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 2009-03-11

00:02:45pixelmabluebrother: like me...
00:03:21pixelmathe same as me, I meant
00:04:32 Join bertrik_ [0] (
00:05:13 Quit bertrik (Read error: 113 (No route to host))
00:05:20kugelkkurbjun: 16kb boundary?
00:05:23 Nick bertrik_ is now known as bertrik (
00:05:47kugelI don't think I paid attention on that
00:06:13kkurbjunThat's just what I remember from someone else's work, I'm not 100% positive on that.
00:06:14kugelany idea for the size? I wasn't sure how big the ttb would be
00:06:48bertrikIIRC, the ds mentions 4096 entries
00:08:19kkurbjunit should just be 0x4000 bytes if I remember correct. I don't remember the exact terminology, but it depends on how you setup the primary and secondary entries. We only use the most basic framing for the MMU so I think we only use/setup the primary entries.
00:08:28 Join Conic_ [0] (
00:08:57kkurbjunif you use more advanced mmu features you need more space
00:08:58bertrikI just did a half-assed attempt at enabling the MMU on my clip and it still boots ...
00:09:33jhMikeSAre we under feature freeze yet? I wanted to commit the common piece of FS #9708 (ATA DMA) which adds nothing unless HAVE_ATA_DMA is defined (except for the FS #9721 which is apparently a bugfix) and probably for Gigabeat S soon (which shouldn't matter).
00:10:05rasherjhMikeS: We are
00:10:27LloreanjhMikeS: Is it a feature, or a bug fix?
00:10:58pixelmabluebrother (other Germans): wikipedia gives another idea... "Würdigung" but the article is called "Credits (Film)". I'd like to avoid Denglisch though
00:11:24jhMikeSrasher: FS #9721 is a bugfix according to dreamlayers which is included in the ATA DMA patch. Other than that, the common portion changes nothing at all.
00:11:26pixelmaerr... Credit
00:11:39 Quit __lifeless (Remote closed the connection)
00:11:55 Join __lifeless [0] (n=lifeless@
00:11:57saratogajhMikeS: if its code thats not enabled yet it should be pretty safe to commit
00:12:22rasherjhMikeS: I don't think I'm qualified to have an opinion on that. I was simply stating fact, not trying to be harsh
00:12:25 Join tim__ [0] (n=aoeu@
00:12:35LloreanjhMikeS: If it's not "new features" but rather "fixes and improvements" it's probably fine during the freeze anyway.
00:12:49jhMikeSSince gigabeat S isn't technically supported and that bit only affects i.MX31 code specifically, I suppose that's ok, right?
00:13:01rasherThere's a fine line between fixes and improvements.
00:13:01 Quit tyfoo (Read error: 104 (Connection reset by peer))
00:13:20 Quit midgey ()
00:13:33rasherI don't think improvements are appropriate during a feature freeze. But if it's all inactive code, there can hardly be any harm
00:13:47LloreanWell, the Gigabeat S is still up for debate as including as a release target, isn't it?
00:13:57kkurbjunI didn't think that unreleased targets were part of the feature freeze. I've been committing to the MR500 since the freeze.
00:14:18rasherkkurbjun: That's fine
00:14:19Lloreankkurbjun: Targets we aren't releasing aren't part of the freeze. Targets we _may_ release are.
00:14:29 Quit tvelocity (Remote closed the connection)
00:14:57kkurbjunok, that makes sense. (and what I was running under :) )
00:14:59 Quit Conic (Read error: 60 (Operation timed out))
00:15:15jhMikeSok, we still _may_ release on Gigabeat S?
00:15:35rasherI think so
00:15:58LloreanjhMikeS: It's almost entirely down to installation now, I believe.
00:16:04LloreanWhether we can make an installer.
00:16:14 Quit bmbl ("Woah!")
00:16:21LloreanOr rather, get the installer working in Windows, I guess.
00:16:36 Join Conic [0] (
00:17:42 Quit Conic_ (Read error: 60 (Operation timed out))
00:18:32linuxstbWell, there hasn't been much progress on any of the issues - i.e. there's no bootloader released yet, and don't we have a problem with the single-boot bootloader on v1.2 beasts?
00:18:34 Join moos [0] (i=Mustapha@rockbox/staff/moos)
00:18:48linuxstbThe beast hasn't even reached "supported" status yet...
00:19:10rasherNot looking good for the Gigabeat S
00:19:20Lloreanlinuxstb: Well "supported" and "released" have pretty much the same requirements.
00:20:04linuxstbYes, although "release" should come a period of time after "supported" - to make sure it's ready.
00:20:38 Quit timc`` (Connection timed out)
00:20:39linuxstbAlthough the beast is quite special, as I doubt "supported" status will suddenly bring lots of new users.
00:20:44LloreanI'm not sure I see that strong a reason. Especially when it's a case like this where it's been widely used, and we're waiting for the tools, not the build, to be ready for a release
00:21:38linuxstbThere's no point debating until "supported" status is reached anyway...
00:21:43n1sI think we should not include it in the release and just make it "supported" once the installation is worked out
00:21:52jhMikeSI guess I won't add ATA code to support DMA at this point, active or not, unless at least one target uses it though.
00:22:19LloreanWe may as well just say "The Beast is not being released in 3.2" right now, and not even worry about it.
00:22:51*jhMikeS wonders if anyone at all ever tested the bootloader padding patch :)
00:22:55rasherLlorean: +1
00:23:11linuxstbjhMikeS: There's a bootloader padding patch? ;)
00:23:11jhMikeS* someone that actually has the 1.2 issue of course
00:23:24jhMikeSnot in FS but on my ftp
00:23:46*linuxstb stops searching flyspray...
00:23:53jhMikeSit makes it about the size of the retailos just to see if there's anything to the idea
00:24:16linuxstbIs that a patch to beastpatcher/mknkboot or do you just build a huge bootloader.bin?
00:24:18LloreanjhMikeS: I would be perfectly willing to attempt to develop the 1.2 issue if you could tell me how to get my computer to see my Gigabeast in MTP mode. :)
00:24:33 Quit amiconn (Nick collision from services.)
00:24:34 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
00:24:42 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
00:24:42 Quit pixelma (Nick collision from services.)
00:24:52 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)">
00:24:57 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
00:25:48jhMikeSLlorean: did you ever install those libUSB drivers?
00:25:57toffe82Llorean: if you put the gigabeast in recovery mode you should be able to use the updater to version 1.2 (under windows)
00:26:04LloreanjhMikeS: What libUSB drivers?
00:26:08Lloreantoffe82: Even if the OF won't connect?
00:26:34toffe82you have problem with usb ?
00:26:57jhMikeSLlorean: If OF won't connect, it could be one of those on-board fuses got blown out
00:27:00LloreanNeither my, nor I think BigBambi's, Gigabeat S will connect in the original firmware
00:27:03LloreanWindows asks for drivers
00:27:23LloreanjhMikeS: UMS works fine in Rockbox. But windows demands drivers for it in MTP mode.
00:27:53toffe82I have the same problem of usb with one of my S , but going in recovery mode , usb is working
00:28:13LloreanjhMikeS: It's quite strange. Bigbambi's reporting the same behaviour, and some internet searching seems to indicate it's not Rockbox caused at least.
00:28:21Lloreantoffe82: Okay.
00:28:59jhMikeSLlorean: So it just started "doing it" out of nowhere?
00:29:05mcuelenaererasher: this snippet should be able to use the forum as a login backend:">
00:30:12jhMikeSLlorean: Hmmm...reinstall WMP? Does it depend upon the host it's plugged into?
00:30:16 Join planetbeing [0] (
00:30:33 Quit bertrik ("Leaving")
00:30:49perrikwpjhMikeS: I have a Gigabeat S where the single-boot bootloader fails to work
00:31:26 Quit bluebrother ("leaving")
00:31:32jhMikeSperrikwp: you can try the patch out. if that works, we at least have a direction to go.
00:32:34perrikwpjhMikeS: ok I'll try it
00:32:50jhMikeSperrikwp: thanks
00:33:19rashermcuelenaere: not bad! Would be nice if we could get the user's badges as well
00:33:40linuxstbjhMikeS: Your patch just adds 12MB of zeros to the end of bootloader.bin?
00:33:44mcuelenaereoh yes, forgot about that part :/
00:33:47 Join homielowe [0] (n=homielow@unaffiliated/homielowe)
00:33:47jhMikeSlinuxstb: yep
00:34:00linuxstbjhMikeS: OK, so that's easy to replicate in beastpatcher.
00:34:06rasherBTW is the config class a bad idea? It has the advantage over using defines that trying to change the value is a *syntax error*, whereas attempting to change a define just silently fails
00:34:24 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
00:34:32jhMikeSlinuxstb: it uses the linker to add it to bootloader.bin so the CE file is valid
00:34:37toffe82jhMikeS: we should implement an option to force usb on the S, to fix the problem of a blown fuse , If you remeber, you told me how to do a long time ago
00:34:58toffe82it will prevent some usb problem
00:34:59LloreanjhMikeS: I've tried it on two different Windows computers (vista-32 and vista-64) with the same behaviour.
00:35:17LloreanjhMikeS: I'm not actually *sure* it ever connected in MTP mode. I *want* to say it did, but I'm not certain
00:35:34jhMikeSLlorean: How'd RB ever end up installed then?
00:35:40LloreanjhMikeS: Linux computer.
00:36:07jhMikeSah, so technically it does depend on the host in a way
00:36:25toffe82you can upload the 1.2 via linux , with sendfirm , no?
00:36:50mcuelenaererasher: something like this should do:">
00:36:57linuxstbtoffe82: Does that do a full update? I thought there was also a file that gets flashed?
00:37:01LloreanjhMikeS: It depends on the host OS I guess. But BigBambi reported XP didn't work either.
00:37:06 Quit ender` (" But there, everything has its drawbacks, as the man said when his mother-in-law died, and they came down upon him for the f")
00:37:20jhMikeStoffe82: yeah, I recall. That patch wouldn't apply but could be synced.
00:37:21mcuelenaererasher: I don't think the config class is a bad idea, you're using the OO principal anyway..
00:37:30toffe82linuxstb: perhaps you are right
00:38:05toffe82jhMikeS: it would be nice, we just have to decide what would be the key to press to force it on and off
00:38:36linuxstbmcuelenaere: Did you ever look at beastpatcher again, or is your version still not compiling?
00:38:42rashermcuelenaere: it does have the disadvantage that the value must be completely static. It can't be computed (1000*1000 doesn't work, 1000000 does)
00:38:49rasherSo it makes things a bit awkward
00:38:51jhMikeStoffe82: You mean just the bootloader right? I'm not sure what would be acceptable in the firmware.
00:38:55 Join Martyn [0] (n=martinb@
00:39:04mcuelenaerelinuxstb: no, I haven't looked at it since I tried using MingW + VS
00:39:12mcuelenaere(it did compile, but it didn't run)
00:39:32toffe82jhMikeS: I think it was in the bootloader
00:40:00toffe82but it should be also in the firmware if not you cannot connect
00:40:06jhMikeStoffe82: It was, yes. I was wondering if you wanted a way to force it within the firmware.
00:40:16mcuelenaererasher: oh, I didn't know that. Isn't the PHP compiler smart enough to do const bleh = (100*10); -> const bleh = 1000; ?
00:40:16linuxstbtoffe82: What's this problem?
00:40:25Martynfat32 filesystem check works. Where should I post the availability of the plugin?
00:40:52MartynIt won't format a card, yet, but I can at least check fat32 fs and make sure it's not broken
00:40:57*jhMikeS wonders about just ordering fuses from digikey and fixing the board :)
00:40:59mcuelenaerelinuxstb: IMO I think the DLL-way seems the best option (and you could pack it in the executable if you really wanted to have one file)
00:41:11toffe82linuxstb: a fuse must be blown on one of my S and it didn't connect in the OF , but it works in recovery, because recovery force the usb to wait
00:41:15***Saving seen data "./dancer.seen"
00:41:28linuxstbtoffe82: And it doesn't connect in Rockbox either?
00:41:34toffe82jhMikeS: I didn't find the fuse ;)
00:41:59toffe82and it is easier by software, so if somebody has the same problem...
00:42:11linuxstbMartyn: Post a patch to the patch tracker.
00:42:12jhMikeStoffe82: weird, they sell the I-sense resistor and other parts
00:42:20toffe82linuxstb: I don't remebrer but I think not
00:43:22toffe82it would be nice to say , use rockbox to fix your broken S :)
00:44:26linuxstbmcuelenaere: How would you pack it in the executable? I don't care how it's done, but a single .exe would be nice...
00:44:43*rasher thinks a filesystem checker is more useful than a formatter
00:44:50jhMikeSwe can do any key combo so some hard to do by mistake keypress is possible
00:45:13linuxstbrasher: Assuming it repairs as well?
00:45:13toffe82the window key :)
00:45:15mcuelenaerelinuxstb: by using the same principal as Self-Extracting executables do?
00:45:27*mcuelenaere isn't yet sure how to do it, but he knows it is do-able :)
00:45:38linuxstb;) I thought you knew of a tool to do it...
00:46:06mcuelenaereand then you could dynamically load the DLL in the exe
00:46:29mcuelenaereit is a bit ugly, but it should work
00:46:33 Quit BigBambi ("No Ping reply in 30 seconds.")
00:46:53jhMikeStoffe82: a new key to replace that would be nice as well (a custom rockbox logo key :)
00:47:49 Join BigBambi_ [0] (
00:47:51toffe82we need a key to connect and disconnect
00:48:02linuxstbmcuelenaere: A quick google just shows methods that write the .dll to a temporary file on disk...
00:48:19mcuelenaereyeah, I was thinking something like that
00:48:22mcuelenaereusing bin2c
00:48:51mcuelenaereI'm not sure whether that bloats the executable (in RAM) much though
00:48:56Martynlinuxstb: Will do
00:49:06 Join thegeek [0] (
00:49:14jhMikeStoffe82: 3 keys or two? (back+menu+volup)?
00:49:18linuxstbmcuelenaere: I don't think that matters much - beastpatcher is tiny (and so is that dll)
00:49:36toffe82jhMikeS: and the 380
00:49:40 Join Conic_ [0] (
00:50:17jhMikeS380? ??
00:50:23toffe82jhMikeS: how do you do it with one hand while plug in
00:50:23LloreanI don't think we should have a key combination in the main build with a purpose of solving broken hardware issues.
00:50:31toffe82wrong windows ;)
00:50:55 Quit mcuelenaere ("Gnight")
00:51:16perrikwpjhMikeS: do I need to use sendfirm to upload the new nk.bin or can I just replace the current nk.bin in the TFAT partition using Rockbox USB?
00:51:19 Quit Conic_ (Client Quit)
00:51:25 Join Thundercloud [0] (
00:51:50MartynDid the current sansa build work?
00:51:56jhMikeSperrikwp: Just dump it in there with USB.
00:52:13perrikwpjhMikeS: ok thanks
00:52:18linuxstbjhMikeS: Is that a fair test? Wouldn't it be better to use sendfirm?
00:53:00n1sMartyn: you know that you can check the build table, here: right?
00:53:05jhMikeSlinuxstb: makes no difference to tell whether the bootloader actually functions
00:53:31Martynn1s : Oh! no didn't know :)
00:53:44Martynn1s : I have a script that tries to get the build every day, and I saw it failed. Thanks :)
00:54:21toffe82Llorean: do we know if this problem happened on other target ?
00:54:59Martynn1s : Ah, and I see the bright red 'ld' on the chart :) thanks, this will make my life a lot easier.
00:55:07Lloreantoffe82: It's still a hardware problem. Maybe a #define so it's easy to compile for, but we shouldn't put aside a button combination (even a complex one) to deal with issues only on damaged targets.
00:55:15Lloreanif someone *does* accidentally trigger it, they will think they have a bug.
00:55:47toffe82we can change the usb screen sprecailly for this case
00:56:08LloreanThat's even more additional code unnecessary for people with unbroken units.
00:56:25 Join midgey|web [0] (i=8dd5376f@gateway/web/ajax/
00:56:37*amiconn agrees with Llorean
00:56:57amiconnPlain rockbox shouldn't contain code for working around defective hardware
00:56:59jhMikeSit's a fairly trivial task from the bootloader but not so much in the firmware
00:57:01toffe82you say this until you will have the problem :)
00:57:04LloreanI just don't think we should have code for broken hardware compiled in by default. I see no problem with having it ifdeffed out so people with the hardware issue just need to define something to get it available.
00:57:24Lloreantoffe82: If I had the problem, I'd have no problem changing a #define in config-gigabeats.h
00:57:28perrikwpjhMikeS: didn't work, says firmware upgrade or restoration required after attempting to boot
00:57:35LloreanI already compile my GigabeatS build with a different plugin buffer size
00:57:46amiconnPeople still wanting to use such units can always make their own build, thanks to opensource
00:57:51toffe82Llorean: I am thinking about people who are not programmer
00:57:56jhMikeSperrikwp: Thanks. Interesting, I wonder just what it does expect.
00:58:21Lloreantoffe82: A #define doesn't require any programming knowledge. You can be told how to do it fairly simply. A tutorial could be made.
00:58:39 Quit Conic (Read error: 110 (Connection timed out))
00:58:55Lloreantoffe82: If you really feel dedicated to these people, you could provide an unsupported build in the forum for them, so they don't have to compile at all.
00:58:58amiconnIf we start doing such stuff, it'd be hard to tell where to stop. What about partially broken LCDs? Do we need to support adjustable margins because of those?
00:59:24perrikwpjhMikeS: the single-bootloader unpadded works for first boot but the second reboot onward it doesn't, if that helps any
00:59:25Lloreanamiconn: What about buttons that die easily (sticks), do we need to make backup control sets for players with wobbly sticks?
00:59:43linuxstbtoffe82: How common is this issue then? And what's the actual workaround?
00:59:50amiconnYeah, you'll probably find many more examples
00:59:52toffe82amiconn: it is more difficult in the case of the lcd , because you don't know how it is broken
00:59:56kkurbjunLlorean: that's already in the gigabeat fx (keypad sensativity
01:00:14Lloreankkurbjun: That's not a broken hardware workaround.
01:00:42kkurbjunyes it is, at least on mine high sensativity causes erratic button behavior
01:00:49toffe82linuxstb: I had one lwith this problem from 6 or 7 Ibought
01:00:55kkurbjunI thought that was why it was made an option
01:01:16toffe82it is not really a common problem
01:01:18kkurbjunthe problem is that the front plate can get depressed and cause problems
01:02:07Lloreankkurbjun: The high sensitivity can also just be very difficult to use for people with larger fingers trying to press center.
01:02:07 Quit planetbeing ()
01:02:11LloreanIt's like adjustable scroll speed.
01:02:17LloreanThere is no one option that _can_ suit everyone.
01:02:37toffe82I remember 2 person having usb problem on the F , but I don';t know if this ca n fix it
01:02:46jhMikeSperrikwp: does the dual-boot require booting into the OF as well first to work or is it possible to never boot into retailos?
01:02:57kkurbjun:), ok, so it's justified
01:03:29toffe82I have to go.... bye
01:03:34jhMikeSperrikwp: I'm wondering if the presence of those .edb files and others matters
01:03:38 Part toffe82
01:04:43perrikwpjhMikeS: I'm not sure I follow your question, could you rephrase it?
01:05:42jhMikeSperrikwp: Retailos creates a bunch of other files but only if you boot into it at least one. I was wondering if the presence of those files makes the difference.
01:05:50 Quit n1s ("Lmnar")
01:06:33jhMikeSBut if you dumped the bootloader in from a working install, that seems to indicate that it doesn't.
01:07:02perrikwplet me restore the OF, then try the padded single-boot loader again
01:07:18moosjhMikeS: hi, I remenber that when I encountred "File not found" problems with my beast, at 100% of cases, those OF files wasn't created, but I remenber to already pointed you this :)
01:07:54jhMikeSmoos: I've gotten that one myself many times and it's the reason the untar code is still there :)
01:08:11linuxstbmoos: Maybe the problem is that Rockbox doesn't like "virgin" TFAT partitions, but once the OF writes to it, it's OK?
01:08:23mooswhen the OF dirs/files wasn't be created, I was sure that rockbox will refuse to boot even when you advise me to create the dirs with th fles...
01:08:58*linuxstb had that problem once, and thinks he did boot into the OF (after a recovery) and that got it working again
01:09:18 Quit Reptile211 ("ChatZilla 0.9.84 [Firefox 3.0.7/2009021910]")
01:09:54mooslinuxstb: those weeks I have no more beast (hope to get it back very soon). I tried laot of time in different configurations, but I remenber that the crapyness was the "random" bit :(
01:10:35moosjhMikeS: even after untaring, rockbox said me "File not found"
01:10:57*moos thinks we need to get ride of OF totaly :)
01:11:00jhMikeSlinuxstb: I've had it happen many times where the first time a file is created, even in rockbox it can end up lost. But after the first time, it seems to not fail again.
01:12:31moosbut imagine this OF format the thing even if you don't want so
01:13:06perrikwpjhMikeS: ok reloaded OF, ran it once, used sendfirm to send padded single-bootloader, still didn't work
01:13:44moosThat happened to me one time when our USB wasn't 100% reliable, and after USB deconection, BOUM stack overflow on rb folowed by the OF formating process
01:14:04linuxstbIs it possible to downgrade a 1.2 beast back to 1.1?
01:14:27mooslinuxstb: I think so, why not?
01:14:47jhMikeSperrikwp: wats something fancy. it must check the actual nk.bin image for something beyond a valid format.
01:15:29linuxstbmoos: Where would you get the files from? I thought there was more to it than just an nk.bin (i.e. what gets flashed) ? I may be wrong though.,..
01:15:41perrikwpjhMikeS: maybe some checksum or key?
01:15:48jhMikeSWere any of these 1.2 beast previously 1.1 beasts that didn't complain about single boot?
01:15:56 Quit Wraith| (Read error: 110 (Connection timed out))
01:16:30linuxstbperrikwp: Do you have a v1.1 nk.bin you can restore? Or have you tried that?
01:16:35jhMikeSperrikwp: Perhaps it expects certain sections to be present that the single boot bootloader doesn't include?
01:16:45mooslinuxstb: my mind that is the same process with a nk.bin then the ability to recover and then get back to 1.1, but I must be wrong
01:17:05kugelamiconn: custom list would be handy for partially broken lcds
01:17:21*moos is frustrated to not have his beast currently :(
01:17:40jhMikeSit's "how to we support broken hardware" day?
01:17:56perrikwpjhMikeS: I no longer have the original v1.1 restore cd, otherwise I would have tried that
01:18:53 Quit Thundercloud (Remote closed the connection)
01:21:12linuxstbThere was a restore CD?
01:21:52moosjhMikeS: how hard it will to do the same flash rewriter as gigabeat F/X?
01:22:01 Quit jhMikeS (Read error: 104 (Connection reset by peer))
01:22:03perrikwpmine came with some v1 software cd, but I threw it out
01:22:08 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
01:22:16moosjhMikeS: how hard it will to do the same flash rewriter as gigabeat F/X?
01:22:43perrikwpI don't know if it contained restore software or not
01:23:45jhMikeSmoos: I have no idea. I don't have any means at hand to unbrick it.
01:24:23moosno BDM or something to debugg resurect it?
01:25:17moosor JTAG
01:25:30 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
01:25:38jhMikeSnone of that on hand atm
01:26:02*moos thinks that we can at least offer you one beast for experiments :)
01:26:41linuxstbperrikwp: If it was the same as mine, the CD just contained WMP10
01:26:51perrikwpoh ok
01:26:59 Join z35 [0] (
01:27:47perrikwpany idea's on where I could find a v1.1 nk.bin legally to test it out?
01:28:30*linuxstb PMs perrikwp
01:33:19 Quit MethoS- (Remote closed the connection)
01:36:38perrikwpjhMikeS: would I have to use recovery mode to downgrade 1.2 to 1.1 or just use sendfirm during a normal MTP transfer in the OF?
01:38:49jhMikeSperrikwp: I can't imagine it should matter unless the recovery mode is capable of doing something simply sending a file won't do.
01:41:29perrikwpjhMikeS: ok loaded v1.1, confirmed it booted up and changed, re-tried the padded single-bootloader, still a no-go
01:42:42 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
01:43:02jhMikeSI can't see why the nk.bin version would matter unless it would change the flash code
01:43:56perrikwpjhMikeS: maybe something was flashed during the v1.2 upgrade and simply changing the nk.bin version doesn't matter
01:44:57perrikwpjhMikeS: O well, I guess there is not much more I can do to help, I guess I will have to stick to using the slow dual-boot bootloader
01:50:52jhMikeSperrikwp: that's what I've wondered about and whether a previously well behaved beast "turned bad"
01:51:08 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
01:51:15jhMikeSbut thanks for testing this stuff out
01:52:09perrikwpyour welcome
01:58:40 Join vertic39 [0] (
01:59:06vertic39hey still someone here? I rockboxed my Ipod but now the original firmware suddenly booted - what did I wrong?
02:01:21vertic39lmao .. I did a combination ... accidently .. lol
02:03:52 Join vertic23 [0] (
02:04:34 Quit vertic39 (Read error: 104 (Connection reset by peer))
02:05:47UnhelpfuljhMikeS: i've manage to "break" single-boot on my beast by upgrading. i don't think a downgrade is possible, as we don't have an installer for the whole thing, and it's probably a pre-boot flash bit that validates the on-disk firmware
02:06:31 Quit awake_ (Read error: 104 (Connection reset by peer))
02:07:14 Quit moos ("Time to catch some sleep")
02:07:20Unhelpfulit's kind of spotty, installing a single-boot firmware might let me get through the bootloader itself a few times, but seems to trigger restore if i actually install and run rockbox
02:09:31 Quit tim__ (Read error: 104 (Connection reset by peer))
02:09:35 Join tim__ [0] (n=aoeu@
02:09:52 Quit Hadaka (Read error: 60 (Operation timed out))
02:10:44vertic23hmm I can't get rid of the original firmware
02:11:16vertic23I am doing it exactly as told... first hold on hold off ... then play + middle button...
02:11:52 Join moredhel [0] (
02:11:55Unhelpfulvertic23: if you have the newer OF boot flash, i'm not sure we know a way for you to get *rid* of the OF yet
02:12:16vertic23rockbox is installed
02:12:25vertic23but I pressed some buttons :P
02:12:35vertic23and then the original firmware bootet... accidently
02:12:59vertic23and it were... unfortunately the right combination I pressed :P
02:13:03vertic23to load original firmware
02:14:33Unhelpfulto load the OF, you simply need to have the hold switch on while booting, if you've installed a dual-boot RB bootloader.
02:15:12vertic23well I want my rockbox back ;D
02:15:16Unhelpfulyou *can* turn the beast on while the hold switch is on, by the way - it will load the OF and then go to sleep
02:15:36jhMikeSUnhelpful: I must say I almost had a psychic premonition of exactly this sort of thing happening and resolved never to upgrade it :)
02:15:37 Join faemir [0] (
02:15:52vertic23Unhelpful: so what am I doing wrong now? I don't get the point
02:16:18Unhelpfulyou'll need to get a pin or paper clip and turn the battery switch off while it's unlpugged, since you can't turn the OF "off" with the buttons
02:16:31Unhelpfulthen turn make sure the hold switch is off, and turn it on
02:17:37vertic23it's dark where is the switch ;P
02:17:40 Join tmzt_ [0] (
02:18:04 Join soap_ [0] (
02:18:08 Quit perrikwp (
02:18:08 Quit avis (
02:18:08 Quit tmzt (
02:18:08 Quit soap (
02:18:08 Quit Slasheri (
02:18:08 Quit jon-kha (
02:18:12jhMikeSUnhelpful: it can't be excessively picky since it allows the bootloader to be present. I just wonder what aspect needs to be mocked for it to pass.
02:18:19NJoinperrikwp [0] (i=18ac0c41@gateway/web/ajax/
02:18:19NJoinavis [0] (n=ident@pdpc/supporter/student/avis)
02:18:19NJointmzt [0] (
02:18:19NJoinSlasheri [0] (i=miipekk@rockbox/developer/Slasheri)
02:18:19NJoinjon-kha [0] (
02:18:35Unhelpfuli'd love to know, i want dual-boot back ;)
02:18:35 Join Naked [0] (
02:18:41 Quit tmzt (Connection reset by peer)
02:18:46 Nick Naked is now known as Hadaka (
02:18:52UnhelpfulSlasheri: i understand you're the one to pester with tagcache questions? :)
02:18:58vertic23...well so I reinstall rockbox is the fastest, right? :P
02:19:04jhMikeSUnhelpful: I thought you needed single-boot, not dual :)
02:20:34UnhelpfuljhMikeS: well, it'd be nice if the player could not get into needs-a-paperclip state with it turned off and the lock switch engaged. yes, single-boot is what i'd like to get back, but the *whole* point was to see if it was the upgrade ROM that was the trouble, and to have a guinea pig beast handy that's "broken"
02:21:12Unhelpfulvertic23: why would you need to reinstall? you only *started* the OF, unless you've gone and loaded the OF nk.bin again.
02:21:23 Quit moredhel (Read error: 60 (Operation timed out))
02:21:50vertic23well−− wahhhhh what to do?
02:22:05vertic23it's dark can't look for tools or sumthing my gf is sleeping
02:22:10vertic23and I need music ;P
02:23:17vertic23in the FAQ is written I need to reset the ipod (video 30GB) like it said on the apple page..
02:23:29 Quit miepchen^schla (Read error: 110 (Connection timed out))
02:23:51jhMikeSUnhelpful: Needs-a-paperclip state? What's this? I think we're talking about two different players here. :)
02:24:13vertic23I go and reinstall rockbox now :E
02:24:19vertic23maybe the installer is helping me...
02:24:22vertic23will it? :P
02:24:27UnhelpfuljhMikeS: the OF will start with the hold switch engaged - it immediately goes to sleep. you can't get back without a paperclip. :/
02:24:40Unhelpfulvertic23: you really just need to turn it off, as i explained.
02:24:52vertic23read above .. no tools here
02:24:53jhMikeSon a beast?
02:25:04vertic23and I don't find the hole
02:25:25UnhelpfuljhMikeS: yes. or at least, it's done so for me in the past
02:25:45vertic23c'mon.. will the installer help me?
02:27:05vertic23yes / no / maybe
02:28:15*jhMikeS votes "maybe yes, maybe no" (since he has no idea)
02:28:16 Quit arohtar (Read error: 110 (Connection timed out))
02:28:29Unhelpfulvertic23: if you load the bootloader again, the OF should reboot, probably.
02:28:36 Join itcheg [0] (i=62db4767@gateway/web/ajax/
02:29:11vertic23...load the bootloader again - in "steps" this means??
02:30:57vertic23youre name sticks to ya :P
02:32:38vertic23took 10 seconds
02:32:45vertic23to restore the bootloader with the utility
02:33:26Unhelpfulto restore the bootloader that never even needed restored. i told you what i knew would work, i can't tell you if something you guessed at will work.
02:33:56 Quit faemir (Connection timed out)
02:34:19UnhelpfuljhMikeS: yes, on a beast, i just tried it again... power off, hold switch engaged, press the power button and it boots into the OF, then sleeps.
02:34:56jhMikeSUnhelpful: And plugging the charger or USB won't kick it back up?
02:35:40Unhelpfulit's asleep, not off. if you plug it in or whatever, the OF wakes up. you have to flip the battery switch, while it's not plugged in, to turn it off
02:36:04vertic23heh thx anyways
02:36:07vertic23finally music
02:36:25Unhelpfulthe fact that this can happen while the unit is off and locked is quite possibly *more* annoying than the longer load time on dual-boot
02:36:39jhMikeSUnhelpful: what you describe sounds normal
02:37:12Llorean"Hold" *really* shouldn't be the OF trigger on the beast.
02:37:28Unhelpfuli didn't mean to imply it wasn't, only to say that it's pretty annoying
02:37:34Llorean"Hold" should probably just enter a 'charging' mode if a charger is inserted, and USB mode if USB is inserted
02:38:10jhMikeSLlorean: Unless we reverse the loading (which requires a "single boot" bootloader btw), there is no other switch avaiable without keypad port scanning.
02:38:35LloreanjhMikeS: Is there a reason we can't scan for the actual buttons?
02:38:47LloreanRight now it's very easy to plug it into a car charger, and get stuck in the OF until you're home.
02:39:08jhMikeSPerhaps not but the included code would have to have much of the code that button-imx31.c has
02:39:50LloreanI'd rather see that, than a bootloader that can allow you to accidentally make Rockbox unusable when you try to charge it.
02:39:59LloreanBut that's personal opinion
02:40:05LloreanI'm afraid to turn on the hold switch now, basically.
02:40:09 Quit gevaerts (Nick collision from services.)
02:40:21 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
02:40:45Unhelpfuli rather like the idea of the rb boot being able to chain-load the OF nk.bin, but i'm not sure how big a mess that would really be
02:41:19***Saving seen data "./dancer.seen"
02:43:05 Join Strife89 [0] (n=nds@
02:43:35 Nick tmzt_ is now known as tmzt (
02:43:52jhMikeSUnhelpful: Perhaps the bootloader embedded with the OF could actually have a section outside the normal bootloader execution. I think I just got an idea. Have that stub code call some other code on the side and run OF depending on what gets returned.
02:45:16 Quit midgey|web (" ajax IRC Client")
02:48:32 Join sbhsu_ [0] (
02:48:42JdGordonkkurbjun: yo.. nice work!
02:50:07kkurbjun:), thanks
02:50:25kkurbjunit's nice to actually hear sound from this thing
02:50:36JdGordonhow unstable is it? i dont have mine back yet so i cant play
02:51:15kkurbjunI'm not sure what is causing the problem it just resets when I try to play music.. it never plays and then when I start pressing buttons rockbox resets
02:51:29kkurbjunbut doom, rockboy and the sound test all work fine
02:52:13kkurbjunthe cube demo is messed up too, so I think something may not be initialized right when the player is started up
02:53:34 Join sbhsu__ [0] (
03:05:20kugelkkurbjun: ams sansas reboot too on mp3
03:05:31kugelat least those that don't have the codecs in iramn
03:06:12 Join midgey [0] (n=tjross@
03:06:31 Quit sbhsu (Read error: 110 (Connection timed out))
03:06:35 Quit sbhsu_ (Read error: 110 (Connection timed out))
03:06:40Strife89Speaking of Sansas, are there any notable differences in charging with Rockbox versus the OF?
03:07:25kkurbjunkugel, hmm, do you know the reason for the reboot?
03:07:41Strife89My c250 is 19 months old and the battery is showing its age.
03:07:47 Quit Nico_P (Remote closed the connection)
03:07:54Strife89Possibly prematurely.
03:08:30kugelkkurbjun: No, no idea
03:08:39LloreanStrife89: Is it charging differently in the OF than in Rockbox?
03:09:08Strife89Well, that's what I'm trying to find out, Llorean.
03:09:56LloreanStrife89: Isn't that done by charging it a few times in each and comparing the results?
03:10:04Strife89I'm wondering if it's just me and my charging habits or if Rockbox has a role in my battery's aging.
03:10:20LloreanI'd say "it's been in use for nearly two years" has a significant role in its aging.
03:10:44LloreanLithium based batteries are hardly immortal. They're expected to lose a measurable amount of their capacity each year.
03:11:06LloreanA not-insignificant amount either.
03:11:27Strife89Mainly, I've noticed that the highest the battery goes dropped from 94% to 88% over the last two weeks.
03:12:21LloreanAnd has the weather changed?
03:12:47 Quit nuonguy ("This computer has gone to sleep")
03:12:47LloreanDifferent humidity and temperatures can cause this sort of thing. As well, it may just be inaccuracies in the measurements. Has runtime changed significantly too?
03:12:55Strife89This is based on charging with Rockbox, unplugging and powering off at 100%, then reading the level at next power on.
03:13:22Strife89Llorean: Yes, it's gotten cold, then much warmer recently.
03:13:52Strife89Full runtime I have not measured.
03:13:58 Join Taylor_ [0] (
03:15:01Taylor_Hello everyone. Just a quick question - are the Sansa fuzes' supported yet? (And I only ask as a devlopment need, not because Im trying to be a pain :) )
03:15:14 Quit kugel ("ChatZilla 0.9.84 [Firefox 3.0.7/2009021910]")
03:15:15LloreanTaylor_: They would be listed on the front page if they were.
03:15:46 Quit Tristan ("changing servers")
03:15:54Taylor_I just heard some "rumblings" on the forums, so thats why I was a little mixed up. So thanks
03:15:54Strife89Llorean: I'll do a battery bench or two in the near future.
03:16:07 Join kugel [0] (i=kugel@rockbox/developer/kugel)
03:16:32 Join Tristan [0] (
03:17:31LloreanTaylor_: don't worry. We will announce it when it happens. So just have patience, and please just read the front page. Rumors in the forums are often (usually) just people misunderstanding something.
03:17:41Strife89Also, much as I hate it, I'll probably see what happens after I charge with the OF.
03:18:53Taylor_Ok thanks. And as I said, Im not trying to be a pain :P I just need one of these target mp3 players for my project
03:19:04 Quit Strife89 ("To bed.")
03:19:14LloreanTaylor_: You could always work on it. It's never going to happen if people keep sitting around hoping someone else will get the work done.
03:19:34Taylor_Well I am going to. Thats why I asked :)
03:19:53Taylor_I might get one in the near future.
03:20:09Taylor_Wanted to make sure it was "un-hackable" to work on!
03:20:31LloreanI don't even know what you mean by that.
03:21:09Taylor_Umm.. Basically I wanted to know which Sansa was __Unhackable_
03:21:16Taylor_I want to work on that
03:21:19LloreanNone of them, so far.
03:21:26LloreanIf they were unhackable, we couldn't put Rockbox on them.
03:21:41Taylor_Yes; the fuze :)
03:21:59LloreanThere's a pretty significant difference between "no Rockbox port is completed" and "unhackable"
03:22:06LloreanUnless you mean "not yet hacked fully"
03:22:42 Quit jhMikeS (Nick collision from services.)
03:22:44Taylor_just a little confused. I don't care if the fuze has been "sort of hacked" or "fully hacked"
03:22:48 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
03:22:55 Join Llorean1 [0] (
03:23:10Taylor_So im guessing code has been run on the fuze then?
03:24:10Llorean1Any and all work done should be discussed either in the forum thread or the wiki page. I personally don't know the status, I can never remember which CPU is in the Fuze and which in the View
03:24:25Llorean1But I'm *fairly* certain the Fuze is the AMS one, which would mean yes, if it is, code has been run
03:24:32 Quit Llorean1 (Client Quit)
03:25:03 Quit itcheg (" ajax IRC Client")
03:25:09Taylor_well, Im sure I will like to work on it
03:25:31 Join Llorean1 [0] (
03:25:42Llorean1Generally, "unhackable" would mean to me "cannot be hacked"
03:25:47Taylor_I have some Staples money hanging around an Im pretty sure they sell Sansa's :)
03:25:51 Quit Llorean (Nick collision from services.)
03:25:56 Nick Llorean1 is now known as Llorean (
03:26:36FlynDice2G refurb fuzes at Newegg
03:26:52FlynDicewhoops $30.00
03:28:27Taylor_Unfortunately when we dumped the contents of the chip on the 2g, it was utility flash and not the Bootrom. So we didn't get anything much useful other than more encrypted data!
03:28:49Taylor_(Im talking about the 2g nano)
03:40:36 Join midijunkie [0] (
03:56:41 Join sarixe [0] (
04:11:43 Join sbhsu [0] (
04:11:48 Join timc`` [0] (n=aoeu@
04:12:59 Quit midijunkie ("?(???~~)?")
04:14:24 Quit tim__ (Connection timed out)
04:15:09 Quit kugel ("ChatZilla 0.9.84 [Firefox 3.0.7/2009021910]")
04:17:25 Join faemir [0] (
04:17:50 Quit blkhawk (Read error: 60 (Operation timed out))
04:19:36 Quit Tristan ("changing servers")
04:21:41kkurbjunJdGordon, I just got mp3's playing!
04:21:52kkurbjunIt was just a problem with the codec linker script
04:22:14 Join Tristan [0] (
04:22:33Taylor_nice job!
04:22:50 Quit Tristan (SendQ exceeded)
04:23:35 Join Tristan [0] (
04:24:11 Quit Tristan (SendQ exceeded)
04:24:50 Join Tristan [0] (
04:25:13 Join blkhawk [0] (
04:25:21 Join AndyIL [0] (i=AndyI@
04:25:26 Quit Tristan (SendQ exceeded)
04:25:56JdGordonkkurbjun: nice
04:27:02kkurbjunI'll commit it in a bit, but it's working really well. Skipping tracks back and forth causes some sound glitches when it's skipping, but I'm pretty sure it's because the DMA transfers never stop right now
04:27:03 Join Tristan [0] (
04:27:33 Quit Taylor_ ("Leaving")
04:29:04 Quit sbhsu__ (Read error: 110 (Connection timed out))
04:31:40JdGordonkkurbjun: hows the touchscreen? ill try to work on that when i get mine
04:32:21saratogatheres a c4000 dsp core on the mrobe 500 right?
04:32:28kkurbjunIt's just using up, down, left, right and select now. It doesn't select items from lists like it used to
04:32:44kkurbjunyeah, it's a c5409 I think
04:32:58JdGordonoh, what did you decide to d about the binary file? just commit it as it?
04:33:13 Quit Tristan ("changing servers")
04:33:23kkurbjunIt's actually suppisingly easy to program for, I have to thank Cat for doing the initial legwork on it though
04:33:33kkurbjunyeah, it's a binary file that's in SVN
04:33:50saratogakkurbjun: does rockbox do anything with the DSP?
04:33:51 Join Tristan [0] (
04:33:55kkurbjunif you do a normal build from the latest checkout you won't have to compile the dsp code
04:34:02JdGordonok, greayt
04:34:09JdGordonnice work, im heading out now
04:34:43kkurbjunsaratoga: yeah, it uses it to control the DMA transfers from the SDRAM to the DSP memory and then another transfer from the DSP memory to the McBSP (IIS) interface
04:34:57saratogathis is done in assembly?
04:35:28kkurbjunno, it's written in C, TI released a version of their compiler for linux thanks to neuros
04:35:43kkurbjunit is setup so that you can use it in "open source" projects
04:36:08saratogakkurbjun: do you need the TI compiler installed to build for the mrobe then?
04:36:44 Quit AndyI (Read error: 110 (Connection timed out))
04:36:58kkurbjunno, you can build without the TI compiler, initially Cat and I were talking about integrating the DSP code build with the main build, but since we only have a linux compiler it would mess up cygwin devs and the like
04:37:21 Part Martyn
04:37:33kkurbjunso there is just a .h file that has the binary image of the dsp code
04:38:18kkurbjundsp/dsp-image.h in the tms320dm320/dsp directory under target/arm
04:38:33saratogahow well does the c compiler work for generating fast code?
04:38:48saratogafor something like mpeg player would it be useful or would you still want to use assembly?
04:38:53 Join gartral [0] (
04:39:18kkurbjunyou could probably use the C compiler just fine, I think it's just an older version of what they have in the code studio product
04:39:36kkurbjunI havn't looked at the disassembly to see how efficient it is though
04:39:42 Quit jhMikeS (Nick collision from services.)
04:39:48 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
04:39:50kkurbjunI'm not really familiar with the DSP assembly
04:39:54saratogadoes it use plain c or are there extensions to handle the DSP features?
04:40:09saratogaand does gcc have an assembler for it ?
04:41:02kkurbjunthere's some extensions that I'm aware of for mapping addresses but I havn't really looked into fully utilizing the DSP, binutils has support I think.
04:41:22***Saving seen data "./dancer.seen"
04:41:34kkurbjunright now like I said the code is just setting up the dma transfers and monitoring the buffers
04:41:58kkurbjunare you thinking about trying your hand at one of those DSP's?
04:42:15saratogai've wanted to for a while now, but there was no suitable target
04:42:51kkurbjunwell now you have one :-D
04:43:09kkurbjunthe m:robe's are going for around 50-70 depending on the condition
04:43:20saratogayes they seem quite cheap on ebay
04:43:23kkurbjunI just saw a nearly new one sell for about 50 on ebay
04:43:50kkurbjunthe OF on them is really terrible
04:44:57kkurbjunif you go to that neuros link and login there's the manual and the latest versions of the different compilers TI released if you are interested in taking a look at what features they support
04:45:27saratogai'll look through it
04:50:36 Quit Aurix_Lexico (Read error: 110 (Connection timed out))
05:02:10 Join gartral1 [0] (
05:02:10 Quit gartral (Read error: 104 (Connection reset by peer))
05:02:21 Part gartral1
05:12:29 Join sbhsu_ [0] (
05:17:06 Join yhuang [0] (n=yhuang@unaffiliated/yhuang)
05:17:32 Join sbhsu__ [0] (
05:21:03 Quit killan (Read error: 54 (Connection reset by peer))
05:21:18 Join killan [0] (
05:25:55 Quit timc`` (Read error: 104 (Connection reset by peer))
05:26:08 Join timc`` [0] (n=aoeu@
05:30:16 Join gartral [0] (
05:30:24 Part gartral
05:30:25 Quit midgey (Read error: 104 (Connection reset by peer))
05:31:14 Quit sbhsu (Read error: 110 (Connection timed out))
05:31:42 Quit sbhsu_ (Read error: 113 (No route to host))
05:32:07 Join PatWilborn [0] (n=4a20af79@gateway/web/cgi-irc/
05:32:22 Join midgey [0] (n=tjross@
05:34:09 Quit PatWilborn (Client Quit)
05:39:10 Quit Horscht ("Verlassend")
05:55:15 Join StealthyXIIGer [0] (
06:06:45 Quit saratoga ("CGI:IRC (EOF)")
06:31:29 Join nuonguy [0] (
06:36:22 Quit nuonguy ("This computer has gone to sleep")
06:38:33 Quit Ubuntuxer ("Leaving.")
06:41:24***Saving seen data "./dancer.seen"
06:43:07 Quit z35 ("Leaving")
06:54:25 Quit StealthyXIIGer (Read error: 110 (Connection timed out))
06:55:16 Quit agaffney (Read error: 104 (Connection reset by peer))
06:59:04 Join agaffney [0] (n=agaffney@gentoo/developer/agaffney)
07:14:06 Quit agaffney (Read error: 104 (Connection reset by peer))
07:14:28 Join agaffney [0] (n=agaffney@gentoo/developer/agaffney)
07:19:46 Quit timc`` (Read error: 145 (Connection timed out))
07:25:49 Quit cool_walking_ (" ajax IRC Client")
07:29:44 Quit jhMikeS (Nick collision from services.)
07:29:50 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
07:37:34 Join einhirn [0] (
07:39:18 Join timc`` [0] (n=aoeu@
07:56:52midgeykkurbjun: congratulations on the mrobe work
08:00:27 Quit sarixe (Read error: 104 (Connection reset by peer))
08:00:49 Join Wraith| [0] (
08:04:47 Quit jhMikeS (Nick collision from services.)
08:04:53 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
08:06:19 Quit Wraith| (Read error: 104 (Connection reset by peer))
08:08:27 Quit BHSPitMonkey (Remote closed the connection)
08:19:26 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
08:27:19 Join ender` [0] (
08:29:18 Quit timc`` (Read error: 104 (Connection reset by peer))
08:29:26 Join timc`` [0] (n=aoeu@
08:30:52 Join Nitekreeper [0] (n=724cd23d@gateway/web/cgi-irc/
08:31:00NitekreeperAnyone there?
08:31:10 Quit Nitekreeper (Client Quit)
08:41:27***Saving seen data "./dancer.seen"
08:48:13 Join Rob2223 [0] (
08:50:25 Join arohtar [0] (
08:50:58 Join Zagor [242] (n=bjorn@rockbox/developer/Zagor)
08:51:19 Quit jon-kha (Read error: 131 (Connection reset by peer))
08:51:32 Join jon-kha [0] (
08:56:10 Quit faemir (Read error: 60 (Operation timed out))
08:58:59 Join Bagderr [241] (n=daniel@rockbox/developer/bagder)
08:59:04 Join flydutch [0] (
09:03:23 Nick Bagderr is now known as B4gder (n=daniel@rockbox/developer/bagder)
09:03:39 Quit HellDragon (Read error: 104 (Connection reset by peer))
09:05:25 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
09:05:44 Quit Rob2222 (Read error: 110 (Connection timed out))
09:06:16*B4gder spots a Ladies and Gentlemen commit!
09:07:03B4gderis soap's build server having a hw problem again?
09:18:12amiconnLooks like it.
09:19:36amiconnThe Onda is not yet in the binsize table
09:20:00B4gderI forgot to add it there
09:23:08B4gderI gotta do something about the source package too, it takes too long to build
09:23:48 Quit Llorean (Read error: 110 (Connection timed out))
09:24:45B4gderyet it has been downloaded 388 times during March, which indicates it is being used
09:25:17linuxstbThe Onda build or the source?
09:25:25B4gderthe source 7z package
09:30:32 Join midijunkie [0] (
09:36:25 Join Thundercloud [0] (
09:38:12 Join gartral1 [0] (n=gareth@
09:38:20 Part gartral1
09:49:29 Join n1s [0] (n=n1s@rockbox/developer/n1s)
09:54:23 Quit Thundercloud (Remote closed the connection)
09:55:54 Quit midijunkie (Read error: 104 (Connection reset by peer))
09:59:13 Part pondlife
10:13:10 Quit midgey ()
10:22:29 Quit homielowe ()
10:30:22 Quit linuxstb (Read error: 113 (No route to host))
10:30:50 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
10:35:01 Join gartral1 [0] (n=gareth@
10:35:11B4gdergsoc2009 students should be told that a weekly summary need to be posted to the -dev list or something
10:35:53B4gderto encourage frequent and early discussions around the project
10:36:11 Join nibbler_ [0] (
10:36:26scorchesome projects find it useful to ahve the students keep a current blog...i think that the least we can do is a weekly summary and being active in IRC
10:37:01B4gderyes, I think that'd fit us the best
10:40:00daurnmake bi-weekly irc visits mandatory
10:40:42scorchethat is a bit less than i would like, honestly
10:41:17daurnas in, twice weekly minimum
10:41:29***Saving seen data "./dancer.seen"
10:41:39scorcheunless they have a valid, acceptable reason, i dont see why they cant be always in here...they dont have to talk, but at least be here
10:42:04daurndon't want to leave their comp on all night?
10:42:19scorchethey dont have to
10:42:19daurnhave dialup :p
10:42:35B4gderwhat's that?
10:42:44daurnhave a shared computer
10:42:54daurnor something like only have net access @ uni/place of work
10:42:59scorche<scorche> unless they have a valid, acceptable reason...
10:43:04B4gderdaurn: they are still expected to work on a computer so they would use one
10:43:19 Quit arohtar (Client Quit)
10:44:07daurnscorche: in that way/tone you could say anything.
10:44:18scorchewhat is your point?
10:44:26scorchethis is done on a case-by-case basis
10:45:07scorchethis isnt some broad rule that will apply to thousands of people...we will probably have 4-6 students that will be getting close attention
10:54:09 Quit Zambezi (Read error: 104 (Connection reset by peer))
10:58:29 Join Zambezi [0] (
11:02:30 Part gartral1
11:03:30 Join gartral1 [0] (n=gareth@
11:05:34 Join sbhsu [0] (
11:06:10 Quit sbhsu__ (Read error: 60 (Operation timed out))
11:10:14 Quit __lifeless (Remote closed the connection)
11:10:31 Join __lifeless [0] (n=lifeless@
11:10:53 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
11:12:39 Part gartral1
11:19:20 Join _lifeless [0] (n=lifeless@
11:30:07 Join tvelocity [0] (
11:35:05 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
11:35:26 Quit __lifeless (Read error: 110 (Connection timed out))
11:43:02 Join __lifeless [0] (n=lifeless@
11:53:58rasherIs there a problem with the -dev list? I just received a mail from someone unable to post to it (haven't checked yet if he's subscribed, but I'd guess so)
11:54:16rasherMaybe recent subscribers or something?
11:54:35B4gderI'm not aware of any problems
11:54:45rasherWe had the gsoc guy also
11:55:05rasherNo, devcon
11:55:19B4gderyes, but we also get newcomers posting, like yday
11:56:07B4gderrasher: if you tell me the guys email in PM I can see if he's subscribed
11:58:12 Quit _lifeless (Read error: 110 (Connection timed out))
12:09:04pixelmadidn't the devcon guy say he had a general mail problem?
12:09:59rasherI think this one turned out to be "not sending with the subscribed mail as From: header"
12:21:04 Join _lifeless [0] (n=lifeless@
12:21:45 Quit __lifeless (Remote closed the connection)
12:26:57 Join __lifeless [0] (n=lifeless@
12:30:21 Quit nibbler_ (Read error: 110 (Connection timed out))
12:37:18 Join robin0800 [0] (
12:41:10 Quit _lifeless (Read error: 110 (Connection timed out))
12:41:31***Saving seen data "./dancer.seen"
12:46:15 Join nibbler_ [0] (
12:54:36 Quit SirFunk (Read error: 110 (Connection timed out))
12:55:35 Join SirFunk [0] (
13:16:17 Join moos [0] (i=Mustapha@rockbox/staff/moos)
13:25:39 Quit timc`` ("Leaving")
13:27:11 Quit n17ikh (Read error: 104 (Connection reset by peer))
13:27:20 Join kachna [0] (
13:28:15 Join n17ikh [0] (n=n17ikh@
13:29:30pondlifeDid I dream it, or is it now possible to select a track/album for playback from pictureflow? (It seems to work fine during playback now.)
13:36:19pondlifeSeems that's the main thing missing for it to be a usable alternate browser for those who don't care about battery life ;)
13:36:24moospondlife: I think that is on the Unhelpful's TODO list
13:36:29pondlifeAh, ok
13:36:37 Quit obo (Remote closed the connection)
13:36:38pondlifeI thought it might be a keymap missing (on H300)
13:36:59 Join petur [50] (n=petur@rockbox/developer/petur)
13:37:06 Join obo [0] (
13:37:14moosIt need a bit more work IIRC
13:39:28pondlifeIt's certainly a lot faster/more stable than it used to be. I like.
13:41:21moosyup Unhelpful reworked it and amiconn helped too to optimize
13:48:51 Join fyrestorm [0] (
13:53:49 Join LambdaCalculus37 [0] (n=44a04303@rockbox/staff/LambdaCalculus37)
13:54:26*LambdaCalculus37 congratulates kkurbjun on getting sound on the m:robe 500 :)
13:54:54*moos joins LambdaCalculus37 in the dance :)
13:55:09B4gderyeah, really cool
13:55:18LambdaCalculus37Why hasn't a "Ladies and Gentlemen" mail been sent, though?
13:55:22B4gderI wonder how that helps (or not) the creatives
13:55:47*B4gder now detects that the size info for the onda looks screwed up
13:56:07LambdaCalculus37B4gder: IIRC the ZVM also uses a DM320 core, so that means mculenaere might be able to adjust the sound output for the ZVM.
13:56:23B4gderyeah, that's what I'm thinking too
14:00:21 Quit fyrestorm ("ChatZilla 0.9.84 [Firefox 3.0.7/2009021910]")
14:04:27pixelmapondlife: there's a keypad patch in the tracker as there are quite a few problems due to the fact that pf currently uses plugiinlib actionswith combined contexts
14:04:49pixelmaor better: a controls patch, not keymap
14:04:58 Join jgarvey [0] (
14:07:35*LambdaCalculus37 pokes kkurbjun
14:09:29pixelmapondlife: hmm... misread
14:10:57pixelmathe mentioned patch is not for playback from pf, just fixing bugs in the controls on other targets (just let you know then ;) )
14:11:46pondlifeI was very pleased to find I could now run the plugin without interrupting playback, that seems to be the biggest hurdle overcome.
14:12:32pondlifeHaving said that, it seems to corrupt tagcache data (or maybe dircache data) if rebuffering occurs while in pictureflow. Will try to repro and report properly later.
14:15:58 Join Munkie [0] (n=29b13251@gateway/web/cgi-irc/
14:16:07 Quit Munkie (Client Quit)
14:17:07 Join tyfoo [0] (
14:20:16rasherIs ID3v2->v1 fallback supposed to work for individual frames, or just for the entire tag?
14:21:28 Quit kachna (Read error: 110 (Connection timed out))
14:21:58rasherI have a bunch of files with most info in v1 tags, but replaygain (and nothing else) in v2 tags. They show up as untagged in the WPS
14:31:27 Join draft [0] (
14:31:48draftdoesn't Rockbox support iPod Nano (2nd gen)?
14:32:13B4gderwhich is why the front page of says so! ;-)
14:32:35drafthas Apple protected the bootloader somehow? are there any progress in cracking it?
14:33:25scorchethey have encrypted the firmware as well as using new proprietary hardware
14:34:20 Join Lynx_ [0] (n=Lynx@
14:34:28rasherIs it really any more proprietary than the old ones?
14:34:57scorchedoes it get more proprietary?
14:35:00LambdaCalculus37The hardware in the nano 2nd gen is similar to that of the Meizu M3 and M6; same SoC being used.
14:35:26LambdaCalculus37But yes, the real show stopper is the firmware encryption, which no one has been able to figure out.
14:36:21rasherscorche: just think it's not the most important point to stress
14:36:40 Join goffa [0] (n=goffa@
14:36:54B4gderwoo, a speaking player
14:37:18scorcherasher: though r/e new hardware is yet another task that takes a lot of effort..
14:37:28B4gder(the shuffle thing just posted to the dev ml)
14:39:05rasherSure, but ALL targets are using proprietary hardware. It's nothing specific to the new ipods
14:41:32***Saving seen data "./dancer.seen"
14:46:52 Part B4gder
14:48:06 Join fyrestorm [0] (
14:52:16 Quit goffa_ (Read error: 110 (Connection timed out))
14:55:37 Quit draft ("Lost terminal")
14:59:34Unhelpfulpondlife: it caches a lot of tag data at startup, it doesn't hit tagcache again until you enter tracklist view
15:00:46 Join midgey [0] (n=tjross@
15:07:30 Join faemir [0] (
15:13:52pondlifeUnhelpful: The problem I saw wais that the database browser became corrupted. i.e. silly tag names, or only one album listed... Not managed to repro i from a cold boot though.
15:15:56 Join barrywardell [0] (n=barry@rockbox/developer/barrywardell)
15:18:47 Quit robin0800 (Remote closed the connection)
15:19:30 Join MrSpeedy [0] (n=misko@
15:25:11 Join midijunkie [0] (
15:28:46 Join robin0800 [0] (
15:30:55 Join robin0800_ [0] (
15:33:10 Quit yhuang ("Leaving")
15:34:01 Quit midijunkie (Read error: 104 (Connection reset by peer))
15:37:01Unhelpfulpixelma: you'd said long mode press for menu, short for select, in the album list on ondio? it's just that a long press of the select button is the context menu on the targets i have... i think the application menu should maybe go somewhere else?
15:38:21 Join HellDragon [0] (
15:38:51Unhelpfulpondlife: i had some weirdness when i tried to use db queries while PF was running... no corruption, though, just a very obvious pause, bizarrely only on my e200, when each title was loaded :/
15:39:10pondlifeI think it happened after playback rebuffered
15:40:40pixelmacontext menu for a specific item in lists is on long "right" while select is on a short "right" on the Ondio. But that won't be possible in the horizontally scrolling album list, so short "mode" would be my next choice for select and long "mode" is menu almost everywhere (if possible)
15:40:53 Quit avis (Remote closed the connection)
15:41:09pixelmathere is no context menu in the pf album list, is there?
15:41:32 Join jaykay [0] (
15:42:05Unhelpfulpixelma: there's no context menu anywhere, only the app menu, but surely we'll want a context menu for insert/queue, next/last etc?
15:42:28pixelmaon an album?
15:43:17pixelmawhy not though, it is possible in other browsers too
15:43:58Unhelpfulbut, maybe the app menu should *be* the context menu, with settings/quit below all the other stuff?
15:45:38pixelmahmm... maybe a short "off" for the settings menu and long "off" for quit. Just a first idea, but I can't imagine e.g. up or down used for that (or for the context menu), guess I'd never expect it there
15:46:58pixelmawait... pf stops playback on these lowmem targets curently, not sure it will ever change and how much sense a context menu there...
15:48:40 Quit robin0800 (Read error: 110 (Connection timed out))
15:48:45Unhelpfulactually, i'd talked about it with amiconn, and while you couldn't use it to start bunches of albums, it could always say, "i need to exit to play this for you, is that ok?", then clean up everything except a reference to the item to play, release the audio buffer, start playback, and exit
15:49:07 Join toffe82 [0] (n=chatzill@
15:49:14Unhelpfuli... can *use* a long press of power? on my targets, that's shutdown
15:51:12pixelmait is on the Ondio too, but it's already used for e.g. pause vs. stop in the WPS. Shutdown is an even longer press and even though I don't have to wait too long for shutdown I can't remember accidentally switching off when I just wanted to stop playback
15:52:16pixelmadiscussion seems familiar, ah, talked to pondlife about it recently :)
15:52:42pixelmait was about general keymapping problems though, not pictureflow
15:53:04pondlifeI'd still like to bring more consistency across targets where possible...
15:53:58 Join avis [0] (n=ident@pdpc/supporter/student/avis)
15:54:37Unhelpfuli'm working on that, but it basically means going over target keymaps and figuring out how which of their actions will be hidden by stealing l/r for scrolling, and where those can be remapped. it's taking a while :)
15:57:01pixelmasome plugins already use quit from a menu (not directly with a button). That could be another possibilty although I like being able to quit with a button too
15:58:44pondlifeI think longer term we should expect pictureflow to be an (optional?) third browser- i.e files / database-text / database-pictures.
15:58:59Unhelpfuli like having a dedicated quit, but i think the actions need to be prioritized in terms of putting the most important ones on keys that the user can figure out... scrolling first, then select/cancel/context, then app menu, then quit button
16:00:22pondlifeI don't know if it helps to think of it as a slot-in replacement for the database browser when deciding on buttons or not... ;)
16:01:49Unhelpfulit does, actually, i can steal mappings from whichever context that uses ;)
16:02:36pondlifeYou've got scrolling through the (album) list, going down to a list of tracks (and back up), then context menus/play etc. when a track is highlighted.
16:02:44pondlifePlus exit to top level/menu.
16:03:29pondlifeI guess the plugin menu is an extra function though
16:04:42Unhelpfulcan we easily call into a core function to display a context menu for an item we've found in the DB? if not, and PF has to implement the context menu itself, it might as well also be the app menu.
16:04:49Unhelpfulif so, that's a different story :/
16:09:03 Join CaptainKwel [0] (i=2669ecc2@gateway/web/ajax/
16:09:52toffe82Bagder: you have the change for the hdd6330 to test it or do I miss something ?
16:11:43 Join timc [0] (n=aoeu@
16:16:26pondlifeUnhelpful: It's probably not available at the moment, but I'd prefer if you could reuse the main one. Bit of a high-level function for the plugin API though!
16:19:43 Join sarixe [0] (
16:23:23 Join StealthyXIIGer [0] (
16:25:22 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
16:25:29 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
16:25:44mcuelenaerekkurbjun: congrats on the DM320 work!
16:26:16mcuelenaereLambdaCalculus37: yes, the ZVM has a DM320 core; but last time I checked I was unable to access the DSP from the ARM core
16:26:40LambdaCalculus37mcuelenaere: Do you think kkurbjun's work may help you out in getting sound working?
16:27:10mcuelenaereit certainly does, but I should first be able to access the DSP
16:27:37mcuelenaerethen I should adjust the current code to work with the TLV320 and see if it's connected on the same McBSP
16:29:32 Quit nibbler_ (Read error: 110 (Connection timed out))
16:29:51 Join Llorean [0] (
16:33:52 Join MethoS- [0] (
16:33:54 Join goffa_ [0] (n=goffa@
16:34:51LambdaCalculus37mcuelenaere: And kkurbjun hasn't sent out a "Ladies and Gentlemen" mail, the lazy bum. :P
16:35:09mcuelenaerehe did put it in a commit message though :)
16:35:55mcuelenaereshouldn't this be added to the Ladies And Gentlemen wiki page?
16:36:26LambdaCalculus37Yes, but it's that excitement of seeing that email in your box just calling to you, "Ladies and Gentlemen, we have SOUND!"
16:37:24LambdaCalculus37mcuelenaere: I'll add it.
16:37:25markunwe never sent out a Gigabeat "we have sound" email either
16:38:07gevaertsmarkun: fix that now :)
16:38:20markungevaerts: nah, it can wait :)
16:38:53 Quit jaykay (Read error: 110 (Connection timed out))
16:39:22*mcuelenaere wonders whether distorted audio qualifies as 'sound'
16:39:25 Join ucchan [0] (
16:40:34LambdaCalculus37mcuelenaere: It did for jhMikeS, when he was getting the Gigabeast sound working initially. :)
16:40:48LambdaCalculus37His mail said that the first noise it made was an awful pulsing sound.
16:40:56mcuelenaerein that case, the VX747 has had 'sound' for quite some time now
16:41:19mcuelenaeres/has had/has/
16:41:36***Saving seen data "./dancer.seen"
16:41:44 Quit midgey ()
16:42:44 Join kugel [0] (n=kugel@rockbox/developer/kugel)
16:42:59LambdaCalculus37Looks like I'll have to fix this. Should the mail go to both the regular and the -dev ML?
16:43:21kugelkkurbjun: haha, your iramsize=0 trick helps on the fuze too, no reboot on mp3 anymore
16:43:25mcuelenaereLambdaCalculus37: the M:Robe 500?
16:44:14LambdaCalculus37mcuelenaere: Yes.
16:45:20mcuelenaerelooks like it isn't being consistenly sent to both
16:45:31mcuelenaeres/it isn't/most aren't/
16:45:33pixelmaLambdaCalculus37: do you know the first sng played?
16:45:40mcuelenaerethe rover?
16:45:41pixelmaor song
16:45:57scorcheusually it is sent to the dev list
16:46:02LambdaCalculus37pixelma: No, but maybe kkurbjun can respond and tell us.
16:46:47LambdaCalculus37Actually, there are quite a few players where we have no idea what the first song played was.
16:48:11 Quit MethoS- (Read error: 104 (Connection reset by peer))
16:48:23kugelgiven that the first sound counts, he probably heard some pumpgun sounds of doom :)
16:49:23 Quit goffa (Read error: 110 (Connection timed out))
16:49:29LambdaCalculus37kugel: Or the sounds of Mario jumping in Super Mario Land; he did say Rockboy had sound, too. :)
16:49:29ucchanIf it is not a code freeze still, could you commit the following Bug's patches?
16:49:41ucchanFS #9206, FS #9614, FS #9692, FS #9707, FS #9736, FS #9780, FS #9970, FS #9973.
16:49:53kugelLambdaCalculus37: maybe it was "PIKACHUU"
16:50:13 Join MethoS- [0] (
16:50:18kugelucchan: we are in freeze, but pure bug fixes are accepted
16:51:22LambdaCalculus37ucchan: Are those all bugs with fixes, or new features?
16:53:27mcuelenaereucchan: isn't FS #9614 a new feature instead of a bug?
16:53:29ucchanThe above task is all bug only.
16:54:10ucchanFS #9614 is bug.
16:54:31ucchan"make tools" does not succeed.
16:55:16mcuelenaerebecause it doesn't exist
16:55:23mcuelenaereis there a need for 'make tools'?
16:57:03ucchanI think that "make tools" needs.
16:57:37ucchanExecute "make help" shows "tools - builds the tools only"
16:57:50mcuelenaereah, in that case it is a bug
17:01:29 Join Linton [0] (
17:03:08 Quit robin0800_ (Remote closed the connection)
17:03:22kugelucchan: make tools is done when you build a normal build. I guess it has been build before
17:03:36kugeldid you run make veryclean before trying make tools?
17:03:42 Quit Zagor ("Don't panic")
17:05:57 Join robin0800 [0] (
17:08:15 Quit robin0800 (Remote closed the connection)
17:08:25 Join MTee [0] (n=MTee@
17:10:37 Quit bmbl ("Woah!")
17:10:58LintonOn the Sansa Clip, does anyone else have Rockbox running with no display or sound? I suspect a bootloader problem since I don't even see a splash screen.
17:11:53mcuelenaerekugel: running 'make tools' from a newly configured target gives the same "Nothing to be done for 'tools'"
17:12:34mcuelenaere(even with make veryclean)
17:13:29rasherI think it's a documentation bug
17:13:41LambdaCalculus37Linton: Display and sound both work on the Clip.
17:13:53LambdaCalculus37Linton: Can you still boot the OF? Press the Home button as you turn it on.
17:14:03LintonOF boot works fine.
17:14:11ucchanrasher, "make help"'s bug?
17:14:15Linton(USB or Home)
17:14:28rasherucchan: Yes. Since the set of tools is dependent on which target you're working with
17:14:48rasherAh wait.
17:14:51rasherIgnore me.
17:14:52LambdaCalculus37Linton: Well, you do know that the Clip isn't a supported target yet, right? Unless you're looking to help development work, expect a few bumps in the road.
17:15:01*mcuelenaere committed FS #9614
17:15:09LintonRockbox also runs (saves files to disc, button lights work) but no sound or display.
17:15:23rasherI was thinking of "make" in the tools dir
17:15:46rasherOnce you have the makefile we already know which tools are needed of course.
17:16:04LintonLambdaCalculus37:I know it's not going to be all plain sailing, but no display or sound leaves me a bit stuck as to what to do next
17:16:44LambdaCalculus37Linton: Try rolling the bootloader again, and use a different OF for the upgrade (I used the .30 firmware myself).
17:16:53LintonDoes anyone have a known-working bootloader-sansa.clip file, so that I could rule that in or out as a cause
17:17:07LintonI've tried .30 and .29
17:18:36ucchanIs "Generating dependencies" necessary to build the execution file that exists in tools?
17:18:42LambdaCalculus37Linton: I might.
17:18:47LambdaCalculus37Let me check.
17:19:50FlynDicekugel: you got mp3 on fuze?
17:20:04ucchanmcuelenaere, commit thanks!
17:20:33mcuelenaereucchan: what did you do to produce FS #9970 ? (I tried Gigabeat S simulator with logf enabled)
17:23:54ucchanFor FS #9970, When Archos series's simulator was build, I found this bug.
17:25:00*mcuelenaere also can't reproduce FS #9973
17:25:43ucchanIt makes an error of this because "drivers/serial.c" is not a build object when "SMULATOR" is defined.
17:26:05ucchansee firmware/SOURCES.
17:26:55 Join kkurbjunW [0] (n=karlk@
17:27:21kkurbjunWThanks LambdaCalculus37, midgey, and mcuelenaere
17:27:48*mcuelenaere now gets 'error: SYSFONT_WIDTH undeclared here (not in a function)' when doing a simulator build (with logf) for the Archos Studio
17:28:30mcuelenaerewithout logf it builds
17:28:54LambdaCalculus37kkurbjunW: So are you going to send that mail? ;)
17:29:45kkurbjunW:), it would be a follow up to Cat's email over a year ago. He had sound working then, but it didn't use the decoder, it was streaming pre-decoded pcm from a file
17:30:05rasherI think that counts
17:31:07kkurbjunW:), I'll send something out as a followup
17:31:27 Join MethoS-- [0] (
17:31:59ucchanI also will do the build again when for archos player's simulator.
17:33:30pixelmathe Player/Studio is the one with a charcell display, the error sounds related (no sysfont)
17:34:59 Join yhuang [0] (n=yhuang@unaffiliated/yhuang)
17:36:12LambdaCalculus37Linton: What model of Clip do you have?
17:37:13mcuelenaerepixelma: what do you mean? is it a known bug?
17:38:23 Join perrikwp|lab [0] (i=98214c9d@gateway/web/ajax/
17:38:39pixelmanot that I know of but thought it could help finding where to look at
17:40:31mcuelenaereArchos recorder seems to suffer from FS #9970
17:41:39 Join saratoga [0] (n=41becb3b@gateway/web/cgi-irc/
17:44:05mcuelenaereucchan: what error message is FS #9973 supposed to give?
17:44:16 Join Conic [0] (
17:44:39rashergevaerts: Seems start screen set to Database doesn't work either
17:44:50rashergevaerts: When booting with usb attached on pp...
17:45:01ucchanI build succeed the Archos/Studio's simulator under Cygwin/MinGW.
17:45:16mcuelenaereucchan: with logf enabled?
17:45:38ucchanSorry "logf" is not enable.
17:46:01kugelFlynDice: yea, with disabling IRAM though
17:46:54 Quit MethoS- (Read error: 113 (No route to host))
17:47:04LintonLambdaCalculus37: my Clip is a 2GB, V1, without radio - I think the lack of radio is probably the unusual feature here. Have any previous reports of Rockbox working or not working mentioned if a non-FM Clip was used?
17:47:09 Join midgey|web [0] (i=8dd43638@gateway/web/ajax/
17:47:37ucchanWhen "logf" is em
17:48:02ucchanWhen "logf" is enable, then the simulator build faqilure.
17:48:04rashermidgey|web: I nearly missed that translation!
17:48:13 Join SirFunk_ [0] (
17:48:31mcuelenaereucchan: yes, that's what I was talking about
17:48:33ucchanlogf.c:135 undefined reference to 'serial_tx'
17:48:45mcuelenaerehmm no, I commited your fix for that just now
17:49:04midgey|webrasher: what are your thoughts on the arabic update?
17:49:21rashermidgey|web: I posted a comment
17:49:40midgey|webah yes
17:49:48LambdaCalculus37Linton: I'm not very sure. The one developer who did the bulk of the work hasn't been around lately, and I don't know if any other devs with a Clip have tried the build on a Clip without radio.
17:49:52gevaertsDo we have a HWUSB target without bootloader usb?
17:50:04midgey|webrasher: it seems like most of those insertions are simply copying over the english phrases
17:51:01rashermidgey|web: Didn't he say he didn't do them all anyway?
17:51:16mcuelenaereshouldn't '(A)dvanced' in the configure script have a '(B)ootloader' option to enable logf bootloader builds?
17:51:23ucchanSorry my source is old. (base on r 20289)
17:51:30rashermidgey|web: I think we should just cut out the untranslated ones and commit the ones he's translated
17:51:42midgey|webi agree
17:52:15ucchanI will check for current source.
17:52:30mcuelenaereucchan: what error message is FS #9973 supposed to give?
17:52:46kugelkkurbjunW: I just tried enabling mmu again, using mr500 code. doesn't work :(
17:53:01rashermidgey|web: Doesn't look like it'll be too much work actually
17:53:04rasherI'll get on it
17:53:51kkurbjunWkugel: Yeah, I didn't look into it too much, but I think the problem was that when the codecs were run they were overwriting other portions of the iram that they shouldn't be since we weren't properly monitoring how much the core was using versus the codecs. Didn't bertrik say that he got it running yesterday with an attempt?
17:54:47kugelI didn't notice bertrik getting it to work
17:54:57 Join nuonguy [0] (
17:55:07kkurbjunWI think it's in the logs somewhere that he did an initial attempt and it was still booting
17:55:12*LambdaCalculus37 saw the mail :)
17:55:33kugelkkurbjunW: indeed
17:56:16mcuelenaereLambdaCalculus37: you're adding the song to the wiki page?
17:56:41kugelkkurbjunW: weird. maybe it has something to do with the usage of the iram. the fuze doesn't use it much, but the clip heavily for codecs
17:56:54kugelthus, on the fuze the codec buffer is in dram
17:57:00 Quit MrSpeedy ("Leaving.")
17:57:24kugelnot sure if that makes enough difference to have the mmu working on the clip but not on the fuze
17:58:55LambdaCalculus37mcuelenaere: Just did.
17:58:57kkurbjunWI might be wrong, I was just looking at it briefly and it didn't look right to me so I tried changing it and it worked great.
17:59:21mcuelenaereLambdaCalculus37: yeah, I got the 'RobertMenes has been editing the topic for ..' message :)
17:59:26kugelkkurbjunW: this is what I have now
17:59:51kkurbjunWOn the m:robe with the cache enabled and the small size of the iram it doesn't make much sense to pursue the iram that much, at least not in my opinion
18:00:48kkurbjunWI am thinking about completely getting rid of the iram use on the m:robe since I can make all the irq jumps with a branch instead of an LDR which would potentially be less cycles.. I havn't actually counted the cycles yet though
18:00:51 Part Linton
18:00:59LambdaCalculus37kkurbjunW: Just curious... isn't the mr500 another target with 64MB of RAM?
18:01:50kkurbjunWLambdaCalculus37: yep, 64 megs of ram a DSP and a couple of other co-processors to help with mpeg2, mpeg4, and decoding some other codecs
18:01:58kkurbjunWkugel: I'm looking at it now
18:01:59kugelkkurbjunW: we have 320k iram on the ams sansas
18:02:04kkurbjunWyeah, I remember that
18:02:14kkurbjunWand I think you are right that it makes sense to use it there
18:02:24mcuelenaerekkurbjunW: regarding the DM320 and IRQ: have you seen FS #8687 ?
18:02:41kugelwell, particularly because it means over 15% extra ram for the low mem ones
18:03:21 Quit SirFunk (Connection timed out)
18:04:11kkurbjunWmcuelenaere: yeah, I looked at that feature before.. It could save some time on interrupts as well
18:04:42kkurbjunWI would be interested in getting it working eventually
18:05:20 Quit CaptainKwel (" ajax IRC Client")
18:06:40*kugel eagerly awaits gevearts then
18:06:45kugelbertrik, I mean
18:07:23rashermidgey|web: result was about 125 newly translated strings. Not bad
18:07:35midgey|webnot at all
18:07:45kkurbjunWkugel, make sure that the ttb_base calculation matches the app and boot .lds. Enabling the MMU in the code that you are mapping over might be problematic. If you enabled it in the bootloader running from IRAM first it might work better. Otherwise like you said see what bertrik has.
18:09:47kugelkkurbjunW: I put the ttb into iram for this try. there should be nothing in this area on the fuze
18:10:31kkurbjunWI dont know if the ttb can be in the iram the MMU may not be able to access it to determine the mappings.
18:11:13kkurbjunWI'm not positive on that, but I would bet that it can't use the iram. On some chips DMA transfers can't be run from IRAM
18:11:21 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
18:11:28ucchanFor FS #9970, the simulator build failure under MINGW.(No rule target sysfont.h)
18:11:52ucchanBut under Cygwin, this problem does not occur.
18:12:08LambdaCalculus37rasher: Hey, save me some language patches to commit. :P
18:12:30mcuelenaereucchan: I have the same error for native Linux
18:13:14rasherLambdaCalculus37: You could find out what's wrong with FS #9998
18:13:37kugelkkurbjunW: it didn't work when I tried to have the ttb at the end of dram (and I paid attention to not overlap with other regions)
18:14:16ucchanIs this Makefile's bug?
18:15:27mcuelenaereucchan: I don't know what's causing this
18:15:54LambdaCalculus37rasher: The patch doesn't apply cleanly at all, because the path is completely wrong.
18:16:27 Quit perrikwp|lab (" ajax IRC Client")
18:16:31rasherLambdaCalculus37: sure, but even after you do -p5 or whatever it was, it still doesn't apply at all
18:16:40 Quit Horscht ("Verlassend")
18:16:52LambdaCalculus37rasher: I'll try it tonight after work.
18:17:16kkurbjunWDo you know where it fails booting? Does it happen right after you start to enable the MMU? What happens if you just run the map everything over itself?
18:17:23*LambdaCalculus37 assigns the task to himself so he can work on it
18:17:59*mcuelenaere wonders what's faster: a for-loop looking for the first set bit or a call to find_first_set_bit() (on MIPS target, so CLZ is available as instruction)
18:18:15rasherWith a bit of luck, TomerS will have an answer
18:19:23pixelmamany of the new strings seem to miss the user: line completely
18:19:38pixelmain the hebrew lang patch
18:19:46midgey|webrasher: it seems the second file mostly reverts hebrew strings to english
18:19:57rasherAh, sounds like he used an old file as a base...
18:20:43kugelkkurbjunW: my clip doesn't boot either, with or without ttb in dram
18:21:07kugelkkurbjunW: the mcr command to enable the mmu makes it not boot
18:21:12pixelmarasher: and it looks very messy (still the ### present and the serial bitrate phrases look incomplete)
18:21:22midgey|webI think the patch is reversed
18:23:30LambdaCalculus37Yes, it's very messy.
18:23:41rashermidgey|web: that doesn't sound unlikely
18:24:05LambdaCalculus37Maybe we should tell TomerS to do it again, and this time, make a diff from the lang file currently in SVN.
18:24:09rasherMaybe it applies on top of the other one
18:24:22midgey|webI think I have a unified version
18:24:27midgey|webI'll add it to the task
18:25:20LambdaCalculus37midgey|web: Sure; maybe I'll commit it later on tonight.
18:25:22 Join Ubuntuxer [0] (
18:25:28ucchanFor the problem of sysfont.h, it did not occur when 08-Schumacher-Clean.bdf existed in folder fonts.
18:26:31kugelkkurbjunW: looking at other targets code, I don't really see what I'm doing differently or wrong :(
18:27:00 Part pondlife
18:27:23 Quit midgey|web (" ajax IRC Client")
18:28:36 Quit __lifeless (Remote closed the connection)
18:29:31 Join _lifeless [0] (n=lifeless@
18:30:46ScothHmm, ran into another issue on my iPod 60g 5G running yesterday's build. If I have it docked and exit out of the USB connectivity screen and try to play anything, it crashes with a "Data Abort"
18:31:10 Quit linuxstb (Read error: 60 (Operation timed out))
18:32:19saratogai think that was fixed a week ago, are you running an old build?
18:33:22 Quit _Auron_ (Read error: 104 (Connection reset by peer))
18:33:49 Join kachna [0] (
18:34:03Scothsaratoga: Yesterday's build
18:34:14ScothHaven't tried today
18:35:00ucchanSorry very late, for FS #9973: When Archos's simulator builds, output "No rule to make target ../rockbox.ucl needed by build Stop"
18:35:08 Join archivator [0] (n=archivat@
18:35:18 Join _Auron_ [0] (
18:35:41lostlogiclol, ipod finally got some form of voice support :-P
18:35:58rasherlostlogic: What do you mean finally - they're the first in the world!
18:36:21lostlogicrasher: oh, sorry, MY BAD! *rolleyes*
18:40:07kugelgevaerts: ping :)
18:40:19UbuntuxerHi, I would like to improve the game menus and uniform them. I have noticed that the code style is very different between the games. e.g. some guys used always static functions, but I don't see the necessity
18:40:57gevaertskugel: pong
18:40:59kugelUbuntuxer: It's the other way around. you need to look for reasons for functions to be non-static
18:41:23kugelgevaerts: so, what's the status of the meizus? :p
18:41:39***Saving seen data "./dancer.seen"
18:42:21kugelthe forum thread is hardly suitable if you just want a quick overview of the status
18:42:51Ubuntuxerbut then actually in a plugin, which consist of one file, all functions can be static
18:43:02kugelgevaerts: is this table up-to-date?
18:43:05gevaertskugel: basically unchanged since months. We have a mostly working lcd driver for the M3, and we can read its buttons. M6sp and M6sl need work on crt0.s, and they have formerly-working button readout (clock and timing issues make it not work now)
18:43:22gevaertsAh yes, backlight toggling works well :)
18:43:39kugelbut no fading?
18:43:41gevaertskugel: yes, that's correct
18:43:41kugelfail! :)
18:44:40gevaertskugel: we're leaving some little things for you!
18:45:28kugellooks like is out of date
18:46:48markungevaerts: first things are interrupts and then USB?
18:46:49kugelgevaerts: You can imagine I'll be gladly doing the backlight fading work
18:47:20gevaertsmarkun: that's an option, yes. LCD on the M6 would of course also be nice
18:47:38markunof course
18:47:44*gevaerts wonders if we could manage to get a port to supported state without working LCD :)
18:47:56markunkugel: are you thinking of working on the meizu port?
18:48:46kugelmarkun: not without such a device, I guess. I was just curious about the status
18:49:18 Join vertic39 [0] (
18:49:55LambdaCalculus37Someone get kugel a Meizu! :)
18:54:10 Quit ucchan ("Leaving...")
18:55:41UbuntuxerWhat do you think about uniform the menus? Every game should have a game menu similary to chooper, except that resume game isn't displayed, if it isn't available.
18:57:33saratogathat would be nice
18:57:38*kugel wonders why resume should be unavailable
18:58:03gevaertskugel: if the game doesn't implement resume...
18:58:41kugeloh, resume the game? I thought about resuming playback in the playback menus :S
18:58:54*markun thought the same :)
18:58:57Ubuntuxeryes, I have written a very small patch for brickmania 9953
18:59:26 Join {phoenix} [0] (
18:59:44 Quit yhuang ("Leaving")
18:59:57kugelsaratoga: I can play mp3 on my fuze when I disable the IRAM
19:00:21kugelthis is interesting, I just wonder what needs fixing in this case
19:00:40 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
19:03:58 Part Ubuntuxer
19:06:02 Join ibseco [0] (
19:07:10 Quit vertic23 (Connection timed out)
19:07:20 Quit ibseco (Connection reset by peer)
19:07:58 Join PaulJam [0] (
19:08:21 Join ibseco [0] (
19:12:09 Join miepchen^schlaf [0] (
19:19:09saratogakugel: maybe we don't link the codecs right? do the IRAM addresses in your map file corrispond to the real IRAM?
19:19:39 Join bertrik [0] (
19:19:44saratogamp3 uses more IRAM then most codecs, maybe its more sensitive to that
19:22:15 Join CaptainKwel [0] (i=2669ecc2@gateway/web/ajax/
19:23:30*kugel checks the map
19:26:56 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
19:27:15kugelsaratoga: it seems it's using plugin buffer
19:27:33kugelI don't understand maps fully, but the map shows addresses within the plugin buffer only
19:29:38kugelaac too
19:30:21saratogakugel: the plugin buffer is actually the codec buffer
19:30:28saratogai think its because the script is shared with plugins
19:31:46saratogakugel: does IRAM really start at 0x0 ?
19:31:57 Join z35 [0] (
19:32:46kugelsaratoga: yea
19:33:24saratogaaren't there interupt vectors and such there?
19:34:45kugelafaik, no
19:35:24 Join insertnext_us [0] (n=54982c8a@gateway/web/cgi-irc/
19:35:42 Quit insertnext_us (Client Quit)
19:35:50 Join webguest87 [0] (n=54982c8a@gateway/web/cgi-irc/
19:36:00bertrikI thought the MajorChanges page was more or less reserved for supported targets
19:36:43bertrik(having sound on a new port is great though, I don't want to devalue that fact or anything)
19:36:55saratogahmm my build shows most of the IRAM stuff on PP is in DRAM on the e200v2
19:37:02saratogathat seems a bit odd
19:37:10saratogais ICONST not defined on AMS?
19:38:04kugelsaratoga: something is definitely wrong says it's start address is: 0x30680000
19:38:29webguest87hi, H140, using latest current build, if a album/track is "Inserted Next" to the playing playlist it updates the dynamic playlist but actual playback continues with prior playlist
19:38:34kugelwhich should be neither codec nor plugin ram
19:38:49kkurbjunWkugel, I would expect the interupt vectors to be in iram at 0x0, most of the crt0.s code for arm targets stores a copy in dram and copies it to iram when the code starts
19:39:05kugelwe have 0x800000 total ram, plugin should start at 0x700000, codec at 0x620000
19:40:11saratogai wonder if the crashes we get are due to sections being linked on top of one another
19:40:42kugeleh, the other way around
19:41:03kugelplugin should start at 0x720000 and codec at 0x620000
19:42:05kugelbut I can't spot anything obviously wrong in the .lds files
19:42:40webguest87using build r20287-090311 for ^^^
19:43:33kugelwebguest87: really?
19:44:06webguest87yes, I have been able to duplicate it quite readily sith album inserting
19:44:45webguest87I use Insert several times consequestively then I use Inser Next, it updates the playlist but playback is the prior playlist
19:45:20LambdaCalculus37bertrik: No, there are announcements of sound on new targets on the MajorChanges page as well.
19:45:42webguest87Kugel, I use dircahce, file tree browsing
19:46:18PaulJamcould this be related to the problem in FS #7423 ?
19:46:48bertrikLambdaCalculus37, in that case the clip (and probably other ams sansas) are missing :)
19:47:15LambdaCalculus37bertrik: That'll be funman's fault. ;)
19:47:16kugelbertrik: how did you get the mmu running?
19:48:05 Quit gevaerts ("hardware maintenance time")
19:49:10bertrikkugel, not sure if it's actually running, but at least it doesn't crash
19:49:18kugelcan I see a patch?
19:50:37 Quit PaulJam (".")
19:51:33*kugel 's e200 won't connect
19:52:53 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
19:53:16bertrikthere is some asm code in system-as3525 to explicitly enable the caches. I thought this was no longer necessary after enable_mmu() but rockbox seemed slower without it
19:54:25 Quit webguest87 ("CGI:IRC")
19:54:27kugelbertrik: you didn't define a TTB_SIZE. mmu-arm.S needs this macro
19:54:29 Join yhuang [0] (n=yhuang@unaffiliated/yhuang)
19:55:11kugelalso, you don't seem to mark the ram as cachable. I didn't try that, but I'd think we want to cache it, else we wouldn't need the dcache at all
19:55:46bertrikI didn't get any error about a missing TTB_SIZE
19:55:50kugelbertrik: if it was slower, it probably means the icache is inactive too, which would mean enable_mmu failed
19:56:17bertrikcan't even find it in mmu-arm.s :S
19:56:29 Join webguest87 [0] (n=54982c8a@gateway/web/cgi-irc/
19:56:34kugeloh, I was wrong
19:56:57kugelIIRC it needed it before jhMikeS did the assembly rework
19:57:59kugelseems it disappeared in that rework
19:58:17bertrikI think I'll do a comparison between the code in mmu-arm.S and the arm922T datasheet
19:58:27 Join Thundercloud [0] (
19:58:45webguest87PaulJam; Kugel, my issue with Insert next is as described in FS #7423
19:59:50kugelwebguest87: ah so that's an old issue?
20:00:06kugelfor a moment I thought I caused it with recent commits
20:00:08webguest87appears so, it needs some luv
20:01:17kugelbertrik: I think you didn't succeed. I assume neither is enabled if you experienced a much slower rockbox without the asm below your additions
20:01:46 Join Reptile211 [0] (
20:02:09kkurbjunWkugel, I think that the icache is typically enabled in crt0.s for the targets that need the MMU.
20:02:13kugelbertrik: but you can try to read back the dcache register to see if it's enabled
20:02:24 Quit webguest87 ("CGI:IRC")
20:02:48kugelkkurbjunW: the mr500 doesnt do that?
20:03:58kkurbjunWoh, you're right, the mr500 is different
20:04:04kugelkkurbjunW: can you tell why TTB_SIZE doesn't appear to be used anymore?
20:04:52 Quit jhMikeS (Nick collision from services.)
20:04:58 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
20:05:01kkurbjunWI would guess that we just assume the size since it is consistent on all the targets that use the MMU, we don't do secondary mappings on any of the targets
20:05:12kugelhm, it was never used, it seems
20:05:18kkurbjunWbut I don't know for sure, jhMikeS might be able to answer
20:06:06kugelkkurbjunW: I tried to do all the mmu stuff in crt0.S, without success
20:06:19kugelsame no-boot symptoms
20:06:39 Quit fyrestorm (Read error: 131 (Connection reset by peer))
20:07:03kkurbjunWyeah, it looks like that is the assumption in the code, for a 32 bit space you have a known amount of memory needed to store the ttb table
20:07:52kkurbjunWlike I said though that is just to do the top level mapping, if you need to define area's with finer detail than 1 mb, you will need more space for the ttb
20:08:14kkurbjunWbut I don't see a need for that since we are not using any protected memory or anything like that
20:08:49 Join stripwax [0] (
20:10:40*kugel takes a course in hex numbers again :S
20:10:52kkurbjunWkugel, I didn't see the code that bertrik was talking about before in system-as3525, but I would remove that and let enable mmu take care of enabling the i/dcache
20:11:28kugelkkurbjunW: that was his plan. He said rockbox was noticeably slower then
20:11:54kugeland I experienced this too when I removed that code without doing mmu_enable()
20:12:12kugelenable_mmu() even
20:12:29kkurbjunWIt's the same code as in enable mmu.. I don't see why that would be a problem
20:12:56kkurbjunWexperineced the crash on startup, or slower operation?
20:12:57kugelthat'S why I think the enable_mmu call didn't succeed at all. I don't know why
20:13:28kugelhe said rockbox was much slower but didn't crash
20:13:30 Join arohtar [0] (
20:13:47kugelwhich matches my experience without enabling any cache
20:13:54kkurbjunWtry that while enabling sections of memory for caching and see if it's still slow
20:14:18bertrikI could remap a section and see if the memory is really remapped
20:14:25 Join Lynx [0] (
20:15:36kugelsaratoga: forget my talk about wrong addresses.
20:15:53kugelbut I still don't find any iram-like address in
20:16:31*amiconn still doesn't know whether the gfx drawing optimisations should be in MajorChanges
20:16:32kugelah wait, there's some
20:17:15kugelamiconn: why not? major speed ups are major changes, imo
20:17:15bertrikamiconn, how much faster did it get?
20:17:33gevaertsamiconn: according to Buschel they make a real difference on the ipod video, so I'd say yes
20:18:21saratogai think we should put more in the major changes page
20:18:35saratogamost people don't read the SVN logs, but would still like to know when things are improved
20:18:55kugelsaratoga: what does the /DISCARD/ stand for?
20:18:56 Quit _lifeless (Read error: 110 (Connection timed out))
20:19:00LloreanWe have a category in MajorChanges for "Enhancement, improvement or optimization"
20:19:22LloreanI'd say _any_ visible / easily measurable speed up probably qualifies as an optimization.
20:19:35bertrikamiconn, if it's a directly observable improvement for users, I would add it
20:20:04 Quit Seed ("cu, Andre")
20:20:13kugelexactly my opinion too
20:20:52 Join ucchan [0] (
20:21:09amiconnThat reminds me - I should brush up my gfx performance test plugin and commit it
20:21:24saratogakugel: i've seent hat in map files but i don't know what it means
20:21:38amiconnDoes a test plugin count as a feature commit? Guess not (it doesn't touch the core at all)
20:21:57saratogaamiconn: speaking of which, has anyone ever profiled any rockbox themes to figure out where they spend their CPU time?
20:22:29 Join raphi [0] (
20:22:57Lloreanamiconn: Don't test plugins mostly not compile by default anyway, or have we stopped doing that?
20:23:14saratogathey don't compile by default
20:23:17amiconnThey are not compiled by defalut
20:23:28saratogathough i wouldn't mind enabling some of them by default
20:23:43 Join Seed [0] (
20:24:11*kugel would like to have a debug build which just builds them all
20:24:38amiconnI don't think they should be included in official builds, but then not building them means that breakage won't be detected
20:26:30 Quit timc (Read error: 60 (Operation timed out))
20:26:45LloreanCould we just compile them by default, and intentionally exclude them from "make zip" and maybe add "make debugzip"?
20:26:55 Join timc [0] (n=aoeu@
20:27:01saratogathey're so harmless i don't see why we couldn't just distribute them
20:28:08 Quit arohtar (Client Quit)
20:28:21 Join arohtar [0] (
20:28:50amiconnCluttering the list of plugins, questions "what's this for", slower downloads...
20:29:15bluebrotherif they are not in the CATEGORIES file they shouldn't show up in the menu iirc
20:29:46amiconnThey are in the categories file, because they're very cumbersome to reach if they weren't.
20:29:55amiconnAnd some even need to be (viewers)
20:30:00 Quit faemir (Read error: 110 (Connection timed out))
20:30:49kugelsaratoga: moving the iram start (and reducing the size) for codecs throws an undefined instruction
20:31:20 Quit ucchan ("Leaving...")
20:32:28 Join Aurix_Lexico [0] (
20:32:58 Quit Lynx_ (Connection timed out)
20:32:58 Nick Lynx is now known as Lynx_ (
20:34:13 Join goffa [0] (n=goffa@
20:34:31kugelnow prefetch abort TT
20:37:58 Join Munkie [0] (n=29b13251@gateway/web/cgi-irc/
20:40:01Munkiewhen is 2.3 coming out and whats new in the release?
20:40:11 Quit SirFunk_ (Read error: 110 (Connection timed out))
20:40:18gevaerts2.3 came out years ago...
20:40:18n1sMunkie: that was released years ago ;)
20:40:20kugelMunkie: 2.3 is long ago
20:40:27*n1s wins!
20:40:29kugelwe're so funny :)
20:40:36kugeln1s: gevaerts won
20:40:44n1snot i my logs!
20:40:53kugeloh, in my logs I won! :p
20:40:53LloreanMunkie: What's new in 3.2 is everything we have now that wasn't in 3.1
20:40:59gevaertsSpeaking of 3.2, what do we do about USB?
20:41:09 Join SirFunk [0] (
20:41:39Lloreangevaerts: It works "okay" now. We don't _have_ to change it, right?
20:41:41***Saving seen data "./dancer.seen"
20:41:49LloreanThough we should apply that iPod charging patch probably.
20:42:04gevaertsLlorean: it still doesn't work at all for some people, and I have no idea why
20:42:05LloreanI know it's sorta a new feature, but it's also a _very_ important fix if it works right
20:42:20Lloreangevaerts: Are they all using Ubuntu 9.04 (or newer Linux kernels possibly?)
20:43:00gevaertsI've seen reports about recent ubuntu, osx, xp and vista.
20:43:55LloreanI think we should still include it. Rebooting manually is still an option, and wider testing might help track it down.
20:43:55gevaertsThe vista one is what I've looked into most, as there was a dev on the other end. For some reason the controller never seems to send out data, and I don't think it's actually related to the host OS
20:44:04kugelLlorean: I agree
20:44:18gevaertsIf we disable it, I'd only do that for the release and leave it enabled in trunk
20:45:03kugelIf we disable it, we get less testers, and a smaller chance to resolve remaining issues
20:45:05LloreanI think as long as we include in the release notes "USB Enabled. Should work in nearly all cases. If your player refuses to connect, manual dual boot is still available."
20:45:30kugelas long as we don't claim that 3.2 is a bugfix and stability release, there's no reason to exclude features
20:45:35Lloreangevaerts: So are we initializing the controller wrong possible? What hardware was it?
20:45:35LambdaCalculus37I agree there.
20:45:38 Quit goffa_ (Read error: 110 (Connection timed out))
20:45:47gevaertsLlorean: e200
20:46:31gevaertskugel: "more testers" isn't a valid reason to enable things in a *release* IMO
20:46:41gevaertsin trunk, sure
20:46:42Lloreangevaerts: Did you have him try booting the OF briefly, then going back into Rockbox?
20:47:05gevaertssaratoga: ping
20:47:07kugelgevaerts: it always is. it's vital for open source
20:47:15gevaertskugel: not in a *release*!
20:47:25Lloreankugel: The whole point of a release is that it's not supposed to be "testing"
20:48:15Lloreangevaerts: USB can be disabled at the last minute anyway.
20:48:27gevaertsLlorean: true. We have 12 more days
20:48:37JdGordon|its a point release so it can be not cmpletly rock solid
20:48:56LloreanBut I think unless we have evidence that we're still doing something wrong, we should probably go with "it's expected to work"
20:49:00saratogagevaerts: back
20:49:04 Quit Munkie ("CGI:IRC")
20:49:28gevaertssaratoga: if you have time, can you try disabling the queue_broadcast() lines in usb_storage,c?
20:49:47*gevaerts just thought of that one...
20:50:17kugelsaratoga: I get prefetch aborts if I shrink the iram for codecs (in the same way it's done on pp)
20:50:34gevaertsIt may increase latency by other threads being more active (see also FS #10006). You never know...
20:50:46kugelwhat's a prefetch abort anyway?
20:51:02saratogagevaerts: all of them?
20:51:13saratogakugel: i think its when you try to fetch an op from an invalid location
20:51:23gevaertssaratoga: only the READ and WRITE ones. The others should be harmless
20:52:05gevaertsAlthough nothing actually makes use of them yet, so disabling them all won't break anything
20:53:11gevaertskugel: speaking of usb event broadcasts, have you seen FS #10006?
20:53:20MTeelinuxstb : hello ?
20:53:27saratogakugel: I don't really know enough about this kind of thing to speculate, but it really sounds like we're not seeting up memory correctly
20:53:29kugelgevaerts: any broadcast will interrupt backlight fading IIUC
20:55:14*gevaerts thinks that on the one hand that is a bug, but on the other hand these USB broadcasts may actually be abusing the broadcast system a bit
20:55:14kugelfading can only happen if the queue is timed out
20:55:46gevaertswhy? By design, or an implementation decision?
20:55:55kugelboth, more the latter
20:56:05kugelit was jhMikeS's idea to make it timeout based
20:56:19 Quit LambdaCalculus37 ("CGI:IRC 0.5.9 (2006/06/06)")
20:56:25kugelit works well as long as nothing is constantly broadcasting, of course
20:56:56gevaertsas this shows... How often can something broadcast and not stop backlight fading?
20:57:02 Quit Reptile211 (Read error: 104 (Connection reset by peer))
20:57:27saratogagevaerts: the disk never mounts with those two lines commented out
20:57:36saratogawindows reports that the device isn't recognized
20:57:47gevaertsthat's weird...
20:57:48kugelgevaerts: actually, when it starts fading, no broadcast should be happen, since it will do 1 fading step on each consecutive timeout
20:58:23gevaertskugel: well, you can't guarantee that...
20:58:30kugelI know
20:59:31MTeeDo I need to learn about vlc, vlc tables, sub-band coding..etc. to be able to use the source of the cook decoder ?
20:59:44gevaertssaratoga: can you try plain r20139?
21:00:02kugelto be honest, I was never a fan of the timeout way, but there were problems with too many posts going the "normal" way
21:00:31gevaertskugel: I guess an occasional broadcast at a bad moment just makes that particular fade a bit jerky. Are there other consequences?
21:02:05gevaertsok. so as long as I don't send more often than the biggest timeout you use, we should be mostly fine.
21:02:32*gevaerts now sends up to every 2 or 3 miliseconds. That's probably too often...
21:03:00kugelif a broadcast happens, the queue_wait_w_tmo will be interrupted. then the next queue_wait_w_tmo will start
21:03:12 Quit stripwax ("")
21:03:20gevaertswell yes. We're not absolutely fine, only mostly so :)
21:03:44kugelah, well, mostly might be correct :)
21:03:47gevaertsHow long can your timeouts realistically get?
21:04:27kugeltypically between 20 and 30ms, max is 40
21:05:07LloreanAnother option is just not to fade when USB is connected. If we've got USB charging, it might make sense to simply leave the backlight on, or have a button to toggle it on/off manually.
21:05:08gevaertsOK. I was thinking of only sending status updates every half-second or second or so. We should be fine then
21:05:09kugelHZ/, I mean
21:05:52 Quit balou (Read error: 110 (Connection timed out))
21:05:52gevaertsso between 25 and 50 ms?
21:05:55 Quit bzed (Read error: 110 (Connection timed out))
21:06:11gevaertshm, no...
21:06:18*gevaerts can't think straight
21:06:47kugelthe whole formula is a mess
21:07:14kugelfor e200, it's supposed to be HZ/25
21:07:25 Quit Thundercloud (Remote closed the connection)
21:07:27gevaerts(HZ/(MIN(_FADE_DELAY, 40))) means that you'll always wait longer than HZ/40, right?
21:08:23kugelno, never shorter than HZ/40
21:08:49 Quit Lynx_ (" HydraIRC -> <- The professional IRC Client :D")
21:09:10gevaertsThe MIN() is after the /, so you divide by it. That means an upper limit on time, or am I confused?
21:09:37kugelthe higher the divisor is, the lower is the number of ticks to wait
21:09:38 Quit JdGordon| (" ajax IRC Client")
21:09:48 Join itcheg [0] (i=62db4767@gateway/web/ajax/
21:09:59gevaertsyes, and you limit the divisor to 40. It can't get higher
21:10:02*gevaerts thinks more
21:10:31*gevaerts sees that the biggest MAX_BRIGHTNESS_SETTING for reasonably finished targets is 32
21:10:34kugelthe divisor can get lower. and lower divisors mean more ticks
21:11:03LloreanThe divisor can get higher.
21:11:09LloreanMIN(X,40) means the smallest it can be is 40
21:11:29*Llorean is apparently not thinking well either
21:11:31LloreanIgnore me
21:11:38kugelif X is <40, then X is taken
21:12:07gevaertslet's just work this out... MAX_BRIGHTNESS_SETTING goes from 12 to 32 on involved targets
21:12:14kugelX is 25 for the e200. This the number of ticks to wait between each fade step is HZ/25 (40ms)
21:12:18 Quit yhuang ("Leaving")
21:12:41 Quit raphi ("Verlassend")
21:13:16kugelif you had X==20, it would be HZ/20 > 50ms
21:13:39gevaertsMIN(_FADE_DELAY, 40) is always between 25 and 40, so FADE_DELAY is always between 2.5 and 4, so between 25ms and 40ms
21:14:18kugelyea, until a target with <= 10 brightness levels comes in, that's true
21:14:18*gevaerts hopes that he's right this time...
21:14:37gevaertstrue, but we can solve that when we get there...
21:15:05 Join JdGordon| [0] (i=836b0055@gateway/web/ajax/
21:15:36 Join ibseco_ [0] (
21:15:56*gevaerts can't rememver now, was fade length configurable, or was it always one step per timeout?
21:16:05kugelit's not configurable
21:16:15kugel1 step per timeout
21:16:23JdGordon|death to configurability!
21:16:25gevaertsok, so the longest fade right now would be 800ms?
21:17:54kugelseems correct
21:18:27kugelgevaerts: doing 1s would probably aid bandwith messurements too (if someone implements it)
21:19:24gevaertskugel: I was thinking about having the other side worry about all that, but I was probably wrong
21:20:06saratogagevaerts: do I need logf for this?
21:20:30gevaertssaratoga: not really. If it still goes wrong, we know where...
21:20:37gevaerts(just not why)
21:20:40 Join MethoS- [0] (
21:23:11 Join [1]shawn [0] (n=shawn@
21:23:28 Quit [1]shawn (Client Quit)
21:24:14 Join ch4os [0] (n=ch4os@gentoo/user/ch4os)
21:24:28saratogagevaerts: no change
21:24:43saratogamounts but eventually disconnects during long transfers
21:24:52 Join bzed [0] (
21:25:38kugelgevaerts: I think I'll get rid of the calculation there, it's not so much cases anyway
21:28:51gevaertsThe problem with this USB issue is that it *does* work for a while, so a disconnect can easily cause filesystem corruption...
21:31:07 Quit ibseco (Read error: 113 (No route to host))
21:31:07 Nick ibseco_ is now known as ibseco (
21:31:38saratogagevaerts: for what its worth, after about 15 of these my disk still passes check disk just fine
21:31:46saratogamaybe just luck though
21:31:51gevaertssaratoga: yes, but you're just reading
21:32:00saratogano i mean about 15 writing
21:32:05saratogaeventually i stopped doing that
21:32:08gevaertsah ok
21:32:33saratogaafter I reformatted I didn't want to risk complicating our tests
21:32:42saratogabut before that i did many write tests, apparently without corruption
21:32:51 Join BdN3504 [0] (n=55b23de0@gateway/web/cgi-irc/
21:34:57*gevaerts re-reads last night's log
21:35:36BdN3504Hey, i have something strange to report: I'm using vmware and have compiled a build and when i change to that build directory and open the rockbox-info.txt file there is this line: r20252M-090311 which is weird because the vm tells me i checked out revision 20293
21:36:40gevaertssaratoga: is that a dual-boot machine? i.e. are the working linux and xp setups the same hardware?
21:37:12 Quit MethoS-- (Read error: 113 (No route to host))
21:37:20 Join Reptile211 [0] (
21:37:32 Quit timc (Connection timed out)
21:39:33gevaertssaratoga: did you ever see problems while actually writing, or only when reading?
21:39:34 Quit StealthyXIIGer (Connection timed out)
21:41:10 Join timc [0] (n=aoeu@
21:45:34 Quit Horscht ("Verlassend")
21:46:30 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
21:46:31 Quit BdN3504 ("CGI:IRC (EOF)")
21:46:40 Join nibbler_ [0] (
21:48:10saratogagevaerts: reading and writing had the exact same issues
21:48:18saratogai don't have linux on this machine but I could probably find a live cd
21:48:22 Join tessarakt [0] (
21:48:48gevaertsthat would be helpful. I'd really like to know if this is OS or hardware dependent
21:50:45saratogaok i will try this live cd i found right now
21:51:35 Quit saratoga ("CGI:IRC (EOF)")
21:54:50FlynDicebertik: kugel: I tried bertik's mmu patch. It seems the function calls are inside the "#ifdef BOOTLOADER" section. I moved them out and I can boot but the action is very slow as if dcache got disabled in the process
21:55:18FlynDiceer icache I mean
21:55:31 Quit Reptile211 ("ChatZilla 0.9.84 [Firefox 3.0.7/2009021910]")
21:56:21 Quit barrywardell (Remote closed the connection)
22:05:31 Join saratoga [0] (n=41becb3b@gateway/web/cgi-irc/
22:09:53 Nick BigBambi_ is now known as BigBambi (n=alex@rockbox/staff/BigBambi)
22:10:14mcuelenaerekkurbjunW: I'm not sure whether 'I2C driver 10%' (OlympusMR500Info) is correct, the IC driver on the ZVM works rather good; there could be some issues with wrong clock setup but other than that it should work fine
22:11:25kugelFlynDice: yep, that's what I assume too
22:11:57gevaertsBagder: is the build master stuck?
22:12:06kkurbjunWmcuelenaere: I was just filling it out with a low number for now since I am not in front of the source to check. If there are other numbers that should be updated I am fine with that
22:12:26 Join efyx [0] (
22:12:31kkurbjunWI was mainly just getting a place holder in there for now to start documenting areas that interested parties could look
22:12:52mcuelenaerekkurbjunW: I'm not really asking to correct it, I was just wondering whether you knew of the DM320 IC driver
22:13:23Bagdergevaerts: looks like it, but it'll restart itself in a short while
22:13:45gevaertsah ok. I didn't follow development on that side very much lately...
22:13:46kkurbjunWno, I haven't used it, I don't know if there is anything on the i2c interface for the m:robe 500 offhand
22:14:06kkurbjunWthanks for the pointer in either case
22:14:35kugelBagder: did you add
22:14:53Bagderno, I missed that. We'll look into it immediately
22:15:18Bagderwhat port number?
22:15:35 Join jaykay [0] (
22:15:52mcuelenaerekkurbjunW: oh, isn't there anything connected through IC with the M:Robe 500?
22:16:19JdGordon|the touchscreen is isnt it?
22:16:20kugelBagder: good question. I don't think I have specified a port. I never use a port when I ssh to my machine
22:16:26kkurbjunWI'm not sure offhand, I think most things use the SPI interface
22:16:37Bagder"ssh: connect to host port 22: Connection refused"
22:16:40kkurbjunWJdGordon: I thought that was with SPI
22:16:49kugelBagder: yea, it's offline right now :)
22:16:49JdGordon|you're probably right
22:16:49*mcuelenaere would guess the RTC is, but he will check the source
22:16:55Bagderkugel: oh
22:17:03kkurbjunWI think the RTC is SPI too
22:17:06Bagderkugel: all compilers and sdl installed?
22:17:08kugelI'm right now turning it on, will take some minutes
22:17:26mcuelenaereyeah, RTC is SPI too
22:17:44gevaertsAre there many non-mips-enabled servers left?
22:17:59Bagdergevaerts: 10 something
22:18:17mcuelenaereanyway, it would be nice if you could check whether my IC implementation is correct; that way I could eliminate at least one problem with the PIC on the ZVM :)
22:18:18Bagderbut we only have two mips builds atm, it's not a big problem
22:18:41*gevaerts has a mips-enabled build server that's turned off...
22:18:44mcuelenaereBagder: the VX767 could get added too..
22:18:49kkurbjunWmcuelenaere: I'll check it out sometime - I just need to find something that uses it first :-D
22:18:57mcuelenaereok, no hurry :)
22:18:57Bagdermcuelenaere: I'll be happy to!
22:19:05mcuelenaerewell it doesn't compile currently :)
22:19:16mcuelenaerethe problem is, the only one I know with such a target is wpyh
22:19:20Bagderah then I won't be as happy...
22:19:23JdGordon|kkurbjunW: feel like working on the flash loader?
22:19:32mcuelenaereand I didn't got the LCD working with the little testing I did
22:19:53mcuelenaere(it isn't easy to debug issues when not having the DAP next to you)
22:20:02Bagdergrep -v "^#" | grep -c :mipsel
22:20:18Bagdergrep -v "^#" | grep -c :arm
22:21:00kkurbjunWJdGordon|: We need to do alot more initialization work first. As it stands I would be surprised if the LCD, sound, or external memory worked without the OF
22:21:10kkurbjunWI do want to do it though
22:21:34 Quit JdGordon| (" ajax IRC Client")
22:21:45kkurbjunWbut I think we need to get more of the lowlevel stuff done before it's worth trying to tackle
22:21:51bertrikwhat kind of fm chip does the mrobe have?
22:21:55kugelmeh, why doesn't it work?
22:21:58bertrik(assuming it has one)
22:22:18 Join _lifeless [0] (n=lifeless@
22:22:24 Join JdGordon| [0] (i=836b0055@gateway/web/ajax/
22:22:30JdGordon|ah ok
22:23:03kkurbjunWbertrik: to my knowledge the m:robe 500 doesn't have one
22:23:08kkurbjunWthe 100 might have one
22:23:47kkurbjunWmcuelenaere: what is the pic used for on the ZVM?
22:24:08mcuelenaerekkurbjunW: the pic used for what?
22:24:18FlynDicekugel: I'm getting in a bit over my head here, can we massage the TTB_BASE_ADDR or map_section variables to make this work?
22:24:32kkurbjunWoh there are multiple pics on the zvm?
22:24:32mcuelenaeres/the pic used for what/what pic/
22:24:37mcuelenaereoh you mean the PIC?
22:24:56*mcuelenaere assumed pic -> picture :P
22:25:34mcuelenaerekkurbjunW: it's used for almost all user input (so buttons, touchpad, USB insert, power charger insert, etc) + backlight control and probably other stuff
22:25:43mcuelenaereI think it's described on CreativeZVMPort
22:26:04 Join buk_ [0] (
22:27:14kkurbjunWoh gotcha
22:29:54 Join MethoS-- [0] (
22:32:25 Quit saratoga ("CGI:IRC (EOF)")
22:36:23 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
22:41:29 Join taylor_ [0] (
22:41:45***Saving seen data "./dancer.seen"
22:42:49JdGordon|LambdaCalculus37: did you get around to sending the ipod cable yet?
22:42:59 Quit arohtar (Client Quit)
22:43:12 Join faemir [0] (
22:43:22LambdaCalculus37JdGordon|: I'm going to mail it tomorrow.
22:43:33LambdaCalculus37Didn't have a chance to.
22:43:43JdGordon|ok, cool... no worries
22:43:51JdGordon|i havnt got my ipod back yet anyway so no rush
22:43:58JdGordon|just wondering if it was lost in the mail or what
22:44:13JdGordon|Ill fix up paypal tonight if i get a chance
22:44:31 Join saratoga [0] (n=41becb3b@gateway/web/cgi-irc/
22:45:59LambdaCalculus37midgey: Ping
22:46:39 Quit MethoS- (Read error: 113 (No route to host))
22:48:16 Quit archivator ()
22:49:04saratogagavaerts: same thing in ubuntu
22:49:09saratogahere is the dmesg output:
22:49:34 Quit kugel (Nick collision from services.)
22:49:39 Join kugel [0] (n=kugel@rockbox/developer/kugel)
22:49:54 Quit tvelocity ("Αποχώρησε")
22:50:24gevaertssaratoga: Thanks. That at least means that I still understand a bit of this (i.e. that the logf output doesn't match something triggered by host OS behaviour)
22:52:02saratogaok i have to run
22:52:05 Quit saratoga ("CGI:IRC")
22:52:37LambdaCalculus37rasher: midgey's new patch at FS #9998 applies clean. I'm going to commit it now.
22:53:12taylor_Hello Again, all. I was wondering, does Rockbox have an app that can "talk" to DFUs of mp3 players? (thats either on the web or I can take a look at?) thanks
22:55:23Lloreantaylor_: If it's a "real" DFU mode (as in, as per the spec) there should either be existing tools, or at least the specification that you can investigate.
22:55:52LambdaCalculus37We have meizu_dfu in SVN, which can talk to the DFU mode of the Meizu M6 and M3 players.
22:56:00taylor_ok thanks. Do you know of any link readily available? or will I need to do a little googling? :P
22:56:12taylor_ok cool thanls!
22:57:22taylor_do you know if these are for linux? I guess ill take a look....
22:58:46Bagderopenmoko also has dfu tools
23:00:29taylor_and may I ask also...What exactly can you do with a DFU? other than obviously updating firmware
23:01:35LloreanWe don't really use DFU in most cases. As far as I know ,all of our supported targets at least have non-DFU update modes, so far.
23:02:21 Part moos ("Rockbox rules the DAP world")
23:03:45LloreanAgain, the DFU spec would be where to go to find out what you're *supposed* to be able to do in DFU
23:04:05LloreanBearing in mind that the idea of someone strictly following the spec, outside of cases like OpenMoko, is probably pretty laughable.
23:05:15taylor_hmmmm can seem to find meizu_dfu. What branch is it under? (Im in SVN)
23:05:51Bagdertaylor_: utils/meizu_dfu/
23:06:33taylor_Ah ha.. Thanks Bagder! I was looking right at it ;)
23:07:49 Join Thundercloud [0] (
23:11:57 Quit miepchen^schlaf ()
23:12:19markuntaylor_: hi, we can use the dfu mode to flash our firmware, but we can also abuse it to run our code directly without flashing
23:12:50taylor_Really? How so?
23:12:51markunat least it works like that for the Meizus
23:13:09markunit waits for 2 images. The first is the flash writer, the 2nd is the firmware
23:13:28markuninstead of the flash writer we can just send our test code and it gets executed
23:13:54mcuelenaereBagder: the VX767 should compile now, so you can add it to the build table
23:13:54taylor_I wonder if it would be that easy on the ipods. Prob not
23:14:20markunI guess not, but in the meizu case we are also lucky because nothing is encrypted
23:14:41taylor_ipods are encrypted
23:14:42markunand since we have pretty much all the documentation it's a wonder we haven't finished the port yet :)
23:14:42taylor_too bas
23:14:44mcuelenaeretaylor_: AFAIK the iPods use a custom version of DFU
23:14:47 Join Kohlrabi [0] (
23:15:31markunthe dfu tool from OpenMoko didn't work with the meizus
23:15:57Lloreanmarkun: Do the meizus follow the DFU spec, or was there some changes from it?
23:16:18markunLlorean: I'm not sure actually. Maybe gevaerts knows?
23:17:01LambdaCalculus37markun: I was able to get meizu_dfu to notice my 2nd gen nano by replacing the USB IDs.
23:17:02LloreanMaybe we should be careful about what we call "DFU mode" in cases where it's not actually DFU mode as described as a USB standard.
23:17:52 Join MethoS- [0] (
23:18:31*gevaerts has only committed meizu_dfu, he didn't make it work...
23:18:54LambdaCalculus37Yes, that was wpyh's wizardry that made it work.
23:18:58 Nick taylor_ is now known as davidc__ (
23:19:15 Nick davidc__ is now known as taylor__ (
23:19:33*Llorean just suddenly realized he hasn't yet gotten a response from
23:21:05 Quit {phoenix} (Remote closed the connection)
23:21:05Bagderit is annoying when organizations fulfill those bad expectations you have...
23:21:06 Quit taylor__ ("Leaving")
23:21:30n1sLlorean: afaiu (from what gevaerts have said about it) the dfu spec is only about the transfer protocol, not what is transferred or what the device does with it
23:21:36 Join taylor_ [0] (
23:22:24 Join timc`` [0] (n=aoeu@
23:22:45Lloreann1s: That would still be a standard though. I'm just curious if our "dfu" tool should be abstracted rather than being device specific, or if we should remove the "dfu" from the name.
23:23:57Lloreankugel: might be of interest.
23:24:04 Quit jaykay (Read error: 110 (Connection timed out))
23:24:08 Quit MethoS-- (Read error: 60 (Operation timed out))
23:25:00 Part CaptainKwel
23:26:49 Quit BUMBACL0T ()
23:29:51 Quit avis (Remote closed the connection)
23:30:48kugelLlorean: he came here today, and he pretty much said FS #7423 is his issue.
23:31:13kugelI cannot reproduce it just by inserting and inserting as next in random sequence
23:31:24Lloreanbut the post is newer than that and says the 7423 patch doesn't help
23:31:31 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
23:31:33LloreanAre you following his recipe there?
23:32:39 Quit LambdaCalculus37 ("hungry!")
23:32:44 Quit jgarvey ("Leaving")
23:34:06mcuelenaerewhat unit does cycles have in __timer_set() ?
23:35:16kkurbjunWmcuelenaere: is usb working on the ZVM?
23:35:27mcuelenaerekkurbjunW: no
23:35:40kugelLlorean: I cannot reproduce using his recipe
23:35:43mcuelenaereit does yield some packets on enumeration
23:35:46mcuelenaerebut it stops for some reason
23:36:01kugelI can only reproduce FS #7423
23:36:42mcuelenaerekkurbjunW: how's USB connected with the core on the M:Robe? On the ZVM this is through EMIF
23:37:15kkurbjunWI'm not really sure at all, USB is an unfamiliar territory for me
23:37:28 Quit intrados (Read error: 60 (Operation timed out))
23:37:28kkurbjunWI'm trying to figure out where I would start to implement it
23:37:56kkurbjunWthere's another chip on the board that sits between the SOC and the computer
23:38:07mcuelenaerethe computer?
23:38:20 Quit tyfoo (Read error: 104 (Connection reset by peer))
23:38:29kkurbjunWwell, you have the computer, a USB cable, then a controller, and then the DM320
23:38:43kkurbjunWis EMIF an interface on the DM320?
23:38:48gevaertsCould be a separate PHY
23:38:56Lloreankugel: I've asked him to clarify the recipe. He's slightly imprecise.
23:39:18mcuelenaerekkurbjunW: EMIF = External Memory Interface
23:39:25mcuelenaeredon't you have the DM320 data sheets?
23:39:28 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
23:39:34 Quit timc (Connection reset by peer)
23:39:36kkurbjunWgevaerts: it's this Renesas device:
23:39:39kkurbjunWnot on hand
23:40:07 Quit itcheg (" ajax IRC Client")
23:40:31mcuelenaerekkurbjunW: 'not on hand' -> is that to me?
23:41:09kkurbjunWI would guess it is connected as an external memory interface based on the data/address lines from the device.. mcuelenaere, yeah, I don't have them available at the moment
23:41:30mcuelenaerebut you do have access to them?
23:43:09kkurbjunWwell, I have access to what I have in other places.. as far as the total scope of the DM320 documents potentially available, I am not sure specifically what I have.
23:43:52gevaertskkurbjunW: that's apparently the full USB controller. Nice to see that you have a full datasheet :)
23:44:37mcuelenaerekkurbjunW: but you have access to the NDA datasheets, right? If not, you should mail Catalin/Neuros
23:44:53mcuelenaereI mailed Catalin and he forwarded me to someone at Neuros, who gave me them
23:45:34Bagdergevaerts: quite possibly there are even linux drivers for it
23:45:50kkurbjunW:), gevaerts, what does that mean in terms of implementing the software usb interface? I've been looking at the existing implementations and I'm not sure exactly where to start..
23:46:34gevaertskkurbjunW: you should ask mcuelenaere. I've never started a rockbox driver from scratch...
23:47:10gevaertsBagder: true. I guess I should also look there to see if I find hints about this PP USB issue saratoga and others are seeing...
23:47:52mcuelenaerekkurbjunW: the first step is finding out how to communicate with the USB controller, then try receiving some EP0 SETUP packages
23:49:40kugelLlorean: more precise than the usual recipes though
23:49:44mcuelenaerekkurbjunW: CS3 starts at 0x50000000, CS4 at 0x60000000
23:50:32kkurbjunWmcuelenaere: I'll have to ask cat about talking to it, he had done some initial USB work on linux for the renesas controller.
23:50:44mcuelenaereoh so you already know how to talk to it
23:51:00kkurbjunWI'm not sure exactly how far cat got with it
23:52:03*mcuelenaere looks at cat's site
23:56:02kkurbjunWThere's this branch, but I'm not sure what was changed offhand, when I get home I'll check it out relative to another branch and see what the difference is:
23:56:34gevaertshm, the linux driver for this ARC controller doesn't do as much error handling as we do...
23:57:43 Join MethoS-- [0] (
23:57:43 Quit timc`` (Connection timed out)
23:58:15*mcuelenaere wonders why everyone (except neuros) seems to pick an external USB chip instead of using the one already on-board in the DM320..
23:58:58*Bagder recalls neuros osd1 and usb problems being a frequent subject

Previous day | Next day