| 00:00:00 | | Join kushal_12_27_200 [0] (n=kushal@12.169.180.134) |
| 00:00:35 | advcomp2019 | Narc4746751, i think when the sansa in manufacturer mode the operating systems reads it as something else but i do not remember right now |
| 00:00:58 | scorche | the site will start off clean slate as originally intended...i dont feel like editing hundreds of text files to make them have the correct info for the new site |
| 00:01:03 | | Quit fyrestorm (Read error: 104 (Connection reset by peer)) |
| 00:01:05 | Narc4746751 | narc, it reads as some USB device. |
| 00:01:09 | | Join fyre^OS [0] (n=fyre@cpe-68-173-163-201.nyc.res.rr.com) |
| 00:01:17 | Narc4746751 | ok, I took off the front cover |
| 00:01:33 | Narc4746751 | and on haxx.se there is a picture of the board |
| 00:01:38 | Narc4746751 | and the picture has a TI L2 chip |
| 00:01:43 | Narc4746751 | which is missing on my player |
| 00:01:58 | | Join cbr|w [0] (n=cbr@212.98.160.130) |
| 00:02:16 | scorche | Narc4746751: please express yourself in complete thoughts on a single line...it makes things much easier to read than splitting it up liek that |
| 00:02:30 | Narc4746751 | Will do, scorche. |
| 00:02:59 | advcomp2019 | Narc4746751, do you know if it is v1 or v2.. i can find a photo of a v2 from the forums if needed |
| 00:03:01 | Narc4746751 | I've removed the front cover on my e260 and looked at the circuit board. On haxx.se there is a picture of a TI L2 chip on the front of the board. Mine seems to be missing. |
| 00:03:05 | Narc4746751 | it is a v1 |
| 00:03:51 | Narc4746751 | lol |
| 00:04:01 | * | Narc4746751 puts a big greasy fingerprint right on the LCD |
| 00:04:37 | advcomp2019 | look at these photos and see if they look like yours.. http://forums.rockbox.org/index.php?topic=14064.msg114410#msg114410 |
| 00:05:43 | Narc4746751 | advcomp2019, the front board does not match. I don't think it's a v2. |
| 00:06:45 | advcomp2019 | o ok.. it was just an idea |
| 00:06:54 | Narc4746751 | it's OK |
| 00:07:11 | Narc4746751 | I wonder what that TI chip does |
| 00:07:28 | Narc4746751 | it's U14 on the circuit board |
| 00:08:17 | mud-rb_ | could it be the chip for the radio? that's the only thing i know of that some v1 have that others don't |
| 00:09:38 | Narc4746751 | mud-rb_, apparently it's a CMOS driver |
| 00:09:50 | Narc4746751 | http://www.digchip.com/datasheets/parts/datasheet/477/SN74LVC244A.php |
| 00:10:22 | Narc4746751 | it doesn't show any signs of having been removed either |
| 00:11:05 | scorche | Bagder: isnt it logical that the M3 be placed before the M5 and X5 in the builds page? |
| 00:11:25 | Bagder | yes I guess |
| 00:13:18 | | Join jrsharp [0] (n=jrsharp@c-68-52-226-233.hsd1.tn.comcast.net) |
| 00:13:42 | | Join nub [0] (n=drumsocc@cpe-24-169-134-13.rochester.res.rr.com) |
| 00:14:44 | jrsharp | hey all... my father-in-law is blind and I've recently started considering getting him an mp3 player of some kind and installing the rockbox firmware with speech stuff... my question is, is there a particular supported device that is recommended? |
| 00:15:35 | nub | anyone know how to get a vid onto a ipod 1st gen? |
| 00:15:41 | Bagder | jrsharp: I'd consider asking on the mailing list, where most of our blind users seem to hang out |
| 00:16:13 | Bagder | I'd guess that the ones with actual real buttons are preferred |
| 00:16:34 | jrsharp | over, say, an iPod with a click wheel? |
| 00:16:36 | * | domonok1 recomends jrsharp a device with actual buttons instead of a slidewheel or alike.. |
| 00:16:43 | Bagder | jrsharp: right |
| 00:17:08 | jrsharp | ok... so a 1st gen iPod would count, right? they didn't have the click wheel, right? |
| 00:17:21 | DerPapst | yes. |
| 00:17:43 | jrsharp | ok, cool |
| 00:17:51 | domonok1 | but 1gen ipod isnt really stabel with rockbox till now, am i right ? |
| 00:17:52 | jrsharp | of course, I'm not stuck on an iPod... |
| 00:17:54 | amiconn | Llorean, scorche: It might be useful to separate the M3 themes, as the M3 allows to choose e.g. the font freely, while the M5/X5 remotes are coupled to the main font |
| 00:18:25 | jrsharp | domonok1: oh? I figured 1st gen ought to be well supported... |
| 00:18:46 | domonok1 | http://www.rockbox.org/twiki/bin/view/Main/TargetStatus :-) |
| 00:18:48 | scorche | amiconn: yes...i already have a separate section for the M3 and am not planning on tying in the RWPSs to the M3 section |
| 00:18:50 | | Quit CyBergRind|w (Read error: 110 (Connection timed out)) |
| 00:18:50 | DerPapst | tough i wouldn't recommend an iPod at all, though i can use mine blindy very well since it has a case and the clickwheel is now 1mm below the average case heigh |
| 00:18:53 | BigBambi | jrsharp: You want a non-portalplayer in my opinion |
| 00:19:00 | amiconn | 1st/2nd Gen should be stable now, but the lack of suspend isn't nice |
| 00:19:13 | Llorean | amiconn: The interesting thing is, though, that any font chosen for the "main" font, would still have to work with the .rwps file on M5/X5 anyway, so they'd be valid. The only problem is I imagine it'd load the .wps instead of the .rwps for the remote screen on the M3, so that's the real problem (in my view) |
| 00:19:32 | jrsharp | domonok1: ahh... I see that now |
| 00:19:33 | Llorean | If it loaded .rwps, it wouldn't look as good as it could, but it'd still be designed for that screen / font combo |
| 00:20:06 | nub | anyone: do you guys know who to put a video onto a 1st gen ipod nano. i am confused by the directions given |
| 00:20:14 | BigBambi | iriver H100, H300, iaudio X5, M5, (M3?), or for most stability but MP3 only one of the Archos' |
| 00:20:16 | scorche | theme makers can just submit their own for the M3 and if a person wants (and knows about it), they can take the RWPS out of the M5/X5 sections...i could put a notice at teh top of the M5 portion as well |
| 00:20:19 | BigBambi | nub: Cpy and paste? |
| 00:20:23 | BigBambi | *copy |
| 00:20:24 | domonok1 | jrsharp: you may also want to look at: http://www.rockbox.org/twiki/bin/view/Main/TargetStatus?topic=BuyersGuide |
| 00:20:29 | mud-rb_ | nub: which part specifically? |
| 00:20:49 | jrsharp | DerPapst: why a non-portalplayer? |
| 00:20:57 | amiconn | I think a 1st..3rd Gen ipod would be better for a blind user than the newer ipods, because these older ones don't have the buttons integrated into the wheel (i.e. no clickwheel, but a separate touchwheel (2nd and 3rd Gen) or a mechanical scrollwheel (1st Gen) respectively |
| 00:21:00 | jrsharp | domonok1: thanks, I'm checking that out |
| 00:21:20 | BigBambi | jrsharp: I assume that was at me, not DerPapst |
| 00:21:49 | jrsharp | BigBambi: yeah, sorry... my eyes got crossed.. >) |
| 00:22:09 | amiconn | Ah yes, there's that detection problem on early 1st Gens... |
| 00:22:11 | jrsharp | amiconn: yeah, tha's a good point |
| 00:22:25 | Llorean | scorche: Notice is probably best. "While X5/M5 themes don't directly work with the M3, the RWPS can be taken and used as an M3 WPS with small modifications to the .cfg" or some such |
| 00:22:26 | BigBambi | jrsharp: And because of battery life, stability and completeness of features. There isn't much in it now, but still |
| 00:22:38 | * | amiconn should post his test bootloader... |
| 00:22:42 | DerPapst | i wouldn't recommend a 3G at all |
| 00:22:53 | DerPapst | because the hold switch is the only real "button" |
| 00:23:19 | amiconn | DerPapst: The buttons below the lcd are touch stuff on 3rd Gen? |
| 00:23:26 | DerPapst | yes |
| 00:23:33 | jrsharp | so do the voice features of rockbox work equally well on all ports, then |
| 00:23:34 | jrsharp | ? |
| 00:24:03 | amiconn | Oh, so that's also bad. 1st/2nd Gen should have better battery runtime than 3rd Gen (purely because of the higher battery capacity) |
| 00:24:05 | domonok1 | yes, only the old archos are a bit more limited.. |
| 00:24:21 | amiconn | More limited, but stable..... |
| 00:25:14 | jrsharp | ok |
| 00:25:26 | DerPapst | i thought the 3G has a better battery |
| 00:25:29 | amiconn | The 1st/2nd Gen buttons are mechanical, and they protrude from the case |
| 00:25:57 | amiconn | DerPapst: 3rd Gen battery capacity is around half that of the 1st/2nd Gen - with the same CPU... |
| 00:26:12 | amiconn | The drawback of the 1st/2nd Gen is that they're firewire only |
| 00:26:15 | DerPapst | oh... |
| 00:26:16 | * | domonok1 still need someone to tell him why this talkfile: http://www.retrospektiwe.de/The%20Last%20Supper.mp3.talk is cut of in rockbox, but decodes fine with rbspeexdec .. |
| 00:26:46 | amiconn | 1230mAh (1st/2nd) vs. 630mA (3rd) |
| 00:26:49 | * | jhMikeS still wonders why any would be at all |
| 00:27:05 | nub | ok dude im confused say i have a .mp4 vid |
| 00:27:12 | nub | i cant just put that into the ipod right |
| 00:27:45 | domonok1 | nub, yes you have to convert, take a look at the wikipage of mpegplayer |
| 00:27:47 | mud-rb_ | nub: no, you can't. you need to transcode it to a different codec. |
| 00:27:58 | nub | to mpeg correct |
| 00:28:02 | nub | mpeg-2 |
| 00:28:04 | nub | right |
| 00:28:13 | domonok1 | to mpeg and the right screensize.. |
| 00:28:14 | mud-rb_ | correct |
| 00:28:18 | BigBambi | or 1 |
| 00:29:49 | domonok1 | is there some header which could be wrong in talkfiles ? or are they completly raw ? |
| 00:30:12 | jrsharp | thanks everybody... you've been very helpful |
| 00:31:59 | jhMikeS | domonok1: raw speex |
| 00:32:26 | nub | after converted |
| 00:32:31 | nub | copy and paste? |
| 00:32:50 | domonok1 | jhMikeS: any idea how i could debug this problem ? ( the talkfile was generated by rbutil, and i want find the bug) :-) |
| 00:33:49 | jhMikeS | make sure talk.c is sending everything. the voice thread should also be decoding everything until libspeex indicates the clip has ended. |
| 00:34:51 | jhMikeS | I take it this depends on if music is running or not? |
| 00:35:10 | domonok1 | i only tried it without music.. |
| 00:35:15 | Narc4746751 | What's the easiest way to remove lint and fibers from the inside of a player's case? |
| 00:35:23 | Narc4746751 | i have access to a cleanroom if necessary |
| 00:35:36 | domonok1 | its also cut off in the sim... but a decoded wav file with rbspeexdec is fine with it.. |
| 00:36:08 | jhMikeS | I'd try with music. that may point to the pcm buffer not playing every frame |
| 00:36:39 | Narc4746751 | Maybe I'll try a can of duster. |
| 00:36:51 | domonok1 | on target its fine with musik.. |
| 00:37:37 | domonok1 | so it probably just stops playback to early ? |
| 00:37:39 | | Quit phinze () |
| 00:38:45 | jhMikeS | that indicates talk.c and voice_thread.c are doing it right but makes me suspect the pcmbuf.c fails to play all the inserted data |
| 00:41:08 | | Quit nub () |
| 00:41:10 | * | domonok1 is a bit lost in this code .. :-/ |
| 00:43:13 | | Join BumAz [0] (n=drumsocc@cpe-24-169-134-13.rochester.res.rr.com) |
| 00:43:25 | | Quit m0f0x () |
| 00:43:42 | BumAz | untagged tracksss? how can i change that |
| 00:43:49 | scorche | tag them |
| 00:44:01 | BumAz | how |
| 00:44:31 | scorche | by using a tag editor like mp3tag, the godfather, etc (assuming you are using windows and mp3s) |
| 00:44:59 | | Join mokkurkalve [0] (n=eivind@062016139251.customer.alfanett.no) |
| 00:45:35 | gevaerts | Narc4746751: just reading the backlog. Did you run lsusb and e200tool as root ? |
| 00:45:42 | BumAz | scorche: why does it do that |
| 00:46:03 | scorche | "it" doesnt...you didnt tag them or wherever you got them from didnt |
| 00:46:40 | BumAz | scorche: pretty gayy if u ask me |
| 00:46:45 | scorche | ... |
| 00:46:54 | Narc4746751 | gevaerts, yes |
| 00:47:04 | scorche | this is *your* fault here for not tagging your files... |
| 00:47:07 | gevaerts | ok. that's not the problem then... |
| 00:47:09 | Narc4746751 | gevaerts, i opened the cabinet and found a missing IC. |
| 00:47:22 | BumAz | scorche, well its itunes fault |
| 00:47:29 | scorche | then whine to them |
| 00:47:56 | | Quit jrsharp () |
| 00:48:06 | Narc4746751 | gevaerts, the TI LC24 to be exact |
| 00:49:53 | * | amiconn now has a bunch of rather interesting curves |
| 00:50:17 | * | Narc4746751 puts on sunglasses and gazes at amiconn sideways |
| 00:50:41 | | Quit saratoga ("CGI:IRC (Ping timeout)") |
| 00:50:55 | BumAz | i will |
| 00:51:51 | BumAz | ok scorche, can u use other programs to dl music too then? |
| 00:51:59 | scorche | that is offtopic here |
| 00:52:19 | BumAz | o ok |
| 00:53:44 | | Join phinze [0] (n=phinze@pcp027324pcs.jesres.mu.edu) |
| 00:56:45 | gevaerts | Could some gsoc mentors have a look at the latest comment in Corey Shaffer's usability application ? I think someone with more rockbox-knowledge than me should reply to that |
| 00:58:37 | | Quit Rincewind ("bye") |
| 00:59:34 | scorche | I commented...there are quite a few other things about his application I think needs to be discussed besides that, so it would just be better in IRC |
| 01:00 |
| 01:00:14 | | Quit Chipsaru (Read error: 104 (Connection reset by peer)) |
| 01:00:20 | gevaerts | great |
| 01:00:34 | | Quit phinze () |
| 01:03:57 | *** | Saving seen data "./dancer.seen" |
| 01:06:25 | linuxstb | scorche: Just a thought - if we have discussions with students in IRC, one of us should try and remember to post a link to the logs as a comment attached to the application, so mentors that miss the chat in IRC know it existed. |
| 01:06:43 | linuxstb | (or even the student could...) |
| 01:06:52 | | Join GPT [0] (n=GPT@unaffiliated/gpt) |
| 01:07:34 | scorche | linuxstb: that is a good idea, however there are some folks who are in here routinely and we would end up with a hundred or so links... |
| 01:08:03 | linuxstb | True... |
| 01:08:39 | | Quit ender` (" Today's BOFH excuse: broadcast packets on wrong frequency") |
| 01:09:00 | scorche | there are also discussions about certain ideas that may not have been about the student's proposal specifically |
| 01:09:14 | | Quit tessarakt ("Client exiting") |
| 01:09:19 | | Join tessarakt [0] (n=jens@e180066037.adsl.alicedsl.de) |
| 01:16:12 | | Quit Narc4746751 (Read error: 110 (Connection timed out)) |
| 01:19:05 | | Quit midkay ("Leaving") |
| 01:20:27 | | Join FOAD_ [0] (n=dok@dinah.blub.net) |
| 01:21:32 | | Join midkay [0] (n=midkay@rockbox/developer/midkay) |
| 01:23:37 | | Quit kugel (Remote closed the connection) |
| 01:23:51 | | Join kugel [0] (n=kugel@unaffiliated/kugel) |
| 01:27:45 | | Quit gevaerts ("sleeping time") |
| 01:29:07 | | Quit Breakage ("Leaving") |
| 01:35:04 | | Quit amiconn (Nick collision from services.) |
| 01:35:11 | | Join amiconn [50] (n=jens@rockbox/developer/amiconn) |
| 01:37:18 | | Quit FOAD (Read error: 110 (Connection timed out)) |
| 01:37:19 | | Nick FOAD_ is now known as FOAD (n=dok@dinah.blub.net) |
| 01:37:23 | | Quit roolku () |
| 01:37:38 | | Join mud_rb__ [0] (n=mud-rb@dialup-4.156.9.62.Dial1.Boston1.Level3.net) |
| 01:39:04 | | Join Pio__ [0] (n=sean@64.251.10.106) |
| 01:46:32 | | Quit domonok1 (Read error: 104 (Connection reset by peer)) |
| 01:48:39 | | Join SirFunk [0] (n=Sir@206-159-155-246.netsync.net) |
| 01:50:18 | | Quit moos (Read error: 110 (Connection timed out)) |
| 01:50:31 | | Quit Pio (Connection timed out) |
| 01:50:43 | | Join Chronon [0] (n=knoppix@c-71-236-147-148.hsd1.or.comcast.net) |
| 01:56:09 | | Quit mud-rb_ (Read error: 110 (Connection timed out)) |
| 03:00 |
| 03:00:51 | | Quit n1s () |
| 03:02:48 | | Quit tessarakt (Read error: 110 (Connection timed out)) |
| 03:03:18 | | Join tessarakt [0] (i=nobody@vpn-cl-160-120.rz.uni-karlsruhe.de) |
| 03:04:15 | | Part kslater |
| 03:07:00 | | Quit tessarakt (Client Quit) |
| 03:10:52 | kugel | what's the best way to prepare the source code to be distributed without all the .svn folders? |
| 03:11:26 | | Quit tvelocity (Remote closed the connection) |
| 03:11:39 | DerPapst | kugel: look how it is done here: http://pastebin.ca/962361 |
| 03:12:14 | | Join psycho_maniac [0] (i=psycho_m@ppp024.hk.centurytel.net) |
| 03:13:36 | | Nick fxb__ is now known as fxb (n=felixbru@h1252615.stratoserver.net) |
| 03:13:45 | psycho_maniac | congrats on getting into the 2008 GSOC |
| 03:13:46 | kugel | I don't quite get the script. what will it exactly do? |
| 03:15:11 | kugel | DerPapst: ^ |
| 03:16:11 | kugel | DerPapst: I don't see where the .svn folders are taken out |
| 03:16:14 | DerPapst | it will copy the entire svn tree to a folder and 7zip it up. |
| 03:16:48 | kugel | with the .svn folders? |
| 03:16:55 | kugel | I do not want them |
| 03:16:55 | DerPapst | kugel: it doesn't. svn ls -R doesn't include them |
| 03:18:26 | DerPapst | the line including the svn ls -R part copies the entire source tree to another directory skipping the .svn folders. |
| 03:19:35 | kugel | ah ok |
| 03:20:06 | kugel | and if I don't want 7z but normal zip? |
| 03:23:02 | linuxstb | A simple way would just be (in your Rockbox dir) "svn ls -R | xargs zip mysource.zip" |
| 03:23:16 | DerPapst | replace the 7zr command with the zip one. |
| 03:23:38 | DerPapst | meh.. too easy |
| 03:24:22 | kugel | DerPapst: Yea, that idea came just into my mind after I asked :/ Too late allready for me |
| 03:24:22 | | Quit psycho_maniac (" HydraIRC -> http://www.hydrairc.com <- Now with extra fish!") |
| 03:24:46 | kugel | linuxstb, DerPapst: Thanks. I didn't know about "svn ls -R" |
| 03:25:45 | kugel | It seems to delete.rej and .org files as well. |
| 03:25:59 | DerPapst | s/delete/skip :-P |
| 03:26:33 | DerPapst | it skips all files not added to svn. So in order to add new source files you need to svn add them first. |
| 03:27:31 | kugel | ah ok |
| 03:27:36 | kugel | yea, I've done that |
| 03:29:29 | Nico_P | kugel: there's also svn export |
| 03:31:04 | kugel | DerPapst: The files I added aren't in the zip |
| 03:32:50 | linuxstb | Nico_P: Does that work with local changes though? |
| 03:33:29 | | Join Kage [0] (n=delphiki@ppp-88-217-76-109.dynamic.mnet-online.de) |
| 03:33:37 | DerPapst | kugel: and you did "svn add ./path/to/file.c" ? |
| 03:33:44 | kugel | DerPapst: sure |
| 03:34:04 | Nico_P | linuxstb: apparently if you don't specify the rev local changes are kept |
| 03:34:13 | DerPapst | hrmm.... don't know then |
| 03:34:19 | Nico_P | I've never tried though |
| 03:34:24 | * | DerPapst isn't a svn wizard |
| 03:35:23 | * | DerPapst tries |
| 03:36:12 | | Quit DerDome (Nick collision from services.) |
| 03:36:13 | | Join DerDom1 [0] (n=DerDome@dslb-082-083-216-218.pools.arcor-ip.net) |
| 03:36:23 | | Nick DerDom1 is now known as DerDome (n=DerDome@dslb-082-083-216-218.pools.arcor-ip.net) |
| 03:38:16 | DerPapst | indeed, it's missing |
| 03:38:45 | | Nick fxb is now known as fxb__ (n=felixbru@h1252615.stratoserver.net) |
| 03:38:50 | DerPapst | no idea why though |
| 03:40:43 | DerPapst | svn export does the trick. |
| 03:43:52 | kugel | yea |
| 03:47:31 | | Quit Kage ("return(EXIT_SUCCESS);") |
| 03:47:50 | | Quit Nico_P (Remote closed the connection) |
| 04:00 |
| 04:04:01 | *** | Saving seen data "./dancer.seen" |
| 04:04:03 | | Nick TaylorKillia1 is now known as TaylorKillian (n=Administ@d60-65-59-145.col.wideopenwest.com) |
| 04:14:42 | | Quit Thundercloud (Read error: 104 (Connection reset by peer)) |
| 04:21:59 | | Quit Rob2223 () |
| 04:26:10 | | Quit mud_rb__ (Read error: 104 (Connection reset by peer)) |
| 04:26:16 | | Quit midkay ("Leaving") |
| 04:27:14 | | Join miepchen^schlaf [0] (n=miepchen@p54BF6E90.dip.t-dialin.net) |
| 04:27:57 | | Join CaptainKewl [0] (i=jds@207-237-172-5.c3-0.nyr-ubr4.nyr.ny.cable.rcn.com) |
| 04:29:01 | | Join mud_rb__ [0] (n=mud-rb@dialup-4.156.9.62.Dial1.Boston1.Level3.net) |
| 04:42:28 | | Quit miepchen^schlaf_ (Read error: 110 (Connection timed out)) |
| 04:50:24 | | Join safetydan [0] (n=safetyda@rockbox/developer/safetydan) |
| 04:54:14 | | Quit kushal_12_27_200 ("This computer has gone to sleep") |
| 04:59:05 | | Join Rob2222 [0] (n=Miranda@p4FDCD0FC.dip.t-dialin.net) |
| 04:59:53 | | Quit DerPapst (Nick collision from services.) |
| 04:59:56 | | Join DerPapst [0] (n=DerPapst@p5B23DD3D.dip.t-dialin.net) |
| 05:00 |
| 05:05:35 | | Join kushal_12_27_200 [0] (n=kushal@12.169.180.134) |
| 05:07:16 | | Quit Chronon (Read error: 110 (Connection timed out)) |
| 05:16:28 | | Quit CaptainKewl ("( www.nnscript.de :: NoNameScript 4.02 :: www.XLhost.de )") |
| 05:30:53 | | Quit rotator () |
| 05:34:44 | | Join Dr_Pressure [0] (n=1816b83b@gateway/web/cgi-irc/labb.contactor.se/x-7624a32f3dab139e) |
| 05:37:00 | Dr_Pressure | I have a question about power supplies for an Archos. Would it be better suited in #rockbox-community or in here? |
| 05:40:00 | | Quit kugel ("ChatZilla 0.9.81 [Firefox 2.0.0.13/2008032514]") |
| 05:40:06 | | Nick JdGordon_ is now known as JdGordon (n=jonno@rockbox/developer/JdGordon) |
| 05:42:28 | | Join phinze [0] (n=phinze@pcp027324pcs.jesres.mu.edu) |
| 05:49:49 | | Quit phinze () |
| 05:51:58 | | Join phinze [0] (n=phinze@pcp027324pcs.jesres.mu.edu) |
| 06:00 |
| 06:02:26 | | Join jjjj [0] (n=ttrrr@pool-71-118-171-95.lsanca.dsl-w.verizon.net) |
| 06:04:05 | *** | Saving seen data "./dancer.seen" |
| 06:08:36 | | Quit phinze () |
| 06:10:08 | | Join sarixe [0] (n=sarixe@68.239.199.99) |
| 06:14:56 | | Join piga [0] (n=leonardo@200-161-96-59.dsl.telesp.net.br) |
| 06:16:48 | | Quit jjjj ("—I-n-v-i-s-i-o-n— 2.0 Build 3515") |
| 06:25:45 | | Quit ctaylorr (Read error: 110 (Connection timed out)) |
| 06:30:39 | | Join BHSPitLappy [0] (n=stephen@adsl-68-88-140-226.dsl.rcsntx.swbell.net) |
| 06:32:19 | | Join mud-rb_ [0] (n=mud-rb@dialup-4.156.9.58.Dial1.Boston1.Level3.net) |
| 06:44:31 | | Join goffa_ [0] (n=goffa@216.220.23.105) |
| 06:45:40 | | Quit mud_rb__ (Connection timed out) |
| 06:47:10 | | Quit piga (Remote closed the connection) |
| 06:48:34 | | Quit goffa (Read error: 110 (Connection timed out)) |
| 06:51:53 | | Join InHisName [0] (n=Administ@c-71-225-221-149.hsd1.pa.comcast.net) |
| 06:54:37 | | Quit Dr_Pressure ("CGI:IRC (EOF)") |
| 07:00 |
| 07:00:52 | | Join aliask [0] (n=aliask@rockbox/developer/aliask) |
| 07:00:57 | | Join phinze [0] (n=phinze@pcp027324pcs.jesres.mu.edu) |
| 07:09:42 | | Quit HellDragon (Client Quit) |
| 07:11:23 | | Join ranjith [0] (n=sailesh@202.83.54.56) |
| 07:11:53 | ranjith | hi everyone |
| 07:12:40 | DerPapst | lo ranjith |
| 07:15:55 | | Quit mud-rb_ (Remote closed the connection) |
| 07:17:45 | | Join hd [0] (i=jd@unaffiliated/helldragon) |
| 07:27:06 | | Join ranjith_ [0] (n=sailesh@202.83.54.56) |
| 07:27:08 | | Quit ranjith (Read error: 104 (Connection reset by peer)) |
| 07:31:21 | | Join ranjith__ [0] (n=sailesh@202.83.54.56) |
| 07:31:25 | | Quit ranjith_ (Read error: 104 (Connection reset by peer)) |
| 07:31:58 | | Join MagicFab [0] (n=magicfab@ubuntu/member/magicfab) |
| 07:33:27 | | Quit hd (Client Quit) |
| 07:36:25 | | Quit aliask ("ChatZilla 0.9.81 [Firefox 2.0.0.12/2008020710]") |
| 07:39:08 | | Quit Kanos (Read error: 104 (Connection reset by peer)) |
| 07:41:40 | | Join tedrock [0] (n=tedrock@d235-159-75.home1.cgocable.net) |
| 07:46:04 | | Join jhulst_ [0] (n=jhulst@unaffiliated/jhulst) |
| 07:54:47 | | Join hd [0] (i=jd@unaffiliated/helldragon) |
| 07:55:39 | | Quit phinze () |
| 08:00 |
| 08:04:07 | *** | Saving seen data "./dancer.seen" |
| 08:09:30 | | Quit GPT () |
| 08:14:46 | | Part toffe82 |
| 08:17:31 | * | amiconn would think 'lightweight' would be for slow processors, not fast ones... |
| 08:18:50 | | Quit kushal_12_27_200 (Read error: 104 (Connection reset by peer)) |
| 08:22:23 | DerPapst | morning amiconn :-) |
| 08:22:39 | | Join Rob2223 [0] (n=Miranda@p4FDCEB03.dip.t-dialin.net) |
| 08:23:07 | amiconn | mo0ning |
| 08:23:10 | | Join kushal_12_27_200 [0] (n=kushal@12.169.180.134) |
| 08:27:01 | | Quit ranjith__ (Read error: 110 (Connection timed out)) |
| 08:28:21 | | Join Zarggg_ [0] (n=z@216-15-73-111.c3-0.eas-ubr6.atw-eas.pa.cable.rcn.com) |
| 08:32:16 | | Quit sarixe ("(EE) Failed to load "quit" module") |
| 08:37:46 | | Quit Zarggg (Read error: 113 (No route to host)) |
| 08:39:13 | | Quit Rob2222 (Read error: 110 (Connection timed out)) |
| 08:42:04 | | Join davina [0] (n=davina@cpc1-sout6-0-0-cust616.sotn.cable.ntl.com) |
| 08:52:48 | | Quit kushal_12_27_200 (Read error: 104 (Connection reset by peer)) |
| 08:53:52 | | Join RedDraken [0] (n=RedDrake@adsl-69-149-214-192.dsl.wchtks.swbell.net) |
| 08:59:45 | jhMikeS | not at all. the point is to have drivers that keep a fast processor doing nothing as much as possible and let peripherals do the work instead instead of cycle buring in loops. |
| 09:00 |
| 09:01:27 | | Quit RedDraken ("KVIrc 3.2.4 Anomalies http://www.kvirc.net/") |
| 09:02:05 | amiconn | You mean just let the hardware do something and sleep meanwhile *without* yielding? |
| 09:03:03 | jhMikeS | yield by just sleeping the calling thread which lasts for the entire duration of whatever is being done |
| 09:03:33 | | Join kushal_12_27_200 [0] (n=kushal@12.169.180.134) |
| 09:04:09 | amiconn | Why not a plain yield() in this case? |
| 09:04:34 | jhMikeS | then the processor can't enter a sleep state |
| 09:04:55 | amiconn | hmm |
| 09:06:00 | amiconn | So this is actually almost the same as sleep(0) if I'm not missing something |
| 09:06:10 | jhMikeS | no |
| 09:06:50 | jhMikeS | it's not tick based at all, it's immediate wakeup of the thread when the interrupt signals it |
| 09:07:21 | amiconn | Yes, that's why I said 'almost'Ä |
| 09:07:24 | amiconn | -Ä |
| 09:07:44 | jhMikeS | all explicit wakeups are immediate so they can happen faster and with far less latency than the tick |
| 09:07:52 | amiconn | sleep(0) sleeps until the end of the tick. This one just sleeps until the next interrupt |
| 09:09:51 | amiconn | Well, my first thought might actually be useful on all targets: a micro-sleep *without* yielding, instead of those really short delay loops. Requires the respective driver code to use interrupts though |
| 09:09:53 | jhMikeS | the thread isn't nescessary put pack into the run list on the next interrupt, many interrupts could happen before the handler wakes it. |
| 09:11:03 | amiconn | But that idea doesn't need support in the scheduler, and is better inlinied into the driver code |
| 09:11:26 | jhMikeS | micro sleep? like udelay? :) |
| 09:11:55 | amiconn | Yes and no. udelay() uses busy-looping |
| 09:14:04 | amiconn | Hmm, the PP502x has that suspend-for-n microseconds feature, which could (should?) be used for udealy... |
| 09:14:09 | amiconn | *udelay |
| 09:14:47 | jhMikeS | I don't think it's sensitive to interrupts though even if bit 31 is set |
| 09:15:33 | amiconn | What was the limit, again? There are some udelay calls waiting a quite high number of microseconds |
| 09:15:52 | jhMikeS | 256 of any unit |
| 09:16:11 | jhMikeS | 255 |
| 09:16:37 | amiconn | Hmm, so the wait has to be split anyway |
| 09:17:51 | jhMikeS | where are the big udelays used for anything other than init code (and the e200 radio driver)? |
| 09:18:20 | amiconn | I only know some inits (e.g. lcd) |
| 09:18:31 | amiconn | Those should better use sleep() |
| 09:18:51 | amiconn | But then we need a working sleep() for bootloaders |
| 09:19:01 | jhMikeS | yep |
| 09:20:27 | amiconn | The udelay() could use the core suspend feature with just a tiny bit more code: |
| 09:21:17 | jhMikeS | It has to be short enough to allow FIQ servicing though |
| 09:21:18 | amiconn | Instead of the empty while(){} body, make it suspend for MAX(stop-USEC_TIMER, 255) |
| 09:21:42 | amiconn | Erm, MIN() of course |
| 09:23:00 | amiconn | How deep is the sample fifo? |
| 09:24:03 | jhMikeS | 16 |
| 09:24:24 | jhMikeS | it's allow to drop to 4 before being refilled |
| 09:26:52 | amiconn | So 12 samples. Does the PP hardware support 88200Hz sample rate? |
| 09:28:01 | jhMikeS | It's supports a huge sample clock range and the wm codecs can support 96KHz I think |
| 09:28:52 | amiconn | OK. Hmm, actually the latency is 4 samples max, because the udelay() could be entered when the buffer is already quite low |
| 09:29:22 | amiconn | Still, 40 us should be safe (at 96kHz) then |
| 09:29:30 | jhMikeS | IIS master mode has a large range. A master codec determines the IIS clock otherwise |
| 09:31:33 | amiconn | At 80MHz, 80 us equals 3200 clock cycles during which the cpu could sleep before rechecking USEC_TIMER... depending on how often this is used in drivers, it save quite a bit of battery power |
| 09:31:51 | amiconn | Insert "could" where appropriate |
| 09:32:06 | amiconn | And iI meant 40 us, bah |
| 09:35:43 | jhMikeS | FIQ is allowed from IRQ mode so it should be ok even from inside an IRQ handler (I hope that's actually working) |
| 09:39:09 | jhMikeS | the scheduler has lower latencies across sleeps and context switches than 3200 clocks |
| 09:40:04 | JdGordon | jhMikeS: why is therre IF_COP() being used instead of the full #ifdef ? isnt that frowned on? |
| 09:40:41 | jhMikeS | #ifdef everywhere makes the code bloody unreadable |
| 09:41:29 | | Quit midgey () |
| 09:41:41 | jhMikeS | it's confined to one-liners only but if the code police must enforce it I'll change it |
| 09:43:22 | jhMikeS | 3200 clocks is about 3 to 4 messages received and responded to at 80MHz |
| 09:46:01 | amiconn | The scheduler might have lower latency, but the code in the schedules threads might not |
| 09:46:14 | amiconn | *scheduled |
| 09:47:39 | amiconn | The 3200 cycles are for 40 us, and some driver code relies on udelay() being precise enough |
| 09:48:09 | amiconn | There's at least one I know of - the brightness setting code for iPod Video+Nano |
| 09:49:36 | jhMikeS | that's the point - the scheduler is pure friction code and performs no useful work. those number didn't include disabling ticks or other periodic threads like power. |
| 09:49:47 | | Quit ceclin (Read error: 104 (Connection reset by peer)) |
| 09:54:15 | amiconn | But sleeping 3200 cycles is still better than busy-looping the same time, that was my point |
| 09:54:31 | * | amiconn should probably just try it |
| 09:56:01 | jhMikeS | quite true |
| 09:56:28 | amiconn | Is udelay() called from COP? |
| 09:57:30 | jhMikeS | I'm not aware of any instances right now. |
| 09:58:59 | amiconn | Hmm, still something it should be secured against, I think |
| 09:59:15 | amiconn | Maybe it makes sense to un-inline it then |
| 09:59:45 | amiconn | Something that is meant for waiting doesn't need to be super-fast after all :) |
| 09:59:50 | jhMikeS | CURRENT_CORE is only a 3-instruction inline |
| 10:00 |
| 10:00:53 | amiconn | Yeah, but checking that every loop iteration doesn't make sense |
| 10:01:24 | jhMikeS | the curent core is always the current one |
| 10:02:06 | amiconn | Of course. Thing is, you need to use a different register for suspending depending on which core enters udelay |
| 10:02:29 | amiconn | So udelay would need to be expanded into 2 different while() loops |
| 10:02:30 | jhMikeS | use PROC_CTL(core) = nnn |
| 10:03:13 | jhMikeS | that just makes a str r0, [r1, r2, lsl #2] kind of thing |
| 10:03:14 | amiconn | Then you do check every interation... |
| 10:03:22 | amiconn | hmm |
| 10:04:08 | *** | Saving seen data "./dancer.seen" |
| 10:10:07 | * | amiconn thinks we should document all the RE'd PP details in our wiki |
| 10:11:34 | jhMikeS | another thing is I think the USEC_TIMER wraps way before 2^32-1, trying to use it for more extended measurement seemed to indicate that but I never actually ran a test |
| 10:12:31 | amiconn | Hmm, I don't think it does |
| 10:12:44 | amiconn | If it would, it would cause occasional freezes |
| 10:13:32 | jhMikeS | I tried measuring longer intervals in the e200 scrollwheel driver and it broke after about 20 minutes of inactivity |
| 10:16:03 | amiconn | Putting the USEC timer into the ports debug screen and then watch it for a while should tell. Weel, that 'while' would be quite long |
| 10:16:37 | | Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul) |
| 10:16:52 | amiconn | 4295 seconds, i.e. 1:11:35 |
| 10:17:42 | jhMikeS | just have it splash or write the ticks to a file when it sees a value less than the previous? |
| 10:18:31 | amiconn | But if it wraps earlier, it could cause freezes, because then stop = USEC_TIMER + usecs could calculate a value that is never reached |
| 10:21:43 | jhMikeS | the udelay imp |
icon identifies that the person is a core developer (has commit access).