**** BEGIN LOGGING AT Wed Oct 22 02:59:57 2008 Oct 22 10:43:26 BobOfDoom: how GSM_POWER behaves? Oct 22 10:43:38 it's low all the time Oct 22 10:43:57 but when using GSM, it's high? Oct 22 10:44:02 and GSM_WAKE? Oct 22 10:44:26 just switching on/off from time to time? Oct 22 10:53:02 in linux-hnd? Oct 22 10:53:16 IIRC I made it power up the GSM when the serial port is opened by some userspace app Oct 22 10:53:36 can't do this in 2.6.277 as there's no hooks Oct 22 10:54:08 BobOfDoom: I want to find that GPIOs Oct 22 10:54:24 I'm asking about its behaviour to find it easily Oct 22 10:54:31 ah Oct 22 10:54:32 nah Oct 22 10:54:34 it stays on Oct 22 10:54:37 while the GSM is no Oct 22 10:54:39 both? Oct 22 10:54:58 sorry my brains not working today Oct 22 10:55:05 * BobOfDoom stops and rereads what Sleep_Walker said Oct 22 10:55:09 right Oct 22 10:55:10 yes Oct 22 10:55:18 damn Oct 22 10:55:20 GSM_POWER is high when phone is on, low when phone is off Oct 22 10:55:52 GSM_WAKE high, send some bytes, GSM_WAKE low Oct 22 10:55:54 hmm Oct 22 10:56:00 I can't remember how I found it now Oct 22 10:56:19 it might have been that it wasn't configured as an output until the phone was on Oct 22 10:56:41 :( Oct 22 10:57:11 Sleep_Walker: but it may not exist on the 680, it might be different with USB Oct 22 10:57:24 have you tried that AT+CFUN=1 ? Oct 22 10:57:24 well, who knows if there is USB Oct 22 10:57:28 oh Oct 22 10:57:33 I thought you said you found it Oct 22 10:57:37 I have tried AT+CFUN=0 Oct 22 10:57:39 and could do AT and get OK Oct 22 10:57:43 cause I didn't remembered :b Oct 22 10:57:52 and got no response Oct 22 10:58:04 AT gives OK Oct 22 10:58:09 but AT+CFUN=0 gives no response? Oct 22 10:58:11 I can get about 4 ERROR lines Oct 22 10:58:18 in good case Oct 22 10:58:30 (probably reaction for minicom initialization) Oct 22 10:58:38 oh Oct 22 10:58:54 but I got AT/OK dialog only twice yet Oct 22 10:59:16 so I'm trying to find proper GPIOs :/ Oct 22 10:59:40 hmm Oct 22 11:00:06 and I'm not sure what can I do then :b Oct 22 11:00:22 I mean - with GSM_WAKE, but without GSM_POWER :b Oct 22 11:00:40 how can I indicate that GSM_POWER is low? Oct 22 11:00:44 (on UART) Oct 22 11:01:06 does Treo650 have red led? Oct 22 11:01:43 it has a green flashing LED.. but thats setup by software (programmed flash speed) via ASIC Oct 22 11:02:01 hmmm :( Oct 22 11:02:10 red flashing LED == connecting to the network Oct 22 11:02:18 green flashing LED == connected Oct 22 11:02:24 uh Oct 22 11:02:30 so it has red led Oct 22 11:02:47 yes Oct 22 11:02:55 but it's just conencted to the ASIC Oct 22 11:02:58 same as green LED Oct 22 11:03:00 ok Oct 22 11:03:43 BobOfDoom: I'm not sure if I have all that GPIOs of FFUART set correctly Oct 22 11:03:43 oh.. hey.. my 680 is not resetting itself today Oct 22 11:03:47 :) Oct 22 11:04:44 there's also power controls in USB host controller IIRC Oct 22 11:04:49 they may or may not be used Oct 22 11:05:16 when I have a look on GPIO Alternate Function Table, there are multiple FF* on different places Oct 22 11:05:55 hrmm Oct 22 11:06:07 9c000054: 0x103 Oct 22 11:06:15 9c000058: 0x100 Oct 22 11:06:18 9c00005c: 0x100 Oct 22 11:06:32 host port 1, 2, 3 respectively Oct 22 11:06:35 FFRXD, FFTXD, FFCTS, FFRTS, FFDTR, FFDSR, FFDCD, FFRI... Oct 22 11:06:37 but 1 means something is connected Oct 22 11:06:41 bit* Oct 22 11:06:48 BobOfDoom: on Treo680 Oct 22 11:06:49 ? Oct 22 11:06:52 BT Oct 22 11:06:54 yes Oct 22 11:06:59 thats with BT and phone on Oct 22 11:07:04 * BobOfDoom tries turning BT off Oct 22 11:07:28 port 2 is USB client Oct 22 11:07:37 port 3 is still empty for me :( Oct 22 11:07:40 yes... Oct 22 11:07:46 and port 1 goes off when you turn off BT Oct 22 11:08:20 that's weird Oct 22 11:08:35 that they power port 3? Oct 22 11:08:49 unless they power off USB while it's idle Oct 22 11:08:49 hmm Oct 22 11:09:25 is it controlable in linux userspace ? Oct 22 11:10:13 there's no userspace interface AFAIK Oct 22 11:10:29 I was messing with it just by mmaping /dev/mem and writing directly to the register Oct 22 11:10:36 you can use that devmem2.c tool Oct 22 11:10:38 to do that Oct 22 11:10:54 ./devmem2 0x5c0000xx w 1234 Oct 22 11:11:00 or whatever Oct 22 11:11:05 I see Oct 22 11:12:19 hmm Oct 22 11:12:25 UP3OCR = 0 in palmos Oct 22 11:12:28 this is really strange Oct 22 11:12:43 maybe that driver is for something else Oct 22 11:13:02 or maybe only certain revision of the 680 have phone connected via serial Oct 22 11:13:03 hmm Oct 22 11:13:12 I need some palmos terminal emulator Oct 22 11:14:04 UP3OCR = 0; what is wrong about that? Oct 22 11:14:40 it just means its not configured as host Oct 22 11:15:22 0x0 = Host controller port 3 transceiver with external host controller Oct 22 11:15:48 a bit cryptic sentence Oct 22 11:16:01 oh.. Oct 22 11:16:26 oh Oct 22 11:16:30 good opint Oct 22 11:16:36 I was looking at the wrong ting Oct 22 11:16:49 UP2OCR has different bits... Oct 22 11:17:02 yep Oct 22 11:26:46 BobOfDoom: I pushed to linux-hnd my current status to sync with you Oct 22 11:27:14 GSM_* GPIOs are not known for sure... Oct 22 11:28:08 thanks :) Oct 22 11:28:25 hmm Oct 22 11:28:30 * BobOfDoom has an idea for experiment Oct 22 11:29:01 I'll stuff up the USB host settings, then call it and see it rings :p Oct 22 11:31:50 yeah, ok I'm pretty sure its not on USB Oct 22 11:31:56 that's weird Oct 22 11:32:07 I wonder why tehre's a driver then Oct 22 11:32:37 maybe it's something to do with bluetooth headsets or whatever Oct 22 11:36:35 althoguh Oct 22 11:36:39 waut you talked to it? Oct 22 11:36:43 man I'm so confused Oct 22 11:37:08 AT Oct 22 11:37:09 OK Oct 22 11:37:13 oh Oct 22 11:37:17 you talked to it on FFUART not USB? Oct 22 11:37:17 only successfull dialog Oct 22 11:37:19 aaah Oct 22 11:37:21 yes Oct 22 11:37:23 I misunderstood Oct 22 11:37:23 ha Oct 22 11:37:26 ok Oct 22 11:37:36 what baud did you use? Oct 22 11:37:37 there is no device for dialog Oct 22 11:37:42 460800 Oct 22 11:37:46 same as you Oct 22 11:37:50 k Oct 22 11:38:22 do you have USB networking working? Oct 22 11:38:53 yes Oct 22 11:39:03 excellent Oct 22 11:39:04 I disabled USB Host on port 2 Oct 22 11:39:14 * BobOfDoom clones linux-hnd Oct 22 11:45:21 30 kb/s Oct 22 11:45:28 that's gonna take a while Oct 22 11:45:54 sheesh.. traffic to the US is really slow here in the evenings Oct 22 11:48:01 it's not hosted on US, is it? Oct 22 11:48:18 s/on US/in US/ Oct 22 11:50:14 bb in hour or so Oct 22 11:50:52 k Oct 22 11:50:58 I'm pretty sure thor is in the US Oct 22 11:57:50 IP address registered in Seattle Oct 22 12:02:23 btw. I suggest not to use palmt680-gsm for now Oct 22 12:02:39 no? Oct 22 12:02:44 it enables pxa ffuart Oct 22 12:02:59 so you can do everything manually by yourself Oct 22 12:03:06 (in combination with gpioed) Oct 22 12:03:22 ah ok Oct 22 12:03:35 and palmt680-gsm try with found GPIOs values :b Oct 22 12:03:44 did you commit it? Oct 22 12:03:52 yes Oct 22 12:03:53 is it in mach-pxa/palmt680/ Oct 22 12:03:54 hmm Oct 22 12:03:55 setl 79 && seth 79 &&setl 87 40 39 34 && seth 87 40 39 34 && set2 41 && set1 34 && seti 35 34 14 Oct 22 12:04:02 this does something Oct 22 12:04:07 oh Oct 22 12:04:08 right Oct 22 12:04:13 it's jsut in palmt6580.c Oct 22 12:04:15 gotcha Oct 22 12:04:26 setl X = echo "l X" > /proc/gpio Oct 22 12:04:34 seth X = echo "L X" > /proc/gpio Oct 22 12:04:48 seti X = echo "d X" > /proc/gpio Oct 22 12:04:55 seto X = echo "D X" > /proc/gpio Oct 22 12:05:00 set2 X = echo "2 X" > /proc/gpio Oct 22 12:05:04 set1 X = echo "1 X" > /proc/gpio Oct 22 12:05:08 set0 X = echo "0 X" > /proc/gpio Oct 22 12:05:33 (since I haven't good keymap, I prefer use short commands :b ) Oct 22 12:06:12 I'm not sure now with d/D, but it should be obvious what it is Oct 22 12:06:18 and now really bbl :) Oct 22 12:06:45 :) Oct 22 12:51:57 back Oct 22 12:54:13 hrmm Oct 22 12:54:18 cocoboot is just resetting Oct 22 12:54:23 :( Oct 22 12:55:29 BobOfDoom: I'll send you my cocoboot Oct 22 12:55:49 Sleep_Walker: did you modify it to work better? Oct 22 12:56:19 I found that cocoboot is reseting quite often, but it gets better after hard reset Oct 22 12:56:39 also, if it resets - run cocoboot which will remain in RAM Oct 22 12:56:48 it helps somehow Oct 22 12:56:56 and Oct 22 12:57:32 after one successfull boot if you run cocoboot from RAM it makes next boot run also fine Oct 22 12:57:50 O.o Oct 22 12:59:23 I also reduced kernel+initrd size to 1.2 megs which helped Oct 22 13:01:08 http://sleepwalker.hackndev.com/cocotreo.diff Oct 22 13:01:15 this is modification I made Oct 22 13:01:23 I can't say if it helps much Oct 22 13:01:44 ah Oct 22 13:01:54 I'm not using initrd anyway Oct 22 13:01:56 gah Oct 22 13:02:03 it's doing it's resetting thing again now Oct 22 13:02:07 (reset loop) Oct 22 13:02:15 ROM version? Oct 22 13:02:24 some AT&T one Oct 22 13:02:26 I dunno Oct 22 13:02:30 I think it's recent Oct 22 13:02:35 how do you tell? Oct 22 13:02:44 ooh it's alive again Oct 22 13:02:48 there is some phone info in phone application Oct 22 13:02:50 no... gone Oct 22 13:02:54 must be when it warms up Oct 22 13:03:06 I'll leave it sit for a while Oct 22 13:03:48 and yeah - that resets because of cocoboot are really fucked :( Oct 22 13:04:14 Sleep_Walker: yea ... and BobOfDoom originally wrote cocoboot ;) Oct 22 13:04:29 Marex: why originally? Oct 22 13:04:35 it's his work.. Oct 22 13:05:07 I can't pursuit bob for not working on Tre680 - he hadn't any... Oct 22 13:05:09 (translated - BobOfDoom puvodne napsal cocoboot) Oct 22 13:05:51 Sleep_Walker: it wasnt nice comment from you basically ... Oct 22 13:06:47 Marex: it's ok, he's right, those resets are really annoying Oct 22 13:07:10 BobOfDoom: that's rather palmos's issue Oct 22 13:07:23 yeah Oct 22 13:07:24 but anyway - wont it help relocating the kernel with mmu off ? Oct 22 13:07:55 ie. you load the kernel into ram while still in palmos's protected mode, turn off mmu and move kernel to 0x8000 ? Oct 22 13:08:05 I think that's what it does Oct 22 13:08:07 I dont recall how it works now Oct 22 13:08:10 * BobOfDoom checks Oct 22 13:08:28 * Marex is at school Oct 22 13:08:45 Marex: why it wasn't nice comment? Oct 22 13:08:45 (btw. I found anime shop nearby - I already bought two figures and a tshirt) Oct 22 13:09:01 it's his work = Bob wrote cocoboot Oct 22 13:09:17 Sleep_Walker: just ... forget what I said Oct 22 13:09:18 as I said it's not Bob's fault Oct 22 13:09:20 yep, that's what it does, disables interrupts and MMU then copies it into place Oct 22 13:09:44 BobOfDoom: can pxa set identities in memory ? Oct 22 13:09:50 Marex: maybe my English is bad and I said something else Oct 22 13:10:06 Marex: you mean identity mapping? Oct 22 13:10:07 Sleep_Walker: even your czech is Oct 22 13:10:07 Marex: I was just curious about your word originally Oct 22 13:10:24 Marex: you're as kind as asshole as usual Oct 22 13:10:24 Sleep_Walker: use dictionary please ;) Oct 22 13:10:36 Sleep_Walker: same goes for you ... Oct 22 13:10:40 BobOfDoom: right Oct 22 13:10:56 BobOfDoom: dunno if linux would be able to cope with it though Oct 22 13:10:57 Marex: great chat... thanks :) Oct 22 13:11:09 Sleep_Walker: no problem .... Oct 22 13:11:10 yeah Oct 22 13:11:26 Linux expects MMU to be off and turns it on Oct 22 13:11:34 I dunno what happens if you turn it on while it's already on Oct 22 13:11:40 BobOfDoom: you know what I mean, right ... mapping 0x0 to identity with some other page, which would be already preloaded with linux kernel Oct 22 13:12:05 well that was discussed some time ago on lakml iirc Oct 22 13:12:09 * Sleep_Walker goes cook Oct 22 13:12:13 interesting Oct 22 13:12:14 you have to comment some stuff in boot code which is no-go Oct 22 13:12:23 yeah Oct 22 13:12:35 s/\(comment\)/\1 out/ Oct 22 13:13:03 BobOfDoom: so it seem they patched palmos's memory manager ? Oct 22 13:13:22 I doubt there is some thread checking if memory is in consistent state Oct 22 13:13:35 I think what we may ahve to end up doing is just to add tiny MMC driver to cocoboot and load the zImage with the MMU off so palmso can't get in the way and reset us Oct 22 13:14:13 BobOfDoom: tiny pxamci ? you really are crazy ;) Oct 22 13:14:24 wow, another wbsd driver for Palm TT3? D: Oct 22 13:14:24 I haven't looked at it Oct 22 13:14:25 :D Oct 22 13:14:26 is it Oct 22 13:14:28 really bad? Oct 22 13:14:39 you are talking about pxamci here ... THAT pxamci everyone hates ... Oct 22 13:14:50 XD Oct 22 13:14:51 dang Oct 22 13:15:09 yea ... it has loads of hardware issues itself Oct 22 13:15:12 or try and get u-boot working and use it's MCM driver Oct 22 13:15:17 MMC& Oct 22 13:15:18 there is already a pile of workarounds in the driver Oct 22 13:15:19 bleh Oct 22 13:15:34 BobOfDoom: you think you can load small binary like uboot then ? Oct 22 13:16:37 I dunno whether that's the 680 problem Oct 22 13:16:47 but that'd help with the kernel size issue in general Oct 22 13:17:00 I'll have to play more with 680 and figure out what's going on Oct 22 13:17:11 I've really no idea where the problem is atm Oct 22 13:17:31 BobOfDoom: btw. seen this http://thread.gmane.org/gmane.linux.kernel/744513 ? Oct 22 13:17:53 ha Oct 22 13:17:54 yeah Oct 22 13:17:55 :b Oct 22 13:17:56 I saw it in LWN Oct 22 13:18:03 :D Oct 22 13:18:10 BobOfDoom: yes, but you'll need individual binary of uboot for each and every device Oct 22 13:18:17 BobOfDoom: jeez ... it hit even LWN ?! Oct 22 13:18:41 Marex: you're popular vibrator guy Oct 22 13:19:13 Marex: http://lwn.net/Articles/303276/ Oct 22 13:19:21 oh my goddess ... Oct 22 13:20:02 Im leaving for a class. bye Oct 22 13:22:58 gnome-cups-manager README :D http://dirac.uos.ac.kr/cgi-bin/dwww?type=file&location=/usr/share/doc/gnome-cups-manager/README Oct 22 13:23:54 ... Oct 22 13:24:08 nice Oct 22 13:25:26 * BobOfDoom wonders how long til someone starts demanding Android port in hackdev forums now that the source has been released Oct 22 13:25:55 is android in OE? Oct 22 13:26:20 it seems not Oct 22 13:26:36 that can be good answer :) Oct 22 13:26:53 no Oct 22 13:27:00 it uses it's own build stuff Oct 22 13:27:07 it might use maven actually Oct 22 13:27:14 it's mostly java stuff Oct 22 13:27:23 android.kernel.org Oct 22 13:27:33 I haven't really looked much yet Oct 22 13:28:02 I can't imagine it will work with my Treo680 and it will be usable for calling... Oct 22 13:28:24 it'd probably boot Oct 22 13:28:37 but yeah, you'd need to write a driver for the GSM probably Oct 22 13:28:45 oh Oct 22 13:28:52 actually it needs ahacked kernel IIRC Oct 22 13:29:04 they added a bunch of extra power management and other stuff that it requires Oct 22 13:29:19 meh Oct 22 13:29:20 ok Oct 22 13:29:22 I'm off Oct 22 13:29:23 gn Oct 22 13:30:01 gn Oct 22 13:43:13 BobOfDoom: AT+CFUN=1 got OK Oct 22 13:43:34 +CREG: 2 Oct 22 13:43:36 +CGREG: 2 Oct 22 13:43:37 *MRDY: 4 Oct 22 13:43:39 *MTSMENU: "Vodafone", 0, 3 Oct 22 13:43:40 1:"Info&Zabava",0 Oct 22 13:43:42 2:"M-Banka",0 Oct 22 13:43:43 3:"M-Platby",0 Oct 22 13:43:45 *MTZ: 1, "VodafoneCZ" Oct 22 13:43:47 *MTZ: 2, "10/22/2008, 15:43:08+08", 1 Oct 22 13:43:48 +CREG: 1 Oct 22 13:43:51 +CGREG: 1 Oct 22 13:43:55 *MTZ: 1, "VodafoneCZ" Oct 22 13:43:56 :) Oct 22 15:43:32 yo Oct 22 16:34:07 bye Oct 22 20:43:00 *MRDY: 1 Oct 22 20:44:16 *MRDAT: 2 Oct 22 20:48:52 marex: heh, alone with your vibrator? Oct 22 20:49:30 :E~ Oct 22 20:49:46 strange behaviour of flow control, but it seems to work Oct 22 20:52:22 lucky you have plain text communication, not everyone is so lucky Oct 22 20:53:14 not all the time Oct 22 20:58:11 QSt� Oct 22 20:58:24 it doesn't look like plain text :/ Oct 22 21:02:13 dump in hexa Oct 22 21:02:17 od -x or hd Oct 22 21:11:32 MRDY = Message Ready? Oct 22 21:12:56 module ready status Oct 22 21:20:20 modem ready Oct 22 21:22:53 *MRDY:1 Module is ready. Oct 22 21:23:13 hmm, dialing shows "NO CARRIER" Oct 22 21:26:27 RING :D Oct 22 21:33:14 marex: do you know if it is configurable which pins are used for UART? Oct 22 21:33:55 I would think it is not, but I can't understand multiple pins in GPIO alternate function table (PXA27x manual) Oct 22 21:34:31 I'd like to find/setup flow control pins... Oct 22 21:34:43 (I can't even find them in sources :/ ) Oct 22 21:34:45 set it to software Oct 22 21:34:54 pxa doesnt have those pins on uarts iirc Oct 22 21:35:02 or maybe ... errr Oct 22 21:35:15 it does have, but they should be preset by linux kernel itself Oct 22 21:36:02 if it is configurable - I haven't found configuration part in PXA27x manual Oct 22 21:36:22 if it is not, I'd like to see code where it is set up :/ Oct 22 21:36:34 well, I want to see it in any case :b Oct 22 21:42:16 see mfp_pxa27x.h or something Oct 22 21:42:36 (just my guess, it'd be somewhere around - XXUART_RTS XXUART_CTS) Oct 22 21:44:22 (and the code will be in some cpu init routine) Oct 22 21:44:24 gnight Oct 22 22:07:43 Sleep_Walker: :) congrats Oct 22 22:07:55 BobOfDoom: thanks Oct 22 22:08:01 it seems that I have power Oct 22 22:08:15 but I'm not sure about wak Oct 22 22:08:18 wake Oct 22 22:08:33 so does it stop responding after a while (a minute or two) if you leave it idle after doing CFUN=1? Oct 22 22:09:17 they could have changed it to use CTS/RTS instead of wake I suppose as well Oct 22 22:09:18 it doesn't matter which command is called, it just falls asleep Oct 22 22:09:29 ah ok Oct 22 22:09:52 I wonder if they use other pins Oct 22 22:09:55 FFRI Oct 22 22:10:12 FFDCD, FFDTR Oct 22 22:10:15 etc Oct 22 22:18:05 I've got *MRDY: 1 or *MRDY: 4 Oct 22 22:18:38 (1=Module is ready,4=SIM card removed) Oct 22 22:29:41 I thought 4 was SIM card inserted Oct 22 22:30:44 http://eng.euromobile.ru/files/documents/EDGE_EE54_AT_Commands_Rev_C.pdf page 31 Oct 22 22:31:13 it looks like it lost sync from time to time Oct 22 22:31:23 and I start to receive "QSt�" Oct 22 22:31:49 or it is some mode I don't know Oct 22 22:32:54 hmm Oct 22 22:33:18 I think that doc must be wrong Oct 22 22:33:33 cause you got MTSMENU right? Oct 22 22:33:42 which means it must have read the SIM car Oct 22 22:33:43 d Oct 22 22:34:07 agreed Oct 22 22:34:35 I dunno what QSt? means, I've never seen that Oct 22 22:38:03 pushed Oct 22 22:51:26 heh Oct 22 22:51:57 I'm not sure if that strange intervals of inactivity aren't caused by UDC code and ssh Oct 22 22:51:59 :b Oct 22 22:54:52 I've got *MRDY: 4 and *MTSMENU at the same moment :/ Oct 22 22:55:20 ATD reports ERROR Oct 22 22:55:37 ERROR ? Oct 22 22:55:43 that's odd Oct 22 22:55:45 yes Oct 22 22:55:58 usually it tells you more than that Oct 22 22:56:07 and I can't call in most of tries Oct 22 22:56:45 nothing but error Oct 22 22:57:23 maybe you have to force registration Oct 22 22:57:24 hmm Oct 22 22:57:27 but it gave you the timezone Oct 22 22:57:28 weird Oct 22 22:57:58 < Sleep_Walker> +CREG: 1 Oct 22 22:58:03 that means it registered Oct 22 22:58:18 +CREG: 2 Oct 22 22:58:29 it changes unpredictably Oct 22 22:59:47 "not registered, but ME is currently searching a new operator to Oct 22 22:59:47 register to Oct 22 22:59:49 " Oct 22 23:00:12 QH8� Oct 22 23:00:14 hmmm Oct 22 23:00:30 I still miss something important Oct 22 23:07:36 do I have to unlock something? Oct 22 23:08:40 "AT+CPIN?" (including the question mark) will tel you if you need to unlock Oct 22 23:08:50 thx Oct 22 23:08:58 were you successful in booting? Oct 22 23:09:38 nah, haven't tried today, just got up. I won't have a chance to try until tonight Oct 22 23:11:37 ok Oct 22 23:13:55 AT+CPIN? Oct 22 23:13:57 ERROR Oct 22 23:15:59 can your palmt6*-gsm code affect CTS/RTS so minicom lost sync? Oct 22 23:16:32 I dunno Oct 22 23:16:35 is it echoing ok? Oct 22 23:16:45 that's weird Oct 22 23:17:28 no Oct 22 23:17:28 I've got set up local echo Oct 22 23:18:52 ooh Oct 22 23:18:56 you're getting no echo? Oct 22 23:19:04 try issueing ATE1 Oct 22 23:19:55 that should turn on echo Oct 22 23:20:03 should then be able to see if you're losing characters or something Oct 22 23:20:33 btw this is of course with AT+CFUN=1 .. it will return ERROR to just about everything if AT+CFUN=0 Oct 22 23:26:29 oh Oct 22 23:26:38 it seems that it timeouted Oct 22 23:26:44 it switched offline Oct 22 23:26:48 and online again Oct 22 23:26:53 and then there was Oct 22 23:27:00 QSt� Oct 22 23:27:18 ooh Oct 22 23:29:11 heh Oct 22 23:29:22 palmt680_gsm_set_txrx: timed out waiting for host wake Oct 22 23:29:23 palmt680_gsm_configure: power off Oct 22 23:29:25 palmt680_gsm_configure: power on Oct 22 23:29:26 palmt680_gsm_set_txrx: timed out waiting for host wake Oct 22 23:29:31 dmesg is full of that Oct 22 23:29:38 maybe wrong gpios Oct 22 23:29:39 ah Oct 22 23:32:44 GEDR? Oct 22 23:33:07 edge detect status register Oct 22 23:33:36 I know that GPIO changes Oct 22 23:33:54 when 87 is low, 14 is also low Oct 22 23:34:13 when 87 is high, 14 is high and can fall Oct 22 23:40:55 I don't understand much that code :/ Oct 22 23:41:15 _gsm ? Oct 22 23:41:18 yup Oct 22 23:41:20 kill it Oct 22 23:41:35 hm Oct 22 23:41:44 maybe I can learn from that Oct 22 23:42:01 I don't even remember that reset stuff Oct 22 23:42:02 txrx is bool? Oct 22 23:42:02 that's weird Oct 22 23:42:20 in set_txrx Oct 22 23:42:31 ooh Oct 22 23:42:37 right Oct 22 23:42:44 so it waits until it wakes pup Oct 22 23:42:45 hm Oct 22 23:42:49 yes Oct 22 23:42:52 yes Oct 22 23:42:57 oh wait Oct 22 23:42:58 no Oct 22 23:43:05 it's flas Oct 22 23:43:06 but it waits for GSM_WAKE Oct 22 23:43:06 not bool Oct 22 23:43:20 int new_wake = (txrx & PXA_SERIAL_TX) ? 1 : 0; Oct 22 23:43:29 flags* Oct 22 23:43:50 new_wake can be 1 or 0 Oct 22 23:44:02 if new_wake is zero - bye Oct 22 23:44:03 right Oct 22 23:44:20 if new_wake is 1 and old_wake is 0 Oct 22 23:44:25 wait for wake Oct 22 23:44:41 it sets GSM_WAKE to 1 Oct 22 23:44:50 (new_wake is funny cryptic :b ) Oct 22 23:46:30 oh Oct 22 23:47:47 GFER or GRER sets GEDR on change so it waits for change and can timeout, that's all Oct 22 23:50:36 hm, but it seems that my GSM_WAKE doesn't cause GSM_HOST_WAKE change Oct 22 23:50:42 maybe it's wrong GPIO Oct 22 23:53:30 hmmm, you're right, I won't kill it whole but I'll try to ignore that timeout stuff Oct 22 23:53:50 it almost look like that is that part of code causing delays :b Oct 22 23:54:35 heh, well there is nothing else in that driver :D Oct 22 23:54:39 ok Oct 23 00:03:08 hmmm, I have to reduce size again :( Oct 23 00:11:16 hmmm, very weird Oct 23 00:11:32 It works quite fine when also Palm Treo 650 is selected Oct 23 00:11:48 but without it there is reset, reset, reset Oct 23 00:12:06 probably some not obvious dependency :/ Oct 23 00:16:27 :/ Oct 23 00:27:18 ATE1 Oct 23 00:27:20 ERROR Oct 23 00:27:37 what.. Oct 23 00:27:50 how about AT+CMUX=? Oct 23 00:27:56 (including question mark) Oct 23 00:28:07 that's one that even works in AT+CFUN=0 mode for me Oct 23 00:28:12 hmmm, probably in next dialog Oct 23 00:29:34 there is 34 always high and 35 always low :/ Oct 23 00:30:51 (34-RXD, 35-CTS) Oct 23 00:30:59 35 is probably OK Oct 23 00:31:23 hmm, maybe 34 is ok too Oct 23 00:52:30 gotta go sleep Oct 23 00:52:38 good luck tonigh (tomorrow) Oct 23 01:59:20 Anyone know if I can I use "qemu-system-arm -M versatilepb" to install an emulated debian and use it to "natively" compile a kernel (using the tree from git://git.hackndev.com/linux-hnd ?) and debootstrap a rootfs to boot with cocoboot (from kedar.palmlinux.cz/test ?) on a Treo 680? Is this a sane approach for a debian sid user? Oct 23 02:09:28 I've a centro, ld, e2 and z72 I'll probably look at later if I get off and running, and some experience building live debian systems for i386/amd64/powerpc ;-) With my old phone dying I've an itch to scratch (and a spare 680) but I'm also game to see if I can leverage my other experience to automate building some useful debian images, and have some machines to test things with **** ENDING LOGGING AT Thu Oct 23 02:59:57 2008