00:00.31 | cr2 | hm, not sharp. hitachi |
00:00.36 | cr2 | G1000 |
00:03.53 | Kevin2 | I think I found it. |
00:05.39 | Kevin2 | cr2: http://pastebin.ca/496000 |
00:07.55 | cr2 | hehe. found it. |
00:07.58 | cr2 | http://www.phonenews.com/phones/sprint/hitachi/g1000/G1000SDIO.CAB |
00:08.31 | cr2 | Kevin2: what does the table say ? |
00:09.00 | Kevin2 | cr2: table? |
00:09.17 | cr2 | ATIWxxxx |
00:09.25 | cr2 | +0x2000 |
00:09.30 | cr2 | that makes sense. |
00:10.25 | Kevin2 | These are writes to physical 10202000 |
00:10.41 | Kevin2 | So it is +202000 |
00:12.02 | cr2 | 202000 ? |
00:12.53 | cr2 | ok. |
00:12.57 | Kevin2 | Right. These writes aren't to the first 1Meg ATI range. They're to the third meg of the range. |
00:13.56 | cr2 | 0924b058.... |
00:14.05 | cr2 | this is the call address ? |
00:14.25 | cr2 | 09b9f180: e593a200 ldr sl, [r3, #512] # r10=02000000 r3=48010200 |
00:14.36 | cr2 | 0924b058: e19330b1 ldrh r3, [r3, r1] # r3=00000300 r3=4c202044 |
00:14.45 | Kevin2 | Yes, that is the address of the instruction that made the access. I was just thinking the same thing - we could disassemble that area of ram. |
00:14.58 | cr2 | lsmod first. |
00:15.14 | cr2 | and look where is the ACE_DDI.dll mapped. |
00:15.53 | cr2 | imvho it should be ace_ddi |
00:15.57 | cr2 | but we will see. |
00:16.08 | Kevin2 | 1076 fl=00000000 mid=83760da0 pid=00000000 gusg= 11 pusg=00004018 base=01B40000 size= |
00:16.08 | Kevin2 | 00104000 hmod=83760DA0 mod=ace_ddi.dll exe= |
00:16.48 | cr2 | 01b4* -> 09b9* |
00:16.55 | cr2 | looks reasonable. |
00:17.25 | cr2 | it's a big driver, so the 5*64K offset may be ok. |
00:19.18 | cr2 | yeah, it's 1MB dll. |
00:19.26 | Kevin2 | ps: pid=837aa78e ppid=0 pmem=08000000 tcnt=109 perm=00000008 procname=device.exe |
00:20.31 | cr2 | you can make the raw (ROM?) dump around the call address |
00:20.33 | Kevin2 | The 09b* should be in "device.exe" memory area. |
00:21.01 | Kevin2 | I don't understand. |
00:21.19 | cr2 | yes, but the 32M offsets separate the "virtual" slots. |
00:21.38 | cr2 | the "zero" slot is for the running process. |
00:21.48 | cr2 | 0x01b4* is a zero slot address. |
00:22.23 | cr2 | and the "process 4" is device.exe |
00:22.50 | cr2 | it has it 32MB virtual space at +4*32M |
00:25.09 | Kevin2 | What were you saying about raw/ROM stuff? |
00:26.28 | cr2 | where is the ace_ddi.dll in ROM ? |
00:27.03 | Kevin2 | How do I find that out? |
00:27.11 | cr2 | i can look at my binary, but i'm not sure we have the same dlls. |
00:27.39 | cr2 | hmod=83760DA0 |
00:27.51 | cr2 | which is the physical address for it ? |
00:28.22 | cr2 | it's a virtual address for MODULE * structure |
00:28.44 | Kevin2 | HaRET(49)# show v2p(0x83760DA0) |
00:28.44 | Kevin2 | 0x33d60da0 |
00:29.18 | cr2 | ok, so it's in RAM. obviously ;) |
00:31.22 | cr2 | you can look by dumping this structure where is this dll located. |
00:31.31 | cr2 | the testwm5 stuff. |
00:33.28 | cr2 | AhiInit(0x11); |
00:33.38 | Kevin2 | I'm not sure what you're looking for. The program device.exe is in ram from 0x08000000 - 0x09999999. Doesn't this mean we know what process it is? |
00:34.54 | cr2 | these are virtual addresses. |
00:35.38 | cr2 | you mean it will be 0x09b4*, whereever it is mapped from ? |
00:35.54 | cr2 | AhiDevEnum(...) |
00:36.16 | cr2 | AhiDevOpen(ptr, |
00:36.41 | cr2 | ptr,"SD",0); |
00:36.48 | Kevin2 | There is a PID feature on ARM that remaps addresses. The wi stuff only shows the addresses after they've been remapped. So, the 0x09b4* address is the MVA address (modified virtual address). |
00:37.09 | cr2 | AhiDevOpen(ptr,ptr,"SD",0); |
00:37.19 | cr2 | yes, PID is the 32MB switch. |
00:37.37 | Kevin2 | Right. That only applies to addresses <= 0x01ffffff |
00:38.41 | cr2 | i'm a bit slow now, it's around 3:00 here ;) |
00:39.23 | cr2 | ret = AhiDevOpen(&ctx, asic, "haret", 0); |
00:39.34 | cr2 | hm. looks good. |
00:40.02 | Kevin2 | :) |
00:41.50 | cr2 | and then SrmGetProperty... |
00:45.53 | cr2 | 2010, 2040, 2044 . do you see them ? |
00:46.06 | cr2 | yes :) |
00:46.22 | Kevin2 | cr2: I'm a bit lost. I don't know what SrmGetProperty is.. |
00:46.35 | cr2 | ok, so the SD on older ATIW32xx is the same. or very similar. |
00:46.54 | cr2 | it's some function deep in the ACE_DDI.dll |
00:47.00 | cr2 | it does not matter. |
00:47.15 | cr2 | 2014 |
00:47.30 | cr2 | 200c and 2000 |
00:48.24 | cr2 | r3=r0+0x37c |
00:48.48 | cr2 | ldr r3, [r3,0x20] |
00:49.14 | cr2 | str r2, [r3,r1] |
00:49.24 | cr2 | r1 takes the values above. |
00:49.26 | Kevin2 | I see 4c202000 - 4c202060 |
00:50.03 | cr2 | and i see the r1 values explicitly hardcoded in the driver. |
00:50.32 | Kevin2 | 4c202060 seems to be the register where we actually transfer data. |
00:50.34 | cr2 | r0 is probably 'asic' variable, or something like that. |
00:50.39 | cr2 | or ctx. |
00:51.00 | cr2 | it's located in RAM , and can be dumped in order to check. |
00:51.10 | cr2 | it#s visible in ATIDBG command. |
00:53.00 | cr2 | +0x2000 is called SD_COMMAND |
00:54.26 | cr2 | hm. with 0x3c offset. |
00:54.51 | cr2 | demo version exited. |
00:55.01 | cr2 | ok, but at least i know where to look now. |
00:55.04 | cr2 | good night :) |
00:55.29 | Kevin2 | Good night. |
01:00.31 | *** join/#htc-linux TBBle (n=tbble@2001:388:f000:0:0:0:0:2d) |
01:06.08 | Kevin2 | cr2: I think you're right. I modified lsmod to take a pid number and I get: |
01:06.10 | Kevin2 | 1076 fl=00000000 mid=83760da0 pid=837aa78e gusg= 11 pusg=00000006 base=01B40000 size=00104000 hmod=83760DA0 mod=ace_ddi.dll exe= |
01:07.53 | Kevin2 | This was for "device.exe" (pid 0x837aa78e). It indicates ace_ddi.dll should be between 0x01b40000 - 0x1c44000. This (in MVA) is 0x09b40000 - 0x9c44000. Which is right in the instruction address range. |
01:13.35 | Kevin2 | Wait no - most SD access are in 0x0924* range. This equates to: |
01:13.43 | Kevin2 | 1076 fl=00000000 mid=8319997c pid=837aa78e gusg= 1 pusg=00000001 base=01240000 size=00014000 hmod=8319997C mod=sdhc_sc2410.dll exe= |
02:14.51 | Kevin2 | cr2: Do a "cvs up" followed by: |
02:14.57 | Kevin2 | HaRET(21)# addr2mod 0x09b9f180 |
02:14.57 | Kevin2 | Address 09b9f180 in process: device.exe (08000000 - 0a000000) |
02:14.57 | Kevin2 | <PROTECTED> |
02:14.57 | Kevin2 | HaRET(22)# addr2mod 0x0924f05c |
02:14.57 | Kevin2 | Address 0924f05c in process: device.exe (08000000 - 0a000000) |
02:14.57 | Kevin2 | <PROTECTED> |
03:18.35 | *** join/#htc-linux Whacko (n=Whacko@s5593e8a2.adsl.wanadoo.nl) [NETSPLIT VICTIM] |
03:19.36 | *** join/#htc-linux drno (n=drno@eki.student.utwente.nl) [NETSPLIT VICTIM] |
03:20.03 | *** join/#htc-linux hlbot (n=adm@iclem.net) [NETSPLIT VICTIM] |
03:20.04 | *** join/#htc-linux KimuSan^ (n=kimusan@ip205.tankegang.dk) [NETSPLIT VICTIM] |
03:41.03 | *** join/#htc-linux apt (i=ibot@pdpc/supporter/active/TimRiker/bot/apt) |
03:41.03 | *** 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 | <cr2> let's define a common setup. |
04:50.02 | *** join/#htc-linux scorpio16v (n=scorpio1@ACB39B4B.ipt.aol.com) |
05:34.28 | *** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com) |
07:37.10 | *** join/#htc-linux RoEn_PC (n=roen@p54A65E6B.dip.t-dialin.net) |
07:42.24 | *** join/#htc-linux rmoravcik (n=rmoravci@adsl-d181.84-47-56.t-com.sk) |
07:57.53 | *** part/#htc-linux scorpio16v (n=scorpio1@ACB39B4B.ipt.aol.com) |
08:06.24 | *** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be) |
08:11.27 | *** join/#htc-linux the_sys0p (n=the_sys0@cpe-66-75-196-173.bak.res.rr.com) |
08:16.58 | *** join/#htc-linux ashkan (n=ashkan@unaffiliated/askywalker/x-0000002) |
08:30.09 | *** join/#htc-linux freshr (n=freshr@89.113.75.13) |
08:46.34 | *** join/#htc-linux lilliput (n=Touriste@82-47-146-102.cable.ubr11.brad.blueyonder.co.uk) |
08:58.37 | *** join/#htc-linux scorpio16v (n=scorpio1@ACB39B4B.ipt.aol.com) |
09:02.53 | *** join/#htc-linux psokolovsky (n=psokolov@82.193.98.15) |
09:10.18 | *** join/#htc-linux tsdogs (n=twostupi@84.18.144.39) |
09:13.09 | tsdogs | morning all |
09:27.13 | goxboxlive | hi tsdogs |
09:27.19 | goxboxlive | what is this: <5>ds2760-battery ds2760-battery.0: not enough power to charge |
09:27.37 | goxboxlive | Could this be the reason qtopia wouldnt launch? |
09:27.38 | tsdogs | it means that the battery is not charging... |
09:27.52 | tsdogs | I don't think so |
09:27.58 | goxboxlive | I tried oyt your opie2 yesterday and it launched. |
09:28.21 | tsdogs | You get this when youdiscconect the cable |
09:28.33 | goxboxlive | I know something with my battery is screwed up. Because the battery indicator doesnt work with WInCE either |
09:28.37 | tsdogs | or if there is not enough power to charge. |
09:28.45 | goxboxlive | even though the battery is charged. |
09:28.59 | tsdogs | have you tryed changing the USB cable? |
09:29.30 | goxboxlive | no actually i havent. I dont know where i have the original usb cable. |
09:30.12 | tsdogs | It might be just a guess. but I had this with a thin usb cable too. |
09:30.15 | goxboxlive | Now i have rebuilded qtopia-4.2.0 with a new gcc compiler. And now even that version wouldnt launch |
09:30.30 | tsdogs | It's some config. |
09:30.33 | goxboxlive | ok, my USB cable is thin :-) It's a cheap short one |
09:30.37 | goxboxlive | yes same config |
09:30.52 | tsdogs | Have you tryed removing the root/Settings ? |
09:31.17 | goxboxlive | And to launch your Opie2 image used my own kernel sImage and modules and also startup script. |
09:31.40 | goxboxlive | yes, the image i am trying to launch now is a new one, i builded it today |
09:31.52 | tsdogs | ok |
09:32.30 | tsdogs | you tryed my opieII also right? it does not start either ? |
09:33.19 | tsdogs | http://84.18.144.44/tsdogs/opieII/opieII-tsdogs-20070514.tar.bz2 ?? |
09:33.46 | goxboxlive | tsdogs: Yes your starts |
09:34.02 | goxboxlive | I think this might be something with either helox, alsa or gnokii |
09:34.04 | tsdogs | ok, so it must be something in the new image I guess. |
09:34.13 | tsdogs | you built helix ? |
09:35.47 | tsdogs | then I think it's helix |
09:35.48 | goxboxlive | yes |
09:35.56 | goxboxlive | maybe |
09:36.06 | tsdogs | gnokii does initilize right? |
09:36.10 | goxboxlive | But BabelO have had it running with helix |
09:36.18 | tsdogs | you still using the old version of BabelO ? |
09:36.21 | goxboxlive | I think it does |
09:36.25 | goxboxlive | i cant see |
09:36.37 | tsdogs | gnokii I mean |
09:36.46 | tsdogs | does it start at startup ? |
09:36.58 | tsdogs | still talking about gnokii... |
09:38.05 | tsdogs | goxboxlive: as far as I remember he (babelO) had some CPU issue with mediaserver |
09:43.31 | goxboxlive | ok |
09:45.20 | goxboxlive | I'm not sure if gnokii is starting. It doesnt spam the output as i did before. |
09:45.20 | goxboxlive | hmm strange. Well i have to go now. bbl |
09:46.04 | goxboxlive | root@htcuniversal:/opt/Qtopia/bin$ ldd qpe |
09:46.05 | goxboxlive | <PROTECTED> |
09:46.23 | goxboxlive | so it finds gnokii lib |
09:46.49 | tsdogs | ok |
09:54.45 | *** join/#htc-linux pH5 (n=ph5@p5485CAB3.dip.t-dialin.net) |
10:02.59 | goxboxlive | Well it starts gnokii |
10:03.38 | goxboxlive | and it hangs here: Unable to open '/usr/share/zoneinfo/America/New_York' |
10:03.38 | goxboxlive | QTimeZone::data Can't create a valid data object for 'America/New_York' |
10:05.50 | goxboxlive | I tried to just start a single application and it hangs with this. It runs it in loops: |
10:05.53 | goxboxlive | gettimeofday({864040167, 397569}, NULL) = 0 |
10:05.53 | goxboxlive | select(9, [8], [], NULL, {0, 0}) = 0 (Timeout) |
10:06.28 | goxboxlive | Because i can see that gnokii is running as it shall. |
10:07.04 | tsdogs | ok |
10:07.29 | tsdogs | we need some more Logging then |
10:07.39 | goxboxlive | http://pastebin.no/1656 |
10:08.27 | goxboxlive | it's all the same the whole way. The thing i have edited since last time is custom.cpp and htcphonehardware.cpp. I am using the latest ones at sf.net (xanadux) |
10:08.27 | tsdogs | copy the Log.conf from /opt/Qtopia/etc/default/Trolltech into Settings/Trolltech |
10:09.31 | tsdogs | I believe BabelO removed the gnokii part from there... |
10:09.37 | tsdogs | custom.cpp |
10:09.38 | tsdogs | but ok |
10:10.04 | tsdogs | change a few settings in the log, and let's see where it hangs... |
10:10.32 | tsdogs | You have BT and Irda built in right ? |
10:11.09 | goxboxlive | yes |
10:11.35 | goxboxlive | and then copy it back to /opt/Qtopia/etc/default/Trolltech ? |
10:11.44 | tsdogs | ok then enable BT logging |
10:12.00 | tsdogs | no simply put it in local conf /home/root/Settings/Trolltech |
10:12.10 | tsdogs | it will override default |
10:12.20 | goxboxlive | ok |
10:12.31 | tsdogs | DBUS ... |
10:12.47 | tsdogs | Hardware / Infrared |
10:12.53 | goxboxlive | ok, i'll enable DBUS |
10:13.14 | tsdogs | and finnally QtopiaServer |
10:13.18 | tsdogs | that's all for now. |
10:13.31 | tsdogs | then start qpe and let's see where it hangs |
10:17.56 | goxboxlive | http://pastebin.no/1657 seems like it's looping true this |
10:19.49 | goxboxlive | gnokii is running no doubt about that. |
10:20.55 | goxboxlive | but anyway, i have to get going now. I'm allready late |
10:21.31 | tsdogs | ok, cu, I thinkg there is something strage here... |
10:22.15 | goxboxlive | Anyway, this is the last logging after i stoped qpe: http://pastebin.no/1658 |
10:22.22 | goxboxlive | bbl |
10:24.02 | tsdogs | ok, you changed the Storage right? |
10:26.04 | goxboxlive | yes, i changed it in htcphone before building |
10:26.06 | goxboxlive | bbl |
10:55.40 | *** join/#htc-linux asylumed (n=insanity@dsl-245-191-214.telkomadsl.co.za) |
11:04.26 | *** part/#htc-linux asylumed (n=insanity@dsl-245-191-214.telkomadsl.co.za) |
12:04.54 | *** join/#htc-linux SmallR2002 (n=SmallR20@88-107-107-83.dynamic.dsl.as9105.com) |
12:18.24 | *** join/#htc-linux TripleDE1 (n=sergio@151.pool85-50-250.dynamic.orange.es) |
12:24.14 | *** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
12:32.54 | *** join/#htc-linux TripleDES (n=sergio@10.pool85-50-145.dynamic.orange.es) |
12:36.49 | *** join/#htc-linux WizMaui (n=WizMaui@62.112.90.250) |
12:54.27 | tsdogs | BabelO: you here? |
12:54.44 | BabelO | hi tsdogs ;) |
12:54.51 | tsdogs | Hi BabelO |
12:55.02 | tsdogs | I sent you a patch for htckbdhandler... |
12:55.17 | tsdogs | kernelkeymap does not work :( |
12:55.33 | tsdogs | rephrease: I have not been able to make it work till now. |
12:55.52 | BabelO | i don"t understan how it is working :( |
12:56.08 | tsdogs | so I decided to have at least htckbdhandler see the passed argument in the environment |
12:56.40 | tsdogs | did you know, that qtopia is not passing the argument after the : |
12:56.53 | tsdogs | only to internal handlers. |
12:57.44 | tsdogs | you'll probably get it in the spam folder (still have to find out why gmail considers me spam :/) |
12:58.29 | BabelO | yes, spam folder at gmail |
12:58.42 | tsdogs | :) |
13:00.11 | BabelO | got it |
13:00.15 | tsdogs | ok |
13:00.36 | tsdogs | have a look, and testit on BA, so then we at least have something working... |
13:00.44 | BabelO | i spent night until 4:00am to have all my test running in a single exe :( |
13:01.04 | *** join/#htc-linux lilliput (n=Touriste@82-47-146-102.cable.ubr11.brad.blueyonder.co.uk) |
13:01.06 | tsdogs | gosh :( |
13:01.39 | BabelO | doing a svg image for qtopia is really strange too ;) |
13:02.07 | tsdogs | didn't the one you made work? |
13:03.33 | BabelO | they duplicate picture with a step and then move a Qrect over |
13:04.41 | tsdogs | I think this is pretty normal |
13:04.46 | *** join/#htc-linux TripleDE1 (n=sergio@11.pool85-50-254.dynamic.orange.es) |
13:05.22 | BabelO | did you have a switch event to detect headphone on universal ? |
13:05.49 | tsdogs | hmm, don't think so, let me check. |
13:06.38 | BabelO | ok, keyboard compiled |
13:07.16 | tsdogs | No I don't have a SW. |
13:07.54 | BabelO | maybe it is usefull to have switch for this king of entry |
13:08.25 | BabelO | so we can get noticed about event and don't poll for it |
13:08.54 | tsdogs | why are you polling? |
13:09.07 | tsdogs | should not the mixer handle it |
13:09.35 | tsdogs | I have bluetooth to, so I could have a BT headphone too |
13:10.17 | BabelO | tsdogs: ok,don't know if mixer handle it, we can drive with mixer sure |
13:11.04 | tsdogs | Uni, does not have a mature mixer right now. also sound support ain't the best |
13:12.13 | BabelO | ok trying plugin ;) |
13:13.29 | tsdogs | :) I think I have an event for sound plug in, I just tryed and snd_htcuniversal_audio_set_headphone_detect:1 jumped out |
13:15.27 | BabelO | ok, but how can it notice me ? |
13:15.57 | BabelO | touchscreen is working for you ? |
13:16.15 | tsdogs | yes. |
13:16.46 | tsdogs | you putted the string htckbdhandler:/dev/input/eventX right? |
13:16.59 | tsdogs | otherwise it defaults to event1 (?) |
13:17.21 | tsdogs | we should also find out a way to detect through the sysfs |
13:17.27 | BabelO | tsdogs: need to do the same for button plugin |
13:18.00 | tsdogs | buttons are a bit more coplicated as you have 2 events for it |
13:18.26 | tsdogs | I only need 1 I think |
13:19.08 | BabelO | ok, but maybe can load plugin two times with different param |
13:19.46 | tsdogs | It should work... |
13:20.15 | tsdogs | but we should know what the mapping is... |
13:20.22 | tsdogs | for the specified device. |
13:20.39 | *** part/#htc-linux scorpio16v (n=scorpio1@ACB39B4B.ipt.aol.com) |
13:20.56 | tsdogs | so the entry would look like something htcbuthandler:/dev/input/event2:GPIO |
13:20.58 | BabelO | where sensor is mapped on universal ? |
13:21.12 | tsdogs | for headphone you mean? |
13:21.24 | BabelO | i think param are separated with coma |
13:21.39 | BabelO | no for keyboard backlight |
13:21.49 | tsdogs | ok wait |
13:22.32 | BabelO | good it wokrs ;) |
13:22.43 | tsdogs | keyboard right? |
13:23.06 | BabelO | qwerty but wokrs ;) |
13:23.37 | tsdogs | <PROTECTED> |
13:23.59 | tsdogs | I think we could put the mapping in some file too. |
13:24.19 | BabelO | thx, i ll comit plugin |
13:24.24 | tsdogs | ok |
13:24.42 | BabelO | so sensor is not a SW event ? |
13:25.23 | tsdogs | I have the hinge sensor is that's what you mean. |
13:25.30 | tsdogs | ? |
13:25.40 | BabelO | because when a key is pressed we can switch backlight with a timer, but on uni you have a good sensor gadget |
13:26.21 | tsdogs | Ho, they removed the handling, it was causing strange things with touchscreen (if I remember right) |
13:27.13 | tsdogs | it was in event1 |
13:28.46 | tsdogs | activation is on /sys/class/leds/htcuniversal:kbdbac/brightness |
13:29.12 | *** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be) |
13:30.51 | tsdogs | <PROTECTED> |
13:33.15 | tsdogs | BabelO: I gg now, cu tonight. |
13:34.49 | BabelO | ok |
13:39.37 | *** join/#htc-linux TripleDES (n=sergio@224.pool85-50-237.dynamic.orange.es) |
13:45.45 | *** join/#htc-linux rob_w (n=bob@p85.212.39.113.tisdip.tiscali.de) |
13:58.57 | *** join/#htc-linux Ro_En_PC (n=roen@p54a66bd1.dip.t-dialin.net) |
14:01.43 | *** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com) |
14:16.56 | goxboxlive | hi |
14:17.21 | BabelO | hi |
14:18.18 | goxboxlive | BabelO: Does this look familar to you? http://pastebin.no/1658 Do you have any clue whats going on? |
14:19.19 | goxboxlive | The Sigterm at the end is because i ended qpe my self. It is just running in loops with like shown |
14:20.55 | BabelO | i don t know, but good news, i ll remove gnokii from qtopia link binary |
14:21.06 | BabelO | it is an external application |
14:24.00 | goxboxlive | ohh good |
14:24.55 | goxboxlive | But i know it's not gnokii wich causes this, because it is running perfectly in the background. You can see that at the start of my pastebin. |
14:25.26 | goxboxlive | I am using the latest htcphonehardware.cpp at sf.net and tsdogs battery patch. |
14:30.09 | BabelO | did you try to remove section by section in defaultbuttons.conf |
14:32.08 | goxboxlive | BabelO no, i havent tried that. Hold on i try |
14:32.38 | goxboxlive | I didnt know that was an option |
14:35.37 | *** join/#htc-linux scorpio16v (n=scorpio1@ACB39B4B.ipt.aol.com) |
14:39.46 | *** join/#htc-linux TripleDE1 (n=sergio@152.pool85-50-138.dynamic.orange.es) |
14:43.26 | *** join/#htc-linux AlGe (n=alge@chello080109231226.4.uni-klu.teleweb.at) |
14:49.14 | goxboxlive | didnt help that either. I'll try to revert a version of htcphonehardware.cpp |
14:52.14 | *** join/#htc-linux TripleDES (n=sergio@179.pool85-50-155.dynamic.orange.es) |
14:57.39 | *** join/#htc-linux psokolovsky (n=psokolov@82.193.98.15) |
15:00.46 | goxboxlive | BabelO: When i do some changes to custom.cpp, shouldnt something be rebuildt? |
15:01.16 | goxboxlive | It just ran all over and said nothingto be done when i have revrted custom.cpp to previous version |
15:01.17 | BabelO | yes, when you do make, it rebuild qpe |
15:01.45 | goxboxlive | hmm |
15:05.06 | *** join/#htc-linux lilliput (n=Touriste@82-47-146-102.cable.ubr11.brad.blueyonder.co.uk) |
15:20.50 | AlGe | Hi all, |
15:20.50 | AlGe | after some time I want to give my blueangel another chance to switch to a better OS :-) |
15:20.50 | AlGe | Has anyone a pointer to a working htcblueangel defconfig file for the latest hh kernel 2.6.21-hh2? |
15:20.50 | AlGe | I tried a little bit, of course with asic3_mmc included, but I get stuck at the kernel not finding mmcblk0p2 ... |
15:25.14 | BabelO | AlGe: for kernel do a make htcblueangel_defconfig |
15:25.51 | BabelO | and for haret command line it is something like root=/dev/mmcblk0p1 or root=/dev/mmc/blk0/p1 |
15:28.48 | AlGe | Thanks, all try this. |
15:28.48 | AlGe | Another question: what are the mmc partitions numbered, strting with 0 or 1? |
15:29.53 | BabelO | it is partition number like you see under fdisk |
15:31.00 | *** join/#htc-linux rob_w (n=bob@p85.212.57.246.tisdip.tiscali.de) |
15:31.20 | *** join/#htc-linux bd2 (n=cbou@89.113.85.104) |
15:38.19 | *** join/#htc-linux TripleDE1 (n=sergio@232.pool85-50-204.dynamic.orange.es) |
16:21.21 | Foxdie | BabelO: sorry for dissapearing yesterday, I had family troubles |
16:21.34 | BabelO | no problem ;) |
16:25.07 | Foxdie | how are you? |
16:25.43 | Foxdie | I am still away today by the way, I apologise, my partner is moving into my home and we've got to tidy the place up, you know how messy geeks live ;) |
16:25.46 | BabelO | fine, going to supermarket soon |
16:27.59 | *** join/#htc-linux psokolovsky (n=psokolov@82.193.98.15) |
16:58.40 | Kevin2 | bd2: Are you around? |
17:05.56 | cr2 | hi Kevin2. i've finally reached my computer. |
17:06.10 | Kevin2 | Hi cr2. |
17:07.00 | *** join/#htc-linux TripleDES (n=sergio@22.pool85-50-202.dynamic.orange.es) |
17:13.56 | AlGe | now with a defconfig generated by make htcblueangel_defconfig I still get kernel panic saying "cannot open root dev "mmcblk0p2" |
17:13.56 | AlGe | the last line before the panic is "asic3_mmc: ASIC3 MMC/SD Driver, controller at 0xe000000". |
17:13.56 | AlGe | shouldn't then a line follow which detects the partitions on mmcblk0? |
17:15.08 | cr2 | AlGe: yes. do you have mmc_blk driver compiled in ? |
17:15.37 | cr2 | CONFIG_MMC_BLOCK=y |
17:15.41 | cr2 | hm. |
17:16.01 | AlGe | I ll have a look, wait ... |
17:16.02 | cr2 | filesystems ? |
17:16.22 | cr2 | EXT2 yes EXT3 yes. |
17:16.28 | Kevin2 | cr2: Did you see my note last night about sd? It looks like most of the work is done in sdhc_sc2410.dll |
17:16.28 | cr2 | it should work ;) |
17:16.43 | cr2 | Kevin2: yes, i'll look at sdhc_sc2410.dll |
17:17.26 | cr2 | i knew that this driver supports both SD controllers, but didn't know that raw register access is also ther.e |
17:18.08 | cr2 | AlGe: try to add 'ro' option to the CMDLINE, and look if it brings your further. |
17:19.03 | BabelO | cr2: maybe it is also rootdelay ? |
17:19.11 | AlGe | CONFIG_MMC_BLOCK=y, CONFIG_EXT2_FS=y and CONFIG_EXT3_FS=y are there |
17:19.55 | cr2 | AlGe: do you have rootdelay= in the CMDLINE ? |
17:20.18 | cr2 | BabelO: we should edit the old BA wiki at hh.org |
17:20.38 | AlGe | no, no rootdelay |
17:20.48 | cr2 | add it. that's the problem. |
17:21.20 | BabelO | AlGe: my command line look like this root=/dev/mmcblk0p2 console=tty0 mem=128M rootdelay=5 |
17:21.22 | AlGe | ok, I'm just wading thru the wm reinstall ... |
17:25.22 | BabelO | cr2: yes maybe having same look as universal |
17:29.30 | bd2 | Kevin2, yup |
17:29.53 | Kevin2 | Hi bd2. I had some questions on the adc stuff. |
17:30.25 | Kevin2 | In ad7877.c don't we need to reinit the completion structure every sense? |
17:31.47 | bd2 | Kevin2, well.. not sure, need to look into completion code, or find some example. It does not work? |
17:32.02 | Kevin2 | I haven't tried it yet. I was reviewing the code. |
17:33.25 | Kevin2 | A couple other questions I had -- the h4000 and the custom pins thing looks really odd. Wouldn't it be simpler to just define an adc driver for h4000 that just wraps around the base ad7846 call? |
17:34.13 | Kevin2 | (BTW, everything looks good in general - it really cleans things up - thanks for doing this work.) |
17:36.09 | bd2 | Kevin2, h4000 custom pins not looks really odd, but also works odd (not wrong, but complicated ;-). Anyway, another ADC refactoring step is planned by Paul and me, for LKML submission. Maybe at that time we'll find more cleaner way |
17:36.45 | bd2 | feel free to remind your ideas when you'll see ADC RFCs |
17:37.00 | Kevin2 | Okay. I also think we should remove pin_id, and instead just keep a pointer to adc_classdev in adc_sense. |
17:37.06 | psokolovsky | hi! |
17:37.15 | psokolovsky | what exactly looks "odd"? |
17:37.30 | psokolovsky | that's just how it is, that's all. nothing odd, really trivial |
17:37.42 | psokolovsky | don't have them - don't use them ;-) |
17:37.54 | Kevin2 | psokolovsky: Wouldn't it be simpler to just define an adc driver for h4000 that just wraps around the base ad7846 call? |
17:38.20 | psokolovsky | Kevin2: defining another driver is not simpler, apparently |
17:38.33 | rmoravcik | hi psokolovsky! |
17:38.55 | rmoravcik | psokolovsky: did you get my mail? |
17:39.18 | psokolovsky | also, it seems that you think it's h4000 thing. no. it's used in h1910 too, and I possibly in other devices too |
17:39.24 | bd2 | Kevin2, um... I think we can remove pin_id (i.e. wrap adc_classdev into adc driver specific struct, and do container_of, to get pin id), but... most drivers we have using pin_id, so we'll not gain much from it |
17:39.51 | psokolovsky | rmoravcik: Hi. Saw your mail, great work. don't know when i'll have time to look into it though ;-( |
17:41.15 | Kevin2 | bd2: Most drivers don't need pin_id. They can just do acdev - acdevs[]. |
17:41.29 | bd2 | (regarding hx4700 flash problem) quiz: does ioremap *really* returns virtual address? I've looked into docs, and they say "yes". But virt_addr_valid() macro returns 0 for ioremaped PXA CS region! and that's why we getting oops |
17:41.56 | Kevin2 | Also, both tsc2200 and ads7846 don't want pin_id at all - they want the TSC2200_DATAREG_BAT1 type thing. |
17:42.07 | bd2 | Kevin2, yes, true. It is optimization. But every optimization have drawback: code becomes non obvious. :-( |
17:42.49 | bd2 | Kevin2, that's another argument, strong one. If we really have few drivers don'w wanting pin_id, we should remove it from generic api, indeed |
17:43.01 | Kevin2 | You may be right. I'd still prefer to have adc_classdev * in adc_sense -- even if it means callers do sense->adc->pin_id. |
17:44.48 | Kevin2 | Well 2.6.21 boots on apache, but touching the screen causes kernel panic. |
17:45.08 | bd2 | apache is ad7877 user? |
17:45.19 | Kevin2 | bd2: Yes. I'll look into it. |
17:50.47 | Kevin2 | Heh, my sd card works again. :-) |
17:55.53 | bd2 | congrats ;-) |
17:56.04 | bd2 | psokolovsky, any idea on ioremap q? |
17:56.29 | psokolovsky | bd2: sorry, no |
18:01.05 | cr2 | bd2: ioremap returns the virtual address. it's used all over the code. |
18:03.06 | bd2 | cr2, no matter what phys addr given? and what means if virt_addr_valid(ioremaped_addr)) returns 0? |
18:03.31 | bd2 | ioremapped_addr != 0, i.e. ioremap() done w/o errors |
18:11.27 | Kevin2 | bd2: s/INIT_COMPLETION(/init_completion(&/ |
18:11.34 | Kevin2 | And it works great! |
18:12.16 | bd2 | Kevin2, wow, great! ;-) |
18:12.32 | Kevin2 | It even fixes my sd card! :-) |
18:12.39 | cr2 | bd2: what was the value given to virt_addr_valid ? |
18:13.11 | bd2 | cr2, PXA_CS0_PHYS |
18:13.59 | bd2 | oh no.. it is virt_addr_valid(ioremap_cached(PXA_CS0_PHYS)) |
18:14.24 | bd2 | actually, _cached or _nocache does not matter :-/ |
18:15.03 | cr2 | well, but was what the result of ioremap_cached(PXA_CS0_PHYS) ? |
18:16.34 | bd2 | 0xd5000000 |
18:16.35 | cr2 | i guess we don't have discontig_mem |
18:16.37 | cr2 | #define virt_addr_valid(kaddr) ((unsigned long)(kaddr) >= PAGE_OFFSET && (unsigned long)(kaddr) < (unsigned long)high_mem |
18:17.24 | cr2 | #define PAGE_OFFSET UL(0xc0000000) |
18:17.42 | cr2 | high_mem ? |
18:17.59 | bd2 | high_mem? |
18:18.07 | cr2 | how big is that. |
18:18.27 | bd2 | PXA_CS_SIZE.. let me find number.. |
18:18.51 | bd2 | 0x04000000 |
18:19.20 | cr2 | then you will get zero as a result... |
18:19.50 | bd2 | hmmm |
18:20.02 | *** join/#htc-linux tsdogs (n=twostupi@62.123.180.130) |
18:20.26 | bd2 | why, and then all pxa drivers in drivers/mtd/ |
18:20.30 | bd2 | are wrong |
18:21.07 | bd2 | for example mainstone-flash.c, it's pretty close to hx4700-flash.c |
18:22.19 | cr2 | ./arch/arm/mm/consistent.c: BUG_ON(!virt_addr_valid(start) || !virt_addr_valid(end - 1)); |
18:22.26 | bd2 | cr2, fck. <cr2> how big is that. - I misunderstood it. I told you "how big" is remapped (CS) area, not high mem |
18:22.43 | cr2 | ./mm/page_alloc.c: VM_BUG_ON(!virt_addr_valid((void *)addr)); |
18:22.44 | bd2 | cr2, exactly. if I remove that BUG_ON then flash works |
18:23.01 | cr2 | ./mm/slab.c: if (!virt_addr_valid(objp)) { |
18:23.12 | bd2 | that BUG_ON appeard in 2.6.21 |
18:23.27 | cr2 | i don't see any other references to virt_addr_valid ?? |
18:24.45 | cr2 | i think this code makes sense only for RAM allocations. |
18:27.27 | bd2 | heh.. so, should I post q to LAKML regarding that? |
18:30.32 | *** join/#htc-linux TripleDE1 (n=sergio@94.pool85-50-141.dynamic.orange.es) |
18:31.42 | BabelO | tsdogs: htcbuthandler is ok ;) |
18:33.55 | *** part/#htc-linux scorpio16v (n=scorpio1@ACB39B4B.ipt.aol.com) |
18:33.55 | tsdogs | perfect |
18:34.33 | BabelO | now we can add many event to button driver ;) |
18:34.45 | BabelO | i fix a latest bug and it is ok |
18:35.14 | BabelO | we can also use GPIO0 to drive backlight on/off or power on/off |
18:35.16 | tsdogs | :) nice work |
18:36.55 | tsdogs | did you also merge the 2 asic and gpio readKey funcionts? |
18:37.14 | BabelO | tsdogs: no :( |
18:37.41 | BabelO | maybe we can ;) |
18:38.33 | tsdogs | I think so, they only handle different keycodes, but on universal I think that key mappigs are all on event2 (except for power= |
18:38.44 | tsdogs | s/=/)/ |
18:39.55 | AlGe | BabelO/cr2: thanks for the rootdelay hint, I had to enlarge it a bit, but now it boots. Thanks. |
18:39.55 | AlGe | now the next problem seems to be the touchscreen, it deosnt't responde so I'm stuck at the inital gpe-new-user screen. |
18:39.55 | AlGe | I guess I should now try to login vie usb/g_ether to fix it? |
18:40.07 | BabelO | another thing, have an issue with battery now |
18:40.38 | tsdogs | BabelO: what issue? |
18:41.08 | BabelO | AlGe: yes log over ssh and then modprobe -r ts-adc-debounce |
18:41.27 | BabelO | and after modprobe tsc2200_adc_ssp; modprobe ts-adc-debounce |
18:41.47 | BabelO | it seems it should be a problem with loading order |
18:42.06 | BabelO | tsdogs: battery status does not display :( |
18:42.14 | tsdogs | We should take over that code and replace it with ours in htcphonehardware (functions are already there, just need to be hooked) |
18:42.41 | tsdogs | what's the value of Charging ? |
18:43.25 | BabelO | 0 and false ... that's problem, i trace it when finished button |
18:44.03 | tsdogs | hmm, was working before the patch right? |
18:44.10 | BabelO | right |
18:44.51 | tsdogs | that's strange. we only added the full thing ... |
18:45.16 | tsdogs | I'll look into the code and see what happened. Got to put baby to sleep bbl. |
18:48.59 | tsdogs | BabelO: btw can you try reverting the patch and see if it works? |
18:49.20 | BabelO | no it works for you ? then i ll look |
18:52.16 | tsdogs | It didn't but with the patch it is.... |
19:05.17 | BabelO | tsdogs: ok i ll comit your patch and my button handler |
19:05.33 | BabelO | merge it later |
19:07.55 | *** join/#htc-linux goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com) |
19:12.08 | BabelO | tsdogs: commited, can you check ? |
19:13.18 | tsdogs | ok checkedout |
19:19.18 | AlGe | isn't 192.168.0.206 the ip address to connect to via usb? I can not connect to it. |
19:19.19 | AlGe | where is it set? maybe I need to configure it first somewhere in my rootfs on sdcard? |
19:19.31 | goxboxlive | 192.168.0.202 |
19:19.59 | goxboxlive | '/etc/network/interfaces |
19:26.09 | tsdogs | BabelO: htcbuthandler.cpp:68: error: expected primary-expression before ‘<<’ token |
19:26.13 | tsdogs | ?? |
19:26.45 | BabelO | you do a cvs up ? |
19:27.13 | BabelO | check into file you have maybe <<<< 1.2, something like this |
19:27.35 | *** join/#htc-linux psokolovsky (n=psokolov@82.193.98.15) |
19:27.40 | tsdogs | cheking ... |
19:28.57 | goxboxlive | total 7 files right? |
19:29.09 | BabelO | goxboxlive: yes |
19:29.16 | goxboxlive | ok, i'll compile |
19:30.58 | goxboxlive | yes it compiled, just to lame i cant test it though :-) |
19:34.51 | tsdogs | BabelO: syntax is: htcbuthandler:/dev,asic3 ? |
19:35.23 | BabelO | tsdogs: yes ASIC3 ... i ve not set it to lower case , sorry |
19:36.19 | tsdogs | this should do right? htcbuthandler:/dev/input/event2,ASIC3:/dev/input/event1,GPIO |
19:36.37 | BabelO | tsdogs: yes |
19:37.04 | BabelO | i m merging event to one, it is a test ;) |
19:37.47 | tsdogs | :) |
19:38.42 | tsdogs | it's looping somewhere, as it does not respond anymore :/ |
19:39.31 | goxboxlive | tsdogs: What? Same as mine? |
19:40.10 | tsdogs | goxboxlive: nope, didn't even print missing locale (could not even login again...) |
19:40.36 | goxboxlive | ok |
19:42.56 | *** join/#htc-linux TripleDES (n=sergio@5.pool85-50-135.dynamic.orange.es) |
19:42.57 | tsdogs | BabelO: didn't you remove gnokii init from opie init ? |
19:43.16 | tsdogs | Initialization failed (11) |
19:43.16 | tsdogs | Telephone interface init failed: Command timed out. |
19:43.16 | tsdogs | Quitting. |
19:43.21 | BabelO | not at the moment, if you want just omit define at top |
19:44.40 | tsdogs | no define anymore here ... (?) |
19:45.07 | BabelO | in htcphonehardware ? |
19:45.19 | tsdogs | yep |
19:46.12 | tsdogs | I'll take older one with define |
19:47.34 | goxboxlive | tsdogd: Are you using the cvs version of htcphonehardware ? It hasent been changed for 4 months? Is that correct? |
19:47.57 | BabelO | not correct ;) |
19:48.01 | tsdogs | using cvs yes. |
19:48.17 | tsdogs | ho right, you sent it to me, not in cvs yet right ? |
19:48.24 | BabelO | tsdogs: i ll comit mine instead |
19:48.31 | tsdogs | okstarted anyway |
19:49.37 | tsdogs | goxboxlive: with new kernel, lcd is flashing now |
19:49.44 | tsdogs | ok needs more testing but it is working |
19:50.00 | tsdogs | also battery... |
19:51.32 | goxboxlive | tsdogs: I noticed when i tried out your Opie2 version |
19:51.37 | tsdogs | BabelO: I have the feeling I changes something else to make battery work |
19:52.08 | tsdogs | with -hh0 it wasn't. I think it is the HTCUNIVERSAL_LCD |
19:52.35 | tsdogs | in -hh0 could not be selected... |
19:53.00 | tsdogs | or maybe it is conflicting with the TS2 |
19:53.16 | tsdogs | (which is marked as old) (?) |
19:53.21 | goxboxlive | BabelO: what do i have to do to start with gnokii? |
19:53.52 | tsdogs | got to eat now, bbl |
19:55.04 | BabelO | goxboxlive: i remove it, i m doing a standalone app |
19:55.23 | AlGe | now I'm trying with (laptop) 192.168.0.200 <-> 192.168.0.202 (ba), still no luck. |
19:55.23 | AlGe | I get "NETDEV WATCHDOG: usb0: transmit timed out" messages on the laptop(2.6.15/dapper) side ... wasn' there an issue whith the cdc_ether module, do I need to blacklist something? |
19:56.22 | goxboxlive | AlGe: have you did: ifup usb0 on your handheld? |
19:56.40 | goxboxlive | BabelO ok |
19:58.37 | AlGe | how should I do this? I'm stuck in the gpe-new-user dialog without working touchscreen or keyboard. |
19:58.38 | AlGe | does it mean that I should add an init script which just ifups usb0? |
19:59.41 | BabelO | AlGe: yes, if when you plug BA on cradle you didn't see usbnet line in dmesg |
20:03.31 | AlGe | if I plug it into my laptop the usbnet module gets loaded, (and the g_ether stuff on the ba side shows some initializiation message very early), anyway I'll add the ifup script ... |
20:07.05 | BabelO | tsdogs: maybe you have to try this key "Key_Flip" |
20:07.33 | BabelO | http://doc.trolltech.com/qtopia4.2/syscust-keypadarch.html |
20:12.59 | goxboxlive | AlGe or you can just add 'auto usb0' to your interfaces file |
20:13.21 | goxboxlive | Nope, the new files didnt help either. |
20:15.03 | AlGe | so it works the standard debian way (ifup...) |
20:15.06 | goxboxlive | it's actually starting to p*ss me off. |
20:15.11 | goxboxlive | yes |
20:21.17 | goxboxlive | [pid 1254] ioctl(12, KDGKBENT, 0xbed7647c) = -1 EINVAL (Invalid argument) |
20:22.23 | tsdogs | goxboxlive: try setting the QWS_KEYBOARD=None |
20:22.31 | goxboxlive | ok |
20:23.42 | tsdogs | BabelO: yes I'll need those, like in opie and also the hinge thing... |
20:24.54 | tsdogs | though I don't like the answer/hangup call ... |
20:25.22 | tsdogs | I don't want it to answer when I just want to check who is calling... |
20:25.23 | tsdogs | :) |
20:26.15 | goxboxlive | it didnt help. It's the same as happend before today |
20:26.42 | tsdogs | hmm, it's Qtopia 4.2.0 or 4.2.2 ? |
20:26.55 | goxboxlive | I am using qtopia-4.2.0 now |
20:27.24 | tsdogs | Let me check the source code, I saw that ioctl code somewhere ... |
20:28.02 | BabelO | tsdogs: can you check out again button, test and add your missing buttons |
20:28.20 | tsdogs | BabelO: ok |
20:29.57 | tsdogs | BabelO: didint' update anything |
20:30.21 | BabelO | tsdogs: wait a little maybe |
20:33.02 | goxboxlive | it compiled |
20:35.40 | AlGe | finally I was able to ssh into my ba: there was no ts_adc-debounce module installed, I have to add kernel-module-ts-adc-debounce to my machine config file (and maybe some more modules)... thanks |
20:36.54 | goxboxlive | tsdogs: Send a mail to martin-s and ask himif you can join xanadux. |
20:37.51 | goxboxlive | there can be a lag for about 30 min for users to get the latest files. As a developer you will have access to the at one |
20:41.46 | tsdogs | BabelO: buttons do work. I'll add what it is missing, though I have the call/hangup on the keyboard not on buttons... |
20:41.55 | tsdogs | goxboxlive: ok I'll do it. |
20:43.29 | BabelO | tsdogs: but i think it works with my last commit, it call same funciton with all keycode |
20:44.36 | BabelO | ok i ll play qasteroids, it works ;) |
20:44.52 | tsdogs | :) |
20:45.02 | BabelO | tsdogs: ah ok, i understand only now, on keyboard ;) |
20:45.13 | tsdogs | yes on different event. |
20:45.56 | tsdogs | and handler. But I still think we should sort out the tty0 thing, and only have things like sensors in here... :) |
20:48.09 | tsdogs | goxboxlive: it is something related to the keyboard, but haven't found where it is called yet |
20:49.13 | goxboxlive | ? what do you mean? |
20:49.28 | tsdogs | found it: kernelkeymapkbdhandler.cpp: if (ioctl(kbdFD, KDGKBENT, &kbe) != 0) |
20:51.04 | tsdogs | goxboxlive: you sure you are not setting somewhere to use the kernelkeymapkbdhandler ? |
20:51.24 | goxboxlive | hmm i dont have any itemcontrol in qtopie-4.2.0 source |
20:52.16 | tsdogs | it's only for appearence, I think it was added in 4.2.2 |
20:52.23 | goxboxlive | I wasnt when you said i should use =none. I quited and loged in again . Also deleted Settings |
20:52.34 | tsdogs | ok |
20:52.59 | goxboxlive | is there anything in custom.cpp wich depend on itemcontrol? |
20:53.08 | goxboxlive | or anywhere else? |
20:53.29 | goxboxlive | Maybe that is why it hangs and loops anround the battery stuff. |
20:53.32 | tsdogs | don't think so, it's only apearence app dependency (I think) |
20:53.38 | goxboxlive | ok |
20:54.01 | tsdogs | can you read battery status ? |
20:54.49 | tsdogs | though I think that that "loop" is normal, as it is cheking every now and then to update the status .. |
20:54.56 | cr2 | Kevin2: are you here ? |
20:55.54 | tsdogs | goxboxlive: just to have a try, can u rename the libkernelkeymaphandler.so in opt/Qtopia/qt_plugins/kbddrivers ? |
20:56.15 | goxboxlive | tsdogs: Yes i can read the battery status |
20:56.21 | tsdogs | ok |
20:56.56 | goxboxlive | tsdogs: I renamed all of them yesterday, but it still wouldnt launch |
20:57.19 | goxboxlive | I'll try out opie2 now (cvs ) |
20:57.19 | tsdogs | ok |
20:57.32 | tsdogs | ok. |
20:58.41 | goxboxlive | cr2: Do you think AThena wiill have better support with linux than Uni? |
20:58.53 | cr2 | g_pATIHWContext * inside sdhc. +0x37c offset +0x24 . here should be the base address of SD controller on hermes. |
20:58.56 | *** join/#htc-linux TripleDE1 (n=sergio@120.pool85-50-201.dynamic.orange.es) |
20:59.04 | cr2 | goxboxlive: of course not. |
20:59.37 | cr2 | the main problem right now is the disk. |
20:59.39 | goxboxlive | cr2: Why is that of course? |
20:59.44 | goxboxlive | ok |
20:59.58 | goxboxlive | Have you been looing at the cam yet? |
20:59.58 | cr2 | camera will not be supported in any near future. |
21:00.04 | goxboxlive | right |
21:00.14 | cr2 | no, i'm looking at hermes code now. |
21:00.22 | goxboxlive | ok, how is it going? |
21:00.27 | cr2 | camera on universal is a technicality. |
21:00.49 | cr2 | it's the code from a780. plus some pxa gpios that i've already documented. |
21:00.50 | goxboxlive | and by that you mean? difficult to make it work? |
21:01.00 | goxboxlive | okk so the opisite :-) |
21:01.00 | cr2 | which one ? |
21:01.04 | goxboxlive | uni |
21:01.11 | cr2 | should be easy ;) |
21:01.32 | goxboxlive | ok, good to hear. |
21:01.46 | cr2 | +0x2050 |8000 - 1bit path |
21:02.00 | goxboxlive | do you know if there is any cam apps we can use ? other than in Opie2? |
21:02.01 | cr2 | +0x2050 ~8000 - 4bit path |
21:02.31 | cr2 | look into the openezx archive. they have patches and some discussion. |
21:02.56 | goxboxlive | ok |
21:03.14 | goxboxlive | is the FIC phone aplied with cam? |
21:03.39 | cr2 | +0x2068 and +0x2070 ack sdio interrupt. |
21:03.55 | cr2 | FIC phone is _very_ simple ;) |
21:04.26 | cr2 | ~1 and |1 |
21:04.37 | *** join/#htc-linux Cheescake (n=Remmy@82-32-129-75.cable.ubr05.hawk.blueyonder.co.uk) |
21:04.37 | goxboxlive | yes i know, but it has a killer screen i've been told |
21:05.03 | cr2 | i have visually compared it to the universal. |
21:05.20 | goxboxlive | yes and it's better right? |
21:05.36 | cr2 | you can't detect the difference in dpi, but the size is too small. imho. |
21:06.14 | goxboxlive | ok, while we talks about screen. Do you know if there are any settings in the A780 so i can get the backlight a little brighter? |
21:06.15 | cr2 | with a much slower cpu and a builtin framebuffer it's not going to be fast. |
21:06.35 | goxboxlive | hmm, and openmoko is slow on Uni :-) |
21:06.36 | cr2 | unifortunately universal does not have ati. |
21:07.21 | cr2 | athena is great, but very big, worse dpi than universal and €€€ |
21:07.58 | cr2 | but with 128MB RAM and ATI video it should be nice :) |
21:08.31 | goxboxlive | yes it's realy expensive. I read a test of it and they said that it is a killer machine, but it is _BIG_ they said to have it in the innerpocket |
21:08.39 | goxboxlive | yes |
21:08.41 | cr2 | and sirf3 gps. not that abomination on neo. afaik they don't have a gpsd still. |
21:08.52 | goxboxlive | why is the HDD so har d to hack? |
21:08.59 | goxboxlive | isent it usb2 or something? |
21:09.06 | cr2 | mdoc g4 - no driver. |
21:09.23 | cr2 | miniSD - on the ATI chip, the same problem as on hermes. |
21:10.26 | goxboxlive | oh |
21:10.27 | goxboxlive | i c |
21:10.28 | cr2 | the 8GB disk - most probably on the EPSON chip (IDE==CF interface) needs a completely new driver. |
21:10.43 | goxboxlive | ok |
21:11.11 | cr2 | but if the EPSON driver will be written, then you'll have a CF HDD. which is much faster than miniSD: |
21:11.56 | cr2 | don't forget about the i2c accelerometer :) |
21:12.10 | cr2 | and the TV out. |
21:12.19 | Kevin2 | Hi |
21:12.20 | goxboxlive | yes that's riht |
21:12.22 | BabelO | cr2: what this sensor is doing in a phone ? |
21:12.25 | goxboxlive | hi Kevin2 |
21:12.36 | BabelO | hi Kevin2 |
21:13.03 | cr2 | Kevin2: look in the logs. i've written down some obvious ATI SD registers. |
21:13.14 | cr2 | probably should put it into wiki. |
21:13.35 | cr2 | BabelO: some webbrowser "innovation" |
21:14.22 | Kevin2 | Hi all! |
21:14.36 | Kevin2 | cr2: Your notes are above my head. |
21:16.07 | cr2 | Kevin2: have you seen the ATIDBG code ? |
21:16.08 | Kevin2 | I see sd irq acks on 48010200 and 4c010204 |
21:17.05 | Kevin2 | cr2: In gnuharet? |
21:17.05 | cr2 | yes. |
21:17.05 | cr2 | there are 2 undocumented structures. |
21:17.06 | cr2 | Asic and Ctx |
21:17.19 | cr2 | one of then is part of the other. |
21:17.36 | cr2 | there are some hints about their size. |
21:18.39 | cr2 | your tracings say that the SD base address is +0x200000 |
21:19.18 | cr2 | according to the code, this number should be available in the structures. |
21:19.31 | cr2 | i just want to check it. |
21:19.48 | cr2 | the locations are known. |
21:23.35 | cr2 | Kevin2: what is mapped at 0xac* ? |
21:23.46 | cr2 | RAM ? |
21:24.33 | Kevin2 | ac000000 | 30000000 | 1MB section | D=0 AP=1 |
21:24.48 | cr2 | ok. |
21:25.40 | cr2 | 0x9370 and 0x9374 keep some SD status variables. |
21:25.51 | cr2 | you may try to track them. |
21:26.48 | cr2 | AhiInit(0x11) also on hermes. |
21:27.17 | cr2 | AhiDevEnum() |
21:28.37 | cr2 | AhiDevOpen(ptr,ptr,"SD",2); |
21:28.44 | cr2 | not 0, though. |
21:29.25 | cr2 | AhiPwrModeSet(ptr, 3,4,0); |
21:31.44 | cr2 | AhiDevEnum(ptr,ptr,0); |
21:33.09 | cr2 | +0x2038, +0x203c, +0x2040 |
21:33.19 | cr2 | all 16bit. |
21:35.34 | Kevin2 | There is quite a bit of activity to those three 16 bit registers. |
21:35.52 | cr2 | on card remove +0x2038 ~8 |
21:35.58 | Kevin2 | +0x2060 seems to be a 32bit data register |
21:36.54 | cr2 | +0x21c0 is used. |
21:37.15 | cr2 | yes, some of then are ldr/str |
21:38.50 | BabelO | tsdogs: it is better with button ;) |
21:39.16 | cr2 | found SDSetClockRate code. |
21:39.32 | tsdogs | BabelO: :) almost finished . |
21:41.01 | BabelO | tsdogs: good ;) did you set power button too ? |
21:41.17 | BabelO | tsdogs: with latest plugin ? |
21:41.23 | cr2 | AhiPwrModeSet(ptr,3,0,0); on sleep |
21:41.35 | tsdogs | BabelO: why is not there ? |
21:43.24 | BabelO | tsdogs: it is there called PWR_BUTTON |
21:43.40 | cr2 | Kevin2: i think we will be able to locate all the registers. |
21:43.41 | tsdogs | it's perfect the way it is. |
21:43.56 | BabelO | tsdogs: it is not working as they said, missing something |
21:43.56 | cr2 | and what is written there on which occasion. |
21:44.30 | tsdogs | hmm, neither backlight, I think it is calling something in htcphonehardware which is not working ... |
21:44.37 | tsdogs | we need some debugging ... |
21:45.50 | cr2 | +0x2048 |100 - start SD clock |
21:45.50 | BabelO | tsdogs: backlight is ok, it call qpe_brightness |
21:46.12 | cr2 | tst 0x100 - wait for the clock |
21:47.09 | tsdogs | BabelO: I added (the only missing key there) the toggle_backlight, which is not working (like power button). For brightness regulation yes it is working (but not the on/off switch)... |
21:47.22 | cr2 | +0x203c tst 0x4000 - check bus status |
21:48.55 | cr2 | SD_CMD. that's good. |
21:48.59 | tsdogs | BabelO: http://www.pastebin.ca/497493 (also fixed a couple of warnings ...) |
21:49.10 | tsdogs | but it is not working... |
21:51.02 | Kevin2 | cr2: Cool. Will you be able to write a driver with the info? |
21:51.36 | BabelO | ok looking if we can add |
21:55.43 | cr2 | Kevin2: we must, but it's not going to be easy. otherwise hermes is useless. |
21:55.58 | *** part/#htc-linux Cheescake (n=Remmy@82-32-129-75.cable.ubr05.hawk.blueyonder.co.uk) |
21:56.14 | cr2 | unless you'll fix the nand driver, and we can use nand samsung flash :) |
21:56.41 | cr2 | Kevin2: does the current hh.org CVS boot on hermes. i know it compiles cleanly. |
21:57.03 | Kevin2 | I haven't tried yet. I'll run a "cvs up" now. |
21:57.34 | cr2 | ok. |
21:58.32 | cr2 | hmm. it's +0x380 on athena. |
21:58.39 | cr2 | for the struct offset. |
22:02.39 | tsdogs | BabelO: maybe I found out why battery is not correctly displayed! |
22:03.34 | tsdogs | <PROTECTED> |
22:03.58 | tsdogs | <PROTECTED> |
22:03.59 | tsdogs | <PROTECTED> |
22:04.06 | tsdogs | from greenphone ... |
22:13.16 | tsdogs | BabelO: ok I found out why my uni didn't turn off light, I removed the apm -s which was cheking for value <=1 I changed it to < 1 (as don't want it to suspend after few seconds... |
22:16.41 | BabelO | tsdogs: i fix percentRemain |
22:18.54 | cr2 | +offset is the first parameter to AhiDevOpen() |
22:21.07 | cr2 | int (*AhiDevOpen)(struct Ctx**, Asic*, const char* name, int); |
22:21.55 | Kevin2 | cr2: It doesn't look like my irq changes are in hh.org cvs. |
22:22.49 | cr2 | Kevin2: can you make a diff with your current tree ? |
22:23.23 | Kevin2 | with hh.org current tree? |
22:23.48 | cr2 | hmm. Asic comes from Enum. &Ctx from Open. |
22:24.00 | cr2 | we may take a challenge in haret. |
22:24.23 | cr2 | unfortunately the offsets depend on the ATI chip used. |
22:26.58 | cr2 | ok, but ATI does not change the code pattern. that's good. |
22:27.43 | Kevin2 | hh.org head does not boot on hermes. I'll forward port the irq stuff and see if that helps. |
22:28.20 | cr2 | ok. |
22:33.50 | Kevin2 | That didn't work either - we'll need to pull out the irda stuff again. |
22:37.40 | cr2 | http://www.softclub.co.jp/~zoro/sig3/ace_ddi.html |
22:38.03 | cr2 | the systranbox translation is not bad :) |
22:38.39 | BabelO | cr2: systran works with japanese ? |
22:39.31 | BabelO | tsdogs: it is the test about charging which is wrong |
22:39.44 | cr2 | yes. |
22:39.47 | cr2 | name the suitable character string, is good with anything, is. Ctx* is struct of the 0x168 byte at the portion of Asic*. |
22:39.47 | tsdogs | ho |
22:40.12 | cr2 | AhiDevOpen (out Ctx *, Asic*, const char* name and int 0) |
22:40.29 | BabelO | tsdogs: we are looking for "Not" to set charging variable, so it is reverse |
22:40.37 | cr2 | a bit outdated, though. but good anyway. |
22:41.13 | cr2 | i've see the sigmarion3 sdk somewhere ... |
22:41.21 | tsdogs | I made the change to that (but probably you did not see) should have done the patch I changed it to ! read("Not") |
22:41.53 | tsdogs | That's why mine is working ... |
22:41.54 | BabelO | ahh, sorry |
22:42.40 | tsdogs | BabelO: you know what cut/paste is ;) |
22:42.59 | BabelO | yes :( |
22:44.23 | tsdogs | BabelO: I'm trying to find out what does qtopia do with the resulting code for the PWR_BUTTON, you know where the call comes from ? |
22:44.56 | tsdogs | wrong question, just ignore it (processEvent ,..) |
23:00.31 | cr2 | hmm. nothing ATI-specific there. |
23:01.14 | BabelO | tsdogs: did qss2 build for you ? |
23:01.57 | tsdogs | qss2 ? I only have qss did not know there is also a qss2 |
23:02.18 | BabelO | i see it in subtree with a readme.txt |
23:02.50 | tsdogs | I have it under tools ... |
23:03.18 | tsdogs | plugindecoder.cpp:22:33: error: qtopia/qtopiaglobal.h: No such file or directory |
23:03.56 | tsdogs | It did not build by default |
23:04.30 | tsdogs | so maybe config did not prepare it correctly... |
23:05.27 | tsdogs | You get the same missing include path error ? |
23:06.28 | tsdogs | I think they did not correctly migrate it to new 4.2.1 |
23:06.49 | tsdogs | maybe changing it to qtopiabase/qtopiaglobal.h will work. |
23:15.46 | tsdogs | BabelO: I cannot find where this Key_F34 / F35 are used ... |
23:16.13 | BabelO | tsdogs: i use grep F34 * -R |
23:16.28 | tsdogs | thx |
23:16.40 | BabelO | it is used in src/server/phone/phoneserver.cpp or something |
23:17.22 | *** part/#htc-linux rmoravcik (n=rmoravci@adsl-d181.84-47-56.t-com.sk) |
23:22.47 | tsdogs | I think that either the use is not implemented, or it is ignored 'case sould be handled by the kernel ... |
23:29.51 | tsdogs | Well, we should write a keyboard filter anyway (for the hinge at least, so we might want to intercept it there and do the actions...) |
23:32.15 | BabelO | yes maybe same for keyboard backlight |
23:32.46 | tsdogs | yep. |
23:33.36 | BabelO | you find nothing about Fxx funtion of hinge ? |
23:34.26 | tsdogs | Nope, only the FLIP thing, but as I sayd don't want to trigger a phone event with it. |
23:35.37 | BabelO | ok so flip is managed by opie |
23:37.06 | tsdogs | according to documentation yes. Also hinge events are switches, and need at least 2 mappings (closed/open and roatate) |
23:38.29 | tsdogs | In opie (not opie2) it is handled in something like the htcbuthandler, but keyboard filter I think could be a nice place to gather all the actions there. |
23:39.38 | tsdogs | You know what this is? Key_Mode_switch = 0x0100117e, // Character set switch |
23:40.02 | tsdogs | never mind international related |
23:40.51 | BabelO | no don't know |
23:42.05 | tsdogs | BabelO: you also mapped JOG_DOWN and JOG_UP (on mine it is a Volume Up/Down event) |
23:42.06 | tsdogs | I found this keys. Key_VolumeDown Key_VolumeUp |
23:42.06 | BabelO | yes, i map they to F5 and F6 that are vol up/down |
23:42.31 | BabelO | ah ok ... i ll look why i do that |
23:45.47 | tsdogs | BabelO: yes Key_Flip is handled by phone/dilaer code, so won't be using that one... |
23:47.18 | BabelO | i ll map task switch to windows key on ba |
23:47.50 | tsdogs | ok |
23:48.30 | tsdogs | Enough for today. bon nuit BabelO |
23:48.42 | BabelO | yes good night |