**** BEGIN LOGGING AT Fri Oct 17 02:59:57 2008 Oct 17 04:48:50 WOOHOOO Oct 17 04:48:52 I did it! Oct 17 04:51:02 it seems the internal pull down is too strong Oct 17 04:51:36 but we can hack around that by using the pull-up from the USB client controller Oct 17 04:52:37 ahahah Oct 17 04:57:06 Sleep_Walker: you're a genius, it works! Oct 17 04:57:18 now i just gotta make it reliable Oct 17 04:57:28 but we know its possible! Oct 17 05:19:58 marex: dude! Oct 17 05:20:04 I just got USB host working on treo Oct 17 05:20:15 650 Oct 17 05:20:21 via the normal USB cable Oct 17 05:20:30 BobOfDoom, congratulations ;-) Oct 17 05:20:56 I think it's doable on any of PXA27x devices that use the PXA's USB controller Oct 17 05:21:25 so now we can add wifi and whatever we want :D Oct 17 05:21:40 cool! Oct 17 05:59:07 BobOfDoom: You rock! Oct 17 05:59:13 BobOfDoom: What about SDIO ? Oct 17 06:00:29 Speedy2: I assume SDIO would "just work" if the Linux driver supports it Oct 17 06:00:39 I don't have any SDIO cards though Oct 17 06:00:56 Where are you located? US ? AU? Oct 17 06:01:02 AU Oct 17 06:01:05 :( Oct 17 06:01:14 I have a SD WiFi card I'd send, but that doesn't make sense Oct 17 06:01:22 hehe Oct 17 06:01:30 I might get one just to try it Oct 17 06:01:53 but yeah, that's a good option for wifi Oct 17 06:02:11 So dumb that Palm won't release a WiFi driver for SD Oct 17 06:02:30 Now we just need to find away to conviently run Palm programs. Maybe some hacking on the PalmEMU Oct 17 06:09:14 http://www.avecmobile.com/index.php?id=1047 Oct 17 06:09:15 hrm Oct 17 06:12:16 http://www.access-company.com/products/gvm/index.html Oct 17 06:40:51 BobOfDoom: no, you're genius! great! Oct 17 06:40:56 gm Oct 17 06:41:20 Sleep_Walker: gm Oct 17 06:42:01 Sleep_Walker: low speed devices work fine Oct 17 06:42:20 Sleep_Walker: "full" speed devices (USB 1.1) need a hack Oct 17 06:42:31 have to turn on pullup register Oct 17 06:42:59 cause plug in detection is stuffed for full speed devices Oct 17 06:43:09 because the hardware is wired to be a client Oct 17 06:43:18 so you have to to trick it Oct 17 06:44:02 so it means user will have to enable USB _after_ plugging in device (if before it'll error out cause it thinks theres something there but there isn't) Oct 17 06:44:10 anyway that's not too much of a problem Oct 17 06:44:21 and every usb device I've tried has worked Oct 17 06:44:26 omg Oct 17 06:44:32 few different flash drives, card readers, camera, keyboard, mouse Oct 17 06:44:56 * BobOfDoom uploading video Oct 17 06:45:11 :) Oct 17 06:45:31 I wonder if we could hack the Access GVM for Nokia to run on Treo-Linux Oct 17 06:45:41 another prove that PalmOS sucks and Linux rocks Oct 17 06:46:19 gah Oct 17 06:46:22 youtube takes forever Oct 17 06:46:23 The dumb thing is *Palm* could release a Linux OS for Treo's and then provide a VM to run Garnet apps. Oct 17 06:46:30 But of course that would be too smart. Oct 17 06:46:40 Sleep_Walker: try this http://meshy.org/~ato/tmp/treo650_usb_host.asf Oct 17 06:46:49 it's in whatever random format the camera spits out though Oct 17 06:48:11 USB keyboard :D Oct 17 06:48:23 mouse?! Oct 17 06:48:38 ah here we go: http://au.youtube.com/watch?v=fWEzaPk-baM Oct 17 06:48:43 same video on youtube Oct 17 06:48:59 so I can use just hub and play? Oct 17 06:49:55 Sleep_Walker: gender changer + powered hub would probably work Oct 17 06:50:00 hmm Oct 17 06:50:03 I should try that actually Oct 17 06:50:10 I wonder if I have a powered hub somewhere Oct 17 06:50:37 maybe the one in my monitor is powered Oct 17 06:52:03 I have to read more about that switching USB Host/USB Client :/ Oct 17 06:53:36 hmm Oct 17 06:53:42 it doesn't like the hub in my monitor Oct 17 06:53:48 that's the first thing I've tried tht hasn't worked Oct 17 06:55:01 USB lines have pu/pd on them I think to indicate type of device Oct 17 06:55:06 D+ / D- Oct 17 06:55:08 it didn't work with the freerunner either Oct 17 07:03:05 do you know which suspend is called by APM? Oct 17 07:06:10 mem probably Oct 17 07:07:58 I have to fix this as top priority cause it's very annoying :b Oct 17 07:16:52 I'm still trying to avoid reading step by step disassembled bootloader Oct 17 07:17:03 there is interesting thing Oct 17 07:17:20 sed -n 's/.*\(0x[a-fA-F0-9]\{8\}\).*/\1/gp' bl.asm Oct 17 07:17:41 this will dig out all addresses which are used Oct 17 07:18:00 cut -d\ -f3 memorymap.txt | tr '-' ' ' > map.txt Oct 17 07:18:14 this will prepare map of memory Oct 17 07:18:22 sed -n 's/.*\(0x[a-fA-F0-9]\{8\}\).*/\1/gp' bl.asm | sort -u | while read val; do while read min max; do if [ $((val)) -ge $((0x$min)) ] && [ $((val)) -le $((0x$max)) ]; then echo -e "0x$min\t$val\t0x$max"; fi; done < map.txt ; done Oct 17 07:18:41 and this should show, which memory addresses are mapped in PalmOS Oct 17 07:18:55 the result is interesting Oct 17 07:19:03 0x40000000 0x40000000 0x41800000 Oct 17 07:19:07 only this one Oct 17 07:19:46 DMA controler register Oct 17 07:23:43 hmm Oct 17 08:04:40 Sleep_Walker: ah ha! Oct 17 08:04:46 ? Oct 17 08:04:55 Sleep_Walker: 20000000-21c00000 -> a2000000-a3c00000 Oct 17 08:04:58 (680) Oct 17 08:05:06 Sleep_Walker: 20000000-20d00000 a0000000-a0d00000 Oct 17 08:05:09 (650) Oct 17 08:05:25 20000000 == palm card Oct 17 08:05:40 first word of palm card is reset vector (according to the format) Oct 17 08:05:54 so a2000000 may be your address Oct 17 08:06:01 compiling :) Oct 17 08:06:40 booting Oct 17 08:07:17 still booting :b Oct 17 08:08:33 still booting Oct 17 08:09:11 starting OPIE Oct 17 08:09:47 starting opie-console Oct 17 08:10:10 suspended Oct 17 08:10:33 stucked Oct 17 08:11:18 noes Oct 17 08:11:19 I'll try to add mmc-card trigger before suspend to see if the problem isn't elsewhere Oct 17 08:11:35 or heartbeat trigger Oct 17 08:11:48 gotta go to work now Oct 17 08:12:30 ok Oct 17 08:12:31 thanks for hint, I'll test it on the way... Oct 17 08:12:31 cya Oct 17 09:25:54 * Sleep_Walker has now gender changer and powered hub Oct 17 09:26:08 I hope that it won't break anything Oct 17 09:26:59 BobOfDoom: If I have it set incorrectly, can I break something? (it seems that only port1 is powered, port2 and port3 not) Oct 17 09:28:04 * Sleep_Walker has prepared USB mouse Oct 17 09:29:35 Sleep_Walker: I think it's pretty safe Oct 17 09:29:47 it does not work Oct 17 09:29:55 I have to enable that ports probably Oct 17 09:30:04 yeah Oct 17 09:30:07 and pullup Oct 17 09:30:32 I have to find out, how to do that :/ Oct 17 09:30:42 I was using PalmOS values to init OHCI Oct 17 09:31:43 btw. maybe I have some larger misconfiguration of PM Oct 17 09:32:03 it sometimes can't resume even from standby Oct 17 09:32:10 Sleep_Walker: http://meshy.org/~ato/tmp/ohci-hack.diff Oct 17 09:33:02 oh actuall Oct 17 09:33:18 but without UP2OCR_DPPDE | UP2OCR_DMPDE; Oct 17 09:33:20 and Oct 17 09:33:37 after plugging in, turn on the USB pullup GPIO Oct 17 09:33:40 or Oct 17 09:33:51 probably booting with USB enabled would work Oct 17 09:34:02 (select USB console in cocoboot before booting) Oct 17 09:34:12 make sure you're plugged in before booting Oct 17 09:34:14 that might work Oct 17 09:34:33 I was toggling the GPIO manually to make it detect insertion Oct 17 09:34:36 BobOfDoom: I've got USB pullup as part of UDC code Oct 17 09:34:50 so when it detect USB, it does pullup Oct 17 09:36:32 UDC.. Oct 17 09:36:35 oh Oct 17 09:36:46 I haven't tried compiling in both UDC and OHCI Oct 17 09:37:00 I guess it works Oct 17 09:37:07 it has to Oct 17 09:37:08 as long as you give the port to OHCI Oct 17 09:37:13 since it works in palmos Oct 17 09:37:29 yeh Oct 17 09:37:35 but we're trying to route UDC signals to OHCI Oct 17 09:37:39 for the multiconnector Oct 17 09:38:04 anyway.. I guess UDC will just stop working when you do that Oct 17 09:38:07 so UP2OCR = UP2OCR_HXS | UP2OCR_HXOE; Oct 17 09:38:13 routes UDC signals to OHCI Oct 17 09:38:34 UHCHR &= ~(UHCHR_SSE | UHCHR_SSEP2); Oct 17 09:38:50 turns on poewr to host port 2 (== UDC pins once that routing is done) Oct 17 09:39:03 aussies?! http://www.theinquirer.net/gb/inquirer/news/2008/10/15/australians-censor-internet Oct 17 09:39:22 UHCRHDA |= UHCRHDA_NOCP; <-- no over current protection Oct 17 09:39:33 yeah Oct 17 09:39:50 both major parties are into that internet censoring rubbish Oct 17 09:40:26 the liberal party made some software you could download if you wanted it Oct 17 09:40:39 but it sucked (of course) and was easily worked around by 10 year olds etc Oct 17 09:40:47 so labor gets into government Oct 17 09:40:55 and decides well lets do it at the ISP level Oct 17 09:48:18 it's pretty screwed up Oct 17 09:48:31 it seems Oct 17 09:48:47 *sigh* Oct 17 09:49:01 if there will be prohibited anonymizers, it will be end of freedom Oct 17 09:50:27 "Illegal is illegal and if there is infrastructure in place to block it, then it will be required to be blocked -- end of story." Oct 17 09:50:48 * BobOfDoom rolls eyes Oct 17 09:50:57 anyone who really cares will just proxy around it Oct 17 09:51:06 so it's pretty pointless if you ask me Oct 17 09:51:49 if it won't be prohibited by law ;) Oct 17 09:53:26 heh Oct 17 09:56:01 actually austraia is pretty crap for "legal" freedoms Oct 17 09:56:07 eg we have no free speech laws Oct 17 09:56:55 our copyright has just been made way more restrictive than the US Oct 17 09:57:02 thanks to free trade agreements Oct 17 09:57:08 we got the US's DMCA like stuff Oct 17 09:57:24 but without as many exeptions and "fair use" stuff Oct 17 09:58:14 omg Oct 17 09:59:55 but fortunately we seem to have less corrupt governments than many places.. well, at least they're less obviously corrupt anyway :p Oct 17 10:02:23 "Australia’s laws on Internet censorship are, theoretically, amongst the most restrictive in the Western world. However, the restrictive nature of the laws has been combined with almost complete lack of interest in enforcement from the agencies responsible . Oct 17 10:02:28 haha Oct 17 10:02:29 yeah Oct 17 10:02:38 actually it used to be illegal to record TV Oct 17 10:02:44 omfg Oct 17 10:02:48 or record radio Oct 17 10:02:51 for later listening/watching Oct 17 10:02:58 but nobody ever bothered to enforce it Oct 17 10:04:55 oh games, for example the original Duke Nukem 3D had to be edited (to remove violence and sex) before release in austraila Oct 17 10:04:58 same with GTA Oct 17 10:05:08 Fallout 3 as well Oct 17 10:05:23 that's fine they didn't suppress 0verkill Oct 17 10:05:29 omg, Fallout 3 is out? Oct 17 10:05:47 * Sleep_Walker missed something again Oct 17 10:06:04 it's not out Oct 17 10:06:08 I don't think Oct 17 10:06:12 but it's already banned :p Oct 17 10:06:25 oh Oct 17 10:06:27 I see Oct 17 10:06:27 "A revised version of the game was resubmitted to the OFLC and reclassified as MA 15+ on August 7, 200" Oct 17 10:06:28 :b Oct 17 10:06:32 ah Oct 17 10:06:43 so it's they've allowed a censoed versoin Oct 17 10:06:52 that removes some of the drug references Oct 17 10:07:36 but meh... what can you do? Oct 17 10:08:23 the only people who can get into power in most countries are god-type religious people who are into all that censorship and commands from above stuff Oct 17 10:09:01 my state government were going to legalise gay marriage Oct 17 10:09:11 but the federal government blocked it Oct 17 10:11:58 god bless European Union :/ Oct 17 10:12:58 BobOfDoom: why you don't set much PM registers? Oct 17 10:13:05 it is already set by PalmOS? Oct 17 10:13:32 or it is in some pxa generic code... Oct 17 10:16:10 which ones? Oct 17 10:16:25 I think the main ones were set by palmos Oct 17 10:16:43 like the power states Oct 17 10:16:52 in linux-hnd IIRC I'm leaving the GPIOs as they are in wake Oct 17 10:16:59 and just setting up some wake interrupts Oct 17 10:19:09 I'd rather commit my changes to be able to review it Oct 17 10:19:56 it just worked, so I didn't bother looking at the other registers ;P Oct 17 10:20:13 :/ Oct 17 10:20:46 is it possible that return address change affects also standby mode? Oct 17 10:22:14 mm Oct 17 10:22:16 hmm Oct 17 10:22:43 I suppose it might be powering off one of the RAM banks Oct 17 10:22:57 check PSTR Oct 17 10:23:00 o hwait Oct 17 10:23:12 you mean would putting that return address break standby? Oct 17 10:23:17 hmmm don't think so Oct 17 10:25:57 well, then there is at least another problem :/ Oct 17 10:34:20 git updated Oct 17 10:34:35 I tried to dump PM settings from PalmOS Oct 17 10:34:46 but I may create some mistake Oct 17 10:35:52 PSTR = 0; Oct 17 10:35:57 right so RAM is staying on Oct 17 10:36:35 PUCR = 0; -- well no u-sim reader so it doesn't matter Oct 17 10:37:20 PSLR = 0xff000004; Oct 17 10:37:23 maximum delay Oct 17 10:37:53 plus PI power domain active Oct 17 10:38:34 PSSR = PSSR_OTGPH | PSSR_SSS; Oct 17 10:38:39 on the go alive? hmm Oct 17 10:39:07 clear sleep status Oct 17 10:40:04 PEDR = PWER_RTC; Oct 17 10:40:20 clears RTC edge detect.. probably unnecesary but shouldn't hurt Oct 17 10:40:51 PEDR - it's edge detect wake up source, isn't it?! Oct 17 10:41:07 it's stautus Oct 17 10:41:15 it tells you what woke you up Oct 17 10:41:18 writing ot it just clears it Oct 17 10:41:43 PCFR = PCFR_FVC | PCFR_PI2CEN; Oct 17 10:41:57 of right Oct 17 10:42:01 oh right Oct 17 10:42:06 I2C power manager enabled Oct 17 10:42:33 FVC freq/volt change power manager thinginess Oct 17 10:42:59 ie let the power manager know when we're clock scaling Oct 17 10:43:13 so it can adjust amount of power supplied I guess Oct 17 10:43:55 OMG Oct 17 10:44:14 I just removed/inserted battery and Treo stayed off Oct 17 10:44:29 oh dear Oct 17 10:44:48 you battery is not flat is it? Oct 17 10:44:56 that means? Oct 17 10:45:14 so it's stuck off? Oct 17 10:45:19 could the battery have gone empty Oct 17 10:45:25 while you were doing linux stuff Oct 17 10:45:30 and it has not enough power to boot? Oct 17 10:45:34 try plugging in charger Oct 17 10:45:42 well it was all the time in USB Oct 17 10:45:44 oh Oct 17 10:45:45 right Oct 17 10:45:50 maybe with USB port off Oct 17 10:46:10 it's now on power Oct 17 10:46:12 ah Oct 17 10:46:18 red led is lighting Oct 17 10:46:28 stay tuned :b Oct 17 10:46:34 XD Oct 17 10:46:50 * Sleep_Walker has almost heart attack Oct 17 10:46:58 got it :) Oct 17 10:47:02 it's alive Oct 17 10:47:40 btw. that also means we need to enable USB Client as power source Oct 17 10:48:14 hmm Oct 17 10:51:28 so you can't find the problem too? Oct 17 10:51:46 I'll try to do some more PM dumps with on/off Oct 17 11:01:18 hmmm, PCFR does not have same value all the time Oct 17 11:01:33 it enables FVC lately Oct 17 11:01:39 maybe after resume? Oct 17 11:04:03 don't think it should hurt to try without it Oct 17 11:04:12 I think it's just to obtain lower power usage Oct 17 11:04:15 in suspend Oct 17 11:04:30 and wtf that i2c? Oct 17 11:04:42 it confuses me a bit Oct 17 11:06:12 |miska|: do you have serial cable with MUC? Oct 17 11:06:39 <|miska|> No Oct 17 11:06:48 I2C talks to the power management chip Oct 17 11:06:53 <|miska|> I think I've got spare MUC Oct 17 11:07:03 so it can power off some of the CPU power lines Oct 17 11:07:09 which aren't needed in suspend Oct 17 11:07:10 I think Oct 17 11:07:21 there is some i2c chip? :/ Oct 17 11:07:35 PM is so much mess for me ;/ Oct 17 11:07:52 probably this one: http://datasheets.maxim-ic.com/en/ds/MAX1586A-MAX1587C.pdf Oct 17 11:10:09 ok Oct 17 11:11:22 BobOfDoom: does it need some driver? Oct 17 11:12:20 Sleep_Walker: no, the CPU handles it automatically Oct 17 11:12:24 I think Oct 17 11:12:39 fine Oct 17 11:12:42 although maybe you can tweak some things from kernel too Oct 17 11:12:51 but you don't need to Oct 17 11:13:04 ok FCV really changed nothing Oct 17 11:13:11 :( Oct 17 11:14:13 hm, maybe serial cable for debugging PM will be necessary Oct 17 11:14:51 |miska|: well, I don't want to just take it - is it really spare? Oct 17 11:15:53 BobOfDoom: I've got I2C drivers enabled, can that cause problem? Oct 17 11:16:58 Sleep_Walker: no idea Oct 17 11:17:04 I've never tried enabling them Oct 17 11:17:10 ok, compiling Oct 17 11:17:40 well, in fact, they wasn't even probably loaded :( Oct 17 11:19:06 ah Oct 17 11:29:58 hmmm, colegue give me a hint - I'll add i2c dumping :b Oct 17 11:30:08 lets see what it does Oct 17 12:23:21 BobOfDoom: do you have somewhere save PCMDx and PVCR registers? Oct 17 12:23:32 could you send it to me? Oct 17 12:23:35 thx Oct 17 12:34:59 from 650? Oct 17 12:35:04 no but I'll get it now Oct 17 12:35:05 just a sec Oct 17 12:35:19 thx Oct 17 12:35:57 I think that is that missing part Oct 17 12:36:15 I dumped mine and I hope I've got it whole Oct 17 12:36:31 booting to proove Oct 17 12:36:56 PCMD0 = 0000_0410 Oct 17 12:37:09 PCMD1 = 0 Oct 17 12:38:05 all the rest of PCMDx are zero Oct 17 12:38:05 hmmm Oct 17 12:38:15 I've got 0x0416 Oct 17 12:38:19 and rest zeroes Oct 17 12:38:29 0x0400 means it is last command Oct 17 12:38:38 lets try and hope Oct 17 12:38:52 PVCR = 0000_0014 Oct 17 12:40:42 hmm Oct 17 12:40:50 0x10 == 1.1V Oct 17 12:41:08 according to MAX1586 datasheet Oct 17 12:41:18 0x14 - The seven-bit address of the external regulator’s I2C module. Oct 17 12:41:26 PVCR (due to PXA doc) Oct 17 12:41:40 address on I2C bus I think Oct 17 12:41:55 mm Oct 17 12:41:56 0x0416 did not work Oct 17 12:42:07 I'll try 0x410 Oct 17 12:42:41 hmm Oct 17 12:42:48 0x0416 = 1.25 V Oct 17 12:42:52 core voltage Oct 17 12:43:27 i2cdump ? i2cdetect ? Oct 17 12:43:36 are you looking for a chip on i2c bus or what ? Oct 17 12:44:06 marex: the power manager Oct 17 12:44:32 yea, I know what that chip is, it's in TX as well Oct 17 12:45:13 Sleep_Walker: my guess is it's something else.. all that stuff should still be configured by Palm OS Oct 17 12:45:39 marex: Sleep_Walker is trying to debug suspend Oct 17 12:45:47 marex: standby is being unreliable Oct 17 12:46:39 yea, max should be preset by palmos and linux wont touch it at all Oct 17 12:47:03 or rather it's set to power-on defaults and palmos does some little adjustments when needed Oct 17 12:47:12 it has nothing to do with suspend iirc Oct 17 12:47:52 moreover if you wont touch i2c, it should just work Oct 17 12:52:35 marex: could you please review the code? Oct 17 12:53:07 what code ? Oct 17 12:53:47 marex: http://git.hackndev.com/?p=linux-hnd;a=blob;f=arch/arm/mach-pxa/palmt680/palmt680_pm.c;h=bd637963889d64fa13106e442aa78a56a6517298;hb=1c499de4f53bf46ba4f93fce3959a2f94fb6fbce Oct 17 12:54:21 ad i2c - you're probably right Oct 17 12:54:44 sometimes it switch red led on so processor did something Oct 17 12:55:24 red led can be hooked to gpio or asic Oct 17 12:55:27 can that be that return address? :( Oct 17 12:55:40 dont forget that asic has pwm driven led outputs Oct 17 12:55:41 marex: it's probably connected to BCM2133 chip Oct 17 12:55:58 as far as we can tell it doesn't have the ASIC Oct 17 12:56:11 it's not necessary Oct 17 12:56:12 not at all ? Oct 17 12:56:13 since it's using USB Oct 17 12:56:17 lucky :) Oct 17 12:56:26 yep Oct 17 12:56:33 that ASIC is a big pain Oct 17 12:56:39 indeed Oct 17 12:57:03 Sleep_Walker: actually, does the LED ever flash in suspend in palm os? Oct 17 12:57:08 or is it just straight on? Oct 17 12:57:20 I can't say Oct 17 12:57:30 you've never seen it? Oct 17 12:57:30 hmm Oct 17 12:57:31 then Oct 17 12:57:48 red LED is the only one you can't find control for huh? Oct 17 12:57:50 BobOfDoom, the whole gsm thingamie is connected over usb ? Oct 17 12:57:56 what is it used for in palm os? Oct 17 12:57:59 just to show charging? Oct 17 12:58:12 it might just be wired to battery charger directly I guess Oct 17 12:58:28 marex: yeah, it's a serial over USB Oct 17 12:58:39 BobOfDoom: I have found GPIO which sometimes turns on/off red led Oct 17 12:58:41 marex: bluetooth is wired that way too Oct 17 12:58:44 in the 680 Oct 17 12:58:55 so there is even usb hub ? Oct 17 12:58:55 Sleep_Walker: ah ok Oct 17 12:59:03 that's pretty :) Oct 17 12:59:15 marex: I don't think so, PXA has 3 USB ports, my guess is 2 in host mode 1 in client mode Oct 17 12:59:29 also Sleep has got the bluetooth chip to show up Oct 17 12:59:32 but not GSM Oct 17 12:59:34 * marex recalls something like that ... Oct 17 12:59:37 so probably GSM is on the other port Oct 17 12:59:43 if there were hub he'd see both of them Oct 17 12:59:46 and the hub in lsusb Oct 17 12:59:54 maybe I need GPIO to power it up Oct 17 13:00:01 if the gsm was powered ... yes Oct 17 13:00:14 Sleep_Walker, it should be powered after bootup though Oct 17 13:00:18 right, but if you had it on in Palm OS it should be powered Oct 17 13:00:19 yeah Oct 17 13:00:34 it should be even connected to network if you dont reset it somehow Oct 17 13:00:46 as I recall universal had some qualcomm modem connected over usb too Oct 17 13:01:10 and what I recall from my hacking on hw6515d gsm, it has more than two gpios (unlike bluetooth) to drive it Oct 17 13:02:35 hmm.. although weird that it doens't show in your UHC traces Oct 17 13:02:38 http://sleepwalker.hackndev.com/GPIO/UHC-bt-on.txt Oct 17 13:02:41 BT one shows it Oct 17 13:02:45 0x9c000054 0x00030100 Oct 17 13:02:45 0x9c000058 0x00000100 Oct 17 13:02:45 0x9c00005c 0x00000100 Oct 17 13:02:57 54 = port 1 status, 58 = port 2, 5c = port 3 Oct 17 13:03:03 port 2 == client usually Oct 17 13:03:16 so we assume BT is on port 1 Oct 17 13:03:27 which is why it just works when you install ohci driver Oct 17 13:03:30 are you sure the GSM runs over usb too ? Oct 17 13:03:41 it can be connected via pxa-uart ... Oct 17 13:03:50 that's the most command approach Oct 17 13:03:56 marex: to which one? Oct 17 13:03:57 s/command/common/ Oct 17 13:04:21 UsbRadioSerialBulverde.prc Oct 17 13:04:41 btw. dmitry confirmed that and it has some well called PRC in ROM :) Oct 17 13:04:43 BobOfDoom, oh well ... you took it apart ? Oct 17 13:05:12 marex: I don't have one yet... I've got ebay order in though :p Oct 17 13:05:18 marex: I'll take it apart as soon as I get it Oct 17 13:05:31 can make out some chips here http://www.palminfocenter.com/images/treo-68-guts-5-L.jpg Oct 17 13:06:01 top right is bluetooth, unreadable but it matches some other pics I found of it in other places Oct 17 13:06:07 bottom right is WM9712 Oct 17 13:06:14 bottom left is maxim.. I assume power manager Oct 17 13:06:35 Top left, I dunno Oct 17 13:06:45 one under that is RF transciever for the radio Oct 17 13:07:02 and the one under the shield is also GSM related, there's a shield like taht in the 650 Oct 17 13:07:06 BobOfDoom, top left - probably bluetooth Oct 17 13:07:33 there were bcm2035s packaged like that Oct 17 13:07:54 Treo680 has BCM2045A Oct 17 13:08:17 Sleep_Walker, yea, but the package is apparently similar (and the chip internals wont be that different too) Oct 17 13:08:17 good pics on the other side of the PCB: http://www.inkhorn.net/site/log/Entries/2007/5/30_Interior_of_the_Treo_680.html Oct 17 13:08:25 evil M-SYstems Oct 17 13:08:36 the GSM baseband processor itself Oct 17 13:08:44 the RAM + Flash for the GSM CPU Oct 17 13:08:46 BobOfDoom, we can do raw-reads and writes from the DoCs ;) Oct 17 13:08:56 BobOfDoom, there is a driver for that Oct 17 13:09:06 marex: ooh.. really?! Oct 17 13:09:07 where Oct 17 13:09:11 I know there was read-only one Oct 17 13:09:24 but can do writes now? Oct 17 13:09:29 it can do writes, but Im not brave enough Oct 17 13:09:33 hehe Oct 17 13:09:54 the universal one ... I cleaned it and ported to .27 so it can be found on odin somewhere Oct 17 13:10:23 nice Oct 17 13:14:45 actually come to think of it, with that trick kedar and sleep_walker use, read only is still pretty good, as long as we can find where TFFS has thrown the data Oct 17 13:15:31 wtf TFFS? Oct 17 13:15:52 check nftl and friends ... they are not that simple ;|) Oct 17 13:16:00 TFFS = M-System's proprietary flash format Oct 17 13:16:09 well Oct 17 13:16:10 filesystem thingy Oct 17 13:16:15 TrueFlashFileSystem Oct 17 13:16:17 for doing wear-leveling and whatever Oct 17 13:16:56 TFFS Transaction Flash File System Oct 17 13:17:05 TFFS True Ffs Oct 17 13:17:07 hmmm Oct 17 13:17:09 ok Oct 17 13:17:40 I don't need internal flash until I've got SDIO Wifi working in linux :b Oct 17 13:18:00 hmm Oct 17 13:18:15 do you reckon LifeDrive screen is compatible with PalmTX ? Oct 17 13:18:43 BobOfDoom, not only that ... I personally tested it, but dont tell |miska| Oct 17 13:18:45 I have a spare LifeDrive screen (from my old too-hacked LD) and there's TX with broken screen on ebay Oct 17 13:19:01 oh dear.. he's in the room :p Oct 17 13:19:07 awesome Oct 17 13:19:14 * BobOfDoom places a bid Oct 17 13:19:15 (oh damn ;-) ) Oct 17 13:22:18 marex: so, have you find some reason why not to work? Oct 17 13:26:09 Sleep_Walker, huh? Oct 17 13:27:04 [14:52] marex: could you please review the code? Oct 17 13:27:05 [14:53] what code ? Oct 17 13:27:07 [14:53] marex: http://git.hackndev.com/?p=linux-hnd;a=blob;f=arch/arm/mach-pxa/palmt680/palmt680_pm.c;h=bd637963889d64fa13106e442aa78a56a6517298;hb=1c499de4f53bf46ba4f93fce3959a2f94fb6fbce Oct 17 13:27:48 Sleep_Walker, I havent bothered to look at it yet Oct 17 13:28:41 Sleep_Walker, you should configure your keyboard using MFP Oct 17 13:29:26 and beware - PGSRs's states are kept while sleeping Oct 17 13:29:39 (are might be reconfigured by bootloader, causing some hardware to hang) Oct 17 13:29:42 since keyboard sometimes makes red led on, that probably is not problem Oct 17 13:30:00 PGSR are configured same way as in PalmOS Oct 17 13:30:05 also - are lines 110 - 114 really needed ? Oct 17 13:30:26 well ... that never worked for me on T5 :-B Oct 17 13:30:43 (configuring PGSRs the same way as in palmos) Oct 17 13:31:02 you should properly pinpoint what causes the wakeup and set only those bits Oct 17 13:31:25 or try what I used for the 650 Oct 17 13:31:28 PGSR0 = GPLR0; Oct 17 13:31:29 etc Oct 17 13:31:32 keep current states Oct 17 13:31:45 to minimize changes during sleep Oct 17 13:32:17 BobOfDoom: that was first thing I tried Oct 17 13:32:17 http://marex.hackndev.com/patches-mainline/0004-Palm-Tungsten-T5-Suspend-to-RAM-support.patch Oct 17 13:32:22 but ok Oct 17 13:33:07 Sleep_Walker: so what happens exactly, yyou go into standby nd it refuses to wake? Oct 17 13:33:08 see the patch, that's how it works ... Oct 17 13:33:13 Sleep_Walker: but sometimes red LED comes on? Oct 17 13:33:21 yes Oct 17 13:33:27 hmm Oct 17 13:33:29 no other activity found Oct 17 13:33:34 red LED could be the bootloader doing something Oct 17 13:33:38 BobOfDoom, suspicious - it's like if it wasnt totally sleeping but rather stuck in loader Oct 17 13:33:52 red LED come son only after you wake it? Oct 17 13:34:11 no Oct 17 13:34:18 also sometimes during boot Oct 17 13:34:19 it's on while asleep too? Oct 17 13:34:24 ah Oct 17 13:34:30 I don't know about that :( Oct 17 13:34:33 boot of palm os? Oct 17 13:34:36 ok, so not Oct 17 13:34:41 boot of linux Oct 17 13:35:47 and GPIO 20 has something to do with LEDs? Oct 17 13:36:06 yes Oct 17 13:36:13 it turns green on Oct 17 13:36:34 sometimes switching green on causes red off Oct 17 13:36:36 "Red/Green led switching Oct 17 13:36:37 sometimes not Oct 17 13:36:40 ah Oct 17 13:36:43 hmm Oct 17 13:37:04 that's confusing Oct 17 13:38:14 oooh Oct 17 13:38:40 % strings Troubleshooter_enUS.oprc| grep "red LED" Oct 17 13:38:41 test the red LED Oct 17 13:38:43 heh Oct 17 13:47:15 jeez ... Oct 17 13:47:32 but still better than russian comments in palmLD's rom Oct 17 13:48:43 .. Oct 17 13:48:46 my window just got egged Oct 17 13:48:53 bastards Oct 17 13:49:15 that Oct 17 13:49:26 s one problem with warmer whether Oct 17 13:50:04 more arseholes hanging out at night Oct 17 15:29:07 * Sleep_Walker just tested USB host too! Oct 17 15:29:17 great! Oct 17 17:41:14 hello, is somebody here? i have a problem with my palm treo 650 , i try this http://hackndev.com/node/212 but when i ran Boot! on cocoboot, it wrote that loading /zImage ... Ok , loading /initrd.gz ... Ok ... and it stopped, i cant do anything, before it i try hard reset, without solutions :( sorry my english is bad but i have to have linux on my PDA, i am a developer of new desktop for PDA computers Oct 17 18:52:48 BobOfDoom: I discussed with collegue and I've got something interesting... Oct 17 18:53:13 1] there is some USB wifi which internally changes 5V into 3V Oct 17 18:53:47 2] he drawed me some simple circuit which will change 3.3V into 5V Oct 17 18:54:10 so it would do complete USB Host without any external power source Oct 17 22:16:57 good morning, Bob :) Oct 17 23:51:50 BobOfDoom: it seems that address in PM code is correct because other addresses causes malfuction of standby mode Oct 17 23:52:44 ...but I thought that standby mode is working out of box... :( Oct 18 00:40:00 what? Oct 18 00:40:06 the address should have nothing to do with standby mode Oct 18 00:40:12 cause standby doesn't lose state Oct 18 00:46:28 I'm also confused Oct 18 00:48:18 hmm Oct 18 00:48:37 I tested it 3 times with different addresses Oct 18 00:48:55 I create workaround for OPIE Oct 18 00:49:28 instead of binary /usr/bin/apm there is just script calling "echo standby > /sys/power/state" Oct 18 00:50:17 Sleep_Walker: http://git.hackndev.com/?p=linux-hnd;a=blob;f=arch/arm/mach-pxa/pxa27x.c Oct 18 00:50:30 for TX, T5 and LD in linux-hnd we workaround it Oct 18 00:50:37 if(machine_is_xscale_palmld() || machine_is_xscale_palmtx() || Oct 18 00:50:37 156 machine_is_xscale_palmtt5()) { Oct 18 00:51:03 cause on those it's raher dangerous to go into mem sleep Oct 18 00:51:14 on TX you have to desolder the batteyr (or wait for it to drain) to get our Oct 18 00:51:15 out* Oct 18 00:52:59 hmmm, hardcoded PEDR ?! Oct 18 00:55:22 what? Oct 18 00:55:30 I keep telling you PEDR is a status register Oct 18 00:55:39 you're right Oct 18 00:55:46 sorry Oct 18 00:55:48 :) Oct 18 00:56:01 :b Oct 18 00:56:04 3 o'clock Oct 18 00:56:07 almost sleeping Oct 18 00:56:34 so there is another posibility - pure coincidence Oct 18 00:57:04 unfortunately it doesn't push me further :b Oct 18 00:57:43 BobOfDoom: btw. I've got running USB 2.0 device Oct 18 00:57:50 it was detected successfully Oct 18 00:58:05 (despite info in your blog) Oct 18 01:00:13 and - do you think that camera is attached to quick capture interface? Oct 18 01:19:14 * Sleep_Walker enabled that if accidentally does have some supported chip Oct 18 01:27:19 hm, mplayer without TV support Oct 18 01:51:35 standby mode stress test Oct 18 01:51:37 gn **** ENDING LOGGING AT Sat Oct 18 02:59:56 2008