00:00.13 | cr2 | good night. |
00:01.46 | lkcl | you out cr2? |
00:06.23 | cr2 | yes. |
00:25.27 | lkcl | k dude. end of long day. |
01:01.47 | Kevin2 | Hi |
01:20.59 | Speedy2 | Hey all |
02:28.59 | gw280 | in case anyone's wondering how to get a windows mobile to act as a modem nicely for a linux computer, I've documented it at http://blog.gwright.org.uk/articles/2006/11/22/using-a-windows-mobile-phone-to-connect-to-the-internet |
03:38.38 | lkcl | g4 read is a hell of a lot simpler |
03:38.59 | lkcl | ta gw280. i'll wait until i have linux on the phone first, though :) |
03:39.45 | Speedy2 | lkcl: What is your local time? |
03:39.50 | lkcl | GMT |
03:39.53 | Speedy2 | Seems like you've been up all night! |
03:40.01 | lkcl | no, i just got up for a bit again |
03:40.11 | gw280 | lkcl: heh |
03:40.19 | lkcl | have small baby - it's baby feeding time. again. |
03:42.31 | ljp | been there. doing that :) |
03:43.59 | lkcl | hehe |
03:44.26 | Speedy2 | ljp: Hey, I didn't know you had a child |
03:46.21 | ljp | 15 months old now :) |
03:46.29 | Speedy2 | Congrats. |
03:46.33 | ljp | thx |
04:20.21 | gw280 | LJP!!!!!!! |
05:18.16 | ljp | :) hi |
07:09.29 | *** join/#htc-linux psokolovsky (n=psokolov@237.usernat.ip.net.ua) |
08:11.32 | *** join/#htc-linux Fossi (n=cmewes@195.177.48.115) |
08:30.13 | *** join/#htc-linux rob_w (n=rob_w@p549B993F.dip0.t-ipconnect.de) |
11:11.16 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
12:42.09 | *** join/#htc-linux WizMaui (n=WizMaui@62.112.90.179) |
14:00.33 | lkcl | another phone arrived, haha |
14:01.48 | cr2 | lkcl: g4 ready ? :) |
14:02.28 | lkcl | naah dude. i only just woke up 20 mins ago |
14:02.51 | lkcl | took a look at the other listing when i woke up at 4am - it's a _lot_ simpler than the g3 one |
14:03.15 | cr2 | :) yes. |
14:03.25 | cr2 | but them we will have both. |
14:03.37 | lkcl | yep. |
14:03.43 | lkcl | i'll do it this evening. |
14:04.14 | lkcl | i have to take baby in to be photographed. can't quite believe that the british govt requires that children aged 6 weeks require their own passport |
14:08.14 | *** join/#htc-linux pH5 (n=ph5@e178231089.adsl.alicedsl.de) |
14:08.32 | Fossi | at least the passport doesn't have to be machine-readable or does it? |
14:09.33 | cr2 | lkcl: i thought you don't have passports in the UK ? hmm, you want to Holland. |
14:10.16 | lkcl | yep. actually we'll be having fascist ID cards with a database for criminals to bribe government officials and obtain biometic information from |
14:10.55 | cr2 | Fossi: with the current trends they will want to have the fingerprints of the 6 day olds ;) |
14:13.28 | Fossi | i think the uk's best move so far is actually paying people for reporting crimes they've seen on the surveilance cams that are around almost everywhere |
14:14.34 | Fossi | big brother makes you watch each other... |
14:15.10 | cr2 | i think i'll update my id card before they will start with this fingerprinting scam. |
14:54.20 | *** join/#htc-linux rob_w (n=rob_w@p549B993F.dip0.t-ipconnect.de) |
14:55.50 | cr2 | pH5: are you here ? |
15:04.10 | pH5 | I am. |
15:04.48 | cr2 | i have modified the tsc2046 driver to be much more reusable, but there is one ugliness there. |
15:05.19 | cr2 | the platform_device structure is created in the .probe function ;) |
15:06.11 | cr2 | can you have a look how we can painlessly move it "upwards" ? |
15:06.33 | cr2 | then the driver should work on all known devices. |
15:06.40 | cr2 | ok. thanks. |
15:06.48 | cr2 | htcuniversal_ts2.c |
15:06.59 | pH5 | cr2: are you still using the same interleaved x/y readout pattern that hx4700 uses? |
15:07.00 | cr2 | i'll remove htcuniversal_ts.c now. |
15:07.15 | pH5 | alternating x/y, I mean |
15:07.16 | cr2 | i use the old hx4700 .12 driver |
15:07.31 | cr2 | yes, because the newer does not work for me. |
15:07.37 | pH5 | ok, I'll have a look. |
15:07.41 | cr2 | the newer is nonlinear. |
15:07.50 | cr2 | the same old problem. |
15:07.54 | pH5 | it doesn't work for the magician |
15:08.11 | cr2 | ok, you have the same parameters as universal. |
15:08.28 | cr2 | btw, do we really need 24bit spi ? |
15:08.49 | pH5 | I don't know about the 24bit, it's just what the wince driver does. |
15:08.50 | cr2 | i think your generic spi driver approach is the best. |
15:09.06 | pH5 | If only I could find time to follow up on that. |
15:09.12 | cr2 | i have seen the sample wince code for tsc2046, and they don't do 24bit. |
15:09.35 | cr2 | well, if you want to do 24bit, a fix to the pxa2xx-spi driver is needed. |
15:09.55 | cr2 | + /* HTC Universal */ |
15:09.55 | cr2 | + ts->port=1; |
15:09.55 | cr2 | + ts->clock=CKEN23_SSP1; |
15:09.55 | cr2 | + ts->pwrbit_X=1; |
15:09.55 | cr2 | + ts->pwrbit_Y=1; |
15:10.04 | cr2 | i have such code now |
15:10.20 | cr2 | since only these 4 parameters differ beteen the phones. |
15:10.41 | pH5 | you should know that only platform_drivers are allowed on the platform bus |
15:11.07 | pH5 | you should convert all device_drivers to platform_drivers, otherwise it'll break during the transition to 2.6.19 |
15:11.16 | cr2 | yes. ok. |
15:11.36 | cr2 | i think i have copied this driver from h4000 . blame psokolovsky :) |
15:12.13 | cr2 | i also need to add the ohci init. you are the only one so far who does it right. |
15:12.37 | pH5 | cr2: I guess I copied that from the zaurus kernels |
15:12.44 | cr2 | ok. |
15:15.12 | psokolovsky | Hi! |
15:15.17 | pH5 | cr2: but I assume you don't want to export the whole touchscreen_data structure, right? |
15:15.25 | cr2 | psokolovsky: hi |
15:15.30 | pH5 | hey psokolovsky |
15:15.31 | psokolovsky | we should convert *all* (which is pertinent, of course) |
15:15.39 | cr2 | pH5: that's why i'm asking an expert to help me :) |
15:15.40 | psokolovsky | and start with asic3_base ;-) |
15:16.15 | cr2 | psokolovsky: can we get rid of the &foobar_asic3.dev ugly pointer ? |
15:16.39 | cr2 | is there a more sensible way to ask the parent device about it ? |
15:16.43 | psokolovsky | cr2, why is it ugly and what ideas you have? |
15:16.52 | pH5 | yeah, that would be of interest to me to - I face the same problem with the magician cpld |
15:17.01 | psokolovsky | cr2, we could fish in device structure, I guess. |
15:17.37 | cr2 | it's ugly because it's an extern in the header and is copied over from device to device. |
15:17.45 | cr2 | that will be nice. |
15:21.04 | psokolovsky | pH5, I guess, we don't want *all* drivers to be platform, right. Some dev-independent thing can be a normal driver, or maybe even module, if it's not really a driver, right? |
15:22.16 | cr2 | psokolovsky: the _udc is probably not a platform driver, only its parent is. |
15:22.26 | pH5 | psokolovsky: right, only the drivers that attach to the platform bus have to be platform_drivers |
15:22.36 | cr2 | but magician does not have asic3. |
15:22.55 | psokolovsky | cr2, IIRC, _udc canoot be driver/module at all. only static. the rest is decoration ;-) |
15:23.05 | cr2 | ok. |
15:23.28 | cr2 | but it does not make sense to pull asic3 gpios if the asic3_base is not loaded. |
15:23.51 | psokolovsky | pH5, well, I guess not only. if we really need to pass something in via platform_data, we can "attach" it to that bus too, right? ;-) |
15:24.18 | pH5 | psokolovsky: I didn't understand that. |
15:24.24 | psokolovsky | cr2, yes, and it all goes again to static irq base, getting rid of "child devices", etc. |
15:24.48 | cr2 | imho, it should be possible to run the device with the asic3_base unloaded. |
15:25.22 | cr2 | one more standardization question. |
15:25.36 | cr2 | there is a .shutdown platform_device function. |
15:25.47 | psokolovsky | pH5, I see 2 purposes of platform_devices now: 1) direct, for devices which are attached to "dumb" "platform" bus, 2) and for passing extra info to tehir drivers via platform_data member. we can (ab)use the 2nd on occasion. |
15:25.47 | cr2 | is somebody using it ? |
15:26.07 | psokolovsky | cr2, what is it for? ;-) |
15:26.23 | cr2 | don't know myself :) but it's in the header. |
15:27.01 | psokolovsky | cr2, so, I guess it's extra hook in add/remove susp/res sequence |
15:27.01 | cr2 | i see planform_bus as a replacement for pci_bus for us small people :) |
15:27.19 | psokolovsky | so far, we seem to live w/o it ;-) (.shutdown) |
15:27.42 | cr2 | yes. maybe i should grep in the kernel tree. |
15:30.00 | cr2 | psokolovsky: ./arch/arm/mach-pxa/h4000/h4300_kbd.c |
15:30.19 | cr2 | has .shutdown |
15:30.21 | psokolovsky | cool ;-) |
15:30.26 | cr2 | lol |
15:30.28 | psokolovsky | and it's just suspend, right? |
15:30.47 | cr2 | .suspend = h4300_kbd_suspend, |
15:30.49 | psokolovsky | so, I guess it's called before driver removal to shutdown hardware |
15:30.56 | cr2 | .shutdown = h4300_kbd_shutdown, |
15:31.20 | pH5 | cr2: something like this: http://en.pastebin.ca/255000 |
15:31.28 | pH5 | cr2: and put tsc2406_mach_info in its own header |
15:31.39 | cr2 | back in 10min |
15:32.39 | psokolovsky | gawd, we need some kind of kernel defconfig control solution |
15:33.50 | psokolovsky | pH5, so, do you think my idea about platform_data usage is valid? |
15:36.19 | pH5 | psokolovsky: could you give an example for that? |
15:37.18 | psokolovsky | pH5, nope, just thoughts. I didn't really hit that yet (well, we don't have many cross-device drivers yet). |
15:37.37 | psokolovsky | pH5, so well, just fyi, maybe you'll remember that later if needed ;-) |
15:38.20 | psokolovsky | just consider, say SPI driver which needs to be parametrized with device-specific data |
15:38.24 | psokolovsky | how to do that? |
15:38.42 | psokolovsky | (that's rhetorical q for now) |
15:40.30 | pH5 | Ok, I'll try to keep that in mind :) |
15:43.28 | cr2 | pH5: thanks. looks great. now we need to solve the problem with the asic3 irq dynamic base, and the horrible driver duplication will come to an end :) |
15:47.09 | psokolovsky | pH5, hey, btw I guess I just was reading your thoughts, that patch is exactly about passing bits down to TS driver, right? |
15:49.48 | cr2 | psokolovsky: it is still not the final stage. actually the platform_device here is the pxa2xx-spi |
15:50.07 | psokolovsky | ok, so that's what I mean ;-) |
15:50.34 | cr2 | pH5 started to implement it for magician, but we use 24bit spi transfers, which are not included in pxa2xx-spi |
15:50.45 | psokolovsky | nevermind, I won't bother you, will look what you will put together, that's good thing to do |
15:51.37 | cr2 | ads7648 is the predecessor of tsc2046, it should be 100% compatible in some way. |
15:56.10 | psokolovsky | yep, exactly (or so I heard) |
15:56.46 | psokolovsky | so, we should try to commonize it all, but hx4700+universal would good start, of course |
15:57.46 | cr2 | yeah, they share most of the hardware. |
15:58.24 | cr2 | the hx4700_power is the next in the clean process. |
15:58.53 | cr2 | to get asic3_owm as it should be. |
15:59.07 | psokolovsky | ok, nice plan |
15:59.42 | psokolovsky | I won't go in that direction then, to not interfere |
15:59.56 | psokolovsky | so far in my q is to clean UDC for hx4700 ;-) |
16:01.00 | cr2 | i think pH5 was planning to do something with hx4700_power. |
16:01.40 | cr2 | i need to find out the conversion coefficients for the universal. for some reason i don't get the right battery data on the universal. |
16:01.54 | cr2 | with the hx4700_power values. |
16:03.39 | pH5 | cr2: look at the ds2762 data sheet, afaik there are only two possibilities |
16:28.16 | lkcl | CMP R0, #3 and BCS loc_2C8 |
16:28.26 | lkcl | is that if R0 < 3 goto loc_2c8 |
16:28.35 | lkcl | or is it if R0 <= 3 goto loc_2c8 |
16:28.36 | lkcl | ? |
16:34.01 | cr2 | i think < |
16:34.13 | cr2 | why carry will be set on 0. |
16:37.36 | cr2 | evil. if RO > 3 goto loc_2c8 |
17:06.05 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
17:10.04 | cr2 | goxboxlive: hi. i have found a bug in the touchscreen driver (!) by cleaning the code. i'm surprised it did not crash the machine. can you compile and test it ? |
17:12.11 | goxboxlive | of course. |
17:18.07 | *** join/#htc-linux pleemans (n=peter@d51A5E421.access.telenet.be) |
17:27.06 | *** join/#htc-linux toi_ (n=peter@d51A5E421.access.telenet.be) |
17:46.06 | *** join/#htc-linux rob_w (n=bob@p85.212.177.66.tisdip.tiscali.de) |
17:50.00 | psokolovsky | Ok, so here |
17:50.32 | psokolovsky | 's confirmation: sound doesn't work on hx4700 with 2.6.18, familiar 0.8.4 |
17:51.01 | cr2 | ok. how can we check that dma works ? |
17:51.17 | cr2 | i2c is more or less ok, and gpios are right. |
17:52.45 | *** join/#htc-linux TripleDES (n=sergio@84.77.161.251) |
17:52.57 | psokolovsky | dmesg: |
17:52.57 | psokolovsky | ALSA sound/core/pcm_native.c:1510: playback drain error (DMA or IRQ trouble?) |
17:53.12 | psokolovsky | dunno, need try older kernels, likely |
17:53.13 | cr2 | the same error here ;) |
17:53.56 | psokolovsky | will try to play with low-level alsa utils first |
17:53.59 | cr2 | it can't be irq trouble ? i don't think there is some irq involved. |
17:54.11 | cr2 | so it should be dma. |
17:54.20 | cr2 | lcd uses its own. |
17:54.52 | psokolovsky | with speaker-test: |
17:55.04 | psokolovsky | was set buffer_size = 16384 |
17:55.04 | psokolovsky | <PROTECTED> |
17:55.04 | psokolovsky | Write error: -5,Input/output error |
17:55.04 | psokolovsky | xrun_recovery failed: -5,Input/output error |
17:55.04 | psokolovsky | Transfer failed: Operation not permitted |
17:55.08 | psokolovsky | ALSA sound/core/pcm_lib.c:1714: playback write error (DMA or IRQ trouble?) |
17:55.54 | cr2 | it looks like dma timeout. |
17:56.26 | psokolovsky | wonder, are there big differences in DMA between pxa25x & 27x? can we suspect cpu part or only ak codec? |
17:56.57 | cr2 | pxa27x has twice more channels. |
17:57.16 | cr2 | ak codec was working in .16 afaik. |
17:57.23 | cr2 | says Giorgio |
17:57.40 | cr2 | were there any dma api changes in between ? |
17:57.49 | psokolovsky | but anything more funky in 27x? (think UDC diffs) |
17:58.02 | psokolovsky | we have report even .17-hh2 working |
17:58.09 | lkcl | http://hands.com/~lkcl/hp6915/g4.c |
17:58.16 | cr2 | ok, .17-18 ? |
17:58.21 | lkcl | i'm getting lost of i2c timeouts on the htc sable |
17:58.27 | lkcl | s/lost/lots |
17:59.06 | cr2 | lkcl: then you didn't powerup/reset the codec. it's like with wifi. |
17:59.15 | lkcl | argh ok. |
17:59.26 | cr2 | speakers/mic has nothing to do with it. |
17:59.56 | lkcl | yes i possibly found another ASIC3 gpio after doing a diff of dump asic3gpio |
18:01.04 | cr2 | lkcl: the first part of your g4.c is setting up the PXA gpios. it's something for the SableGPIO (Init column) |
18:03.01 | *** join/#htc-linux asylumed (n=insanity@196.211.116.2) |
18:05.23 | *** join/#htc-linux rob__w (n=bob@p85.212.140.219.tisdip.tiscali.de) |
18:06.36 | asylumed | rob_w: hey |
18:07.09 | lkcl | ok. hooray. |
18:08.40 | cr2 | foodtime. |
18:16.02 | *** join/#htc-linux BabelOued (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
18:16.24 | goxboxlive | cr2 ts works |
18:19.20 | goxboxlive | still no usbhost |
18:22.00 | psokolovsky | ok, 2.6.17-hh3 gives me the same result (doesn't work) |
18:41.41 | goxboxlive | I remember there where a lot thathappend between 16-hh2 -> hh4. We had problem compiling the universal, and there where some drivers that didnt work. Also ACX has become more unstable that it was before. |
18:42.35 | goxboxlive | I tried out the xanadux-.12 kernel some days ago, and i didnt had any failures with ACX, but with late 16, .17 and now .18 it failes when it is loaded. |
18:43.39 | goxboxlive | Begore i couls install applications using WLAN, but that failes a lot now, especially if it is lot of dependencies that also needs to be installed. |
18:45.50 | goxboxlive | ( idont mean it failes when the acx modules is loaded, but when it is transfering/recivien a lot of bits. ) |
18:50.17 | lkcl | you know: i'm _really_ tempted to go back to that 2.6.12 kernel tree. |
18:50.21 | psokolovsky | ok, so here come the miracles, 2.6.16-hh8 doesn't work for me either |
18:50.29 | psokolovsky | lol |
18:51.54 | lkcl | psoko: you're trying an hx4700, right? |
18:51.57 | psokolovsky | yep |
18:52.33 | psokolovsky | So, here's the question: anybody at all ever had sound working on hx4700 with *haret* |
18:52.46 | psokolovsky | and that's q to Kevin2, probably |
18:58.24 | psokolovsky | so, .16 and .17 have those differerent ak4641.c versions, but they behave the same |
19:07.46 | lkcl | btw what's OWM stand for, again? |
19:09.13 | lkcl | you mean, psoko, has anyone ever had sound working on the hx4700 when booting linux using haret.exe, as opposed to starting up after replacing wince and starting linux from the bootloader? |
19:10.19 | psokolovsky | one-wire, find your meaning for M, maybe module |
19:10.23 | psokolovsky | owm == w1 |
19:10.33 | psokolovsky | unless I'm hallucinating, of course |
19:10.53 | psokolovsky | lkcl, yes, exactly regarding haret and sound |
19:11.02 | lkcl | ok. ta. |
19:11.15 | lkcl | and EXTCF? (all on the asic3 page) |
19:11.46 | lkcl | i'm looking at _loads_ of differences in the asic3 mem-map and i'm wondering if some of them are related to sound failure and acx failure on the sable. |
19:13.22 | *** join/#htc-linux g3gg0_ (n=g3gg0@ppp-82-135-15-150.dynamic.mnet-online.de) |
19:14.53 | psokolovsky | something new: "ACR reset |
19:14.53 | psokolovsky | " |
19:17.27 | *** join/#htc-linux doepiedoe (n=doepiedo@84-53-96-51.fiber.unet.nl) |
19:17.55 | psokolovsky | ok, ACR is UART stuff |
19:18.31 | psokolovsky | the most annoying actually not that it doesn't work |
19:18.42 | cr2 | psokolovsky: isn't ACR accumulated current something ? |
19:18.45 | psokolovsky | but that almost no errors are thrown |
19:18.51 | psokolovsky | lol |
19:18.54 | psokolovsky | dunno ;-) |
19:19.00 | *** join/#htc-linux buz (n=buz@80-218-107-221.dclient.hispeed.ch) |
19:19.04 | psokolovsky | at least it's apparaently not sound related |
19:19.16 | cr2 | i think its hx4700_power battery message. |
19:19.54 | psokolovsky | cr2, ack, thanks |
19:21.37 | cr2 | lkcl: the "haret sound" is a blueangel-specific problem. |
19:22.15 | cr2 | asylumed: i'll commit the (probable) fix to the xanadux .12 CVS, can you test it ? |
19:22.31 | psokolovsky | cr2, can you confirm any pxa27x & ak4641 device sound work under haret? |
19:23.07 | cr2 | i have only the universal. Kevin did not test sound afaik, on apache. |
19:24.14 | asylumed | cr2: hey, just tuned in... the kernel or haret? |
19:24.16 | psokolovsky | cr2, so, what about uni? |
19:24.17 | cr2 | asylumed: the fix is for the kernel. after that sound should work with any haret. |
19:24.50 | asylumed | cr2: kewl... sure i'll test it... |
19:24.53 | *** join/#htc-linux wewe (n=bob@p85.212.162.185.tisdip.tiscali.de) |
19:24.57 | psokolovsky | cr2, I'm going to commit error diagnostic to sound code, any concerns? |
19:24.59 | cr2 | psokolovsky: i have not tried myself, but goxboxlive says it does not work. the error message is the same that you have. |
19:25.04 | cr2 | no. |
19:25.18 | cr2 | it does not work anyway ;) |
19:25.19 | psokolovsky | cr2, .12? |
19:25.30 | cr2 | .12 was working. |
19:25.36 | psokolovsky | cr2, from haret? |
19:25.43 | cr2 | mic and speakers and headphone. |
19:25.54 | cr2 | yes, we don#t have other bootloader. |
19:26.06 | cr2 | i'd like to have one, though. |
19:26.14 | psokolovsky | Kevin2: can there be any issues with DMA and haret boot? |
19:26.35 | psokolovsky | cr2, so, no concerns regarding sane error reporting? |
19:27.05 | cr2 | psokolovsky: we switch DMA off in haret, but the kernel should be clever enough to setup it properly. |
19:27.10 | cr2 | no. |
19:27.19 | psokolovsky | ok |
19:27.42 | cr2 | we need to repair the ak4641 sound, all methods are good for it :) |
19:28.28 | cr2 | hmm. i'll checkout the .12 kernel from sf.net |
19:29.50 | doepiedoe | cr2, can't we use uboot_load too? ;) |
19:30.11 | doepiedoe | ahh it's for omap |
19:36.45 | cr2 | doepiedoe: if you'll port it ;) |
19:37.04 | cr2 | we can easily adapt the hx4700 bootldr |
19:45.03 | doepiedoe | sounds nice :) |
19:45.16 | doepiedoe | i can do some c ;) |
19:45.28 | doepiedoe | would you boot from a binary i'll give you? ;) |
19:47.23 | lkcl | btw cr2 you know what EXTCF stands for? |
19:48.00 | *** join/#htc-linux _buz (n=buz@84-73-95-139.dclient.hispeed.ch) |
19:55.23 | cr2 | extrnal CF, but i'm not sure. |
19:57.09 | cr2 | goxboxlive: i've modified the touchscreen again. will commit the changes soon. we are close to the generic tsc2046 touchscreen driver. hmm. want the static asic3 irq base :) |
19:58.43 | cr2 | actually there should be a platform_bus init, like the pci_bus where you can assign the irqs. |
19:59.15 | cr2 | then you can connect 10 asic3 chips to the cpu and be happy . |
19:59.46 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
20:00.47 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
20:01.46 | *** join/#htc-linux iggy (n=iggy@12.45.184.235) |
20:03.26 | cr2 | goxboxlive: added ohci. |
20:04.19 | goxboxlive | cr2: what do you mean? |
20:04.55 | cr2 | <PROTECTED> |
20:06.53 | cr2 | it's a big update now. compile the ohci into the kernel. check the touchscreen. |
20:08.41 | *** join/#htc-linux _buz (n=buz@84-73-95-139.dclient.hispeed.ch) |
20:10.49 | cr2 | goxboxlive: do you see the patch ? |
20:11.26 | goxboxlive | not yet, it's not ready for users yet. Only in cvsweb |
20:11.33 | cr2 | ok. |
20:11.45 | goxboxlive | now it is i guess |
20:12.28 | cr2 | lkcl: does your touchscreen work ? |
20:12.59 | *** join/#htc-linux _buz (n=buz@84-73-95-139.dclient.hispeed.ch) |
20:15.28 | Kevin2 | Hi |
20:16.09 | Kevin2 | psokolovsky: I've not tried sound on apache. I've been waiting for universal to work first. :-) |
20:16.15 | goxboxlive | hi Kevin2 |
20:16.33 | Kevin2 | Are you concerned about dma and haret? Have you tried gnu-haret / earlier haret? |
20:16.47 | psokolovsky | Kevin2, hi! |
20:16.51 | cr2 | Kevin2: i'm waiting for hx4700 :) |
20:16.53 | psokolovsky | yes, concerned |
20:17.01 | psokolovsky | no, didn't try yet |
20:17.05 | psokolovsky | will try |
20:17.24 | psokolovsky | old mainline haret had issues with UDC, I assume due to DMA |
20:17.41 | psokolovsky | gnuharet and haret-0.4.3 don't have UDC probs |
20:17.57 | Kevin2 | Okay, I suggest you try gnu-haret. The code to disable dma is slightly different between haret and gnu-haret. gnu-haret tries to zero everything out, while mainline is now just disabling. |
20:18.18 | cr2 | Kevin2: it seems your pxa27x keyboard driver goes into mainline. it seems i had a good idea to put it into driver/input/keyboard :) |
20:18.49 | Kevin2 | I was reading the pxa docs, and the gnu-haret (and old mainline) are definitely in violation of the spec. (They zero things while dma might still be active.) However, it may be worthwhile to zero after disabling. |
20:19.29 | *** join/#htc-linux buz_ (n=buz@84-73-95-139.dclient.hispeed.ch) |
20:20.14 | psokolovsky | ALSA sound/arm/pxa2xx-i2sound.c:434: Started DMA on channel 10 |
20:20.14 | psokolovsky | ALSA sound/core/pcm_native.c:1510: playback drain error (DMA or IRQ trouble?) |
20:20.14 | psokolovsky | ALSA sound/arm/pxa2xx-i2sound.c:302: DMA IRQ |
20:20.31 | Kevin2 | cr2: cool. I only modified someone else's code. |
20:20.51 | lkcl | cr2: well i get interrupts, and cat /dev/input/ts0 gets data but opie fails miserably after calibration |
20:21.05 | cr2 | i know, and you wanted to put it into the apache subdir. |
20:21.08 | lkcl | other than that _yes_ it works :) |
20:21.33 | cr2 | lkcl: you need to know the "power bits". |
20:21.36 | Kevin2 | cr2: Heh - thanks. |
20:21.40 | lkcl | ahh |
20:22.41 | cr2 | lkcl: there are devices that use 0 (hx4700) , 1(universal and magician) and 3(beetles i think). |
20:23.20 | cr2 | you should run "set trace " on the spi data buffer. |
20:23.30 | cr2 | then you |
20:23.36 | cr2 | will know. |
20:23.39 | goxboxlive | cr2: Yes. TS works |
20:23.46 | cr2 | great! |
20:23.49 | goxboxlive | But |
20:24.12 | goxboxlive | HWat is the unknown product when loading the usbserial module? |
20:24.48 | cr2 | do you have the usb host startup messages ? |
20:25.04 | goxboxlive | hold |
20:25.07 | goxboxlive | on |
20:25.08 | Kevin2 | cr2: Did you get a chance to retry haret on jornada? I've discussed this with the egcc folks - it looks like the problem is egcc is calling GetCommandLine in the startup code and that is a wm3 function only. |
20:25.44 | cr2 | Kevin2: do you have a working binary, i can try it right now. |
20:26.28 | cr2 | i want to put jornada820 into hh.org CVS, so i need a bootloader . |
20:26.53 | cr2 | don't ask me how we booted it 3 years ago ;) |
20:27.04 | Kevin2 | haret.exe-gccmain from svn -- this is a hacked version that bypasses egcc init but still calls gccmain init. |
20:27.10 | cr2 | ok. |
20:28.11 | cr2 | we used hpcboot from netbsd and it was a PITA. |
20:30.31 | cr2 | it works. "Looking at machine Sable" and crash. |
20:33.11 | goxboxlive | cr2: Here is the USBHOST dmesg http://pastebin.ca/255290 |
20:33.43 | goxboxlive | it works |
20:33.51 | goxboxlive | take a look here: |
20:34.06 | goxboxlive | root@htcuniversal:~$ lsusb |
20:34.06 | goxboxlive | Bus 1 Device 2: ID 05c6:3197 Qualcomm, Inc. CDMA Wireless Modem/Phone |
20:34.06 | goxboxlive | Bus 1 Device 1: ID 0000:0000 |
20:34.06 | goxboxlive | root@htcuniversal:~$ |
20:34.22 | goxboxlive | root@htcuniversal:~$ lsusb -t |
20:34.22 | goxboxlive | cannot open /proc/bus/usb/devices, No such file or directory (2) |
20:34.36 | cr2 | mount usbfs |
20:34.45 | goxboxlive | I have not loaded the usbserial module yet |
20:35.14 | cr2 | strange. |
20:35.30 | cr2 | Kevin2: what do you think about it ? |
20:35.42 | goxboxlive | root@htcuniversal:~$ lsusb -t |
20:35.43 | goxboxlive | Bus# 1 |
20:35.43 | goxboxlive | `-Dev# 1 Vendor 0x0000 Product 0x0000 |
20:35.43 | goxboxlive | <PROTECTED> |
20:35.43 | goxboxlive | root@htcuniversal:~$ |
20:35.54 | goxboxlive | Should i load the usbserial modules? |
20:36.19 | cr2 | gxoxboxlive: yes. with vendor=xxx product=yyyy debug=1 |
20:36.30 | goxboxlive | ok |
20:37.29 | goxboxlive | With caption V or small v in vendor and product (p/P) |
20:37.41 | cr2 | small. |
20:37.43 | Kevin2 | cr2: can you post full log somewhere (eg, pastebin)? |
20:37.45 | goxboxlive | ok |
20:38.01 | cr2 | Kevin2: yes, i'll copy it from CF. |
20:39.06 | goxboxlive | http://pastebin.ca/255293 |
20:39.51 | goxboxlive | And this gives me: root@htcuniversal:~$ ls -la /dev/ttyUSB |
20:39.51 | goxboxlive | /dev/ttyUSB0 /dev/ttyUSB1 |
20:40.41 | cr2 | looks good. |
20:40.44 | cr2 | <7>drivers/usb/serial/usb-serial.c: the device claims to support interrupt in transfers, but read_int_callback is not defined |
20:41.02 | cr2 | maybe it's just this generic driver. |
20:41.25 | cr2 | <7>drivers/usb/serial/usb-serial.c: found interrupt in on endpoint 0 |
20:41.25 | cr2 | <7>drivers/usb/serial/usb-serial.c: found bulk in on endpoint 1 |
20:41.25 | cr2 | <7>drivers/usb/serial/usb-serial.c: found bulk out on endpoint 2 |
20:42.17 | psokolovsky | cr2, what exactly is DREQing bytes for sound? |
20:43.03 | cr2 | i'm not an usb expert, is the USB0 for in, and USB1 for out ? |
20:43.18 | cr2 | psokolovsky: RAM <-> I2S |
20:43.39 | cr2 | I2S should have its own channel. |
20:43.41 | goxboxlive | cr2: Is it a safe to load this module and have it started at boot? |
20:44.00 | cr2 | goxboxlive: is the phone running ? |
20:44.08 | goxboxlive | No not now |
20:44.15 | psokolovsky | cr2, i.e. is it i2s requesting it? or codec itself? or something else? |
20:44.25 | cr2 | I2S |
20:44.38 | goxboxlive | should i start the phone? |
20:44.54 | psokolovsky | ok |
20:45.11 | cr2 | goxboxlive: yes, and then you can run 'od -x' on /dev/ttyUSB0 |
20:45.21 | cr2 | i don't know how to init video call. |
20:46.00 | cr2 | maybe it's some command sent over usb. i have not seen any special AT commands. |
20:46.40 | cr2 | psokolovsky: the codec is not memory mapped, how can it dma ? |
20:47.24 | psokolovsky | cr2, I mostlly interrested what generates clocks which go to DREQ |
20:47.42 | cr2 | hmm. |
20:47.59 | cr2 | it must be CF on the hx4700 |
20:48.04 | cr2 | from asic3. |
20:48.16 | psokolovsky | (well, I have to say that I have only draft knowledge how DMA works in modern systems, but I hope basic idea is the same as it was on 8080 ;-) ) |
20:48.36 | cr2 | my knowledge is from pdp11 times ;) |
20:49.18 | cr2 | i have hard time to get the 8080 way of doing things, including the shitty assembler ;-) |
20:49.54 | goxboxlive | cr2: ttyUSB0 and ttyUSB1 disapeared when i started the phone |
20:50.14 | cr2 | yes, you should run AT@USBPOWER=1 |
20:50.37 | cr2 | you should init the phone. not exactly what wince does, but the useful things. |
20:51.55 | goxboxlive | yes now they are back |
20:52.53 | goxboxlive | yes now what i am running od -x /dev/ttyUSB0 |
20:52.55 | goxboxlive | and it hangs |
20:53.40 | goxboxlive | cr2 it gave me this: |
20:53.43 | goxboxlive | <7>drivers/usb/serial/usb-serial.c: serial_open |
20:53.43 | goxboxlive | <7>drivers/usb/serial/generic.c: usb_serial_generic_open - port 0 |
20:53.43 | goxboxlive | <7>drivers/usb/serial/usb-serial.c: serial_ioctl - port 0, cmd 0x5401 |
20:53.43 | goxboxlive | <7>drivers/usb/serial/usb-serial.c: serial_close - port 0 |
20:53.43 | goxboxlive | <7>drivers/usb/serial/generic.c: usb_serial_generic_close - port 0 |
20:53.45 | goxboxlive | <7>drivers/usb/serial/generic.c: generic_cleanup - port 0 |
20:53.47 | goxboxlive | <7>drivers/usb/serial/generic.c: usb_serial_generic_read_bulk_callback - port 0 |
20:53.49 | goxboxlive | <7>drivers/usb/serial/generic.c: usb_serial_generic_read_bulk_callback - nonzero read bulk status received: -2 |
20:53.56 | cr2 | the phone hangs ? |
20:54.00 | goxboxlive | no |
20:54.03 | goxboxlive | the od |
20:54.06 | cr2 | the unibversal ? |
20:54.11 | goxboxlive | no |
20:54.14 | cr2 | it's normal. it blocks. |
20:54.34 | goxboxlive | ok, i ended it wiit ctrl+c |
20:54.38 | cr2 | we don't know what is going on ther.e, we need to trace it in wince. |
20:54.40 | cr2 | ok. |
20:54.49 | cr2 | Kevin2: http://pastebin.ca/255310 |
20:54.54 | *** join/#htc-linux psokolovsky_ (n=psokolov@ip.85.202.124.214.dyn.sub-9.broadband.voliacable.com) |
20:55.46 | goxboxlive | Ok, i'll try to reboot it and see if it loads up at boot |
20:56.23 | cr2 | goxboxlive: it does not make much sense. we need a wince trace first. |
20:56.47 | goxboxlive | so i should use usbseral then ? |
20:56.57 | psokolovsky_ | ALSA sound/arm/pxa2xx-i2sound.c:434: Started DMA on channel 8 |
20:56.58 | psokolovsky_ | ALSA sound/arm/pxa2xx-i2sound.c:302: DMA IRQ |
20:56.58 | psokolovsky_ | ALSA sound/arm/pxa2xx-i2sound.c:443: Stopping DMA on channel 8 |
20:56.58 | psokolovsky_ | ALSA sound/arm/pxa2xx-i2sound.c:302: DMA IRQ |
20:57.14 | psokolovsky_ | that's from h4000, working |
20:57.28 | cr2 | ok. |
20:57.33 | cr2 | you have uda1380 ? |
20:57.36 | psokolovsky_ | missing 2 irqs, apparently of buffer xfer completion |
20:57.38 | psokolovsky_ | yep |
20:57.42 | cr2 | ok. |
20:58.05 | psokolovsky_ | As DMA dummy as me, I would think DREQs are not generated, as first thing ;-\ |
20:59.18 | cr2 | we have a generic tsc2046 driver now. only the touchscreen, i don't know what to do with the battery. SoC ? |
20:59.35 | cr2 | and it suffers from dynamic irqs, of course. |
21:00.04 | cr2 | here is a new idea: ask the SoC parent about its base irq. |
21:00.21 | Kevin2 | cr2: Do you need to put SetKMode() around accesses to coprocessor registers on the jornada? |
21:00.25 | psokolovsky_ | cr2, that's what Spyro proposed in his RFC |
21:00.31 | goxboxlive | Well anyway i tested it. It loads ttyUSB0 and ttyUSB1 druing boot now. I guess it not's hurt to use it, right? |
21:00.46 | psokolovsky_ | at least it will be independent from specific SOC |
21:01.06 | *** join/#htc-linux g3gg0__ (n=g3gg0@ppp-82-135-15-150.dynamic.mnet-online.de) |
21:01.07 | cr2 | Kevin2: yes. wince < 3.0 was running in user mode. |
21:03.26 | cr2 | psokolovsky_: i think it's a good idea. then handling multiple SoCs will be easy. |
21:06.21 | psokolovsky_ | yes, if we can't just drop dynamic stuff at whim, that would be a solution |
21:06.21 | psokolovsky_ | and we can turn CPU into SOC too ;-) |
21:06.22 | cr2 | unfortunately, this tsc2046-ts driver is pxa-specific. because it messes directly with the spi registers. |
21:06.35 | cr2 | i'd like to have the static SoC irq, but i creates problems with multiple SoCs. imho. |
21:10.29 | cr2 | psokolovsky_: it's actually a pc bios illusion that the irqs are static. |
21:11.36 | psokolovsky_ | cr2, well, I agree, all those irq solutions are "select 2 of 3" type |
21:11.49 | cr2 | yeah. |
21:11.49 | Kevin2 | cr2: Do you think it would be okay to just run "SetKMode(true)" at haret startup? That would certainly simplify things. |
21:12.00 | cr2 | Kevin2: yes. |
21:12.02 | psokolovsky_ | so if we can work out soc-independent solution, let's have that better |
21:12.03 | Kevin2 | psokolovsky_: Any thoughts on the above? |
21:12.51 | cr2 | Kevin2: haret does nasty crash-friendly things anyway. |
21:13.08 | psokolovsky_ | Kevin2, I have no idea about SetKMode() much. but I guess, it's not immediately necessary to abuse wince insecurity ;-) |
21:13.35 | cr2 | it's only wince < 3.0 relevant |
21:13.52 | cr2 | newer wince always runs in the kernel mode. |
21:14.55 | psokolovsky_ | nothing can surprise me about wince for some time already ;-) |
21:15.03 | cr2 | lol. |
21:20.23 | Kevin2 | cr2: try new haret.exe-gccmain (rev 128). This version sets kmode early - I think we were failing on access to cp15. |
21:21.05 | cr2 | ok. |
21:23.48 | cr2 | Kevin2: yes. now it works. the window is modal (i will not be able to close it, time to add the exit button). |
21:24.04 | cr2 | it says : Found machine |
21:26.43 | doepiedoe | cr2, should i define a ramdisk of size 14336K in the ramdisk configuration of the kernel? or just do some values and define it in the bootconfig of haret |
21:27.17 | cr2 | Kevin2: if i put toolhelp.dll on the machine, it registers kill and ps. cool. |
21:28.13 | *** join/#htc-linux psokolovsky (n=psokolov@ip.85.202.124.214.dyn.sub-9.broadband.voliacable.com) |
21:28.30 | cr2 | doepiedoe: use the actual ramdisk size. if you did not change the standard one, then you know the size. on the kernel command line. |
21:29.03 | doepiedoe | ok, thanks |
21:30.51 | lkcl | cr2: when you say 'power bits' - is that an spi register or something? |
21:30.53 | doepiedoe | get unable to open initial console |
21:31.04 | doepiedoe | are there more things i should enable in kernel? |
21:31.11 | doepiedoe | ramdisk does get found now |
21:31.18 | cr2 | lkcl: no, it's the bits in the data request command. |
21:31.25 | lkcl | o right. |
21:31.32 | lkcl | ok. i think i have an spi trace already... |
21:31.43 | cr2 | 0x9X0000 for X, and 0x9Y0000 for Y. |
21:32.09 | lkcl | ok that was for the gps spi port ha ha |
21:32.11 | cr2 | we use 24bit transfer, so the lower 2 bytes are 0. |
21:32.22 | cr2 | lol |
21:33.09 | lkcl | cr2: que - do you mean i2s rather than spi? |
21:33.12 | cr2 | i don't know why it's not 16bit, because the command is 8bit and the ADC response is 12bit. |
21:33.23 | lkcl | cr2: this is touchscreen, right? |
21:33.29 | cr2 | spi, SSP. |
21:33.34 | cr2 | yes. |
21:33.49 | *** join/#htc-linux psokolovsky__ (n=psokolov@ip.85.202.124.214.dyn.sub-9.broadband.voliacable.com) |
21:33.50 | lkcl | ok there are bits that you set that tell the ADC what 'mode' to be in. |
21:34.03 | cr2 | i2c is the camera (you should trace me the i2c id :), and the sound. and the wifi eeprom. |
21:34.15 | cr2 | and maybe your hw RTC. |
21:34.32 | lkcl | and what ranges of samples to 'average' to avoid bounce. |
21:34.54 | cr2 | lkcl: look for 0x9X0000 and 0x9Y0000 |
21:35.04 | lkcl | cr2: remember, i did that about... 8 days ago. got everything but the camera, because it's still sticking out at an angle of 30 degrees from the back :) |
21:35.17 | lkcl | ok |
21:35.17 | cr2 | ok, i have forgot ;) |
21:35.32 | *** join/#htc-linux psokolovsky__ (n=psokolov@ip.85.202.124.214.dyn.sub-9.broadband.voliacable.com) |
21:35.47 | cr2 | damn. 0xdY0000 |
21:36.02 | lkcl | cr2: got an address for me so i don't get confused with all these different TLAs (three letter anachronyms)... |
21:36.07 | cr2 | these will be the most visible commands, all the rest is random. |
21:36.12 | lkcl | k. |
21:36.16 | cr2 | SSP1 or SSP2 ? |
21:36.24 | lkcl | err.... |
21:36.27 | lkcl | burble. |
21:36.33 | cr2 | look at your driver |
21:36.34 | lkcl | SSP2 is... touchscreen |
21:36.39 | cr2 | ok. |
21:36.50 | cr2 | looking in the pxa-regs.h |
21:37.48 | lkcl | oh now i know what you're talking about - i thought you were talking about sound |
21:37.56 | lkcl | no need i can find it |
21:38.13 | lkcl | P1. |
21:38.30 | cr2 | ok. |
21:38.35 | lkcl | 0x41000010 |
21:39.29 | doepiedoe | strange |
21:39.37 | lkcl | virtual 0xaf000010 |
21:39.42 | lkcl | ok. |
21:39.45 | doepiedoe | copy a gzipped ramdisk to my windows box and windows is showing the actual size of the ramdisk |
21:39.45 | doepiedoe | ;) |
21:41.35 | cr2 | lkcl: SSP2 is 0x4170 |
21:41.38 | *** part/#htc-linux WizMaui (n=WizMaui@62.112.90.179) |
21:41.44 | lkcl | am using SSP1 |
21:41.47 | lkcl | sorry :) |
21:43.00 | lkcl | huh. no trace output. hmmm.... |
21:44.26 | cr2 | are you touching the screen ? |
21:44.29 | lkcl | hmh. getting irqs - no data. |
21:44.30 | lkcl | yes. |
21:44.41 | lkcl | HaRET(7)# set trace 0xaf000000 |
21:44.41 | lkcl | HaRET(8)# set tracemask 0x1f |
21:44.41 | lkcl | HaRET(9)# wi 2 |
21:44.53 | lkcl | Finished installing exception handlers. |
21:44.53 | lkcl | 00030f8b: irq 149(gpio 115) |
21:44.53 | lkcl | 00083f95: irq 149(gpio 115) |
21:44.53 | lkcl | 000c0d18: irq 149(gpio 115) |
21:44.53 | lkcl | 000c98a7: irq 149(gpio 115) |
21:44.54 | lkcl | 001691a9: irq 149(gpio 115) |
21:44.58 | lkcl | Restoring windows exception handlers... |
21:45.00 | lkcl | burble. |
21:45.10 | cr2 | set trace 0xaf700000 |
21:45.17 | lkcl | yep. good idea. |
21:45.53 | lkcl | _whoa_.... |
21:46.02 | lkcl | eep. |
21:46.06 | cr2 | pastebin. |
21:46.09 | lkcl | that was a lot of trash onscreen |
21:46.19 | cr2 | lol |
21:47.40 | lkcl | http://hands.com/~lkcl/hp6915/ssp2.trace.txt |
21:47.44 | cr2 | set the mask only for 1 register. |
21:48.04 | cr2 | 91 |
21:48.24 | cr2 | and d1 |
21:48.33 | cr2 | the same as on universal. |
21:48.51 | cr2 | wtf we are using 24bit ??? |
21:49.04 | lkcl | que? |
21:49.15 | *** join/#htc-linux Fossi (n=Fossi@e176115211.adsl.alicedsl.de) |
21:49.21 | lkcl | remember that was trace on _all_ the registers not just data. |
21:49.44 | cr2 | 00b10000 ? |
21:49.57 | lkcl | oh. right. ok. just the data register. |
21:50.00 | lkcl | 1sec... |
21:50.11 | *** join/#htc-linux psokolovsky__ (n=psokolov@ip.85.202.124.214.dyn.sub-9.broadband.voliacable.com) |
21:51.00 | cr2 | need a tsc2046 datasheet. |
21:51.15 | lkcl | need a brain |
21:51.29 | cr2 | the commands are 8bit and the responses are 16bit as you can see. why 24bit spi ? |
21:51.38 | lkcl | o god. that's great. |
21:52.18 | lkcl | that's not possible. you _have_ to exchange the same amount of read data as write data |
21:53.23 | cr2 | then use 16bit |
21:53.51 | cr2 | maybe we should try. |
21:55.40 | lkcl | ok just because wince is _reading_ 16-bit doesn't necessarily mean... |
21:55.41 | lkcl | :) |
21:58.40 | lkcl | ok updated |
21:58.57 | lkcl | some 00d10000s some 91s and some b1s |
21:59.56 | lkcl | a 00c00000.... |
22:00.14 | cr2 | b is z1-position. |
22:00.23 | doepiedoe | hey ;) |
22:00.33 | doepiedoe | looks like my SD slot is working in linux |
22:00.33 | doepiedoe | :D |
22:00.46 | cr2 | c is z2-position |
22:00.55 | lkcl | ahh |
22:01.15 | cr2 | 9 is y |
22:01.17 | cr2 | d is x |
22:01.31 | cr2 | 1 means ADC is on. |
22:01.39 | cr2 | hx4700 is most powersaving. |
22:02.18 | cr2 | doepiedoe: and not in wince ? |
22:02.46 | cr2 | now i need to remember what is z1 and z2. |
22:03.42 | cr2 | it's for the pressure measurement. |
22:06.57 | lkcl | ok i am getting data by using SSDR_P2 but it is so inaccurate that i can't calibrate the screen. |
22:08.34 | lkcl | gonna try htcsable_ts2 although last time it failed miserably |
22:09.53 | doepiedoe | cr2, i loaded mmc_block and after that i loaded asic3_mmc |
22:10.10 | lkcl | hooray, doepiedoe! |
22:10.13 | doepiedoe | asic3_mmc: ASIC3 MMC/SD Driver, controller at 0xc000000 |
22:10.17 | doepiedoe | :) |
22:10.23 | doepiedoe | mknod /dev/mmcblk0 b 254 0 |
22:10.31 | doepiedoe | should find the SD card then right? |
22:11.02 | cr2 | try it. |
22:11.03 | lkcl | yep. looks right. |
22:11.09 | doepiedoe | yes but it doesn't |
22:11.23 | doepiedoe | mkfs.ext3 /dev/mmcblk0 |
22:11.30 | cr2 | then your asic3 is toast |
22:11.33 | doepiedoe | mkfs.ext3: No such device...... |
22:11.37 | lkcl | highfield:/home/lkcl/xda# ls -altr /mnt/initrd/dev/mmcblk0p2 |
22:11.38 | lkcl | brw-r--r-- 1 root root 254, 2 Nov 1 18:54 /mnt/initrd/dev/mmcblk0p2 |
22:11.47 | lkcl | NO doepiedo |
22:11.53 | doepiedoe | ? |
22:11.57 | cr2 | fdisk /dev/mmcblk0 |
22:12.13 | lkcl | that will make the entire device an ext3 drive - you want to create a partition first |
22:12.16 | doepiedoe | unable to open /dev/mmcblk0 |
22:12.27 | lkcl | ahh joy. |
22:12.30 | lkcl | run dmesg |
22:12.34 | lkcl | what does it say at the end? |
22:12.43 | doepiedoe | nothing in dmesg |
22:12.47 | cr2 | it should detect the card. |
22:12.49 | doepiedoe | asic3_mmc: ASIC3 MMC/SD Driver, controller at 0xc000000 |
22:13.00 | lkcl | hmmm... |
22:13.05 | cr2 | hmm. we may also export the hw id in /sys |
22:13.08 | doepiedoe | when i insert in windows it says battery to low can't use sd card |
22:13.14 | cr2 | it is useful for some uses. |
22:13.21 | doepiedoe | theres no /sys |
22:13.23 | lkcl | HOORAY |
22:13.31 | lkcl | touchscreen works. |
22:14.38 | doepiedoe | asic3_mmc isnt used too |
22:14.40 | cr2 | lkcl: can use set it to do 16bit spi ? and cut one byte from the command ? |
22:15.06 | lkcl | ah come on cr2 let me savour the moment a bit where i can actually use this friggin device |
22:15.14 | cr2 | :)))))))))))))) |
22:15.27 | lkcl | and run the alsa mixer etc. |
22:15.37 | cr2 | write the lcd and bl driver ;) |
22:15.59 | lkcl | PM: Adding info for No Bus:i2c-0 |
22:15.59 | lkcl | i2c: error: exhausted retries |
22:15.59 | lkcl | i2c: msg_num: 0 msg_idx: -2000 msg_ptr: 0 |
22:15.59 | lkcl | i2c: ICR: 000007e0 ISR: 00000002 |
22:15.59 | lkcl | i2c: log: [00000442:000007e0] |
22:16.00 | lkcl | PM: Adding info for i2c:0-0051 |
22:16.02 | lkcl | i2c: error: exhausted retries |
22:16.04 | lkcl | i2c: msg_num: 0 msg_idx: -2000 msg_ptr: 0 |
22:16.06 | lkcl | i2c: ICR: 000007e0 ISR: 00000002 |
22:16.08 | lkcl | i2c: log: [00000442:000007e0] |
22:16.10 | lkcl | i2c: error: exhausted retries |
22:16.12 | lkcl | on modprobe snd-htcsable_audio |
22:16.34 | cr2 | 0x51 is eeprom |
22:16.40 | cr2 | where is 0x12 ? |
22:17.19 | cr2 | if it's not there, then the codec is not powered up. |
22:19.09 | lkcl | HA this is great. |
22:19.11 | lkcl | haha |
22:19.18 | lkcl | lemme do a scan... |
22:22.20 | lkcl | ha ha - when audio is enabled, i2cdetect shows UU |
22:22.35 | lkcl | when i rmmod snd-htcsable_audio, it shows up! |
22:22.52 | lkcl | /usr/sbin # rmmod snd_htcsable_audio |
22:22.52 | lkcl | rmmod snd_htcsable_audio |
22:22.52 | lkcl | /usr/sbin # ./i2cdetect 0 |
22:22.52 | lkcl | ./i2cdetect 0 |
22:22.52 | lkcl | WARNING! This program can confuse your I2C bus, cause data loss and worse! |
22:22.53 | lkcl | I will probe file /dev/i2c/0. |
22:22.55 | lkcl | I will probe address range 0x03-0x77. |
22:22.59 | lkcl | Continue? [Y/n] y |
22:23.01 | lkcl | <PROTECTED> |
22:23.03 | lkcl | 00: XX XX XX XX XX XX XX XX XX XX XX XX XX |
22:23.05 | lkcl | 10: XX XX 12 XX XX XX XX XX XX XX XX XX XX XX XX XX |
22:23.07 | lkcl | 20: XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX |
22:23.19 | lkcl | what's the guessing that the enable/disable is the wrong way round? |
22:23.29 | cr2 | then you do something wrong in htcsable_audio. |
22:23.53 | lkcl | don't look at me!! |
22:25.52 | doepiedoe | cr2, should asic3_mmc leave a message ni dmesg when inserting card? |
22:28.13 | cr2 | yes. |
22:28.40 | cr2 | you should better insert the card and then 'modprobe asic3_mmc' |
22:29.03 | psokolovsky__ | cr2, do you know if i2s is anyhow smart? i.e. does it sense that receiver actually receives data? sorry for dumb q ;-) |
22:29.42 | doepiedoe | i did |
22:29.43 | doepiedoe | :S |
22:32.37 | cr2 | psokolovsky_: don't know. |
22:32.44 | psokolovsky__ | ok |
22:33.26 | cr2 | the clock and data can go nowhere. frame too. what else ? |
22:35.08 | lkcl | it's good to have an actual ... like.... input device |
22:37.03 | cr2 | :) |
22:37.13 | cr2 | lkcl: how is opie on 240x240 screen ? |
22:37.36 | lkcl | good - the icons and toolbar still aren't big enough for my liking. |
22:37.43 | cr2 | should not be nice with the universal's VGA image. |
22:37.46 | lkcl | but it's pretty good. |
22:38.00 | *** join/#htc-linux skodde (n=skodde@85-18-136-66.fastres.net) |
22:38.00 | cr2 | icons are small ?? |
22:39.46 | lkcl | icons are fine - the toolbar icons are too small (personal taste) and i HATE crosses in the 'top right corner' |
22:40.09 | lkcl | fine for a mouse. really awkward for touchscreens. |
22:40.16 | lkcl | anyway. enough celebration. |
22:40.48 | cr2 | ok. |
22:42.15 | psokolovsky__ | cr2, do you have an idea why sysfs doesn't show resources of platform_device's under /sys/devices/platform ? |
22:43.57 | lkcl | it would help if i knew the reset asic3 pin... |
22:44.17 | lkcl | ok - how am i gonna detect this... |
22:44.20 | doepiedoe | hey, how can i watch GPIOD5 with haret? |
22:44.37 | lkcl | specifically and only GPIOD5, doepiedoe |
22:44.38 | lkcl | ? |
22:44.41 | doepiedoe | yeah |
22:44.48 | lkcl | wasic3gpio <seconds> does everything |
22:44.49 | doepiedoe | want to check my sd card slot for activity |
22:44.54 | doepiedoe | yeah i know ;) |
22:44.59 | lkcl | on linux or wince? |
22:45.04 | doepiedoe | on wince |
22:45.06 | doepiedoe | via haret.... |
22:45.15 | cr2 | psokolovsky__: no idea. |
22:45.17 | doepiedoe | on linux should be nice to |
22:45.27 | lkcl | well you can always do a set trace 0xa920000 (or wherever it is) |
22:45.29 | psokolovsky__ | stupid linux ;-) |
22:45.37 | doepiedoe | 0x0020 ? |
22:45.39 | lkcl | and then wi 10 (for 10 seconds) |
22:45.41 | doepiedoe | set trace 0x0020 |
22:45.54 | lkcl | set trace 0xa9200180 or thereabouts |
22:46.13 | lkcl | check your asic3 address, add the offset for the register (see xda-dev ASIC3 page) |
22:46.26 | lkcl | then trace that memory address. that will give you D0 to D15 only |
22:46.32 | doepiedoe | offset == bit? |
22:46.48 | lkcl | http://wiki.xda-developers.com/index.php?pagename=ASIC3 |
22:47.00 | lkcl | find your device's physical mem address. |
22:47.28 | doepiedoe | <PROTECTED> |
22:47.29 | lkcl | convert it to virtual (check your device mem-map page) |
22:47.35 | doepiedoe | damn..... |
22:47.38 | lkcl | add 0x198 |
22:47.40 | lkcl | GPIOD_Status +0x0330 +0x198 Read here |
22:47.44 | lkcl | from the ASIC3 page |
22:47.58 | lkcl | and do set trace 0xxxx0198 |
22:48.20 | lkcl | then do some set iis because otherwise it's gonna get _really_ boring watching RTC interrupts etc. |
22:48.37 | lkcl | something like this: |
22:48.39 | lkcl | set ii(43) 1 |
22:48.39 | lkcl | set ii(7) 1 |
22:48.39 | lkcl | set ii(30) 1 |
22:48.39 | lkcl | set ii(10) 1 |
22:48.46 | doepiedoe | lol |
22:48.49 | lkcl | do a wi 2 first |
22:48.51 | doepiedoe | i really don't get it |
22:48.53 | lkcl | for a couple of seconds |
22:49.24 | lkcl | you're going to do a trace - a software watch - on the ASIC3 GPIOD 0 to 15 memory area |
22:49.55 | lkcl | any time this memory changes, the wi command will print out a debug log message telling you the ARM instruction that was accessing that area of memory. |
22:50.22 | doepiedoe | set trace 0x000198 |
22:50.32 | lkcl | what's your device called? |
22:50.33 | doepiedoe | set trace 0xc00198 |
22:50.39 | doepiedoe | HTC Universal |
22:50.40 | doepiedoe | :) |
22:50.45 | lkcl | ahh universal. ok. |
22:50.55 | lkcl | find the universal memory-map page. |
22:51.10 | doepiedoe | asic3 gpio base 0x10000000 |
22:51.31 | doepiedoe | GPIOD5 0x0020 SD card detect in/ou |
22:51.41 | lkcl | no, find the universal memory map page - on wiki.xda-developers.com |
22:51.44 | lkcl | <PROTECTED> |
22:51.57 | lkcl | <PROTECTED> |
22:52.00 | doepiedoe | 0xa9200000-a9300000 0x10000000-10100000 ASIC3 GPIO & misc |
22:52.04 | lkcl | sorry - yes, you're right. 0x10000000 |
22:52.10 | lkcl | ok. GREAT. |
22:52.12 | doepiedoe | lol |
22:52.16 | doepiedoe | how can i :P |
22:52.20 | doepiedoe | :) |
22:52.29 | doepiedoe | im always wrong |
22:52.30 | doepiedoe | :D |
22:52.31 | lkcl | so, translate that phys addr to virtual - 0xa9200000 |
22:52.36 | lkcl | burble :) |
22:52.40 | doepiedoe | :P |
22:52.49 | doepiedoe | set trace 0xa9200000 wi 2 |
22:52.58 | lkcl | ok, now look on the ASIC3 page again, for GPIOD 'read' offset. |
22:53.10 | lkcl | and then add that offset to 0xa9200000 |
22:53.11 | doepiedoe | thats 0x020 |
22:53.16 | lkcl | that's two separate commands |
22:53.23 | lkcl | the first command sets the address to watch |
22:53.30 | doepiedoe | <PROTECTED> |
22:53.34 | doepiedoe | jups |
22:53.36 | lkcl | the second command (wi 2) says 'start watching for two seconds' |
22:53.39 | doepiedoe | and then wi is watch ... |
22:53.43 | lkcl | ok where did you get 0x0020 from? |
22:53.58 | doepiedoe | from http://wiki.xda-developers.com/index.php?pagename=UniversalASIC3,GPIOD5 , SD card detect in/out |
22:54.46 | lkcl | ok that just tells you about D5. |
22:54.46 | lkcl | ok it tells you which bit of the 16 values, D0 to D15, in the memory address 0xa9200198, is D5. |
22:54.46 | doepiedoe | so 0xa9200020? |
22:54.48 | lkcl | no |
22:54.50 | lkcl | <PROTECTED> |
22:54.54 | doepiedoe | ahhhh |
22:55.05 | lkcl | D_Status is _all_ 16 GPIO D registers, as a bit-field. |
22:55.14 | lkcl | 1<<5 equals 0x0020 |
22:55.34 | lkcl | otherwise it's a hell of a waste of memory addressing |
22:56.17 | doepiedoe | so 0xa9200198 should be only for SD? |
22:56.29 | lkcl | ok. i gotta do an asic3 trace as well, get me a reset pin. |
22:56.42 | doepiedoe | GPIOD8 0x0100 radio reset |
22:56.45 | lkcl | no - 0xa9200018 is for ALL the GPIOD pins - all of them from 0 to 15 |
22:56.46 | doepiedoe | GPIOD7 0x0080 HW reboot |
22:57.00 | doepiedoe | GPIOB8 0x0100 radio reset? |
22:57.19 | lkcl | & all of those together, and you get 0xffff, right? |
22:57.34 | lkcl | that means you have 16-bits at that address. |
22:57.45 | lkcl | come on, dude - this is standard computer science stuff! |
22:58.07 | doepiedoe | lol |
22:58.24 | lkcl | ok lemme find you some code to look at... |
22:58.46 | doepiedoe | i just wanted to know if i could detect if my sd card responds..... |
22:58.57 | doepiedoe | but this isnt the right way i think :) |
22:59.41 | lkcl | the easiest way is to do WASIC3GPIO <seconds> and then grep the output looking for D5 |
23:00.22 | doepiedoe | allright! |
23:00.24 | doepiedoe | thats perfect |
23:00.25 | doepiedoe | GPIOD5 S 0 SDIO_DET 1606568 |
23:00.29 | doepiedoe | GPIOD5 S 1 SDIO_DET 1608299 |
23:00.33 | doepiedoe | so it does something |
23:00.33 | doepiedoe | ;) |
23:00.38 | lkcl | that's the stuff |
23:00.47 | lkcl | ok. |
23:01.11 | doepiedoe | :) thanks lkcl |
23:01.20 | lkcl | i gotta do some wi tracing because i am looking for some very fast 'reset' that is well below the poll threshold of WASIC3 oops... |
23:01.25 | lkcl | no problem dude |
23:03.05 | doepiedoe | your talking arabic to me |
23:03.14 | doepiedoe | ;) |
23:03.59 | doepiedoe | you know if theres any way to flash the wince image to a flash card>? |
23:06.06 | lkcl | yes there is - using the wince bootloader. |
23:06.14 | lkcl | d2s i think is the command you're looking for. |
23:06.26 | lkcl | you'll need at least a 256mb flash card |
23:06.35 | doepiedoe | ahhh ;) |
23:07.06 | doepiedoe | got a and there was a tool to talk via usb |
23:07.14 | doepiedoe | mtty ;) |
23:08.41 | lkcl | bloody hell, what you using windows for, dude??? |
23:08.51 | lkcl | go download kubuntu 6.10 _right_ now |
23:08.57 | doepiedoe | don't like linux on a desktop |
23:08.57 | doepiedoe | ;) |
23:08.59 | doepiedoe | lol |
23:09.08 | doepiedoe | only use it for servers |
23:09.23 | lkcl | ahh, then you haven't tried kubuntu with my kroller.sez, then, have you |
23:09.33 | doepiedoe | can't use photoshop.... no outlook express.... ;) |
23:09.37 | lkcl | http://www.kde-look.org/content/show.php?content=31081 |
23:09.42 | doepiedoe | got kubuntu with xgl on my work on 6 displays :) |
23:10.00 | lkcl | you _can_ use photoshop!! actually - you can use a much better program called xara xtreme |
23:10.06 | doepiedoe | yeah wine |
23:10.07 | doepiedoe | ;) |
23:10.07 | lkcl | SIX??? are you MAD?? :) |
23:10.14 | doepiedoe | ;) |
23:10.20 | doepiedoe | 3 different pc's |
23:10.25 | doepiedoe | with 2 monitors |
23:10.30 | doepiedoe | with xgl 3d |
23:10.44 | doepiedoe | but only use it for monitoring purposes |
23:10.50 | doepiedoe | kind of for show thingie |
23:11.03 | doepiedoe | i'll make a picture of it tomorrow :) |
23:11.06 | doepiedoe | looks nice |
23:11.34 | doepiedoe | but we have a lot of windows machines at work |
23:11.47 | lkcl | kde 3.5.x is a _big_ step up - but only if you use things like kroller.sez (a macosx program-bar) |
23:11.58 | lkcl | as the kde task bar is SHIT |
23:12.28 | doepiedoe | yeah ;) |
23:12.30 | doepiedoe | really.... |
23:12.41 | doepiedoe | but to be honest, i really like toe look of XP |
23:12.41 | doepiedoe | ;) |
23:12.43 | doepiedoe | the |
23:12.44 | doepiedoe | ... |
23:13.24 | lkcl | you are sick. get to the doctor's _right_ now. |
23:13.30 | lkcl | ask for a CAT scan |
23:13.34 | lkcl | quick |
23:13.35 | doepiedoe | haha |
23:13.35 | doepiedoe | ;) |
23:13.48 | lkcl | darn it's not that one... *sigh*... |
23:13.51 | doepiedoe | you know whats sick ;) |
23:13.55 | lkcl | gotta expand the watch area |
23:13.58 | doepiedoe | we use windows as host for vmware machines... |
23:14.07 | lkcl | yeh, using fvwm2 for 7 years |
23:14.17 | lkcl | that's not sick, it's sad |
23:14.44 | doepiedoe | http://www.ramsinks.com/images/bartixbig.jpg |
23:14.48 | doepiedoe | this looks nice 2me |
23:16.25 | Kevin2 | cr2: I'll add in a description of "Default" for the generic case. I guess this means we found the problem - I'll see if I can get the cegcc guys to fix it. |
23:16.31 | Kevin2 | cr2:BTW, does it boot linux? |
23:18.14 | cr2 | Kevin2: i have not tried, but i can. btw, why is the machine not detected ? |
23:19.23 | cr2 | hmm. it will not work. |
23:19.34 | cr2 | because of the famous OHCI clock. |
23:19.43 | Kevin2 | There is no machine class defined for your platform. |
23:20.05 | Kevin2 | So, it falls back to the default no-op class. |
23:20.19 | cr2 | i need to disable OHCI clock in the companion SA-1101 chip. and the busmaster DMA. |
23:20.59 | Kevin2 | Do you know the detection code for the platform? If so, we should add a new class that detects the platform and does the necessary hardware stuff. |
23:21.43 | cr2 | can you print the wince IDs in the debug window, or log ? |
23:22.30 | cr2 | otherwise i need to look for the PCMCIA ethernet card, eth hub, cable and so on. |
23:22.48 | Kevin2 | cr2: They should be there already. |
23:22.59 | cr2 | where ? |
23:23.07 | Kevin2 | Hrmm. I guess not - they'll print if you telnet it, however. |
23:23.57 | cr2 | add Log() there. |
23:24.16 | cr2 | it's a good general feature. |
23:24.44 | lkcl | cr2 - what's this A0 and A1 'sound related' in SableAsic3? |
23:24.59 | cr2 | lkcl: they are ref's in wavedev |
23:25.37 | Kevin2 | cr2: Yeah, I'll have it in the next rev. |
23:25.40 | cr2 | check the headphone. hmm, you don't have it ;) |
23:25.54 | lkcl | ahh joy. |
23:25.57 | lkcl | ok i'll try some/ |
23:26.23 | Kevin2 | The welcome prompt when you telnet in will have. Is this okay, or do you need a new haret.exe? |
23:26.27 | lkcl | random bit-flipping... *sigh* |
23:26.40 | cr2 | lkcl: forcing powerdown on universal didn't show the gpio change. |
23:26.56 | cr2 | Kevin2: i'd like to have it in the debug window. |
23:27.07 | cr2 | otherwise i need to setup ethernet. |
23:27.09 | lkcl | yes i noticed that, cr2. |
23:27.12 | lkcl | same here. |
23:27.39 | cr2 | Kevin2: and there are no wifi drivers in wince. |
23:28.14 | cr2 | and pcmcia is 16bit, and only very special pcmcia cards are supported by wince ;) |
23:28.57 | Kevin2 | Okay - grab haret.exe-gccmain rev 129. |
23:29.03 | cr2 | thanks :) |
23:29.16 | lkcl | brb - baby-feeding. then i gonna try the 'random guess' approach to finding the right pins... |
23:30.12 | Kevin2 | Is it possible to detect this is a "strongarm" chip and then react accordingly? Or is it better to just code up for your particular machine? |
23:32.49 | cr2 | Kevin2: you can detect strongarm. but the machine-specific hooks are also needed. |
23:33.16 | cr2 | ok. it crashed after Found machine. and didn't show anything |
23:33.51 | cr2 | btw, there is one more visual glitch. the program icon is not detected. but i had this problem with vlc too. |
23:40.51 | Kevin2 | cr2: It crashed this time? That is odd. |
23:41.16 | *** join/#htc-linux florian_ (n=fuchs@87.193.64.133) |
23:42.45 | Kevin2 | Did it show: Trying to detect machine (Plat='%ls' OEM='%ls') |
23:43.10 | Kevin2 | It'll be in the haretlog.txt file. |
23:43.18 | psokolovsky | cr2, well, so there appears to be some issue on DMA-I2S link |
23:43.38 | cr2 | Kevin2: no. |
23:43.54 | Kevin2 | What was the last line in the file? |
23:44.29 | cr2 | Detected machine executing startup.txt |
23:45.44 | cr2 | psokolovsky: ok. another question: can i get the &foorbar_asic3.dev pointer in the child device from asic3 somehow ? |
23:46.01 | Kevin2 | It should be much earlier in the log. Before we print all the "looking at machine xxx" logs. |
23:46.34 | psokolovsky | cr2, there're all kind(s) of parent links |
23:46.36 | cr2 | Kevin2: i'll check again. the evil pword.exe does not do line breaks in this file ;) |
23:47.13 | Kevin2 | cr2: If you can use "total commander", I'd recommend it. (It has a simple file viewer that works well.) |
23:47.21 | psokolovsky | cr2, so I have no idea what can be wrong with ak driver |
23:47.46 | psokolovsky | cr2, so, it just hangs in DMA xfer, and being killed by timeout by ALSA |
23:48.31 | psokolovsky | and fsck, we done even have /proc/dma, to look what channels are in use at all |
23:49.57 | cr2 | Kevin2: it was my fault. now it works. copying the file. |
23:50.29 | cr2 | psokolovsky: strange. then the DMA descriptors should be damaged or something. |
23:50.31 | psokolovsky | cr2, can you point me to source which was known to work with haret? |
23:50.44 | *** join/#htc-linux Kmarc (i=kari@markos.biz) |
23:51.10 | psokolovsky | cr2, damaged descriptors would get fault irq, I guess |
23:51.26 | cr2 | psokolovsky: i'll try to find it in my e-mail. |
23:52.13 | psokolovsky | cr2, I mean, if you can tell me what tag in xanadux repo marks known working code, ... |
23:52.56 | cr2 | -un0 |
23:54.37 | psokolovsky | ok, will have a look after peering at mainline git first |
23:54.53 | cr2 | Kevin2: Trying to detect machine (Plat='Jupiter' OEM='HP, Jornada 820,, 8262, German, SAA') |