | lpotter: we are tuned, but need the source :) |
| 22:16.42 | goxboxlive | lpotter: looking good :-) I guess it will be released monday. one week after the consumers |
| 22:21.32 | lkcl | http://www.pierrox.net/G500/20061205/linux-2.6.19-phe-g500.diff |
| 22:21.57 | lkcl | oh, good grief. cr2: there's a patch which you need to integrate otherwise the s3c kernel won't link. |
| 22:22.06 | lkcl | http://www.spinics.net/lists/arm/msg12639.htm |
| 22:22.12 | lkcl | http://www.spinics.net/lists/arm/msg12638.html |
| 22:22.50 | lpotter | supposed to have been 3 days after... |
| 22:22.54 | cr2 | yes, something like that. |
| 22:23.05 | goxboxlive | lpotter: ok |
| 22:23.27 | lpotter | difficulties with the package |
| 22:24.13 | goxboxlive | off topic: My wife bought a mp3 player, it was only aplied with a usb charger: But since my Universal has a AC charger, could she use that one? Shouldnt that be the same since it usb? |
| 22:24.29 | cr2 | hmm. lost the cracked battery somewhere. |
| 22:24.40 | pH5 | goxboxlive: should work |
| 22:24.50 | cr2 | goxboxlive: there are gazillion of usb chargers. |
| 22:24.51 | pH5 | lkcl: you shouldn't have started with the hammer |
| 22:24.57 | lkcl | teehee |
| 22:25.22 | cr2 | universal AC charger grounds the OTG pin. |
| 22:25.30 | cr2 | a780 does not like it. |
| 22:25.44 | lkcl | that's because a780 does enhanced-usb |
| 22:25.49 | cr2 | and i can't charge universal with a780 AC charger at all. |
| 22:25.51 | goxboxlive | hmm, so it might not be safe then. |
| 22:26.08 | cr2 | normal usb works ok. |
| 22:26.09 | lkcl | and it detects charging and asks the charger to supply 2.5amp instead of 0.5 |
| 22:26.22 | cr2 | goxboxlive: it's safe, bt it may not work. |
| 22:26.25 | lkcl | don't for god's sake use the a780 charger on anything other than an a780 |
| 22:26.31 | goxboxlive | ok, thx |
| 22:27.25 | goxboxlive | well, it worked, thanks. |
| 22:28.27 | cr2 | lkcl: but i don't want to carry 4 sorts of usb chargers around ;) |
| 22:28.32 | lkcl | *snort* |
| 22:28.37 | BabelO | goxboxlive: try this great song on your mp3 player :) http://linuxbox2.free.fr/blueangel/01 - Electronica - Der Ententanz.mp3 |
| 22:28.50 | lkcl | what, you carry 4 phones around??? |
| 22:29.14 | cr2 | 3, now 5 :) |
| 22:29.47 | cr2 | they all have their own function. |
| 22:30.43 | lkcl | good grief... |
| 22:30.49 | lkcl | you're going to get fried |
| 22:30.50 | cr2 | a780 as a phone + route planning, ba for gps tracks and normal maps on normal screen. |
| 22:31.05 | goxboxlive | <BabelO> :-) Yes i remember that song, it was so popular here in Norway in the 80's. I also know the dance to it :-) |
| 22:31.57 | BabelO | i prefer this than the french version :) |
| 22:35.12 | lkcl | Babel0: i found a bluetooth audio device. |
| 22:35.19 | lkcl | i've filled in the page a bit more. |
| 22:36.57 | BabelO | lkcl: eh eh we'll have all working soon :) |
| 22:37.08 | lkcl | *snort* |
| 22:39.19 | BabelO | lkcl: no Ronald this evening, wine + beef :) |
| 22:39.43 | lkcl | he he. Babel0: can i write you some recipes for good food? |
| 22:40.39 | goxboxlive | cr2: Still awake? |
| 22:40.50 | goxboxlive | Did you get my mail today? |
| 22:41.10 | BabelO | lkcl: no thanks, have a lot of recipes here when got time :) |
| 22:41.59 | lkcl | ok then what's your address, i'll order you a chinese or indian takeaway? |
| 22:44.00 | BabelO | lkcl: i think i ve test some indian when i go to Farnham :) |
| 22:52.23 | cr2 | goxboxlive: it was meant for blueangel. we don't have write code for the universal yet. |
| 22:55.32 | goxboxlive | cr2? The driver or the mail. He was asking about the Universal. If we dont have any driver for it to test i'll remove it from wiki. |
| 23:02.50 | cr2 | goxboxlive: the driver is for blueangel. |
| 23:03.10 | goxboxlive | ok |
| 23:03.28 | goxboxlive | ok i answer him then. |
| 23:08.41 | *** join/#htc-linux g3gg0_ (n=g3gg0@ppp-82-135-86-240.dynamic.mnet-online.de) |
| 23:12.22 | lkcl | oh you live in guilford, then? |
| 23:12.38 | lkcl | cr2: ok, asic3.cpp is cvs committed for haret-gnu |
| 23:13.35 | cr2 | ok. |
| 23:16.14 | lkcl | cr2: i'm just applying pierrox's patch to a local hh.org tree. |
| 23:16.21 | lkcl | what i might do is create a branch of the modified files. |
| 23:16.30 | lkcl | i'll let you know how to check them out. |
| 23:16.44 | lkcl | it could involve doing cvs co -rHEAD or something daft |
| 23:18.03 | cr2 | ok. |
| 23:18.11 | cr2 | some files are harmless. |
| 23:18.28 | cr2 | defconfig, and everything strictly g500-specific. |
| 23:19.03 | cr2 | maybe create the eten-g500 subdirectory. |
| 23:34.29 | *** join/#htc-linux ralith_ (n=ralith@saunderssbs.saundersinstruments.com) |
| 23:44.30 | *** join/#htc-linux apt (i=ibot@pdpc/supporter/active/TimRiker/bot/apt) |
| 23:44.30 | *** topic/#htc-linux is HTC Linux Channel: Find logs at http://apt.rikers.org/%23htc-linux/ | please check http://handhelds.org/moin/moin.cgi/HTC_2dPhones | http://wiki.xda-developers.com/index.php?pagename=Xanadux |
| 23:44.36 | lkcl | hiya psoko |
| 23:45.06 | lkcl | can i patch in the s3c stuff that will make g500 and other s3c projects work? |
| 23:45.25 | psokolovsky | lkcl, I really suggest pierrox or you first post patches to k-d, especially if they require changes to common code... |
| 23:45.45 | lkcl | there's two patches already which have been posted to arm-linux-kernel |
| 23:45.46 | psokolovsky | lkcl, afaik, rx3000 does pretty work... |
| 23:45.58 | lkcl | which he requires to get things to link. |
| 23:46.03 | lkcl | but nobody has applied those patches. |
| 23:46.28 | lkcl | http://www.spinics.net/lists/arm/msg12639.html |
| 23:46.29 | psokolovsky | lkcl, please posta bout that to k-d, so we all were in loop... |
| 23:46.33 | lkcl | ok. |
| 23:46.45 | lkcl | http://www.spinics.net/lists/arm/msg12638.html |
| 23:47.37 | lkcl | they're pretty simple things. Kconfig switches on power management irrespective of cpu |
| 23:48.03 | psokolovsky | lkcl, I believe I saw something like that. That guy ports rx1950 iirc. still, pls posta t least notices that you apply them... |
| 23:48.11 | lkcl | okie. |
| 23:52.51 | *** join/#htc-linux rmoravcik (n=rmoravci@icm2-orange.orange.sk) |
| 23:57.04 | cr2 | rx3715 needs defconfig, then everybody will see that is does not compile without the patches. |
| 23:58.11 | cr2 | btw, is there any samsung cpu doc where they discuss gpio and alt gpio settings ? the file i have seen was light on details. but maybe i'm spoiled by dec/intel docs ;) |
| 23:59.58 | asylumed | ds1wm works on the ba :) |