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 2009-10-10

00:00:19 Quit Dege (Read error: 110 (Connection timed out))
00:00:37mc2739kugel: pong
00:00:42 Quit Horscht ("Verlassend")
00:01:46kugelmc273: can you give a test, to see if it fixes the scrollwheel problems?
00:02:40TheSeveni assume "B" in the stack view means that the thread is blocked?
00:02:48TheSevenwhat does a "+" in front of it mean?
00:07:48amiconn+ means the thread in question requested a cpu boost
00:08:34amiconnSee debug_menu.c line 120..226
00:12:15 Quit Tristan (Nick collision from services.)
00:12:34TheSevenok, so the codec and audio threadys are blocked, and audio is requesting a boost
00:13:27 Join _Tristan [0] (
00:15:42 Join FlynDice [0] (n=FlynDice@
00:16:15*TheSeven can't think any more and will go to bed now
00:26:58 Part froggyman
00:39:32 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
00:41:26 Quit JackWinter (Read error: 104 (Connection reset by peer))
00:41:52***Saving seen data "./dancer.seen"
00:46:03 Quit mt (Read error: 113 (No route to host))
00:47:20mc2739kugel: scrollwheel patch works quite good
00:47:57kugelmc2739: how's it compared to ?
00:48:27mc2739no more jumping, good slow response, good acceleration
00:49:11mc2739I'd say it is probably as good as FS #10603
00:50:36kugelI'll post a patch to it which works very well for the fuze (including hold button. finally)
00:50:40 Join JackWinter [0] (
00:51:12mc2739good work!
00:51:58bertrikkugel, oh you figured out the hold button? what was the problem?
00:52:45bertrikmy guess wass that the hold button just takes longer than the other buttons to reach its final value after the "precharge"
00:53:13kugelyea, i did
00:53:37kugelbertrik: exactly, I need a delay. But, I also need to set bit 19 (tri-state) of dbop_ctrl
00:55:12 Join Hillshum [0] (
00:56:06kugelbertrik: btw, why are you incrementing/decrementing dbop_locked? that requires reading it
00:56:56*kugel wonders what impact reading dbop_din has on the lcd framerates
00:57:19 Join GeekShado_ [0] (
00:58:35bertrikkugel, it was meant as a protection against nested locks, so that it only really unlocks when reaching 0 again
00:59:09 Quit JackWinter (Read error: 60 (Operation timed out))
00:59:19bertrikI expect the impact of reading to be very small, it's basically just a few cycles (write cycle + read cycle + extra delay for hold button)
01:00:11bertrikit can be probably be done in just over 1 microsecond
01:01:54 Quit antil33t ()
01:02:28kugelI don't think nested locks are possible
01:02:43kugelsince drawing is only done/allowed in the main thread
01:05:24bertrikthe case I was thinking of was that lcd_update might lock and then call a lcd sub-function that also calls dbop_lock (and dbop_unlock), causing the lock to be released too early
01:05:47 Join JackWinter [0] (
01:07:10kugelthe sub-functions don't do locking (at least not in SVN)
01:07:58bertrikthis was just defensive programming, just in case someone might modify the lcd code and not realise this potential problem
01:08:56 Quit DerPapst ("Leaving.")
01:09:18bertrikI'm with making it a simple boolean-like variable, if you insist :)
01:09:29kugeljust saying, reading the value makes the overhead a tiny bit higher
01:09:42kugelI don't think it matters too much
01:09:42bertrikI meant, I'm OK ...
01:11:43bertrikHow are you doing the delay? I think an elegant method would be to simply read the dbop twice (and discarding the first read), so the read delay is based on PCLK timing and relatively independent of CPU frequency / boost
01:11:47 Quit robin0800 (Remote closed the connection)
01:12:37kugelsimple busy wait
01:12:43bertrik(I'm still hoping for the possibility of a "universal" DBOP read method)
01:12:48kugelI don't find reading it twice elegant at all
01:13:24bertrika simple busy wait suffers from variable delay due to variable CPU frequency
01:14:00kugelyes, but I'd rather find a way to do udelay() instead reading it twice
01:14:17 Quit JackWinter (Remote closed the connection)
01:14:40 Quit GeekShadow (Connection timed out)
01:14:56kugelbut we are at a point now where we cannot be much faster (max cpu freq, caches working, max lcd framerate), so the delay shouldn't be needed to increased
01:16:41 Join JackWinter [0] (
01:17:06bertrikif the busy wait delay is just long enough to work correctly at Fcpu=248 MHz, it will waste unneeded time at Fcpu=62 MHz
01:17:29kugelthat's true for all delays
01:18:20bertrika DBOP read cycle is timed with PCLK, which is a constant
01:18:39kugelmaybe we can have a udelay, using the tick interrupt (handling wrap, but capping it to 5ms to keep it simple)
01:20:54bertrikhow much delay do we need really? I think the tick interrupt is way too slow for this
01:26:36 Quit FlynDice (Read error: 60 (Operation timed out))
01:27:40CIA-43New commit by kugel (r23059): Fuze/e200v2: Fix inaccuracy when moving the wheel slowly (fixes FS #10667).
01:28:36kugelbertrik: why is it too slow?
01:29:13kugelI'm not thinking of waiting for the tick interrupt to fire of, but busy waiting his actual value
01:29:17bertrikmost buttons (except hold) require on the order of a microsecond to read, the tick interrupt runs at 5 or 10 ms
01:29:45bertrikah ok, you want to read some counter value
01:30:15kugelyou just need to handle at most 1 wrap if you cap it
01:32:20 Quit MethoS- (Remote closed the connection)
01:33:10kugelor watchdog
01:35:27kugelwatchdog seems doable actually
01:36:28amiconnkugel: Nothing *guarantees* that the main thread is the only one ever drawing to the lcd. That said, no other thread should draw to the lcd during normal operation (by convention). Technically all thread could draw without confusing the lcd controller (this would probably produce gibberish in the framebuffer though)
01:37:06kugelI knew that
01:37:39amiconnBut the main thread is allowed to hand over lcd control to another thread. Iirc mpegplayer is drawing from another thread. And the greylib even does something you should never do during normal operation: it draws from an isr
01:38:18amiconnThe latter would instantly confuse the lcd controller when mixed with normal drawing from a thread
01:41:20kugelhrm, no, watchdog isn't clocked high enough
01:43:02amiconnbertrik, kugel: Can't you make the delay depend on cpu frequency?
01:43:50bertrikamiconn, I'd rather make it depend on the peripheral clock frequency, which is constant regardless of boost
01:44:35bertriklike implementing the delay by doing a dummy DBOP read (which is clocked by the peripheral clock PCLK) :)
01:44:41 Quit GeekShado_ ("The cake is a lie !")
01:45:10kugelbertrik: the dbop clock may change too in the future
01:45:49bertrikwe'll fix that problem in the future then :)
01:46:30amiconnOn coldfire some drivers are adjusting delays based on cpu clock
01:46:34kugelwe can do that with a plain delay also
01:47:34kugelamiconn: can one have a almost-accurate udelay() using nops and bein dependant on the frequency?
01:47:51amiconnThe bit-banging i2c driver for the pcf50606 adjusts the number of iterations of delay loops, and the bit.banging remote lcd drivers even switch transfer routines at a certain frequency
01:48:21 Quit ender` (" Engineers think that equations approximate the real world. Scientists think that the real world approximates equations. Mat")
01:49:08amiconnkugel: No, since there is no guarantee the loop won't be interrupted. But it's still better than consistently waiting for too long at lower clock frequencies
01:50:15kugeldoesn't the problem with being interrupt apply to other udelay()s as well (even those using a separate timer)?
01:50:17 Quit togetic ("WeeChat 0.3.0")
01:50:28 Join togetic [0] (n=togetic@unaffiliated/ibuffy)
01:51:08kugelwell, those using a separate timer wouldn't be wrong if the delay was bigger than the isr needs time
01:52:45amiconnIf you need guaranteed delays you need to disable interrupts, but that has to be done very carefully
01:52:54amiconnInterrupts should never be disabled for too long
01:53:04bertrikI think we basically just need a *minimal* time delay for DBOP
01:53:07kugelI think being accurate most of the time is sufficient enough
01:53:23bertrikso a little extra latency from an interrupt should not be a problem
01:53:24kugelas long as it's not shorter that requested
01:53:44 Join FlynDice [0] (n=FlynDice@
01:55:12 Quit bertrik ("sleep")
01:55:50amiconnIf you only need a certain minimum delay, an ordinary loop is sufficient. Changing the loop count according to cpu clock then is a matter of efficiency if the routine containing the delay is called very often
01:56:05amiconn...and the cpu frequency is variable, of course
01:57:11amiconnThe pcf i2c on H300 and X5/M5 used called every tick so efficiency was important
01:58:18amiconn(as long as no remote is connected it's not that critical nowadays, because it's then only called every tick if at least one button is pressed)
02:03:32 Quit togetic ("WeeChat 0.3.0")
02:03:33 Join togetic [0] (n=togetic@unaffiliated/ibuffy)
02:08:49kugelactually, I think I've managed to make a udelay
02:10:48 Quit Thundercloud (Remote closed the connection)
02:12:58kugelamiconn: does that look sane?
02:13:23 Quit StealthyXIIGer (Client Quit)
02:14:46 Join AndyIL [0] (n=pasha_in@
02:22:34 Quit pixelma (Nick collision from services.)
02:22:35 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
02:22:47 Quit amiconn (Nick collision from services.)
02:22:51 Join amiconn___ [0] (
02:22:54 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
02:22:58 Nick amiconn___ is now known as amiconn (
02:25:51 Quit AndyI (Read error: 110 (Connection timed out))
02:37:22 Quit Horscht ("Verlassend")
02:41:56***Saving seen data "./dancer.seen"
02:44:11 Quit efyx_ (Remote closed the connection)
02:52:47 Join Hillshum_ [0] (
02:53:03 Quit chandoo ("Leaving")
02:53:49 Quit kkurbjun (Read error: 110 (Connection timed out))
02:54:54 Quit Hillshum (Read error: 113 (No route to host))
02:58:04 Quit gevaerts (Nick collision from services.)
02:58:16 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
02:59:17 Quit Zoxc ()
03:01:21 Quit JackWinter (Read error: 131 (Connection reset by peer))
03:03:58 Join JackWinter [0] (
03:04:35 Nick Hillshum_ is now known as Hillshum (
03:05:13 Join midgey [0] (n=tjross@rockbox/developer/midgey)
03:05:50 Join kkurbjun [0] (
03:28:47 Join efyx [0] (
03:34:09 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
03:53:18 Join IceGuest_75_ [0] (
03:53:55IceGuest_75_with jailbreak for the ipod touch will rockbox be ported to ipod touch
03:54:03HillshumNot likely
03:54:17IceGuest_75_no love for the ipod touch
03:54:22HillshumIceGuest_75_: If you do it then yes
03:54:47IceGuest_75_not a programmer
03:55:08IceGuest_75_but i hear you
03:55:37HillshumPlatforms like the iPod Touch are much more suited for Rockbox running as an app rather than as a firmware replacement, and getting that to work takes a good deal of work
03:55:54IceGuest_75_good point
03:58:05 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
03:59:30IceGuest_75_have you seen jailbroken ipod?
03:59:48IceGuest_75_does it make ipod touch much better?
04:01:07HillshumIceGuest_75_: It means we don't have to deal with getting into the App Store, but that's about it
04:01:34IceGuest_75_sounds like no big deal
04:02:55HillshumWhat do you mean by that?
04:03:24IceGuest_75_sounds like jailbroken ipod touch is not that cool
04:03:53 Join Revolution [0] (
04:04:44HillshumOh. If I had one I'd jailbreak it quick, but this is getting off-topic anyway
04:05:10Revolutionwat was the question?
04:05:49HillshumRevolution: regarding iTouches
04:06:15Revolutionoh i hav 1
04:06:22Revolutionthey r cool
04:06:30Revolutionmines jailbroken
04:06:57 Quit TheSeven (Nick collision from services.)
04:07:15 Join The_Seven [0] (
04:07:21Revolutionand i think my nano 2g is almost as cool as itouch as soon as i put rockbox on it
04:07:27 Nick The_Seven is now known as TheSeven (
04:07:37IceGuest_75_i have a nano
04:07:43IceGuest_75_it's the new one though
04:07:54IceGuest_75_can't put rockbox on it right?
04:08:17HillshumOnly the first-generation
04:08:31IceGuest_75_i miss rockbox
04:08:54Revolutionsecond 2
04:09:01Revolutionnano 2g works now
04:09:12Revolutionlook it up in the build
04:10:12Revolutionwell i hav a self built debug build on my nano 2g
04:10:39 Quit Revolution (Read error: 104 (Connection reset by peer))
04:10:39 Join Revolution_ [0] (
04:10:45IceGuest_75_with the advent of 2g is 3g and 4g in the works?
04:10:49 Nick Revolution_ is now known as Revolution (
04:11:01Revolutionso it works 4 me
04:11:17IceGuest_75_with the advent of 2g is 3g and 4g in the works?
04:11:42Revolutionyep 3g and 4g is in the works
04:11:46HillshumIceGuest_75_: Not really. 3g is I think somewhat similar to the 2g, not sure about the 4g
04:11:59Revolutionthey hav execution on 4g
04:12:06Revolutionnot on 3g though
04:12:17IceGuest_75_really, is there a timeline on 4g?
04:12:24Revolutionnot really
04:12:32 Quit kugel ("Ex-Chat")
04:12:37Revolutioni think they wanna finish 2g first
04:13:10 Quit Revolution (Read error: 104 (Connection reset by peer))
04:13:10 Join Revolution_ [0] (
04:13:23IceGuest_75_thx guys
04:13:25 Nick Revolution_ is now known as Revolution (
04:13:27 Part IceGuest_75_
04:15:09 Quit Strife89 ("Bed.")
04:18:39 Quit Revolution (Remote closed the connection)
04:27:49 Quit intrados3 (Connection timed out)
04:38:34 Quit Rondom (Nick collision from services.)
04:38:45 Join Rondom [0] (
04:41:58***Saving seen data "./dancer.seen"
05:08:37 Join JackWinter2 [0] (
05:12:00 Join Strife89 [0] (
05:21:00 Quit JackWinter (Read error: 110 (Connection timed out))
05:23:16LambdaCalculus37I would not recommend using Rockbox on the 2nd gen nano just yet.
05:23:36LambdaCalculus37It's still very buggy in terms of audio playback, and some functions have not been implemented yet.
06:11:40 Join dmb [0] (n=Dmb@unaffiliated/dmb)
06:24:36 Quit midgey ()
06:24:45 Join froggyman [0] (
06:26:32 Quit efyx (Remote closed the connection)
06:28:06 Join panni__ [0] (i=hannes@
06:29:04 Quit JackWinter2 (
06:29:04 Quit _Tristan (
06:29:04 Quit panni_ (
06:29:04 Quit pcc1 (
06:29:04 Quit r00s (
06:30:02NJoinJackWinter2 [0] (
06:33:25 Join Tristan [0] (i=tristan@
06:42:00***Saving seen data "./dancer.seen"
06:47:36 Quit Kohlrabi (Remote closed the connection)
06:47:40 Join Kohlrabi [0] (
06:56:45froggyman /join #rockbox-community
06:57:41 Part froggyman
06:57:49 Quit LambdaCalculus37 ("Fwump")
07:24:48 Join Grahack [0] (
07:28:01 Quit panni__ (Read error: 131 (Connection reset by peer))
07:33:23 Quit Lss (Read error: 110 (Connection timed out))
07:43:38 Quit Hillshum ("Ex-Chat")
07:56:13 Join robin0800 [0] (
07:58:44 Quit FlynDice (Read error: 60 (Operation timed out))
08:01:04 Join stoffel [0] (
08:22:33DEBUGEOF from server (Connection timed out) (snapshot: netstuff.c line 545)
08:22:33***Saving seen data "./dancer.seen"
08:22:36***Started Dancer V4.16
08:22:36***Connected to on port 6667
08:22:36***Logfile for #rockbox started
08:22:38***Server message 501: 'logbot :Unknown MODE flag'
08:22:38Mode"logbot :+i" by logbot
08:22:38 Join logbot [0] (
08:22:38 Join yosafbridge [0] (
08:22:38 Join shai [0] (
08:22:38 Join robin0800_ [0] (
08:22:38 Join tmzt [0] (
08:22:38 Join fxb__ [0] (n=felixbru@
08:22:38 Join polobricolo [0] (n=paul@
08:22:38 Join batman52 [0] (
08:22:38 Join Llorean1 [0] (
08:22:38 Join Bob_C [0] (
08:22:38 Join hd [0] (n=jd@Wikipedia/HellDragon)
08:22:38 Join stoffel [0] (
08:22:38 Join Grahack [0] (
08:22:38 Join Kohlrabi [0] (
08:22:38 Join Tristan [0] (i=tristan@
08:22:38 Join JackWinter2 [0] (
08:22:38 Join dmb [0] (n=Dmb@unaffiliated/dmb)
08:22:38 Join Strife89 [0] (
08:22:38 Join Rondom [0] (
08:22:38 Join TheSeven [0] (
08:22:38 Join kkurbjun [0] (n=kkurbjun@rockbox/developer/kkurbjun)
08:22:38 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
08:22:38 Join amiconn [0] (
08:22:38 Join pixelma [0] (i=quassel@rockbox/staff/pixelma)
08:22:38 Join AndyIL [0] (n=pasha_in@
08:22:38 Join togetic [0] (n=togetic@unaffiliated/ibuffy)
08:22:38 Join Galois [0] (
08:22:38 Join krazykit [0] (
08:22:38 Join BlakeJohnson86 [0] (
08:22:38 Join Xerion [0] (
08:22:38 Join lyngaas [0] (
08:22:38 Join elcan [0] (
08:22:38 Join shodanX [0] (
08:22:38 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
08:22:38 Join Topy [0] (
08:22:38 Join GodEater_ [0] (n=bibble@rockbox/staff/GodEater)
08:22:38 Join pjm0616 [0] (n=user@
08:22:38 Join tha [0] (
08:22:38 Join ch4os [0] (n=ch4os@gentoo/user/ch4os)
08:22:38 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
08:22:38 Join n1s [0] (n=n1s@rockbox/developer/n1s)
08:22:38 Join faemir [0] (n=faemir@
08:22:38 Join Zarggg [0] (
08:22:38 Join saratoga [0] (i=98039f25@gateway/web/freenode/x-lqpprvomqheekqip)
08:22:38 Join Unhelpful [0] (n=quassel@rockbox/developer/Unhelpful)
08:22:38 Join liar|netbook [0] (n=liar@
08:22:38 Join linuxguy3 [0] (
08:22:38 Join niekie [0] (
08:22:38 Join gtkspert_ [0] (
08:22:38 Join FOAD [0] (
08:22:38 Join Rob2223 [0] (
08:22:38 Join Res1 [0] (
08:22:38 Join evilnick [0] (n=evilnick@rockbox/staff/evilnick)
08:22:38 Join SUSaiyan [0] (
08:22:38 Join freqmod_qu [0] (
08:22:38 Join thegeek [0] (
08:22:38 Join preglow [0] (
08:22:38 Join Hadaka [0] (
08:22:38 Join mikroflops [0] (
08:22:38 Join parafin [0] (
08:22:38 Join bubsy [0] (n=bubsy@
08:22:38 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
08:22:38 Join mc2739 [0] (n=mc2739@rockbox/developer/mc2739)
08:22:38 Join tarbo [0] (n=me@unaffiliated/tarbo)
08:22:38 Join markun [50] (n=markun@rockbox/developer/markun)
08:22:38 Join killan_ [0] (
08:22:38 Join xavieran [0] (
08:22:38 Join HBK- [0] (
08:22:38 Join shadearg [0] (
08:22:38 Join feisar-_ [0] (
08:22:38 Join Zambezi [0] (i=Zulu@
08:22:38 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
08:22:38 Join jvd [0] (n=syscrash@poipu/developer/syscrash)
08:22:38 Join Torne [0] (
08:22:38 Join blithe [0] (
08:22:38 Join ej0rge [0] (
08:22:38 Join dionoea [0] (n=dionoea@videolan/developer/dionoea)
08:22:38 Join n17ikh [0] (
08:22:38 Join jds2001 [0] (n=jds2001@fedora/jds2001)
08:22:38 Join sbhsu [0] (
08:22:38 Join aidy [0] (
08:22:38 Join rvvs89 [0] (n=ivo@pdpc/supporter/base/rvvs89)
08:22:38 Join YPSY [0] (n=ypsy@
08:22:38 Join J-23 [0] (
08:22:38 Join jordan`` [0] (i=gromit@
08:22:38 Join Dhraakellian [0] (
08:22:38 Join Tuplanolla [0] (n=jani@unaffiliated/tuplanolla)
08:22:38 Join jfc [0] (
08:22:38 Join kadoban [0] (
08:22:38 Join Shaid [0] (
08:22:38 Join janne_oksanen [0] (n=janne@
08:22:38 Join volkmar [0] (n=volkmar@gentoo/developer/volkmar)
08:22:38 Join crashd [0] (
08:22:38 Join scorche [50] (n=scorche@rockbox/administrator/scorche)
08:22:38 Join rasher [50] (n=rasher@rockbox/developer/rasher)
08:22:38 Join maraz_ [0] (
08:22:38 Join Bagder [241] (n=daniel@rockbox/developer/bagder)
08:22:38 Join avacore [0] (
08:22:38 Join Beta2K [0] (
08:22:38 Join goffa [0] (n=goffa@
08:22:38 Join rphillips [0] (
08:22:38 Join zu [0] (
08:22:38 Join CIA-43 [0] (n=CIA@
08:22:38 Join lostlogic [50] (n=lostlogi@rockbox/developer/lostlogic)
08:22:38 Join DaCapn [0] (
08:22:38 Join AB3JU [0] (
08:22:38 Join meermanr [0] (
08:22:38 Join alexbobp [0] (n=alex@
08:22:38 Join scorche|sh [50] (n=scorche@rockbox/administrator/scorche)
08:22:38 Join fish_ [0] (
08:22:38 Join redfox [0] (
08:22:38 Join Slasheri [0] (i=miipekk@rockbox/developer/Slasheri)
08:22:38 Join bzed [0] (
08:22:38 Join jon-kha [0] (
08:22:38 Join UserFriendly [0] (n=qball@unaffiliated/qball)
08:22:38 Join @ChanServ [0] (ChanServ@services.)
08:22:38 Join w1ll14m| [0] (
08:22:38 Join TDJACR [0] (n=TDJACR@Wikipedia/Thedjatclubrock)
08:22:38 Join courtc [0] (n=court@unaffiliated/courtc)
08:22:38 Join crwl [0] (
08:24:54CtcpVersion from freenode-connect!freenode@freenode/bot/connect
08:24:55***Server message 477: 'logbot #rockbox :[freenode-info] if you're at a conference and other people are having trouble connecting, please mention it to staff:'
08:25:11shaiIs it common to replace iPod's firmware with Rockbox?
08:25:17 Quit Strife89 ("If you hold a Linux shell to your ear, you can hear the C.")
08:25:52robin0800_shai: no but you can
08:26:09shairobin0800_, After using the Rockbox Utility, I still have the ability to boot into iPod's original firmware and I have leftovers in the filesystem too...
08:27:57robin0800_shai: like what?
08:31:13shaiDirectories called "iPod Computer Transfer" and "iPod_Control"
08:32:06shaiI don't know if Calendars, Contacts etc. are also from original or came from Rockbox
08:32:25shaiI also have Photos and Notes
08:33:17robin0800_shai: I think most of your questions are answered here and all those folders are ipod ones
08:34:00 Join Rob2222 [0] (
08:34:46shairobin0800_, I will go read :)
08:36:23 Quit Topy (Read error: 104 (Connection reset by peer))
08:36:38 Join Topy44 [0] (
08:44:55 Join lifeless_ [0] (n=lifeless@
08:51:40 Quit Rob2223 (Read error: 110 (Connection timed out))
08:52:25 Join snookums [0] (
09:12:33 Quit snookums ()
09:12:36 Quit GodEater_ ("leaving")
09:12:44 Join snookums [0] (
09:12:48 Join GodEater [0] (
09:17:29CIA-43New commit by nls (r23060): Clean up a bogus array and avoid some copying of static strings
09:21:07 Part jon-kha ("[IRSSI] the cure for the common client")
09:22:13 Join jon-kha [0] (i=jon-kha@
09:27:15 Quit snookums ()
09:29:50polobricolodoes the rockbox still has some freezes on the nano2g
09:37:10amiconnn1s: Why did you remove the second 'const'?
09:38:04n1samiconn: eh, oops; i'll put it back
09:38:45amiconnstatic const char * const freq_strings[]
09:39:09amiconnThe pointers are also constant - 6*4 bytes which end up in flash instead of ram when running from rom
09:39:33amiconnNice spot though (12 -> 6)
09:41:52CIA-43New commit by nls (r23061): Add second 'const' qualifyer accidentally removed in r23060
09:43:30amiconnThis could be made even more efficient by using a 2-dimensional char array instead of an array of char pointer, but that may be too ugly to be worth it
09:44:58n1sthat const pointer thing can probably be done in a few more places
09:46:55CIA-43New commit by nls (r23062): Deinline fuction and some whitespace tweaks
09:47:47n1samiconn: are there any targets that can still run from rom?
09:49:32pixelman1s: the Player, OndioSP and the H120s
09:49:32amiconn3 targets can, although I still didn't do it on the 3rd: Archos Player, Ondio SP, and iriver H1x0
09:50:25amiconnThe other archoses can as well, if they're modded with a 512KB flash rom
09:50:37n1sbtw i think we should make that h100 flash support more official, i.e supported by an official bootloader and described in the manual
09:50:40linuxstbpolobricolo: You tell us! TheSeven made a commit last night which should help a little, but I think there are still problems.
09:50:50amiconnAfaik [IDC]Dragon was the only one who ever did this mod though
09:51:31amiconnn1s: The official bootloader is still awaiting testing by LinusN :(
09:51:36polobricololinuxstb: ok i'll update my rockbox.ipod file and i tell you
09:51:46pixelmaI guess the OndioSP is the closest to lose rombox of those 3
09:52:46n1samiconn: i thought it was pretty ell tested by others by now or is that a different one than the version Slasheri released?
09:54:47amiconnAfaik Slasheri's version doesn't contain the CF related fixes, and probably other storage related ones. Imo we should release a bootloader which is based on current driver code - but testing such a bootloader may need a bdm for recovery
09:55:39amiconnI'm not sure, there might have been some more things to check as well. Quite some time has passed :\
09:57:03n1syes, the h300 bootloader is in a similar state, the released one is buggy and no new one in sight
09:58:12pixelmaI thought a newer one had trouble on some devices (e.g. petur's)
09:58:23pixelmaH300 bootloader that is
10:00:32n1syeah i think so
10:01:52 Join snookums [0] (
10:02:45 Quit snookums (Client Quit)
10:05:45amiconnOne problem I know about the H300 bootloader is that the current official one produces a nasty click when entering bootloader usb, but not during normal boot - on my H300
10:06:18amiconnLinusN's fix for that made it click on normal boot instead on mine
10:11:39n1sit also starts the OF when ac is inserted which was intentional but is very annoying and it turns off immediately if powered on with the non lcd remote
10:11:51n1sand i think it has problems with CF too
10:12:08 Quit Grahack ("Leaving.")
10:14:00amiconnVery likely. It's as old as the H1x0 bootloader
10:20:36 Join intrados3 [0] (
10:21:06 Join snookums [0] (
10:22:40***Saving seen data "./dancer.seen"
10:29:15CIA-43New commit by nls (r23063): Make function arg const pointer
10:37:13 Quit AB3JU (
10:37:43NJoinAB3JU [0] (
10:51:09 Join bertrik [0] (
11:08:33 Join DerPapst [0] (
11:16:20 Join funman [0] (n=fun@2001:0:53aa:64c:207f:2ca3:a9b7:5870)
11:16:48 Nick funman is now known as Guest98412 (n=fun@2001:0:53aa:64c:207f:2ca3:a9b7:5870)
11:17:19 Nick Guest98412 is now known as funman (n=fun@2001:0:53aa:64c:207f:2ca3:a9b7:5870)
11:22:36funmanZagor: can you fix permissions ? (i see mkamsboot-1.1RC is back ?)
11:24:18TheSevencan anybody tell me where i can find the code that is run if i click on a music file in the file browser?
11:27:20funmanTheSeven: ft_enter() in filetree.c is a good start i think
11:27:47funmanthen filetype_get_plugin() and plugin_load()
11:32:34 Quit polobricolo (Read error: 131 (Connection reset by peer))
11:40:01 Join Horscht [0] (n=Horscht2@
11:42:25amiconnfunman, TheSeven: Starting a music file doesn't call plugin_load() for sure.
11:42:31amiconnPlayback is core functionality
11:44:09funmanTheSeven: did you see ?
11:44:37TheSevenyes, i did
11:45:30TheSevenbut i'll go for more serious things first before fixing cosmetic issues like that one
11:46:16funmanamiconn: i can't see where extension comparison is made elsewhere
11:46:49TheSevenfunman: it's done with switch ( file->attr & FILE_ATTR_MASK ), whatever this does
11:48:10amiconnCorrect, and the case that's executed for an audio file is case FILE_ATTR_AUDIO
11:55:19 Join Thundercloud [0] (i=thunderc@
11:55:32CIA-43New commit by nls (r23064): Don't cast away constness
11:56:49 Join Grahack [0] (
11:57:05CIA-43New commit by nls (r23065): Slight const police and avoid copying a lang string
11:57:31funmanwho has access to the download server except zagor and Bagder ?
11:58:56bertrikn1s, I saw your commit de-inlining a function and that makes me wonder when to use inline and when not?
11:59:34n1sbertrik: explicit inline is for speed critical things IMO
11:59:49bertrik(I prefer to not use inline, unless for a specific reason)
12:01:54n1susing inline on functions that aren't critical for speed usually wastes space (unless the function is static and called few times in which case the compiler might decide not to emit the function body)
12:02:14n1ss/few/times/from few places/
12:03:02bertrika quick look at the rockbox code suggests to me that there's a lot of places where it's not really needed
12:03:42n1sfixing that might save some code then
12:04:56n1sasm stuff is of course special
12:06:20 Join barrywardell [0] (n=barrywar@rockbox/developer/barrywardell)
12:06:35amiconnIf a static function is only used once, gcc auto-inlines it anyway
12:07:12amiconnExplicit inline should only be used for speed reasons, as it makes the code larger
12:07:26n1sthat's what i was thinking
12:07:51amiconn*Sometimes* that may even make things slower (larger code may cause more cache misses), so it always needs testing
12:10:38 Quit funman ("free(random());")
12:13:01TheSevenok, I narrowed down that freeze to fd = open(trackname, O_RDONLY);
12:13:58TheSevenbtw, something's garbling part of the logf buffer
12:17:03 Quit barrywardell (Read error: 104 (Connection reset by peer))
12:17:52 Join barrywardell [0] (
12:21:50amiconnTheSeven: File system - as I kinda expected
12:22:00*amiconn guesses it's related to SECTOR_SIZE
12:22:09TheSevenyeah, half of it was icache, half of it FS
12:22:16TheSevenno, it's getting stuck somewhere in the FTL
12:22:42***Saving seen data "./dancer.seen"
12:22:52TheSeven(and thus holding the mutex, which will make the rest of rockbox lock up soon)
12:23:10amiconnThe worm that came from the can, remember? ;)
12:23:28TheSevenno, i don't even think that it's one of those worms
12:23:33 Quit faemir ("Leaving")
12:23:52TheSeventhat bug rather looks like it is coming from the FTL itself
12:24:12TheSevenooooh yeah
12:24:20TheSevendouble-locking a mutex :-/
12:24:57*TheSeven needs to look at the kernel, how this is handled
12:26:22TheSevenso far that can seems to only have contained a single worm
12:26:31bertrikin theory, a thread can claim a mutex it has already locked without blocking
12:26:58TheSevenyes, but double-unlocking that afterwards may make things go crazy
12:27:30bertrikit should only really unlock when the number of unlocks is the same as the number of locks, otherwise there might be a bug in the kernel
12:31:37TheSevenok, had a look at the kernel, this seems to be true
12:36:00TheSevenok, the lockup is happening in the lowlevel NAND driver
12:36:06 Join JackWinter3 [0] (
12:37:16TheSevenso this is evidence of either failing timeouts or forgetting to unlock a mutex somewhere earlier
12:37:19TheSevenlet's see
12:38:27*TheSeven spotted a place where the latter may happen
12:40:51TheSeventhat didn't fix it, but it's definitely a NAND mutex issue
12:44:02bertrikare you sure that the mutex is initialised at all? this isn't currently checked in the kernel
12:44:24TheSevenbertrik: I already hit (and caught) that earlier
12:44:43TheSevenand now i also caught this one
12:44:55TheSevenwriting a nand page didn't release the mutex if it succeeded
12:45:31*TheSeven wonders why that much could ever work then
12:46:15TheSevenwell, yes, at long as it was all being done from the same thread, this worked
12:50:00 Quit JackWinter2 (Read error: 110 (Connection timed out))
12:51:08CIA-43New commit by theseven (r23066): Fixed some mutex issues in the iPod Nano 2G NAND driver and FTL.
12:59:11linuxstbTheSeven: What's the state of Rockbox now? Are you still getting freezes?
13:00:28TheSevennot when opening tracks
13:00:43TheSevenstill need to see if i manage to reproduce these wmcodec i2c freezes
13:01:53TheSeventhey are still there
13:03:11TheSevenso there are 2 major issues left: these i2c problems (and fixing that interrupt loss properly) and the clickwheel problems after releasing the hold switch
13:03:18linuxstbDo you think it's worth adding a list of known bugs on the iPodNano2G wiki page?
13:04:29 Join FOAD_ [0] (
13:05:06TheSevenprobably yes
13:05:08bertrikI don't understand the while (...) wakeup_wait() construct in the i2c driver
13:05:59TheSevenbertrik: that's to poll it occasionally to keep it from blocking if an int gets lost (which happens regularly on nano2g right now for a still unknown reason)
13:07:16bertrikok, I still don't understand it and it looks suspicious
13:08:05TheSeveni just added a timeout of 10msec to the wakeup wait, and then check if it's done. if not, it'll wait again
13:09:41TheSeveni want to get rid of that in the long term, but if i remove it right now, it just locks up as soon as the first I2C request is issued after starting audio playback
13:12:32 Join petur [0] (
13:12:37bertrikI wrote the initial i2c driver, but to be honest I don't remember anymore why the INT_IIC just disables the interrupt instead of clearing it directly ...
13:13:09tmztbertrik: this is i2c host periph, not bitbang gpios?
13:13:10bertrikTheSeven, you could try to go back to the first revision of the s5l8700 driver, that one was simply polled instead of using interrupts
13:13:19bertriktmzt, yes
13:13:22TheSevenyes, that one works
13:13:36TheSevenand that's why i did that polling-interrupt-hybrid now
13:14:39 Join einhirn [0] (
13:14:45TheSevencrap. i just removed it, and it still works
13:15:00bertrikremoved what ?
13:15:07TheSeventhat while ... wait code
13:15:20TheSevenso this was probably once again some side-effect of a bug we fixed in the meantime
13:15:37bertrikI hope so ... :)
13:16:14bertrikI would really like some more debugging/checks in the kernel sometimes
13:17:13bertrikWe could make a 'debug' target with lots of checks (but possibly slower) and a 'release' target without them (basically what we have now)
13:17:26TheSevenbertrik: it already is like that
13:17:58bertrikMaybe during the pre-release freeze we could make debug version binaries available for testing
13:18:14TheSevenKERNEL_ASSERT will only do something if KERNEL_OBJECT_CHECKS is defined, which is usually the case with DEBUG defined and in the sim
13:18:44TheSevenso we should just scatter those all over the place
13:18:45*bertrik should have looked better ...
13:18:48 Quit Thundercloud (Remote closed the connection)
13:19:48TheSevenhaving them check for uninitialized mutex/semaphore/wakeup/queue/whatever structs would be a good idea, i think...
13:20:47bertrikI agree. Maybe "someone" (could be me) should make a patch as a proposal for this.
13:20:56 Quit FOAD (Read error: 110 (Connection timed out))
13:20:56 Nick FOAD_ is now known as FOAD (
13:21:17CIA-43New commit by funman (r23067): mkamsboot: support Fuze OF v 1.02.28
13:23:00TheSevenbertrik: bad news. the i2c trouble keeps coming back
13:26:54TheSevenbertrik: clearing the int properly indeed makes rockbox behave very weird
13:27:14TheSevenit seems to just read garbage some of the time
13:30:15bertrikthe driver might need a rewrite then...
13:32:34 Nick fxb__ is now known as fxb (n=felixbru@
13:34:04 Join mcuelenaere [0] (
13:39:03pixelmabarrywardell: did you read logs a bit this week?
13:39:32barrywardellpixelma: i had a quick look at them earlier today
13:39:50barrywardellpixelma: so you got usb prober working?
13:39:58 Quit yosafbridge ("Coyote finally caught me")
13:40:08 Join yosafbridge [0] (
13:42:24pixelmayeah. I posted the log too - it only shows tiny difference in the hex numbers under "Raw Descriptor", the fun part was that with USB prober running, the HID functions worked properly but still no drive appearing. Someone pointed towards a known MacOS 10.4 bug with composite devices
13:42:35pixelmalog is here:
13:45:18barrywardellpixelma: did you also try using the logger?
13:45:36 Join T44 [0] (
13:45:55barrywardellit would report any errors
13:46:33CIA-43New commit by dave (r23068): Add crypt_firmware plugin for Nano2G - this uses the hardware crypto unit to encrypt (or decrypt) a firmware image for writing to a Nano 2G's firmware ...
13:47:41pixelmabarrywardell: no, didn't know about it and was happy to get the USB prober running at all. My knowledge about MacOS is very very limited
13:48:01 Join fdinel [0] (
13:48:20linuxstbbertrik: Was that for r23068?
13:49:07barrywardellpixelma: the logger is the last tab in usb prober
13:49:40barrywardellit might report some error that would indicate what the problem might be
13:49:46pixelmaok, will try next time
13:50:23*linuxstb now has two tasks left before we have a nice user-friendly install for the Nano2G - finish ipodpatcher support for the Nano2G and write a Rockbox bootloader...
13:52:00barrywardellpixelma: on the plus side, your log on pastebin seems to suggest that the device is enumerating properly, although I don't know the significance of those two bytes being swapped
13:52:02 Quit Topy44 (Read error: 60 (Operation timed out))
13:52:41bertriklinuxstb, yes
13:54:14linuxstbbertrik: Again, mostly TheSeven's work, but yes, it's nice ;)
14:02:14TheSevenbertrik: I had a look at i2c again, and not clearing the int in the handler seems to be just the right thing
14:11:00 Join Dege [0] (n=bla@
14:12:07 Nick fxb is now known as fxb__ (n=felixbru@
14:22:46***Saving seen data "./dancer.seen"
14:23:40 Quit TheSeven (Read error: 104 (Connection reset by peer))
14:23:58 Join The_Seven [0] (
14:24:10 Nick The_Seven is now known as TheSeven (
14:27:19 Quit TheSeven (Nick collision from services.)
14:27:35 Join The_Seven [0] (
14:27:47 Nick The_Seven is now known as TheSeven (
14:33:48 Join teru [0] (
14:48:30 Quit barrywardell ()
14:49:18 Quit TDJACR (Connection reset by peer)
14:49:18 Quit intrados3 (Read error: 110 (Connection timed out))
14:53:54 Join Lynx_ [0] (
14:56:34 Join JackWinter4 [0] (
14:58:05 Join ender` [0] (i=krneki@
15:00:50 Join gb_master [0] (i=82c03212@gateway/web/freenode/x-vkclvpbpfqfaznwh)
15:08:40 Join froggyman [0] (
15:11:01 Quit JackWinter3 (Read error: 110 (Connection timed out))
15:19:24*linuxstb looks at apps/debug_menu.c and is horrified!
15:23:09CIA-43New commit by dave (r23069): Implement __dbg_ports for S5L870X targets (Nano2G and Meizus)
15:24:21 Join Strife89 [0] (
15:27:28 Quit gb_master ("Page closed")
15:34:35 Join efyx [0] (n=efyx@
15:59:47 Quit Strife89 ("Ah, got a job to do ....")
16:01:15 Join kugel [0] (n=kugel@rockbox/developer/kugel)
16:04:28kugelbertrik: I think I have a working udelay()I
16:15:15 Join darkham [0] (
16:16:18 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
16:17:17 Join panni_ [0] (
16:20:54 Part froggyman
16:21:03 Quit stoffel (Read error: 113 (No route to host))
16:22:50***Saving seen data "./dancer.seen"
16:27:37bubsyhey mcuelenaere
16:27:49bubsyI opened my Zen Vision:M some days ago, sure has got some powerful chips in there :)
16:31:30mcuelenaereyeah it does
16:32:33bubsywhere can I grab the lastest rockbox source again? The "general" one, so I can grab the ZVM source and compile it
16:32:41bubsycan't see any links in the ZVM wiki on
16:35:28 Quit bubsy ("Party time, EXCELLENT! *wring wring wring!!* Party on, Wayne! Party on, Garth!")
16:38:33 Join bubsy [0] (n=bubsy@
16:39:21 Quit darkham ("Sto andando via")
16:43:53bubsyyes, I'm there now
16:43:59bubsysvn:// rockbox
16:44:03bubsyTortioise SVN says:
16:44:13bubsyError URL 'svn://' doesn't exist
16:44:33bubsynevermind :)
16:44:39linuxstbPeople very often have problems using Tortoise - I would suggest using the command-line SVN
16:45:35 Join Gun_Smoke [0] (n=G@fsf/member/GunSmoke)
16:46:58Gun_SmokeGoing shopping today.... Any personal favorites? around 4G, FM, cheaper end?
16:48:48 Join sunflash2 [0] (
16:49:32 Join Lynx [0] (
16:50:04 Nick Lynx is now known as Guest61553 (
16:50:25 Join froggyman [0] (
16:50:53linuxstbGun_Smoke: I don't think there are any devices you can buy in a local shop which run Rockbox.
16:51:09linuxstb(see for the list of targets)
16:51:15sunflash2hey guys i just got an mp4 player out of China via ebay one of the nano ripoffs and was wondering if rockbox will work on it im iffy on trying it only cause i dont know how to restore the original firmware
16:51:17Gun_SmokeYeah I'm looking at that.
16:51:43linuxstbsunflash2: No. All devices are different. This is what is needed to run Rockbox on a new device -
16:52:00Gun_SmokeWell I'm not opposed to ordering something. While back there was a nice drop on the mailing list of a newegg deal.
16:53:03sunflash2@linuxstb thanks ill look into getting a spare to play with
16:53:19TheSevenlinuxstb: lowering the i2c bus freq seems to help
16:53:48TheSeven(even though I would think that 12kHz is ridiculously low)
16:54:12 Quit Lynx_ (Read error: 60 (Operation timed out))
16:54:13 Nick Guest61553 is now known as Lynx_ (
16:54:24mcuelenaerebubsy: ?
16:54:24bubsyI can only find Creative ZEN Vision:M 60 GB Firmware 1.21.02e (For Europe only)
16:54:31 Part sunflash2 ("Leaving")
16:54:31bubsybut I have the 30GB one :/
16:54:51mcuelenaerelook for 'Older players'
16:55:13bubsyfound it :)
16:56:16*mcuelenaere spots new Creative DAPs
16:58:09TheSevenbertrik, linuxstb: 12kHz I2C is rock solid, 24kHz already starts getting flakey. and even 96kHz seems to work for everything except changing volume while music is playing
16:59:13*TheSeven has the suspicion that calling wakeup_wait is taking longer than the actual I2C transfer
16:59:19*bubsy really wants mcuelenaere to continue hack on the ZVM port
16:59:20DBUGEnqueued KICK bubsy
16:59:38linuxstbTheSeven: Any idea what speed the Nano2G OF uses?
16:59:51CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
16:59:51*TheSeven will have a look
17:00:20*linuxstb gets frustrated in the maze of USB #defines, trying to add support for "reboot to disk mode" on the Nano2G
17:00:42 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
17:00:50 Quit Grahack ("Leaving.")
17:04:12TheSevenlinuxstb: reboot to disk mode if usb is connected, or just do it at all?
17:04:54linuxstbYes, reboot to disk mode if usb is connected.
17:05:13TheSevenwell, do you know how to detect if usb is connected?
17:05:45linuxstbProbably not, but there's a GPIO pin that toggles when USB is inserted (but that's probably just charging/power)
17:05:56TheSevenwhich one is that?
17:06:02linuxstbPCON14 & 0x8
17:06:16linuxstb(see the new "View I/O ports" debug screen)
17:06:20TheSevennice to know - i want to have a look at proper charging soon
17:06:45TheSevenrebooting to disk mode is the easier part... write "hotstuffdiskmode\1\0\0\0" to 0x2202BF00 and reset :-)
17:07:55linuxstbYes, that part is working. For some reason my detect code doesn't seem to be though...
17:18:25 Part linuxstb ("Leaving")
17:18:35 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
17:18:57linuxstbTheSeven: Also, headphone detection is PCON14 & 0x20
17:19:13 Quit teru ("Quit")
17:19:53TheSevenlooking at the apple firmware tells me that there are 2 i2c controllers, we're writing to the wrong one and it works nevertheless, and there are undocumented bits in the regs
17:20:50 Quit Gun_Smoke (Read error: 110 (Connection timed out))
17:21:30*linuxstb thinks he should use PDAT14, not PCON14... ;)
17:23:35 Join intrados3 [0] (
17:26:40CIA-43New commit by dave (r23070): Nano2G - reboot into disk mode when USB is inserted.
17:27:42bertrikkugel, nice ...
17:27:43kugelamiconn: ping
17:28:30linuxstbbertrik: Is anything holding up the Meizu ports, or is it just a lack of time/motivation ?
17:28:31bertrikTheSeven, must be some kind of timing problem then between the interrupt and the main thread, I can't imagine that the i2c hardware is unable to run at 96 kHz
17:28:33kugelbertrik: it's not very accurate for us < 10 though (timer frequency too low
17:29:11 Join ChristW [0] (
17:29:13bertriklinuxstb, I lack the time/motivation to really look into the Meizu/Samsung FTL
17:29:57bertrikkugel, I'm about to change my mind again on how to do the DBOP button read
17:30:13bertrikthe fuze is just so cumbersomely designed to use the DBOP port
17:30:34ChristWI'm looking into obtaining the source code, but the Wiki is a bit unhelpful. It doesn't tell me the svn url!
17:30:58linuxstbbertrik: Ah yes, I forgot the FTL...
17:31:12bertrikthe fuze needs a high scrollwheel polling rate AND apparently a lot of time to read HOLD button
17:31:20CIA-43New commit by bluebrother (r23071): Don't put the complete HTML source in the trace when checking for a new version. The recognized filenames is enough.
17:31:32bubsybut bertrik, I know you want to look more into the Creative Zen Vision:M!
17:31:34bubsywe're come far
17:31:41bubsysome things left, is sound and PIC
17:31:48bubsyand something with the filesystem/hdd thingies :/
17:32:06linuxstbbluebrother: Can I just check, if the Nano2G bootloader is bootloader-ipodnano2g.ipodx, is that going to be problem for rbutil?
17:32:13kugelbertrik: the wheel works perfectly, and no lcd glitches, with the patch I uploaded (the one that has the two tweaks applied). I'm not sure why you're still nagging
17:32:19linuxstbbluebrother: (an extra "x" at the end)
17:33:10kugeludelay(2); works for the hold button, 1 would probably also work but my udelay doesn't do 1us in that case
17:33:35kugelthe huge delay is in SVN too, btw
17:34:02bertrikkugel, so you mean it's still an improvement over SVN :)
17:34:10bluebrotherlinuxstb: won't be a problem at all, the bootloader classes are done to handle this :)
17:37:07CIA-43New commit by dave (r23072): Nano2G - implement "Pause on headphone unplug" feature
17:39:12 Quit Horscht ("Verlassend")
17:39:41kugelTheSeven: I believe a fix for the scrollwheel-after-hold problem for the older ipods was committed a few weeks agop
17:48:03linuxstbkugel: Thanks, but we know that. Problem is we don't know how to re-init the clickwheel (which was the fix on the older ipods).
17:51:27 Join midgey [0] (n=tjross@rockbox/developer/midgey)
17:54:18*TheSeven is back
18:01:29 Join Lss [0] (
18:02:29TheSeven780kHz I2C works, 1.5MHz fails sometimes
18:04:07 Join intrados4 [0] (n=intrados@
18:04:07TheSeven1MHz seems to be stable
18:04:57 Quit intrados4 (SendQ exceeded)
18:06:03TheSevenlinuxstb, bertrik: my suspicion is that wakeup_wait gets interrupted by the I2C int before everything is in place
18:06:21 Join intrados4 [0] (
18:06:25TheSeven(which points towards polling probably being faster than interrupt-driven I2C
18:06:45TheSevenan I2C request takes less than 1000 CPU clock cycles
18:11:02 Quit intrados3 (Read error: 110 (Connection timed out))
18:11:08TheSeveneven though the kernel code looks like it should be able to handle the wakeup coming in before even waiting for it
18:12:14bertrikI have no problem with changing it into a busy-wait, IIRC the Meizu OF does that too, I just tried to improve on it
18:12:40TheSeventhe apple OF is doing it really weird. they are polling for the interrupt handler to be called
18:13:07bertrikoh, I think we can certainly do better than *that* :)
18:13:48TheSeveneven though i would really like to find out what's causing the trouble we're seeing right now
18:14:12TheSevenbtw, your I2C is running at 48kHz, is the OF doing that, too, or are they faster?
18:14:33bertrikTheSeven, regarding max frequency, I think nearly all i2c device can handle 100 kHz and most can handle 400 kHz but I wouldn't go any higher
18:14:49TheSeveni haven't figured out what apple is using, but 1MHz seems to work flawlessly
18:16:23bertrik48 kHz wasn't my intention, I think I copied what the meizu OF did
18:17:03TheSevenwell, I assumed PCLK = 100MHz for that calculation, but I doubt it will be higher than that ;-)
18:18:08 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
18:18:16CIA-43New commit by dave (r23073): Squash a warning when FTL_READONLY is defined
18:20:16kugelTheSeven: : " they are polling for the interrupt handler to be called" <−− that's not much different from wakeup_wait, isn't it? I think wakeup_wait yields, but that's all
18:20:33TheSevenno, that's real thread blocking
18:20:52 Join toffe82 [0] (
18:20:55TheSevenand apple doesn't even yield, so it's completely pointless to poll for an int, rather for a HW reg
18:21:44TheSevenkugel: but yes, the kernel of course polls for the thread to get woken up or the timeout to expire when doing task switches ;-)
18:22:53***Saving seen data "./dancer.seen"
18:25:07TheSevenbertrik: is it ok with you if i commit a polling I2C driver running at 390kHz?
18:25:26TheSevenor should I ifdef the clock speed?
18:26:54 Quit froggyman (Read error: 104 (Connection reset by peer))
18:26:59linuxstbTheSeven: Does this finally make audio playback stable?
18:27:24bertrikI'm not fully sure that all i2c devices do indeed handle 400 kHz (meaning I haven't checked, but I'm fairly confident)
18:28:00TheSevenseems so. i haven't noticed any freezes when talking to the wmcodec at 500kHz, so 390kHz should be safe.
18:28:12bertrikI can add an #ifdef later if it turns out that 400 kHz is too fast for some other device (on the meizu or samsung yp-s3)
18:28:24TheSevenlinuxstb: that volume changing bug resolved by this seems to have been the last one
18:31:19linuxstbTheSeven: Cool! If I can get a bootloader (I'm working on it now) and ipodpatcher finished, I can't think of anything else to stop us releasing...
18:31:46linuxstbI _think_ the Nano1G manual can be used "as-is"...
18:32:41TheSeventhe FTL also seems to be a lot more stable now that I caught that mutex bug
18:32:58linuxstbWell, first step of "releasing" is moving it up to "unstable".
18:33:34*linuxstb wonders if "Usable" is a better definition than "Unstable"
18:34:46*bertrik agrees
18:34:52linuxstb(and moving some of the "stable" ports down a category, such as the earlier ipods (no charging))
18:35:15*linuxstb would like a "top-tier" category for the very best ports of Rockbox
18:35:18TheSevenah, yes, charging. another thing i want to have a look at today.
18:36:56CIA-43New commit by theseven (r23074): Revert the S5L870X I2C driver to busy waiting because of some issues with interrupts, and increase the bus speed from 24kHz to 390kHz.
18:37:13*linuxstb hopes the bootloader works first-time, as the compile, copy to ipod, encrypt using Rockbox, write to device with ipodpatcher, test cycle is a bit slow...
18:37:40TheSevenlinuxstb: test it using ibugger
18:38:06TheSeven(and boot that one through iloader)
18:38:18linuxstbBut I can't use IRAM then?
18:38:33TheSevenyou can, you'll just need to reboot ibugger afterwards to make it work again
18:38:39TheSevenbut that's like 10 seconds
18:39:32TheSeventhe rockbox crt0.S should kill iBugger on the interrupt side quite early, long before writing to IRAM
18:39:54TheSevenyou won't be able to use it for debugging, but it's still nice for uploading
18:41:26linuxstbYes, that's probably better...
18:44:47 Join froggyman [0] (
18:52:53CIA-43New commit by bluebrother (r23075): Rework configuration dialog for TTS / Encoder values. ...
18:54:31 Part ChristW
18:55:14linuxstbTheSeven: I guess we should commit that stack increase. My crypt_firmware plugin also needs it, as well as Rockboy and pictureflow
18:55:33TheSevenOK, I'll go for it
18:55:51TheSevenI'll do that for S5L870X as the other ones have even more IRAM
18:56:11linuxstbI'm not sure about that - perhaps use SECTOR_SIZE
18:56:28TheSeventhe issue doesn't seem to depend only on the sector size
18:57:03TheSeventhe pictureflow stack was way too full to work with 8KB stack even at 512
18:57:58linuxstbIt works fine on other targets with a 8KB stack though.
18:58:27linuxstbMaybe its simply the NAND/FTL code is using a lot of stack in general?
18:58:36linuxstb(or something else target-specific)
18:58:39CIA-43New commit by theseven (r23076): Fix panic caused by r23074
19:01:05TheSevenanyways, I expect that to happen on the other S5L870Xs, too, and we have more than enough IRAM. We may try to lower that later, though.
19:01:20TheSevenAny idea how to find out what's actually causing it?
19:06:14TheSevenlinuxstb: I would estimate your plugin at ~5KB of stack usage
19:06:28TheSeven(supposed the FAT driver takes about half of that)
19:06:42TheSevenoh, wait, there is another sector buffer
19:07:03TheSevenbut things should still have worked with 12KB, which they didn't
19:07:20 Quit robin0800_ (Remote closed the connection)
19:10:41 Quit linuxstb (Read error: 104 (Connection reset by peer))
19:13:01 Quit hd (Read error: 110 (Connection timed out))
19:13:38 Quit bertrik (Read error: 113 (No route to host))
19:13:56 Quit efyx (Read error: 131 (Connection reset by peer))
19:15:52*TheSeven will add a timeout for I2C requests
19:16:09TheSevenI suppose 20ms is OK?
19:16:14 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
19:18:54 Part froggyman
19:19:55 Quit linuxstb (Read error: 131 (Connection reset by peer))
19:21:09 Join lifeless__ [0] (n=lifeless@
19:21:41 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
19:22:08CIA-43New commit by theseven (r23077): Add a timeout for I2C transfers (S5L870x)
19:22:29TheSevenlinuxstb: is there a way to disable that reboot to disk mode on usb connect feature?
19:22:32 Join Hillshum [0] (
19:22:50TheSeveni somehow don't like it at all
19:23:19gevaertsTheSeven: disable in what way?
19:23:42TheSevennot have it reboot to disk mode when I plug it (e.g. to charge it)
19:24:11gevaertsthere should be a button mapped to do that
19:26:14linuxstbYes, you hold MENU to charge.
19:26:23linuxstb(I haven't tested, but that's how it works on other ipods)
19:26:59 Join Grahack [0] (
19:27:10TheSevendoesn't work
19:29:25linuxstbHmm, you're right, it doesn't...
19:29:49TheSevenI'm reverting that locally for now, as it blocks me from playing around with charging
19:29:56 Quit Hillshum (Read error: 104 (Connection reset by peer))
19:30:48 Quit Dege ("Peace and Protection 4.22.2")
19:33:23linuxstbTheSeven: You could just change it to return "false" in the USB detect function in usb-s5l8700.c
19:33:31gevaertslinuxstb, TheSeven: I think you need HAVE_USB_POWER
19:34:08linuxstbgevaerts: Ah yes...
19:34:50 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
19:35:05CIA-43New commit by rob (r23078): Initial support for runtime detection of the PMU used in newer D2+ models (PCF50635). The backlight and battery monitoring should work, but the RTC ...
19:35:14 Join adambrower [0] (
19:36:06 Quit lifeless_ (Read error: 110 (Connection timed out))
19:36:15adambroweranyone one here
19:37:11gevaertslots of people
19:38:43 Quit adambrower (Client Quit)
19:39:32 Join Strife89 [0] (
19:39:45 Nick Kohlrabi is now known as kohl^fi (
19:42:38CIA-43New commit by rob (r23079): D2: Don't read touchscreen co-ordinates when hold is enabled.
19:42:44 Join r00s [0] (
19:48:10 Nick fxb__ is now known as fxb (n=felixbru@
19:48:20 Join Hillshum [0] (
20:13:13TheSevenlinuxstb: how did you find that GPIO?
20:13:27TheSevenjust look at the values while plugging?
20:14:21 Nick Llorean1 is now known as Llorean (
20:14:24linuxstbTheSeven: Yes.
20:15:28linuxstbTheSeven: If it's like earlier ipods, then that pin will probably actually indicate "charging" or "charger attached". To distinguish between USB power only and a USB data connection you need to talk to the USB controller.
20:16:53saratogais anything known about the USB controller?
20:17:11TheSevenyes, of course, but I'm actually dealing with the charger right now, and I'm wondering if there are some interesting pins beneath
20:17:31TheSevensaratoga: it's the same one as used on the S5L8900 iphone soc
20:18:34linuxstbsaratoga: TheSeven's "ibugger" tool communicates over USB. So he already has working code for it...
20:19:23saratogaI'm really amazed at the speed of the nano2g port
20:20:03TheSevensaratoga: once we got that bootrom dump, the hardest part was done
20:20:25linuxstbAnd fortunately the encryption wasn't secure...
20:20:33saratogafunny that "porting rockbox" wasn't the hard part
20:20:38saratogausually it is
20:20:50TheSeventhe hardest part definitely was breaking into it the first time
20:20:59linuxstbWell, that took about 3 years!
20:21:23saratogawhat about the Ipod 6G/Classic?
20:21:27TheSevenusing a buffer overflow exploit on an almost unknown system... and if we hadn't found that (crippled) JTAG, we would still be trying around
20:21:54TheSevensaratoga: we can execute code on that one, but nobody has bothered actually writing code for it yet
20:22:09saratogais the same true of the nano3g?
20:22:12linuxstbTheSeven: Only the 1st gen Classic though (there are 3 generations now).
20:22:20*linuxstb has the 2nd gen Classic (120GB)
20:22:26TheSeventhe nano3g is still resisting...
20:22:56***Saving seen data "./dancer.seen"
20:23:06*TheSeven suspects they may have used the same GID key on 1st/2nd gen classic
20:23:13TheSevenand even if not, it should be bruteforcable
20:23:27*TheSeven has no idea what's going wrong with our 3g attack
20:25:17liar|netbookTheSeven: how many nano3g nanotrons are currently bruteforcing?
20:26:18linuxstbsaratoga: A lot of code came for free from other ports though. The Meizu work helped get it started, including most of the work for the audio driver. apps/ code didn't need adapting as it's the same button layout and LCD size/depth as the 1st gen Nano. The button driver is merged with the PP ipod clickwheel driver, so the wheel acceleration etc came for free.
20:26:45linuxstbIt's the same audio codec as the 1st gen nano as well.
20:30:15TheSevenliar|netbook: one, but it's not finding anything, even though we almost completely scanned it for 2 firmware revisions
20:30:56liar|netbookTheSeven: so its maybe just not working proplery
20:31:29TheSeventhe results seem sane though
20:37:33liar|netbookmaybe i can borrow a nanotron from a classmate, just to see if it gives different results
20:37:41liar|netbooka nano3g*
20:37:50liar|netbookand build a nanotron i meant
20:38:31*TheSeven will do GPIO bruteforcing now
20:43:31TheSevenno reaction from the charger :-(
20:46:25 Quit intrados4 (Read error: 60 (Operation timed out))
20:47:14 Join intrados4 [0] (
20:51:57 Quit evilnick (Read error: 113 (No route to host))
20:56:53 Quit Rob2222 (Read error: 54 (Connection reset by peer))
20:57:37 Join Rob2222 [0] (
21:00:53*pixelma wonders about the timestamp on the red line of the build table
21:03:10 Join Dege [0] (i=Dege@
21:03:23linuxstbTheSeven: I don't know if this will give you any clues, but it's the task related to charging on older ipods -
21:06:25 Quit Strife89 ("If you hold a Linux shell to your ear, you can hear the C.")
21:07:43TheSevenlinuxstb: yep, I'm searching for the corresponding GPIO
21:11:07 Quit intrados4 (Read error: 110 (Connection timed out))
21:16:30 Quit bluebrother (Nick collision from services.)
21:16:33 Join bluebroth3r [0] (n=dom@rockbox/developer/bluebrother)
21:19:08 Join Thundercloud [0] (
21:24:30*TheSeven found that gpio ports 3 and B have an influence on battery current
21:24:50TheSevenzeroing port 3 increases drain a lot, zeroing port B decreases it a little
21:26:25liar|netbooklinuxstb: "Pause on headphone unplug" does not work for me
21:26:39linuxstbliar|netbook: Have you enabled it? It's a setting (and off by default).
21:27:03liar|netbooklinuxstb: oh did not knew that
21:27:14*linuxstb hands liar|netbook a copy of the Rockbox manual ;)
21:30:34liar|netbookok it works, but its a bit out of sync
21:30:43*liar|netbook is going to read the rockbox manual
21:31:40linuxstbWhat do you mean "out of sync" ? Out of sync with what?
21:32:52liar|netbookits takes 1 second to pause
21:34:00 Join AquaFox [0] (
21:34:52linuxstbI guess that's just how it works - I've never used it before. I don't know if that's by design or not though.
21:35:00 Quit shai ("Leaving")
21:35:21 Join froggyman [0] (
21:35:34 Join Strife89 [0] (
21:36:26liar|netbookand in audio playback the counters work strange too(the right one counts down ~0.5secs after the left one counts up)
21:36:45AquaFoxis there any idea if rockbox could work on a Creative Nomad Jukebox? (i have one) this is the player that runs on 4AA batteries and looks like a personal CD player and has a standard notebook harddrive in it
21:37:25liar|netbookbut thats just a "cosmeticbug"
21:42:38 Join nosa- [0] (
21:43:26 Nick nosa- is now known as nosa-j (
21:43:59linuxstbAquaFox: It _could_ work on most things. No-one has done any of the work needed for that device though. See here -
21:46:16 Part Grahack
21:47:08 Part froggyman
21:47:16 Join merbanan [0] (
21:47:24 Join JackWinter [0] (
21:48:54 Join zu_ [0] (
21:49:18 Join avacore^ [0] (
21:52:15 Join CIA-85 [0] (n=CIA@
21:58:19 Quit nosa-j (Read error: 104 (Connection reset by peer))
21:58:32 Join nosa-j [0] (
21:59:32 Quit zu (Connection timed out)
21:59:54 Quit JackWinter4 (Connection timed out)
21:59:56 Quit avacore (Read error: 110 (Connection timed out))
22:01:10 Join moonscapex [0] (
22:01:22 Quit CIA-43 (Read error: 110 (Connection timed out))
22:02:23moonscapexhas the problem with album art overlapping images been fixed yet?
22:03:44 Quit moonscapex (Client Quit)
22:12:49linuxstbTheSeven: Can I now remove the "I2C driver is unreliable, causing freezes" known bug from the IPodNano2GPort wiki page?
22:13:59TheSevenyes, even though I'm not happy with the solution
22:14:11TheSevenI2C just sometimes fails, but this is handled more gracefully now
22:14:49TheSevenbtw, PCON3 |= 0x80000000 causes a heavily increased current draw
22:20:03TheSevenPCONB |= 0x01000000
22:20:23TheSeven...and battery current fires up like a rocket :-P
22:23:00***Saving seen data "./dancer.seen"
22:23:20TheSeven(just wrote some small python tool that connects to ibugger, tinkers with GPIOs, and monitors the battery voltage and current)
22:33:54 Quit merbanan (Read error: 110 (Connection timed out))
22:37:00TheSevenok, what next? the clickwheel?
22:37:39linuxstbI guess so. Or figure out why the stack usage is so high...
22:37:51*linuxstb can't see why his bootloader isn't working...
22:38:25TheSevenhow does it fail?
22:41:01linuxstbJust freezes on the Apple logo. I've now re-installed the Apple firmware and am using iBuggerLoader...
22:41:19TheSeveni would suggest the iloader-ibugger way, it's faster
22:41:29TheSevendid the bootloader run in ibugger? upload 08000000 bootloader.bin && execute 08000000 0a000000
22:42:05linuxstbI'm just using ibuggerloader to upload the code to 0x08000000 and then executing it. That seems as fast as it could be.
22:42:15linuxstbYes, that's exactly what I'm running.
22:42:32TheSeveni mean booting ibugger loader through the apple firmware is really slow
22:42:50TheSevenand you'll need to start up ibugger core before uploading
22:43:05TheSevenelse some usb buffers will trash uploaded data
22:43:34linuxstbAh yes, I remember now... So how can I use ibugger with iloader?
22:43:34 Join nosa- [0] (n=k@
22:44:07TheSevenif you want to do it the quick way, just dd one of my iloader images to the ipod, this already contains ibugger loader
22:44:23TheSeventhen just boot it using the menu button and run the following startup; upload 08000000 bootloader.bin && execute 08000000 0a000000
22:47:04linuxstbIs there an alternative? e.g. loading ibugger from /iloader/ ?
22:47:38TheSevenyes, you can install iloader in whatever way and feed ibugger as /iloader/ibugger.bin or a "ibgr" firmware partition image
22:47:52TheSeven(the latter is how it's done in my image)
22:48:50linuxstbWhere/what is ibugger.bin ? I can't see it in your archive. Is it just core.bin ?
22:49:15TheSevenno, it's loader.bin
22:51:34linuxstbThanks, I'll try that. Just to check, my bootloader gets loaded to 0x08000000 and then remaps IRAM to 0x0 and relocates itself there. Should that work fine?
22:52:03TheSevenyes, at least if interrupts are disabled during that, which they should be
22:52:29linuxstbI'm guessing they are when the NOR bootloader loads my bootloader. But with ibugger?
22:53:04TheSevenwith ibugger, only the USB int will be active, but you'll need to shut that one down, too (but I think crt0.S is doing this early enough)
22:54:07TheSevenwith ibugger, only the USB int will be active, but you'll need to shut that one down, too (but I think crt0.S is doing this early enough)
22:54:33linuxstbOK, this is weird. When I run those three commands, I see the iBugger loader, then I'm taken back to iloader...
22:54:41linuxstbs/iBugger loader/iBugger logo/
22:55:15TheSeventhat's really weird.
22:55:30TheSevenwithout an apple logo in between?
22:55:58TheSeventhen you somehow didn't upload your bootloader, and ibugger didn't corrupt iloader enough to stop it from booting :-P
22:56:56linuxstbYes, it seems running the commands in one command-line is too fast for it. Running them manually, and my bootloader just crashes (i.e. it stays on the iBugger logo)
22:57:19linuxstb(it displayed a "No iBugger found" error which I missed...)
22:57:48TheSevenfunny, this should work
22:58:27TheSeveni suggest changing line 268 and using the run command
22:59:43linuxstbYes, that's better.
23:00:02TheSevenand does things even faster, but without all those reconnects in between
23:00:46linuxstbSorry, no it's not better...
23:01:29TheSevenrun has always worked for me
23:01:51linuxstbDoing it immediately after startup isn't working. If I wait a short time, it's fine.
23:01:56TheSevendoes it still just boot to iloader or spit out some error message?
23:02:13TheSevenstartup is supposed to wait 2 seconds for the ipod to reconnect
23:02:19 Quit nosa-j (Connection timed out)
23:02:30TheSevenyou may possibly need to increase that, but it always worked for me (on windows) gives an error - no iBugger device found
23:03:03linuxstbYes, I'm on Linux. I'll increase that wait.
23:03:05TheSeventhen increase line 260
23:03:36linuxstbAh, you meant "run" without using startup?
23:03:46 Join evilnick [0] (
23:04:15TheSevenrun will do it all magically
23:04:26linuxstbThen yes, that's fine.
23:07:42amiconnReally strange build line...
23:10:06 Join ifonefox [0] (
23:11:25linuxstbamiconn: Hmm, the non-existent builds?
23:11:53TheSevenheh, that was my commit! :-P
23:12:01 Quit ifonefox (Client Quit)
23:12:40amiconnYes, including the timestamp
23:14:08 Quit DaCapn (Read error: 113 (No route to host))
23:14:15 Join DaCapn [0] (
23:21:33TheSevenwhat about finally adding nano2g to the build system?
23:21:34TheSevenare these builds only done for checking, or can one download them somewhere?
23:21:38 Join ShapeShifter499 [0] (
23:21:56linuxstbTheSeven: Yes, they're all downloadable.
23:22:07TheSeventhen that could be the solution to that problem
23:22:23linuxstbI'm not really sure what the criteria are for adding builds there. Apart from compiling without any warnings (which it doesn't at the moment)
23:22:37TheSevenbtw, i just spotted a shutdown freeze (caught by the 20sec timeout) :-/
23:22:49TheSevenfixing those yellows shouldn't be too hard
23:22:57linuxstbBut the Gigabeat S is there, and that's in the same category as the Nano2G - "Unusable" ;)
23:23:26TheSevenI'll have a look at that core_sleep one
23:23:36linuxstbSo yes, I would say that once we fix the warnings, it could be added.
23:23:59linuxstbAnyone know the mechanics of adding new builds? gevaerts ?
23:24:11gevaertsadd them to the builds file :)
23:24:32linuxstbDoes that then need to be installed somewhere by someone?
23:24:45linuxstbs/someone/some Swede/
23:25:38*linuxstb stares blankly at the builds file...
23:25:55*gevaerts can do that bit
23:26:30 Quit Lss (Read error: 110 (Connection timed out))
23:26:58*bluebroth3r wonders who on earth "The Rockbox Crew" on cnet is
23:27:05linuxstbgevaerts: OK ;) What's the "−−type" parameter?
23:28:43gevaertslinuxstb: that's the build type, i.e. n for normal, b for bootloader, s for sim, c for checkwps
23:28:54linuxstbI guessed that, but the number?
23:29:23gevaertsthat's another column :). It's the score
23:29:55linuxstbHow difficult the build is?
23:30:33linuxstbOK. It's a bit clever that build system...
23:31:04ShapeShifter499which audio format takes up less space that can be read by rockbox?
23:31:38TheSevenShapeShifter499: mainly depends on what your quality assumptions are :-)
23:32:15gevaertsfor the nano2g normal build, score is 22606
23:32:46ShapeShifter499quality would be so-so, as long there are no horrid clicking or static in the music
23:33:01linuxstbShapeShifter499: If you want a low bitrate codec, HE-AAC could be worth testing - the Nano2G's CPU might handle it.
23:33:29ShapeShifter499no I want more music in less space
23:33:48linuxstbDon't you mean more music in the same space?
23:35:06gevaertsTheSeven, linuxstb: the nano2g bootloader doesn't seem to compile
23:35:13linuxstbgevaerts: No, it doesn't exist yet...
23:35:23gevaertsah, ok
23:35:26linuxstb(we're using "iloader" - TheSeven's own bootloader)
23:35:33*gevaerts thinks that those people are doing things backwards :)
23:35:35*linuxstb is working on a Rockbox bootloader ATM
23:36:17TheSevenis there any place in rockbox where you may need current measurements?
23:36:24ShapeShifter499linuxstb: an how is the rockbox loader coming?
23:36:28TheSeveni haven't found such a thing in the powermgmt apis
23:36:29 Join ifonefox [0] (
23:36:58linuxstbShapeShifter499: Doesn't work... I'm currently trying to debug it.
23:37:07linuxstb(but it doesn't get as far as being debuggable...)
23:37:14ShapeShifter499linuxstb: oh
23:37:28 Quit JackWinter (Read error: 60 (Operation timed out))
23:37:54TheSevenlinuxstb: what did you do to it? it has worked before, hasn't it? (at least as far as entering main)
23:39:01 Join JackWinter [0] (
23:39:03linuxstbI guess the relocation to IRAM and remapping IRAM to 0x0 is new.
23:39:20linuxstbBut I haven't tried running it for a long time.
23:39:27 Quit ifonefox (Client Quit)
23:39:59 Join ifonefox [0] (
23:40:16 Quit ifonefox (Client Quit)
23:40:56*ShapeShifter499 is away playing with his hacked wii
23:41:12 Join efyx [0] (
23:41:23TheSevenlinuxstb: wow the beast battery debug screen is nice :-)
23:41:41linuxstbI don't think I've seen it. Torne has my beast...
23:41:58TheSeveni just saw the source - awesome!
23:42:14TheSevenmonitoring things like regulator heat dissipation, different temperatures, etc. :-)
23:42:54linuxstbYes, jhMikeS (who did most of the Beast port) is thorough... Sadly he hasn't been around recently.
23:43:48 Join intrados4 [0] (
23:44:08 Join HellDragon [0] (
23:51:28ShapeShifter499what is the beast port?
23:51:37TheSevengigabeat s
23:52:00midgeypixelma: around?
23:52:54gevaertsTheSeven, linuxstb: when do you plan to fix the warnings? Before or after adding it to the build system?
23:53:09 Join stoffel [0] (
23:53:33linuxstbgevaerts: I think we should fix before.
23:54:10*TheSeven wonders why nano2g battery runtime is higher when playing music than when idle
23:54:34TheSevenit should be around ~15 hours right now
23:55:14linuxstbHow are you measuring runtime?
23:55:28TheSeveni'm just believing what the PMU tells me
23:55:41TheSevenwe're at 27mA when playing music (backlight off)
23:56:19AquaFoxidle time should be similar to a cell phone, which most models can run for a week or two in idle
23:57:19TheSevencore_sleep only makes 3mA difference
23:57:27TheSeveni suspect most of the current draw is outside the SOC
23:57:46HillshumBut DAPs get turned off when idle
23:57:55linuxstbUseful that it can be measured though. IIUC, other targets require linking up a meter to the device.
23:58:02AquaFoxmoms phone plays mp3s and stuff and it lasts about 12 days between charges with the occasional call lasting about a minute

Previous day | Next day