--- Log opened Wed Sep 08 00:00:28 2010
00:00 -!- fmibot [~fmibot@static.225.178.40.188.clients.your-server.de] has joined #freemyipod
00:00 -!- funman [~fun@rockbox/developer/funman] has joined #freemyipod
00:05 < TheSeven> hm, but those mismatching locations seem to be read-only
00:06 < TheSeven> one is next to an lcd_detect pin
00:06 < TheSeven> the other one is totally unknown
00:12 -!- mulenmar_ [~mulenmar@74-132-43-158.dhcp.insightbb.com] has joined #freemyipod
00:13 -!- mulenmar [~mulenmar_@74-132-43-158.dhcp.insightbb.com] has quit [Ping timeout: 264 seconds]
00:22 < TheSeven> if only stooo would be around now, he could poke at that board and check what that pad is connected to
00:24 < mulenmar_> I probably could, if I hadn't already been away 23 hours, had a visual diagram of what to look for, and only needed a multimeter.
00:26 < TheSeven> getting that beast out of its case isn't an easy task at all
00:27 < TheSeven> mulenmar_: there's exactly one option left: let the battery run down until the power manager enters deep sleep mode
00:27 < TheSeven> and hope that this changes something
00:29 < mulenmar_> Ugh.
00:30 < TheSeven> this will take about 4 hours
00:35 < TheSeven> maybe a bit more if it doesn't have much to do
00:35 < TheSeven> so it's probably best to just leave it alone for the night
00:36 < TheSeven> it will shut off at about 2.9V battery voltage
01:13 -!- TheSeven [~TheSeven@rockbox/developer/TheSeven] has quit [Ping timeout: 265 seconds]
01:13 -!- funman [~fun@rockbox/developer/funman] has quit [Quit: bbl]
01:18 -!- TheSeven [~TheSeven@rockbox/developer/TheSeven] has joined #freemyipod
01:28 -!- funman [~fun@rockbox/developer/funman] has joined #freemyipod
01:35 -!- Dreamxtreme [~Dreamxtre@92.30.83.33] has quit [Read error: Connection reset by peer]
01:38 -!- Dreamxtreme [~Dreamxtre@92.30.156.178] has joined #freemyipod
01:50 -!- mulenmar_ [~mulenmar@74-132-43-158.dhcp.insightbb.com] has quit [Quit: Leaving]
02:14 -!- TheSeven [~TheSeven@rockbox/developer/TheSeven] has quit [Ping timeout: 240 seconds]
02:20 -!- TheSeven [~TheSeven@rockbox/developer/TheSeven] has joined #freemyipod
05:00 -!- clustur [~logger@c-76-127-58-39.hsd1.ga.comcast.net] has quit [Remote host closed the connection]
05:00 -!- clustur [~logger@c-76-127-58-39.hsd1.ga.comcast.net] has joined #freemyipod
05:01 -!- liar [~liar@clnet-p09-185.ikbnet.co.at] has joined #freemyipod
05:15 -!- soap [~soap@rockbox/staff/soap] has quit [Ping timeout: 255 seconds]
05:28 -!- soap [~soap@cpe-76-181-78-156.columbus.res.rr.com] has joined #freemyipod
05:28 -!- soap [~soap@cpe-76-181-78-156.columbus.res.rr.com] has quit [Changing host]
05:28 -!- soap [~soap@rockbox/staff/soap] has joined #freemyipod
07:24 -!- funman [~fun@rockbox/developer/funman] has quit [Quit: free(random());]
07:25 -!- perror [~fleury@aldebaran.labri.fr] has joined #freemyipod
09:33 -!- [Saint] [S_a_i_n_t@203.184.0.26] has joined #freemyipod
09:34 * [Saint] wonders what's happening, if anything, with the bricked Nano2G
09:34 < [Saint]> (progress wise...)
09:37 < TheSeven> [Saint]: no idea, mulenmar isn't around
09:38 < TheSeven> he's trying to let the battery run down tonight
09:38 < [Saint]> Bugger, I'll be somewhat pissed if he falls of the scene because he "doesn't have a device anymore"
09:38 < [Saint]> this could be a valuable learning tool.
09:39 < TheSeven> if i manage to grab our hardware guru, I'll have him check what that write protect pin is connected to
09:40 < TheSeven> maybe that could supply an explanation for the misbehavior
09:40 < [Saint]> From what you saw....nothing uses it in the OF, correct? or did I read that incorrectly?
09:40 < TheSeven> what bugs me even more is that we've seen similar behavior a few days ago, that went away after a couple of reboots
09:40 < TheSeven> so this might not be the only ipod where this will happen
09:41 < [Saint]> Argh! :-S
09:41 < TheSeven> we've tried uninstalling the bootloader booting the OF directly, no success
09:41 < TheSeven> so it isn't something we're screwing up on boot, but something persistent
09:42 < TheSeven> the apple bootloader will just reset or show that empty battery logo
09:42 < TheSeven> diskmode always resets
09:42 < [Saint]> weird...
09:43 < TheSeven> this proves that those things can be bricked by only messing up the nand, without touching the nor
09:43 < TheSeven> what i'm wondering is how this relates to embios. it has happened two times since embios was released, and we've never seen such a thing before
09:44 < TheSeven> on the other hand i have no idea how embios could affect it
09:46 < TheSeven> [Saint]: btw, did you have success with your menu change animation experiment?
09:47 < [Saint]> It seems rather strange indeed...no offence to the guy, but, my dealing with him in the past few days lead me to believe there's a chance he could have done something overwhelmingly silly...but, perhaps I'm being overly synical.
09:47 < [Saint]> TheSeven: Limited...
09:47 * TheSeven can't think of *any* way to cause such a thing, let alone a silly one :)
09:47 < [Saint]> what're you using to convert to .UCL?
09:48 < TheSeven> (well, apart from screwing with the hardware)
09:48 < [Saint]> I seem to be getting garbled conversions.
09:48 < TheSeven> the ucl2e10singleblk tool from the freemyipod svn
09:48 < [Saint]> aha...
09:48 < TheSeven> you can just forget about ucl for now and test with the plain bitmap file
09:49 < TheSeven> that'll just increase loading times a bit
09:49 < [Saint]> UCl will load faster right? I'm wanting to get load/blit()  arounf the 0.1sec mark at least...anything below that looks rather crap from experience with RB
09:50 < [Saint]> *around
09:50 * TheSeven would prefetch the whole thing before starting to show it
09:50 < [Saint]> Ah...right, that's a lot smarter than the way I was planning on doing it. I only have a few basic idea on how I *could* do it for mow jotted down.
09:51 < TheSeven> as long as it's <14MB you can just load a file consisting of concatenated bitmaps to 0x09000000 when entering the menu (while the user wonders what to choose)
09:51 < [Saint]> I'm just thinking of "fancy" menu transitions.
09:51 < TheSeven> and then just renderbmp()/blit() those once the user chooses to go to another menu
09:52 < [Saint]> wow...114MB will be plenty ;) I was worried how much room I had to play with, *one* fullscreen image is ~60Kb in .bmd...let along in .ucl so I think I don't have to worry anymore :P
09:52 < TheSeven> (the way I'm doing it with the "booting <whatever>" bitmaps)
09:52 < [Saint]> err...14MB rather
09:53 < [Saint]> argh .bmp, even
09:53 < TheSeven> ucl won't buy you anything if you don't want to decompress the bitmaps individually
09:53 < TheSeven> if you ucl it, you would load the whole thing to 0x08000000 and then decompress it to 0x09000000
09:54 < [Saint]> The thing that gave me the idea for it was that I can also control the backlight, so, if I really put my mind to it I could come up with some pretty awesome effects.
09:54 < [Saint]> (nice touch by the way)
09:54 < TheSeven> you basically have 30MB of ram (from 0x08000000 to 0x09dfffff) to play with, and i'd reserve the lower 16MB as scratch space and to load the actual firmware
09:55 -!- funman [~fun@rockbox/developer/funman] has joined #freemyipod
09:56 < TheSeven> iloader resides in the 0x09e00000 to 0x09f7fffff range, and embios in 0x09f80000 to 0x09ffffff
09:57 < [Saint]> TheSeven: well, when I finish all the images I'm intending to use...and the sequence I intend to use them, I'll probably give you a yell to make sure my code is correct...as I understand it I could wind up with a brick fucking it up, and, the code is a little foreign to me.
09:58 < TheSeven> you can't brick it that way
09:58 < [Saint]> Ah, I must have misread.
09:58 < TheSeven> (or at least the chance to end up with a brick isn't higher than when rolo'ing a new build apparently)
09:59 < [Saint]> ;)
09:59 < TheSeven> you can just get it into a state where you can't enter disk mode easily :)
09:59 < TheSeven> but using some tools you can always fix that
10:00 < [Saint]> I understand I can leave a key combo assigned (while testing) my code that'll force it to reboot?
10:00 < TheSeven> and if you do some basic safety precautions (have a button combination in the vector table point to 0xffffffff) you can use that to enter the fallback menu, which allows you to enter disk mode
10:00 < [Saint]> ah, yeah..that's what I was thinking about.
11:00 -!- clustur [~logger@c-76-127-58-39.hsd1.ga.comcast.net] has quit [Remote host closed the connection]
11:00 -!- clustur [~logger@c-76-127-58-39.hsd1.ga.comcast.net] has joined #freemyipod
11:53 -!- funman [~fun@rockbox/developer/funman] has quit [Quit: free(random());]
12:48 * TheSeven tries enabling the hid and reducing the main stack size, now that a bunch of bugs have been fixed
12:55 < TheSeven> hm
12:58 -!- Dreamxtreme [~Dreamxtre@92.30.156.178] has quit [Quit: Don't follow me]
12:59 < [Saint]> Hm?
13:01 < TheSeven> hm, the hid doesn't even get detected
13:03 -!- Dreamxtreme [~Dreamxtre@92.30.156.178] has joined #freemyipod
13:08 -!- MrShlee [~Default@182-239-156-197.ip.adam.com.au] has joined #freemyipod
13:28 -!- user890104 [~Venci@212.233.135.74] has quit []
13:28 -!- user890104 [~Venci@212.233.135.74] has joined #freemyipod
13:29 -!- user890104 [~Venci@212.233.135.74] has quit [Client Quit]
13:29 -!- user890104 [~Venci@212.233.135.74] has joined #freemyipod
13:35 -!- MrShlee [~Default@182-239-156-197.ip.adam.com.au] has quit [Quit: Leaving]
15:16 -!- benedikt93 [~benedikt9@unaffiliated/benedikt93] has joined #freemyipod
15:19 -!- n1s [~n1s@rockbox/developer/n1s] has joined #freemyipod
16:23 -!- Dreamxtreme [~Dreamxtre@92.30.156.178] has quit [Quit: Ex-Chat]
16:42 -!- benedikt93 [~benedikt9@unaffiliated/benedikt93] has quit [Ping timeout: 276 seconds]
16:47 -!- benedikt93 [~benedikt9@unaffiliated/benedikt93] has joined #freemyipod
16:59 -!- clustur [~logger@c-76-127-58-39.hsd1.ga.comcast.net] has quit [Remote host closed the connection]
17:00 -!- clustur [~logger@c-76-127-58-39.hsd1.ga.comcast.net] has joined #freemyipod
17:23 -!- perror [~fleury@aldebaran.labri.fr] has quit [Quit: Bye all !]
17:58 -!- Dreamxtreme [~Dreamxtre@92.30.189.198] has joined #freemyipod
19:20 < fmibot> New commit by 3theseven (r211): Get rid of some leftover on-stack sector buffers
19:20 < fmibot> r211 build result: All green!
20:40 -!- liar [~liar@clnet-p09-185.ikbnet.co.at] has quit [Ping timeout: 258 seconds]
20:44 -!- benedikt93 [~benedikt9@unaffiliated/benedikt93] has quit [Quit: Bye ;)]
20:59 -!- n1s [~n1s@rockbox/developer/n1s] has quit [Quit: Lämnar]
21:50 -!- mulenmar [~mulenmar_@74-132-43-158.dhcp.insightbb.com] has joined #freemyipod
21:58 -!- Dreamxtreme [~Dreamxtre@92.30.189.198] has quit [Ping timeout: 255 seconds]
22:01 -!- Dreamxtreme [Dreamxtrem@92.30.25.36] has joined #freemyipod
22:02 < mulenmar> TheSeven: Was finally able to get back the Nano2G issue. Last night I left it at the "Cannot find iLoader theme" menu so the batter would completely drain until everything shut off.
22:03 < TheSeven> ok... now how does it behave today?
22:04 < mulenmar> Still can't get to disk mode of course, tested from embios-debug.
22:04 < mulenmar> I did another GPIO dump, diff said it was different.
22:04 < mulenmar> I'll load it to Rapidshare
22:04 < TheSeven> hmm
22:05 * TheSeven still waits for a response from our hardware guru
22:06 < mulenmar> http://rapidshare.com/files/417932861/gpio-powerdrained.bin
22:06 < mulenmar> MD5: 16029ECC5FF80A470FA7DBD5D4C22A9F 
22:06 -!- Dreamxtreme [Dreamxtrem@92.30.25.36] has quit [Quit: Never look down on someone unless you're helping them up.]
22:08 < mulenmar> Should I try the embios-wipe again?
22:09 < TheSeven> you can try, but i doubt it will help
22:13 < TheSeven> disk mode still resets?
22:15 < mulenmar> Seems to, with embios-debug. Goes to a black screen with backlight on.
22:15 -!- davide [~davide@c-67-161-123-167.hsd1.wa.comcast.net] has joined #freemyipod
22:15 < TheSeven> a black screen?
22:15 < TheSeven> never seen that.
22:15 < mulenmar> embios-wipe does nothing but go to the that old no-iLoader-theme menu.
22:16 < mulenmar> I assume it's DFU, since lsusb sees the iPod.
22:16 < TheSeven> if anything, it usually freezes with a white screen or with the previous screen contents
22:17 < TheSeven> black => something is actively driving the lcd => can't be dfu
22:17 < mulenmar> Yep, it's DFU -- ipoddfu.py embiosldr-ipodnano2g.dfu makes emBIOS load.
22:18 < TheSeven> can you try going back to dfu and running resetftl.bin this time?
22:18 < TheSeven> how long does it take?
22:18 < TheSeven> or does it even get stuck?
22:18 < mulenmar> Instanaeously
22:18 < TheSeven> uh oh, no good
22:18 < mulenmar> oh, wait with resetftl.bin
22:18 < mulenmar> Gimme a sec
22:19 < TheSeven> does it say anything below the progress bar?
22:19 -!- Dreamxtreme [~Dreamxtre@92.30.25.36] has joined #freemyipod
22:19 < TheSeven> once resetftl ran, try running norboot.dfu or whatever we called it
22:21 < mulenmar> Says nothing below progress bar, takes ~9 seconds. After a moment, goes to a white screen.
22:21 < TheSeven> ok
22:21 < TheSeven> now norboot
22:23 < mulenmar> It's got the black apple logo.
22:23 < mulenmar> And that's slowly faded away.
22:24 < TheSeven> ok, once again...
22:24 * TheSeven thinks this time it might show a battery :P
22:24 < mulenmar> Good guess. 8~J
22:25 < TheSeven> really?
22:25 < mulenmar> Yes.
22:25 < TheSeven> dammit
22:25 < TheSeven> so we're basically stuck where we were yesterday
22:25 < mulenmar> Any luck from the new GPIO dump?
22:26 < TheSeven> a couple of changed bits
22:26 < TheSeven> some of them probably related to the charger circuitry being running
22:26 < mulenmar> Let me guess, just battery level stuff then.
22:27 < TheSeven> one of the bits i spotted yesterday now has the same value as on mine
22:27 < TheSeven> so that one is not very helpful either
22:28 * mulenmar is starting to get even more suspicious of that stupid, impossible-to-be-the-problem nvram bit
22:28 * TheSeven still wonders what could have made that flash chip go nuts
22:29 < mulenmar> Well this thing is a few years old, and I did do some tests sending 1.5GB to, and then from, the device to test throughtput between OFW and Rockbox. . .
22:30 < mulenmar> I was intending to test a patch that supposedly increased the throughput, and wanted numbers to compare to.
22:30 < mulenmar> Maybe that finally sent one of the flash chips over the edge.
22:31 -!- Dreamxtreme_ [~Dreamxtre@92.27.234.219] has joined #freemyipod
22:31 < mulenmar> Or that, combined with the Rolo running without a smooth shutdown inbetween, caused some screwup.
22:31 -!- Dreamxtreme [~Dreamxtre@92.30.25.36] has quit [Ping timeout: 252 seconds]
22:32 -!- Dreamxtreme_ is now known as Dreamxtreme
22:36 * mulenmar is starting to believe that any music player dependent on non-replacable flash media is a bad design
22:40 < fmibot> New commit by 3theseven (r212): Fix i2cwrite monitor command writing garbage
22:40 < fmibot> r212 build result: All green!
22:42 * TheSeven has a suspicion where the backlight issue might be coming from
22:44 < TheSeven> and my ipod works perfectly fine with your nvram contents
22:45 * mulenmar is now out of even stupid ideas
22:48 * TheSeven tries writing your whole pmu register values to his ipod
22:48 < TheSeven> including things that should be read-only
22:49 * mulenmar takes no responsibility for if his iPod's register values ruin your iPod as well
22:49 < TheSeven> no noticable reaction besides the backlight brightness changing
22:50 < TheSeven> and it probably messing up my RTC's time and date
22:58 < mulenmar> Hmm.
22:58 < mulenmar> Maybe it simply failed from age.
22:59 < TheSeven> flash chips shouldn't fail in such a weird way
22:59 < mulenmar> I don't suppose there's a way to figure out which chip failed, completely ignore it's existence, and fix things to work on another flash chip, is there?
22:59 < TheSeven> and i haven't seen a single ipod that was even close to the bad block count limit that the ftl can handle
22:59 < mulenmar> Better than nothing.
22:59 -!- clustur [~logger@c-76-127-58-39.hsd1.ga.comcast.net] has quit [Remote host closed the connection]
23:00 -!- clustur [~logger@c-76-127-58-39.hsd1.ga.comcast.net] has joined #freemyipod
23:00 < TheSeven> there are two chips within the same package, and from what i can tell, both switched into that weird read-only mode at the same time
23:01 < mulenmar> Ah, I see -- this isn't Star Trek so I didn't expect that idea to work. 8~P
23:02 < TheSeven> [23:14] <mulenmar> Wipe 235:0, increase number of bad blocks by one. Wipe 235:1, increase by one. Repeat forever.
23:02 < TheSeven> at least this is telling me that it isn't only one of them that failed
23:04 * TheSeven wonders what to do now
23:04 < TheSeven> wait for our hardware guru to respond and tell me where that write protect line (which both chips share) ends up?
23:04 < TheSeven> open up that bugger and measure the voltage on that line?
23:04 < TheSeven> just leave everything as is?
23:05 < TheSeven> flash something that doesn't need the storage? :P
23:05 < TheSeven> basically you have a nice little device with 176KB of SRAM, 32MB of SDRAM and 1MB of flash now
23:11 < mulenmar> Wonderful.
23:13 < mulenmar> At this point, if I knew what I was doing I'd hack together a way to use an SD card instead of the flash chips. But I don't, so all I'm left with is looking for my missing iPod Touch and waiting for the guru.
23:13 < TheSeven> you could in theory use usb-based storage :)
23:13 < TheSeven> the controller can handle a host role on the bus
23:14 < TheSeven> it would of course need modified cables and lots of coding...
23:14 < teuf> TheSeven: do you think you'd have more luck with the device if you got your hands on it?
23:14 < TheSeven> i doubt it
23:14 < TheSeven> i've poked at basically everything i could think about, and we still don't even have a glimpse of an explanation what could have happened
23:16 < mulenmar> My hypotheses: Age or a weird interaction between the Rockbox update code and the FTL, or both. Both ideas pretty much shot to pieces.
23:16 < TheSeven> it might be anything ranging from some kind of unknown gpio somewhere, to hardware damage because something flipped a direction bit of a gpio port from in to out, or just a flash chip failing in a really weird way
23:17 < TheSeven> if you had the equipment (and we would have verified that the flash chip is the culprit), you could theoretically swap that
23:17 < mulenmar> I have a multimeter and a grounded soldering pencil. 10W, I think.
23:18 < TheSeven> you would at least need a heatgun and some SMD equipment
23:18 < mulenmar> I've used it to turn a DVD-ROM drive into an Xbox1 DVD drive, install modchips, and a few other things...
23:18 < TheSeven> that way you could even upgrade it to 16GB
23:18 < TheSeven> (or even more if you find a bigger chip, I could just add it to the driver)
23:19 < TheSeven> mulenmar: this is the pcb: http://www.freemyipod.org/w/images/f/fe/Bot_annote.jpg
23:19 < TheSeven> the flash chip resides in the bottom left corner
23:19 < TheSeven> you can clearly see the two rows of pads (with that smaller set in between)
23:20 < TheSeven> those pins are a bit tiny :P
23:20 < TheSeven> or a picture with the components soldered to it: http://www.freemyipod.org/w/images/6/63/Nano_2g_bck_a.jpg
23:21 < mulenmar> Ever tried to solder the top d0 point on an Xbox? I can handle tiny. ;~P 
23:21 < mulenmar> (barely, but still)
23:22 < TheSeven> i doubt your ipod will survive it if you try that
23:23 < TheSeven> if you really want to mess with the hardware, just open it and check the voltage level on one of those pins (/WP)
23:23 < TheSeven> it's pin 19 of the package
23:24 < TheSeven> so the 6th one from the bottom left corner
23:24 < TheSeven> measure the voltage of that to either a ground pin of the chip or the dock connector case or whatever else is connected to gnd
23:25 < TheSeven> but please be really careful
23:25 < TheSeven> the hold switch ribbon is the single most likely thing to break
23:29 < TheSeven> it's probably best to follow an ifixit guide for the opening process
23:31 < TheSeven> if the voltage of that pins is somewhere around 3V, i'd blame the flash chip (one should still check some other things like whether there are /WE pulses before attempting to replace it)
23:32 < TheSeven> if not, it's something else... but what?
23:36 * TheSeven will go to sleep
--- Log closed Thu Sep 09 00:48:30 2010