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 2015-01-01

00:00:42 Quit igitoor (Changing host)
00:00:42 Join igitoor [0] (igitur@unaffiliated/contempt)
00:27:11 Nick [Franklin] is now known as [HappyNewYear] (~franklin@unaffiliated/franklin)
00:27:40 Nick [HappyNewYear] is now known as [Franklin] (~franklin@unaffiliated/franklin)
00:27:57 Nick [Franklin] is now known as [HappyNewYear] (~franklin@unaffiliated/franklin)
00:47:41***Saving seen data "./dancer.seen"
01:00:28 Quit shmibs (Quit: obythur!)
01:02:53 Join shmibs [0] (~shmibs@
01:03:05 Quit [HappyNewYear] (Ping timeout: 264 seconds)
01:11:45 Join [HappyNewYear] [0] (
01:18:15 Join pystar89 [0] (
01:29:22 Quit ZincAlloy (Quit: Leaving.)
01:29:25 Join zoktar_ [0] (
01:31:01 Quit zoktar_ (Changing host)
01:31:01 Join zoktar_ [0] (~zoktar@unaffiliated/zoktar)
01:35:58 Quit zoktar (*.net *.split)
01:35:58 Nick zoktar_ is now known as zoktar (~zoktar@unaffiliated/zoktar)
01:50:25 Quit burgobianco (Remote host closed the connection)
01:52:51 Quit AlexP (Remote host closed the connection)
01:54:34 Join burgobianco [0] (
01:54:47 Quit [HappyNewYear] (Ping timeout: 245 seconds)
02:03:37 Join [HappyNewYear] [0] (
02:04:10[HappyNewYear]Happy new year, everybody!
02:10:11 Quit [HappyNewYear] (Changing host)
02:10:11 Join [HappyNewYear] [0] (~franklin@unaffiliated/franklin)
02:17:06 Nick [HappyNewYear] is now known as [Franklin] (~franklin@unaffiliated/franklin)
02:17:52 Nick [Franklin] is now known as [HappyNewYear] (~franklin@unaffiliated/franklin)
02:18:00 Nick [HappyNewYear] is now known as [Franklin] (~franklin@unaffiliated/franklin)
02:26:06sakaxhappy new year guys! wishing you all the best in hunting the usb bug for nano2g :)
02:26:36 Join Strife89 [0] (
02:47:43***Saving seen data "./dancer.seen"
03:03:59[Franklin]foolsh: do you want to help me define some more units for unitconv?
03:08:35 Quit pamaury (Ping timeout: 240 seconds)
03:47:12 Join krabador [0] (~krabador_@unaffiliated/krabador)
03:53:24 Join Riviera [0] (Riviera@2a03:b0c0:1:d0::10:b001)
03:55:43foolsh[Franklin]: Sure, but right now, I'm going to go lay down and feel the room spin with my eyes closed
04:04:58 Quit RiD (Quit: A good plan today is better than a perfect plan tomorrow.)
04:14:44 Quit tchan (Quit: WeeChat 1.0.1)
04:26:54 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
04:39:33 Quit lebellium_ (Quit: ChatZilla [Firefox 35.0/20141222200458])
04:47:46***Saving seen data "./dancer.seen"
04:53:30 Quit Scall (Ping timeout: 265 seconds)
04:59:18 Join Scall [0] (~chat@unaffiliated/scall)
05:13:00 Quit [Franklin] (Ping timeout: 255 seconds)
05:13:21 Join JdGordon [0] (
05:13:21 Quit JdGordon (Changing host)
05:13:21 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
05:15:15 Quit JdGordon_ (Ping timeout: 255 seconds)
05:23:19 Quit TheSeven (Ping timeout: 272 seconds)
05:24:35 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
05:39:15 Quit krabador (Quit: Take the time.)
05:57:44 Join JdGordon_ [0] (~jonno@rockbox/developer/JdGordon)
05:58:43 Quit JdGordon (Ping timeout: 250 seconds)
06:00:10the-kyleHappy New Year!
06:15:20 Quit Provel (Quit: Provel)
06:18:42 Join Provel [0] (
06:47:50***Saving seen data "./dancer.seen"
07:29:45 Join ungali [0] (
07:29:46 Quit ungali (Changing host)
07:29:46 Join ungali [0] (~ungali@unaffiliated/ungali)
07:48:08 Quit Provel (Ping timeout: 256 seconds)
07:49:01 Join Provel [0] (
08:37:18 Quit JdGordon_ (Ping timeout: 245 seconds)
08:47:52***Saving seen data "./dancer.seen"
08:55:08 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
09:33:08 Quit Strife89 (Ping timeout: 244 seconds)
10:10:33 Quit ungali (Quit: ungali)
10:47:53***Saving seen data "./dancer.seen"
10:59:25 Quit Zambezi (Ping timeout: 244 seconds)
11:14:07 Join Zambezi [0] (
11:52:23 Join sobukus [0] (
11:53:15sobukusI seem to have trouble with rtfm ... can I get some printf-style debugging of a plugin on the console with the UI simulator?
11:56:33*sobukus idling around
12:08:53 Join AlexP [0] (~alex@rockbox/staff/AlexP)
12:24:34 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
12:36:17pamauryhappy new year everyone :)
12:47:54***Saving seen data "./dancer.seen"
12:54:02 Join lorenzo92 [0] (
12:54:23lorenzo92guys, happy new year !!! :)
13:00:49lorenzo92kugel: do you think we still need mount bindings to /.rockbox and /Playlists ?
13:02:57lorenzo92also, don't forget g#707
13:03:02fs-bluebotGerrit review #707 at : yp-r0: improve the charging code by Lorenzo Miori
13:22:03 Join thomasjfox [0] (~thomasjfo@rockbox/developer/thomasjfox)
13:25:08sobukuslorenzo92: you got a hint about how to debug plugin code in the simulator?
13:26:55lorenzo92well, if you just need to do some printfs, do them
13:26:57lorenzo92it works
13:27:08sobukusHm, fprintf(stderr, ... did nothing
13:27:22lorenzo92huh! It should work indeed
13:27:24sobukusI though I remember it did.
13:28:28lorenzo92otherwise, you might also want to enable logf when configure is run + #define LOGF_ENABLED in the source code you want to debug. Every logf("printf format") will be then hopefully displayed in the console
13:28:44lorenzo92I'm not really in the simulator business, but it should work
13:31:58thomasjfoxDEBUGF() certainly works with out-of-the-box simulator builds
13:32:05sobukusHm, undef reference to logf ... rb->debugf also doesn't work.
13:32:29sobukusIt just does nothing. Something is strange since I updated my source tree. I think I remember having debug output before.
13:33:12thomasjfoxin the last days I also remember not seeing logf() output from main code even when enabled sim + logf during configure
13:33:24thomasjfoxI haven't investigated though since debugf() worked
13:33:52 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
13:34:24sobukusHm, I'll figure it out.
13:35:06*sobukus thanks for hints and leaves keyboard
13:35:15gevaertsOn a plain sim build with no extra options, printf(), fprintf(stdout). fprintf(stderr), and DEBUGF() all work
13:36:13sobukusgevaerts: OK, thanks for that data point, too.
13:36:37lorenzo92I can definitely agree, printf works fine on hosted targets
13:36:54thomasjfoxgevaerts: do you know by chance how to enable dircache in a simulator build? It's usually disabled in firmware/export/config.h for sims
13:37:13gevaertsthomasjfox: I'm not sure you can
13:37:30gevaertsIsn't that fairly intricatly linked to the underlying FAT implementation?
13:37:34thomasjfoxcrap. I'm pretty sure I've spotted a bad bug and wanted to verify it
13:38:36gevaertsIIRC pamaury once built a sim with disk image support
13:38:43gevaertsDoing that might help
13:39:48thomasjfoxor I wait for jhMikeS to wake up and ask about his code refactoring ;)
13:40:45thomasjfoxI think "struct dircache_runinfo" lacks the "static" keyword. Therefore it might contain all kinds of random memory
13:41:01thomasjfoxI noticed that while looking through all buflib users calling core_alloc
13:41:19thomasjfoxEspecially it will contain a random value for the shrink callback
13:42:06thomasjfoxsee firmware/common/dircache.c
13:42:12gevaertsWhy would that need static?
13:42:19thomasjfoxotherwise that memory is not initialized to zero
13:42:31gevaertsThat's a global
13:43:26thomasjfoxah crap. Tripped over that again
13:43:55gevaertsStatic wouldn't *hurt* of course, it doesn't need to be visible elsewhere
13:44:40*thomasjfox doesn't like implicit initialization
13:44:57*gevaerts does :)
13:46:20pamaurygevaerts: yeah but there were problems with disk support in the sim, though I still think it would be very useful to have this
13:46:33 Join davek [0] (
13:52:16fs-bluebotBuild Server message: New build round started. Revision 812406f, 255 builds, 28 clients.
14:08:09 Join Misanthropos [0] (
14:18:37 Join pimaster [0] (~pimaster@
14:25:45*lorenzo92 also testing bluetooth module on YPR1 and gets some possibly HCI commands out of a test program (chip init + UART)
14:33:27 Join ZincAlloy [0] (
14:37:42pamauryhaving support for bluetooth in rockbox would be nice, a few devices have bluetooth
14:47:58***No seen item changed, no save performed.
14:50:53 Quit Misanthropos (Ping timeout: 244 seconds)
15:00:56lorenzo92I see it completely doable, and I also found a lightweight lib implementing the useful stuff like HCI and some higher protocols
15:02:28ZincAlloy+1 on bluetooth. And I guess DLNA would be nice to have as well
15:05:40 Quit davek (Quit: ChatZilla 0.9.87 [Iceape 2.7.12/20130119143918])
15:05:41lorenzo92do you know whether or not the other devices are using UART-like interfaces to the bt chip? SPI might be used as well
15:06:40pamaurythe thing is that depending on devices, it's implemented completely differently, if I remember correctly:
15:07:06pamaury- zen x-fi 3: uart but uses a AT-like protocol, so don't have a HCI interface, much higher level
15:07:25pamaury- samsung i-don't-remembe-which-one: HCI over uart
15:07:56pamauryalso if you have HCI, you can implement any profile (up to hardware support and wiring) but for high levels it's not the case
15:09:55 Join ender` [0] (
15:11:05pamauryI don't have the time right now, but the API must clearly be thought to handle both cases. For example, the public API would be high level, with one part to enumerate/connect with devices and then some functions for supported profiles.
15:11:17pamauryAnd then a generic implementation based on HCI for the HCI chips
15:20:15 Quit AlexP (Remote host closed the connection)
15:44:56 Join RiD [0] (
15:49:05lorenzo92pamaury: thanks for the info! I can sketch something on gerrit later, in the meantime maybe it would be cool also to continue regarding USB
15:49:22pamauryyeah usb is the priority
15:51:55lorenzo92pamaury: chances to know if the samsung mod. i-don't-remembe-which-one is a BCMxxx?
15:52:41lorenzo92(oh c'mon look,
15:53:04lorenzo92speaker set called rockbox
15:53:44pamauryit's a BCM2048
15:53:53pamauryit's not the only device with bluetooth
15:54:19pamauryBCM2048 too
15:54:52lorenzo92definitely, HCI over UART. I have found the complete datasheet for BCM2070 (the one in R1, doesn't have FM tuner inside)
15:55:12pamaury but I don't know the chip
15:55:12lorenzo92but they are similar in power up pins for example (read similar as same)
15:56:27pamauryI guess most of those are HCI of UART, with a few custom commands to change UART speed and you need to upload the firmware on startup
15:59:21lorenzo92are you sure that I need to upload a firmware? hum it seems to me that BCM2070 has an eeprom, and it doesn't appear to exist a firmware file in the rootfs: I might be wrong.
16:00:57lorenzo92i'm now reading T9 schematics
16:04:01lorenzo92T9: BTTZ0502SA
16:04:54lorenzo92wiki is now informed of this ;)
16:05:54lorenzo92pamaury: it seems that BCM2070 does UART speed autodetecting because I read the same characters both using 9k6 and 115k2
16:07:35pamauryI'm just guessing, it depends on the chip
16:07:47lorenzo92yes, indeed ;)
16:08:14 Join lebellium [0] (
16:08:48lorenzo92lebellium: did you just feel interesting topics? :P
16:11:01lebellium? Haven't read the logs yet :)
16:11:11lorenzo92pamaury: aha look the BTTxx named chip contains a CSR - xxx device
16:28:01thomasjfoxhmm, "Haas Surround" directly in the "Sound settings" menu feels like overkill (regarding g#922)
16:28:03lebelliumlorenzo92: okay I read the logs ;) BTW, I will probably buy a YP-T9 in a few time
16:28:07fs-bluebotGerrit review #922 at : three new DSPs by Chiwen Chang
16:28:07lebelliumand Happy New Year :)
16:28:19lorenzo92thanks ;)
16:28:47 Join Misanthropos [0] (
16:29:09thomasjfoxalso "haas surround" doesn't give many google results in case people want to learn about it...
16:42:35 Quit Misanthropos (Ping timeout: 240 seconds)
16:48:00***Saving seen data "./dancer.seen"
17:00:53 Join [Franklin] [0] (~franklin@unaffiliated/franklin)
17:01:15[Franklin]hey guys, happy new year!
17:05:14thomasjfox[Franklin]: same to you :)
17:05:22thomasjfox[Franklin]: here's a good start:;type=cowond2#prob1
17:07:43[Franklin]looking into it
17:08:28thomasjfoxthe logic looks funny :) "xx == !yy". And "xx != !zz"
17:10:27[Franklin]xx != !yy
17:10:54[Franklin]I think foolsh meant x != y :)
17:11:15[Franklin]also, I think it should be CONFIG_KEYPAD, not CONFIG_KEYMAP :)
17:12:30thomasjfoxha, I didn't even spot that one.
17:12:43[Franklin]I didn't 'til just now :)
17:12:51[Franklin]anyway, pushed as G#1093
17:12:56fs-bluebotGerrit review #1093 at : XWorld: fix some typos in keymaps.h by Franklin Wei
17:13:07[Franklin]First Gerrit change of 2015 :D
17:13:15 Join Strife89 [0] (
17:14:07thomasjfox[Franklin]: did you compile test a f.e. Cowon sim?
17:14:15 Join ploco [0] (dce9b7f9@gateway/web/freenode/ip.
17:14:29[Franklin]I will now
17:15:02[Franklin]hardware, not sim
17:15:10plocothomasjfox: Try google "Haas Effect"
17:15:22[Franklin]it'll be a while
17:15:23thomasjfoxploco: ha, you read the logs ;)
17:16:17thomasjfoxploco: I'm not sure mortal users will make the connection between "haas surround" and "haas effect"
17:17:15[Franklin]"haas sound"
17:17:24[Franklin]gives the same top result
17:17:49[Franklin]with G#1093 applied?
17:17:53fs-bluebotGerrit review #1093 at : XWorld: fix some typos in keymaps.h by Franklin Wei
17:18:07thomasjfoxccache + "make -j8" really speeds up the build
17:18:21[Franklin]This is my first D2 build
17:18:29[Franklin]plus beta's down
17:19:35[Franklin]thomasjfox: I see the error
17:19:47[Franklin]it's OR'ing the results of the inequalities
17:19:59[Franklin]so it always evaluates as true
17:20:08[Franklin]it should be AND'ing them instead, I think
17:22:17thomasjfoxwith the AND change, it compiles without error
17:22:21[Franklin]ok, pushed it
17:23:01thomasjfoxI'll try a DX50 sim next
17:24:51thomasjfoxok, DX50 chokes because of missing Android SDK. Back to D2...
17:26:54thomasjfoxONDAVX777 is fine, too
17:26:58thomasjfoxI'll push it
17:28:23fs-bluebotBuild Server message: New build round started. Revision b0277e4, 255 builds, 28 clients.
17:31:41plocothomasjfox: btw, those Dsps does free memory when disable now.
17:32:34 Quit RiD (Quit: A good plan today is better than a perfect plan tomorrow.)
17:32:35thomasjfoxploco: I've seen it. Pretty cool
17:32:46thomasjfoxanyway, I've got to run, friends just visited
17:33:19plocosure. I'm off to bed too.
17:33:29 Quit ploco (Quit: Page closed)
17:35:09thomasjfox[Franklin]: I'll hope the build will be fine now ;) cu
17:35:22[Franklin]thomasjfox: bb and HNY!
17:35:29 Join RiD [0] (
17:35:39 Join rela [0] (~x@pdpc/supporter/active/rela)
17:36:12 Quit thomasjfox (Quit: Konversation terminated!)
17:36:29fs-bluebotBuild Server message: Build round completed after 486 seconds.
17:37:03[Franklin]ok, all green except for the usual suspects :)
17:40:28 Join AlexP [0] (~alex@rockbox/staff/AlexP)
17:45:48*foolsh hangs his head
17:46:06foolshstill have to rewrite that
17:46:40[Franklin]foolsh: fixed ;)
17:48:59 Join scorche [0] (~scorche@rockbox/administrator/scorche)
17:51:38 Quit scorche`` (Ping timeout: 256 seconds)
17:54:24[Franklin]there's 3 kind of "gallons" >:
17:54:51[Franklin]I'll just do US gallons for now, because that's all I care :)
17:55:46foolshI would suggest only the most popular archaic units, or else the useful plugin you're writing starts to fall in to the useless catagory ;)
17:56:28[Franklin]hey, at least pounds are base-16 :)
17:56:44[Franklin]sort of
18:02:43 Quit [Franklin] (Ping timeout: 245 seconds)
18:04:58 Join [Franklin] [0] (
18:37:02pamaurygevaerts: do you have any why config.h defines USB_STATUS_BY_EVENT for bootloader usb ?
18:37:07pamaury*any idea
18:37:38pamauryand also why this check is enclosed in SWCODEC ????
18:43:05pamauryalso it seems to me that HAVE_BOOTLOADER_USB_MODE implies HAVE_USBSTACK
18:43:09 Quit RiD (Quit: A good plan today is better than a perfect plan tomorrow.)
18:45:06 Join RiD [0] (
18:48:03***Saving seen data "./dancer.seen"
18:50:54pamauryI really feel we are lacking "soc config files"
18:51:33pamaurywe end up having plenty of things duplicated in all target config files or in config.h where they do not belong
18:51:56pamauryand soc header files are not included in config.h which make then useless for some purposes
19:11:21pamauryalso I'm a bit puzzled by the use of usb_powered() in powermgmt.c and CURRENT_USB
19:11:36pamaurymost targets define CURRENT_USB as 500mA, and it is used for current estimation
19:14:02pamauryalso usb_powered() is highly miss-named, if I understand correctly it means "POWERED state", aka "usb is inserted but for charging only"
19:51:06 Quit [Franklin] (Ping timeout: 245 seconds)
19:56:07 Join [Franklin] [0] (
20:20:42 Join bertrik [0] (~bertrik@
20:20:43 Quit bertrik (Changing host)
20:20:43 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
20:26:53 Quit [Franklin] (Ping timeout: 245 seconds)
20:32:25 Join [Franklin] [0] (
20:37:05 Nick bertrik is now known as bertrik_trein (~bertrik@rockbox/developer/bertrik)
20:37:47 Nick [Franklin] is now known as [HappyNewYear201 (
20:37:54 Nick [HappyNewYear201 is now known as [Happy2015] (
20:38:12 Quit [Happy2015] (Changing host)
20:38:12 Join [Happy2015] [0] (~franklin@unaffiliated/franklin)
20:38:24 Nick [Happy2015] is now known as [HappyNewYear] (~franklin@unaffiliated/franklin)
20:41:15 Quit bertrik_trein (Ping timeout: 240 seconds)
20:48:07***Saving seen data "./dancer.seen"
20:53:01pamaurygevaerts: ping
20:57:29 Join advcomp2019_ [0] (~advcomp20@unaffiliated/advcomp2019)
21:00:16 Quit advcomp2019__ (Ping timeout: 245 seconds)
21:12:00 Join bertrik_trein [0] (~bertrik@
21:12:26 Quit RiD (Quit: A good plan today is better than a perfect plan tomorrow.)
21:16:53 Join RiD [0] (
21:39:16 Quit bertrik_trein (Quit: Lost terminal)
21:52:54 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
21:55:44 Quit rela (Ping timeout: 265 seconds)
21:56:22 Quit Jinx (Quit: reboot)
21:59:11 Quit [HappyNewYear] (Ping timeout: 244 seconds)
22:02:45 Join thomasjfox [0] (~thomasjfo@rockbox/developer/thomasjfox)
22:04:49 Join [HappyNewYear] [0] (
22:04:54 Join krabador [0] (~krabador_@unaffiliated/krabador)
22:06:59TheSevenplease let me know if anything is fishy with my buildclient or nightly/android builds, the HDD on which that stuff was running has failed a few days ago
22:07:32TheSeveneverything should be back to normal now, but if anything is still acting up, please notify me
22:07:41 Join rela [0] (~x@pdpc/supporter/active/rela)
22:12:50 Quit rela (Ping timeout: 272 seconds)
22:25:43 Quit [HappyNewYear] (Changing host)
22:25:43 Join [HappyNewYear] [0] (~franklin@unaffiliated/franklin)
22:37:36 Join einhirn [0] (
22:40:40 Quit krabador (Quit: Take the time.)
22:48:08***Saving seen data "./dancer.seen"
22:48:51 Quit [HappyNewYear] (Ping timeout: 240 seconds)
22:55:18 Quit lorenzo92 (Ping timeout: 264 seconds)
23:01:48 Quit Strife89 (Ping timeout: 250 seconds)
23:03:08 Join AlwaysHi- [0] (~AlwaysHig@
23:03:26 Join soap_ [0] (~soap@rockbox/staff/soap)
23:03:42 Join Scromple_ [0] (~Simon@
23:03:42 Quit Scr0mple (Read error: Connection reset by peer)
23:04:02 Quit Zambezi (Ping timeout: 250 seconds)
23:04:36 Join skx_netb [0] (~sakax@unaffiliated/sakax)
23:05:30gevaertspamaury: pong?
23:05:45 Quit Scall (Ping timeout: 256 seconds)
23:05:45 Quit sakax (Ping timeout: 256 seconds)
23:05:45 Quit AlwaysHigh (Ping timeout: 256 seconds)
23:05:45 Quit soap (Ping timeout: 256 seconds)
23:05:46 Join lebellium_ [0] (
23:05:48 Quit thomasjfox (Ping timeout: 264 seconds)
23:05:48 Join Zambezi_ [0] (
23:05:58 Join Scall- [0] (~chat@unaffiliated/scall)
23:05:59 Nick Scall- is now known as Scall (~chat@unaffiliated/scall)
23:06:00pamauryI have a few thoughts on USB, I'd like your opinion
23:06:18 Join HoloIRCUser1 [0] (
23:06:29pamauryfirst, I don't understand what usb_powered() means in the current code
23:06:33 Nick HoloIRCUser1 is now known as HoloIRCUser1hf (
23:06:41pamauryit returns true if usb_state == USB_POWERED
23:07:04pamauryversus usb_inserted() which checks for usb_state == USB_POWERED or USB_INSERTED
23:07:57pamaurythe problem is that my understand of usb.c is that USB_POWERED is now only a transitional state when a host is present, or a permanent state when charging only, so it does not represent the usb "powering state"
23:08:05 Part HoloIRCUser1hf
23:08:34*gevaerts nods
23:09:04pamauryso my question is: did usb_powered() changed meaning without anyone noticing when usb rework was done ?
23:09:10 Quit lebellium (Ping timeout: 250 seconds)
23:09:11 Nick lebellium_ is now known as lebellium (
23:10:41gevaertsNot that many things call usb_powered()
23:10:55pamauryyeah which brings my second question
23:11:01gevaertsIt probably did subtly change
23:11:12pamaurythe only use of usb_powered() is strange
23:11:23gevaertsThere's another one, in apps/main.c
23:11:38gevaertsOh, and in battery_bench
23:11:42pamauryyeah, this one I don't understand, maybe there is something subtle there
23:12:04pamaurythe one in the plugin clearly shows that usb_powered() seems to have changed meaning
23:12:15 Quit bluebrother (Disconnected by services)
23:12:20 Join bluebrother^ [0] (~dom@rockbox/developer/bluebrother)
23:12:27gevaertsI don't think it's changed meaning as such
23:12:39gevaertsIt's just that we got a better way to do things along the way
23:13:31pamaurywhat do you mean ?
23:13:55gevaertsI think usb_powered() still returns 1 in exactly the same circumstances it did before
23:14:31 Quit fs-bluebot (Ping timeout: 265 seconds)
23:14:43pamauryhum, I'm not so sure, or else it's weird because if you plug usb and the usb stack kicks in, the usb_state is USB_INSERTED so usb_powered() returns false even though the device is powered/charging from usb
23:16:07gevaertsYes, but the bit in powermgmt.c seems to explicitly handle that, I think
23:16:16 Join fs-bluebot [0] (
23:17:09gevaertsAnd that's been untouched for ages
23:17:36pamauryok, then the name is really misleading
23:18:05pamauryit has nothing to do with the powering state
23:18:11pamaurywell not much
23:18:40gevaertsThere's also the wps tag that's not documented :)
23:18:48gevaerts[Saint]: what exactly does %bu mean?
23:19:32gevaertsI'm not saying you're wrong, it's just an inconsistency that predates me, so I don't know...
23:20:52pamauryI'd just like to get the documentation right, and possibly rename the function :)
23:22:49pamaurydo you have targets in mind which can do usb but cannot charge or be powered from it ?
23:23:03gevaertsOne issue with things like USB_INSERTED is that they're a return value from usb_detect() and a state in usb.c, where it means subtly different things
23:23:11gevaertsThe archoses, at least
23:23:26pamauryI built a list of defines here: pamaury/b3db692c1880ef97a0f2/raw/f4a2740a351bf8304b6abcb9d5666c4978b1d9e4/usb_defines.txt">
23:23:27gevaertsMaybe not the ondios, not sure about those, but definitely the player and the recorder
23:24:41pamaurystrange, the player and recorder don't define HAVE_USBSTACK my tool says
23:25:00gevaertsAh, you mean software USB?
23:25:32pamauryah I forgot about hardware USB again
23:27:38[Saint]gevaerts: %bu is, iirc, "is USB plugged? <yes|no|fucked if I know>"
23:28:14gevaerts[Saint]: what should it return if USB is in mass storage mode?
23:28:29gevaerts(apart from "Who cares? You can't use skins then anyway!")
23:28:44[Saint] /can/...
23:28:50[Saint]Just, not reliably.
23:30:24gevaertspamaury: in the software USB world, your defines list also says some RK27xx things, some ondas, and the ZVMs don't use USB power
23:30:38gevaertsThose might be bugs though
23:30:54[Saint]gevaerts: regarding your question, I'm not sure it matters...
23:30:59[Saint]"10:12JdGordon%bu is being "deprecated".. enter %up
23:30:59[Saint]10:12JdGordonbecuase the line lengths are not static"
23:31:06[Saint]it seems its just a hanger on.
23:31:13pamauryit looks like a bug, on the other hand, HAVE_USB_POWER is so nearly useless no one noticed
23:31:22 Join thomasjfox [0] (~thomasjfo@rockbox/developer/thomasjfox)
23:31:24[Saint]leave it to do whatever it may or may not do...
23:31:29gevaerts[Saint]: the thing is, we want to figure out what usb_powered is supposed to mean, and since %bu is just that... :)
23:31:48[Saint]well, I think %up is what you want now.
23:32:16[Saint]you'd likely have to ask JdGordon, I've lost track of what he considers viable/deprecated.
23:32:35[Saint]There's a fair few tags that exist still, but /really. shouldn't be used.
23:33:17[Saint]IIRC he bailed half way through the USB skinning debacle, so I'm not confident about wither.
23:33:38gevaerts[Saint]: that's where you're mistaken. We want to know what usb_powered() means, we don't care about what tags you should or shouldn't use :)
23:33:45pamauryit seems to mean that the charging source should be left to power_input_status()
23:34:25pamauryhowever main targets probably wrongly return POWER_INPUT_MAIN_CHARGER instead of POWER_INPUT_USB_CHARGER
23:34:54gevaertsCould well be
23:35:42pamaurywell, the doc in power.h is ambiguous, it *seems* to say that main == usb if those are not discernable
23:36:13gevaertspamaury: only the archoses have CURRENT_USB set to something meaningful
23:37:31gevaertsSo the powermgmt bit for relevant targets really is usb_inserted() || usb_powered(), where the usb_powered() really is redundant anyway
23:37:31pamauryI'm not sure the value is *useful* because runcurrent() doc says it estimates the consumption for the run time, clearly it doesn't consume 500mA just to run, or I don't understand the meaning of it
23:37:43gevaertsWell, no
23:37:56gevaertsBut it probably assumes the disk is spinning all the time while connected
23:40:11pamauryhum, if usb_inserted() || usb_powered() is true (which is the case when usb is plugged), then current = CURRENT_USB = 500mA
23:40:24pamauryso runcurrent() will returns something >= 500mA
23:40:43pamaurythen it does CURRENT_MAX_CHG - runcurrent()
23:40:44 Join [HappyNewYear] [0] (
23:40:47 Quit ender` (Quit: History is a set of lies agreed upon. -- Napoleon Bonaparte)
23:41:10pamaurywhich is 350mA by default
23:41:17pamauryand the archos don't define it, so it's the default
23:42:20gevaertsRemember that if those are charging, it's not over USB
23:44:38pamaurybut then the check is just usb_inserted() so my point still holds ?
23:44:51pamaury(because they don't have HAVE_USB_POWER)
23:45:09gevaertsGood point
23:45:22pamauryso I don't understand what CURRENT_USB is supposed to represent
23:46:20pamauryit seems to me it should be "consumption when usb plugged (exclusing charging)" and some target define it as "maximum usb current draw"
23:47:09gevaertsSo usb_powered() is irrelevant in powermgmt.c, and can almost certainly be replaced (and maybe should be) by usb_inserted() in the skin engine and battery_bench, so the only remaining case is main.c
23:47:11pamauryI'm sorry to bother you, I'm just trying to understand this mess :)
23:47:34pamauryyeah, we should probably remove usb_powered(), now in main.c I don't understand the comment
23:47:34gevaertsI *never* looked into powermgmt, so your guess is as good as mine :)
23:48:11pamauryI'm not sure I understand the goal of this loop in main.c
23:48:28gevaertsSo-called early USB
23:49:05pamauryit prevents full init if USB is plugged because with threads being created it would be mess
23:49:18pamauryalso with disk mounting
23:49:25gevaertsIt allows USB before the disk has been touched
23:49:32pamaurymake sense
23:49:40gevaertsBasically the equivalent of bootloader USB on bootloader-less targets
23:49:54gevaertsWhich, come to think of it, might be the archoses again :)
23:50:12gevaertsAlthough it *could* be flashed rockbox on the irivers, not sure about those
23:50:13fs-bluebotBuild Server message: New build round started. Revision 9076b43, 255 builds, 28 clients.
23:50:32pamauryI *think* this usb_powered() is correct: if usb is plugged to a charger only, usb_detect() is also USB_INSERTED and the loop would never stop
23:50:50pamaurybut if there is no host (USB_POWERED) then you can safely continue booting
23:52:14pamauryhowever this seems slightly buggy because USB_POWERED is also a transitional state
23:52:40pamauryor does the main loop also needs to ack the SYS_USB_CONNECTED message ?
23:52:53gevaertsI think
23:53:23*gevaerts isn't sure
23:53:57pamauryhum, that's nearly ok but there is a race condition I think:
23:54:26pamaury1) when the host is detected, usb_state is set to USB_POWERED and SYS_USB_CONNECTED is broadcast
23:54:29gevaertsHmmm, it probably does
23:54:43gevaertsIt depends on the timeout there I think
23:54:48gevaertsWhich is rather risky
23:54:49pamaury2) so button_get_w_tmo() returns SYS_USB_CONNECTED and usb gui is started
23:55:14pamaurybut if usb_powered() is seems before the call to the button, or a button is pressed before the timeout, that's a potential disaster
23:57:51gevaertsI suspect you're right
23:58:36pamauryI thought there was a timeout on host detection, was I wrong ?

Previous day | Next day