Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

#rockbox log for 2014-06-23

00:05:02 Join Provel [0] (
00:13:30 Quit yosafbridge (Ping timeout: 240 seconds)
00:15:22 Quit lebellium (Quit: ChatZilla [Firefox 31.0/20140616143923])
00:21:53 Join yosafbridge [0] (~yosafbrid@
00:31:22 Join Xerion [0] (
00:37:32 Quit Guest70275 (Ping timeout: 240 seconds)
00:38:01***Saving seen data "./dancer.seen"
00:56:16 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
00:59:02 Quit bluebrother^ (Ping timeout: 240 seconds)
00:59:16 Quit fs-bluebot (Ping timeout: 272 seconds)
01:01:30 Join fs-bluebot [0] (
01:03:34 Quit ZincAlloy (Quit: Leaving.)
01:24:41 Join jhMikeS [0] (~jethead71@rockbox/developer/jhMikeS)
01:40:20 Join JdGordon [0] (~JdGordon@
01:40:27 Quit JdGordon (Changing host)
01:40:27 Join JdGordon [0] (~JdGordon@rockbox/developer/JdGordon)
01:56:40 Quit RiD (Quit: A good plan today is better than a perfect plan tomorrow.)
02:09:48 Quit zoktar (Read error: Connection reset by peer)
02:12:13 Join zoktar [0] (~zoktar@unaffiliated/zoktar)
02:35:20 Join Misanthropos_ [0] (
02:36:46 Quit Misanthropos_ (Client Quit)
02:38:02***Saving seen data "./dancer.seen"
02:41:51 Join Guest70275 [0] (
03:00:02 Quit AlexP (Remote host closed the connection)
03:52:22 Join us`0gb [0] (
04:08:35 Join ygrek [0] (~user@
04:18:17 Quit amiconn (Disconnected by services)
04:18:17 Quit pixelma (Disconnected by services)
04:18:17 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:18:17 Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:18:17 Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
04:18:17 Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
04:38:04***Saving seen data "./dancer.seen"
05:06:37 Join steffengy [0] (
05:07:33 Quit TheSeven (Ping timeout: 272 seconds)
05:09:06 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
05:09:33 Quit steffengy1 (Ping timeout: 240 seconds)
06:10:15[Saint]Regarding exFAT vs FAT32 in modern Android handsets, isn't the FS compatibility argument pretty much dead these days?
06:11:50[Saint]I mean, I use F2FS for my external storage (and internal, actually), and MTP makes it completely irrelevant that the host may or may not support it.
06:33:05 Join pretty_function [0] (~sigBART@
06:33:28 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
06:38:08***Saving seen data "./dancer.seen"
07:13:06 Quit pamaury (Ping timeout: 245 seconds)
07:53:38 Quit pretty_function (Remote host closed the connection)
07:54:10 Join fragilematter [0] (~fragilema@unaffiliated/fragilematter)
08:00:25 Quit ygrek (Ping timeout: 264 seconds)
08:00:39 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
08:13:40 Quit user890104 (Ping timeout: 244 seconds)
08:15:55 Join pretty_function [0] (~sigBART@
08:16:27 Join AlexP [0] (~alex@rockbox/staff/AlexP)
08:18:09 Join mortalis [0] (~kvirc@
08:22:07 Join kugel [0] (
08:22:07 Quit kugel (Changing host)
08:22:07 Join kugel [0] (~kugel@rockbox/developer/kugel)
08:22:39 Join ender` [0] (
08:35:25 Nick SuperBrainAK is now known as DormantBrain (~andy@
08:38:09***Saving seen data "./dancer.seen"
08:41:10 Join ygrek [0] (~user@
08:41:19 Join petur [0] (~petur@rockbox/developer/petur)
08:44:30 Quit pretty_function (Remote host closed the connection)
08:53:41 Join Sparkie [0] (
08:54:00 Nick Sparkie is now known as Guest51238 (
08:57:04 Part Guest51238
09:01:11 Join ygrek_ [0] (~user@
09:03:56 Quit ygrek (Ping timeout: 245 seconds)
09:04:37 Join einhirn [0] (
09:26:01 Quit us`0gb (Quit:
09:32:27 Join pystar89 [0] (
09:34:10[Saint]I've been looking at
09:34:31[Saint]saratoga_: have you seen this? Is it useful to us?
09:36:29[Saint]Well...shit. That's somewhat creepy.
09:36:47*[Saint] just found the backlog where saratoga found the exact same thing
09:41:25 Join Sparkie_ [0] (
09:42:45Sparkie_I installed rockbox on my 7th gen ipod classic and my battery life is appalling when i have the screen on, 1% every 5 seconds, but it stayed on playing music on shuffle with no display for around 15 hours over night and today. How do i fix the battery life?
09:43:32 Quit JdGordon (Remote host closed the connection)
09:45:25[Saint]The battery indicator is based on voltage, so, its likely your battery is just fucked.
09:45:52Sparkie_should i wipe rockbox and put apple back on and ask for replacement?
09:46:34[Saint]Hahaha. Good luck with that. Its a decade old.
09:46:41[Saint]If not more.
09:48:16Sparkie_oh no i bought it two days ago
09:48:47kugelthe ipod 7th gen is not a decade old
09:49:46kugeland apple still sells it
09:50:07kugelSparkie_: if it's that new the battery is probably fine
09:50:39kugelperhaps the calibration data is just inaccurate
09:50:54[Saint]I don't see how it could be anything other than the battery.
09:51:05[Saint]Its possible to configure it poorly, but, not *that* poorly.
09:51:11kugeldid you try if the "1% every 5 seconds" continues until the battery is completely empty?
09:51:53[Saint]And, yes, the calibration data is bullcrap, but I haven't ever seen the battery on any of my Classics plummet at a rate anywhere close to 1% every 5 seconds.
09:52:03Sparkie_the battery stays on 84% for a while and then at 0% for about 15 minutes with the screen on
09:52:43Sparkie_im thinking it could be to do with the calibration because when it charges it goes really fast then really slow
09:52:53Sparkie_and stays on %'s for a decent amoutn of time
09:52:58[Saint]Even if I configure my Classic as badly as I possibly can, I can't get the batery to drop that fast.
09:53:02Sparkie_oh and it fluctuates
09:53:10[Saint]I set the LCD to always on, and the HDD spindown time to 10 minutes.
09:53:16 Join edhelas [0] (~edhelas@2001:981:e7ba:1:863a:4bff:fe85:8a3c)
09:53:18[Saint]And its still not dropping anywhere near that rate,
09:54:08[Saint]Even other things I can think of that add to it like maxing out all the EQ bands, and enabled accessory power supply don't come close to making the batteyr drop as fast as OP claims his does.
09:54:19[Saint]And my batteries are very well used.
09:54:59kugelSparkie_: i think you should check if it's bad with the OF too
09:55:22[Saint]Hmmmmm...what theme are you using?
09:55:32Sparkie_flying spaghetti monster
09:55:40[Saint]Its highly possible the author fucked up the display.
09:56:03[Saint]Improbable, but, certainly possible.
09:57:07Sparkie_ok i will restore back to apple firmware through i tunes
09:57:40kugelSparkie_: i don't think the battery is bad at this point, but better be sure by checking with the OF
09:58:19Sparkie_ok yes thank you very much
09:58:26[Saint]Hmmm, nope, the theme is displaying battery in a sane way.
09:59:08[Saint]Odd indeed.
09:59:43Sparkie_hold on guys, my ipod seems to be charging fine now
09:59:51Sparkie_at a proper rate i mean
09:59:56Sparkie_maybe it calibrated by itself
10:00:06[Saint]Doesn;t work that way.
10:00:18Sparkie_yer yer i know bu ti have no idea what is happening
10:00:37kugelSparkie_: I also think you need a few more charge cycles to reliably tell that the battery runs empty fast. could be that it wasn't fully charged
10:01:23Sparkie_yer sure thank you and i will do
10:01:55Sparkie_my battery is charging at a seems to be normal rate now instead of some stupid % per 2 seconds
10:02:43[Saint]Do you mean *discharging*?
10:02:49Sparkie_no no chargin now
10:03:04[Saint]Ok. Well, that's new information.
10:03:52[Saint]Are we positive about the "loses 1% every 5s" statement?
10:04:02Sparkie_it was before
10:04:12[Saint]Simple math seems to suggest that you'd be getting 8 minutes of battery at that rate.
10:04:19[Saint]I'm willing to believe that wasn't the case.
10:04:54Sparkie_when the screen was on. when it wasn't 80% to 30% shuffling music through earphones for about 15 hours straight
10:05:25Sparkie_saint, it was doing that for a while and then slowing down every now and again and stayed on 0% for about 15 minutes
10:06:45[Saint]How long is "for a while"? Even if it did it for a period of 30s, that's 6% battery lost.
10:06:58[Saint]Is it possible at all that the perception of time was warped?
10:07:46Sparkie_Oh my god dude. I am saying that before, the battery was decreasing fast. But sometimes it would decrease slower
10:08:41 Quit ygrek_ (Ping timeout: 255 seconds)
10:09:31Sparkie_It charged super fast aswell.
10:10:44[Saint]The impression you gave, was that whenever the screen was on the battery would reliably discharge at a rate of ~1% ever ~5s, I'm just trying to ascertain how close to reality that is. Because as noted, at that rate, you've only got 8 minutes of battery with the screen on.
10:11:26Sparkie_yes ok . It was doing that for times, but not all the time
10:11:37Sparkie_that's why i said it fluctuates
10:11:42Sparkie_it's puzzling me too
10:13:41[Saint]It definitely puzzles me. I attempted to configure my iPod in the worst possible way and I can't get anywhere near that kind of discharge rate.
10:13:47[Saint]Something odd is going on.
10:14:11Sparkie_i orignally thought it was a battery that could hold very little charge but i think it must be something else.
10:14:24Sparkie_also do you happen to know how many miiliamphours 7th gen ipods hold
10:14:56[Saint]600, IIRC.
10:15:12Sparkie_ok i think it said 500 on battery stats
10:15:15Sparkie_on rockbox
10:15:38[Saint]That setting isn't related to any battery level displayed to the user at any time.
10:15:48[Saint]That setting is only related to the estimated runtime.
10:15:54[Saint]Which is hilariously inaccurate anyway.
10:16:32[Saint]So, errr, what I'm saying is, changing it will have absolutely no effect.
10:16:39Sparkie_ok yep
10:17:04Sparkie_my ipod charging seems allright now. It might have just taken some cahrge/discharge cycles as kugel said
10:19:01TheSevenanother thing on my todo list...
10:19:24TheSevenfinally figure out where apple is hiding that current sensor so that we can get more accurate battery readings
10:20:09 Join JdGordon [0] (
10:20:18Sparkie_Oh cool. theseven are you part of dev team?
10:20:36TheSevenI'm the ipod nano 2g & ipod classic port maintainer
10:20:50Sparkie_oh my that's really cool
10:21:02Sparkie_ty for the rockbox :0
10:22:09TheSevenSparkie_: I'd *guess* that the battery runtime will be fine, just the meter is probably a bit off
10:22:43TheSevenso if it works for several hours while pretending that the battery would be empty, I just wouldn't worry about that at this time
10:23:05TheSevenif it's very new, it might also be that the battery just hasn't been run in properly yet
10:23:30TheSevenespecially if it has (somewhat likely) been sitting in a warehouse for years before being sold
10:23:50Sparkie_yer I guess. I'll keep rockbox on for like another week or so then change back to apple and if still hasnt fixed i will take back in 30 day warranty babyyyy
10:23:57Sparkie_yer that's true
10:24:44TheSevenanother thing is that the battery won't even try to charge before it's (actually, even if the meter says otherwise) below 80% or so
10:25:11TheSevenhowever plugging it in (and thus removing the load on the battery) will make the meter fluctuate quite a bit even in that area
10:25:42Sparkie_Ok, it just went from 100% to 95% in abuit 10 seconds
10:26:06TheSevenwhat does the voltage reading in the battery debug screen say?
10:26:13TheSeven(system => debug => battery)
10:27:04TheSeventhe percent gauge is nowhere near accurate, it has >10% tolerance in a lot of situations, especially shortly after changing the power source
10:27:18TheSevenit will vastly overestimate the battery state while charging, and slightly underestimate it while discharging
10:27:25Sparkie_view battery?
10:27:46TheSevenso if you're just playing with the 80-100% range, this kind of behavior is just normal at this point
10:28:09Sparkie_it's 3.977V atm
10:28:27TheSevenok, it should actually start charging at that point if you plug it in
10:28:59TheSevenhowever the meter will likely just jump straight to 100% if you do so, even though fully charging it from that point would take an hour or so
10:29:24Sparkie_ok it is jumping all over tha place on the view battery
10:29:27TheSeventhis is something that's somewhat likely to get fixed during the next weeks, as I'm currently trying to clean out the last few remaining quirks of the ipod classic port
10:29:36Sparkie_oh cool
10:30:06TheSevenSparkie_: I'd expect it to jump to about 4.15-4.20V if you plug it in, and back to about 4.00V if you unplug it
10:30:15Sparkie_where will that be available from?
10:30:38Sparkie_yer you are right
10:30:45Sparkie_i meant the voltage is jumping
10:30:52TheSevenok, you're just seeing the normal battery meter confusion then ;)
10:31:21Sparkie_pwr status is discharging when it is plugged in?
10:31:22TheSevenyou can't properly tell the exact state of charge from just the voltage, but without access to the current sensor it can't do much better
10:31:57TheSevenyes, that's also just not implemented yet. it will always say discharging.
10:32:08TheSevenanother thing that I already have a fix for, that just didn't go into the official builds yet
10:32:15Sparkie_oh ok. so are you saying my problem is pretty normal but will most likley be fixed next patch?
10:32:31Sparkie_oh coool dude thanks for that
10:33:13TheSevenI'd expect this kind of things to be fixed in the development builds in about 1-2 months
10:33:30Sparkie_how long should i keep it charging after 100%?
10:33:39TheSevenand once we've reached that point there will very likely also be a new emCORE release which embeds these newer builds
10:34:06TheSevenSparkie_: about an hour should be sufficient to get it close to actual 100%, however it's safe to just keep it plugged in overnight ;)
10:34:32Sparkie_ok yer i might do that
10:35:32TheSevenalso note that if the voltage was above ~4.00-4.05V while plugging in, it won't start charging (you'll see that the voltage goes back up to about 4.1V, but not close to 4.2V)
10:37:40Sparkie_ok so, one last time, the reason my batteyr was discharging fast and then slow may have been because it wasn't fully charged and therefore not at the given percentage?
10:38:12***Saving seen data "./dancer.seen"
10:41:15Sparkie_is that right?
10:41:21 Join pretty_function [0] (~sigBART@
11:01:24 Join ygrek_ [0] (~user@
11:04:07TheSevenSparkie_: techincally not 100% correct, but the behavior is similar, yes
11:04:40Sparkie_ok cool. so what do you think i should do from here?
11:04:46TheSevenactually the battery meter is slightly underestimating the state of charge while discharging, and massively overestimating it while charging (but only if it actually charges)
11:05:08Sparkie_ok and the bettery meter is trying to coirrect itself hence the fast discharge?
11:05:18TheSeventhe meter will "jump" (not instantly, but during a few minutes) between those different estimates
11:05:57TheSevenonce we have current sensor readings, the meter can actually calculate how big these over/underestimations are and correct them
11:06:42*kugel wonders why other targets can get largely-reliable battery readouts without current sensor
11:07:28TheSevenbecause they aren't dealing with 300+ mA discharge current spikes I guess ;)
11:07:42TheSevenor just because nobody cared about the meter accuracy
11:08:25TheSevenbut even on nano2g adding the current sensor to compensate for the battery's internal resistance vastly improved the readings
11:08:52TheSevenespecially during charging, as you just can't tell the battery state from just the voltage if it's charging in the CV phase
11:09:45TheSeventhis one is probably just as "largely reliable" as the others ;)
11:10:53*TheSeven spotted some code in disk mode yesterday that measures the USB D+/D- lines and thus must be dealing with the mux in front of the ADC that is also responsible for battery current measurements
11:13:02TheSeven(this would also enable us to properly detect USB charging spec compliant chargers btw)
11:21:36Sparkie_where will i find the rockbox update if it come sout next month?
11:22:31copperwhere do you usually "find" it?
11:22:38TheSeven will always point to the newest development builds
11:22:56TheSevenand if we consider the latest changes somewhat complete and stable, we'll release them on
11:25:31Sparkie_where would be the ipod classic 7th gen one be though
11:26:12TheSevenhm, looks like the link is missing indeed
11:26:46Sparkie_oh ok ty
11:26:54Sparkie_and this is the latest one with the battery issue/
11:27:12copperthere's a battery issue?
11:28:08TheSeventhis is the most recent development build, which doesn't have any battery meter and charging state detection improvements yet
11:28:23TheSevenI guess these will go in there within the next month though
11:28:49Sparkie_ok yer i had this on before but i changed to the one from rockbox utility after the battery issue
11:29:54TheSeventhe rockbox utility will do nothing else, it will just download that file ;)
11:30:29Sparkie_hmm ok. looking at the graph on view battery on my ipod, the gradient at which the battery is decreasing is really steep
11:31:33Sparkie_even after charging for over anhour after 100% is this normal?
11:33:16Sparkie_oh wait that's voltage my bad
11:35:55Sparkie_anywaays im of ty for the help. I'll look forward to seeing new rockbox. goodbye
11:41:23 Quit Sparkie_ (Ping timeout: 255 seconds)
12:10:37 Quit JdGordon (Ping timeout: 272 seconds)
12:15:30 Quit mc2739 (Read error: Connection reset by peer)
12:16:04 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
12:16:38 Join user890104 [0] (Venci@unaffiliated/user890104)
12:31:42 Join krabador [0] (~krabador@unaffiliated/krabador)
12:38:15***Saving seen data "./dancer.seen"
12:44:11 Quit pretty_function (Remote host closed the connection)
13:04:23 Join JdGordon [0] (
13:04:27 Quit JdGordon (Remote host closed the connection)
13:15:27 Join wodz [0] (
13:17:07wodzwhat is the difference between target_id in configure and MODEL_NUMBER define in target specific config file?
13:20:47 Quit bcobco ()
13:29:35wodzhmm, we have duplicates in MODEL_NUMBER - hm60x, clip zip 79; xfi2, hm801 82; xfi3, ma9 83; ma9c, yp-z5 84; zenv, ihifi760 92; android, nokian8xx, nokian900, pandora, yp-r0, sdlapp 100
13:29:43wodzI guess this is not good
13:29:50mortaliswodz: IIRC MODEL_NUMBER used in bootloader to check compatibility of bootloader and main binary, and target_id is the same thing but for voicefiles
13:31:54wodzso we shouldn't have duplicates in MODEL_NUMBER
13:32:06wodzit doesn't make much sense for app builds also
13:34:21 Join bcobco [0] (~bcobco@
13:35:29wodztarget_id is exported as TARGET_ID
13:39:10mortalischanging MODEL_NUMBER might be a prblem for some targets, as it will require to release new bootloaders
13:42:53wodzI wonder why there are two distinct numbers while vast majority of targets have this unique
13:46:41 Join ZincAlloy [0] (
14:07:41 Quit bcobco (Remote host closed the connection)
14:08:06 Join bcobco [0] (~bcobco@
14:10:54pamaurywodz: modelnum is for the bootloader and target_id for apps
14:11:16wodzwhat you mean apps?
14:11:42pamaurycodecs, plugins, voicefiles
14:11:52pamauryI think each codec and plugin has the target id in the header
14:12:19wodzok, but why it needs to be different then MODEL_NUMBER?
14:12:36wodzwhy two separate numbers
14:13:17 Join Rower [0] (
14:13:25pamauryno good reason I think
14:14:22wodzpamaury: btw. I looked at cp3 OF and there is gpio toggling in codec related functions. I tried this but without luck. I need to study more gpio setup :/
14:15:02pamauryok, so there might be power/hp gates controlled by gpio ?
14:15:28wodzcould be
14:16:14pamauryalso the model_num is in two places: configure and target config file
14:16:32pamaurybecause some targets don't use it except for rolo
14:17:28pamauryalso the scramble tool doesn't always make use of the model_num values, like for RKW I think
14:19:34 Quit Rower (Ping timeout: 240 seconds)
14:20:26 Join Rower [0] (
14:25:08kugelpamaury: i pushed the 24bit driver
14:25:39wodzpamaury: in configure there is target_id in target config file there is MODEL_NUMBER
14:30:33 Quit ygrek_ (Ping timeout: 240 seconds)
14:32:17 Quit copper (Ping timeout: 240 seconds)
14:32:35wodzTheSeven: (HEAD + g#843 + g#844) with rockbox official compiler
14:32:38fs-bluebotGerrit review #843 at : Fix cache coherency on ARM940T (and other ARMv4T cores). by Michael Sparmann
14:32:38fs-bluebotGerrit review #844 at : usb-designware: New USB driver for Synopsys DesignWare USB OTG core. by Michael Sparmann
14:33:08wodz+ bunch of 'function declaration isn’t a prototype' which I fixed locally
14:37:37 Join Sparkie_ [0] (
14:38:19***Saving seen data "./dancer.seen"
14:42:17 Quit Sparkie_ (Ping timeout: 255 seconds)
14:49:26pamaurywodz: but the modelnum is duplicated in configure when given to scramble as an argument
14:49:45pamaurykugel: yeah I saw, I modified the zen to use it, it works fine
14:49:54pamauryI'll push it soon
14:50:17wodzpamaury: I don't understand
14:50:54pamauryin configure, there are lines like "$rootdir/tools/scramble -rkw -modelnum=92"
14:51:31pamauryand also in tools/scramble.c, the modelnum are written for each target, that's many places to keep in sync, pretty bad
14:54:25 Quit cmhobbs (Ping timeout: 264 seconds)
14:54:57pamaurythe simplest approach would be to change the target_id, I *think* it won't cause any problem
14:56:42 Join copper [0] (~copper@unaffiliated/copper)
14:57:49 Quit bcobco (Remote host closed the connection)
14:58:13 Join bcobco [0] (~bcobco@
14:58:57 Join amayer [0] (
15:44:41 Quit petur (Quit: *plop*)
16:01:56 Quit krabador (Quit: Sto andando via)
16:07:13 Join ygrek_ [0] (~user@
16:24:56 Quit kugel (Quit: Lost terminal)
16:38:23***Saving seen data "./dancer.seen"
16:44:56 Join kugel [0] (~kugel@
16:44:57 Quit kugel (Changing host)
16:44:57 Join kugel [0] (~kugel@rockbox/developer/kugel)
16:48:50 Quit bcobco ()
16:53:16 Join bcobco [0] (~bcobco@
17:00:00 Join petur [0] (~petur@rockbox/developer/petur)
17:03:39 Quit edhelas (Ping timeout: 260 seconds)
17:04:36 Quit kugel (Read error: Connection reset by peer)
17:10:00 Quit bcobco ()
17:12:34 Quit mortalis (Ping timeout: 240 seconds)
17:13:46 Join kugel2 [0] (~kugel@
17:17:32 Quit GodEater (Ping timeout: 255 seconds)
17:18:34 Join GodEater [0] (~whoknows@
17:18:34 Quit GodEater (Changing host)
17:18:34 Join GodEater [0] (~whoknows@rockbox/staff/GodEater)
17:19:28 Join bcobco [0] (~bcobco@
17:23:50TheSevenwodz: hm, weird errors...
17:24:43wodzTheSeven: I made it compile but didn't have time to test
17:38:35 Quit mc2739 (Ping timeout: 240 seconds)
17:39:47 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
17:51:09 Join kugel [0] (~kugel@rockbox/developer/kugel)
17:51:46 Quit petur (Quit: *plop*)
17:58:43 Join edhelas [0] (
18:04:48TheSevencopper: which way does that adapter get inserted into the ipod?
18:04:48TheSevenwith the CF card towards the front or the back?
18:06:12copperI'm trying to figure out a way to explain it
18:06:27copperbut basically the ZIF interface should give you a clue
18:06:35copperit's not perfectly in the middle
18:07:09copperthat side should go against the back
18:07:17copperthe back case*
18:07:43copperyou won't be able to plug in the ZIF ribbon if you install the adapter the wrong way
18:09:01copperin other words, the adapter should be facing you like in the picture, with the front side of the ipod (with the display) against the table
18:09:19copperdunno if that's clear enough :P
18:10:02copperthe top left end of the adapter will make it fit snuggly in the front case
18:10:38copperit's designed to fit the case so it doesn't move around
18:11:06copperyou're supposed to remove the two blue bottom "bumpers"
18:11:09TheSevenok, so unlike the HDD, the adapter gets plugged with the contacts of the ZIF ribbon cable facing towards the top of the connector?
18:11:49coppernot sure what you mean, but yes, it's installed differently from the HDD
18:12:04copperyou'll know it when it fits
18:12:44copperI took pictures but I was stupid and deleted them
18:13:11TheSevenok, that little piece of metal that holds it in its position makes it more clear
18:14:38copperI don't suppose you're going to close the ipod anyway?
18:15:10TheSevenyeah, just didn't want to insert it the wrong way round, as the ZIF connector might have pins on both sides
18:15:25TheSevenand it's kinda convenient that the battery is accessible on the top side
18:15:29copperlike I said, it's conveniently off-center
18:16:13copper16:15:26 UTC <TheSeven> and it's kinda convenient that the battery is accessible on the top side
18:16:16copperwhat do you mean?
18:16:28TheSevener, damn
18:16:32TheSevenwhat did I type
18:16:35copperthe battery should be glued to the back case, not the front case
18:16:36TheSevenI meant the CF card
18:16:47TheSevenwas just plugging the battery while typing that :P
18:16:59TheSevenhm, surprise
18:17:02copperyes, so you can insert an SD card without removing the adapter
18:17:12TheSevenI just tried to plug a microdrive into it just out of curiosity, and it doesn't work (ata error 5)
18:17:44copperit wouldn't be any fun if it worked!
18:18:17TheSevenhm, interesting
18:18:27TheSevenwith just the adapter without a card, I don't even get an error
18:18:28[Saint]Seems a bit of a unnt gesture to get east access to the card without renovi the adapter considering how you have to get to the adapter in the first place.
18:18:30TheSevenit pretends to work
18:19:09coppertoo many typos!
18:19:10[Saint]s/unnt/cute/ no idea what autocomplete did there.
18:19:55*TheSeven needs to remove that empty battery detection code again
18:20:36TheSeventoo inconvenient to actually connect the battery! :P
18:22:13copperI can't remember, did you determine if it was the ZIF adapter, or the CF adapter, that used up too much current?
18:22:28TheSeventhe ZIF adapter is completely passive
18:22:32TheSeventhat's just a connector adapter
18:22:52copperhave you tried a CF card?
18:22:56copperwithout the CF adapter
18:23:12TheSeventhat's one of the things that I'll experiment with
18:23:22copperfor a while I couldn't decide if I should buy a CF card instead of an SD card
18:23:44copperbut the CF cards are a lot more expensive, even when used
18:23:47copper(second hand)
18:25:00copperand I figured that I didn't need the extra performance, since I would only access it over USB anyway
18:26:26[Saint]If you're not doing bulk transfers often, the extra performance is rather wasted.
18:26:52[Saint]Roxkbox would certainly never make use of it at all during playback.
18:27:32[Saint]Its rather surprising how slow "fast enough for lossless realtime" actually is.
18:38:26***Saving seen data "./dancer.seen"
18:50:32TheSevennothing connected: ata error 4, microdrive connected: ata error c, tarkan adapter without card: no error
18:51:24TheSeventarkan adapter with weird card: no error
18:52:02 Join bertrik [0] (
18:52:03 Quit bertrik (Changing host)
18:52:03 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
18:53:10[Saint]"weird card"?
18:54:13TheSevensome old minisd thing that works in basically no device
18:54:32TheSevenreading sector 0 on that succeeds, but returns what looks like an ATA identify response
18:54:39TheSeven"Rev 1.4 FC-1307 SD to CF Adapter V1.4"
18:55:06copperI never suspected so much could go wrong with that thing
18:55:16copperI thought, it's just an adapter
18:56:37TheSevenstorage init error 0 with tarkan adapter + some old noname 256MB card. that worked in the card reader, we have something here!
18:58:42TheSeven256MB CF card: no init error, raw access seems to work, but filesystem accesses don't. maybe just too small?
18:59:30TheSevenyeah, probably too small for valid FAT32 with 4K sectors
19:05:02copperthat's downright ancient!
19:14:34[Saint]I have a 6MB SD card somewhere
19:16:05*TheSeven has a 16MB card somewhere :P
19:22:06 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
19:23:17 Join lebellium [0] (
19:25:46 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
19:26:58 Quit APLU (Ping timeout: 272 seconds)
19:31:24 Quit pamaury (Ping timeout: 272 seconds)
19:31:32TheSeventarkan adapter + 256MB noname card: works after sorting out some FS issues
19:32:20TheSevennote: it seems like the emcore installer messes things up if there's already a FAT filesystem on the card and it's not an initial installation
19:32:49TheSevenyou need to manually tell it to reformat the data partition in that case, or it will fail to mount
19:34:39 Join model [0] (
19:37:35 Quit bertrik (Ping timeout: 240 seconds)
19:38:03 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
19:40:21 Join RiD [0] (
19:42:48 Quit Rower (Ping timeout: 272 seconds)
19:43:56 Quit jhMikeS (Ping timeout: 245 seconds)
20:21:39 Quit y4n (Disconnected by services)
20:21:43 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
20:26:21 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
20:32:05 Quit ygrek_ (Ping timeout: 240 seconds)
20:38:20*TheSeven spots some rather idiotic bugs that might have wasted several days of his time
20:38:29***Saving seen data "./dancer.seen"
20:40:59TheSevenok, microdrive compatibility fixed
20:41:54TheSevencopper: how did you manage to get hold of tarkan?
20:43:57wodzTheSeven: <- this is my 'fix' for official rb gcc version. The problem is the driver doesn't work on my n2g (with and without g#842)
20:43:59fs-bluebotGerrit review #842 at : Revert "Change control handling to start expecting host packets before sending data... by Michael Sparmann
20:44:38TheSevenwodz: but with the cache coherency patch?
20:44:40TheSevenhow does it fail?
20:44:52wodzTheSeven: yes with coherency fix
20:46:35wodzTheSeven: (with HID turned off and g#842 applied)
20:46:37fs-bluebotGerrit review #842 at : Revert "Change control handling to start expecting host packets before sending data... by Michael Sparmann
20:47:42TheSevenok, so it survives enumeration...
20:48:20wodzTheSeven: not always it seems
20:48:37TheSeventhe behavior suggests that there are STALLs
20:49:12TheSevenwhich could be a sign of cache trouble
20:49:28wodzTheSeven: the other time I got
20:50:02TheSevenyep, that's EP0 stalls
20:50:13TheSeventypical behavior that I was observing *without* the cache coherency fix
20:50:45wodzTheSeven: could you give me build with your gcc so I could test if this my n2g or gcc + fix + whatever
20:51:15wodzTheSeven: 'Fix cache coherency on ARM940T (and other ARMv4T cores).' is definitely applied
20:59:47 Join APLU [0] (~mulx@2a01:e34:ee29:12b0::10)
20:59:57wodzTheSeven: your build enumerates correctly but drive doesn't pop up
21:00:01wodzno error in dmesg
21:00:17wodzah, and device hard freezes
21:00:39wodzand then there are errors
21:01:45[Saint]Yep. Same here.
21:02:04[Saint]Enumerates, never mounts, hard freeze.
21:02:32wodzTheSeven: the same on w7 and linux
21:02:59[Saint](I used the supplied build too...cos, honestly, fuck sorting out getting the patch set to compile with the Rb toolset.
21:03:43[Saint]But the backlog suggests wodz got a lot further than I did in battling compile errors.
21:03:51[Saint]And, to no avail it seems.
21:05:06TheSevenhm, I'll have to have another look after my replacement LCD arrives
21:07:39wodzTheSeven: sorry to disappoint you :/
21:07:52[Saint]It mounted!
21:07:59TheSevenwodz: better catch that sooner than later
21:08:10[Saint]Buuuuuut, this time the host froze. :)
21:09:37wodzTheSeven: can usb_dw_get_max_transfer_size() be called agains ep0 ? If so this is plain wrong.
21:09:59TheSevenno, that function is completely counter-intuitive
21:10:16TheSevenit returns the number of packets that can be sent at once on a bulk endpoint
21:10:21TheSevennot the number of bytes
21:11:03TheSevenwell, not quite sure what it's supposed to do in rockbox, but that's what it's doing, and it seems to work on the classic ;)
21:12:34wodzTheSeven: anyway makes it compile cleanly with rb gcc and I think should be correct
21:12:57TheSevenyeah, as kugel said we should get rid of those bitfields regardless
21:13:01TheSevenor at least the hw-related ones
21:13:23TheSeventhanks for pinpointing the cause of that error though :)
21:14:27wodzTheSeven: yeah, I am all for this but this is a lot of work to debitfildize the driver
21:18:21TheSevencopper: seems like I can reproduce your behavior with a 32GB sandisk extreme SDHC card
21:18:41TheSeventhat's the only card that I have that exhibits this problem though
21:19:30gevaertsIs it the only one that claims to be extreme?
21:19:40gevaertsYou can only expect such cards to be outliers :)
21:22:22wodzgevaerts: Could you comment maybe my question why there are two different numbers describing target in our build system? (I mean MODEL_NUMBER in target's config .h file and target_id in configure)
21:22:38gevaertswodz: I have no idea
21:22:44copperTheSeven: do you mean that, you have other sandisk cards that work?
21:23:10copperor rather: what cards did you test, that do work?
21:23:11TheSevenno, it's the only sandisk card that I have
21:23:16TheSevenfrom what I can tell the CF adapter's firmware is locking up somehow
21:23:30TheSevenit doesn't recover until a power cycle even if I swap the card
21:23:30wodzgevaerts: Who may I ask about this?
21:23:41gevaertswodz: someone who was around before me :)
21:23:47gevaertsZagor, maybe?
21:23:50TheSevennot sure if it's supposed to handle swapping cards during operation though
21:23:58gevaertsPossibly amiconn?
21:25:09wodzamiconn is rather inactive and hard to catch
21:31:15*TheSeven wonders how it can be possible that we lock up that adapter with only some SD cards, but it still works fine with the very same card in a card reader
21:32:38copperwhat works fine
21:32:42copperthe CF adapter?
21:32:55TheSevenif I put it into a CF card reader, it works just fine with the sandisk card
21:33:00TheSevenlike it did for you with the original firmware
21:33:09TheSevenbut emcore somehow locks it up
21:33:19TheSevenbut only with the sandisk card. hm...
21:33:19copperthere's still the tarkan adapter between the two
21:33:33TheSeventhat adapter isn't more than a few pieces of wire
21:33:34copperyou didn't answer my question: what cards work for you?
21:33:47TheSevensome old 256MB noname crap
21:34:18TheSeventhe 2GB minisd card doesn't work in a card reader either and behaves differently (as if no card was plugged in at all) in emcore
21:34:39wodzwait isn't tarkan cf<->sd thing? If so there must be mcu inside (or cpld/fpga)
21:35:02copperit's zif → cf → sd
21:35:04TheSevenwodz: it's a passive ZIF => CF adapter + active CF => SD adapter
21:35:19TheSevenand the CF => SD card works in a card reader with the very same card that fails in the classic
21:35:33wodzoh, thats weird
21:36:21copperthat fails with emcore*
21:36:32 Quit bcobco ()
21:37:12TheSevenit must be some compatibility thing that involves *both* sides of the adapter
21:37:14copperalso note that the first thing I did, was to format the entire sd card as FAT32, labeled "iPodClassic", then I put the card into the adpater
21:37:51TheSevencopper: the working lexar card had roughly the same capacity as the failing sandisk one, both way above 32GB, right?
21:38:08TheSevendisabling DMA doesn't help at all
21:38:20copperthose are the cards I tried
21:38:24coppertop row: fails
21:38:27copperbottom row: works
21:38:43TheSeventhe top center one is the one that I'm experimenting with right now
21:38:51copperthe final 128 GB Lexar card (which worked) isn't on the picture
21:39:57TheSevenso we can rule out any capacity / sdhc / sdxc based thing as well
21:43:29wodzTheSeven: Do you have any documentation/reveng which sdio signals map to which ata lines?
21:44:00TheSevenwodz: no, I'd have to probe that
21:44:29TheSevenhm or actually I can't probe it, because I don't have a CE-ATA ipod
21:45:04TheSeven[Saint]: do you have the required tools to figure out where the pins inside the CE-ATA ZIF cable end up?
21:52:56 Quit RiD (Ping timeout: 255 seconds)
21:55:05 Quit y4n (Quit: PÆNTS ØLF!)
21:56:08 Join RiD [0] (~RiD@
21:57:16 Join Misanthropos [0] (
21:57:39 Quit Misanthropos (Read error: Connection timed out)
21:58:03 Join Misanthropos [0] (
22:02:06TheSeveninteresting... I can trigger the very same behavior by unplugging and replugging an otherwise working card on the fly
22:02:51TheSevenanother side note: the adapter doesn't seem to support LBA48
22:03:41TheSevenso 128GB is the maximum card size that it can possibly support
22:04:27 Quit ender` (Quit: #define sizeof(x) ((rand() % 100 == 42) ? sizeof(x)-1 : sizeof(x)))
22:10:39 Join wo [0] (594b9570@gateway/web/freenode/ip.
22:13:15woTheSeven: As it seems that ipod side is the same in thick and thin I'd be more interested to see where sdio signals are mapped on 40pin connector
22:13:32 Quit Misanthropos (Ping timeout: 272 seconds)
22:13:45TheSevenwo: yes, but I only have access to one half of that mapping
22:13:53TheSeveni.e. I don't know where which SDIO signal is on the ipod side
22:15:08woTheSeven: why do you need to know this? You need to know where it ends up on 40pin 'hdd' flex' side
22:15:23TheSevenyeah, but I can only track ipod => 40pin here
22:15:31TheSevenbut what I actually need to do is track ceata => 40pin
22:15:50TheSeventhe only way how I can do this is tracking ceata => ipod => 40pin
22:16:10wocan't you drive sdio and look with scope on 40pin?
22:16:54TheSevenpossibly, if I had a scope that's fast enough. it's also a bit hard to tell produce any signal at all on the data pins as long as nothing responds on the CMD pin
22:16:58 Join Misanthropos [0] (
22:17:08TheSeventell it to produce*
22:18:20wook, so the only way is to compare mapping of thick and thin flex cable
22:18:40TheSevenI can figure out the mapping of the thick one, and where the CEATA identify pin is
22:18:44TheSevener, thin one
22:18:51TheSevenI need [Saint] for the thick one
22:18:55 Join rela [0] (~x@pdpc/supporter/active/rela)
22:23:40 Quit Misanthropos (Ping timeout: 272 seconds)
22:25:45 Join RiDD [0] (Ghost@
22:29:49 Quit RiD (Ping timeout: 255 seconds)
22:30:10 Nick RiDD is now known as RiD (Ghost@
22:31:46 Quit model (Read error: Connection reset by peer)
22:34:22 Join Gallomimia [0] (
22:37:55 Quit rela (Read error: Connection reset by peer)
22:38:33***Saving seen data "./dancer.seen"
22:56:12 Quit wo (Quit: Page closed)
22:56:14 Quit TheSeven (Ping timeout: 252 seconds)
22:57:56 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
23:03:12 Quit amayer (Quit: Leaving)
23:10:24TheSevenhm, sometimes I'm getting it to hand badly enough to not even allow access to the registers anymore
23:15:52 Quit wodz (Quit: Leaving)
23:29:30 Quit nyanpasu (Ping timeout: 240 seconds)
23:39:35 Quit pamaury (Ping timeout: 240 seconds)
23:44:35 Quit TheSeven (Disconnected by services)
23:44:52 Join [7] [0] (~quassel@rockbox/developer/TheSeven)
23:49:51 Quit GodEater (Ping timeout: 244 seconds)
23:53:12kugelhm the build client fetches all remotes between each build
23:53:21kugelwhy that?
23:54:40[7]hm, as in checks for new versions of them?
23:55:22 Join GodEater [0] (~whoknows@
23:55:22 Quit GodEater (Changing host)
23:55:22 Join GodEater [0] (~whoknows@rockbox/staff/GodEater)
23:55:51kugelyes. i don't know if it would also actually check out new commits that happen in the meantime though
23:56:22kugeli have 4 remotes, and fetching them takes a significant amount of time if the internet connection is congested due to build uploads
23:57:25[7]hm right, between each build is nonsense, I was thinking at the beginning of each round, which seemed to make sense
23:57:53 Join Basstard` [0] (

Previous day | Next day