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 2010-06-29

00:01:19 Quit kachna (Ping timeout: 260 seconds)
00:01:25 Quit stripwax (Read error: Connection reset by peer)
00:04:19 Join stripwax [0] (
00:05:11 Quit stripwax (Client Quit)
00:08:33 Join wombat23 [0] (
00:08:49 Quit efyx (Remote host closed the connection)
00:18:27 Quit Luca_S (Quit: CGI:IRC)
00:21:07 Quit evilnick_B (Quit: Page closed)
00:25:10 Quit TheSeven (Ping timeout: 252 seconds)
00:29:44 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
00:37:05 Quit petur (Quit: Leaving)
00:39:47CIA-6New commit by 03alle (r27166): Simplify the 'blind' macro and imrove spacing where it's used
00:40:38 Quit ender` (Quit: Connection Reset by Gypsies with Wire Cutters)
00:41:32CIA-6r27166 build result: All green
00:42:05 Quit bertrik (Quit: :tiuQ)
00:42:30***Saving seen data "./dancer.seen"
00:42:40 Quit jgarvey (Quit: Leaving)
00:48:29 Quit [blue] (Quit: leaving)
00:54:27 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
00:54:29 Join kramer3d [0] (~kramer@unaffiliated/kramer3d)
00:55:36 Quit domonoky1 (Ping timeout: 265 seconds)
01:04:11 Part domonoky
01:16:47 Quit dfkt (Quit: -= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.)
01:18:22 Quit MethoS- (Remote host closed the connection)
01:19:15 Join MethoS- [0] (~clemens@
01:20:04 Quit DerPapst (Quit: Leaving.)
01:35:00 Quit Rob2222 (Quit: Rob2222)
01:47:52 Quit GeekShadow (Quit: The cake is a lie !)
01:48:03 Quit kramer3d (Read error: Connection reset by peer)
01:48:28 Join kramer3d [0] (~kramer@unaffiliated/kramer3d)
01:50:03 Quit Zigtown (Remote host closed the connection)
01:57:03 Join kramer3d_ [0] (~kramer@unaffiliated/kramer3d)
01:57:17 Quit kramer3d (Disconnected by services)
01:57:22 Nick kramer3d_ is now known as kramer3 (~kramer@unaffiliated/kramer3d)
01:57:24 Nick kramer3 is now known as kramer3d (~kramer@unaffiliated/kramer3d)
02:02:48 Join JdGordon_ [0] (6e15015a@gateway/web/freenode/ip.
02:20:49 Quit r0b- (Ping timeout: 248 seconds)
02:27:44 Quit kramer3d (Read error: Connection reset by peer)
02:28:08 Quit rado1 (Ping timeout: 240 seconds)
02:28:10 Join kramer3d [0] (~kramer@unaffiliated/kramer3d)
02:28:39 Join binaryhermit [0] (
02:28:47 Quit simonrvn (Read error: Connection reset by peer)
02:30:33 Join simonrvn [0] (
02:33:32 Quit mc2739 (Ping timeout: 260 seconds)
02:33:36 Join chai16 [0] (
02:34:51chai16i have ipod 5g 30gb and when i plug to computer usb, it never stays on "multimedia" keypad it always just mounts as a usb volume
02:35:04chai16what setting do I need to change?
02:36:19 Quit JdGordon_ (Ping timeout: 252 seconds)
02:42:32***Saving seen data "./dancer.seen"
02:42:46S_a_i_n_tchai16: What version of Rockbox are you running?
02:43:04S_a_i_n_tA release build or one of the daily builds?
02:43:46 Join rado1 [0] (
02:45:49chai16uh, i just used the installer and hit complete install
02:46:12S_a_i_n_tYes, but did you choose a current build, or a release build?
02:46:31 Join Zigtown [0] (
02:46:42 Join r0b- [0] (~nnscript@
02:47:17S_a_i_n_tHmmmm...a current build should definitely be using Rockbox's USB mode.
02:47:45S_a_i_n_tYou're *positive* it's not 3.5/3.51 or 3.6?
02:47:56chai16oh nvm
02:48:01chai16it's 3.6
02:48:08S_a_i_n_tThere's the problem.
02:48:27chai16ok so just reinstall it with the daily build?
02:49:12S_a_i_n_tThe release builds use the Apple disk mode as charging isn't implemented properly. Charging is fully implemented in current SVN, so, yes...install a current build and it will use Rockbox's USB mode.
02:49:54chai16great, thanks. i chose archived build instead of the "up to the minute build"
02:49:57S_a_i_n_tPropper charging support for the iPods was implemented some time after 3.6 was released.
02:50:54S_a_i_n_tWhen 3.7 is released, then you can choose "Release" build and it will use the Rockbox USB mode instead of the Apple disk mode.
02:51:23S_a_i_n_tBut, for now, use the current build. It is plenty stable enough for everyday use.
02:54:50 Quit antil33t1 (Ping timeout: 276 seconds)
03:09:22 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
03:32:35 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
03:33:07 Quit Zarggg (Read error: Connection reset by peer)
03:34:53 Join Zarggg [0] (
03:45:50 Join Rob2222 [0] (
03:55:42 Quit MethoS- (Remote host closed the connection)
03:57:51 Quit Rob2222 (Quit: Rob2222)
04:06:28 Part chai16
04:06:56 Join komputes [0] (~komputes@ubuntu/member/komputes)
04:18:48 Join JdGord [0] (~jd@
04:22:42 Quit pixelma (Disconnected by services)
04:22:44 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:22:45 Quit amiconn (Disconnected by services)
04:22:47 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:23:00 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:23:05 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:28:13 Quit z35 (Ping timeout: 265 seconds)
04:30:41 Join z35 [0] (
04:34:17JdGordS_a_i_n_t: what's up with %pE?
04:40:19 Join R31D [0] (~email@
04:42:36***Saving seen data "./dancer.seen"
04:48:28 Quit TheSeven (Ping timeout: 260 seconds)
04:49:37 Join Barahir_ [0] (
04:52:23 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:52:54 Quit Barahir (Ping timeout: 260 seconds)
04:53:27 Quit sbhsu (Ping timeout: 260 seconds)
04:53:44 Join sbhsu [0] (
04:57:27 Join Rob2222 [0] (
04:58:18 Join Buschel [0] (
05:07:16 Quit Buschel (Ping timeout: 240 seconds)
05:07:33 Quit JdGord (Quit: Bye)
05:14:23 Quit binaryhermit (Ping timeout: 260 seconds)
05:20:21 Quit rado1 (Quit: Leaving.)
05:24:40 Join JdGordon_ [0] (6e15015a@gateway/web/freenode/ip.
05:26:42 Quit z35 (Ping timeout: 265 seconds)
05:27:58 Join Strife89DS [0] (
05:30:35 Quit Horscht (Quit: Verlassend)
05:30:49 Quit MagusG (Ping timeout: 260 seconds)
05:38:56 Join z35 [0] (
05:54:25 Quit scorche (Disconnected by services)
05:54:34 Join scorche` [0] (~scorche@rockbox/administrator/scorche)
06:01:24S_a_i_n_tJdGordon: I can't really talk right now, but I'll give you the basics of it. "%?pE<A|B>" works fine, but "%?pE<%s%ac%Sx(Next Track:)|%?it<%s%ac%it|%s%fn>>" fails.
06:01:37S_a_i_n_tIt only shows the "false" condition.
06:02:26S_a_i_n_twhen it is "true", it either fails to update, or just doesn't switch. It will always show the "false" condition.
06:02:41S_a_i_n_tIt seems to be a problem with nesting conditions inside of it.
06:02:54 Quit panni_ (Quit: ( :: NoNameScript 3.81 :: ))
06:02:57S_a_i_n_t("it" being %?pE)
06:04:43JdGordon_extraordinary! :p
06:04:55JdGordon_-a maybe?
06:05:16S_a_i_n_tnope, you git it right ;)
06:05:26S_a_i_n_thahaha...classic. "got"
06:06:20S_a_i_n_tAs I said earlier, I know you have a lot on your plate with the new parser etc. etc. but if you could look into it at some point in time I'd appreciate it.
06:07:37S_a_i_n_tThe fix I had to do to to achieve the same effect really isn't pretty. I had to use conditional viewports in a situation where I really should have only one (non-conditional) viewport, with the info inside it being conditional to %?pE.
06:08:13JdGordon_replace %pE with %mh... if that doesnt work then something is seriously wrong
06:08:19S_a_i_n_tbut *nooooooo*, it decides to be a dick about it and now work for me :P
06:08:23JdGordon_if it does then that is just odd!
06:09:05S_a_i_n_tHmmm..ok, I'll experiment with it a bit more later on. A little busy right now.
06:09:08S_a_i_n_tI'll let you know.
06:09:15S_a_i_n_tAnd, thanks.
06:10:33 Quit Zigtown (Quit: Leaving...)
06:15:35JdGordon_you know the %pE with no number has a tiny timeout? 1s I think?
06:19:14 Join CGL [0] (~CGL@
06:20:14 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
06:22:37 Join bieber [0] (
06:42:38***Saving seen data "./dancer.seen"
06:44:33bieberJdGordon: In conditionals, the branches that aren't displayed still need to be evaluated?
06:44:58bieberErr, JdGordon_
06:45:13JdGordon_not really... images and a few other tags might need to be disabled
06:45:29JdGordon_if you store the last value you can easily only eval that branch
06:46:21bieberSo some tags do work even in unselected branches, but others don't?
06:47:03JdGordon_they dont work, but they might need to do something to be turned off
06:47:24JdGordon_peakmeter, images (the area gets cleared), i think that might be it
06:48:27bieberSo if I just ignore the unselected branch for now, I should still get something that's pretty darn usable, no?
06:48:54bieberI just want to get something together that more or less works first, and then I'll go back and smooth it all out
06:50:21 Join rado1 [0] (
06:52:54 Quit JdGordon_ (Ping timeout: 252 seconds)
06:54:50 Quit Strife89DS (Quit: Bed.)
06:55:33 Quit scorche` (Read error: Connection reset by peer)
06:55:35 Join scorche [0] (~scorche@rockbox/administrator/scorche)
06:56:30 Join JdGordon_ [0] (6e15015a@gateway/web/freenode/ip.
06:57:36 Quit kramer3d (Quit: Leaving)
06:57:48saratogamaybe we should disable rebooting into the OF on USB insert from within rockbox on the fuzev2 until its understood why it sometimes hard locks the player
07:07:22 Quit scorche (Disconnected by services)
07:07:33 Join scorche [0] (~scorche@rockbox/administrator/scorche)
07:09:15pixelmaamiconn: but MazezaM uses the now simplified pluginlib action (only one context, directions, select/long select, exit) and it only chacks for IPOD_3G_PAD. If this needs extra-handling then it sounds to me like the chasen mapping for the actions is wrong. And asn't the 3G pad define used on 1st/2nd gen Ipods as well currently?
07:09:40pixelmachecks too
07:10:14 Quit scorche (Read error: Connection reset by peer)
07:10:29 Join scorche [0] (~scorche@rockbox/administrator/scorche)
07:11:31pixelmaand chosen... * needs more coffee *
07:11:59 Quit JdGordon_ (Ping timeout: 252 seconds)
07:12:55 Join esperegu [0] (~quassel@
07:15:41 Join JdGordon_ [0] (6e15015a@gateway/web/freenode/ip.
07:16:41saratogais there a way to get the SVN changes since 3.4?
07:16:50saratogavia the web interface
07:18:22S_a_i_n_tJdGordon: JdGordon_: Actually, the default timeout for %pE is 10 seconds.
07:18:24JdGordon_do you want one file? or the whole lot?
07:19:15S_a_i_n_tAnd %pS (or, whatever the other one is)
07:19:54S_a_i_n_tI thought of that myself, I was doing %pE(10) untill I realised it was irrelevant :p
07:19:55 Quit sevard (Ping timeout: 260 seconds)
07:20:23 Join sevard [0] (sev@
07:23:24CIA-6New commit by 03bieber (r27167): Theme Editor: Rearranged and separated tag execution and rendering code
07:25:11CIA-6r27167 build result: All green
07:28:15 Join LinusN [0] (linus@rockbox/developer/LinusN)
07:41:18 Quit scorche (Read error: Connection reset by peer)
07:56:44 Quit Zarggg (Ping timeout: 240 seconds)
07:57:06 Join GodEater [0] (
07:57:06 Quit GodEater (Changing host)
07:57:06 Join GodEater [0] (~bibble@rockbox/staff/GodEater)
08:02:28 Join Buschel [0] (
08:06:00 Join Zarggg [0] (
08:09:31 Join scorche [0] (~scorche@rockbox/administrator/scorche)
08:14:02 Join ender` [0] (
08:14:38CIA-6New commit by 03bieber (r27168): Theme Editor: Beginning to work on conditional rendering, just made fix to line following code to speed up horizontal scrolling
08:15:48 Join Meekrab [0] (
08:16:17CIA-6r27168 build result: All green
08:17:24MeekrabThrough an exploit in the game menu on my Zune 30, I got a WinCE window to open with a few words of choice
08:17:51 Quit BHSPitMonkey (Remote host closed the connection)
08:20:11bieberJdGordon_: Is there a list somewhere of all the tags that scale their range of values to the number of conditional branches? Is it just all the tags that return percentage values?
08:21:10MeekrabIts running a winCE executable hello world I made in VC
08:23:34Meekrabbut theres a slight problem
08:23:45Meekrabits obscured by the zune loading pulse...
08:26:43MeekrabAnyone want to help pick up the pace in the zune rockbox deployment?
08:27:19Meekrabafter all, the only thing really stopping me from tearing apart the code like I did on my borrowed gigabeatS is the encryption
08:27:33Meekrabbut I found a bug in the XNA development studio
08:29:21 Quit TheSeven (Ping timeout: 245 seconds)
08:33:29JdGordon_bieber: I doubt there is a list, and it very likely isnt even all the percent ones
08:33:41JdGordon_I know for sure %pv does, no clue about the other ones
08:34:36bieberWoo, my first run at conditional rendering is turning out horribly
08:35:56MeekrabI am seriously astonished that I would find such an interesting bug in such an odd place
08:36:07Meekrabwhy not just develop a game as a rockbox loader?
08:37:42MeekrabI'm sad...I don;t know arm coding that well, X86 and Mips are my forte
08:38:38bieberOh wait, I'm displaying images in the viewport they're declared in, not the one they're displayed in
08:38:49bieberThat's gotta be what's screwing things up the most :/
08:39:09bieberHey JdGordon_, can images ever be displayed in more than one viewport?
08:39:35Meekrabnot trying to be mean, but is there anyone really caring about the zune?
08:40:00JdGordon_each viewport pass marks all images as disabled, then in each viewport any %xd will draw that image in the current viewport at the viewport offset
08:40:02MeekrabI wish I could find some people with the skills I have to build a working rockbox port not that I found away around the executable problem
08:40:24JdGordon_Meekrab: i dont tihnk any rockbox dev owns a zune, so getting help is going to be tough
08:41:07MeekrabWell, I have 6 working zune 30s right here
08:41:15Meekraband all they need is a format and some USB cables
08:42:37Meekrabif any willing dev wants one, the shipping weight is 0.4kg
08:42:41***Saving seen data "./dancer.seen"
08:43:09 Quit literal (Remote host closed the connection)
08:45:49Meekrabor what about the leap year overflow bug? that otta hold some water
08:46:50 Quit amiconn (Disconnected by services)
08:46:52 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
08:46:57 Join literal [0] (
08:46:57 Quit pixelma (Disconnected by services)
08:46:59 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
08:47:10 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
08:47:14 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
08:47:47Buschelwodz: thanks for testing
08:48:47Buschelcan anyone with a MCF5250 and a S5L870X CPU test FS #11445. I expect a large speed up of aac decoding with the MCF5250.
08:50:37 Quit JdGordon_ (Ping timeout: 252 seconds)
08:50:39amiconnMCF5250 will differ from MCF5249?
08:51:05bieberMeekrab: I've gotta ask, how did you end up with 6 Zunes?
08:51:20bieberDid your friends all buy them and subsequently hate them? :P
08:51:39amiconnAh, more iram usage
08:53:46 Quit Buschel (Ping timeout: 264 seconds)
08:54:09 Join Guest58341 [0] (
08:54:09 Quit Guest58341 (Changing host)
08:54:09 Join Guest58341 [0] (bjst@rockbox/developer/Zagor)
08:55:33 Nick Guest58341 is now known as Zagor (bjst@rockbox/developer/Zagor)
08:57:49bieberI've got all kinds of conditional image display going on in the Theme Editor's preview window
09:01:09Meekrabbeiber: Chinese refurbs are cheaper then brand new ones, $20 USD versus $60 for a brand new one
09:01:36Meekraband most of the time they're just open box or scratched.
09:04:15bieberI see
09:05:43 Join flydutch [0] (
09:11:26 Quit rado1 (Quit: Leaving.)
09:14:50 Join JdGordon_ [0] (6e15015a@gateway/web/freenode/ip.
09:15:42CIA-6New commit by 03bieber (r27169): Theme Editor: Implemented conditional rendering, most conditionals should work correctly now
09:16:22JdGordon_bieber: should I keep working on the assumption you wont need the callback?
09:16:42bieberYeah, I don't think I'll use them
09:16:55bieberAlthough I must admit, I still haven't figured out exactly what it all does
09:17:14JdGordon_and did you get a chance to go over the parser change diff? (skipping a branch in the conditional)
09:17:14CIA-6r27169 build result: All green
09:17:45JdGordon_it looks like it works but I'd like a second opinion :)
09:17:47bieberIf you have surplus free time, feel free to explain it and save me the time checking out the code ;) I'm guessing you're doing something like attaching a function pointer to tags to be called when they're encountered?
09:18:05bieberI'll apply the patch and make sure it still works, as long as it doesn't break anything it should be fine
09:18:59JdGordon_I'm using a single callback that happens for each element to attach something to the elements data element...
09:19:42JdGordon_trying to figure out the best way to integrate the parser in... I've mostly given up on walking the tree and converting to a token array because it apparently doesnt make things easie
09:21:10JdGordon_but the problem with doing that is the parser needs to be told to ignore branches if the target s=doesnt have the feature so unknown tags dont fail
09:21:18JdGordon_which is why that branch skipping bit is needed
09:22:18 Join petur [0] (d408b802@rockbox/developer/petur)
09:22:43bieberTags are defined independently for different targets, I take it?
09:23:58bieberHmm, I'm seeing a slight difference, I don't know why though
09:24:30JdGordon_maybe 1% of tags change depending on the target (that includes the few which outright artn accepted)
09:24:41bieberIt's making all my conditional tags render one value ahead of what they were before
09:25:03bieberi.e. Stop is rendering as Play, Play is rendering as Pause, etc.
09:26:30JdGordon_when was before?
09:27:08bieberOn the last commit I made
09:27:39bieberI just realized that I forgot to set that flag to true, it's working as normal now
09:28:10JdGordon_ah with my patch?
09:28:10bieberSo as long as you make sure that feature_available is always true when it's being called from the theme editor, it won't ruin anything
09:28:24JdGordon_yeah, obviously true will be the default
09:29:10bieberOkay, looks like it's working fine, then :)
09:29:31bieberOn a sidenote, I had no idea the standard C library had a bool type
09:29:43CIA-6New commit by 03jethead71 (r27170): Gigabeat S: The forced shutoff voltage level can be lowered a tiny bit it seems.
09:31:18JdGordon_:) dont go changing the parse function return values to them... I'm going to use the ints later :)
09:31:31CIA-6r27170 build result: All green
09:34:08bieberNo worries, I don't plan on touching the parser unless I find something broken
09:34:16bieberOr you go changing the theme language again ;)
09:35:23JdGordon_would I do that :D
09:36:49JdGordon_and you shold get more blame for changing the langauage than me :D
09:37:08 Quit Meekrab (Quit: I love my HydraIRC -> <-)
09:41:27bieberHaha, fair enough
09:41:51bieberbtw, is there any particular reason that the Theme Site has themes in the new format for some targets, and the old format for others?
09:42:43JdGordon_the theme site is two now.. one with the old format and one with the new.... there shouldnt be any old types on the site
09:43:55 Quit GodEater (Quit: leaving)
09:44:16bieberiirc, I stumbled upon one target (it was one of the touch screen targets) that had themes in the new format
09:44:23bieberI've been using them for testing
09:44:43 Join n1s [0] (~n1s@rockbox/developer/n1s)
09:46:24 Join GodEater [0] (
09:46:25 Quit GodEater (Changing host)
09:46:25 Join GodEater [0] (~bibble@rockbox/staff/GodEater)
09:46:59bieber320x240 screen
09:47:10 Join Buschel [0] (
09:47:15 Join Rob2223 [0] (
09:48:04bieberOkay, it's the Cowon D2. On the normal theme site, its themes are in the new format, for some reason
09:48:12jhMikeSMeekrab: is it free Zune giveaway to devs day?
09:50:18jhMikeSoops, Meekrab already split
09:50:30 Quit Rob2222 (Ping timeout: 252 seconds)
09:51:43pixelmabieber: on the "normal" theme site all target's themes should be in the new format. If you want old ones, there is also an "old themes" page
09:52:01bieberOh, okay
09:52:10pixelmaneeded as the last release was still with the old format
09:52:38bieberOh, woops, I read JdGordon's message in reverse
09:53:23pixelmayou should read it from beginning to end ;)
09:55:19 Join efyx [0] (
09:56:37 Quit JdGordon_ (Ping timeout: 252 seconds)
10:08:39 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
10:08:40 Quit Zarggg (Read error: Connection reset by peer)
10:11:50 Join swilde [0] (
10:21:04 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
10:36:28 Join Mich [0] (
10:38:16 Join DerPapst [0] (~Alexander@
10:40:10MichI was just goofing around on my DAP tried to open my test mpg, it locked my DAP up. so I started rolling back daily builds til I hit r27147 and it started working again. I look at svn from r27147 to r27162 and I don't see anything that jumps out at me that might cause the problem.
10:41:19 Quit Buschel (Ping timeout: 276 seconds)
10:41:59MichIt didn't lock up trying to play the file, it locked up before it got to the mpeg-player menu.
10:42:01pixelmawhich player do you have?
10:42:42***Saving seen data "./dancer.seen"
10:44:43pixelmamaybe 27151?
10:45:09pixelmait changes something in the c200's lcd driver
10:45:55pixelmahmm, only for the v2s if I see correctly
10:47:18 Join bor_ka [0] (
10:47:54Michyeah i have a v1
10:49:24pixelmabut if you can build yourself, it would help if you find out the revision that broke it by trying and binchopping (taking last known working revision and first known broken - build a revision that's in the middle between the too, try - repeat till you can pinpoint the one commit that broke it)
10:49:24*jhMikeS was unable to use mpegplayer on his Fuze v2 (hard freeze)\
10:49:46 Quit TheSeven (Ping timeout: 276 seconds)
10:50:23 Join kachna [0] (
10:50:48MichI have a linux system, ill work on getting build env setup
10:51:54bor_karanma, hello! I've tested your verify-read patch, it panicks with the mSD card: that is drive = 1, and on = 0.
10:52:20jhMikeSmpegplayer froze my gigabeat S just now (guess there really is a problem...perhaps the libmad stuff)
10:52:44bor_kawill try it now without the mSD. But the unpatched first - since these glitches could be caused by the mSD
10:53:08bor_kalibmad.... is it mp3?
10:53:29n1sbor_ka: yes
10:53:52n1sused for decoding audio in mpegplayer as well as the mp3 codec
10:53:55Michmpeg-player started working again back to r27147
10:53:56bor_kaI have glitches with mp3 playback, can it be caused by the recent changes in the libmad?
10:54:33bor_kaThat is, both "my" mp3 glitches, and Mich's mpegplayer
10:54:53n1sbor_ka: it could, what type of glitches are they and when did they show up
10:55:22bor_kan1s, when playing mp3, there is part about 1 sec from another track
10:55:40bor_kadon't know, though, if such a glitch can be caused by the codec
10:55:58bor_kait is more likely to low-level reading problems
10:56:18n1swhen is this 1 sec played and is it from the previous track or something else?
10:56:28pixelmajust wanted to suggest Buschel's libmad changes in r27153
10:57:12bor_kafrom another album - since I get woman's vocal where it shouldn't be at all
10:57:13pixelmaand following
10:57:15jhMikeSmpegplayer uses things the core codec doesn't
10:57:27n1sbor_ka: could you try a build from before r 27153 and see if that works ok?
10:58:20n1sbor_ka: that sounds like disk corruption, but a test wouldn't hurt. yes r27152 would work for testign
10:58:51bor_kaBTW, they appeared between 26581, that is Ok, and 27128
10:59:03bor_kathere were 4 libmad changes
10:59:30bor_kaso it is not 27153's change
10:59:32n1sbor_ka: so if you install r 26581 it's ok now?
10:59:39bor_kan1s, yes
10:59:49bor_kaand if I install 27128 - broken
11:00:05jhMikeSyep, tried again...frozen beast
11:00:09n1sok, would be nice if you could work out which change broke it and report a bug
11:00:15bor_kamore - if I disable 4bit widebus - by ranma - in 27128, it is Ok
11:00:52bor_kaBut he doesn't understand how it can affect reads...
11:01:06bor_kaSo, libmad's changes were just a guess
11:01:12n1soh, this is an unstable port, yeah, those read problems are much more likely to cause your problem
11:01:52n1sthe same problem has happened when a players' disk got corrupted by some buggy usb transfers etc
11:01:57jhMikeSbor_ka: it's ok? nothing to do with gigabeat S, the 4-bit bus and mpegplayer worked there not many revisions ago
11:02:37bor_ka4bit bus caused disk corruption, not only USB, but rockbox own wirites as well
11:03:34bor_kajhMikeS, I have fuze v1, so, I think it is not related to gigabeat
11:05:11jhMikeSthere's definitely something up though, could be my end ... time to go checking that
11:06:45 Join einhirn [0] (
11:15:41ranmabor_ka: Ok, didn't try the read debug patch with microsd, just internal
11:18:18 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
11:19:48ranmabor_ka: BTW, is the glitch random or always the same, i.e. always at the same part of that mp3?
11:20:33bor_karanma, got read verify failed: sec=14578223 n=10
11:21:00bor_kabtw, sometimes there is panic when booting, sometimes not, panic is not related directly to mSD
11:21:08bor_kabut always with on == 0
11:22:59bor_kaas for same part of mp3 - not 100%, but, I'd say that I've got one song, where there are always glitches, plus-minus 20 secs, and got an album that plays ok
11:24:16bor_kareboote, the same song - read verify @ 14578340, n=11
11:25:22bor_karebooted again, another song, @ 11968039, n = 10
11:25:41bor_kait happens in about 1 sec after the playback start
11:26:44bor_kastarted flac - all plays ok
11:26:53*bor_ka understands nothing at all
11:27:19bor_kamay be it libmad issue, it corrupts some internal structures?
11:33:24 Quit [keanu] (Ping timeout: 240 seconds)
11:34:37bor_kaflac, playback ~10 min - no read verify errors
11:37:15ranmaMay be interesting to try half the clock speed. I.e. replace "MCI_CLOCK(drive) = MCI_HALFSPEED;" with "MCI_CLOCK(drive) = MCI_QUARTERSPEED;" if it's the internal drive
11:38:58ranmaAlso, it would be interesting to stay in 4bit mode for the verify, to see if a second run at 4bit will return a different result too. The controller should be checking data crcs internally, but who knows...
11:39:23bor_karanma, wavepack plays Ok too....
11:39:47jhMikeSr27159 breaks mpegplayer on gigabeat S (confirmed)
11:39:48ranmaOTOH since it's from a completely differnt part of the disk, most likely the sector number gets messed up somehow...
11:39:58bor_kaI think I try to build without thumb - may be libmad was not fixed correctly for the thumb support
11:40:19ranmabor_ka: I doubt it's a libmad issue
11:40:41ranmaYou could try without thumb, sure.
11:40:53ranmaBut thumb seems to work just fine on my C200v2
11:41:19bor_kahmm, you are right, I've started mp3 that is "close" to the wavepack-files, all is Ok
11:41:19ranmaWhat was your player again? I think it's probably not even compiled as thumb on yours?
11:41:28bor_kafuze v1
11:42:27ranmaThat's not compiled thumb unless you request it
11:43:11bor_kaand yes, when I selected mp3 that is close to previous, that aborts, it aborted too..
11:43:21bor_kaok, trying without thumb first
11:43:59ranmabor_ka: Do a "grep thumb Makefile" in your build dir
11:44:09ranmaIf it doesn't match it isn't compiled as thumb
11:44:22bor_kaI've reconfigured already :)
11:44:53bor_kaIt was compiled as thumb becasuse I've manually instructed it :)
11:46:41*ranma wonders if the command argument is crc protected
11:47:15ranmaHmm, yes. Should be crc protected.
11:47:22bor_karanma, I have a patch from flyspray, so in any case I've got a "M" build
11:48:06bor_kapatch for tagcache, so it shouldn't affect playback and crashes
11:48:31ranmaWell, that you get a verify failure is definitively a problem that shouldn't happen.
11:50:31ranmaAlso, in the case where the set_speed panics, what does it print as values for ret and response? (if you used the updated patch)
11:50:41bor_kaAnd I've enabled thumb, since it is enabled for AMSv1 with 2GB - I don't see, why 8GB can't use it
11:51:20bor_kaok, when I get to speed - I will not it
11:51:28bor_kaI will note it
11:51:31ranmaYou gain very little on 8MB and thumb is supposed to be a bit slower (but can be faster)
11:53:19jhMikeSreverting the changes in just frame.c (the zeroing in mad_frame_mute) seems to fix it
11:53:19ranmaI think it would be very interesting to add some stats about crc errors and show those in the debug mneu
11:53:20bor_kaI thought that lower memory footprint -> less disk acess -> longer playback time, but I may be wrong
11:58:17bor_karanma, build withou thumb seems to work Ok - will test for some time with your patch, though
11:58:43bor_kashould we find thumb problem, or just define that thumb is a no-go for fuze ?
11:59:15ranmaWell, _if_ it is a thumb problem it would be nice to find it ;)
12:00:20 Join dfkt [0] (~dfkt@unaffiliated/dfkt)
12:01:24bor_kaOk. What you think - 1st test several reads in the 4bit mode, then test lower clock speed? In that order?
12:01:56ranmaI'd say first try if the verify in 4bit mode also detects the errors.
12:03:14ranmaThen try the patch I'm working on now, which will keep some statistics about errors. If it happens because of a corrupted cmd not caught by the crc checksum, a lot of crc failures should show up there.
12:03:16bor_kabtw, why is there a 100% panic with set_widebus off:tran, when using mSD?
12:03:44ranmaI'm wondering about that too, I should check in my player with uSD
12:04:02*bor_ka will be back in a hour
12:07:06CIA-6New commit by 03jethead71 (r27171): r27159 broke mpegplayer but not the mpa codec because its libmad init wasn't also updated to be compatible with the changes. Fix that.
12:09:08CIA-6r27171 build result: All green
12:21:36 Nick fxb__ is now known as fxb (
12:34:22ranmabor_ka: I can reproduce the set_widebus erron on uSD. SD_SEND_STATUS in failed (return value -1), but the second seems to have worked and indicates state 5 (data), ready for data, and csd overwrite error (response 0xB00)...
12:40:35 Join JdGord [0] (~jd@
12:42:43***Saving seen data "./dancer.seen"
12:44:05 Quit TheSeven (Ping timeout: 240 seconds)
12:56:10 Quit mt (Read error: Network is unreachable)
12:56:46 Join mt [0] (~mtee@rockbox/developer/mt)
13:03:36ranmaHmm, I'm seeing a problem again where (invalidate|dump)_dcache_range doesn't do as it should.
13:05:38CIA-6New commit by 03uchida (r27172): text_viewer: callback functions are changed to the function that returns int value. ...
13:06:31ranmaThat might also explain bor_ka's problem, as 4bit mode is faster and in 1bit mode another thread might trash the cache during the wakeup_wait, but doesn't have time to run in 4bit mode.
13:06:43 Join ucchan [0] (
13:07:14CIA-6r27172 build result: All green
13:08:08 Join Ingrid_en_Henk [0] (
13:11:02ranmaI have found one file on microSD, that if I try playing it after boot, I reliably get a read verify failed.
13:11:26 Join anewuser [0] (anewuser@unaffiliated/anewuser)
13:11:34ranmaIf I look at the memory contents via jtag, both buffers are identical, still memcmp failed => cache issue
13:12:04ranmaUsing invalidate_dcache() makes it work, I just don't get why the _range version fails.
13:16:40bor_karanma, ok. So no point now in testing two consecutive reads in 4bit mode now? Or?
13:17:16 Join Luca_S [0] (
13:17:59Luca_Scaching issues could also explain the random freezes that happen (rarely) on amsv2, I suppose
13:18:23Luca_Ssince they wouldn't be limited to sd access, isn't it?
13:18:50bor_kaLuca_S, yes, I've got these issues with the internal memory
13:22:23ranmaHmm, weird. If I add 'dump_dcache_range(aligned_buffer, transfer*512)' before the sd_transfer_sectors call in the verify path it works, but sd_transfer_sectors should already do that...
13:22:33 Join Meekrab [0] (
13:23:13MeekrabAfter much screwings with my zune, I have produced a working way to send a working Windows CE ARM native app to my device
13:23:18ranmabor_ka: Can you try if replacing the 'dump_cache_range(..)' with 'invalidate_dcache()' makes your problem go away?
13:23:21 Join panni_ [0] (
13:23:46Meekrabnow, if I could only figure out how to run the rockbox GigabeatS port on it
13:23:55ranmabor_ka: My uSD card doesn't like switching either BTW, it just stops responding to commands...
13:24:17Meekrabanyone want to see it running CEFUN?
13:24:58ranmaI'd say, remove the two sd_set_buswidth from the verify path, change the dump_dcache_range to invalidate_dcache and try if it works
13:26:27MeekrabWjho's in charge of the GigabeatS project?
13:27:50n1sMeekrab: i'd say jhMikeS although we don't have any such official positions
13:27:53 Quit Luca_S (Quit: CGI:IRC)
13:28:01 Join Luca_S [0] (
13:28:29MeekrabIm so freaking excited
13:28:29ranmabor_ka: Hmm, wait. Still something wrong there. *grr*
13:28:40MeekrabI found a bug in the games menu
13:29:25Meekrabone that you could drive a mac truck through
13:30:05Meekrabbut that stupid loading logo freezes up the controls
13:30:30n1sMeekrab: does it look like a bootloader or complete alternative firmware could be installed and run directly without exploiting the Original Firmwae?
13:31:18jhMikeSn1s: I'm in charge? Do I get an office on the top floor by the window overlooking the bay?
13:31:41gevaertsjhMikeS: yes
13:31:42n1sjhMikeS: what, you haven't gotten your office yet!?
13:31:49gevaertsYou don't get to specify which bay though
13:31:59MeekrabI used a small game loader bug, so its a sploit of sorts
13:32:05jhMikeSnope, just this little room (not the ebay either :)
13:32:23MeekrabI'm using freely available deployment tools to build and compile my files
13:32:46bor_karanma, stopping...
13:33:11jhMikeSI did have a dump of the internal i.MX31 boot roms (somewhere), not sure if that would help get at anything
13:34:18MeekrabIn the GigabeatS's files...what would be a good file to try to load into the machine? I tried NK.EXE and shell.exe
13:34:51pixelmahasn't there been some other project revealing something about the Zune? I think I saw a link semi-reecently
13:35:11MeekrabI pulled a copy of Energy Bliss 240x320 from an old zune firmware....coincidence?
13:35:53ranmaIt doesn't make any f-ing sense at all. If I have a single invalidate_dcache() before the while(count) in sd_transfer_sectors I get a verify failed
13:36:20ranmaBut that should be more than enough unless I'm missing something stupid
13:36:38jhMikeSMeekrab: not sure, you'd want to load a gigabeat S's stuff onto a Zune? maybe try the bootloader itself? (eboot.bin)
13:36:46ranmaIf I instead add it before the call to sd_transfer_sectors() it works
13:36:58MeekrabFnarg, all of my old windows CE programming stuff is in C++
13:36:58MeekrabHow do I get it to become C# files
13:36:58Meekraboh...duh theres a "Import" option
13:37:21ranmaEven adding dump_dcache_range() there instead of the big invalidate_dcache() hammer works also.
13:37:45bor_karanma, timing issue?
13:37:52MeekrabWelp, with the game exploit, theres a way to run my helloworld.exe, a few simple dialogs, and etc
13:38:00CIA-6New commit by 03uchida (r27173): text viewer: move and delete #include.
13:38:21pixelma I can't tell how useful it is though
13:39:34CIA-6r27173 build result: All green
13:40:50 Join teru [0] (
13:40:56 Join wodz [0] (
13:42:16MeekrabThe last wince device I messed with was X86 based and had its NK.bin and etc on a compactflash lol
13:42:20Meekrabthat was easy
13:42:29 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
13:42:36Meekrabnow i'm working with an encrypted system and whooole bunch of new ARM code I dont know
13:42:53bor_karanma, is it correct, that if you add dump_dcache before sd_transfer_sectors, then in fact it is called twice in a row - second in the if(aligned) ... branch
13:43:25*ranma has an inkling...
13:44:23wodzare there any informations about keymap files other than files itself?
13:44:51 Quit shai (Quit: Leaving)
13:45:10MeekrabIf I only knew how to tell it to go into disk mode so I could dump out the firmwares contents
13:45:12jhMikeSMeekrab: ARM's pretty simple, and you might want to familiarize yourself then :)
13:45:25 Join shai [0] (
13:45:28ranmaHmm, or not...
13:45:43MeekrabOn a side note: Would rockbox for PSP be worth the effort?
13:46:14Meekrabmain advantages would be solid documented hardware
13:46:42wodzwhat this setting do? global_settings.hold_lr_for_scroll_in_list ?
13:46:43S_a_i_n_tWe already had one guy talk (and he was all talk, and no action) about a PSP port.
13:46:50S_a_i_n_tIMO, it'd be no use at all.
13:47:09S_a_i_n_tRaaA for PSP however.
13:47:47MeekrabWell, I know a fair bit about PSP coding, so I'll think I'll start a 1-man coding venture during the summer when I'm not working
13:47:47S_a_i_n_tif it could be loaded and run by the PSP, then, that'd be cool...but, as a replacemt FW I think it would be pretty bloody useless.
13:47:53 Join AndChat| [0] (~jd@
13:48:14 Quit JdGord (Read error: Connection reset by peer)
13:48:23ranmabor_ka: The really weird thing is, if put dump_dache_range() in front of the sd_transfer_sectors call it works.
13:48:25MeekrabWell, a simple LOB plugin in a custom firmware would suffice
13:48:35Meekrabthen you could just boot with R button held and turn it off
13:49:03S_a_i_n_tDoesn't the PSP already have a fairly decent media player though?
13:49:07Meekrabpersonally, custom EQs would be a godsend
13:49:16Meekrab:) and a decent doom port
13:49:22S_a_i_n_tAny games the rockbox has are bound to seem FAR inferior to the PSP :)
13:49:26ranmaIf I then move the if(aligned) { if(write() clean... else dump } block at the beginning of sd_transfer_sectors I get verify errors again
13:49:38ranma*to the
13:51:40Meekrabto get at the zune firmware, would a decapping suffice in finding the contents of the firmware???
13:52:05Meekrabor will that be futile, cause I'm not too savvy with the nitric acid :P
13:52:06bor_kaand if you add dump and then sleep before the while in the sd_transfer? if not, then it is not timings, but some other calls...
13:52:14 Quit AndChat| (Client Quit)
13:53:01ranmabor_ka: Can you try this?
13:53:12ranmaWith it I currently don't get read verify errors
13:53:23jhMikeSMeekrab: who knows. maybe think "unconventional". who knows, maybe you could play the firmware out the i2s audio port :)
13:53:35ranmaI'm interested in your cmd errors count and if you still get parts of the wrong mp3...
13:56:50Meekrabhow do you call for another EXE from within a running one?
13:57:10bor_karanma, ok, building now
13:58:17jhMikeSMeekrab: if it's windows, I'd think that would be CreateProcess or ShellExecute, or some sort of DLL. Is the code running in supervisor mode?
14:00:17MeekrabI think so, considering it'll happily load any arm coded app for its ver of CE
14:01:20pixelmawodz: turning the ability to force left and right scrolling of the lists (menus, browsers) by holding the left and right button. It's a hidden setting though (only settable through the config file and I'm not sure if it is still needed at all). There was the need to be able to turn it off in the beginning because it interfered with other functions on holding these buttons (especially on Ondio as a "long Right" calls the context menu on the item, the
14:01:20pixelma Ondio doesn't have a Select button)
14:01:49 Join JdGordon_ [0] (6e15015a@gateway/web/freenode/ip.
14:02:04pixelmanow the forced scroll got a different button mapping (a combo)
14:02:50wodzwhat do You mean by scrolling?
14:03:19jhMikeSMeekrab: if you can read cpsr, and see the correct mode bits. I'm sure it'll load app code but you won't have access to everything without supervisor.
14:04:51jhMikeSif you can root kit a wintel machine, perhaps that's doable there
14:05:20jhMikeSit would be ironic if a zune were more secure than "big windows"
14:05:23Meekrabthis is what I'm using to load any helloworld type program
14:05:51 Nick Meekrab is now known as DJFender_Meekrab (
14:06:57pixelmawodz: make the test "move" left or right. I'm not sure if your MPio has the button mappings but e.g. on your Mini you can see it, I think - go to the menu or a browser where an entry is longer than the display could show at once, then hold the "Next" or the "Prev" buttons
14:07:39DJFender_MeekrabI'm tired, I'm gonna log off for a few hours, or go buy a coke
14:07:43DJFender_Meekrabbrb I found money :P
14:10:05wodzpixelma: I see.
14:10:44 Quit wodz (Quit: Leaving)
14:14:39DJFender_Meekrabyay coke
14:18:10n1sDJFender_Meekrab: if you can run any code, couldn't you just dump the firmware to a file and put the HD in an adapter and read it back from a computer?
14:18:27n1s(i assume zune is mtp only)
14:18:43 Quit teru (Quit: Quit)
14:18:44DJFender_Meekrabit uses some stupid MTP-Z protocol for sync
14:19:04DJFender_Meekrabotherwise some awesome dude would have just plugged it into his linux box long time ago
14:19:37n1syeah, but has anyone tried just plugging the HD into a computer? That was done on the gigabeat S
14:20:07DJFender_MeekrabIt has a odd partition
14:20:10jhMikeSdidn't they put it in an ipod or something
14:20:14 Join hebz0rl [0] (
14:20:16DJFender_MeekrabI put it in my ipod 30
14:20:24DJFender_Meekrabbut when I put it back it got formatted
14:21:05jhMikeSwhat sort of odd partition?
14:21:27amiconnipods have their own idea of "proper" partitioning
14:21:49amiconnA usb enclosure or internal ata adapter would be a better idea
14:22:08DJFender_Meekrabwell, it has 2 FAT32 partitions, but with Transaction FAT boot record
14:22:12n1samiconn: surely the emergency disk mode should be usable?
14:22:16bor_karanma, there is still read verify panic
14:22:26n1sDJFender_Meekrab: that's the same as the gigabeat then
14:22:31bor_karanma cmd 18 of 267000
14:22:47bor_kar/w 0 of 8417
14:22:57DJFender_Meekrabwhats the name of the EXPLORER.EXE file in wince?
14:23:05DJFender_Meekrablike the my computer option
14:23:09jhMikeSput it in a rockboxed gigabeat s then
14:23:15bor_karanma cmd error # does not change over time, or at least seems to be
14:23:23DJFender_Meekrabcause maybe I can tell it to show its drives
14:23:43CIA-6New commit by 03uchida (r27174): text viewer: some members of preferences change to bool. And simplify menu setting functions.
14:23:46jhMikeSwell, not quite that, but...perhaps a beast could be useful there
14:25:29CIA-6r27174 build result: All green
14:25:34 Quit kachna (Ping timeout: 248 seconds)
14:25:54DJFender_Meekrabif I could only write a program that would utilise the wifi or the USB capabilities and force it into a sort of disc mode
14:26:43jhMikeSit uses nk.bin as well on the boot partition or some other name?
14:28:32DJFender_Meekrabwould that help?
14:28:56jhMikeSmaybe...I know little about the device to be honest
14:28:58DJFender_MeekrabI wiresharked my network connection while updatng the fames
14:29:20 Quit flydutch (Quit: /* empty */)
14:29:32DJFender_Meekrabwell, I guess I'll have to fiddle with it
14:29:47DJFender_MeekrabI'll get something going by hopefully after canada day
14:29:59DJFender_Meekrabthat is if I dont have a bad hangover
14:30:47n1sisn't the problem that we need to find a way to boot unsigned code since the hole used in the beast was patched?
14:31:30DJFender_MeekrabSofar I havent found a way to load unsigned from boot...yet
14:31:40DJFender_Meekrabbut from the games menu I can load all kinds of shizz
14:32:09DJFender_MeekrabShizz (adj.) A lot of things (Webster 1994)
14:33:01 Join Kitr88 [0] (
14:33:46DJFender_MeekrabWell, I have NK.bin, Eboot, recovery and
14:35:24jhMikeSlooks like a familiar trio
14:35:42 Quit Kitar|st (Ping timeout: 240 seconds)
14:36:21DJFender_Meekrabjhmike; you wanna look at nk.bin? I'm extracting it to bare files
14:36:22jhMikeS*no "" here
14:36:34jhMikeS*on the beast, I mean
14:37:45 Quit Kitr88 (Ping timeout: 260 seconds)
14:37:59DJFender_Meekrabyeah, is what came with my update
14:39:12 Quit ucchan (Quit: Leaving...)
14:39:15DJFender_MeekrabD:\Nkbintools>D:\Nkbintools\viewbin.exe nk.bin
14:39:16DJFender_MeekrabViewBin... nk.bin
14:39:16DJFender_MeekrabImage Start = 0x88200000, length = 0x00D82D8C
14:39:16DBUGEnqueued KICK DJFender_Meekrab
14:39:16DJFender_Meekrab Start address = 0x88201000
14:39:16DJFender_MeekrabChecking record #75 for potential TOC (ROMOFFSET = 0xFFBFC548)
14:39:16***Alert Mode level 1
14:39:16DJFender_MeekrabChecking record #107 for potential TOC (ROMOFFSET = 0x00000000)
14:39:16***Alert Mode level 2
14:39:16DJFender_MeekrabFound pTOC = 0x88f81a60
14:39:17***Alert Mode level 3
14:39:17DJFender_MeekrabROMOFFSET = 0x00000000
14:39:18***Alert Mode level 4
14:39:30Tornewhat this comes down to, unless I am misunderstanding, is that there's a way to run arbitrary user mode code
14:39:42Torneso you need another exploit now, to get supervisor mode access :)
14:39:50Torneor a boot chain compromise.
14:40:03DJFender_MeekrabI'm looking for a hole in the eboot
14:40:36Torneright, but that's not particularly likely
14:40:48Torneand you should be able to discover if it has a problem fairly quickly..
14:41:02Tornethe part of pmcboot_secure which handles image loading is not htat long :)
14:41:41Tornewell, if you assume the crypto implementation works properly
14:41:43DJFender_MeekrabSo what I'm getting from the community is...rockbox is meant to replace not supplement
14:41:53Tornewhich it probably does, it's not like MS had to write a new one
14:42:10DJFender_Meekrabso going the games exploit is not nessicarily the way to go
14:42:19DJFender_Meekrabbut its good for figuring out how it runs
14:42:20Torneright now, we have only the most basic way of running as an application
14:42:34DJFender_Meekrablike what I got at>
14:42:38Tornein the future this might be better developed, and someone could port rockbox-as-an-app to WinCE
14:42:46***Saving seen data "./dancer.seen"
14:42:48 Join Kitar|st [0] (
14:42:55Torneit's being worked on for GSoC (rockbox as an app in general)
14:42:58 Quit hebz0rl (Read error: Connection reset by peer)
14:43:19Tornebut from the point of view of running rockbox as an OS, the games exploit is useless without a kernel exploit as well
14:43:46Tornewhich might well exist, but the surface to search is rather large
14:43:58 Join teru [0] (
14:44:01DJFender_Meekrabdumprom -d D:/Dumprom -v -5 nk.nb0
14:45:13 Quit TheSeven (Ping timeout: 260 seconds)
14:45:27 Join hebz0rl [0] (
14:48:38Tornei assume you've tried the HaRET methods, though?
14:49:08Tornei don't think it's possible any more in CE6
14:49:19***Alert Mode OFF
14:50:37DJFender_MeekrabI have the zune NK.bin dumped and visible...who wants to look at it?
14:50:51Tornehasn't that been available for ages?
14:50:59Torneand probably nobody
14:51:11DJFender_MeekrabWell... I need help too :P
14:51:36Torneif you're seriously wanting to hunt for a kernel exploit in CE6 then you're not likely to get a lot of interest from people here, because that's an extremely tedious and unfruitful endeavour
14:51:44Tornea huge amount of hard work for quite probably no payoff at all :)
14:52:06Tornepoking through pmcboot_secure looking for a boot chain exploit is less work because there's a lot less code involved
14:52:12Tornebut.. there probably isn't one ;)
14:52:12jhMikeSEboot isn't encrypted
14:52:34TornejhMikeS: yeah, i was gonna say, i don't think any of it is encrypted, is it?
14:52:39Tornejust signed..
14:52:56Tornethe pmcboot code i've looked at doesn't decrypt stuff
14:53:16jhMikeSI didn't look at nk.bin for this, but Eboot, definitely plaintext
14:54:05Tornewell, yeah. if it boots the same way as the beast, which it's reasonable to assume it does, the eboot file is a copy of pmcboot_secure, waiting to be flashed to NOR
14:54:53Tornethe cursory look i had at the zune stuff suggested it was all just the same as the beast, except it did a better job of bounds checking the image load :)
14:55:03Tornei didn't actually look at the code myself
14:55:13 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
14:55:59DJFender_Meekrabeboot.bin contains an nk.exe...
14:56:21Tornenot really
14:56:38Tornethe CE bootloader just is a copy of CE
14:56:45DJFender_Meekrabwell, thats what my tools extracted.
14:56:54Tornewell, that's probably useless
14:57:29Tornethe bootloader is just another copy of CE, it includes a bunch of the kernel code and so on
14:57:47Tornebut it doesn't load any other binaries, and its entry point is the hardware-specific bootstrapping code first
14:58:09DJFender_MeekrabI'll stick it in my helloworld
14:58:14Torneif you go through a disassembly you will probably find there's a huge chunk of hardware setup, it relocates itself to RAM, enables the MMU and then jumps to the bit that the rom tools think is nk.exe
14:58:14DJFender_Meekrabsee if it'll just reset
14:58:24Torneyou realise nk.exe is not, ever, an executable? :)
14:58:28Torneit's not a CE binary
14:58:36Torneeven the actual one used by the system
14:58:45Torneit's only called .exe by convention
14:59:04Torneyou can't run it, except from its own bootstrap
14:59:37DJFender_Meekrabperhaps I need two eggs and no chickens to accomplish this feat of awesome
14:59:44jhMikeSwhat checks the signing?
14:59:49TornejhMikeS: the bootloader
15:00:05jhMikeSso if you mess up the booloader in some way?
15:00:15Tornehm? how?
15:00:24Tornemeeting, back later
15:00:55jhMikeSto bypass those checks (probably carefully)
15:01:20DJFender_MeekrabBasically the only real thing that will go awry with messing with the loader is a formatted HD
15:03:20DJFender_MeekrabOk that is odd
15:03:33DJFender_MeekrabLoading the older zune nk.exe made the games button dissapear
15:03:42DJFender_Meekrabbut when I reset it it comes back
15:05:03DJFender_Meekrabzuneroots.p7b < Verisign certificate file....useful?
15:05:07 Quit mt (Ping timeout: 252 seconds)
15:05:57DJFender_Meekraball SHA1RSA and MD5RSA type hashes...
15:06:26 Join mt [0] (~mtee@rockbox/developer/mt)
15:13:31ranmabor_ka: Weird, if you don't have errors on command submission then I don't know how the sector number gets corrupted, which seems the only logical reason for me to get the read verify error (_if_ caching is no longer an issue)
15:14:19bor_karanma, I get 18 errors - if playing without verify abort
15:14:47bor_kabut if I get an abort... should compile version without read verify check and look at cmd errors?
15:15:06ranmaI get the 18 errors too, that's probably somewhere during card initialization
15:15:19ranmabor_ka: Does the abort prevent you from playing the mp3s?
15:15:34bor_kayes, of course, it panicf :)
15:15:48bor_kawill compile with it undefined
15:16:10ranmaWell, I assumed the abort is not directly after boot, but rather sometimes during playing where it would otherwise start playing part of the other mp3 :)
15:16:34bor_kaI get an abort at the beginning of the song, usually
15:16:49bor_kaSo, I can't look at the debug screen and start playing
15:17:33bor_kamoreover, as soon as I get cmd error, I will get an abort, so I will not see it in the debug menu?
15:17:57ranmaNo, the abort doesn't happen on cmd error, only if the verify fails.
15:18:17ranmacmds are retried a few times before giving up, so a single error should make the read fail
15:18:54ranmaOnly if you were getting lots and lots of cmd errors, but not enough to exceed the number of retries for command submission...
15:19:27 Join komputes_ubuntu [0] (~komputes@ubuntu/member/komputes)
15:19:49ranma...then the crc7 protection of cmd+argument might fail to catch an error and might cause the bug you are seeing.
15:23:06bor_kaI've compiled build without read verify - no cmd errors on the song, that aborts with verify enabled
15:23:11bor_kaSo, it is not cmd
15:24:45bor_kabtw, may be it is stupid - I have changed sd_transfer_sectors call like this
15:25:14bor_kait aborts just the same, but my config file just got trimmed in a strange way
15:26:30bor_kathe last lines from config.cfg including font etc was missing - that is why I noticed it
15:26:32ranmaWhere does it abort then if you removed the verify part?
15:27:00bor_kaif I remove the verify part, it does not abort
15:27:55 Quit Mich ()
15:28:03bor_kawhy part of the file was missing if I transfer it by pieces? weird....
15:28:15 Join Zigtown [0] (
15:29:19balintx6 days >_<
15:30:28 Quit hebz0rl (Read error: Connection reset by peer)
15:31:17ranmabor_ka: This should rule out any cacheing issues:
15:32:07 Join hebz0rl [0] (
15:32:23ranmaIt always takes the 'unaligned buffer' path in sd_transfer_sectors and uses memcpy to/from the uncached mapping of the dma buffer.
15:32:34 Part LinusN
15:33:21CIA-6New commit by 03teru (r27175): some improvements for simple viewer. ...
15:34:01 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
15:34:11ranmabor_ka: BTW if you removed the dump_dcache_range() in the verify path in your modification that might explain the verify error with the modified version
15:34:51bor_kano, I have not removed it
15:34:54CIA-6r27175 build result: All green
15:35:43bor_kawhat I posted was my attempt to make reads by UNALIGNED_NUM_SECTORS sectors - may be it got corrupted only on long buffer?
15:36:03bor_kaso, this code was a replacement for the first "main" read
15:39:43bor_kaok, got your patch - read verify failed....
15:41:11bor_karanma, are you sure you uploaded the correct patch? I don't see any changes in sd_read_sectors
15:43:52bor_kaok, I see in verify code
15:44:42bor_kabut I think that the cause is the first read being erroneous
15:53:43ranmaWell, we'd need to dump the two buffers and compare them to tell what went wrong :)
15:55:02ranmaI thought for a while I could reproduce your problem since I saw verify errors, but mine were just cache coherence related (probably reusing the aligned_buffer for verify was problematic cache-wise)
15:56:35ranmaSince you still get these errors with the patch that uses a different buffer (compare_buffer) and forces all reads/write to go through the uncached memcpy ("aligned = 0;" at the beginning of sd_transfer_sectors), your problem must still have some other root cause.
16:01:08ranmabor_ka: You said using −−no-thumb fixes it, right? If it's some weird compiler error that could make it hard to track down...
16:02:09CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
16:02:09*bor_ka have seen once a compiler error on a VAX... took me 2 days and studying some VAX assembly
16:02:12 Quit JdGordon_ (Ping timeout: 252 seconds)
16:02:23 Join mk429_ [0] (
16:03:00bor_karanma, but 26581 was Ok with thumb
16:04:37 Quit teru (Quit: Quit)
16:04:48ranmaYes, but so far you've been the only one seeing problems I think and you explicitly compile your target using −−thumb, so... On 2MB targets (compiled thumb by default) it may not be an issue because of the memory size.
16:05:18 Quit mk429 (Ping timeout: 240 seconds)
16:05:47ranmaI'll be back in germany in about two weeks, I'll try to get an AMSv1 player with more ram then...
16:07:36ranmaMaybe I can reproduce it then, so for now I'd just ask you to build without thumb and see if it's still fine with longer usage.
16:07:39bor_kaif they still are sold, v1 I mean
16:07:48ranmaWell, there's ebay :)
16:08:26bor_kaoh, I have forgot it... :)
16:11:22ranmaSansa apparently weren't too popular in japan, but on there are more than enough listed I'd say
16:13:26*bor_ka got strange result with a nothumb build, starts compiling it anew
16:13:44*DJFender_Meekrab twiddles thumbs
16:25:51bor_karanma, made a fresh build with the HEAD revision and your latest patch without thumb - it still gives read verify fail
16:26:26n1sbor_ka: is it happening on the same file all the time or on different files?
16:26:34 Quit bluebrother (Disconnected by services)
16:26:34bor_kaso, the only difference is that there are audible glitches with thumb, and there are no audible glitches withiu thumb...
16:26:37 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
16:26:42bor_kan1s, different files
16:27:14bor_kaI'd say it is a cluster of files that are located nearby: I recently formatted internal memroy and copied all musci from the PC
16:28:03bor_kaso, files, mostly, a physically located one after another
16:28:09bor_kaor I believe so
16:28:58*bor_ka makes a lot of tpyos ... wonders why
16:30:09bor_kaso, if there is read verify error on a file, it stays so after reboot, if there is no error - it plays always Ok
16:31:20bor_kamay be a defective internal memory chip, that OF can handle in a strange way? Or not defective, but with uncommon timings?
16:32:06ranmaThen it shouldn't be cured by disabling 4bit mode.
16:32:24ranmaDoes internal vs microsd matter?
16:33:03n1sit couldn't be the flash actually going bad?
16:33:15ranmaI doubt it
16:33:37bor_kaI have never heard audio glitches from the microsd, will try a little now
16:34:49bor_kaI've seen with Nintendo DS a situation, where a small batch of NDSes got different touchscreen, and with "official" games all was Ok, but with homebrew apps, touchscreen calibration was terribly broken
16:35:13bor_kaSo, OF "knows" something, what homebrew doesn't
16:35:43bor_kamay be internal memory sometimes cannot work quick enough to handle 4bit ?
16:39:08bor_karanma, weird - started playing music from the mSD, got read verify failed from..... sd0!
16:40:04bor_kadisabled gather runtime data - got read verify failed from sd1
16:40:58bor_kaif there are so many read fails, why binary codes like plugins or codecs are not corrupted?
16:42:50***Saving seen data "./dancer.seen"
16:43:34n1sis there any point in keeping kluges for the old ar toolchain around when they only affect targets that have been switched over?
16:46:51 Join xguto [0] (
16:48:32 Part xguto
16:49:00bor_karanma, can it be so, that when calling sd_transfer_sectors in the verify part results in error, that is ret != 0. It is never check in the #ifdef VERIFY_READ
16:49:53ranmaHmm, yes, might be, can't hurt to check that. :)
16:54:13bor_kahmm, these errors were gone for 1 minute approximately, that is I was able to start playing a song that always made an error
16:54:34bor_kathen it finally got to read verify error, ret == 0
16:56:12bor_kaso, it is not faulting reads in the read verify
16:57:54CIA-6New commit by 03nls (r27176): Clean up a leftover preprocessor symbol and improve the workaround for the 'arm-elf and static functions in iram' quirk so it doesn't affect builds ...
16:59:00n1sthose *_TREMOR_* macros should be renamed
16:59:16 Quit TheSeven (Ping timeout: 240 seconds)
16:59:33CIA-6r27176 build result: All green
16:59:33 Quit shai (Quit: Leaving)
17:00:43bor_kadisabled 4bit bus - still read verify
17:02:32n1ssaratoga: is there any reason to keep around the old tremor mdct in the codec lib?, nothing uses it afaict
17:03:50bor_kabut just once, cannot reproduce it now. Will test a little to see if the 4bit bus is a cause, or just catalyst
17:06:08bor_kaok, reproduced it with the database update. r27175, no thumb, no 4bit bus, ranma's latest verify reads patch
17:06:30bor_ka18 cmd errors on the internal drive, 1170 cmd errors on the uSD
17:06:46ranmaYeah, I see more cmd errors on the uSD too for some reason
17:07:00ranmaAs long as they don't increase while playing...
17:07:53bor_kaWhen I was updating the db, cmd was 18 and don't increase...
17:09:27bor_kain fact we need somebody with the fuze v1 to check this behaviour
17:09:40bor_kaIIRC funman has fuze v1?
17:10:23 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
17:11:44ranmabor_ka: Here's a last patch, which will also add add a partial buffer dump to the panic:
17:12:43ranmaIf it's something simple like the 16 byte skew due to the dma fifo, that I found for the write bug, it should be visible.
17:12:57ranmaBut given that you hear parts of the wrong song I doubt this will help... :/
17:16:53 Quit komputes (Ping timeout: 245 seconds)
17:17:15 Quit hebz0rl (Ping timeout: 265 seconds)
17:19:35 Part Zagor
17:22:43 Join MethoS- [0] (~clemens@
17:25:07 Join Keripo [0] (
17:25:35 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
17:27:29 Quit komputes_ubuntu (Quit: I haven't slept for ten days, because that would be too long.)
17:29:36 Part Keripo
17:35:01 Join shai [0] (
17:36:28 Join mt_ [0] (~mtee@
17:36:37bor_kagot panic... wil try to copy/paste it here
17:37:06 Quit mt (Ping timeout: 240 seconds)
17:37:12 Nick mt_ is now known as mt (~mtee@
17:37:35bor_ka0: 53CB285B542C29...
17:43:27bor_kaI don't see a pattern. Will get tomorrow a good photo of the panic screen - my phone now can't do macro
17:44:40*bor_ka got to go home
17:44:44 Quit bor_ka (Quit: CGI:IRC 0.5.9 (2006/06/06))
17:45:49 Join AzureWurk [0] (
17:46:03 Quit TheSeven (Ping timeout: 245 seconds)
17:47:50saratogan1s: yes it needs to go eventually, but I'd like to leave it in current svn a bit longer since mt is still debugging the mdct and its nice to have them side by side
17:48:02n1ssaratoga: ah, ok
17:51:33 Join FF [0] (
17:55:37FFHey, I just got a Sansa Clip+ and was wondering who I could contact to contribute to the RockBox port?
17:56:45 Quit petur (Quit: *plop*)
17:57:32n1sFF: if you have any questions, they are best asked in here
17:59:31FFn1s: I'd like to help out in the Sansa Clip+ RockBox port. The last post in the thread said developers could be reached via the IRC.
18:00:58Torneyes. you are on irc, you have reached them :)
18:01:53mtsaratoga: Forgot to tell you yesterday that the sampling rate is in fact 48KHz
18:02:08mtin the failing samples
18:02:51n1sFF: i don't know the clips myself but what is it that you want to help out with?
18:05:40FFn1s: I don't have much development experience, but I can help out with testing and debugging the builds that are being released.
18:07:34BSOD90FF: there are many people currently using rockbox on clipv2 and clip+ as it already runs quite stable
18:07:43 Join hebz0rl [0] (
18:07:53TorneFF: you don't need our blessing to do that, just do it :)
18:09:35n1sFF: so you probably want to start with getting a toolchain set up so you can make your own builds
18:09:52TorneFF: just update to the latest current build reasonably regularly, and report issues you have..
18:10:09Tornereading the forums is also useful, since if someone else mentions a problem you can try and reproduce it
18:10:30 Join Zarggg [0] (
18:17:37 Join funman [0] (~fun@rockbox/developer/funman)
18:17:40n1sas well as testing reported bugs to try to add more information to them
18:18:38 Quit z35 (Ping timeout: 265 seconds)
18:20:05 Quit pamaury (Remote host closed the connection)
18:20:50FFCool, thanks for the guidance. Looks like I'll probably have to go through the 129pg, 3 year thread to catch up. =D
18:21:24funmanthe thread has a lot of obsolete info
18:21:27Tornethat thread has very little if any current information in it
18:21:30Tornei suggest you ignore it entirely
18:21:35Torneread new threads on the forum instead :)
18:22:01funmanor check the SansaAMS wiki page, there's a list of known problems
18:29:41 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
18:30:58 Join z35 [0] (
18:34:17balintxhow can I emulate in the sim that an usb connected?
18:35:43mc2739balintx: press "u"
18:36:41balintxlol, the sim wasn't crash like the real player :D
18:38:29 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
18:39:21 Join rado1 [0] (
18:40:18 Quit CGL (Remote host closed the connection)
18:40:36 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
18:41:53 Quit Luca_S (Quit: CGI:IRC)
18:42:52***Saving seen data "./dancer.seen"
18:43:30balintxmc2739: any idea how to use like this...?
18:49:13 Join pamaury [0] (
18:49:13 Quit pamaury (Changing host)
18:49:13 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
18:50:57FFfunman: Will do. Is anyone heading the project or is it just a bunch of people that implement whatever they want?
18:52:46funmanthe latter although there are some people which are more involved in Sansa AMS ports (i am one of them)
18:55:04balintxfunman: what about disabling usb in fuzev2? :/
18:55:48 Quit swilde (Quit: ERC Version 5.3 (IRC client for Emacs))
18:55:50funmanbalintx: what about it ?
18:56:11balintxit almost kills the player
18:56:22funmanno problem for me
18:56:33balintxam waiting for full discharge now... because of that
18:56:59simonrvni don't have issues either
18:57:14funmanbalintx: don't use unstable software, or if you want to complain don't forget to attach a patch which fixes the issue to your complaint
18:57:34simonrvnusing the svn build even
18:57:50funmanand afaik although it's suspected that plugging USB causes this problem we're not sure yet
18:57:59funmanand I can't disable you from plugging USB, whatever I do
18:58:22balintxwhy not sure?
18:58:27balintxi am sure
18:58:36balintxi plugged in, then it happened
18:58:42balintxit IS sure
18:58:44funmanit could happen on other occasions
18:59:44 Quit MethoS- (Remote host closed the connection)
19:00:05simonrvnwhether it's shut off, in RB, or OF, i plug it in, and works fine
19:00:34balintxsimonrvn: fuze v2?
19:00:59Torneyou need to actually demonstrate causation, is the thing..
19:01:05Tornea couple of isolated incidents is not really any help
19:01:14Tornethere's no hint on what to look at..
19:01:16simonrvnfor sure
19:01:26balintxwhat do you mean?
19:01:40balintxafter the week when it discharged, do it again
19:01:49Tornebalintx: it works perfectly fine for a reasonably large number of people
19:02:31balintxthe same hardware.
19:02:37bertrikoh, I seem to remember a remark where someone said it happened at USB plugin before
19:02:48 Quit r0b- (Ping timeout: 264 seconds)
19:03:09balintxsomeone said
19:03:23balintxbut it is a real problem
19:03:44Tornebalintx: quite probably, but just because two people have had the problem once, that's not a very good reason to disable it for everyone
19:03:50Torneit works fine, over and over, for almost everybody :)
19:04:01ranmabertrik: Re: "what is the correct way now to make sure that several ascodec reads or writes are done in sequence": disable_irq_save()
19:04:06Torneif we disabled everything that had ever failed to work for anyone, rockbox wouldn't do anything at all
19:04:44bertrikTorne, I believe there isn't really much that "works" at USB plugin on the fuze v2
19:04:58balintxTorne: I do not agree.
19:05:18balintxif it is the state, what you said
19:05:26simonrvnit just kicks in into usb charge/mount in the OF ATM
19:05:28ranmabertrik: It would be possible to extend ascodec-as3525.c so that's not needed, but I doubt it's worth the additional complexity
19:05:36balintxthen in the wiki pages should be write Working
19:05:43 Join Jerom [0] (~jerome@
19:05:55Tornebalintx: 1) the port isn't done yet, it's listed as unstable, things ar enot guaranteed to work
19:05:55balintxbut.. No work is written
19:06:03Torne2) the wiki is not always up to date
19:06:19bertrikranma, ok
19:06:22TorneIt's not a stable port, USB may indeed not work reliably
19:06:29Tornebut it does, generally, work for most people
19:06:40Torneso, why would you disable it?
19:07:05balintxnot me, someone noticed it.
19:07:37simonrvnfor sure. besides, how do you expect to get it stable if you disable things? the devs need people to test stuff.
19:07:42Tornei use "you" in the abstract sense
19:07:43Torneof "someone"
19:08:16balintx" Comment by MichaelGiacomelli (saratoga) - Tuesday, 29 June 2010, 05:14 GMT
19:08:16balintxSince rebooting for USB seems to occasionally cause problems, we should probably just disable it. We'll hopefully have our own USB mode soon enough."
19:09:02balintxsimonrvn: how can I test it with a dead player?
19:09:20Tornebalintx: I don't know why you think quoting someone else's opinion is going to change mine
19:09:25TorneI have, indeed, already read that bug report
19:09:37Tornei also don't know why you are trying to change my opinion :)
19:10:20simonrvnbalintx: the wiki and other places show you how to fix a "dead" player.
19:11:00balintxsimonrvn: I already doing that. but it is slow to discharge the battery.
19:11:09simonrvndo you expect tests and such to work? you expect too much. it's nice that they work, but it's *unstable*.
19:11:18simonrvnbalintx: it's called patience.
19:11:50balintxI am patient, but also want see my device living.
19:13:59Tornebalintx: anyway, nobody made it do that on purpose. it's only been reported in the last week or so, it's not like we've know about this the whole time
19:14:12Torneand it obviously *does* work for most people or we'd be hearing about it a lot more
19:14:31Torneand the port is not finished, and is listed as unstable, and doesn't claim that USB works.
19:14:46Torneso, i just don't see the point in disabling it, when it doesn't cause any permanent issues, just annoyance
19:15:09simonrvnpretty much what i was also trying to get at
19:15:15Tornedisabling it doesn't help it get fixed, either. and will the problem even stop if it's disabled? we don't knowl
19:15:31Tornemaybe the crash is actually triggered by some setup we have done, and it happens as soon as the electrical connection is made
19:15:33simonrvnthat too
19:15:36balintxOtherwise, a question. Does the warranty extend to a custom firmware caused ruination? Any experience?
19:15:52Tornebalintx: Most manufacturers are of the opinion that flashing a nonstandard firmware voids the warranty
19:16:01Tornewhether that would hold up if you took them to court, i have no idea :)
19:16:19Tornewhether you try to claim ont he warranty anyway is up to your own judgement :)
19:16:19funmanranma: on fuze1 i read some bytes from /dev/by-path/pci-0000:00:04.1-usb-0:6:1.1-event-mouse only when i unplug the fuze
19:16:55saratogaregarding the USB reboot feature, I think we should disable it because its apparently not entirely stable and also not all that useful
19:16:56simonrvnno you need to stick on the unpatched OF first then return it. i got a store-based insurance, so i just clean it up a bit internally, then return it. i did already ;). but i hadn't had RB on it yet, and it was the OF that crashed.
19:17:52funmanhm there's no hid keymap for fuze thoguh
19:18:36 Quit dfkt (Read error: Connection reset by peer)
19:18:52funmanc200v2/e200v2 use the same keymap than v1 so there was no change needed
19:19:42 Quit DerPapst (Quit: Leaving.)
19:20:42ranmafunman: Weirdly your usb trace looks a bit better than mine though, at least up to the part where it starts talking to the usb storage part, haven't really looked at the remainder...
19:21:32 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
19:22:38funmanhow better ?
19:23:53 Quit z35 (Ping timeout: 265 seconds)
19:24:14funmanpixelma: is USB HID keymap documented in the manual ?
19:24:20ranmaAfter "s 00 09" (set config) I get two setup packets in quick succession in the trace ("s 80 06" and "s a1 fe") and linux somehow doesn't bother to read the response for the first one. That part appears to have worked fine on your device.
19:24:44funmani'm running 2.6.32-23-generic #37-Ubuntu
19:24:52ranmaFor me that upsets the fifo, which is why I re-added an attempt to work around that by flushing the fifo prior to send
19:25:10ranmaLinux navi 2.6.35-rc3 #33 PREEMPT Thu Jun 17 14:13:37 JST 2010 i686 GNU/Linux
19:25:19 Quit fxb (Ping timeout: 248 seconds)
19:25:39 Join fxb [0] (
19:25:51 Quit Kohlrabi (Ping timeout: 264 seconds)
19:27:24 Quit solrize (Ping timeout: 264 seconds)
19:28:25funmanranma: works fine now!
19:28:43ranmaAfter adding hid keymap?
19:28:56CIA-6New commit by 03funman (r27177): Fuze: USB HID keymap copy/pasted from e200
19:29:05funmanthough no errors in dmesg this time
19:29:06 Quit Zigtown (Remote host closed the connection)
19:29:47 Nick fxb is now known as fxb__ (
19:29:52ranmaOk, c200v2 just uses the c200 keymap, which already had hid keymap entries
19:30:35CIA-6r27177 build result: All green
19:32:05 Join Luca_s [0] (
19:33:46 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
19:34:04funmantomers had added the keymap for clip though
19:34:08 Join komputes [0] (~komputes@ubuntu/member/komputes)
19:34:39 Nick fxb__ is now known as fxb (
19:35:22 Join Kohlrabi [0] (
19:36:54 Join z35 [0] (
19:37:11pixelmafunman: I don't know, I would have to look it up all the same
19:40:36 Join solrize [0] (
19:40:37 Quit balintx (Read error: Connection reset by peer)
19:41:23 Join balintx [0] (
19:42:36funmanworks fine on clip too
19:42:57funmanalthough the USB logo fits badly on the screen
19:47:10 Join Xerion [0] (
19:48:52 Join u42p [0] (
19:49:00u42pmy sansa clip+ is dead now...
19:49:28u42pyay for waiting for the battery to die
19:50:11 Quit balintx (Remote host closed the connection)
19:50:19funmanwhat did you do to have it in this state ?
19:51:10u42pa database update was running in the background
19:51:18 Join balintx [0] (
19:51:21u42pi think i pressed the home button during playback while the screen was off
19:53:48u42pthis sucks :(
19:53:57u42pis there any way to accelerate the battery consumption?
19:54:26gevaertsHolding power doesn't help?
19:54:39 Join jgarvey [0] (
19:54:49 Quit R31D (Quit: I-n-v-i-s-i-o-n 3.1.3 (May '10))
19:55:35 Join Buschel [0] (
19:55:58u42pyou are my hero
19:56:08u42pi hold power for 1 minute earlier when it happened with no luck
19:56:14u42pbut now it worked
19:56:20u42pthanks a lot! :))
19:57:57n1spixelma: there is a bug in the keymap thoug, in the fm screen the down button both decreases volume and opens the radio menu
19:58:02 Join bor_ka [0] (
19:58:12u42pi never noticed it but now the player is a bit noisy (high-pitched battery noise that cheap power adapters or charging batteries sometimes make). is that normal?
19:59:33bertrikn1s, I vaguely remember seeing a bugfix for that recently, perhaps for a different target
19:59:34bor_karanma, here is a screenshot, cough.., of panic on fuze v1
19:59:45Luca_syup, it was for fuzev2
19:59:56bor_kafunman, you have a fuze v1, don't you?
20:00:03n1sbertrik: ah,this was with the 3.6 release so perhaps it's fixed in svn
20:00:10simonrvnu42p: sometimes holding power for up to 1, 2 minutes is needed. literally a minute, not just "about a minute", which is usually around 30 sec)
20:00:22u42pheh, that is true
20:00:27pixelman1s: I noticed last week, kugel introduced it with the radio quickscreen patch and he didn't answer my question if he wants to bring it back (as it was removed again). I think the fix will look a bit differently depending on his answer
20:00:39u42pis it some kind of lowlevel rockbox emergency shutdown?
20:00:40pixelmathe bug is still there
20:00:49pixelmajust not the radio quickscreen
20:01:50simonrvnit's a hardware reset
20:02:05n1spixelma: hah
20:02:07pixelman1s: removing the radio quickscreen action (which I guess bertrik is referring too) did not fix it
20:02:08n1seh aha
20:02:41ranmabor_ka: Hmm, that is more interesting then I expected.
20:02:47n1syou seem to have a grasp of the situation so i'll leave it alone then
20:03:22funmanbor_ka: yes
20:04:28bor_kafunman, can you test with ranma's patch, if you get read verify error?
20:04:50bor_kait is something very strange there....
20:04:54ranmaBecause the mismatching part from the compare_buffer contains "icroSD1>/Rock/Gary"
20:05:27 Join binaryhermit [0] (
20:05:30 Join kachna [0] (
20:06:23ranmaCan you do "dd if=/dev/your_player_partiton seek=14421088 count=1 | od -x" ?
20:06:49 Join DerPapst [0] (
20:07:04bor_kaI get it when I started playing a file from internal memory. And it was not Gary Moore from uSD
20:07:39ranmaThe read may be accessing the database or something
20:07:40bor_kaoops, I am not on a Linux box - it is inside the vmware thing, at work
20:07:57*bor_ka wonders how to read a sector in a windows
20:08:27*ranma is sure there are hex editors on windows too, but not too sure how many support accessing raw disks :)
20:09:01pixelmathere's a standalone dd.exe available somewhere and hex editors too
20:10:07bor_kayou need sector from the internal memory, or from the uSD?
20:10:29ranmaInternal memory, since it says sd0.
20:11:00*ranma is just wondering which of the reads returned the correct data.
20:11:27funmanbor_ka: where is the patch ?
20:11:51ranmaAnd if you can search for a hex string on the drive that might be intersting to find the sector it read instead
20:12:07ranmaSee if it's a single-bit error in the sector address or something...
20:12:50bor_karanma, the sector looks like a directory entry
20:13:18bor_kafunman, at
20:13:41ranmaDoes it start with "2e2020202020" or with "6963726f5344"?
20:13:50bor_kaone sec
20:14:35funmanranma: please run svn diff from root folder :(
20:15:49bor_kaand what should I search - "icroSD1>/Rock/Gary" ?
20:16:11funmandatabase update seems very slow
20:16:30balintxcan rb deal with office files?
20:16:43balintx.doc, .xls or sy like that
20:16:55bor_kafunman, yes, there is double check for every read
20:17:05 Quit hebz0rl (Quit: Ex-Chat)
20:17:13funmani'm playing from internal and updating DB with µSD inserted
20:17:38ranmaIt's a directory, so the second read failed (I really should add a return code check there just in case...)
20:17:45bor_kafunman, for me on some songs there is a panic, on some - not. Try the first song from all albums
20:19:07 Quit rado1 (Quit: Leaving.)
20:19:26funmani had a panic with screen off
20:19:40funmana few r/w error on internal
20:20:31bor_karanma, found it at sector 193695, but not at the start...
20:20:41ranmaThe weird thing is on my C200v2 all works fine, I've been listening to mp3s the last two hours with that...
20:21:18ranmafunman: Question is, does it work fine in 1bit mode?
20:21:50bor_ka209310, 209311
20:22:05bor_karanma, should I search at the start of the sector, or everywhere?
20:22:05funmanranma: i don't know, i don't have much time to spend on testing
20:22:12ranmabor_ka: Only at the start
20:22:13funmani didn't even follow the whole discussion
20:22:19bor_karanma, Ok
20:23:37funmanpanicfbuf is an interleaving of the bytes of the 2 different transfer ?
20:24:09funmanyou could have added some spaces here and be kind with my eyes :)
20:24:19ranmaI wanted to cram as much on the screen as possible :)
20:24:23bor_kafunman, short story long: sometimes consecutive reads return results that differ from the first one. It happens often with the 4bit mode, and rarely in 1 bit mode
20:24:35 Join hebz0rl [0] (
20:24:38ranmaSo it does happen in 1 bit mode?
20:24:44bor_karanma, yes
20:25:10bor_kabut I don't understand, why there is no panics or like - when reading plugins, database etc
20:25:52ranmaWell, there _may_ be another retry on a higher level, I haven't looked at that...
20:26:23bor_kabut ret=0, how the higher level knows it should retry?
20:27:01ranmaWell, in my patch the read result of the verify read is never checked, so...
20:27:20ranmaAnd the original read with ret=0 worked just fine in your screenshot
20:29:13bor_kabut I have tried adding ret from the verify read to the panic message - it always was 0
20:29:24ranmaHmm, ok.
20:30:33amiconnn1s, pixelma: That radioscreen "down" bug also exists on clip+
20:32:14 Join r0b- [0] (
20:32:27bor_kaneed to add it again to your latest patch, to get both data and the ret value
20:33:21funmanthe wrong sector seems to be 9 sectors before the one we wanted
20:33:27funman(or after)
20:34:04bor_karanma, one more interesting question - if there was error, why USB mode worked Ok and I was able to read all this data?
20:34:13bor_kafunman, let's see...
20:34:28ranmabor_ka: Yeah, it's really quite weird
20:34:50ranma9 sectors is pretty close to UNALIGNED_NUM_SECTORS though...
20:35:35 Join wodz [0] (
20:37:16n1samiconn: afaiu adding a ACTION_NONE thingy for BUTTON_UP or simply dropping the REL thing should solve it
20:37:58n1sbut pixelma mentioned something about quickscreens so i don't know
20:38:17n1seh s/UP/DOWN/
20:38:57bor_kafunman, what do you mean by "wrong sector" - I have looked at the sectors +/- 16 from 14421088 - there is no signs of Rock/Gary of the beginning
20:39:24funmanbor_ka: there's no 'Rock/Gary' in my fuzev1 so it should be ok ;)
20:39:30pixelman1s, amiconn: the REL was introduced by kugel because there also was an action for the repeat event (this quickscreen thing)
20:39:45bor_kabut the song I've tried to play indeed was nearby, I've seen its id3 tag. And I think that 14421088 was a directory, where it was located
20:39:49simonrvni did that on mine, and i just get a buncha 0's
20:40:04pixelmayou get an idea if you read up on the SVN history a bit
20:41:18n1spixelma: makes sense, but that action was removed, so should the quickscreen action come back?
20:41:59bor_karanma, I'm doing a complete search through the internal memory - no matches yet, but no panics as well... So all reads are Ok and verified correctly?
20:42:13ranmaApparently, yes.
20:42:46bor_kaso... player thing - may be codec - messes internal buffers?
20:42:54***Saving seen data "./dancer.seen"
20:43:05ranmaIt must be some interaction problem, writes inbetween, or, well, something :)
20:43:10bor_kaor something inside the rockbox - database, etc?
20:43:18bor_kahm, writes - yes
20:43:21pixelman1s: unfortunately kugel didn't answer and my opinion about the quickscreen is biased (if it was me, it could go everywhere ;) )
20:43:22wodzpixelma: comming back to the global_settings.hold_lr_for_scroll_in_list - I can't understand the setting. At least in simulator lines which are too long are always scrolled
20:44:14ranmabor_ka: Hmm, valid thing to test for. Maybe replace the wakeup_wait() with polling that won't yield to other threads. Then nothing can mess with the buffers in the meantime... (except interrupts)
20:45:34pixelmawodz: you can force them to scroll quicker when holding left or right (on some targets). Just stating what I know, I don't need it and it's a very questionable setting anyways
20:46:09wodzI think I drop this in mpio keymap
20:46:23wodzs/I drop/I'll drop/
20:47:40wodzwhat is tree context?
20:47:57 Quit n1s (Quit: Lmnar)
20:47:58bor_karanma, got the wrongly read sector - it is 1816770
20:48:43bor_kaanother thread was walking through the database and corrupted read buffer?
20:48:53funmannow the 2 different sectors are spaced by 2 sectors
20:49:05bor_kathis 181. sector is 100% database
20:49:14pixelmawodz: the menus and browsers
20:49:24funmanranma: ah right
20:49:32funmanwe could move mutex locking outside sd_transfer also
20:50:18bor_ka1850242 also matches
20:50:26wodzhow this is different from list context?
20:51:58 Join Patsy [0] (
20:52:22funmanthat means current write verify in svn is broken
20:53:11funmanhm oh well i never know with cooperative threading
20:53:15ranmaI'm not sure it's possible to hit that though, unless mutex_unlock() will switch threads
20:53:39 Join stripwax [0] (
20:53:57funmanwell it does
20:53:58 Quit stripwax (Client Quit)
20:54:30funmangraa, undefined instruction in "threads" (ibss)
20:55:27*funman gives up
20:58:12CIA-6New commit by 03pamaury (r27178): as3525v2-usb: fix bug in ep listing macro, rewrite EP0 handling using a state ...
20:59:38CIA-6r27178 build result: 18 errors, 0 warnings (pamaury committed)
21:00:09 Quit Luca_s (Quit: CGI:IRC)
21:01:41Jeromn1 :D
21:01:57 Part u42p ("Leaving")
21:02:40BuscheljhMikeS: thanks for spotting the missing change in mpegplayer after my libmad change. and of course thanks for fixing it :)
21:02:54CIA-6New commit by 03funman (r27179): usb-drv-as3525v2.c: fix compilation (typo)
21:03:22 Join MethoS- [0] (~clemens@
21:04:39CIA-6r27179 build result: 0 errors, 6 warnings (funman committed)
21:05:00funmanpamaury: warnings
21:06:08 Quit funman (Quit: free(random());)
21:08:47 Quit BlakeJohnson86 (Ping timeout: 248 seconds)
21:10:02pixelmawodz: if I remember correctly the actions in "tree" context are for controlling playback from the menus and browsers, list is for navigating the list itself. The most important "tree" context action is the one that's used for resuming playback (something with play in it, I believe)
21:10:46pixelmaor wps
21:11:53pixelmaI'm not a 100% sure at the moment
21:12:20*wodz is confused with all this keymap magic
21:12:59CIA-6New commit by 03ranma (r27180): mutex_lock/unlock must be moved up a level to make the verify code in sd-as3525.c safe
21:13:17 Join BlakeJohnson86 [0] (~bjohnson@2002:1876:a27b:0:227:13ff:fe65:1262)
21:13:39bor_kafunman, ranma, soryy - was away
21:13:47bor_kawhat are these patches?
21:14:18pixelmawodz: they say it's better than what there was before ;)
21:14:28CIA-6r27180 build result: 0 errors, 6 warnings (ranma committed)
21:15:09ranmabor_ka: Another thread corrupting the buffer is quite likely, the lock has to be moved up
21:15:24ranmaTry before r27180
21:15:31bor_karanma, sorry will check it tomorrow - baby at last fell asleep, should be gone too...
21:15:37ranmaOk :)
21:15:41bor_kaok, bye
21:15:51 Quit bor_ka (Quit: sleeeeepppp...)
21:16:38wodzpixelma: did they say also that they were so amused with new keymap thing that forgot to comment this properly? :-)
21:17:58*pixelma looks at JdGordon... ;)
21:20:35pamaurysorry I was afk, I'll fix the warnings
21:21:40pamaurythis is incredible, my commit add some errors because ";" was changed in a "," by git or svn !!
21:22:11JeromI've never heard that
21:23:07pamauryI can tell it, my local code has ";", the one I committed as a ",". Git reports to difference between my code and head
21:23:53JeromI'm feeling stupid now, isn't the codebase on svn ?
21:24:41CIA-6New commit by 03pamaury (r27181): as3525v2-usb: fix yellow
21:24:58pamauryI use git-svn to commit to svn
21:25:14pamauryApparently something went wrong, a cosmic ray :)
21:25:41JeromOf course :p
21:25:48JeromIt's the only sane answer
21:25:50bieberpamaury: Does git-svn just work like Git with a normal repository?
21:26:16CIA-6r27181 build result: All green
21:27:29amiconnBuschel: Seen the MCF5250 results already?
21:27:39wodzbieber: git-svn is like extension that allow you to turn git commits into svn commits
21:28:41Jerompamaury: with your commit I'm getting consistent result
21:29:00pamauryJerom: with this commit, set address should work reliably
21:29:14pamauryhowever, get descriptors freeze randomly on my computer
21:29:15Buschelamiconn: yes, thank you for the test. I expected more speed up :/
21:29:48 Part Patsy
21:29:49JeromI freeze on set Configuration it seems
21:29:58pamauryYes that the best I can get
21:30:11pamauryEither it freeze on set config, either on some random get desc
21:30:29JeromIt's like the controller is getting out of sync
21:30:55pamauryAnd I still don't get any disconnect interrupt
21:31:01JeromMaybe a stupid question but is PLLA a stable clock ?
21:31:45amiconnBuschel: I guess if we want faster aac decoding we need to drop libfaad and port something from ffmpeg
21:31:54pamauryNo idea
21:32:07JeromWill try with PLLB
21:32:07pamaurybertrik: ^
21:32:34wodzkeymap files brokes 80 columns rule
21:32:36bertrikJerom, there is no reason to believe it is unstable, we use the same setting as the OF as far as I know
21:33:09bertrikTrying PLLB with the same setting will not make a change probably
21:34:37JeromBecause it's like the controller isn't responding after 2s or so
21:38:59Buschelamiconn: maybe. but if we can speed up the current included decoder with little effort, we should do it
21:39:24 Nick YPSY is now known as Ypsy (
21:41:04JeromWrong arch...
21:42:28pamaurywith PLLB ?
21:42:58JeromWrong arch 0
21:43:32pamauryit probably means that the controller did initialized
21:43:38pamaury*did not
21:47:03Jerompamaury: what about dumping ep0 status every 5s so you can see if it's still configured as it should ?
21:49:07bertrikweird, cube.c #includes two files, but just in different order depending on #if LCD_DEPTH > 1
21:50:02pamauryWhy not
21:54:39bertrikI don't understand it, so I won't touch it for now
21:55:00pamauryAnyway, I just discovered something strange. If I just rolo rockbox after a controller freeze, then it doesn't work. But after reboot it works.
21:55:25pamaurySo the init code does not do a proper hard reset it seems
21:59:35CIA-6New commit by 03bieber (r27182): Theme Editor: Implemented subline rendering, including conditional subline times
22:00:37JeromWhat is a CCU ?
22:01:16wodzhow windows inform usb device that user clicked 'safe remove'? Does it send special control packet to the device?
22:01:32CIA-6r27182 build result: All green
22:03:13pamauryJerom: for duming of registers every 5sec
22:03:34pamauryJerom: Chip Control Unit
22:04:08pamaurywodz: it just flushes verything to the device and probably unlock lun if it's UMS
22:04:33JeromHow do you read logf if you don't get any disconnect int ?
22:04:59pamauryI go in the debug menu
22:05:07pamaury(I disabled any usb driver)
22:08:13pamauryOk enough for today. If you want to play with it Jerom, I don't see anything wrong after 5sec in the registers. The problem must be elsewhere.
22:08:50 Quit mt (Ping timeout: 240 seconds)
22:10:06bieberHas anyone played around with the Theme Editor preview display so far?
22:10:33 Join Strife89 [0] (
22:10:40pamauryNo but I should have a look as your backup mentor :)
22:10:42 Join mt [0] (~mtee@rockbox/developer/mt)
22:13:24CIA-6New commit by 03bertrik (r27183): Remove some redundant #include's
22:14:51bieberHehe, wouldn't hurt ;)
22:14:53CIA-6r27183 build result: 0 errors, 9 warnings (bertrik committed)
22:15:33bieberI've been testing it with the retroWPS theme, and it's great seeing the various pieces finally starting to come together
22:15:51 Join Strife89DS [0] (
22:17:37bertrikDid I really cause those warnings?
22:20:08 Quit Strife89 (Quit: CGI:IRC (Ping timeout))
22:22:42gevaertsbertrik: no
22:23:30 Quit mt (Ping timeout: 252 seconds)
22:23:45Buschelamiconn: can give a try? another faad patch.
22:24:40 Quit z35 (Ping timeout: 265 seconds)
22:26:21 Join z35 [0] (
22:30:57 Quit esperegu (Remote host closed the connection)
22:31:39 Quit bmbl (Quit: Bye!)
22:36:18 Join yelped [0] (
22:37:04yelpedWow! I was just about to update the wiki for the AMS v1s regarding USB HID, and funman beat me to it.
22:37:50simonrvnwhat's the relevant page?
22:38:01simonrvnSansaAMS or some other one?
22:38:38 Quit Jerom (Quit: Leaving.)
22:39:18simonrvncool thanks
22:39:32 Join Jerom [0] (~heidi@
22:39:46CIA-6New commit by 03bieber (r27184): Theme Editor: Stopped conditionals and sublines from triggering newlines and made combo boxes in config panel return text if not used in a conditional
22:41:17CIA-6r27184 build result: 19 errors, 79 warnings (bieber committed)
22:41:41bieberWhoah, how did that happen?
22:41:50bieberI only committed Theme Editor files :O
22:42:09pixelmasomething about this build client I guess
22:42:20gevaertsbieber: another known build issue...
22:42:32simonrvnah ok
22:42:35gevaertspixelma: no, those two are unrelated
22:42:51biebergevaerts: One that I didn't cause, or did I screw something up?
22:42:57***Saving seen data "./dancer.seen"
22:44:03gevaertsr27183 is because stuff in tools/ ignores the fwrite() return value, which some modern toolchains dislike, and r27184 is because there's some weird dependency issue that shows up every now and then, usually for the Player
22:44:06pixelmagevaerts: I didn't say both builds were done by the same client (and I really didn't mean it), your statement confuses me a bit
22:44:41gevaertsah, ok
22:45:05 Quit yelped (Quit: CGI:IRC (EOF))
22:49:35gevaertshm, those r27183 warnings might be avoidable with -U_FORTIFY_SOURCE
22:51:45gevaertsthey are
22:52:10gevaertsShould we add that, or keep the occasional warning as a reminder to fix them eventually?
22:57:33 Quit komputes (Remote host closed the connection)
22:58:51 Quit evilnick_B (Quit: Page closed)
23:04:49 Join rado1 [0] (
23:06:48 Quit Xerion (Quit: )
23:07:34 Nick fxb is now known as fxb__ (
23:09:31CIA-6New commit by 03bieber (r27185): Theme Editor: Fixed conditional viewport display
23:10:45 Nick Ypsy is now known as YPSY (
23:11:22CIA-6r27185 build result: 0 errors, 9 warnings (bieber committed)
23:15:23kisakI have a DLO Transdock Micro, and I'll be interested how AMSv2 support develops
23:15:41kisakit's too scratched up and abused to get a model off it
23:17:18kisakdoes the USB stack need to develop before accessories are able to be considered?
23:20:34gevaertsAny build system experts around?
23:20:51gevaertskisak: I don't know, but I would expect not
23:21:06CIA-6New commit by 03alle (r27186): Add the glyph for the number sign to 19-Nimbus
23:22:41CIA-6r27186 build result: All green
23:33:13 Quit jgarvey (Quit: Leaving)
23:36:46saratogathe NO_PROF_ATTR directive seems broken in SVN righ tnow
23:39:59 Join arbingordon [0] (~username@unaffiliated/arbingordon)
23:41:26saratogawhere is the documentation for the profiler again? i can't find it on the wiki but i know it exists
23:48:25saratogaheh and its now pretty out of date anyway
23:49:22CIA-6New commit by 03ranma (r27187): sd-as3525.c: handle acmd retries correctly
23:50:39saratogastripwax: when you get a chance could you explain to me how to enable profiling? i keep getting messages that profile_thread can't be found, even though i've defined RB_PROFILE and included profile.h
23:50:49CIA-6r27187 build result: 0 errors, 2 warnings (ranma committed)
23:51:27 Quit Strife89DS (Quit: ClIRC - IRC client for Nintendo DS)
23:57:44 Quit domonoky (Ping timeout: 245 seconds)

Previous day | Next day