Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

#rockbox log for 2004-08-11

00:00:47LinusN3) The i2c drivers were developed when there were no recorders
00:02:23amiconnTo (1): If this is old code, I'll try to move the mas_reset() into the startup code (only), and try if the dsp reset/ codec reset bits in MAS_CONTROL are of any use.
00:02:50LinusNdo so
00:03:08LinusNfunny though, i have had the opposite experience
00:03:29amiconnIn what respect?
00:03:44LinusNi had to go to the recording screen, back to playback and then back to recording
00:04:12amiconnDid you read my bug description at the bottom of the RomBox wiki page?
00:04:15LinusNif i just went to the rec screen, it would record 30 bytes
00:05:10LinusNjust read it now
00:06:51amiconnBtw: I'm just doing a long-term test recording - 24 hours at q=7, 44.1 kHz stereo. Will check that for frame errors
00:07:31amiconnNext will be a frequent time-split recording (several hours in 1 minute chunks)
00:10:07LinusNhow big are your FAT clusters?
00:11:36LinusNamiconn: bad news from Paul on the mailing list
00:13:47amiconnLinusN: fat_cluster_size == 32768 (80 GB drive)
00:14:03amiconnWhat bad news? :(
00:14:25LinusNbig clusters == better recording performance
00:14:34LinusNso you should try smaller clusters
00:15:03LinusNbad news == Paul's recording problems went away when he tried the non-optimized ATA version
00:15:53amiconnIirc 32 KB is the default size for fat32 for partitions >32 GB. I can't choose the cluster size with Win98, and WinXP doesn't even let me format such a large fat32 partition...
00:17:17amiconnConcerning the ATA problem: Did he also try versions with only fast reading/ fast writing disabled? Disk model? hd firmware? hw mask? Archos model?
00:18:04LinusNno more info yet
00:18:52amiconnTo (2): Did you do logic analyses of the i2c drivers? If yes, do you have them still available?
00:19:22LinusNyes i did, but i didn't keep them
00:19:34LinusNi can do new ones, i have my kit with me
00:20:14amiconnIf this doesn't cause too much hassle, it could be interesting. I still have the viewer installed...
00:21:06amiconnAs it gets to logic analyzing: Do you think this may shed some light on the recording init problem after mas_reset() too?
00:21:27LinusNnot sure
00:22:28amiconnRemember that I even managed to trigger the bug once when running from ram. Others did also report that this 30-byte free happens occasionally...
00:22:37amiconns/free/freeze/
00:23:09LinusNi got it a few times when running with gdb
00:24:23LinusNyour latest stack fix, was that to munge the entire stack with DEADBEEF?
00:25:21amiconnYes, because otherwise the main thread panics with stov after using "create directory"
00:25:31amiconns/stov/stkov/
00:25:46LinusNok, but you didn't reply to the bug report
00:26:38amiconnSorry, I didn't realise that there was one. Two users reported it to me on the channel
00:26:45LinusNi'm replying now
00:26:57LinusNnice catch, btw!
00:27:16amiconnI didn't manage to get this myself, because my stack area was a little larger due to using my geek bitswap, so my .iram is smaller
00:27:43amiconnHowever, I got very high stack usage readings after trying that (96..97%)
00:28:05LinusNbadness
00:28:45amiconnI wonder why you changed the linker script in the first place...
00:29:28LinusNthere is a bug in older binutils versions that allows sections to overlap
00:29:46LinusNit is now fixed, so i changed the script to comply with that
00:30:53LinusNfrom the CVS commit log: "Major fixes to please newer versions of the linker"
00:30:55amiconnAh ok. So my fix was "intuitively" right. I also tried defining the start position of the .stack section differently, so that it overlaps with .iram. This worked for me too...
00:32:21amiconnStrangely, I once tried to compile gcc 3.4.1 and current binutils under cygwin (successfully) and then built rockbox with that. This did work too, even without your linker script fixes...
00:33:07LinusNwhich binutils version?
00:33:57amiconn2.15
00:34:04LinusNweird
00:37:05amiconnFor fixing (3): Should there be an additional config define, HAVE_MAS_ENABLE or such?
00:38:14LinusNnah, move the code somewhere else, with #ifdef HAVE_MAS3507D
00:38:47LinusNit'll break the neo, but frankly, i don't care
00:39:03LinusNbut yes, your approach is cleaner
00:39:38amiconnIirc the neo is broken anyway (still builds without warning, but most probably won't run)
00:42:12LinusNand that's why i don't care... :-)
00:42:51***Saving seen data "./dancer.seen"
00:43:56 Quit zeekoe ("quit...")
00:44:17amiconnThe latest bug in the tracker looks like another "classic" rld case...
00:44:44LinusNnot really
00:45:16LinusNwell,maybe it is...
00:47:02amiconnA propos CVS commit log: is there a way to show older entries than those listed on the web page, in chronological order?
00:47:48LinusN"cvs log"
00:48:22amiconnAnd a completely different thing: Why http://rockbox.haxx.se/docs/contributing.html says "Also, avoid using 'const'."
00:48:35LinusNyes
00:49:34amiconn?
00:49:35LinusNcan you guess why it says so?
00:49:42amiconnNo
00:50:13LinusNbecause life is a lot easier if you don't care
00:51:36amiconnImho using 'const' helps to write cleaner code: if you intend to have a read-only variable, declare it 'const', and then later on assign something to it, the compiler will warn you.
00:53:09LinusNi mostly use #define for constants
00:53:11amiconnOther than that, it does have no effect unless the code is running from rom...
00:53:39LinusN...and rockbox was never intended to run from rom
00:55:12amiconn...at the time it was designed. As you said some time ago, thorough checking for proper 'const' use may help in case rockbox gets ported to some other hardware some day
00:55:55LinusNyes, and back then we didn't intend to port rockbox to another platform
00:56:07amiconn...and using 'const' instead of #define may save some memory in case the compiler doesn't do proper constant folding
00:57:21amiconnSo my point is not knowing why the document said so in the past, but why it _still_ does.
00:57:44LinusNbecause we haven't bothered changing them
00:59:09amiconnPerhaps this sentence should be taken out then
01:00
01:01:11LinusNone main problem with "const" is that C++ programmers think that const works the same way in C as well
01:01:30LinusN(i just removed the sentence)
01:06:02 Join sethians [0] (~jirc@200.87.77.112)
01:12:59 Join Nibbler [0] (~andrer@port-212-202-78-112.dynamic.qsc.de)
01:17:42 Quit sethians ("Leaving")
01:20:48amiconnLinusN: (i2c timing) Both mas' want at least 1250 ns SCL high according to the specs. When running from ram, the current implementation of i2c_outb() holds SCL high for only 270 ns (recorders)/ 250 ns (players) [possibly 360 ns/ 333 ns due to pipeline contention]
01:38:59LinusNic
01:40:53zeok so there's no iptables rules, hosts.allow and hosts.deny are set to allow my network, the services are started, the network's up, the box is pingable, nmap accross the network to the box's network ip shows a handful of open ports, nmap on the local box to the box's network ip show all the ports that should be open, all the important services (web, samba, ssh) are set to bind to the box's network ip and are available on the local box
01:41:57LinusNze: wrong channel
01:42:27zewell whats the right one
01:42:29zeheh
01:42:39zethis is the only active one with linux people i'm in
01:42:57zeif you got no idea thats ok, but if you do don't just say wrong channel hehe
01:45:27LinusNze: so what's your problem?
01:45:45zecan't connect to web/ssh/samba from accross the network
01:46:50LinusNwhat happens if you do telnet to port 80?
01:47:21zehmm
01:47:26zewell it was giving me connection refused
01:47:31zeand now it's spontaneously decided to work
01:47:57zeboth ssh and web gave me connection refused every time i tried
01:48:15zeis there something that'd cause that to be like that for some period of time and then after a while it'd kick in? heh
01:48:33zei dunno, it's weird
01:49:01amiconnLinusN: My 24 h recording finished - bad news :(
01:49:28zeafter the box boots up, the only open ports are 135, 139, 445, 1025 (as reported by nmap) ... but nmap'ing locally, or now after whatever spontaneously happened, i get 22, 80, 111, 139, 445, 6543, and 6544 as i should
01:50:24LinusNze: weird indeed
01:50:31LinusNamiconn: corrupt frames?
01:51:42amiconnMp3utility reports a sync error a little above 3 minutes in the 24h file, and right at the beginning of the file started after the 24 h split. Unfortunately it stops analyzing after encountering the first error
01:53:04amiconnVbrfixer tells me there are 4 mpeg version changes in the short file (this is only 314 KB !). It is unable to analyze the big one: runtime error - out of memory
01:53:11zeerr scratch 6534 and 6544.. those only show up locally, which is probably right, but the rest is as i said
01:53:30zeit's like it doesn't start working till the uptime reaches 10 or 15 minutes
01:53:30zeheh
01:55:02LinusNamiconn: could be disk related problems
01:55:29zeer! it's doing it again
01:55:34zeRead from remote host pvr: Connection reset by peer
01:55:34zeConnection to pvr closed.
01:55:44*ze checks the crontab
01:58:20LinusNamiconn: gotta sleep now
01:58:39amiconnnite LinusN
01:59:12zehmm i bet it's this 'shaper' crap whatever it is
01:59:16amiconn(Just checking with mp3fixer, although I know that the windows version isn't always reliable)
01:59:25zeseems to run every 10 minutes and is something about traffic control
02:00
02:00:06zeand all i had to do was wait 20 minutes
02:00:06ze:p
02:00:32zeif thats actually it
02:00:47LinusNok, nite all
02:00:50 Part LinusN
02:01:25zeer doh... well if you see the longs, Thanks and goodnight heh
02:01:31zes/longs/logs/
02:12:13zehaha it might not be though, apparently i'm using an ip address thats already in use :p
02:17:58zeteehee that was it... how stupid
02:18:17zethat'll teach me to leave out an /etc/hosts entry for a new system
02:24:38amiconnnite all
02:25:26 Part amiconn
02:42:54***Saving seen data "./dancer.seen"
02:45:24 Quit AciD (Read error: 104 (Connection reset by peer))
02:46:01 Join AciD [0] (~gni@longchamp44-1-82-67-133-87.fbx.proxad.net)
03:00
03:05:45 Quit Nibbler (Read error: 104 (Connection reset by peer))
03:24:35 Join Heil_Eri1 [0] (~centericq@p508B3B13.dip.t-dialin.net)
03:42:22 Quit Heil_Eris (Read error: 110 (Connection timed out))
04:00
04:42:58***Saving seen data "./dancer.seen"
04:46:08 Join edx{code} [0] (edx@pD9EAB9F0.dip.t-dialin.net)
04:46:44 Join Nibbler [0] (~andrer@port-212-202-78-112.dynamic.qsc.de)
04:57:41 Quit edx{code} ()
05:00
05:02:23 Quit edx (Read error: 110 (Connection timed out))
05:03:55 Quit Hadaka (Read error: 110 (Connection timed out))
05:19:01 Quit Nibbler (Read error: 54 (Connection reset by peer))
06:00
06:42:59***Saving seen data "./dancer.seen"
06:59:28 Join Naked [0] (naked@naked.iki.fi)
06:59:49 Nick Naked is now known as Hadaka (naked@naked.iki.fi)
07:00
07:04:33 Join Nibbler [0] (~andrer@port-212-202-78-112.dynamic.qsc.de)
07:35:24 Quit AciD (Read error: 54 (Connection reset by peer))
07:35:24 Quit Nibbler (Read error: 54 (Connection reset by peer))
08:00
08:23:31 Join Madman [0] (~madmangor@69.107.248.21)
08:24:29Madmancool, you fixed the stkov error when making a dir in the new build
08:24:58Madmani was talking to amiconn and midk about it
08:30:45 Quit Madman ()
08:43:02***Saving seen data "./dancer.seen"
09:00
09:00:29 Join Bagder [241] (~dast@labb.contactor.se)
09:00:40Bagderok amp let me out
09:01:07Bagderhello good people
09:05:23 Join Nibbler [0] (~andrer@port-212-202-78-112.dynamic.qsc.de)
09:07:49 Join amiconn [0] (~jens@pD95D1D06.dip.t-dialin.net)
09:30:04dwihnoHello, fellow Bagder. The amp won't let you go \o/
09:30:05 Quit Nibbler (Read error: 104 (Connection reset by peer))
09:38:51 Join Zagor [242] (~bjst@labb.contactor.se)
09:41:19 Join edx [0] (edx@pD9EAB9F0.dip.t-dialin.net)
09:44:23Bagderhttp://www.opengroup.org/publications/graphics/n900cov.jpg
09:44:33Bagdernow, can I have that on my car? ;-)
09:45:14Zagor:)
09:45:37Zagori would guess both UNIX and LINUX are taken already
09:45:44Bagderyeps
09:46:19Zagorprobably not CURL though ;)
09:46:38Bagderits just a bit too expensive to be worth it, imho
09:46:52Zagorhow much is it?
09:47:07Bagderit used to be 5000 SEK
09:47:15Bagderfor a few years only
09:47:25Bagderbut I know they talked about raising the fee
09:47:26Zagorit's time limited?
09:47:28Bagderyes
09:47:32Zagorweird
09:47:51Zagordeduct it as a marketing expense ;)
09:48:00Bagder:-)
09:49:24Bagderthat plate is available as a mug too ;-)
09:49:30*Bagder considers buying one
09:49:42Zagor:)
09:51:38Bagderhttp://www.opengroup.org/bookstore/catalog/n901.htm
09:51:44Bagderalso a fun product
09:52:02Bagder"don't leave home without it" ;-)
09:52:36Zagorbusiness card cdroms are always fun. rather useless content though
09:53:20Bagderthe SUSV3! ;-)
09:53:36Bagderwhen you buy that in hardcopy, is 11kg!
09:54:45Zagorhaha
09:54:51Zagor...as we all do
09:55:06Bagderstupid google refused my app due to "bad postal address"
09:57:27Bagderbut the mail was in swedish
09:57:42Bagderand they didn't say what was bad with it
09:59:43Zagormaybe it was in their pyramid scheme blacklist
10:00
10:00:32Bagderand then proceeded and said, if you want to fix it, fill in a new application with a DIFFERENT email adress since this one is now connected to the failed application
10:00:42dwihnohmm
10:01:07Zagorhaha, and that failed application can never be fixed?
10:01:12Bagderyeps
10:01:13dwihnosince you guys are ninjas when it comes to virtually anything, do you know what might cause a crc checksum failure with large packets? I'm testing a midlet ssh client which is driving me nuts :)
10:01:22Bagderand since I don't know what the problem was, I'm deemed to repeat it
10:01:22Zagorcute system
10:01:44Zagorthere must be a contact address somewhere
10:01:48Bagderyes
10:01:52BagderI mailed them
10:02:10Bagderdwihno: you're talking about crc problems on the network?
10:03:14dwihnoBagder: Not really... Larger packets seem to trigger the crc failure
10:03:25dwihnoas long as the packets are small, it runs so smoothly!
10:03:33Bagderso the ssh client is what detects the checksum problems?
10:03:43dwihnoyup
10:04:07dwihnothere must be a problem in the client since the Z1010 is the only phone with this specific problem
10:04:26Bagderyes
10:04:38Bagderthe traffic is checksumed already on the TCP layer
10:04:54Bagderit can't be bad when the app receives it, unless the server sent bad data
10:04:59dwihnoyeah
10:05:27dwihnoA couple of larger packets just trigger the internal crc failure
10:05:48Bagderit sounds like a good old application bug
10:06:13dwihnoRather strange it only occurs on my phon.e
10:08:25dwihnoWell. I've contacted the author with my findings. Perhaps he has any ideas.
10:33:33dwihnoUsing the telnet-only version of the client, the same problems arise. I wonder what might be wrong
10:43:05***Saving seen data "./dancer.seen"
10:45:28Bagderpossibly because you have an usual MTU?
10:49:40dwihnoMTU?
10:50:00dwihnoTelia felia lingonbär \o/ :)
10:50:06Zagormaximum transmission unit
10:51:40dwihnoChangeable?
10:53:40dwihnoFixable? :)
10:53:42dwihnoRepairable?
10:54:23ZagorMTU is the maximum block size your device sends. it can be changed, but generally you don't want to do that since it will affect retransmissions etc.
10:54:54dwihnoOkay
10:55:15dwihnoIt is the receiving that breaks, afaik
10:55:20dwihnoPerhaps MTU affects both?
10:56:42Zagoryes
10:59:02BagderI'm just suggesting that the app might have some assumptions in the code that your situation breaks
11:00
11:01:44dwihnothe application assumes larger buffers than the phone has to offer? (otherwise, it should work, huh?)
11:02:32Bagderwell, I refrain from guessing exact details, but since it complains on bad checksums and it really shouldn't be any checksum errors...
11:04:01dwihnoIt isn't just the checksum
11:04:15dwihnoI've tried the plain telnet version of the software and it still bangs
11:05:45dwihnoSo if I was to fix the telnet version, the same fix could most likely be applied on the ssh software
11:07:01Bagderperhaps the bug is in code used independent of the actual transport (ssh or plain telnet)
11:08:17dwihnocould be
11:08:28dwihnomight be
11:08:29 Join Nibbler [0] (~andrer@port-212-202-78-112.dynamic.qsc.de)
11:08:32dwihnomust be
11:08:33dwihno:-)
11:49:50 Quit Nibbler (Read error: 104 (Connection reset by peer))
11:53:58webmindmorning
12:00
12:10:03webmindanyone here know if linux mass storage driver can take 2 disks on 1 usb-ata controller ?
12:13:15amiconnwebmind: I guess this depends on the implementation of the usb-ata bridge. Usb mass storage uses the scsi protocol, so the os driver shouldn't be the limiting factor
12:13:46webmindhmm
12:14:00webmindany idea if it would work on the chip ?
12:36:59Zagori guess it will work, since the ata bus handles two devices and thus the chip does not have to do anything special to support it
12:39:03Zagorhave you checked the data sheet?
12:42:10webmindjust did.. chip claims to be fully ata/atapi compliant..
12:42:14webmindso should be no problem
12:43:09***Saving seen data "./dancer.seen"
12:43:10Zagorare you going to mount dual flash cards, or what is your plan?
12:44:42webmindmy HD on my laptop is dying...
12:44:56webmindso i'm thinking of making a portable raid system
12:45:17webmindraid I can do in software.. so just need 2 disks
12:46:02Zagoraha. battery powered or just movable?
12:46:10webmindbattery powered
12:46:16Zagorneat
12:46:21webminduhuh
12:46:59webmindespecially since I will probably will be able to combine it with my pda
12:48:33Zagora zaurus? or does anyone else have usb host ports these days?
12:57:21webminduhuh
12:57:26webmindzaurus sl-6000L
13:00
13:09:26 Join Nibbler [0] (~andrer@port-212-202-78-112.dynamic.qsc.de)
13:49:25 Quit Nibbler (Read error: 104 (Connection reset by peer))
14:00
14:43:11***Saving seen data "./dancer.seen"
14:49:16 Part Zagor
14:49:27 Join Zagor [242] (~bjst@labb.contactor.se)
15:00
15:09:13 Quit gromit`` (Remote closed the connection)
15:22:33 Join Ericgoh [0] (~kapipo@p29150-adsau12honb7-acca.tokyo.ocn.ne.jp)
15:22:52Ericgohis Zagor here?
15:23:24BagderZagor: BOOO
15:23:41Ericgohah...
15:23:50Ericgohgot something wanna ask him
15:23:55*Bagder tries to scare him awake
15:24:59Ericgohi am trying to find the datasheets for portalplayer IC, heard that Zagor has it
15:25:14Ericgohwonder is anyone here has the datasheet too?
15:25:23Bagderhe doesn't have that
15:25:27Ericgoh??
15:25:29BagderI don't think anyone has
15:25:40Bagderunless they chop off an arm and a leg
15:26:03Ericgohcoz i saw the irc log in rockbox homepage saying that he got it from someone else
15:26:20Bagderthen I'm wrong
15:26:33Bagderbut I don't think I am
15:26:50Zagori'm here
15:26:58*Bagder jumps
15:27:06Bagderdon't scare us like that
15:27:09Zagorbagder is right, i don't have the sheets.
15:27:15Ericgohohh
15:27:32Zagori got an agreement to sign that required me to be their eternal soul slave and sign off my firstborn.
15:27:32Ericgohmaybe i misunderstood
15:27:43Ericgohhaha
15:29:45Ericgohthe stupid portalplayer keep their IC so secret that nobody knows wtf they doing
15:30:18Bagderthe question is why companies buy it
15:31:17Ericgoh??
15:31:54Ericgohdo you mean LSI Logic?
15:32:01Bagderwhy do companies build players based on portalplayer
15:32:34Bagdermaybe this shows why I don't run a company that does that ;-)
15:33:03 Join Nibbler [0] (~andrer@port-212-202-78-112.dynamic.qsc.de)
15:33:22Ericgohoh...
15:33:51Ericgohbut portalplayer eating the battery power like hell
15:34:22Ericgohalso the reason why player based on portalplayer sucks on their playtime
15:34:46Zagori'm guessing the portalplayer chip is part of an attractive package with nearly complete firmware and so on.
15:35:04Ericgoh<−−agree
15:35:04Bagdermost likely, yes
15:35:12Zagorwhich means less work for the manufacturer == less risk == easier to calculate for the bean counters
15:35:19Ericgohportalplayer actually selling their solution but not IC
15:35:26Zagorright
15:35:41Bagderjumping off a cliff is cheap too ;-)
15:36:01Zagor:)
15:36:23Ericgohyap, hope ipod jumping off the cliff one day
15:36:28Ericgohreally hate ipod
15:37:35Ericgoh<−−anti-apple
15:37:45Zagori can't say i hate ipod. i more dislike uncritical consumers
15:38:47dwihno\o/
15:39:04dwihnoRegular hard disk players with NORMAL batteries si teh bestest! :)
15:39:24Bagderthey will be history soon
15:39:56Ericgohhope such day comes as early as possible
15:41:54dwihnoBagder: hush! don't say that!
15:42:25Ericgohipod is now trying to do like microsoft to monopolize the market
15:43:22Bagderevery company would want to do that
15:43:46Bagderbut as Zagor said, its the consumers that let them
15:44:06dwihnoThat reminds me. Any of you guys have a cheap 512 meg memorystick duo for sale? )
15:44:10dwihno:)
15:44:30Ericgohso consumers should think wise before they buy, not just buying through advertisements
15:44:48Ericgoh??MS DUO
15:44:57Ericgohyou want that for what?
15:45:11Ericgohyou never get MS for cheap
15:46:29dwihnoMy cell phone
15:46:34dwihnoI want to keep some music in it.
15:46:40dwihno256 meg chips are affordable
15:47:18Bagderwhen talking about not buying crappy products, memorystick sure is a good thing to mention ;-)
15:47:23Zagordoes your phone use memorystick? is it a sony?
15:47:25Ericgohthat's rare
15:47:45Ericgohhaha, what else other than sony use MS
15:47:47dwihnoZagor: yep. and the chip is 100% covered as well
15:47:55dwihnoEricgoh: Behave :)
15:48:34EricgohMS license almost double of SD Card, only rich company afford to use that
15:48:36Zagorrule #1: never buy sony hardware
15:48:51dwihnoZagor: how come?
15:49:18Bagderproprietary crap
15:49:20Zagorthey always use crappy proprietary things like memory stick to force you to buy everything else from sony too
15:49:20EricgohSony is throwing all the useless rubbish into MS causing it to be so much bigger than SD
15:49:45Ericgohalso increasing the development cost for MS
15:59:39dwihnoZagor: well, I'm really satisfied with my phone anyhow :)
15:59:55Zagorwhich model is it?
16:00
16:01:04dwihnoz1010
16:02:31Ericgohhow many hours of music can you play with it?
16:02:41dwihnoDunno really...
16:02:55dwihnoI only fit 4 tracks on the chip :)
16:03:38Ericgohhaha...wonder what happened to the phone when you find out battery ends becoz of playing music
16:03:52dwihno:-)
16:03:56dwihnoI can live with that
16:05:20Ericgoh??then what the purpose of getting a cell phone?
16:05:40BagderEricgoh: so your phone never runs out of batteries?
16:06:16dwihnoEricgoh: I charge my phone regularly
16:07:15Ericgohmy phone can stand for more than 1week
16:07:32Ericgoheven i talk frequently, it can still stands for 2-3days
16:08:24Ericgohactually i bought a very stupid phone,
16:08:43Ericgohbut the battery life is very long
16:09:24dwihnoWell, people should do as they see fit :)
16:09:37dwihnoAnd not critisize everyone everytime they think differently
16:09:42dwihnoPerhaps I should run for president
16:09:45dwihno;)
16:09:50*Bagder votes for dwihno
16:10:22dwihnoYay!
16:10:27Ericgohthats not critisize, i am just giving opinion
16:10:30dwihnoWatch out Kerry - here I come! :)
16:10:57Ericgohcoz already got lots of customer complaining about the battery life while playing music on cell phone
16:11:31Ericgoheven though they already know how long they can play the music on their cell phone
16:11:31dwihnoKeeping the phone charged and doubling it as a mp3 player works for me
16:12:00Ericgoheveryone thinks like that then we wont get any complain, haha...
16:14:05dwihno:-)
16:22:28Bagdertime to go swimming!
16:22:30 Part Bagder
16:30:43 Part Zagor
16:34:05 Join dstar5 [0] (Lee@ACC2BA2C.ipt.aol.com)
16:43:15***Saving seen data "./dancer.seen"
16:58:54 Quit Nibbler (Read error: 104 (Connection reset by peer))
17:00
17:20:42 Quit dstar5 (Read error: 104 (Connection reset by peer))
17:21:19 Join dstar5 [0] (Lee@ACC2E45A.ipt.aol.com)
17:36:14 Join sethians [0] (~jirc@200.87.78.53)
17:54:42 Quit Ericgoh ()
18:00
18:22:54 Quit sethians ("Leaving")
18:34:43 Join Nibbler [0] (~andrer@port-212-202-78-112.dynamic.qsc.de)
18:43:16***Saving seen data "./dancer.seen"
19:00
19:17:26 Join nevs [0] (nevs@ACBD5A76.ipt.aol.com)
19:20:10nevshi
19:26:29 Quit nevs ()
19:26:45 Quit dstar5 ("Leaving")
19:44:25 Quit Nibbler (Read error: 104 (Connection reset by peer))
20:00
20:43:20***Saving seen data "./dancer.seen"
21:00
21:17:54 Join scott666_ [0] (~scott666@c-24-245-58-48.mn.client2.attbi.com)
21:17:55 Quit scott666 (Read error: 54 (Connection reset by peer))
21:18:00 Nick scott666_ is now known as scott666 (~scott666@c-24-245-58-48.mn.client2.attbi.com)
21:27:26 Join Nibbler [0] (~andrer@port-212-202-78-112.dynamic.qsc.de)
21:44:08 Join jakesir [0] (solaris@pool-70-16-28-73.balt.east.verizon.net)
21:44:26 Quit jakesir (Client Quit)
21:50:51 Join dstar5 [0] (Lee@ACC478EA.ipt.aol.com)
21:55:16 Join [IDC]Dragon [0] (~d90a3255@reladm.kharkov.net)
21:57:28amiconnhi Jörg
21:57:39[IDC]DragonHello Jens
21:58:41[IDC]DragonI'm looking for Jake, he only sticked his head in here and left
22:00
22:00:04amiconnHmm.
22:00:27amiconnDidn't he have some problems with his box?
22:00:50[IDC]Dragonyes, he shipped it to me
22:01:44amiconnInteresting. Isn't this rather expensive (I assume he lives in the states)
22:01:48amiconn?
22:01:49dstar5what was the problem with his box?
22:01:57[IDC]Dragonhe suspected a bad flash, but it turns out to be a bus problem or so
22:02:26[IDC]DragonI've hooked it up to uart_boot, the flash reads OK
22:03:25[IDC]Dragonbut when dumping ROM+flash from revived Rockbox, the files have a stuck 0 bit every 8 KB
22:03:47amiconnHmm. Do you already have an idea what could cause this?
22:04:03[IDC]DragonI ran out of them
22:04:33amiconnExactly _one_ zero every 8 KB?
22:04:40[IDC]Dragonfirst I thought it's a RAM error, I made a test to uart_boot, but this went OK
22:05:07[IDC]Dragonyes, bit 1 on every 8 KB
22:05:23amiconnOne bit only?? Really strange
22:05:45[IDC]Dragonthe address has to end on FF8
22:06:00dstar5well... if you can't fix it... i have a good CPU board he could use it does not have a flashable chip though
22:06:08[IDC]Dragonso, on 0FF8, 2FF8, 4FF8, and so on
22:06:32[IDC]Dragonit's a V2
22:06:49[IDC]Dragonbut CPU boards are welcome!
22:06:54dstar5ohh
22:07:04[IDC]Dragondoes it have USB2?
22:07:09dstar5yes
22:07:11amiconn[IDC]Dragon: Did you try a ram check from within rockbox?
22:07:43[IDC]Dragondssar5: the box of my gf could need such an upgrade
22:08:08[IDC]Dragonamiconn: no, I didn't went through that effort
22:08:53dstar5[IDC]Dragon: i could send it to you
22:09:07[IDC]Dragondstar5: where are you?
22:09:32amiconn[IDC]Dragon: I suppose rockbox crashes every now and then with that error. Does it boot into rockbox from flash, or do you need to load rockbox via uart?
22:09:45dstar5[IDC]Dragon: US (oregon), but he board is so small, it would not cost much at all to ship
22:10:23[IDC]Dragonamiconn: the box starts very unreliable recently, but from flash
22:11:08[IDC]Dragondstar5: the wreck of Jake is from the US, too, I had to pick it up at customs
22:11:37dstar5[IDC]Dragon: at customs? why is that?
22:11:56amiconn[IDC]Dragon: I wonder how this works... The bootloader has to uncompress the ucl image. Doesn't ucl incorporate a checksum?
22:12:47[IDC]Dragondstar5: probably they suspected it could be valuable
22:13:17[IDC]Dragonamiconn: ucl has a checksum, but this is not verfied on decompression
22:13:25dstar5ahh well that sorta is... heh
22:13:26amiconnAh
22:13:55[IDC]Dragoneverything with a checksum fails on that box: RoLo, firmware_flash.rock
22:14:55amiconnSo the problem also exists for the ram... or for the ram _only_??
22:15:50[IDC]DragonI'm puzzled, because my RAM test went OK
22:16:18[IDC]Dragonthen I suspected the RAM on the HD, changed the disk, but same result
22:16:57amiconnIt could well be the ram - one bit cell may be flaky, unable to maintain it's state between refreshes
22:17:42[IDC]Dragonbut I wrote a RAM test...
22:18:04amiconnIirc the refresh cycle is programmable, perhaps the boot loader sets a different value than rockbox?
22:18:13[IDC]DragonI filled the whole with 0x55, verified, filled with 0xAA, verified again
22:19:02[IDC]Dragonthen I filles with 0x00 and wrote a marching 0xFF, verified that this stays, then reverse values
22:19:15[IDC]Dragons/filles/filled
22:19:30dstar5[IDC]Dragon: if you want the board, email me dstar5 at gmail _D0T_ com
22:19:36[IDC]Dragon(to catch address errors)
22:19:37dstar5see you later :)
22:19:48[IDC]Dragondstar5: thanks, yes
22:19:58 Quit dstar5 ("Leaving")
22:23:12[IDC]Dragonafk
22:32:33amiconn[IDC]Dragon: (1) From which memory region the ram test executed? IRAM? (2) The bootloader starts with RAS up mode and burst mode disabled. Rockbox does enable both...
22:36:41amiconn(RAS down and burst)
22:43:23***Saving seen data "./dancer.seen"
22:47:29 Join Bagder [0] (~daniel@1-1-5-26a.hud.sth.bostream.se)
22:49:48 Join undefinedassasin [0] (~Aggressor@undefinedassasin.registered)
22:49:51undefinedassasinh3y
22:49:53undefinedassasinwhere is midk
22:52:51[IDC]Dragonback again
22:53:30undefinedassasinWher3 is midk?
22:53:34[IDC]Dragonamiconn: the test is using minimon in IRAM, bootloader DRAM init
22:54:06[IDC]Dragonin an act of despair, I have just exchanged the DRAM of Jake's box
22:54:15amiconn[IDC]Dragon: You could try setting burst & RAS down mode in your RAM test, then try again...
22:54:43[IDC]DragonI'll try the other RAM when I get back home
22:54:58undefinedassasinPlease help me
22:55:00undefinedassasinmidk is gone
22:55:05undefinedassasinand someone is asking me for him
22:55:10undefinedassasinbut I can't help
22:55:12amiconn[IDC]Dragon: Still at work?
22:55:29[IDC]Dragonamiconn: only for the solder tools
22:55:35[IDC]Dragonleaving now
22:55:46amiconnAhh
22:56:06 Part undefinedassasin ("X-Chat [2.0.10c] Quit.")
22:56:25[IDC]Dragonehatch the sky, we have the Perseides at max now
22:56:32[IDC]Dragonwhatch
22:57:00[IDC]Dragon(meteor storm)
22:57:39amiconnwatch even :)
22:57:48*[IDC]Dragon leaves for falling star gazing
22:58:15[IDC]Dragonyes, watch out for my reckless typing
22:58:41 Quit [IDC]Dragon ("no fate but what we make")
23:00
23:54:36 Join [IDC]Dragon [0] (~idc-drago@pD9E34A79.dip.t-dialin.net)
23:55:40[IDC]DragonJust for the record: the behavior of Jake's box is the same with the new RAM chip. :-(
23:56:38amiconnDid you try your RAM test with burst & RAS down enabled?
23:57:12[IDC]Dragonwhy should I? there's a different RAM chip in in now.
23:58:02[IDC]Dragonin it

Previous day | Next day