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).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2015-01-02

00:00:17gevaertsSo usb power == host as far as rockbox is concerned
00:00:42pamaurythen this check is just redundant ???
00:01:36gevaertsIt probably handles the case where you hold a button while booting
00:02:06gevaertsAll I know about early USB is that almost everyone has accidentally broken it at some point
00:03:18gevaertsWhat I really want to know is if any non-archos target needs it
00:03:30 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
00:03:44pamaurywhy do the archos need it ? because they start from ROM and have no other bootloader usb ?
00:03:45gevaertsamiconn: does anything apart from archos actually need early USB?
00:04:15gevaertsWell, they can start from ROM
00:07:13pamauryI think you are right: usb_powered() will return true only if the usb presses a button on those targets
00:08:33pamauryso usb_powered() is really usb_charging_only()
00:08:58gevaertsI think so...
00:08:59pamauryor usb_powered_only() because powering doesn't imply charging maybe ?
00:09:16gevaertsYes, maybe better
00:09:55pamauryok, let's go for renaming, all instances will get replaced by usb_inserted() except this one, i'll put it on gerrit and we should have a few people test it afterwards
00:10:36pamauryare there targets which can be powered from usb but not charged ? that would be weird
00:10:48 Quit lebellium (Quit: ChatZilla [Firefox 35.0/20141229214612])
00:11:17gevaertsThe ondios, possibly
00:12:45gevaertsThe Logik DAX almost certainly, but that's not at all near being supported
00:13:01gevaertsAs in, it runs on a single AA battery, and it can run off USB as well
00:13:35pamauryah yeah, you can have usb power but the battery cannot be charged by any mean
00:15:37amiconnSame for the Ondios (designed for 3x AAA batteries)
00:16:01 Quit [HappyNewYear] (Changing host)
00:16:02 Join [HappyNewYear] [0] (~franklin@unaffiliated/franklin)
00:21:35pamaurygevaerts: shouldn't usb_charging_maxcurrent_change() be part of usb.h ?
00:21:39pamauryinstead of usb_core.h
00:23:31pamauryhum, on the hand one can argue that it really requires usb_core to make it work, I don't know
00:24:21gevaertsI'd recommend leaving some minor things alone if you ever want to see this cleanup finished :)
00:24:41pamaurywell one could imagine a hardware usb device which reports the maximum current via a pin (I think it exists) so it should be part of usb.h
00:25:41JdGordon[Saint]: woulndt have a clue
00:25:43gevaertsI think technically we can do that
00:26:21gevaertsI mean, on some of the bridge chips, we can specify whether to ask for 100 or 500 mA. We could conceivably ask for 500 and if we don't get it fall back to 100
00:26:30gevaertsI wouldn'
00:26:34gevaertst recommend that though :)
00:26:51 Nick [HappyNewYear] is now known as [Franklin] (~franklin@unaffiliated/franklin)
00:27:17amiconnpamaury: I think both the H300 and M5/X5 do exactly that (i.e. report whether 500mA or only 100mA are allowed)
00:27:43pamauryis that used in any way at the moment ?
00:27:49amiconnWhether we test this I don't remember
00:27:54gevaertsamiconn: isn't that the reverse? As in, we tell the chip what to ask for?
00:28:17gevaertsOr is it a combination of both?
00:28:23amiconnHmm, not sure
00:28:46*amiconn didn't touch rockbox code for a long time (basically since the switch to git)
00:30:04gevaertsAh yes, we set it
00:30:13gevaertsSee firmware/target/coldfire/iaudio/usb-iaudio.c, e.g.
00:30:35amiconnOn M3 we don't - M3 USB is in a bad state :\
00:31:04amiconnOr rather, M3 charging is in a bad state, only part of which is usb power
00:34:53pamaurygevaerts: I have an peculiar question, is there any different somewhere between a target which has hardware USB and a target which does software USB and doesn't define HAVE_USBSTACK (because the driver is buggy say)
00:35:14pamaurymy guess is CONFIG_USBOTG but I'm not sure
00:36:20gevaertsIn the new order those would be USB_NONE or USB_DETECT, but currently I'm not sure
00:41:02gevaertsCONFIG_USBOTG is also set for some devices that (theoretically) do USB host, such as the h300
00:48:10***Saving seen data "./dancer.seen"
00:48:24 Join nialv7 [0] (~nialv7@
01:13:09fs-bluebotBuild Server message: 3New build round started. Revision 7265375, 255 builds, 28 clients.
01:13:19thomasjfoxthere goes another buflib handling bug
01:15:10[Franklin]all this fuss over buflib... why not just emulate an MMU in software?
01:15:57[Franklin]or just drop all the MMU-less targets
01:16:21 Quit y4n (Quit: We're fucking 3LN!)
01:18:56fs-bluebotBuild Server message: 3Build round completed after 348 seconds.
01:19:51thomasjfoxare the build hosts using ccache? Sometimes it's super fast, sometimes quite the opposite.
01:20:11gevaertsMost of them probably are
01:20:25[Franklin]thomasjfox: some drop out occasionally though
01:22:30thomasjfoxJdGordon: the last commit might interest you (shortcuts handling). It was quite complex to get right (and make it look simple)
01:27:16*thomasjfox thought about adding mandatory locking to buflib_get_data() that needs to be released explicitly (less API errors). Just scrolling in the main menu calls ~3700x buflib_get_data()... not an option.
01:39:50thomasjfoxgevaerts: I see you mentored the buflib GSoC project
01:40:09thomasjfoxit might be safer to switch off the movement of buflib buffers if no move callback is provided. It's better if the API user explicitly has to think about what needs to be done on a "move" event.
01:40:29thomasjfox(I'll fix up talk.c tomorrow...)
01:41:17gevaertsThat has the disadvantage of encouraging people to just block movement
01:41:48thomasjfoxcurrently have have rare, random crashing :(
01:42:00thomasjfoxwe have
01:55:06pamauryI don't see the point about the MMU. All a MMU can provide is help in debugging the issues (but that's far from trivial), and many targets don't have a MMU
01:56:19thomasjfoxMMUs that provide virtual addresses can do defragmentation in the background. But I think that's pointless for Rockbox since we must support non-MMU targets, too
01:57:41 Quit bertrik (Ping timeout: 255 seconds)
01:58:46 Join krabador [0] (~krabador_@unaffiliated/krabador)
01:59:18pamaurywell yes and no, you can somehow defragment without moving data, granted. But you are still limited by the virtual address space. And also when data becomes non-contiguous, it makes everything more complicated with DMA. We don't want that in rockbox :) But using it for debug purposes would be nice I admit
02:00:11 Quit ZincAlloy (Quit: Leaving.)
02:01:05thomasjfoxyes, virtual address space is the limit. If you have a 32bit mp3 player with 8 or 16 mb RAM, there should be plenty room for fragmentation :D
02:01:16thomasjfoxanyway, g8 and don't dream of MMUs
02:01:47 Quit thomasjfox (Quit: Konversation terminated!)
02:06:52[Franklin]foolsh: can you give me some help with unitconv?
02:28:53 Quit pamaury (Ping timeout: 244 seconds)
02:45:28 Quit AlexP (Remote host closed the connection)
02:47:15 Join advcomp2019__ [0] (~advcomp20@unaffiliated/advcomp2019)
02:48:13***Saving seen data "./dancer.seen"
02:50:15 Quit advcomp2019_ (Ping timeout: 240 seconds)
03:17:56 Quit RiD (Quit: A good plan today is better than a perfect plan tomorrow.)
03:20:20 Quit [Franklin] (Remote host closed the connection)
03:43:54 Quit JanC (Ping timeout: 264 seconds)
03:57:14 Join JanC [0] (~janc@lugwv/member/JanC)
04:48:16***Saving seen data "./dancer.seen"
04:59:59 Quit Bluefoxicy (Quit: ZNC -
05:01:16 Join Bluefoxicy [0] (
05:11:25 Join JdGordon_ [0] (~jonno@rockbox/developer/JdGordon)
05:12:44 Quit JdGordon (Ping timeout: 245 seconds)
05:15:35 Quit Provel (Read error: Connection reset by peer)
05:17:35 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
05:18:16 Quit JdGordon_ (Ping timeout: 240 seconds)
05:21:52 Quit TheSeven (Ping timeout: 265 seconds)
05:23:10 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
05:27:22 Quit advcomp2019__ (Ping timeout: 240 seconds)
05:30:47 Join Provel [0] (
05:31:02 Join Strife89 [0] (
05:56:57DEBUGLost contact with server (snapshot: dancer.c line 124)
05:56:57***Saving seen data "./dancer.seen"
05:56:59***Started Dancer V4.16
05:56:59***Connected to on port 6667
05:56:59***Logfile for #rockbox started
05:57:01Mode"logbot- :+i" by logbot-
05:57:06***Server message 501: 'logbot- :Unknown MODE flag'
05:57:06 Join logbot- [0] (
05:57:06 Join bp0 [0] (~bp@unaffiliated/bp0)
05:57:06 Join Strife89 [0] (
05:57:06 Join Provel [0] (
05:57:06 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
05:57:06 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
05:57:06 Join Bluefoxicy [0] (
05:57:06 Join JanC [0] (~janc@lugwv/member/JanC)
05:57:06 Join krabador [0] (~krabador_@unaffiliated/krabador)
05:57:06 Join nialv7 [0] (~nialv7@
05:57:06 Join fs-bluebot [0] (
05:57:06 Join bluebrother^ [0] (~dom@rockbox/developer/bluebrother)
05:57:06 Join Scall [0] (~chat@unaffiliated/scall)
05:57:06 Join skx_netb [0] (~sakax@unaffiliated/sakax)
05:57:06 Join Scromple_ [0] (~Simon@
05:57:06 Join soap_ [0] (~soap@rockbox/staff/soap)
05:57:06 Join AlwaysHi- [0] (~AlwaysHig@
05:57:06 Join scorche [0] (~scorche@rockbox/administrator/scorche)
05:57:06 Join pimaster [0] (~pimaster@
05:57:06 Join sobukus [0] (
05:57:06 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
05:57:06 Join Riviera [0] (Riviera@2a03:b0c0:1:d0::10:b001)
05:57:06 Join burgobianco [0] (
05:57:06 Join zoktar [0] (~zoktar@unaffiliated/zoktar)
05:57:06 Join pystar89 [0] (
05:57:06 Join shmibs [0] (~shmibs@
05:57:06 Join greatwolf [0] (greatwolf@gateway/shell/panicbnc/x-amntslmoethxtdeq)
05:57:06 Join Cinos [0] (
05:57:06 Join igitoor [0] (igitur@unaffiliated/contempt)
05:57:06 Join Rondom [0] (
05:57:06 Join ps-auxw [0] (~arneb@2001:470:c807:0:1532:4e5f:2ad3:4123)
05:57:06 Join KotH [0] (~attila@2001:1620:907::1)
05:57:06 Join GodEater [0] (~whoknows@rockbox/staff/GodEater)
05:57:06 Join akaWolf [0] (~akaWolf@unaffiliated/akawolf)
05:57:06 Join Epakai [0] (
05:57:06 Join dfkt [0] (dfkt@unaffiliated/dfkt)
05:57:06 Join byteframe [0] (~byteframe@unaffiliated/byteframe)
05:57:06 Join krnlyng [0] (~liar@
05:57:06 Join MMlosh [0] (~MMlosh@2001:470:6f:23:24f:63ff:fe01:4900)
05:57:06 Join WakiMiko [0] (~WakiMiko@unaffiliated/wakimiko)
05:57:06 Join mikroflops [0] (~yogurt@
05:57:06 Join the-kyle [0] (
05:57:06 Join dongs [0] (
05:57:06 Join Galois [0] (
05:57:06 Join zu [0] (
05:57:06 Join maraz [0] (
05:57:06 Join n17ikh [0] (~n17ikh@unaffiliated/n17ikh)
05:57:06 Join __jae__ [0] (
05:57:06 Join bzed [0] (
05:57:06 Join kugel [0] (~kugel@rockbox/developer/kugel)
05:57:06 Join derf [0] (
05:57:06 Join Guest66888 [0] (Slayer@
05:57:06 Join ParkerR [0] (ParkerR@unaffiliated/parkerr)
05:57:06 Join Unhelpful [0] (~quassel@rockbox/developer/Unhelpful)
05:57:06 Join shamus [0] (
05:57:06 Join markun [0] (~markun@rockbox/developer/markun)
05:57:06 Join ukleinek [0] (~ukl@unaffiliated/ukleinek)
05:57:06 Join rudi_s [0] (
05:57:06 Join @ChanServ [0] (ChanServ@services.)
05:57:06 Join pixelma [0] (~pixelma@rockbox/staff/pixelma)
05:57:06 Join amiconn [0] (~amiconn@rockbox/developer/amiconn)
05:57:06 Join preglow [0] (~thomj@2001:840:4243:3::100)
05:57:06 Join Mir [0] (
05:57:06 Join kvieta [0] (~kvieta@
05:57:06 Join TBCOOL [0] (
05:57:06 Join alexbobp [0] (
05:57:06 Join mykhal_ [0] (
05:57:06 Join rasher [0] (~rasher@rockbox/developer/rasher)
05:57:06 Join froggyman [0] (~frogs@unaffiliated/froggyman)
05:57:06 Join pookie_ [0] (
05:57:06 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
05:57:06 Join charlie [0] (~c@unaffiliated/charlie)
05:57:06 Join [Saint] [0] (~saint@rockbox/staff/saint)
05:57:06 Join x56 [0] (
05:57:06 Join Ketturi [0] (
05:57:06 Join APLU [0] (
05:57:07 Join user890104 [0] (Venci@unaffiliated/user890104)
05:57:07 Join renopt [0] (~renopt@unaffiliated/renopt)
05:57:07 Join gevaerts [0] (~fg@rockbox/developer/gevaerts)
05:57:07 Join endee_ [0] (sid15467@gateway/web/
05:57:07 Join babylonlurker [0] (
05:57:07 Join yosafbridge [0] (~yosafbrid@
05:57:07 Join mazzoo [0] (
05:57:07 Join enyc [0] (
05:57:07 Join yuriks [0] (~quassel@opentyrian/developer/yuriks)
05:57:07 Join williamtdr [0] (uid27909@gateway/web/
05:57:07 Join ender| [0] (krneki@2a01:260:4094:1:42:42:42:42)
05:57:07 Join aevin [0] (eivindsy@unaffiliated/aevin)
05:57:07 Join foolsh [0] (
05:57:07 Join Makinit [0] (
05:57:07 Join Naked_ [0] (
05:57:07 Join Marex [0] (~Marex@
05:57:07 Join Xyem_ [0] (
05:57:07 Join alucryd [0] (~quassel@archlinux/trusteduser/alucryd)
05:57:07 Join Elfish [0] (amba@2001:1608:12:1:13:3:3:7)
05:57:07 Join cronix [0] (
05:57:07 Join michaelni [0] (
05:57:07 Join TD-Linux [0] (~Thomas@about/essy/indecisive/TD-Linux)
05:57:07 Join funman [0] (~fun@rockbox/developer/funman)
05:57:07 Join simabeis [0] (
05:57:07 Join sLite [0] (
05:57:07 Join K1773R [0] (~K1773R@unaffiliated/k1773r)
05:57:07 Join Topy44 [0] (
05:57:07 Join Slasheri [0] (miipekk@rockbox/developer/Slasheri)
05:57:07 Join adnap [0] (
05:57:07 Join Jack87 [0] (Jack87@nasadmin/admin/jack87)
05:57:07 Join suYin`OFF [0] (
05:57:07 Join architekt [0] (
05:57:07 Join ruskie [0] (ruskie@sourcemage/mage/ruskie)
05:57:07 Join knittl [0] (~knittl@unaffiliated/knittl)
05:57:07 Join uwe_mobile [0] (
05:57:07 Join ranmachan [0] (
05:57:07 Join aevin2 [0] (
05:57:07 Join scorche|sh [0] (~scorche@rockbox/administrator/scorche)
05:57:07 Join GeekShadow [0] (~antoine@reactos/tester/GeekShadow)
05:57:07 Join uwe_ [0] (
06:02:30 Join Zambezi [0] (
06:09:31 Join advcomp2019 [0] (~advcomp20@2602:41:83a2:4200:889f:db7:68e8:f6aa)
06:09:33 Quit advcomp2019 (Changing host)
06:09:33 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
06:15:20 Quit advcomp2019 (Read error: No route to host)
06:16:27 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
06:16:42 Join rela [0] (~x@pdpc/supporter/active/rela)
06:21:12 Quit rela (Ping timeout: 244 seconds)
06:25:57 Quit krabador (Quit: Take the time.)
06:33:09 Join rela [0] (~x@pdpc/supporter/active/rela)
06:36:18 Quit Strife89 (Ping timeout: 265 seconds)
06:37:44 Quit rela (Ping timeout: 244 seconds)
06:49:11 Join rela [0] (
06:49:28 Quit rela (Changing host)
06:49:28 Join rela [0] (~x@pdpc/supporter/active/rela)
06:53:45 Quit rela (Ping timeout: 244 seconds)
06:55:47 Quit [Saint] (Read error: Connection reset by peer)
06:56:29 Join [Saint] [0] (~saint@rockbox/staff/saint)
06:56:36 Quit [Saint] (Client Quit)
06:57:10 Join [Saint] [0] (~saint@rockbox/staff/saint)
07:06:52 Join saratoga [0] (47e27552@gateway/web/freenode/ip.
07:08:59saratogalooking at AMSv2 frequency scaling, it seems the GPIO block is clocked directly off PCLK with no divider that I can see
07:09:09saratogaso when PCLK boosts to 96MHz, so does GPIO
07:09:19saratogaon the CLip+ and Zip, this is no problem
07:09:35saratogaon the Clipv2, this seems to cause most devices to lose the ability to read buttons via GPIO
07:09:59saratogaon mine, it acts as if the power button is held down, even through power is a simple GPIO switch (no matrix scan)
07:10:20saratogathat makes it seem like GPIO is working at all, but then i don't understand why it would work fine on other devices with the same SOC
07:12:01[Saint]'most devices'?
07:12:14[Saint]So, there's also some that /don't/ exhibit this behaviour?
07:13:01[Saint]ie. John's ClipV2 and Mary's ClipV2?
07:14:31saratogaskips seems to have similar problems as mine, but possibly worse
07:14:39saratogai don''t think we tested any others
07:15:27saratogai guess voltage could be different, but hte ADC reports all voltages the same between the plus and v2
07:15:42saratogaunless theres some voltage it can't measure
07:16:28saratogaby the way, does anyone have a fuzev2 handy? i put a new test build in the thread
07:16:40saratogacurious is lowering the DBOP clock makes a difference for that device
07:21:40 Join rela [0] (~x@pdpc/supporter/active/rela)
07:26:34 Quit rela (Ping timeout: 265 seconds)
07:27:22saratogalowering PCLK to 64MHz seems to fix the problem
07:34:22saratogai wonder if 96MHz is just too high one some devices
07:35:13saratogaa couple years passed between the clipv2 and the zip, maybe the hardware was revised or yields improved and clock speeds could go a little higher
07:57:00***Saving seen data "./dancer.seen"
08:12:43 Join Guest75680 [0] (Slayer@
08:15:04 Quit Guest66888 (Ping timeout: 244 seconds)
08:31:11 Join advcomp2019_ [0] (~advcomp20@unaffiliated/advcomp2019)
08:32:57 Join advcomp2019__ [0] (~advcomp20@unaffiliated/advcomp2019)
08:34:48 Quit advcomp2019 (Ping timeout: 245 seconds)
08:36:16 Quit advcomp2019_ (Ping timeout: 240 seconds)
08:44:16 Join advcomp2019_ [0] (
08:44:27 Quit advcomp2019_ (Changing host)
08:44:27 Join advcomp2019_ [0] (~advcomp20@unaffiliated/advcomp2019)
08:47:36 Quit advcomp2019__ (Ping timeout: 240 seconds)
09:57:01***Saving seen data "./dancer.seen"
10:00:43 Join ender` [0] (
10:10:19 Join lorenzo92 [0] (~chatzilla@
10:20:42lorenzo92any idea to "bridge" to tty devices in linux? i.e. without modifying anything else, I would like to bin /dev/ttymxc1 to, say, /tmp/ttyGS0...The thing I want to achieve is basically an export of ttymxc1 to ttyGS0
10:20:55lorenzo92without any processing inbetween
10:32:20lorenzo92uhmm not sure it will work, will it you think?
10:33:23marazonly one way to find out for sure! :h
10:34:34maraz/dev symlinks should be just like any other symlinks, and it works for other devices; I see no reason why it wouldn't work for ttys - but then again, I wouldn't exactly call myself a terminal wizard
10:40:22lorenzo92:D let me figure out, the main concern is about device ID
10:42:29 Join mshathlonxp [0] (mshathlonx@
10:44:52user890104lorenzo92: or you can note its minor/major values, and create a new device of the same kind using mknod
10:45:51lorenzo92the thing is, I can eventually modify the kernel module that provides uart over usb and hack the major value, but then I'm not sure it will correctly hook to /dev/ttymxc1
10:47:07user890104it this only about the name, or you need to change something else?
10:48:04lorenzo92yep, that was my first idea actually, unfortunately it's not about the name I fear, there is no module parameter for that...there is only the major / minor mechanisms
10:48:14lorenzo92let me double check
10:48:39 Quit saratoga (Ping timeout: 246 seconds)
10:48:40lorenzo92I was also thinking about doing some getty tricks ^^
10:49:30 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
10:56:02lorenzo92changing major directly in the kernel module does not work of course...
11:00:32 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
11:01:09 Quit bp0 (Quit: Leaving)
11:49:16 Quit pystar89 (Ping timeout: 240 seconds)
11:57:04***Saving seen data "./dancer.seen"
12:01:22 Join AlexP [0] (~alex@rockbox/staff/AlexP)
12:03:32 Join thomasjfox [0] (~thomasjfo@rockbox/developer/thomasjfox)
12:05:08 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
12:05:42 Quit y4n (Read error: Connection reset by peer)
12:06:32 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
12:07:01 Quit Provel (Ping timeout: 250 seconds)
12:12:26 Quit amiconn (Quit: No Ping reply in 64 seconds.)
12:13:20 Quit pixelma (Quit: No Ping reply in 120 seconds.)
12:14:36 Join amiconn [0] (~amiconn@rockbox/developer/amiconn)
12:15:29 Join pixelma [0] (~pixelma@rockbox/staff/pixelma)
12:31:24 Join JdGordon_ [0] (~jonno@rockbox/developer/JdGordon)
12:31:52lorenzo92socat seems to be useful, it's now compiled and running but still some problems to investigate later
12:32:21 Quit JdGordon (Ping timeout: 245 seconds)
12:59:41pamaurygevaerts: am I crazy or is USBPOWER_BUTTON completely unused ?
13:00:00gevaertsHmmm, it might be
13:00:18gevaertsWe moved from requiring a specific button to accepting any button a while ago
13:01:27pamauryyeah, apparently this one was forgotten
13:14:18thomasjfoxgevaerts: regarding buflib callbacks: I think it's kind of irritating that a "struct callback" value of "NULL" supplied to buflib_alloc_ex() enables movement of buffers, but supplying NULL as the actually value in the callback structure doesn't
13:14:45thomasjfoxalso for shrinking it behaves different: Shrinking is enabled only when there is a valid callback function (which makes sense, but still)
13:31:29pamaurygevaerts: shouldn't usb_release_exclusive_storage() be private ? I'm not sure it should in usb.h
13:31:41 Join lebellium [0] (
13:55:28*lorenzo92 starts again trying to make ypr1 a pc bluetooth dongle
13:57:05***Saving seen data "./dancer.seen"
13:57:56 Join pystar89 [0] (
14:03:03lorenzo92yay :D the bluetooth UART is now exported to the computer hehe
14:11:43lorenzo92SUCCESS! :D
14:12:04lorenzo92I can now use ypr1 as bluetooth dongle, isn't that cool? haha
14:41:01[Saint]If there's one thing Rockbox needs...its to support more dead technology. BT checks out. ;p
14:47:48[Saint]Nonetheless, good work.
14:48:23[Saint]Extremely limited use case, but, its not like that's considered a major hurdle.
14:52:48lorenzo92indeed, but I consider this a step in order to make a proper stack ;)
14:55:05[Saint]oh, sure.
14:55:56lorenzo92actually, RFCOMM would be handy to debug USB development ^^
14:57:44[Saint]"My ancient largely dead protocols bring all the boys to the yard"
15:00:36*[Saint] is likely hating on BT more than it deserves.
15:00:48[Saint]But BTLE makes it look virtually prehistoric.
15:01:04alexbobphehe I was just wishing earlier today the fuze had bt
15:01:48[Saint]alexbobp: iirc saratoga was looking for Fuze owners earlier.
15:02:06alexbobpwell I am super jazzed to be a fuze owner again. those things are so hard to find now
15:02:08[Saint][19:16:29] <saratoga> by the way, does anyone have a fuzev2 handy? i put a new test build in the thread
15:02:14alexbobpoh this is fuzev1
15:02:18[Saint]there we go, I was right for a change.
15:02:23[Saint]aha - no problem.
15:02:46alexbobpbut I have a view sitting around in case anyone ever wants a build for that tested~
15:02:49*alexbobp keeps fingers crossed
15:02:52[Saint]and, yes, yes they are.
15:03:15[Saint]I kick myself for not snapping up SanDisk's refurbs a while back.
15:03:19alexbobpsandisk replaced the best mp3 player ever made with a nearly unusable one
15:03:21alexbobpso sad
15:03:46alexbobpI feel the same way
15:03:49alexbobpI would have bought up a dozen
15:03:49[Saint]One posits that was a cost move. But, I'm not confident as to the exact motivation.
15:04:06[Saint]I /guess/ they had a fucktonne of cheap silicon to get rid of.
15:04:17[Saint]Else the move makes little to no sense.
15:04:23alexbobpbut they priced the fuze+ higher in retail right?
15:04:30alexbobpis a clicky wheel really so expensive?
15:04:44[Saint]compared to cheap dome caps, yes.
15:05:45[Saint]though not significantly so if you're a large manufacturer I imagine. I dunno. BOM is hard to judge.
15:05:56alexbobpif we could design a usb button pad that connects onto the fuze+ and works with rockbox for $5 or so that'd be amazing
15:06:09alexbobpsince it's easy to get cheap used fuze+s since people buy them and then realize they are unusable
15:06:15alexbobpbut that's probably a lofty goal
15:06:59[Saint]I have no idea if such a thing is possible.
15:07:17[Saint]To do it in a cost effective and portable format certainly would not be trivial.
15:07:26[Saint]see: constant death of the Lyre project.
15:08:26alexbobpin the days of kickstarter it might actually have legs
15:08:37[Saint]open hardware almost always ends up with a more expensive, less portable, vastly inferior manufactured product.
15:08:41alexbobppono was kickstartered!
15:08:51alexbobphow come?
15:09:19[Saint]its incredibly difficult to manufacture in small batches in high precision at low cost.
15:09:33alexbobpoh I'm not talking about small batches
15:09:49[Saint]well then there's cost to consider.
15:09:57alexbobphence kickstarter
15:10:04[Saint]you need to talk about setting up a manufacturing line. several hundred thousand dollars.
15:10:08alexbobpI wonder if the market for people who want a decent mp3 player firmware without fighting is as big as the market for audiophiles who don't believe in nyquist sampling theory
15:12:02alexbobpwell I can dream
15:12:20[Saint]regarding the market - apparently not.
15:12:33[Saint]hence the lack of such things existing.
15:13:11[Saint]there's basically only two markets for DAPs at all these days.
15:13:26[Saint]high end audiophile gear, and POS disposable Gym Bunny gear.
15:13:48[Saint]smartphones fill the void in the regions leftover.
15:13:52alexbobpwell I like the second kind, but I want rockbox on them and I want physical buttons
15:14:18alexbobpwhat exactly do you mean with high end audiophile gear? are the sound chips on smartphones not as good?
15:14:38alexbobpor is it more about marketing and percieved quality
15:15:52[Saint]depends on ones perception of good I suppose. Good enough for average joe, certainly. Not good enough for the type that'll buy a Pono.
15:16:11alexbobpwell the pono is beyond reality
15:16:17alexbobpthat's for people who deny science and placebo themselves
15:16:44alexbobpI mean good sound chip sure, but the extra sample resolution is pointless
15:17:07alexbobpor do you not agree about that?
15:17:50[Saint]Oh. I certainly do. In many cases, 15 bits is more than enough for wholly accurate reproduction.
15:18:20alexbobpyeah... certainly as accurate as it can be after going through the physical layers anyways
15:18:48[Saint]But my thoughts don't change the fact that these people exist and often have a LOT of expendable funds.
15:19:00alexbobpwell... so percieved quality
15:19:05[Saint]A good marketer doesn't care if his product is bullshit if it sells.
15:19:18alexbobpI wouldn't make a good marketer :(
15:20:23[Saint]I myself have some pretty fancy audio gear. But I don't buy into this "HD" audio baloney.
15:20:40[Saint]All my media is flac or mp3 320.
15:20:46alexbobpI have some nice audio gear but it's pretty low end compared to what hardcore people like
15:21:00alexbobpI go for vorbis 160 when I have the choice
15:21:17alexbobpbecause I'm convinced it's transparent (and thus better than mp3 at basically any realistic bitrate(
15:22:45alexbobpbut yeah I have nice headphones and nice speakers at home but I haven't noticed any issues with the playback quality of my sansas compared to, eg, my various computers... either I'm not audiophile enough to tell or my headphones aren't high end enough
15:24:48alexbobpwell goodnight, I'm gonna go dream of nice open hardware with rockbox, realistic or no :P
15:40:26 Join Jinx [0] (Dojo@unaffiliated/jinx)
15:57:09***No seen item changed, no save performed.
16:08:42 Join krabador [0] (~krabador_@unaffiliated/krabador)
16:39:19 Join Misanthropos [0] (
16:39:25 Join bertrik_ [0] (
16:39:27 Quit bertrik_ (Changing host)
16:39:27 Join bertrik_ [0] (~bertrik@rockbox/developer/bertrik)
16:40:06 Join TheLemonMan [0] (~lemonboy@unaffiliated/thelemonman)
16:40:31 Quit bertrik_ (Client Quit)
16:41:42 Quit krabador (Quit: Take the time.)
16:46:13lebelliumTheSeven: ping
17:02:46 Join [Franklin] [0] (~franklin@unaffiliated/franklin)
17:16:29*[Franklin] discovers GNU units
17:16:51 Quit Misanthropos (Ping timeout: 240 seconds)
17:25:19[Franklin][Saint]: thanks for the super-duper-constructive comment!
17:25:55[Franklin]anyway, I feel like porting something
17:27:12[Franklin] has some interesting stuff
17:28:12[Franklin]REminiscence looks promising
17:28:21[Franklin]but it's GPLv3
17:30:53[Franklin]maybe supertux
17:35:18 Join lunarl0n_ [0] (
17:39:49 Quit [Franklin] (Ping timeout: 250 seconds)
17:43:49lunarl0n_I'm trying to build a simulator of the DX50 port and when running make I get a lot of "CONFIG_PLATFORM" redefined errors and make fails
17:45:22 Join [Franklin] [0] (
17:47:38thomasjfoxlunarl0n_: IIRC when I tried yesterday, it even complained that no Android SDK is installed. So I guess the simulator stuff is not working for Android based targets right now (just a guess though)
17:49:27 Quit [Franklin] (Changing host)
17:49:28 Join [Franklin] [0] (~franklin@unaffiliated/franklin)
17:53:07lorenzo92also, complains about cygwin in the mails...
17:54:39thomasjfoxthough that guy never replied back...
17:55:33[Saint]fuck CygWin.
17:55:50[Saint]No one's supported building on CygWin for years.
17:56:04[Saint]If anyone wants to try it, great, but they're on their own.
17:57:13***Saving seen data "./dancer.seen"
18:04:37[Franklin]lorenzo92: can you help me debug xworld now?
18:04:51 Join dfkt_ [0] (dfkt@unaffiliated/dfkt)
18:05:00 Quit dfkt (Disconnected by services)
18:05:11lorenzo92unfortunately I only have the R1 here with me right now :(
18:05:15 Nick dfkt_ is now known as dfkt (dfkt@unaffiliated/dfkt)
18:05:19[Franklin]does it work on the R1?
18:05:38[Franklin]does it build plugins?
18:05:39lorenzo92later this evening sure
18:05:45lorenzo92i'm back home
18:06:01 Quit bertrik (Remote host closed the connection)
18:06:14[Franklin]ok then
18:06:15lorenzo92first of all at least we have the clue, right?
18:06:22lorenzo92it was about a ... timer?
18:06:52[Franklin]I doubt that a missing timer would cause a crash like that
18:08:02[Franklin]probably memory then
18:08:30[Franklin]YP-R1 seems to build plugins, too
18:09:17thomasjfoxRaaA doesn't really support plugins yet. It's WIP.
18:09:29thomasjfoxand certainly not VIP ;)
18:09:29[Franklin]the weird thing is that YP-R0 has a bigger pluginbuf than codecbuf
18:10:21[Franklin]maybe that's a typo?
18:16:54lorenzo92thomasjfox: RaaA on YPR0 does have all the plugins
18:17:00lorenzo92most of them fully working ;)
18:17:15[Franklin]lorenzo92: how about R1?
18:17:28lorenzo92it doesn't because of the touchscreen
18:17:40lorenzo92and I also did not further developed this way :)
18:17:47thomasjfoxprobably because kugel owns it, too ;)
18:17:48[Franklin]laziness? :P
18:17:57lorenzo92whatever ^^
18:18:01lorenzo92other ports ;)
18:18:07lorenzo92let's say this
18:34:51 Quit MMlosh (Ping timeout: 244 seconds)
18:41:00 Join MMlosh [0] (~MMlosh@2001:470:6f:23:24f:63ff:fe01:4900)
18:41:28 Quit skx_netb (Ping timeout: 245 seconds)
18:43:48lebelliumlorenzo92: I can't build the R0sim anymore. Would you mind trying in your environment?
18:46:03lorenzo92lebellium: sure, let me try
18:55:31 Quit nialv7 (Ping timeout: 244 seconds)
18:57:24fs-bluebotBuild Server message: 3New build round started. Revision 66df5f3, 255 builds, 27 clients.
19:00:32pamaurygevaerts: what do you think about the call to usb_powered() in battery bench ?
19:02:54pamauryit seems a bit useless to me
19:05:32 Quit lorenzo92 (Ping timeout: 244 seconds)
19:13:40pamaurythomasjfox: are you scrutinizing all uses of core_alloc ?
19:13:55thomasjfoxyep :)
19:14:15thomasjfoxreviewed talk.c today, that was quite complex. No bug found :)
19:16:14pamauryimpressive, and very useful :) I'm beginning to think that core_alloc() is too dangerous to exists ^^
19:17:30thomasjfoxthe benefits buflib brings to Rockbox are great though. No more "Please reboot to enable xyz" ;)
19:17:55pamauryyeah I agree, it's just core_alloc() vs core_alloc_ex()
19:19:07thomasjfoxthere are only seven callers of core_alloc() left ;)
19:19:54pamauryyeah, I did the grep myself ;)
19:20:31 Join lorenzo92 [0] (
19:21:30thomasjfox"shortcuts" handling was busted in a very subtile way: It's a kind of single linked list. The handle of the next list entry is stored in the current list entry. When the second core_alloc() call was made, it was possible that it moved the first buffer that should store the result. Outch!
19:22:14pamauryyeah I saw that, subtle indeed
19:24:07 Join bertrik [0] (
19:24:27 Quit bertrik (Changing host)
19:24:27 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
19:24:32thomasjfoxkugel's out-of-tree buflib development tree contains a good bunch of tests for buflib
19:24:47thomasjfoxI want to check the next days if I can make them part of the tree in firmware/test
19:25:57[Saint]aha, that's right, I knew there were tests _somewhere_.
19:26:10[Saint]I thought it was in the rb tree proper, though.
19:27:47thomasjfoxsoon it will :D
19:28:49fs-bluebotBuild Server message: 3New build round started. Revision bce72e6, 255 builds, 27 clients.
19:55:12lorenzo92good thing indeed
19:57:16***Saving seen data "./dancer.seen"
20:06:58TheSevenlebellium: pong
20:08:42 Join nialv7 [0] (
20:23:26 Join iggy_ [0] (6cb4dc55@gateway/web/freenode/ip.
20:23:39 Nick iggy_ is now known as savanrola (6cb4dc55@gateway/web/freenode/ip.
20:23:48 Nick savanrola is now known as savanarola (6cb4dc55@gateway/web/freenode/ip.
20:25:36savanarolaIs there a secure way to download rockbox? The website doesn't have TLS and nothing is gpg signed. I really want to try it out, but I need to be able to authenticate it before I run it.
20:26:31[Saint]https git checkout and build from source.
20:26:51[Saint]though, that's /somewhat/ in the tinfoil hat territory.
20:28:01savanarola[Saint]: What is?
20:28:12savanarolaBuilding from source?
20:28:37[Saint]What's making you do so, moreover.
20:29:44savanarolaI like security. Packet injection is real. I don't want to get owned. Running code that you get over an insecure connection is a bad idea.
20:30:08savanarolaBut thank you, [Saint]. I'll try to build it from source.
20:30:12gevaertsIf the builds were signed, how would you check our good intentions?
20:30:18[Saint]I would suggest hopping off the Internet immediately, then. ;)
20:30:38[Saint]There be spooky ghosts there.
20:31:38savanarolagevaerts: I would still have to worry about your good intentions, but at least then I wouldn't have to worry about anyone upstream injecting into my stream. With crypto, I only have to trust the people that write the software.
20:31:48[Saint]On the non-tinfoil-hat-wearing side...we /could/ at least give an md5 sum.
20:32:10gevaertsWe probably should do that, yes
20:32:16[Saint]That's as many fucks as I'd give to this, though.
20:33:19*foolsh a little is surprised franklin hasn't suggested a badusb plugin yet
20:33:26savanarola[Saint]: That would be nice (maybe SHA256, though?) but it's not worth anything unless you can get the sum securely. It doesn't solve the problem if someone else can forge the sum.
20:33:32[Franklin]foolsh: thanks for the idea! :P
20:34:05[Saint]savanarola: realistically, its not the binaries you should be scared of.
20:34:09[Saint]Its our source.
20:34:28[Saint]That's far more likely to contain easy to overlook nasties in my opinion.
20:34:34[Franklin]yes, I could easily slip some badusb code into xworld :P
20:34:39[Saint]And I think neither is terribly likely.
20:35:04[Saint]If you're going to want secure download, I would expect nothing less than a full audit as well.
20:35:27[Saint]What's the point in securely downloading software you can't vouch for the intentions of?
20:35:48[Saint]We may want to steal all your shit and be obfuscation experts for all you know.
20:35:54gevaerts[Franklin]: forget it
20:35:56savanarolaBut you at least can be held accountable. If someone sees something bad in your code, people know not to use it. Without good crypto, though, we never know whether it was you or someone on the network.
20:36:32[Franklin]savanarola: as long as you can trust gevaerts, rockbox is secure ;_)
20:37:02[Saint]Assuming we've full control over the resources, a wolf crier could be turned into a conspiracy theorist fairly easily.
20:37:02[Franklin]I do think that at least HTTPS should be enabled for the site, though
20:37:07[Saint]but...I digress.
20:37:11savanarolaI'm not trying to tell you guys how to run your project, though. I appreciate what you guys are doing. There are a lot of people like me, though, that want some assurance for the code we run on our machines.
20:37:46[Saint]Its a pity such assurance is often a fallacy.
20:37:50[Saint]But, I see the point.
20:38:00savanarolaOk. Thanks.
20:38:55[Franklin]seriously, why isn't there SSL for the site?
20:39:06foolshany MIM could still change anything they wished with the data between you and the source, https md5 or what ever, if they were persistant enough
20:39:18[Saint]considering a certain swedes involvement, it /does/ seem odd...yes.
20:40:15[Saint]the "why?" is probably something along the lines of " 'cos the shit's hard, yo"
20:40:35[Franklin]it's easy
20:40:48[Saint]Not to get it right it isn't.
20:40:48[Franklin]just go on ArchWiki and look it up :P
20:41:06savanarolaletsencrypt will sort you guys out :)
20:41:44[Saint]savanarola: of interest may be the fact that a prior core developer is nor head of Mozilla and runs the cURL project.
20:41:56lebelliumTheSeven: do you have an idea why the UI sim builds have not been updated for months here"> ?
20:42:08gevaerts[Saint]: huh?
20:42:15gevaertsworks for != head of
20:42:22gevaertsSubtle differences there
20:42:41[Franklin]plus, GPG signing a release isn't that hard
20:43:05savanarola[Franklin]: Exactly. Or you guys could through something into the Debian repo.
20:43:37[Saint]There was some third party debian repo a while back.
20:43:45[Saint]pretty sure it died from obscurity.
20:43:46savanarolaGPG has a trust bootstrapping problem, but it would be very useful.
20:44:17savanarolaTor hidden service maybe? Its not that hard to set one up.
20:45:05[Saint]While all these are true...its not that hard to take some simple precautions to limit exposure to damage and just assume we *aren't* legit as well.
20:45:07gevaertsThe *major* issue rockbox has these days is volunteer time, and volunteer time with trusted access to the servers is especially scarce
20:45:12[Saint]Even if we had such systems in place.
20:46:46[Saint]load up an a DAP with zero user data, on a sandboxed machine...sniff our intentions for a while. etc.
20:46:50[Franklin]gpg: unusable secret key!?
20:47:04savanarola[Franklin]: haha
20:47:44[Franklin]Well, I'll make a new one :P
20:48:09savanarola[Saint]: I get that. Anything you could do would go a long way, though. Sending code over the internet in plaintext is asking to be made a vector in an attack.
20:49:05[Saint]savanarola: oh, I know. It just seems that someone who goes to such efforts probably wouldn't trust us implicitly whatever the setup was, nor should they.
20:49:22[Saint]And those who don't care...don't care.
20:49:38[Saint]I can see the point, but its definitely a niche.
20:50:00savanarolaYeah. But something > nothing. Security gets weird when you get really tin-foil-hat, but that's not a reason not to do it :)
20:50:10foolsh"use the source luke"
20:50:28savanarolaAlso, probably not so much of a niche after Snowden. Linux users tend to care a lot about this stuff right now, I find.
20:50:52[Saint]99% of us fall into said bracket.
20:51:10[Franklin]I doubt that there's a single dev that uses windows
20:51:19[Saint]There used to be.
20:51:22[Saint]Perhaps not now.
20:52:08savanarolaWell security and FOSS are complementary. You can't have the first without the second.
20:53:19[Franklin]seriously, what website doesn't have HTTP SSL in this era?
20:53:43savanarola[Franklin]: lol
20:55:57savanarolaBut you're right. It should be a standard. For more on why every website needs it, check out
20:56:55[Franklin]then from there I guess putting the SHA* sum of the bins would be safe then, no?
20:57:33lorenzo92[Franklin]: I'm inserting some info on the wiki and later I will be more than happy to help
20:57:44[Franklin]ok, thanks
20:58:30savanarola[Franklin]: Yeah, assuming you trust the CA's ( I don't, but most people do). That would make you not the easiest target. People want easy targets. Why inject into an SSL stream when you can do it to a plaintext stream. They'll go after someone ese's website.
20:59:07 Join RiD [0] (
20:59:13[Franklin]savanarola: if you don't trust the CA's then who will you trust? O.o
20:59:36*gevaerts doesn't think this discussion has much to do with rockbox any more
21:00:02[Franklin]so in short, SSL-enable
21:00:30*[Saint] sets /topic to 'all security aspects are now handled in #rockbox-clusterfuck'
21:00:37savanarolaBatman. And decntralized systems like i2p and Tor hidden services. I establish trust with them through over other means. But yeah, that's it in short.
21:02:04savanarolaAnyway. Thanks for your help. Hope you guys implement something soon. If you want painless SSL, the letsencrypt project I linked to earlier will soon be making this free and super easy.
21:02:54 Quit savanarola ()
21:05:29lorenzo92for who doesn't heavily hate ( :P ) Bluetooth,
21:05:46lorenzo92I'm documenting some facts
21:06:50[Saint]'bluetooth is shit. that is all.'
21:07:19lorenzo92hehe yeah perhaps I have to add the "A Fact" heading
21:07:25ukleinek[Saint]: it seems you can contribute that to said page.
21:16:51 Quit Mir (Remote host closed the connection)
21:19:42 Join Mir [0] (
21:32:57lebelliumlorenzo92: you wrote R0 instead of R1, no?
21:33:29lorenzo92lebellium: wops :D my favourite strikes again
21:33:56lorenzo92I'm quite ready to save another revision, with information about the "bt dongle mode testing"
21:38:56 Join krabador [0] (~krabador_@unaffiliated/krabador)
21:54:00pamaurygevaerts: did you see my question about usb_powered in battery bench ?
21:54:39gevaertspamaury: I'd leave that in but have it use usb_inserted
21:54:50pamauryyeah I thought so
21:56:32*[Franklin] takes a stab at optimizing rocklife
21:57:17***Saving seen data "./dancer.seen"
21:57:28TheSevenlebellium: I'll have a look... I thought I had fixed these months ago
21:58:54pamaurygevaerts: should it still be enclosed in HAVE_USB_POWER ? I think so because usb_inserted() doesn't mean you have usb power
21:59:34gevaertsYes, probablu
22:00:05TheSeven> C: printf "archosfmrecorder\n2\na\nw\ns\n\n" | ../tools/configure
22:00:05TheSeven> Run 'make'
22:00:05TheSeven> Run 'make install'
22:00:05DBUGEnqueued KICK TheSeven
22:00:05TheSeven> No rockboxui, clean up and return
22:00:05TheSeven...and similar output for every target
22:01:37lebelliumdoes that mean that one cannot build any UI sim since August and nobody noticed that? :D
22:04:57pamaurygevaerts: what about the skin token ?
22:05:01pamauryJdGordon_: ping
22:05:09gevaertspamaury: same, I'd say
22:05:22 Join ZincAlloy [0] (
22:06:17pamaurygevaerts: why not remove it ? it is deprecated apparently ?
22:06:33TheSevenlebellium: at least nobody informed me about that, not sure if anyone complained in the channel
22:07:02TheSevenoh, while I'm at it:
22:07:03TheSevenThe signer certificate will expire within six months.
22:07:03***Alert Mode level 1
22:07:03TheSevenNo -tsa or -tsacert is provided and this jar is not timestamped. Without a timestamp, users may not be able to validate
22:07:03***Alert Mode level 2
22:07:03TheSeventhis jar after the signer certificate's expiration date (2015-03-10) or after any future revocation date.
22:07:13gevaertspamaury: maybe it is, but I couldn't find any trace of the replacement [Saint] claimed there is
22:07:24pamauryyeah neither me, just looked
22:07:33pamauryI'll rename it USB_INSERTED
22:10:37pamaurygevaerts: what about CURRENT_USB ?
22:11:51gevaertspamaury: no idea. That's entirely out of what I know about...
22:12:17pamauryto me it currently doesn't make any sense, at least for targets which set it to CURRENT_USB
22:12:21pamaury*to 500mA
22:16:20lorenzo92afaik, current is appended to usb descriptors, and of course 500mA is the maximum for USB2.0
22:17:04***Alert Mode OFF
22:17:12lorenzo92official maximum of course, I have also drained more. Read as: Raspberry Pi, other gadgets, accidental shorts ^^
22:17:51pamauryno, CURRENT_USB is only used in power management
22:18:24lorenzo92hum really? I bet I've seen it in usb-core...Most likely I'm wrong
22:19:06pamauryno it's USB_MAX_CURRENT
22:19:11lorenzo92right ;)
22:19:23lorenzo92so yeah, I don't see any utility
22:20:00lorenzo92also because it is a theorical value, i.e. it drops when charge comes to the end...
22:20:04lebelliumwhile I'm looking at"> , does someone know why IRC stats stopped in May 2014?
22:20:07pamauryCURRENT_USB is an estimation of the current draw when usb is plugged, on devices when usb might no power the device
22:21:23pamaurygevaerts: who has a Player or Fm or (Fm)Recorder (V2) ?
22:22:07lorenzo92pamaury: okay, I've seen the code snippet
22:22:14gevaertsGood question :)
22:22:17pamauryamiconn: do you have any of those devices ?
22:22:36TheSevenlebellium: /home/rockbox/dailybuilds/rockbox/firmware/target/hosted/filesystem-win32.c:236:5: error: expected ‘;’ before ‘return’
22:22:36TheSeven return -1;
22:23:35TheSevenmissing ; in line 232 of that file
22:24:06amiconnI only have a Player. LinusN is the only person I know who has (or had some time ago) an FM recorder in working condition
22:24:35pamauryamiconn: the player will do. Could you tell me if the battery remaining time estimation makes sense ?
22:24:55lorenzo92pamaury: but still CURRENT_USB itself is okay, what is not is the conditional < CURRENT_NORMAL because charging is not included. But hey I don't have any idea of those targets, just my 2 cents
22:25:45pamaurywhat is not normal is CURRENT_USB being set to 500mA I think
22:25:46amiconnMakes sense in what way?
22:26:13lebelliumTheSeven Is that easy to fix? I'm not a dev so this doesn't really make any sense to me
22:26:57pamauryamiconn: the code suggest it should be complete garbage
22:26:59amiconnAfair the battery remaining time has been calibrated properly, taking an average discharge curve of many different NiMH cells
22:27:20TheSevenlebellium: well there's at least one issue that's trivial to fix (just an oversight, a missing ; somewhere in the code)
22:27:28TheSevenand it seems to work better after that at least
22:27:44TheSevenbut let's see if it goes all the way through now or just gets stuck somewhere else...
22:28:03pamauryamiconn: the discharge curve is one thing, here I'm interested in the time printed in the System > Info menu (iirc)
22:28:28amiconnIt sure was correct last time I updated rockbox
22:29:14amiconnBtw, FM/V2 is rather different from Player and Recorder V1 regarding battery. Player and Recorder are basically identical (NiMH), while FM/V2 runs off LiIon
22:30:03pamauryok my mistake, I think the battery estimation is wrong in USB mode
22:30:21pamauryyou'll tell me that you can't see the estimated running time in USB mode...hum
22:31:05pamauryI really don't understand this code in powermgmt.c, it makes no sense
22:32:36amiconnEstimated running time in USB mode is rather short, as the archos devices cannot charge from USB, and a continuously spinng 2.5" HDD draws quite a lot of current
22:32:41 Quit byteframe (Read error: Connection reset by peer)
22:33:40amiconnEven with the charger connected the batteries will *discharge* when the HDD is continuously accessed (with HDD spinning but idle it will charge very slowly)
22:34:09pamauryok, so the HDD is really power hungry
22:35:16pamaurygevaerts: last question :) do you have any clue why usb_charging_enable() is defined in power-h300.c ?
22:35:18lorenzo92pamaury: that's why I don't see CURRENT_USB to be useful. A device can be either USB powered or not, otherwise either you are out of USB specs (500mA)
22:35:27lorenzo92or not charging...xD
22:35:48pamaurylorenzo92: some device cannot be powered from USB
22:36:01lorenzo92* yes sorry I mean powered
22:36:27TheSevenlooks like some sims build now, but not all of them
22:36:46pamaurywhen you cannot be powered from USB, in USB the power consumption is quite different because the transceiver is active and the disk is spinning all the time
22:37:09gevaertspamaury: these are *vague* memories, but IIRC the h300 is the only target where we actually only start USB charging when we get the OK from the host
22:37:10lebelliumTheSeven: does the Fuze+ one build?
22:38:07pamaurygevaerts: I'm not sure I understand, isn't usb_core supposed to properly handle usb charging when the target supports it ?
22:38:36 Quit Ketturi (Ping timeout: 245 seconds)
22:38:40pamaurymy table (pamaury/b3db692c1880ef97a0f2/raw/f4a2740a351bf8304b6abcb9d5666c4978b1d9e4/usb_defines.txt"> says that quite a number of targets have HAVE_USB_CHARGING_ENABLE
22:38:42gevaertspamaury: usb_core is suppose to *decide*, but ideally something has to tell the hardware what to do
22:38:51lorenzo92pamaury: got it! Thanks for clarification ;)
22:38:54pamaurybut that's not what this function does
22:39:10gevaertsI think Torne touched that code last
22:39:32pamauryusb_charging_enable() takes the decision and calls usb_charging_maxcurrent_change() which is target specific
22:39:42gevaertsAh, ok
22:40:06gevaertsMaybe it's the only one that did something before Torne updated the mechanism
22:40:58gevaertsOh, wait
22:41:01pamauryI'm ready Torne's commit
22:41:04gevaertsusb_core isn't involved here
22:41:21gevaertsSo I don't know at all
22:41:41pamaurynice catch
22:41:56 Join byteframe [0] (~byteframe@unaffiliated/byteframe)
22:42:04pamauryI think that's it
22:45:25pamaurygevaerts, lorenzo92, everyone ;) : g#1094 g#1095 g#1096 g#1097
22:45:40lorenzo92patch day: the good one! xD
22:45:41fs-bluebot3Gerrit review #1094 at : 3usb: move usb charging function prototype from usb_core.h to usb.h by Amaury Pouly
22:45:41fs-bluebot3Gerrit review #1095 at : 3usb: document usb_status_event and #ifdef it with USB_STATUS_BY_EVENT by Amaury Pouly
22:45:42fs-bluebot3Gerrit review #1096 at : 3usb: document usb states, powering and detection by Amaury Pouly
22:45:43DBUGEnqueued KICK fs-bluebot
22:45:43fs-bluebot3Gerrit review #1097 at : 3Clarify usb_powered() and fix some code. by Amaury Pouly
22:46:01lorenzo92sounds great
22:47:09thomasjfoxdoes gerrit have an easy way to show the whole diff in one window?
22:47:18thomasjfoxfor me it opens a window for each file to diff...
22:47:46pamaurythomasjfox: no, that's a long wanted features, there is a patch pending on the gerrit's bug tracker but it's rotting
22:48:04thomasjfoxclicking on gitweb and then viewing the patch seems to work
22:48:17pamauryg#1098 also
22:48:22fs-bluebot3Gerrit review #1098 at : 3usb: make usb_release_exclusive_storage private by Amaury Pouly
22:48:41 Quit ZincAlloy (Quit: Leaving.)
22:49:24 Join Provel [0] (
22:51:35thomasjfoxit's funny that the USB disconnected mode is called "extracted" instead of just "disconnected"
22:51:46 Join Ketturi [0] (
22:51:54pamaurythomasjfox: it's inserted vs extracted
22:52:37pamauryI think it makes sense because USB cable can be inserted without the device being connected to the host
22:54:30 Quit lorenzo92 (Remote host closed the connection)
22:54:52thomasjfoxwould we detect such a state? Probably not I guess
22:55:10 Quit ender` (Quit: There's only one way to avoid criticism: do nothing, say nothing, be nothing.-- Aristotle)
22:55:18pamauryyes, take a usb charger for instance
22:55:25 Join lorenzo92 [0] (
22:55:38pamaurythere is no host but there is power
22:56:18thomasjfoxgood job on the documentation, still reading it
22:57:00pamaurytell me if anything is unclear so I can fix it
22:57:12thomasjfoxone question about "bool usb_inserted(void);" it says:
22:57:20thomasjfoxcheck whether USB is plugged, note that this is the official value which has
22:57:21thomasjfox+ * been reported the thread
22:57:32thomasjfoxthe "+" sign is from gitweb
22:58:30thomasjfoxI think that should be "reported by" but what I'm really wondering about is that it's reporting a value even though the function is returning a bool
22:58:43thomasjfoxmay be "official state" is easier to understand in this case
22:58:49pamauryit is because there can be a small "lag" between hardware noticing the state change and the thread being notified about it
22:59:54thomasjfoxah ok. May be that can be clarified when you fix the typo :o)
23:00:40pamauryreported by is ambiguous, because it would "reported by *the usb* thread" then
23:01:26thomasjfoxdo you mean "reported to the thread"?
23:01:42gevaertspamaury: they all look OK to me. I'll let thomasjfox handle documentation comments :)
23:01:45pamauryah it's missing the "to", good catch
23:02:08CtcpIgnored 2 channel CTCP requests in 45 minutes and 38 seconds at the last flood
23:02:08*thomasjfox likes documented code
23:02:59thomasjfoxfor me it's a good test since I know next to nothing about Rockbox's USB code
23:03:11pamauryI still need to write the doc of the drivers
23:04:02thomasjfox"important even" -> "important event"
23:05:08thomasjfox"occured" -> "occurred"
23:07:02thomasjfoxregarding "void usb_set_skip_first_drive(bool skip);" I guess it talks about multiple drives, not drivers??
23:08:43thomasjfoxso the "second driver" should also be "second drive"
23:10:19lorenzo92definitely, we need to document the code
23:11:46 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
23:13:33 Quit fs-bluebot (Ping timeout: 245 seconds)
23:15:08 Quit bluebrother^ (Ping timeout: 264 seconds)
23:15:31 Join fs-bluebot [0] (
23:19:36 Join einhirn [0] (
23:19:50 Join chrisb [0] (
23:20:53 Quit nialv7 (Remote host closed the connection)
23:34:32 Quit Provel (Read error: Connection reset by peer)
23:36:00TheSevenlebellium: this commit was very likely the culprit:;a=commit;h=7d1a47cf13726c95ac46027156cc12dd9da5b855 (breaking simulator builds)
23:38:41 Join Provel [0] (
23:38:55lebelliumTheSeven: okay, we already knew about this commit but I didn't think it would break everything :D;a=commit;h=a2136a8
23:41:36TheSevenboth the touched files and the date when the builds broke match that one
23:42:09TheSevenI'm a bit puzzled why some builds seem to still not work though
23:43:34*thomasjfox just built a Fuze+ simulator
23:43:57lorenzo92I wonder if valgrind can be used on ypr0 :P
23:43:59 Quit bertrik (Ping timeout: 256 seconds)
23:46:51lorenzo92mnt/media0/.rockbox/rockbox: /home/pc_path/rockbox/apps/plugins/xworld/resource.c: Network is unreachable
23:47:13lorenzo92hahahahahahahah seriously, we have some serious troubles on hosted target (or R0 at least) haha
23:47:41lebelliumjust enable Wifi on R0. It will work then
23:47:55thomasjfoxpamaury: "when the target has several drivers" -> "when the target has several drives" ;)
23:48:01lorenzo92oh right, I forgot it, pff stupid errors as alwasy...
23:48:05thomasjfoxsorry 'bout that
23:49:40thomasjfoxpamaury: "can the host only supports access" -> should that be "AND the host only supports access"?
23:49:47thomasjfox(lowercase and of course)
23:57:18***Saving seen data "./dancer.seen"
23:57:22lebelliumthomasjfox: do you get a .rockbox directory when making Fuze+ > Sim > make -j ?
23:57:42TheSevenlebellium: run "make install", then look into the "simdisk" directory
23:57:51thomasjfox"make fullinstall" to be precise IIRC
23:58:13TheSevenfullinstall is probably the same + fonts
23:58:27lebelliumthanks :)

Previous day | Next day