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 2012-09-25

00:00:59 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
00:05:08 Join Totalled_ [0] (
00:07:57 Quit sciopath (Read error: Connection reset by peer)
00:08:39 Join sciopath [0] (
00:09:13 Quit Totalled (Ping timeout: 246 seconds)
00:09:13 Nick Totalled_ is now known as Totalled (
00:09:17 Quit soap (Ping timeout: 246 seconds)
00:09:17 Quit LittleCreature (Ping timeout: 246 seconds)
00:09:51 Join soap [0] (
00:10:05 Join LittleCreature [0] (
00:11:03 Quit bertrik (Remote host closed the connection)
00:12:24 Quit soap (Changing host)
00:12:24 Join soap [0] (~soap@rockbox/staff/soap)
00:13:01 Nick LittleCreature is now known as Guest51656 (
00:14:00 Join thegeek_ [0] (
00:16:45 Join kugel_ [0] (~kugel@
00:16:45 Quit kugel_ (Changing host)
00:16:45 Join kugel_ [0] (~kugel@rockbox/developer/kugel)
00:16:47 Join simabeis [0] (
00:17:30 Quit Guest64089 (Quit: Ik ga weg)
00:18:32 Quit simabeis_ (Ping timeout: 245 seconds)
00:18:32 Quit kugel (Read error: Connection reset by peer)
00:18:32 Quit thegeek (Ping timeout: 245 seconds)
00:21:22 Quit shamus (Read error: Connection reset by peer)
00:21:22 Quit ender` (Read error: Connection reset by peer)
00:21:44 Join shamus [0] (
00:24:38 Quit pystar89 (Ping timeout: 268 seconds)
00:26:22 Quit pixelma (Ping timeout: 268 seconds)
00:26:32 Join pixelma [0] (pixelma@rockbox/staff/pixelma)
00:34:49 Quit mgottschlag2 (Read error: Connection reset by peer)
00:39:27 Quit lebellium (Quit: ChatZilla 0.9.89 [Firefox 16.0/20120919065210])
00:42:17 Quit pixelma (Disconnected by services)
00:42:17 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
00:42:19 Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
00:42:42 Join Syconaut [0] (
00:43:17 Quit bootlkjkgf (Read error: Connection reset by peer)
00:43:24 Join bootinfdsds [0] (~Prmhfhfx@
00:44:15 Quit Syconaut^ (Ping timeout: 256 seconds)
00:44:55 Join thegeek [0] (
00:45:54 Join kugel [0] (~kugel@
00:45:54 Quit kugel (Changing host)
00:45:54 Join kugel [0] (~kugel@rockbox/developer/kugel)
00:46:54 Quit ser (Remote host closed the connection)
00:47:39 Quit factor (Ping timeout: 256 seconds)
00:50:55 Quit kugel_ (Read error: Connection reset by peer)
00:51:40 Quit thegeek_ (Ping timeout: 241 seconds)
00:56:35 Join thegeek_ [0] (
01:02:03 Quit thegeek (Read error: Connection reset by peer)
01:02:04 Quit soap (Ping timeout: 252 seconds)
01:03:20 Join soap [0] (~soap@rockbox/staff/soap)
01:12:07 Nick Guest51656 is now known as LittleCreature (
01:12:14 Quit LittleCreature (Changing host)
01:12:14 Join LittleCreature [0] (~fearofmus@unaffiliated/fearofmusic)
01:28:35 Quit Rondom (Disconnected by services)
01:28:54 Join Rondom_ [0] (~rondom@2a01:488:66:1000:b24d:4f2f:0:1)
01:30:08 Join factor [0] (
01:30:29 Join crwll [0] (
01:30:38 Quit crwl (Ping timeout: 272 seconds)
01:36:51 Quit Rondom_ (Remote host closed the connection)
01:38:48 Quit XavierGr (Disconnected by services)
01:38:48 Join XavierGr [0] (XavierGr@rockbox/staff/XavierGr)
01:38:49 Join Rondom [0] (~rondom@2a01:488:66:1000:b24d:4f2f:0:1)
01:49:54***Saving seen data "./dancer.seen"
01:51:55 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
01:51:56 Quit t0rc (Client Quit)
02:10:58 Quit n1s (Quit: Ex-Chat)
02:44:15 Join derf [0] (
02:45:58 Join perrikwp [0] (
03:27:10 Quit XavierGr ()
03:49:56***Saving seen data "./dancer.seen"
04:08:26 Quit sciopath (Read error: Connection reset by peer)
04:08:48 Join sciopath [0] (
04:13:21 Join Scromple [0] (~Simon@
04:21:46 Quit pixelma (Disconnected by services)
04:21:46 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:21:47 Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:21:47 Quit amiconn (Disconnected by services)
04:21:48 Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
04:21:52 Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
04:36:03 Join advcomp2019_ [0] (
04:36:04 Quit advcomp2019_ (Changing host)
04:36:04 Join advcomp2019_ [0] (~advcomp20@unaffiliated/advcomp2019)
04:38:07 Join TheSphinX^ [0] (
04:39:24 Quit advcomp2019 (Ping timeout: 256 seconds)
04:42:16 Quit TheSphinX_ (Ping timeout: 272 seconds)
04:45:15 Quit zoktar (Ping timeout: 260 seconds)
05:27:46 Quit [7] (Disconnected by services)
05:27:55 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
05:37:39 Join JdGord [0] (~AndChat80@
05:40:05 Quit JdGord (Read error: Connection reset by peer)
05:42:48 Join JdGord [0] (~AndChat80@
05:44:58 Quit JdGord (Read error: Connection reset by peer)
05:48:01 Join JdGord [0] (~AndChat80@
05:49:59***Saving seen data "./dancer.seen"
05:50:18 Quit JdGord (Read error: Connection reset by peer)
05:53:02 Join JdGord [0] (~AndChat80@
05:54:59 Quit JdGord (Read error: Connection reset by peer)
05:55:53 Join JdGord [0] (~AndChat80@
05:58:12 Quit JdGord (Read error: Connection reset by peer)
06:00:35 Join JdGord [0] (~AndChat80@
06:02:34 Quit JdGord (Read error: Connection reset by peer)
06:05:43 Join JdGord [0] (~AndChat80@
06:07:57 Quit JdGord (Read error: Connection reset by peer)
06:16:25 Quit sciopath (Read error: Connection reset by peer)
06:16:47 Join sciopath [0] (
06:38:16 Join JdGord [0] (~AndChat80@
06:40:35 Quit JdGord (Client Quit)
06:54:36 Quit soap (Ping timeout: 245 seconds)
06:59:39 Quit nosa-j (Ping timeout: 260 seconds)
07:01:04 Join nosa-j [0] (~m00k@
07:02:19 Join soap [0] (~soap@rockbox/staff/soap)
07:10:22 Quit Scromple (Quit: Leaving)
07:12:44 Join Scromple [0] (~Simon@
07:22:53 Quit the-kyle (Ping timeout: 264 seconds)
07:42:06 Quit Rower85 (Quit: Hmmm...)
07:49:31 Join mortalis [0] (
07:50:00***Saving seen data "./dancer.seen"
07:53:00 Quit factor (Read error: Connection reset by peer)
07:53:54 Join stoffel [0] (
07:58:51 Join kevku [0] (
08:10:12 Join ender` [0] (
08:10:40 Join factor [0] (
08:41:03 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
08:57:09 Quit bertrik (Remote host closed the connection)
08:59:13 Join Zagor [0] (
08:59:14 Quit Zagor (Changing host)
08:59:14 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
09:27:31 Quit scorche` (Ping timeout: 245 seconds)
09:29:27 Quit stoffel (Remote host closed the connection)
09:50:03***Saving seen data "./dancer.seen"
09:53:13 Join LinusN [0] (
10:14:46 Quit kevku (Ping timeout: 272 seconds)
10:16:04 Join einhirn [0] (
10:16:14 Join kevku [0] (
10:22:39 Join lebellium [0] (
10:35:28amiconnZagor: ping
10:36:07amiconnIn case you didn't see it - I repaired my android ndk. Do I need to do anything special to re-enable my client for android builds?
10:36:48Zagorno, it gets unblocked automatically after 10 minutes
10:36:53 Quit mikroflops (Ping timeout: 268 seconds)
10:37:20amiconnah ok
10:38:37 Join mikroflops [0] (
10:44:55 Join wodz [0] (
10:45:36 Quit bootinfdsds (Read error: Connection reset by peer)
10:45:45 Join sciopat [0] (~sciopath@
10:46:15 Join bootinfdsds [0] (~Prmhfhfx@
10:48:53 Quit sciopath (Ping timeout: 240 seconds)
11:01:28 Quit Barahir (*.net *.split)
11:01:29 Quit sinthetek (*.net *.split)
11:01:29 Quit wtachi (*.net *.split)
11:01:36 Join Barahir [0] (
11:01:36 Join sinthetek [0] (
11:01:37 Quit sinthetek (Changing host)
11:01:37 Join sinthetek [0] (~sinthetek@unaffiliated/sinthetek)
11:01:49 Join wtachi [0] (
11:03:17 Join n1s [0] (
11:03:17 Quit n1s (Changing host)
11:03:17 Join n1s [0] (~n1s@rockbox/developer/n1s)
11:04:32n1swodz: thanksd
11:05:00wodzpasswords are standard: root:root and user:user :P
11:21:04 Quit n1s (Read error: Connection timed out)
11:21:41 Join Thra11 [0] (~thrall@
11:24:01 Join Thra11_ [0] (~thrall@
11:26:42 Quit Thra11 (Ping timeout: 265 seconds)
11:29:51wodzmortalis: ping
11:29:58mortaliswodz: pong
11:30:35wodzmortalis: you are not using DWDMA irq for anything other then counting lcd updates, right?
11:32:36mortalisyes, but i think we could use them to know when transfer complete
11:33:03wodzwe could it it works reliably
11:33:54wodzwe could then setup semaphore released in IRQ just as we do in sd driver
11:37:19wodzI still see possible race condition. DMA will signal finish when it transfers last 4 lines to the buffer. This lines still need to be moved to lcd controller. This leaves small time gap where new transfer can corrupt data in lcd controller.
11:40:12[Saint]<AlexP> I think I'm running out of skin buffer now <−− Not technically possible
11:41:31wodz[Saint]: why not possible? mem has finite size
11:41:50mortaliswhy do you think dma send finish signal while transfering last 4 lines and not when they are already transfered?
11:42:12CIA-10Commit afc6b3f in rockbox by Nils Wallménius: opus: asm MULT16_32_Q15 for arm and cf
11:42:20[Saint]<AlexP> As if I use a backdrop viewport then it won't show my backdrop set in Rockbox, I have to load it in the wps using %x correct?
11:42:20[Saint]<AlexP> Well, if I do that I get the fallback <−− This is a bit of a tricky one, but, to do this you must first define a fullscreen viewport, then %x your backdrop at 0,0, then define another fullscreen viewport and continue coding as normal after this.
11:42:30 Quit bluebrother (Disconnected by services)
11:42:36 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
11:42:42wodzThey are transfered but they end up in lcdbuf. lcdif will transfer it internally to lcd controller but this will take some time
11:42:46[Saint]wodz: "not technically possible" in so far as bufflib will allow you to eat all remaining buffer.
11:43:18[Saint]So, you'd have no audio buffer left...making the wps irrelevant as playback wouldn't be possible ;)
11:43:41wodz[Saint]: true but out-of-buf IS technically possible
11:44:29[Saint]well, yeah...I did word it badly. My point was, if you're out of "skin buffer", then you're out of audio buffer the skin buffer doesn't matter :)
11:44:43 Quit n17ikh (Ping timeout: 268 seconds)
11:45:07mortaliswodz: understood, we could add some delay after mutex_lock.
11:45:41 Quit fs-bluebot (Ping timeout: 264 seconds)
11:46:15wodzmortalis: we need to check if it is needed anyway
11:47:05 Join fs-bluebot [0] (
11:49:42wodzmortalis: IMO this while (!(LCDC_STA & LCDC_MCU_IDLE)); should handle the case but when testing I couldnt record any case where it didn't return IDLE state of lcdif
11:50:05***Saving seen data "./dancer.seen"
11:51:56mortalismaybe it's just very rare case.
11:53:54wodzplease update g#306 so I could test it on my DAP
11:53:56fs-bluebotGerrit review #306 at : rk27xx lcd code rework by Marcin Bukat (changes/06/306/2)
11:56:29 Join n1s [0] (
11:56:29 Quit n1s (Changing host)
11:56:29 Join n1s [0] (~n1s@rockbox/developer/n1s)
11:56:59n1swodz: got it
11:58:08wodzn1s: I think it would be great if you post test_codec results for opus
12:01:25kugeln1s: dont you need to reset the emac to its old state?
12:02:08n1swith the latest push the 64 kbps test file i've been testing with is 57.53% realtime on cf (215.54MHz), 85.87% realtime on c200 (93.16MHz) and 221.04% realtime on fuzev1 (112.19MHz)
12:02:33n1skugel: no, everything that uses it is responsible to set it up to what is required
12:04:34wodz57.53% on cf sounds like a lot needs to be done for opus to be usable on this platform
12:05:34n1swodz: still, no iram is used at all, and it helps quite a lot, with my hacky patches it gets to ~90%
12:06:07n1sbut the way they decided to allocate memory makes doing it in a nice way hard
12:08:53n1skugel: the emac macsr is also stored when switching context so each thread can have its own mode
12:09:43 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
12:14:26n1stwo state structs with at least one having a flexible array member allocated in a single buffer and accessed with offsets, and some weird macro voodoo to only memset certain parts of the buffer when clearing the state
12:14:29mortaliswodz: done. I didn't test it on my dap yet, but it should work.
12:14:51 Quit Thra11_ (Remote host closed the connection)
12:15:36wodzmortalis: I'll test it on mine when I get back home
12:18:08 Join Thra11 [0] (~thrall@
12:18:40n1salso writing some better asm functions for the hot paths will probably help quite a bit, since these very short inlines will hit pipeline stalls on both cf and most arm cores better than arm7
12:19:06 Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
12:23:53wodzn1s: I guess this control structs don't fit in iram
12:25:10n1swell, maybe i have not dug through it all but they are not statically sized so i'd need to find the worst case size
12:26:27 Quit mgottschlag (Read error: Connection reset by peer)
12:27:12kugeln1s: you could use buflib to dynamically alloc from iram
12:27:41[Saint]Ooooooh, that sounds like fun.
12:27:55n1sit's not much fun actually
12:28:25[Saint]Well, I was just going to say: "where fun == "not very fun at all"" ;)
12:29:01n1stremor does dynamic iram allocs and quite a lot of thinking was required to get it "right"
12:30:35n1syou need to think about the order of allocs and what to do if one fails and hope the compiler or some global header or something changes somethign so an iram function grows too big so more allocs fair, etc
12:31:16n1ss/changes/doesn't change/
12:31:29n1salso s/fair/fail/
12:32:14n1sanyway, i'd like to try getting it realtime with static allocs for now and look at that stuff if necessary later
12:32:51n1sthe memory constraints on lowmem targets are also something to look at
12:33:24 Part amayer
12:37:37 Quit GeekShad1w (Remote host closed the connection)
12:37:52 Join GeekShadow [0] (
12:42:45 Quit [Saint] (Read error: Connection reset by peer)
12:44:29 Join [Saint] [0] (~saint@rockbox/user/saint)
12:50:44 Join Guest30566 [0] (
12:51:03 Quit kevku (Ping timeout: 260 seconds)
12:51:47 Nick Guest30566 is now known as dfkt_ (
12:51:50 Join Thra11_ [0] (~thrall@
12:52:04 Quit dfkt_ (Changing host)
12:52:04 Join dfkt_ [0] (dfkt@unaffiliated/dfkt)
12:52:17 Quit guymann (Ping timeout: 264 seconds)
12:52:37 Join guymann [0] (~c@unaffiliated/guymann)
12:53:39 Quit Thra11 (Ping timeout: 248 seconds)
12:54:54 Join Thra11 [0] (
12:56:21 Join the-kyle [0] (
12:56:31 Quit Thra11_ (Ping timeout: 252 seconds)
12:59:58 Join kevku [0] (x@2001:470:28:773::3)
13:02:30 Join Thra11_ [0] (
13:05:23 Quit Thra11 (Ping timeout: 248 seconds)
13:05:50freqmodn1s: have you discussed the changes you are doing with the opus-guys
13:06:00freqmodto try to keep the source trees as similar as possible?
13:06:15n1sfreqmod: no, not yet at least
13:06:47 Join petur [0] (~petur@rockbox/developer/petur)
13:07:30freqmodmaybe you could mention it on #opus and ask if they are interrested in cooperating
13:07:41freqmodthere are at least some common nicks on #opus and #rockbox
13:08:43 Nick dfkt_ is now known as dfkt (dfkt@unaffiliated/dfkt)
13:09:15n1sdo you think they would be interested in a patch to add a config option to disable some encoding only code and more of the code only used when CUSTOM_MODES is enabled
13:09:56n1sthat's the only slightly intrusive thing i've done so far
13:10:08freqmodi don't know, but there were some people here which sounded like they worked on it asking what was required for opus support when i was looking at it 1 month ago
13:10:31n1scutting out unused code is important to free up memory on our low mem targets
13:11:09freqmodmaybe an ifdef encoding could be introduced if it doesn't add too much code
13:16:12wodzn1s: what was wrong with crc? mac mode?
13:16:42 Quit the-kyle (Ping timeout: 246 seconds)
13:16:46n1sthe other things will mostly be some inline asm functions that can probably be in separate header files and sprinkling of some rockbox specific attributes and possibly some very rockbox specific memory allocation changes
13:17:27n1swodz: yes, i forgot to set it in my initial attempt
13:18:05n1salso for some reason it took me a while to remember the 31 pos shift
13:18:28n1sso the pushed version gives exactly the same output as the c code
13:21:55 Join the-kyle [0] (
13:25:47*kugel wonders how hard would it be to re-use %ac etc to align images
13:50:08***Saving seen data "./dancer.seen"
13:58:32[Saint]Hmmm...where in the code do I find the preferential ordering of album art?
14:00:01[Saint]Is file based or embedded art preferred?
14:06:11 Quit kevku (Ping timeout: 260 seconds)
14:17:13pixelma*IIRC* the wiki or the manual said, embedded album art was preferred
14:22:14 Join amayer_ [0] (
14:23:02 Quit amayer_ (Remote host closed the connection)
14:25:17 Quit wodz (Quit: Leaving)
14:29:50 Join shai [0] (
14:30:05 Quit shai_ (Ping timeout: 264 seconds)
14:32:14 Join WalkGood [0] (~4@unaffiliated/walkgood)
14:37:29 Join XavierGr [0] (XavierGr@rockbox/staff/XavierGr)
14:47:26 Join einhirn [0] (
15:01:57 Join amayer_ [0] (
15:03:54 Quit Totalled (Quit: PETTAN PETTAN, TSURUPETTAN!)
15:41:51 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
15:46:44 Quit mortalis (Quit: Leaving)
15:50:10***Saving seen data "./dancer.seen"
15:50:41 Join lorenzo92 [0] (~chatzilla@
15:57:49 Join Rower85 [0] (
16:13:55 Join einhirn [0] (
16:35:10 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
16:40:31 Quit lorenzo92 (Quit: ChatZilla 0.9.89 [Firefox 15.0.1/20120907231657])
16:43:58 Join kevku [0] (
17:01:29 Quit Zagor (Quit: Clint excited)
17:04:42 Quit the-kyle (Quit: Leaving.)
17:05:48 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
17:07:48 Join n17ikh [0] (~n17ikh@
17:07:49 Quit n17ikh (Changing host)
17:07:49 Join n17ikh [0] (~n17ikh@unaffiliated/n17ikh)
17:08:38 Quit kevku (Ping timeout: 260 seconds)
17:11:35n1sfreqmod: should i ask anyone in particular in #opus ? just asking the channel didn't seem to give much response
17:18:58derfn1s: What were you asking about?
17:19:18derf(I got knocked offline due to a power outage, so I must have missed it)
17:19:27n1s<n1s> Hi, i'm looking at optimizations for the opus codec for rockbox, ( some of our targets have very little memory so i disabled some encoder only code. Are you guys interested in a cleaned up patch adding some #fidef ENCODER or something to disable this code?
17:19:59derfAh, that's a jmspeex question.
17:20:42derf(I personally think it's a fine idea)
17:20:53n1sok, i'll try asking him then
17:24:13CIA-10Commit 5f60590 in rockbox by Nils Wallménius: opus: put some const tables and structs in iram
17:24:20 Quit petur (Quit: *plop*)
17:47:38 Join mortalis [0] (~mortalis@
17:48:03 Join Buschel [0] (
17:50:11***Saving seen data "./dancer.seen"
17:55:39Buscheldid anyone add opus to the related the manual sections yet?
18:09:05 Join kevku [0] (
18:15:15 Quit CIA-10 (Ping timeout: 250 seconds)
18:37:02 Quit Buschel (Ping timeout: 244 seconds)
18:49:36 Join pedro_angelo [0] (
19:15:32 Join mystica555 [0] (
19:20:31freqmodhopefully this will fix seeking at 0
19:20:51 Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
19:23:36 Join Prodicus [0] (
19:33:35 Join Buschel [0] (
19:35:37 Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)
19:37:31 Quit Buschel (Client Quit)
19:37:58 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
19:42:34 Quit benedikt93 (Quit: Bye ;))
19:43:27 Join dfkt [0] (dfkt@unaffiliated/dfkt)
19:50:12***Saving seen data "./dancer.seen"
19:51:49 Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)
19:59:07 Quit Prodicus (Ping timeout: 256 seconds)
20:00:05 Join Prodicus [0] (
20:05:52 Join scorche [0] (~scorche@rockbox/administrator/scorche)
20:08:47 Join dfkt [0] (dfkt@unaffiliated/dfkt)
20:08:57 Nick dfkt is now known as dfkt_ (dfkt@unaffiliated/dfkt)
20:09:24 Nick dfkt_ is now known as dfkt (dfkt@unaffiliated/dfkt)
20:26:52 Join liar [0] (
20:41:51 Quit eckoit (Quit: eckoit)
20:55:10 Quit sciopat (Read error: Connection reset by peer)
20:55:26 Join sciopath [0] (
20:55:48 Part LinusN
20:58:23 Join lorenzo92 [0] (
21:00:35lorenzo92bertrik: are you the RDS man? :) I'm trying to add RDS to YP-R0 target, in fact it works in various ways I tried but after a certain amount of time rockbox goes into segmentation fault!
21:02:35 Quit liar (Ping timeout: 256 seconds)
21:03:56 Join liar [0] (
21:07:31kugellorenzo92: did you see my remark?
21:09:26 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
21:11:25lorenzo92oh no...wait
21:12:10lorenzo92kugel: uhm the one about threading in RaaA?
21:15:58 Quit kevku (Ping timeout: 260 seconds)
21:33:01 Quit WalkGood ()
21:36:00 Quit mortalis (Quit: KVIrc 4.1.3 Equilibrium
21:36:43kugellorenzo92: yes
21:37:52 Quit Cthulhux (Quit: ne praeteriverit priusquam obesa cantaverit)
21:39:21 Join Cthulhux [0] (cthulhux@piratenpartei/ni/tux)
21:44:01lorenzo92kugel: okay then, I don't know much of this stuff :p perhaps is worth posting there the three ways I tried to achive rds...
21:48:28kugellorenzo92: most likely worth it yes
21:49:03 Quit liar (Ping timeout: 256 seconds)
21:49:57 Join liar [0] (
21:50:15***Saving seen data "./dancer.seen"
21:53:18 Join eckoit [0] (~ryan@
21:55:18lorenzo92in the meantime I'm back debugging usb...
22:02:08kugellorenzo92: the first one might not work
22:02:19kugelbecause of what I said
22:02:45kugelhowever it's supposed to work, there's an issue with the RaaA threading implementation
22:03:20kugelwhat about the second approach?
22:03:48 Quit y4n (Quit: 6,000,000 ways to die — choose one.)
22:06:07lorenzo92the same
22:06:13lorenzo92it crashes after a little time
22:07:22lorenzo92*little = about a minute
22:08:36kugelhm strange
22:08:46kugelyou did try to increase the stack size didnt you?
22:09:00lorenzo92ouch! right...I'll try
22:09:29lorenzo92kugel: using the second way?
22:11:02kugelhm I would expect the second approach to work
22:16:43lorenzo92pff it crashed..
22:18:34lorenzo92I open the serial now to see if segmentation fault or what
22:19:22lorenzo92btw in some weeks I will receive my ttl<->usb converter in order to use R0's hw UART interface
22:22:47kugelwhat serial interface are you talking about?
22:23:27lorenzo92so we have compiled the g_serial usb gadget via usb, but there is also the UART ;)
22:23:36lorenzo92vannix tested it, perfectly working :)
22:24:45lorenzo92kugel: ah! now seems stable uhm
22:25:00lorenzo92kugel: with cable inserted -> backlight always on
22:27:48lorenzo92well now seems working anyways
22:38:13lorenzo92kugel: yep seems really stable now :)
22:41:26lorenzo92kugel: crashed :(
22:42:36kugelcan you find the cause?
22:43:13kugelI once managed to obtain a backtrace with gdb
22:44:18lorenzo92no idea...
22:44:30lorenzo92strange is that
22:44:40lorenzo92for example now is running nicely
22:49:46 Quit Cthulhux (Quit: ne praeteriverit priusquam obesa cantaverit)
22:50:19lorenzo92kugel: I'm pretty sure there is something to investigate in drivers/rds.c
22:51:18lorenzo92look at strcpy, should we do the same trick?
22:53:42 Join the-kyle [0] (
22:54:50 Join Cthulhux [0] (cthulhux@piratenpartei/ni/tux)
22:54:54 Join lebellium_ [0] (
22:56:24 Quit Thra11_ (Ping timeout: 265 seconds)
22:56:38 Quit lebellium (Ping timeout: 246 seconds)
22:56:49 Nick lebellium_ is now known as lebellium (
22:57:07kugellorenzo92: did you also #undef RDS_ISR_PROCESSING ?
22:58:26kugellorenzo92: not sure what trick you mean. the process-4bytes-at-a-time one?
23:00:08lorenzo92kugel: why #undef? where? I was talking about the strcpy of get_ps / get_rt...
23:00:20lorenzo92kugel: was testing now using the string copier!
23:01:16 Quit Cthulhux (Quit: ne praeteriverit priusquam obesa cantaverit)
23:02:33kugellorenzo92: still dont understand what trick you mean
23:02:50kugelcopy the data before passing to rds_process()?
23:03:02lorenzo92nono I was talking about rds.c
23:03:10lorenzo92precisely, /* Need triple buffer so string isn't clobbered while caller is using it */
23:03:11 Quit liar (Remote host closed the connection)
23:03:47kugeli see that. where do you want to apply that?
23:04:29lorenzo92nevermind, it has crashed...okay so back to the @ndef
23:04:40lorenzo92why and where?
23:05:24kugeli wondered if you removed #define RDS_ISR_PROCESSING when testing the second approach
23:05:46kugelshouldnt make a difference actually
23:07:56lorenzo92ah yes yes, RDS_ISR_PROCESSING needs more functions too
23:08:11lorenzo92btw I did not enable it even when using the first approach
23:08:44 Join Cthulhux [0] (cthulhux@piratenpartei/ni/tux)
23:09:02kugelthat's right
23:09:37lorenzo92no idea how to debug the thing :(
23:09:48kugelhm wait
23:10:05kugelusing your first approach you use linux semaphores, not rockbox semaphores
23:10:22kugelah, nevermind
23:11:58lorenzo92i've also used a work queue (when irq is triggered) in the module...but doesn't make any difference
23:12:26kugelthe problem could be in rockbox code
23:13:09lorenzo92actually it should be there indeed...
23:13:43lorenzo92but why after some time? I don't say it's a fixed amount of time, but quite close to be...
23:14:57kugelcan't tell
23:16:17lorenzo92a bug in rds.c?
23:16:24 Join sakax [0] (
23:17:28 Quit the-kyle (Ping timeout: 264 seconds)
23:18:32 Join the-kyle [0] (
23:18:49kugelunfortunately i know nothing about rds
23:19:59 Quit amayer_ (Quit: going ~/)
23:21:25lorenzo92I'm wondering if gdb can be run remotely using the serial port?
23:21:35lorenzo92(I never used gdb so far ^^)
23:23:07lorenzo92kugel: do you have a compiled arm gdb?
23:25:43kugelsomewhere, yes
23:25:46 Quit FOAD (Quit: I'll be back)
23:25:58 Join FOAD [0] (~foad@
23:26:07lorenzo92well I can always compile it :)
23:26:17kugellorenzo92: you should be able to run gdb via serial, but I don't know how
23:26:24kugeldo you have shell access?
23:26:56lorenzo92via usb for now
23:27:14kugelthats great
23:33:51kugellorenzo92: one moment
23:39:01lorenzo92okay, starting the serial, one second ;)
23:40:16lorenzo92I'm in ;)
23:40:33lorenzo92so...what to do now? :p
23:41:57kugelrun rockbox
23:42:44lorenzo92running it...
23:42:57kugelwait until it crashes
23:43:59kugelyou probably need a rockbox binary with debug symbols (I think ../tools/configure > advanced > debug should do it)
23:44:17kugelonce it crashes you can get the backtrace with "bt"
23:50:18***Saving seen data "./dancer.seen"
23:53:15 Quit bertrik (Ping timeout: 244 seconds)
23:58:37 Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)

Previous day | Next day