00:12.09 | oxi__ | lo? |
00:13.08 | oxi__ | :S |
02:12.14 | *** join/#htc-linux tudenbart (n=willi@xdsl-213-196-249-72.netcologne.de) |
05:00.29 | *** join/#htc-linux oxi_ (n=fdhgf@101.Red-83-46-140.dynamicIP.rima-tde.net) |
05:03.37 | *** join/#htc-linux sadeness_ (n=LamersIn@nat1.rlan.ru) |
05:17.15 | *** join/#htc-linux YellowRoze (n=Jonas@c-29f2e155.176-1-64736c12.cust.bredbandsbolaget.se) |
05:20.05 | YellowRoze | Hi folks, I've got a blue angel and some half a year/year ago I got a more or less preinstalled gpe on my memorystick. Now i'd like to see if anything have happened since and if there's a prepared way of getting opie on it. anyone want to shed some light on this? =) |
05:28.33 | *** join/#htc-linux _law_ (n=law@mail.stiftadmont.at) |
06:05.23 | *** join/#htc-linux pleemans (n=peter@d5152D19B.access.telenet.be) |
06:22.26 | *** join/#htc-linux goxboxlive (n=goxboxli@195.159.97.196) |
06:26.26 | goxboxlive | good morning |
06:29.44 | *** join/#htc-linux rob_w (n=rob_w@p549b8474.dip0.t-ipconnect.de) |
06:31.22 | YellowRoze | good morning indeed. |
07:06.34 | YellowRoze | (07:20:16) YellowRoze: Hi folks, I've got a blue angel and some half a year/year ago I got a more or less preinstalled gpe on my memorystick. Now i'd like to see if anything have happened since and if there's a prepared way of getting opie on it. anyone want to shed some light on this? =) |
07:07.38 | *** join/#htc-linux Fossi (n=cmewes@195.177.48.115) |
07:16.07 | *** join/#htc-linux psokolovsky (n=psokolov@82.193.98.4) |
07:18.12 | *** join/#htc-linux drno (n=drno@eki.student.utwente.nl) |
09:14.48 | oxi_ | just for when I return back, the mic routing for the phone you were lately talking about is just for BA, right? (here comes my great question) where is this code supposed to be, and most important, how do I know what piece of code or file, control what hw in my hh? my hh is a universal. I see there are some files that look like modules that are compiled but only one of them is compiled as a module: one that controls sound, i beli |
09:14.48 | oxi_ | eve, something like htcuniversal_ak4641... There's also one called htcuniversal_phone.c (that's the only one I found about phone) but it's not compiled as a module and my phone works well. In short, I'd just like to know in what file/s and how I'm supposed to do the mic routing thingy for my universal. Thanks. BBL :) |
09:19.48 | *** join/#htc-linux psokolovsky (n=psokolov@empty-crossing.volia.net) |
09:33.09 | goxboxlive | oxi_ The mic enable stuff is for the Universal. |
09:33.47 | goxboxlive | You have iut in the latest source. Open alsamixer and you enable it there. |
09:33.56 | goxboxlive | s/iut/it/ |
09:37.16 | *** join/#htc-linux pH5 (n=ph5@e178194135.adsl.alicedsl.de) |
09:54.15 | goxboxlive | hi pH5. |
09:59.57 | pH5 | hi! |
10:01.30 | goxboxlive | pH5: There have been a lot of changes in the libgsmd since you been gone. Semce like your patch in OE doesnt work anymore. |
10:02.46 | pH5 | goxboxlive: yes, looks like harald really started to work on it again. |
10:03.13 | pH5 | I'm just getting back into it, I'll update / drop patches where needed. |
10:15.44 | psokolovsky | Hi! |
10:15.51 | psokolovsky | pH5: welcome back ;-) |
10:16.07 | goxboxlive | good |
10:16.37 | pH5 | hi psokolovsky |
10:17.09 | pH5 | I'm very happy to see anton and your efforts to communicate with upstream! |
10:17.54 | psokolovsky | pH5: I've posted RFC on list regarding changing SoC IRQ initialization per RMK's suggestions. your review is welcome |
10:17.56 | pH5 | goxboxlive: it's just the numeric patch that was applied, the rest still works. I'll update oe.dev |
10:18.00 | psokolovsky | pH5: Thanks! ;-) |
10:20.01 | goxboxlive | good |
10:22.42 | *** join/#htc-linux BabelO (n=root@lun34-2-82-238-28-28.fbx.proxad.net) |
10:24.17 | goxboxlive | pH5 did yo got my mail about the "hangup" issue? |
10:26.39 | pH5 | goxboxlive: yes, I just found it. |
10:27.13 | pH5 | I'll look at the spec - is not supporting ATH0 a bug, or is it optional? |
10:27.49 | pH5 | the problem with the other side aborting the call could be gsmd depending on the PI messages that the qualcomm modem doesn't have. |
10:27.56 | pH5 | s/PI/%CPI/ |
10:32.45 | goxboxlive | i c. If ATH0 is a bug or not is for me unknown. But when i have been useing 'cu' i have noticed that i have to type 'ath' to hang up, and to answer i have to typr 'ata' so guess it's not a bug. |
10:34.00 | goxboxlive | I can take a look at %CPI when i get home to see what output i get when the other part is aborting a call. |
10:36.21 | pH5 | goxboxlive: yes, that would be good. |
10:36.37 | pH5 | hm. ath is defined in ITU-T V.25ter. where do I get that. |
10:38.57 | *** join/#htc-linux cyrill62 (n=cyrill62@12.214.20.81.dynamic.adsl.abo.nordnet.fr) |
10:40.35 | cr2 | pH5: hi. |
10:41.12 | pH5 | hey cr2 |
10:44.20 | BabelO | hi cr2, pH5 |
10:45.25 | cr2 | pH5: i've written "dump pasic3" for haret, and updated the wiki page. have a look :) |
10:46.26 | pH5 | hi BabelO |
10:46.43 | pH5 | cr2: great job! |
10:46.54 | pH5 | 0x06 is not pwm enable, but a voltage to leds bitfield |
10:47.24 | pH5 | no, I'm probably wrong. |
10:47.49 | pH5 | or not. but there are two bits for each led, causing different brightness. |
10:47.58 | pH5 | no idea if this is wiring or pwm related |
10:49.00 | pH5 | 0x28 bit 0x80 en-/disables the ds1wm. |
10:49.29 | BabelO | pH5: about your line discipline driver, can we got a new ttyS line demultiplexed ? |
10:50.06 | BabelO | pH5: i try with your test program from kernel discuss on BA |
10:53.51 | pH5 | BabelO: I don't understand. It's a line discipline that could be used on any tty. |
10:53.55 | pH5 | Doesn't it work on BA? |
10:56.03 | pH5 | cr2: I don't understand the 0x10..0x14 information. what is ba: 0-math-11 supposed to mean? |
11:05.02 | *** join/#htc-linux oxi_street (n=alphavil@80.27.199.226) |
11:05.21 | oxi_street | hi |
11:05.42 | pH5 | hi oxi_street |
11:07.03 | oxi_street | Im on no-qwerty phone. Cant write fast |
11:09.19 | oxi_street | Goxboxlive: i read u before leaving. Anyways, what file/s did u change for the phone mic routing? Im curious. Id like to help |
11:09.35 | oxi_street | With anxthing i can |
11:10.10 | oxi_street | s/anxthing/anything/ |
11:11.56 | cr2 | pH5: it needs more investigation. some complex maths using UMULL and strange numbers. |
11:12.35 | cr2 | pH5: it works on ba and on hima. |
11:13.56 | pH5 | cr2: where is that code? I only know that 0x14 is set to 0x00 on cold boot and in SleepSetting |
11:15.11 | pH5 | and 0x2b somehow depends on the cpld on magician. it has different values depending on the same variable that determines whether the third egpio register on the cpld is used. |
11:15.13 | cr2 | in nk.exe on hima, i think. |
11:15.32 | pH5 | probably something board specific. |
11:15.49 | cr2 | the actual values i've dumped with 'dump pasic3' are a bit different from those in the table, which are from nk.exe |
11:16.08 | cr2 | ok. |
11:17.16 | pH5 | cr2: I think the 0x20 range changes without writing to it. maybe those are some kind of pwm counters that are only written to to reset them. |
11:17.35 | cr2 | i will also need a cpld driver for hermes. with 3-4 16bit gpio regs and an irq register. |
11:17.45 | cr2 | pH5: they might be write-only |
11:17.57 | pH5 | ok |
11:19.24 | cr2 | pH5: will it be difficult to add another line discipline ? |
11:20.00 | pH5 | cr2: no. it will be difficult to make the current one foolproof. it's probably unstable as hell, currently. |
11:20.02 | pH5 | cr2: what for? |
11:20.06 | cr2 | here at the bottom. |
11:20.09 | cr2 | http://wiki.xda-developers.com/index.php?pagename=SablePhone |
11:20.47 | cr2 | but it's not _that_ much different. |
11:21.22 | pH5 | ah, I see. |
11:22.24 | cr2 | btw, the ov9640 works on a780 now. |
11:23.12 | pH5 | nice :) |
11:32.23 | BabelO | pH5: is ds1wm working well for you no freeze ? |
11:33.03 | pH5 | BabelO: on magician, yes. |
11:34.56 | BabelO | with have some problem on BA |
11:35.11 | BabelO | nedd to fix asic3 irq demux problem before |
11:37.04 | rmoravcik | BabelO: hi, i have problem with ds1wm too, sometimes i see irq demux overrun and after while pda freeze |
11:43.16 | BabelO | rmoravcik: hi, is you posting a patch yesturday at kernel discuss about this ? |
11:45.13 | BabelO | rmoravcik: yes it is you :), i can't get your patch with link provided |
11:45.32 | rmoravcik | BabelO: it desn't solve that problem for me, i fix only counting loops |
11:47.05 | goxboxlive | cr2: Have you tested the ov9640 opn A780 yet? |
11:47.16 | rmoravcik | BabelO: ds1wm sometimes return 0x00 and 0xff values on my rx3000 |
11:48.08 | BabelO | rmoravcik: is it ds1wm on asic3 or an exrternal chipset ? |
11:48.18 | cr2 | rmoravcik: the wince driver is full of handlers for these ;) they do a full reset. |
11:48.31 | BabelO | rmoravcik: i add some printk this evening to see values |
11:50.04 | BabelO | rmoravcik: i see thaht you have uda1380 on rx3000 too |
11:57.36 | cr2 | goxboxlive: no, but i've read in the openezx-devel archive that it works. with some minor patches. |
11:57.45 | rmoravcik | BabelO: ds1wm on asic3 |
11:58.38 | cr2 | goxboxlive: i#ve added switch support to asic3_keys, and it works. not in the CVS though, because it will break other devices without an update to the keys table. |
11:59.19 | cr2 | but opie does not react on the switches. probably because you have hardcoded the device numbers. |
12:03.23 | rmoravcik | cr2: i looks like wince is making full reset before every new communication |
12:03.56 | cr2 | rmoravcik: maybe ;) |
12:30.53 | *** join/#htc-linux WizMaui (n=WizMaui@62.112.90.250) |
12:58.13 | bd2 | rmoravcik, cr, if you're experiencing problems with ds1wm, try revert this patch: http://handhelds.org/cgi-bin/cvsweb.cgi/linux/kernel26/drivers/w1/masters/ds1wm.c.diff?r1=1.10&r2=1.11 |
12:58.57 | rmoravcik | bd2: hi |
12:59.27 | rmoravcik | bd2: i had problem with ds1wm before this patch too |
13:00.14 | bd2 | ugh, too bad :-( |
13:00.46 | bd2 | I've seen ds1wm hanging, and doing delays helped me. I just thought, rx3000 would need longer delay |
13:01.07 | rmoravcik | bd2: i don't have idea what is wrong, i only see sometimes asic3 irq demux overrrun |
13:01.13 | bd2 | though, current CVS HEAD state works fine for me.. |
13:01.28 | bd2 | rmoravcik, I recalling irq demux overrun too |
13:01.52 | bd2 | but these were not fatal for hx4700 |
13:02.28 | rmoravcik | bd2: but i don't know why it freeze |
13:02.45 | bd2 | whole pda freeze, or ds1wm? |
13:03.05 | rmoravcik | whole pda and i'm sure problem is with ds1wm module |
13:03.53 | bd2 | this might be unrelated, though... but h5000 was freezing for me, when I tried to use ds1wm. The problem was misconfigured MDREFR (i.e. SDRAM frequency!). |
13:03.55 | rmoravcik | but sometimes is working without any problems |
13:04.01 | bd2 | rmoravcik, are you using cpu freq? |
13:04.10 | rmoravcik | bd2: no |
13:04.46 | bd2 | then most probably it's indeed unrelated :-/ |
13:04.53 | rmoravcik | i compared s3c2440 clock setting and it looks same |
13:06.29 | bd2 | I can't see any code in ds1wm which could freeze whole pda, though.. :-/ |
13:06.44 | *** join/#htc-linux cyrill62_ (n=cyrill62@162.220.20.81.dynamic.adsl.abo.nordnet.fr) |
13:07.00 | bd2 | all stuff executed in userspace, except isr.. and isr is trivial |
13:07.47 | bd2 | thus, maybe it's worth todo stress-testing of asic3 on rx3000. Try write/read to its registers using devmem2 in loop |
13:08.16 | bd2 | like while true; do devmem2 0xsome_asic3ds1wmreg ; done |
13:09.16 | bd2 | on h5000 it was almost surely lead to hang whole pda |
13:09.38 | rmoravcik | i can test it |
13:12.58 | bd2 | while true; do devmem2 0xsome_asic3ds1wmreg > /dev/null ; done would be faster/better |
13:14.04 | cr2 | maybe asic3 should disable the 1w interrupt processing for some time if it's being overflooded ? |
13:21.15 | goxboxlive | cr2: Ok, but i havent hardcoded anything in Opie. Only BT and that's all. |
13:21.50 | rmoravcik | bd2: it didn't freeze |
13:22.05 | goxboxlive | There might be some hardcoding there, but thats some old zaurus stuff i guess. It was Milan who writed the htc_odevice.cpp file. It realy needs some update. |
13:23.23 | cr2 | yes, i think it's htc_odevice.cpp fault. |
13:23.34 | goxboxlive | ok |
13:24.04 | goxboxlive | so i can just aply your patch. It doesnt mather for me if i break other devices localy. I'll try it out when i get home. |
13:24.14 | bd2 | rmoravcik :-(( |
13:24.29 | cr2 | goxboxlive: what is the most urgent thing now ? |
13:25.19 | cr2 | goxboxlive: it's not a complete patch, but i'll add the changes to htcuniversal_buttons.c and will ifdef them. |
13:26.10 | goxboxlive | I dont know now? We have phone working, and i guess pH5 and i will make sms working soon. |
13:26.32 | goxboxlive | What about just adding the camera so we have it . |
13:26.51 | cr2 | is it the most urgent thing ? |
13:27.09 | goxboxlive | I have mension this for you before, but i think it is kevin2 stuff. The keyboard is realy buggy after resume |
13:27.39 | cr2 | do you have a userspace app for the camera ? i think ljp did something... |
13:27.57 | cr2 | btw, i can't suspend at all. |
13:28.32 | goxboxlive | There is something named opie-camera, but it is for sharp devices. IIRC maiceyl made something once |
13:28.50 | cr2 | apm --suspend just hangs in the process. |
13:29.10 | cr2 | sharp driver is closed source, and they were lazy to RE it. |
13:29.33 | goxboxlive | what do you mean, arent you able to suspend? |
13:29.51 | cr2 | i do 'apm --suspend', and it does not suspend. |
13:29.53 | goxboxlive | We have buggy htcuniversal_acx. It hasent been good senice 2.6.16 |
13:30.41 | cr2 | not htcuniversal_acx, but the buggy 'acx' |
13:31.09 | goxboxlive | cr2: If you are doing it form opie-console it wouldnt work. But if you do it via usb or form the shell it works |
13:31.37 | cr2 | it calls function pointers from htcuniversal_acx even if the module is removed and/or they are NULL. |
13:31.50 | goxboxlive | ok |
13:32.01 | cr2 | goxboxlive: does not work from usb too. |
13:32.13 | goxboxlive | then try form the shell. It worked for me yesterday |
13:32.16 | oxi_street | Can i read the channel log in some web or can u paste it somewhere please from my last comment? Im on an old mobile and if i .. |
13:32.19 | goxboxlive | i tried it out |
13:32.48 | cr2 | goxboxlive: ok, but it's unacceptable too. |
13:33.17 | goxboxlive | yes, x11 doesent have the same problem. it works even with GUI part in GPE .-) |
13:33.21 | cr2 | oxi_street: irclog.iclem.net |
13:33.22 | oxi_street | .. Scrollback and someone talks, the scroll goes down again. Thanks |
13:33.47 | oxi_street | Ty |
13:36.57 | cr2 | goxboxlive: i wonder why gsmd can't do sms when gnokii even supports UCS2. |
13:43.21 | goxboxlive | cr2: It's not implemented yet (if that is an english word) |
13:43.50 | goxboxlive | I have to figure out how qualcomm modem stores sms on the sim card. |
13:44.14 | goxboxlive | I havent found the settings in wince yet so i can debugg and see whats going on. |
13:47.51 | goxboxlive | well, time to go. bbl |
13:50.47 | cr2 | hm. i think that the raw sim access is not available on universal. |
13:50.59 | cr2 | maybe on athena. |
14:27.45 | *** join/#htc-linux tudenbart (n=willi@xdsl-213-196-248-46.netcologne.de) |
14:49.06 | *** join/#htc-linux the_sys0p (n=the_sys0@cpe-72-129-85-39.socal.res.rr.com) |
14:57.56 | *** join/#htc-linux Fossi (n=cmewes@195.177.48.115) |
15:50.33 | *** join/#htc-linux goxboxlive (n=goxboxli@206.80-202-161.nextgentel.com) |
16:03.38 | cr2 | BabelO: i'm patching the ba-asoc now. |
16:06.32 | goxboxlive | cr2: Do you know if Kevin2 has the source for only using one exe file for booting linux? |
16:06.51 | goxboxlive | I would like to have it, and make a cab file of it |
16:07.22 | cr2 | yes, it's in the haret source. |
16:09.21 | goxboxlive | ok, have to do a update then. |
16:12.30 | cr2 | you may need an empty ramdisk. |
16:12.37 | goxboxlive | ok |
16:13.06 | cr2 | a new ramdisk will be nice too. |
16:13.25 | goxboxlive | So it's not just to add the zImage, startup.txt and make a exe out of it then |
16:13.57 | cr2 | make linload SCRIPT=myharet.txt KERNEL=zImage INITRD=abc |
16:14.59 | cr2 | you should look at the source to see how to do it without a ramdisk. |
16:15.20 | goxboxlive | i have a new ramdiskbut why do i have to have a ramdisk for that. Shouldnt it be like booting as useual |
16:15.27 | goxboxlive | right |
16:23.56 | goxboxlive | Well it's failing when i just reomve the initrd line |
16:24.08 | goxboxlive | and tries compiling with my own startup.txt |
16:25.08 | goxboxlive | hmm it builded, have totest |
16:29.19 | goxboxlive | it wouldnt execute |
16:31.05 | goxboxlive | and the size seems ok |
16:36.54 | *** join/#htc-linux Olipro (n=Olipro@unaffiliated/olipro) |
16:37.10 | Olipro | wow, there's people in here |
16:37.31 | Olipro | Kevin2: if you're there, I need to pick your brain |
16:37.32 | goxboxlive | seems like i need to have a ramdisk |
16:39.47 | Olipro | by the way |
16:39.53 | Olipro | there's a #xda-devs on here |
16:43.24 | psokolovsky | cr2: committed SW support for asic3_base and cleaned it up a bit, but I really suggest we start to use gpiodev_keys instead. |
16:44.20 | Olipro | Kevin2: pof said you're the man to speak to about overriding the Windows Mobile Kernel |
16:44.36 | goxboxlive | psokolovsky: Seems like the compiling is failing now after the last update. |
16:44.56 | goxboxlive | hold on, i'll give you the output. Have to compile on local machine first |
16:46.25 | BabelO | cr2: ok, good :) you start updating pxa_gpio_mode and some clock setting ? |
16:46.48 | psokolovsky | goxboxlive: fixed |
16:47.01 | goxboxlive | :-) that was fast |
16:47.24 | BabelO | psokolovsky: what devices are already using gpiodev_keys ? |
16:47.58 | psokolovsky | BabelO: no so far in cvs, h5000 is in queue. but all should use it ;-) |
16:48.24 | BabelO | psokolovsky: that's to take so sample of code to do this ;) |
16:49.42 | goxboxlive | psokolovsky: It's still failing |
16:49.51 | goxboxlive | <PROTECTED> |
16:49.51 | goxboxlive | drivers/input/keyboard/asic3_keys.c: In function ‘asic3_keys_asic_handle’: |
16:49.51 | goxboxlive | drivers/input/keyboard/asic3_keys.c:39: error: ‘struct asic3_keys_button’ has no member named ‘type’ |
16:49.51 | goxboxlive | drivers/input/keyboard/asic3_keys.c: In function ‘asic3_keys_probe’: |
16:49.51 | goxboxlive | drivers/input/keyboard/asic3_keys.c:64: error: ‘struct asic3_keys_button’ has no member named ‘type’ |
16:51.17 | goxboxlive | i'll try again |
16:51.49 | psokolovsky | goxboxlive: sorry again, fixed too |
16:51.56 | goxboxlive | :-) |
16:52.00 | goxboxlive | thx |
16:53.32 | goxboxlive | psokolovsky: Now it compiled |
17:04.00 | *** join/#htc-linux rob_w (n=bob@p213.54.179.222.tisdip.tiscali.de) |
17:16.07 | goxboxlive | cr2 / psokolovsky: The updates of asic3_keys and i guess htcuniversal_buttons killed the touchscreen again. |
17:16.30 | psokolovsky | goxboxlive: you use ts-adc-debounce? |
17:16.49 | goxboxlive | no i dont think so |
17:17.00 | goxboxlive | have to put my child to bed. bbl |
17:32.36 | Olipro | pof: nobody wants to talk to me :( |
17:35.58 | *** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be) |
17:43.24 | *** join/#htc-linux Fossi (n=Fossi@e176115206.adsl.alicedsl.de) |
17:46.32 | goxboxlive | pH5: When i abort a upcomming call for my A780 the debugmessage gives me nothing back |
17:47.39 | *** join/#htc-linux BabelO (n=kvirc@lun34-2-82-238-28-28.fbx.proxad.net) |
17:48.06 | *** join/#htc-linux asylumed (n=insanity@vc-196-207-41-253.3g.vodacom.co.za) |
17:48.14 | pH5 | goxboxlive: if we get no unsolicited messages at all, we have to make gsmd periodically poll the connection state :( |
17:49.22 | goxboxlive | When i hangup with my a780 during a active phone call (talking) it gives me "NO CARRIER" when hanging up the phone with my A780 |
17:50.27 | goxboxlive | But a strange thing happend now: when i did: at+cfun=1 it gave med "OK" back? not "0" as it used to be. |
17:54.49 | cr2 | BabelO: i've made a lot of changes in the ba code. |
17:55.58 | goxboxlive | pH5: When i made a call form universal and refused to answer with my A780 it gave me "7" |
17:55.59 | cr2 | the EXTSSP should be ok now. but the CODEC reset and related gpios do not look good. we need more gpio tracing. |
17:55.59 | goxboxlive | not refused but denied the incomming call |
17:56.16 | goxboxlive | cr2; you are back. The TS is broken again after update of the switches |
17:56.40 | cr2 | dmesg ? |
17:57.04 | goxboxlive | htcuniversal_buttons where updated. Psokolovsky asked me if we where using ts-adc-debounce? |
17:57.11 | cr2 | no. |
17:57.27 | cr2 | but it's certainly not the problem. |
17:57.51 | goxboxlive | <5>HTC Universal Touch Screen Driver |
17:57.51 | goxboxlive | <4>htcuniversal: ts_probe |
17:57.52 | goxboxlive | <6>input: htcuniversal_ts as /class/input/input0 |
17:57.52 | goxboxlive | <7>PM: Adding info for platform:asic3-keys.0 |
17:58.12 | cr2 | do you get ts irqs ? |
17:58.58 | goxboxlive | http://pastebin.ca/445058 |
17:59.24 | goxboxlive | doesnt seems like i am |
17:59.39 | BabelO | ok nice cr2 do you have commited your change ? |
17:59.41 | cr2 | asic3_irq_demux: interrupt processing overrun |
17:59.54 | cr2 | it's an asic3 gpio/irq problem |
18:00.06 | cr2 | and has nothing to do with the ts driver itself. |
18:00.17 | cr2 | BabelO: yes. |
18:00.42 | goxboxlive | We had the same problem yesterday. Then what is it to do about it? |
18:01.16 | cr2 | locally i have patched asic3_keys |
18:01.23 | cr2 | i need to rediff the code. |
18:01.28 | goxboxlive | right |
18:01.56 | goxboxlive | could you send me the patch when you are done? |
18:03.08 | cr2 | psokolovsky: i'll add the irq error handling routines to asic3_keys. not fixing buggy drivers when the code is available is just silly. |
18:03.29 | psokolovsky | cr2: please do it like gpio_keys does then ;-) |
18:03.52 | cr2 | ok, i'll have a look. |
18:03.54 | psokolovsky | it mostly useless anyway, but at least we'll get used tow rite code like mainline like sit ;-) |
18:04.01 | psokolovsky | likes it |
18:04.30 | psokolovsky | my real concern was still that we have consistent external interface |
18:04.55 | psokolovsky | (struct *_button's, and their handling) |
18:05.04 | cr2 | if the button table has wrong entries it will silently fail in request_irq |
18:05.27 | cr2 | and then you'll get a flood of free_irq() messages in syslog |
18:06.16 | cr2 | goxboxlive pastebin from yesterday was exactly like that. |
18:06.16 | goxboxlive | yes that's right |
18:07.01 | psokolovsky | cr2: well, that's another matter - borked data. but asic3_keys exists for few months, and known to work well. it would be time to cleanup it, not to add more debug info into it. |
18:07.11 | psokolovsky | well, anyway, do as you see fit |
18:07.29 | psokolovsky | I just wished we didn't do the useless work or same work twice |
18:08.10 | cr2 | i have nothing against backwards compatibility and consistency :) |
18:08.31 | psokolovsky | ok then ;-) |
18:18.06 | cr2 | pH5: did you add the wince hack to asoc pxa2xx-ssp.c ? |
18:18.19 | *** join/#htc-linux _law_ (n=_law_@213.173.86.202) |
18:19.29 | pH5 | cr2: yes. insight/improvements welcome. |
18:20.05 | cr2 | have a look at my update |
18:21.27 | pH5 | cr2: ok. please send changes to asoc stuff to liam, too. though it looks like he is a little busy atm. |
18:22.10 | cr2 | i think we should fix the bugs first. non-working patches are not nice ;) |
18:22.57 | cr2 | you have SSCR0_DataSize(16) |
18:23.08 | cr2 | are you sure it's right ? |
18:23.21 | pH5 | cr2: wait - you set SSCR1_RxThresh bits?? this is a write to SSCR0! |
18:23.41 | cr2 | becasue ba has SSCR0_DataSize(10) |
18:23.48 | cr2 | yes, i'll correct it. |
18:24.04 | cr2 | and i have checked the BA values with the manual. |
18:24.29 | *** join/#htc-linux eremias (n=fdsfs@ACB20932.ipt.aol.com) |
18:24.37 | pH5 | my values are obtained with haret. |
18:25.28 | cr2 | hm. strange. |
18:26.10 | pH5 | but hima/ba are pxa2[56]x, they don't support 32bit transfers, right. |
18:26.59 | pH5 | maybe we should add support for the EDSS bit to DataSize to make things more clear. |
18:27.23 | cr2 | yeas, pxa263 |
18:35.50 | cr2 | goxboxlive: please update from CVS and if the ts will not work, pastebin the dmesg. |
18:35.59 | goxboxlive | cr2: thx, i will |
18:39.42 | oxi_ | ts in uni? |
18:41.14 | goxboxlive | oxi_ yes |
18:41.41 | cr2 | root hus has lost power or was reset |
18:41.43 | oxi_ | i'm going to try too. But wasn't that resolved already? |
18:41.51 | cr2 | s/hus/hub/ |
18:42.14 | goxboxlive | so dead |
18:42.32 | cr2 | the keyboard works for me after resume. |
18:42.45 | cr2 | but resume works only from the console. |
18:43.52 | BabelO | cr2: you do a before summer cleaning on BA code ;) on you set correct battery values ;) much thanks |
18:44.38 | goxboxlive | cr2: http://pastebin.ca/445144 |
18:45.31 | goxboxlive | cr2: Yes the keyboard works. But notice the lag and triple write of lethers. It is not normal behavour before suspend. |
18:47.25 | *** join/#htc-linux TripleDES (n=sergio@89.129.223.62) |
18:50.04 | cr2 | goxboxlive: ok. |
18:51.08 | goxboxlive | s/before/like before/ |
18:53.00 | *** join/#htc-linux WizMaui (n=WizMaui@62.112.90.250) |
18:53.10 | oxi_ | can u point me to a web on setting a crosscompiling environment and maybe on how to use haret .. ? |
18:53.48 | goxboxlive | universal wiki |
18:54.08 | oxi_ | thanks |
18:55.27 | pH5 | goxboxlive: tomorrow I'll visit harald welte - anything I should bring up besides the known gsmd issues? |
18:56.28 | goxboxlive | pH5: Yes, openmoko-dialer should have suppport for taking keyboard input, not just using the touchscreen |
18:56.48 | pH5 | indeed :) |
18:58.15 | BabelO | pH5: how do you run openmoko on small LCD ? |
18:58.30 | BabelO | pH5: i have qvga theme but dialer is too big |
18:58.36 | goxboxlive | sms isent finished yet so it's hard to get any idea of what we miss there. |
18:59.04 | cr2 | BabelO: i vote for ncurses :) |
18:59.32 | BabelO | cr2: yes a little command line, and power consumtion are low |
18:59.33 | pH5 | BabelO: I only have crude WIP patches to make parts of the dialer fit. |
18:59.33 | cr2 | pH5: openmoko-dialer alsa interface needs some written standard. |
18:59.55 | BabelO | pH5: where is this patch ? |
18:59.57 | pH5 | cr2: good one |
19:01.24 | cr2 | pH5: is the LCD on neo1973 controlled over spi ? |
19:01.25 | pH5 | BabelO: http://en.pastebin.ca/445170 |
19:02.01 | pH5 | cr2: I think so. stefan_schmidt would know, I think he played with it. |
19:02.38 | BabelO | pH5: thanks ;) i m trying GSM command over libgsm-tools |
19:02.48 | cr2 | is it a toppoly device ? |
19:03.05 | cr2 | because blueangel6 does it too. |
19:03.46 | pH5 | yes, toppoly. the exact model somewhere on the wiki |
19:03.52 | cr2 | ok. |
19:04.01 | cr2 | i#ll look into the code. |
19:04.22 | cr2 | BabelO: have you tried the sound ? |
19:05.02 | BabelO | cr2: yes, 1mn ago, uda1380 well detected ssp1 mapping ok, i2s mapping ok |
19:05.12 | oxi_ | how do I mount the internal eeprom(whatever :D)? |
19:05.31 | cr2 | oxi_: does not work. |
19:05.40 | oxi_ | :S |
19:05.47 | cr2 | there is no mdoc g3 driver. |
19:06.06 | cr2 | and no usable documentation. |
19:06.37 | cr2 | it may be possible to write a readonly driver. |
19:06.46 | BabelO | cr2: but nothing more |
19:06.53 | cr2 | with the already decoded info. |
19:07.10 | cr2 | BabelO: no sound ? |
19:07.22 | BabelO | cr2: no |
19:07.27 | cr2 | ok. |
19:07.56 | cr2 | BabelO: does the phone work ? |
19:08.22 | goxboxlive | cr2: did you get any info about the non working TS i pastebin? |
19:08.36 | cr2 | i think you don't really need uda1380 for the phone. |
19:08.42 | cr2 | goxboxlive: yes. |
19:08.50 | cr2 | <4>asic3_irq_demux: interrupt processing overrun |
19:08.57 | cr2 | that's the problem. |
19:09.15 | BabelO | cr2: i know that uda1380is not needed for mic but for speaker ? maybe |
19:09.17 | goxboxlive | right the same as yesterday. Is there any thing i can comment out to have it working again? |
19:09.58 | BabelO | cr2: i m trying with little pH5 ti_htc prog... but got a strange problem |
19:10.32 | cr2 | goxboxlive: i know one thing we are not doing right, i'll correct it now. it's asic3-related. |
19:10.50 | goxboxlive | ok, i'll try again |
19:10.56 | oxi_ | I'm going to be honest. You people obviously know your stuff and probably much electronics. I'm just a high level programer and don't have a clue on how to get started into this. Is there somewhere (website, chat, whatever) where I could get started into this, so I could contribute in some way? I do want to contribute, but maybe it's a better idea I don't? Please help me either by pointing me to a good way of getting started in g |
19:10.57 | oxi_ | eneral terms on hardware hacking or maybe just tell me nicely it's not worth the effort for a high level programer without any knowledge on hw. Thanks and sorry for the monologue :D |
19:11.03 | BabelO | cr2: i heard a small click from BA speaker when i try to play sound |
19:11.39 | cr2 | oxi_: it's not worth it. |
19:11.53 | oxi_ | oooh :/ |
19:12.24 | goxboxlive | oxi_ But we need realy help with universal support in opie and x11 |
19:12.25 | oxi_ | well, then maybe I could just test and post results or something? |
19:12.48 | pH5 | BabelO: what is your problem with the ldisc? for gsmd to use it, we probably need to tell it about blueangel. |
19:12.57 | cr2 | oxi_: it's too difficult, and requires 1-2 months of hardcore disassembling and documenting. |
19:13.10 | cr2 | pH5: tell it about himalaya too :) |
19:13.17 | oxi_ | i c |
19:14.09 | oxi_ | goxboxlive, in what way? maybe graphical apps to control specific hardware? :) |
19:14.12 | cr2 | oxi_: there is a linux binary driver for 2.4 |
19:14.50 | goxboxlive | oxi_ exactly we need that. |
19:15.08 | pH5 | cr2: I already did, actually. just add the info to make OE use the plugin patch for ba and hima, too. |
19:15.15 | cr2 | goxboxlive: it is still hardware-related. |
19:15.42 | cr2 | pH5: i wish the magician image should work on all 3 phones. |
19:15.51 | oxi_ | goxboxlive, I have no experience with opie, but have some with qt and Ive been told its alike. Maybe I could help (at last!) :) |
19:15.55 | pH5 | cr2: I know, I didn't forget about that. |
19:16.33 | cr2 | oxi_: do you know qt4 ? |
19:16.36 | goxboxlive | oxi_ Of course you can, there is anapplication named opie-camera. We need it to work with our camera. |
19:17.01 | oxi_ | cr2, no :( |
19:17.56 | oxi_ | goxboxlive, so the drivers for the camera are already working? |
19:19.01 | oxi_ | i'll take a look into it. Thanks :) |
19:19.05 | BabelO | cr2: sound seems to be close to works :) /dev/dsp is accessible, maybe need to check gpio to set speaker |
19:19.48 | BabelO | pH5: i do some test |
19:19.51 | BabelO | <7>n_tihtc_tty_ioctl called 5401 |
19:19.51 | BabelO | <7>n_tihtc_tty_ioctl called 400455c8 |
19:19.51 | BabelO | <7>n_tihtc_tty_read called |
19:19.51 | BabelO | <7>channel rx: ATCMD |
19:19.57 | cr2 | BabelO: i think so. |
19:20.29 | BabelO | cr2: do you think i don't need to start with your haret which play sound ? |
19:21.58 | cr2 | you can always try. |
19:22.39 | BabelO | ok ;) icheck gpio from blueangel_audio.c and after try reboot |
19:25.16 | BabelO | pH5: gsmd log http://pastebin.ca/445213 |
19:26.41 | goxboxlive | <oxi_> There are not yet made, but they are ready. Get known with the code so you are familia with it. But just so you know, non of us are familiar with it. There is a guy named mickeyl wich have made it. He is sometimes at #angstrom |
19:27.52 | oxi_ | goxboxlive, ok thank you :) |
19:27.55 | pH5 | BabelO: thanks, but that log looks incomplete. what happens after submitting ATZ? |
19:30.15 | BabelO | pH5: nothing :( i try to set DEBUG channel in tihtc |
19:30.56 | goxboxlive | oxi_ : http://handhelds.org/cgi-bin/cvsweb.cgi/opie/noncore/multimedia/ |
19:30.57 | BabelO | pH5: but on BA we habe 0x16 and 0x02, maybe it is that ? |
19:31.09 | pH5 | BabelO: so it just hangs? |
19:31.35 | pH5 | and after |
19:31.35 | pH5 | <7>channel rx: ATCMD |
19:31.35 | pH5 | you don't get anything else? |
19:31.51 | oxi_ | thanks |
19:32.08 | BabelO | pH5: i got that after leaving gsmd |
19:32.10 | BabelO | full talloc report on 'GSMD' (total 41 bytes in 6 blocks) |
19:32.10 | BabelO | <PROTECTED> |
19:32.10 | BabelO | <PROTECTED> |
19:32.10 | BabelO | <PROTECTED> |
19:32.10 | BabelO | <PROTECTED> |
19:32.11 | BabelO | <PROTECTED> |
19:32.33 | BabelO | and many lines like this <7>n_tihtc_poll called in dmesg |
19:33.03 | pH5 | BabelO: can you pastebin a complete dmesg dump? |
19:33.21 | pH5 | the many polls are ok, because that is what gsmd uses to look if there is data available. |
19:34.00 | BabelO | pH5: hum i need to restart my BA if you want a clean dmesg dump ;) |
19:34.16 | pH5 | "i try to set DEBUG channel in tihtc" - what do you mean by that? the tihtc plugin has to talk to the ATCMD channel. (0x16) |
19:34.55 | pH5 | BabelO: just running gsmd again and posting a big enough chunk might be good enough for now. |
19:35.05 | BabelO | pH5: i have set channel 0x16 in tihtc and in your sample it is 0x12 and it print DEBUG to me in dmesg |
19:35.12 | pH5 | I'd like to see the part where gsmd sends the first command and what the modem returns. |
19:35.37 | BabelO | pH5: ok, but need to reboot, my battery test is out ;) |
19:35.44 | pH5 | very well :) |
19:36.00 | goxboxlive | pH5: it still seems like the numeric response is not quite ok. It still spam me with a bounch of at+cfun=1 returns "0" |
19:36.11 | pH5 | BabelO: you don't need the sample program at all, gsmd now sets the line discipline itself. |
19:36.44 | BabelO | pH5: hum sure ? what is revison number ? |
19:37.20 | pH5 | BabelO: none, it is the patch in oe.dev. |
19:37.32 | pH5 | plugins.patch |
19:38.07 | BabelO | humm, sorry i use goxboxlive image for universal.. how can i check if patch is here ? |
19:38.27 | pH5 | BabelO: it is there, because universal needs the plugins, too. |
19:38.38 | BabelO | ok so good ;) |
19:39.20 | BabelO | pH5: ok i restart with sim registered |
19:39.30 | pH5 | goxboxlive: that's not good. I'll have to look at the code again. |
19:40.11 | goxboxlive | pH5: look also at the mail laforge sendt us, he said something about setting some at commands us self. dont remeber wich. I'll check |
19:40.58 | pH5 | goxboxlive: he wrote that he didn't apply the '@' part for handling extended replies. I'll have to ask him how he wants to have this done in the vendor plugins. |
19:41.16 | goxboxlive | pH5 : he are saying that we have to set the V0 flag our self |
19:41.34 | goxboxlive | Numeric responses are only considered valid if phone is in V0 mode. The |
19:41.34 | goxboxlive | HTC specific plugin would then need to set the V0 flag, or rather have |
19:41.34 | goxboxlive | the plugin send V0 to the phone, and the core detect any v0/v1 command |
19:41.34 | goxboxlive | and set V0 flag accordingly. |
19:42.33 | BabelO | pH5: full dmesg of BA |
19:42.35 | pH5 | goxboxlive: the problem is that the phone switches to V0 mode behind our back and gsmd doesn't handle those 'invalid' numeric replies in a sane way at all. |
19:42.42 | BabelO | pH5: oups http://pastebin.ca/445243 |
19:42.46 | goxboxlive | right |
19:44.21 | goxboxlive | to have text support for my modem in cu i need to have atv1 and ate1 set (i dont remeber wich is for who). |
19:44.37 | BabelO | pH5: good good good i add -F on gsmd command line and now ;) http://pastebin.ca/445247 |
19:44.56 | pH5 | BabelO: aah, that's much better! :) |
19:45.28 | cr2 | goxboxlive: i've committed a phone fix to CVS. can you check it ? |
19:45.52 | goxboxlive | of course |
19:46.41 | BabelO | pH5: but i think i miss a SMS :( |
19:46.45 | cr2 | that's probably the most important of asic3 fixes. |
19:46.59 | goxboxlive | ok, good |
19:47.09 | cr2 | BabelO: gomunicator/gnokii ? |
19:47.15 | BabelO | pH5: have to get trame and decode |
19:47.29 | BabelO | cr2: no only starting gsmd |
19:47.31 | cr2 | but then you need to disable framing code. |
19:48.16 | pH5 | BabelO: the "<7>starting frame (channel 0x16) at 4" look bogus, so there are certainly bugs in the line discipline. |
19:48.27 | pH5 | but the at frames seem to be decoded correctly. |
19:49.18 | BabelO | pH5: yes there is a lot of 0x2 parasiting :( last time i filter all of this with gnokii |
19:50.06 | goxboxlive | cr2: no, still no touchscreen |
19:50.46 | *** join/#htc-linux WizMaui (n=WizMaui@62.112.90.250) |
19:51.02 | cr2 | goxboxlive: ok. but this was an obvious bug. |
19:51.14 | goxboxlive | <4>asic3_irq_demux: interrupt processing overrun |
19:51.15 | goxboxlive | ok |
19:55.13 | elysion | hey, could someone tell me what that wm6 is that people are talking about on the forum? Is it a newer version of windows or what? |
19:55.14 | cr2 | goxboxlive: can you do the following : |
19:55.16 | cr2 | elysion: yes, it#s a warez beta-copy |
19:55.17 | cr2 | goxboxlive: set_irq_type(base_irq + b->gpio, IRQT_BOTHEDGE); |
19:55.22 | cr2 | change this line to FALLING edge. |
19:55.22 | elysion | ok, so is it wise to install it? and what should i take into consideration when installing it? |
19:55.31 | cr2 | psokolovsky will shoot me now :) |
19:55.32 | elysion | what's the status on the hermes port atm? |
19:55.34 | cr2 | elysion: it' a linux channel. |
19:55.40 | cr2 | it boots. |
19:55.56 | elysion | yeah, sorry, but didn't quite know where else to ask |
19:55.57 | cr2 | but there is more research needed. |
19:56.03 | elysion | ok |
19:56.18 | elysion | so what are you currently working on? |
19:56.23 | cr2 | i don't give a fsck about new versions of wince. |
19:56.36 | cr2 | uni+ba+hima+(hermes/trinity) :) |
19:56.43 | elysion | hehe... well, i'd rather use linux too |
19:56.56 | elysion | quite fed up with wince on my hermes :/ |
19:57.01 | elysion | too buggy |
19:57.25 | cr2 | well, the biggest problem is the s3c2442 cpu. |
19:57.37 | cr2 | it's too different from pxa. |
19:57.37 | elysion | how is that causing problems? |
19:57.39 | elysion | ok |
19:58.08 | cr2 | Kevin2 has written an excellent haret tracing support for it too. |
19:58.37 | cr2 | btw, do you know what "reversed block" means on hermes ? |
19:58.45 | goxboxlive | cr2: Yesi can try, but first of all, where do i add the line? |
19:58.54 | cr2 | linux driver has some hard time with the nand chip on hermes. |
19:59.19 | cr2 | goxboxlive: drivers/input/keyboard/asic3_keys.c |
20:00.03 | cr2 | elysion: but the most urgent thing is to understand why the SD card does not work. |
20:00.09 | goxboxlive | cr2, ok, and where in that file do i add the line? |
20:00.25 | cr2 | search for IRQT_BOTHEDGE |
20:00.46 | goxboxlive | and replave it with the line you wrote? |
20:01.22 | goxboxlive | i found it, should i remove it? |
20:01.24 | cr2 | no. |
20:01.31 | cr2 | and replace it with IRQ_TYPE_EDGE_FALLING |
20:01.46 | cr2 | IRQT_BOTHEDGE -> IRQ_TYPE_EDGE_FALLING |
20:01.53 | goxboxlive | ok :-) |
20:02.29 | goxboxlive | cr2: then it look like this: set_irq_type(base_irq + b->gpio, IRQ_TYPE_EDGE_FALLING); |
20:02.34 | cr2 | yes. |
20:02.37 | goxboxlive | :-) |
20:04.12 | elysion | cr2: would love to help you but don't really understand enough to do anything useful |
20:04.32 | elysion | seems like a few years more studies might be a good idea :P |
20:05.26 | goxboxlive | cr2: still no touchscreen |
20:06.02 | cr2 | elysion: not years :) |
20:06.04 | goxboxlive | <4>asic3_irq_demux: interrupt processing overrun |
20:06.14 | goxboxlive | now at another place |
20:06.36 | goxboxlive | http://pastebin.ca/445291 |
20:07.09 | elysion | how does the touchscreen work? I understood from the datasheet of the processor that it would have somekind of a block for that use |
20:07.25 | cr2 | ok, move this line as follows: |
20:07.37 | elysion | didn't still quite understand how it's supposed to work |
20:07.44 | cr2 | elysion: look at the mach-g500.c code. |
20:08.12 | elysion | cr2: could you maybe give me an url? |
20:08.43 | cr2 | goxboxlive: IRQT_BOTHEDGE for the keys and IRQ_TYPE_EDGE_FALLING for the switches. |
20:08.55 | cr2 | elysion: get the hh.org CVS tree |
20:10.29 | cr2 | goxboxlive: |
20:10.29 | cr2 | - set_irq_type(base_irq + b->gpio, IRQT_BOTHEDGE); |
20:10.30 | cr2 | <PROTECTED> |
20:10.31 | cr2 | + set_irq_type(base_irq + b->gpio, IRQT_FALLING); |
20:10.35 | cr2 | <PROTECTED> |
20:10.39 | cr2 | <PROTECTED> |
20:10.43 | cr2 | <PROTECTED> |
20:10.43 | cr2 | + set_irq_type(base_irq + b->gpio, IRQT_BOTHEDGE); |
20:10.43 | cr2 | <PROTECTED> |
20:10.43 | cr2 | <PROTECTED> |
20:10.44 | cr2 | <PROTECTED> |
20:10.46 | cr2 | then the switch code will not use IRQT_BOTH, |
20:11.10 | cr2 | and for the switches there will be FE. |
20:12.46 | cr2 | imho it should be an option :) |
20:15.46 | *** join/#htc-linux asylumed (n=insanity@vc-196-207-41-253.3g.vodacom.co.za) |
20:18.37 | BabelO | goxboxlive, pH5: where are magician image ? |
20:18.57 | goxboxlive | BabelO: no idea |
20:19.34 | BabelO | ok :( |
20:19.34 | *** part/#htc-linux WizMaui (n=WizMaui@62.112.90.250) |
20:19.52 | BabelO | i want to try openmoko-diaker for 240x320 screen |
20:20.13 | cr2 | linuxtogo.com/~pH5 ? |
20:20.14 | goxboxlive | babelO: http://www.linuxtogo.org/~ph5/ |
20:20.15 | goxboxlive | yes |
20:20.49 | BabelO | cr2: goxboxlive thanks :) |
20:21.11 | pH5 | BabelO: openmoko-dialer is not in the image yet. |
20:21.15 | BabelO | cr2: it is strange gpio seem to be good for sound |
20:21.19 | BabelO | pH5: :( |
20:21.21 | cr2 | lol. |
20:22.18 | pH5 | BabelO: you can't even cancel a call on qvga with that patch :) |
20:22.29 | goxboxlive | cr2: i realy have problem with this. Am i still editing the asic3_keys file? |
20:22.58 | cr2 | yes. |
20:23.08 | goxboxlive | i think i found soemthing |
20:24.09 | BabelO | pH5: but i can't compile openmoko, i didn't have OE, to big for my hardrive (maybe i need to buy a new one before french gouverment tax hardrive at and of june ) ;) |
20:24.33 | goxboxlive | cr2: should i stillhave this one i edited? set_irq_type(base_irq + b->gpio, IRQ_TYPE_EDGE_FALLING); |
20:25.40 | pH5 | BabelO: I can upload some ipks, but I won't include it in the images until at least making/receiving/canceling calls works. |
20:26.36 | BabelO | pH5: ok so it can works in an opie image with Xqt2 (i ve goxboxlive opie image) |
20:27.48 | goxboxlive | booting |
20:28.50 | BabelO | cr2: you fix param_rate to 48000 ? |
20:29.21 | goxboxlive | cr2: still no touchscreen |
20:31.37 | cr2 | BabelO: ? 32 -> default divisor for 44100 |
20:32.08 | BabelO | case 48000: |
20:32.08 | BabelO | <PROTECTED> |
20:32.08 | BabelO | <PROTECTED> |
20:34.30 | *** join/#htc-linux oxiii (n=fdhgf@101.Red-83-46-140.dynamicIP.rima-tde.net) |
20:34.33 | cr2 | BabelO: it's for all cases |
20:34.34 | BabelO | cr2: oups sorry |
20:34.46 | BabelO | cr2: iv' not seen that there is only one break ;) |
20:34.49 | BabelO | sorry |
20:36.59 | goxboxlive | still the same <4>asic3_irq_demux: interrupt processing overrun. |
20:37.11 | goxboxlive | What about the stuff you did yesterday cr2: You comnted out some stuff. Can we just do that? |
20:37.43 | *** join/#htc-linux asylumed (n=insanity@196.30.118.75) |
20:47.08 | cr2 | goxboxlive: comment out the switches. |
20:47.11 | goxboxlive | in asic3_keys? |
20:47.14 | goxboxlive | found it |
20:47.14 | goxboxlive | htcuniversal_buttons.c |
20:51.08 | goxboxlive | cr2: yes, now with the switches commented out touchscreen works |
20:51.09 | pH5 | BabelO: ipks are on ltg.org |
20:51.10 | BabelO | pH5 big thanks, i ll try |
20:51.12 | goxboxlive | pH5: you numeric patch is gone. WHat do i have to change in the gsmd to have the at+cfun=1 working again? |
20:53.07 | cr2 | goxboxlive: you can try to enable switches one after another. |
20:53.30 | cr2 | to find out if it#s a general problem, or just 1 switch. |
20:53.41 | goxboxlive | cr2: Yes, i will.I'll let you know wich one whocauses the probolem |
20:54.01 | pH5 | goxboxlive: try to set g->flags |= GSMD_FLAG_V0; in usock_rcv_phone in usock.c (right before the AT+CFUN=1 atcmd_fill) |
20:54.14 | goxboxlive | pH5 thx |
20:54.16 | pH5 | I'm not sure if that will work. And it's definitely not the correct way. |
20:54.22 | goxboxlive | :-) |
20:55.06 | pH5 | I'll try to discuss this tomorrow, I think the best way would be to remove the GSMD_FLAG_V0 checks in atmcd.c and just accept both numeric and verbose replies as before. |
20:55.21 | goxboxlive | ok |
20:56.04 | cr2 | BabelO: there is a "playsound" command in haret, you can try to trace it too. |
20:56.57 | BabelO | cr2: got this few days ago http://pastebin.ca/440970 |
20:57.21 | BabelO | there is CONTROL_CX in |
21:04.40 | cr2 | nothing spectacular. |
21:04.50 | goxboxlive | pH5: not like this: cmd = atcmd_fill g->flags |= GSMD_FLAG_V0;("AT+CFUN=1", 9+1, this failed compiling. How exactly? |
21:08.16 | pH5 | goxboxlive: I meant you should add that on a new line before the atcmd_fill. And I just notice that g->flags should be gu->gsmd->flags instead. |
21:08.19 | BabelO | cr2: and this from i2c http://pastebin.ca/445413 |
21:12.31 | cr2 | i'm looking for CODECpower/reset |
21:14.12 | BabelO | ok |
21:18.58 | goxboxlive | it didnt help. I'll wait until you have talked to harald tomorrow. |
21:22.37 | pH5 | goxboxlive: ok. I'll stop for today, can't make the openmoko-dialer outgoing window show the cancel button on qvga.. |
21:22.42 | pH5 | good night |
21:23.02 | BabelO | good night pH5 |
21:23.30 | goxboxlive | good night |
21:46.56 | cr2 | BabelO: did not find anything usable. the code structure is different ;) |
21:47.37 | BabelO | cr2: nothing, i found all pxa_gpio_mode |
21:49.14 | BabelO | cr2: does mplayer running nice video on universal ? |
21:52.43 | cr2 | don't know. |
21:53.18 | cr2 | should be slower because of the 4x LCD and main SDRAM. |
21:55.44 | BabelO | i ve tried on BA, it works with a not resized video, but my Film is HD and have some lag |
21:57.32 | oxiii | could somebody test pasting in pastebin. It's not working for me :S |
21:57.47 | oxiii | s/test/try/ |
21:58.00 | BabelO | hum sound is strange... i ve to try to put Codes power/reset in blueangel_ext_control |
21:58.52 | *** join/#htc-linux g3gg0_ (n=g3gg0@ppp-82-135-12-228.dynamic.mnet-online.de) |
22:00.22 | cr2 | BabelO: how do you know which are they ? |
22:00.52 | cr2 | we need to force a suspend/resume event while tracing. |
22:01.08 | cr2 | PWRCTL WAV1 4 |
22:01.17 | cr2 | does not produce anything. |
22:01.40 | cr2 | you may try it on SPK1 and SSP1 |
22:02.50 | *** join/#htc-linux Funklord (n=cow@c-cbd572d5.014-46-73746f28.cust.bredbandsbolaget.se) |
22:02.55 | BabelO | cr2: some off them are written in blueangel_audio.c |
22:03.09 | BabelO | ok need to eat some food before |
22:07.14 | cr2 | i see only SSP1 gpio setup. |
22:10.29 | BabelO | this one are for ssp1 ? |
22:10.31 | BabelO | pxa_gpio_mode(GPIO23_SCLK_MD); |
22:10.31 | BabelO | <PROTECTED> |
22:10.31 | BabelO | <PROTECTED> |
22:11.31 | BabelO | ssp seems to works, because when playing sound with player, i see some debug trace from uda1380 |
22:12.59 | cr2 | yes, but do we setup them now ? |
22:13.26 | cr2 | wince does it dynamically (see wiki). |
22:16.17 | BabelO | what pH5 say about DataSize ? we have 0x10 in old driver too |
22:16.50 | cr2 | i'll commit a real patch now. |
22:17.15 | BabelO | ok |
22:17.57 | cr2 | so, where do we setup the ALT gpio SSP1 pins ? |
22:22.21 | BabelO | it is done in init just after CONTROL_CX in old driver ... |
22:22.59 | cr2 | yes, be we need it in the new driver too. how does magician do it ? |
22:24.01 | BabelO | same place for magician in end of init |
22:24.32 | cr2 | is see only pxa_gpio_mode(GPIO27_SSP_EXT_CLK); |
22:25.16 | cr2 | and imho it should be done by the pxa2xx-ssp module |
22:25.49 | BabelO | it works ;) |
22:25.52 | BabelO | sound ;) |
22:26.15 | BabelO | i try without adding pxa_gpio_mode now |
22:26.38 | cr2 | <PROTECTED> |
22:26.43 | cr2 | this is wrong. |
22:27.20 | BabelO | oups, speed is to slow :( |
22:27.44 | cr2 | hm, maybe only the comment is wrong. |
22:27.50 | BabelO | where is this line ? |
22:28.44 | cr2 | sorry. it's for capture. |
22:29.28 | BabelO | ok, cr2 it is your last comit |
22:29.46 | cr2 | static int blueangel_playback_prepare(struct snd_pcm_substream *substream) |
22:30.01 | BabelO | i remove pxa_gpio_mode line, i heard sound too very slow |
22:30.01 | cr2 | probably the SSP gpios should be setup here. |
22:30.12 | cr2 | which one ? |
22:30.19 | cr2 | pxa_gpio_mode line |
22:30.54 | BabelO | i take GPIO24_SFRM_MD GPIO25_STXD_MD GPIO23_SCLK_MD |
22:31.14 | cr2 | GPIO27_SSP_EXT_CLK for the mode ? |
22:31.22 | BabelO | oki rebbot again to test |
22:32.11 | cr2 | ok. |
22:33.17 | BabelO | it is better to test from fresh boot ;) |
22:34.10 | cr2 | i'll add the modes. |
22:39.29 | BabelO | humm nothing playing now ;) |
22:39.36 | BabelO | i try adding line by line |
22:39.50 | cr2 | <PROTECTED> |
22:39.51 | cr2 | <PROTECTED> |
22:39.51 | cr2 | <PROTECTED> |
22:39.51 | cr2 | <PROTECTED> |
22:39.58 | cr2 | this is useless. |
22:40.19 | cr2 | "GPIOA_CODEC_PWR_ON" is actually CONTROL_CX |
22:40.36 | cr2 | and "GPIOB_CODEC_RESET" is speaker power :) |
22:40.48 | BabelO | ok i see that you have already added if def 0 for these lines |
22:40.56 | cr2 | obviously based on a haret trace. |
22:41.01 | cr2 | i |
22:41.07 | cr2 | 'll remove them. |
22:41.14 | cr2 | completely. |
22:42.34 | cr2 | i think "slow" is not because of EXTSSP. |
22:42.51 | cr2 | maybe lkcl has chosen wrong fs multipliers. |
22:43.50 | BabelO | ok these lines are needed i didn't test all combination but only the first two didn't works |
22:43.53 | BabelO | GPDR(GPIO23_SCLK) |= GPIO_bit(GPIO23_SCLK); |
22:43.53 | BabelO | <PROTECTED> |
22:43.53 | BabelO | <PROTECTED> |
22:43.53 | BabelO | <PROTECTED> |
22:43.53 | BabelO | <PROTECTED> |
22:43.54 | BabelO | <PROTECTED> |
22:44.07 | BabelO | at and of init function |
22:45.29 | BabelO | i says first two for combination of GPDR... adn pxa_gpio_mode if i understand |
22:46.41 | cr2 | i've added them. |
22:46.53 | cr2 | what do you think is wrong with this code ? |
22:46.57 | BabelO | it this line for speed divider ? div = 1411200/params_rate(params) |
22:47.02 | cr2 | yes. |
22:47.16 | cr2 | this number goes into |
22:47.44 | cr2 | SSCR0_P(port) |= SSCR0_SerClkDiv(div); |
22:47.52 | cr2 | whic is |
22:48.49 | cr2 | #define SSCR0_SerClkDiv(x) ((((x) - 2)/2) << 8) /* Divisor [2..512] */ |
22:49.27 | BabelO | div=705600/val; |
22:49.32 | BabelO | in old driver ... |
22:50.46 | BabelO | ok i try again with this divisor |
22:51.07 | BabelO | better ;) |
22:52.20 | BabelO | i need to test different music to know if speed is normal ;) |
22:52.27 | cr2 | LOL |
22:52.41 | BabelO | there is no bass boost :( |
22:53.05 | *** join/#htc-linux twostupiddogs (n=twostupi@62.123.180.130) |
22:53.20 | BabelO | cr2: why in old driver tehre is this two line |
22:53.23 | BabelO | div = 705600/params_rate(params); |
22:53.23 | BabelO | <PROTECTED> |
22:53.34 | BabelO | why substract 1 to result ? |
22:54.46 | cr2 | div << 8 /* SSPEXTCLK/(2*(div+1)) */ |
22:54.48 | BabelO | good ;) Der Ententanz is at the right speed ;) |
22:55.31 | BabelO | ok ;) |
22:55.53 | cr2 | so what divisor is used ? |
22:56.04 | BabelO | the two line |
22:56.13 | cr2 | ? |
22:56.24 | BabelO | 705600/params_rate(params) -1 |
22:56.58 | cr2 | then something is wrong. |
22:57.04 | BabelO | why ? |
22:57.46 | cr2 | can you dump the SSCR0 register in linux ? |
22:57.53 | cr2 | with devmem2 |
22:57.57 | BabelO | and the very good news, is that i boot with standard haret ;) |
22:58.17 | cr2 | ok, that't obvious. |
22:58.19 | BabelO | yes sure, how i can do this ? |
22:58.42 | cr2 | i've patched blueangel_audio.c to switch on the CDEX_SOURCE1 |
22:58.45 | BabelO | i read log about devmem today |
22:58.57 | cr2 | it was not in the .12 kernel |
22:59.15 | cr2 | do you have devmem2 program ? |
22:59.21 | BabelO | ah ah :) you have lot of good idea |
22:59.33 | cr2 | it's good that it works, but i'd like to understand why :) |
22:59.55 | BabelO | is it in opie image from goxboxlive ? |
23:00.17 | cr2 | it's part of busybox afaik. |
23:01.42 | BabelO | not in mine http://pastebin.ca/445574 |
23:02.30 | cr2 | hm. well. |
23:02.54 | cr2 | you can use 'dd' and 'od -x' :) |
23:03.08 | BabelO | there is no /proc/xxxx to do this ? |
23:03.09 | cr2 | but a wrong command will do oops() |
23:03.12 | cr2 | no. |
23:03.53 | *** join/#htc-linux guanxo (n=guanxo@81.203.40.125.dyn.user.ono.com) |
23:04.14 | *** join/#htc-linux psokolovsky_ (n=psokolov@empty-crossing.volia.net) |
23:04.39 | cr2 | dd if=/dev/mem bs=4 count=1 skip=$((0x41000000/4)) | od -t x4 |
23:04.50 | cr2 | i think i've got it right. |
23:04.55 | cr2 | i hope :) |
23:05.26 | BabelO | :) oops() |
23:06.28 | BabelO | od: invalid option -- t |
23:06.34 | cr2 | do you see an error ? |
23:06.40 | cr2 | in the command. |
23:06.52 | BabelO | right no -t with od |
23:08.00 | BabelO | dd if=/dev/mem bs=4 count=1 skip=$((0x41000000/4)) |
23:08.04 | BabelO | this command segfault |
23:08.15 | cr2 | hm. |
23:08.18 | cr2 | strange. |
23:08.51 | cr2 | echo $((0x41000000/4)) ? |
23:09.27 | cr2 | imho it should work. |
23:10.09 | BabelO | ok rebooting to be sure ;) |
23:10.42 | cr2 | the "raw" divisor should be 32 0x2000 |
23:11.15 | cr2 | can you do a diff between the current CVS and your working code ? |
23:11.25 | BabelO | yes |
23:12.05 | cr2 | we also need to understand the headset switching gpios. |
23:12.27 | cr2 | playing music on the main speaker is not cool. |
23:13.26 | BabelO | yes, need to found al lswitch |
23:13.56 | BabelO | diff is here http://pastebin.ca/445591 |
23:15.18 | cr2 | is it the current ? |
23:15.27 | cr2 | well, let's see |
23:16.14 | cr2 | you push 705600/rate-1 into SSCR0_SerClkDiv(x) ((((x) - 2)/2) << 8) |
23:16.31 | BabelO | hum yes |
23:16.54 | cr2 | and get ((705600/rate-1 -2)/2) <<8 |
23:17.08 | BabelO | setting headset with alsamixer set sound on ba speaker and headset |
23:17.17 | BabelO | humm maybe need to remove -1 |
23:17.24 | cr2 | the original code was |
23:17.54 | cr2 | (705600/rate-1) << 8 |
23:18.20 | cr2 | can you explain me how do they match ? |
23:18.26 | BabelO | ouch big problem so |
23:18.28 | cr2 | or i'm completely dumb ? :) |
23:19.17 | cr2 | the manual says |
23:19.33 | BabelO | is there another divider than this one ? |
23:20.03 | cr2 | SSPEXTCLK/(2*(div+1)) |
23:20.15 | cr2 | is the external clock frequency. |
23:20.33 | cr2 | so in the old code. |
23:20.46 | cr2 | div=705600/rate-1 |
23:21.12 | cr2 | for the rate 44100 it gives |
23:21.57 | BabelO | old code say use pll =1 and fs=256 |
23:22.32 | cr2 | SSPEXTCLK/(2*(705600/44100-1+1)) |
23:22.44 | cr2 | what does the new use ? |
23:23.29 | cr2 | it seems you need to play a "known" 44100 sound in wince, and dump the 0x41000000 reg in haret |
23:23.38 | cr2 | then we will know. |
23:23.47 | cr2 | if the divisor is correct. |
23:24.05 | cr2 | pll=1 in the new code. |
23:24.58 | cr2 | .dac_clk = UDA1380_DAC_CLK_WSPL |
23:25.13 | BabelO | ok, but need to be sure to play 44100 sound under wince, that really not sure |
23:26.40 | BabelO | freq=1411200, div=32; |
23:27.19 | cr2 | 32 is not right :) |
23:27.29 | BabelO | so value 1411200 is already here but used for something else ;) |
23:27.54 | BabelO | it is magics not electronics, but sure it is not windows stuff ;) |
23:29.22 | cr2 | well, the only thing that counts is the SSCR0 register bit setting. |
23:30.31 | cr2 | we set |
23:30.35 | cr2 | <PROTECTED> |
23:30.35 | cr2 | <PROTECTED> |
23:30.52 | cr2 | but this bit is not defined on PXA26x |
23:32.08 | cr2 | we need to check the SSCR1 settings too. |
23:32.33 | cr2 | hm. dump pxa27xssp will not work. |
23:33.28 | cr2 | SND_SOC_DAIFMT_NB_NF - this flag is not useful. |
23:34.26 | cr2 | SND_SOC_DAIFMT_CBS_CFS - this one is empty. |
23:34.43 | cr2 | BabelO: can you remove this SND_SOC_DAIFMT_NB_NF flag ? |
23:34.59 | BabelO | yes i try |
23:35.59 | BabelO | strange, i ve the line two time |
23:36.55 | cr2 | i think we should ifdef all PXA27x flags in pxa-regs.h to avoid such confusion. |
23:37.13 | BabelO | works too |
23:37.19 | cr2 | it should. |
23:37.27 | cr2 | because it's useless :) |
23:37.34 | BabelO | :) |
23:38.06 | cr2 | so we need to dump the SCCR0 in wince while playing 44100 sound. |
23:38.41 | BabelO | i need to create a wav file for 44100 hz |
23:39.31 | cr2 | file *.wav |
23:40.06 | BabelO | ok i reboot in wince |
23:40.12 | *** join/#htc-linux cyrill62 (n=cyrill62@162.220.20.81.dynamic.adsl.abo.nordnet.fr) |
23:42.04 | BabelO | ok good i got one 44100 file |
23:42.15 | BabelO | so what is the right command with haret |
23:43.36 | cr2 | 41000000 | 000007df | .... |
23:43.45 | cr2 | i have that on a 44100 mp3 |
23:43.57 | cr2 | pd 0x41000000 4 |
23:44.34 | BabelO | 41000000 | 000007df same for me ;) |
23:44.36 | BabelO | goood |
23:45.19 | cr2 | 0111.1101.1111 |
23:46.49 | cr2 | 07 |
23:47.10 | cr2 | is the SCR |
23:47.15 | Kevin2 | Hi |
23:47.25 | Kevin2 | Olipro: Stil around? |
23:47.25 | BabelO | hi Kevin2 |
23:47.33 | Olipro | oh dear |
23:47.39 | Olipro | this means I'm in for a long night |
23:47.42 | Olipro | hi Kevin2 :P |
23:47.50 | cr2 | bit rate= SSPEXTCLK/(2*(SCR+1)) |
23:47.52 | Kevin2 | Hi all. |
23:48.00 | cr2 | hi Kevin2 |
23:48.22 | cr2 | 44100=SSPEXTCLK/16 |
23:48.47 | BabelO | cr2: so it is good |
23:49.05 | cr2 | hehe: |
23:49.08 | cr2 | 705600 |
23:49.22 | cr2 | so, what is the div |
23:49.26 | Olipro | Kevin2: I've been writing a file into a segment of memory (a SPL) but the memory is utilised by the kernel since after writing, parts get written over again |
23:49.53 | BabelO | yes 1411200 = 705600*2 |
23:50.15 | Olipro | I've successfully patched the IPL to not reload from NAND; so what I need to do is kill enough of the kernel such that I can read the file out into memory so it's not over written, then Branch to the IPL at 0x0 |
23:50.41 | cr2 | (div-2)/2=7 |
23:50.43 | Kevin2 | Olipro: What's your platform? |
23:50.49 | Olipro | HTC hermes |
23:51.01 | cr2 | div-2=14 |
23:51.06 | cr2 | div=16 |
23:51.25 | Kevin2 | You want to run the IPL at 0x0? And then have it launch your SPL at some area in memory? |
23:51.50 | cr2 | BabelO: so the "default" div is 16 |
23:52.05 | Kevin2 | Does your SPL need to be at some specific physical memory address? |
23:53.02 | BabelO | so i replace div=32 by default to div=16, but div is always calculated because of samplerate |
23:53.10 | cr2 | BabelO: then i don't understand the old code. |
23:53.29 | cr2 | 705600/44100 |
23:53.43 | BabelO | this two are closder 1411200 = 705600*2 |
23:53.44 | cr2 | ok, it's 16. |
23:53.53 | Olipro | Kevin2: the SPL has to exist at a fixed physical address... we can write the address, but if we do, the kernel has a spazz because it uses that memory area and eventually the device dies and reboots |
23:54.10 | BabelO | that s why maybe it never work when booting without playing sound :) |
23:54.39 | Kevin2 | Olipro: Are you using HaRET to do this, or are you doing it manually? Are you familiar with the HaRET code? |
23:54.48 | cr2 | BabelO: hm. i'll fix the new, and look at the old code. for the academic reasons :) |
23:55.05 | BabelO | cr2: ok |
23:55.21 | Olipro | Kevin2: it's a hacked up version of Haret; pof tried writing the SPL to memory manually, then getting HaRET to do it's WinCE killing stuff and jump to the IPL |
23:55.46 | BabelO | cr2: so i say to you a very good night, and tomorrow will try to have a working keyboard without repeat and after try to set all input.ouput for sound |
23:56.08 | cr2 | BabelO: good night. way a good day :) |
23:56.19 | BabelO | cr2: yes |
23:56.22 | BabelO | good night |
23:56.34 | Olipro | if you leave the IPL unpatched, it of course boots since the SPL is pulled freshly from NAND into RAM |
23:57.37 | cr2 | Olipro: why do you patch the IPL ? |
23:57.49 | Olipro | so it doesn't load the SPL from NAND ;\ |
23:57.57 | cr2 | why don't you just put your new SPL in place of the wince kernel ? |
23:58.04 | Olipro | and trust me, the patch works |
23:58.16 | Olipro | cr2: what are you talking about |
23:58.24 | cr2 | ok, but it's too risky ;) |
23:58.49 | cr2 | the IPL checks if you are in suspend |
23:59.01 | cr2 | that's how we resume on linux. |
23:59.31 | cr2 | and if you are in suspend, then it jumps to the "wince kernel" location, and not into SPL. |