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 2013-11-23

00:02:44wodzmortalis: (log) please test if g#364 works for you.
00:02:46fs-bluebot3Gerrit review #364 at : 3rk27xx: implement usb driver by Marcin Bukat (changes/64/364/14)
00:02:55 Quit wodz (Quit: Leaving)
00:03:46 Quit Narod ()
00:20:11 Quit benedikt93 (Quit: Bye ;))
00:26:25 Quit krabador (Quit: Sto andando via)
00:30:55 Join krabador [0] (~krabador_@unaffiliated/krabador)
00:49:28 Quit kugel (Ping timeout: 240 seconds)
00:50:12 Join megakacktus [0] (
00:50:26 Part megakacktus
01:10:09bertrikmaybe the radio chip caused a regression, it did at some point in the sansas IIRC
01:10:38[Saint]That wouldn;t explain the Classic.
01:11:10pamaurywould that explain something on the fuze+ ? As far as I know the chip is powered down when unused
01:11:15[Saint]Unless we're actually looking at several regressions.
01:11:32[Saint]Which is highly possible.
01:11:57[Saint]Unfortunately, there was a period about a year ago where commits were added extremely hastily.
01:12:04pamaurybertrik: by the way, not sure you remember, is there a power down mode on the stfm1000 ? Because that might explain the poor battery life on the NWZ: they both have the stfm1000 and no tuner power gpio as far as I can tell
01:12:12[Saint]I believe this is most of the reason why the skin engine is in a giant mess.
01:12:37bertrikpamaury: no I don't remember
01:13:00pamaurywell, a lot could be said and has already been said on the ski engine
01:13:22 Quit bertrik (Remote host closed the connection)
01:13:53[Saint]I try to negate any issues that may arise there these days by using nothing but the fallback theme
01:13:55pamaurythe fact that it breaks usb is very annoying and I should definitely have a look at it but that will mean hard debugging of some code I don't know
01:14:19pamauryI wonder if the use of the MMU for debug purpose could help
01:14:28[Saint]gevaerts and funman did some work in that area.
01:14:33[Saint]skin engine breakages, I mean.
01:14:55pamauryLike having a special define which enhance the memory allocator from kugel with some MMU setup, to trap invalid accesses
01:16:06pamauryI have plenty of ideas in this direction, and I think we could improve the situation much
01:16:32*pamaury wishes he could duplicate himself
01:20:58 Quit chrisjj (Quit: Page closed)
01:32:51 Quit toffe82 (Quit: ChatZilla [Firefox 25.0.1/20131112160018])
01:38:41***Saving seen data "./dancer.seen"
01:40:49 Join treaki__ [0] (
01:43:45 Quit treaki_ (Ping timeout: 272 seconds)
01:58:16 Quit n1s (Quit: Ex-Chat)
02:17:13[Saint],43664.msg221527.html#msg221527 is very interesting.
02:17:18 Nick SuperBrainAk is now known as DormantBrain (~andy@2001:470:8:a61::5f92:59a1)
02:17:41[Saint]The "Prison CLip" apparently has a HW key combo for "nuke firmware".
02:18:29K1773Rand who cleans up the radiation?
02:31:03 Join onder` [0] (
02:46:03 Join chrisjj [0] (561bb732@gateway/web/freenode/ip.
02:49:29chrisjjAnyone still awake? :)
02:50:10[Saint]You know better than that by now - no need to ask such things.
02:50:25[Saint]Ask your question, if anyone can help, they will.
02:50:30chrisjjI'm learning slowly :)
02:51:02chrisjjThanks. Where can I find the daily build of firmware-zen.nk?
02:53:51chrisjjOr for that matter, any build that includes this commit:;a=commit;h=bb8dd053434cbe9c98fe5b9a1095d00ebafb7b61
02:54:00pamaurychrisjj: there is no daily build for firmware-zen.nk
02:54:17pamaurywe don't provide complete daily builds for the bootloaders
02:54:54pamauryhowever if you are looking for the daily build of the main firmware, here it is:
02:55:36pamauryor do you want to update the bootloader ?
02:57:11chrisjjYes, I want to update the bootloader.
02:57:42 Quit krabador (Quit: Sto andando via)
02:57:45pamauryis there an issue with the bootloader ?
02:59:29chrisjjThe issue for which I'm seeking a fix is,13462.msg221277.html#msg221277 .
03:00:39pamauryah I see, ok I will upload a new build at the same address as previously
03:03:15chrisjjThanks A.
03:04:01pamauryI'm uploading, it will take some time with my poor connection here
03:12:27pamaurydamn, one hour expected before completion
03:15:20chrisjjThanks A - I'll look forward to trying it in the morning. Off to bed...
03:16:53 Join krabador [0] (~krabador_@unaffiliated/krabador)
03:27:48 Quit [Saint] (Remote host closed the connection)
03:29:52 Join [Saint] [0] (~saint@rockbox/user/saint)
03:31:12pamaurydammit, dropbox decided to ask for my credential in the middle of the transfer, nice...
03:38:43***Saving seen data "./dancer.seen"
03:43:43 Quit pamaury (Ping timeout: 246 seconds)
03:49:14 Quit krabador (Read error: Connection reset by peer)
03:58:51 Join krabador [0] (~krabador@unaffiliated/krabador)
03:59:08 Quit krabador (Read error: Connection reset by peer)
04:31:36 Quit amiconn (Disconnected by services)
04:31:37 Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:31:40 Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
04:31:48 Quit pixelma (Disconnected by services)
04:31:50 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:31:52 Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
04:52:00 Join qazwsxedc [0] (44bf2bb6@gateway/web/freenode/ip.
04:54:45qazwsxedcHello, I'm having some difficulty installing Rockbox on my player and I didn't find anything helpful in the FAQ:
04:55:43qazwsxedcWhen I try to install, it keeps telling me that I need to provide a version of the device's firmware; however, no guides I found mention this step at all, and their installations don't require it at all.
04:56:20qazwsxedcI have a Sandisk Sansa Clip Zip. I apologize if I missed something in the FAQ but any help would be appreciated.
04:59:34[Saint]The Rockbox manual mentions it clearly, and the installer should point you exactly where you need to go.
05:00:04[Saint]ANy information that doesn't come from Rockbox itself is irrelevant.
05:01:56 Join kevku [0] (~kevku@2001:470:27:773:0:feed:c0f:fee)
05:04:06[Saint]Rockbox Utility should simply be saying "you need an original firmware", it should also provide you with the link to acquire it.
05:04:53[Saint]You should be presented with a page with two downloadable choices, the Sansa Firmware Updater (which we don;t care about), and a Manual Firmware Update - the latter is what you want.
05:06:17[Saint]The Rockbox Clip Zip manual definitely supplies a link to access a firmware file. That much is certain.
05:06:49 Quit TheSeven (Read error: Operation timed out)
05:07:20qazwsxedcThanks a ton, Saint. I'm all set now. I apologize for asking such a simple question; I should have done more thorough research.
05:09:22 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
05:13:52 Join syscrash [0] (~syscrash@poipu/developer/syscrash)
05:24:09 Quit Raptors (Quit: Leaving)
05:35:23 Quit [Saint] (Remote host closed the connection)
05:37:14 Join Raptors [0] (
05:37:28 Join [Saint] [0] (~saint@rockbox/user/saint)
05:38:45***Saving seen data "./dancer.seen"
05:40:19 Quit kevku (Ping timeout: 260 seconds)
05:41:34 Join kevku [0] (~kevku@2001:470:27:773:0:feed:c0f:fee)
05:56:58 Nick DormantBrain is now known as SuperBrainAk (~andy@2001:470:8:a61::5f92:59a1)
06:21:08 Quit qazwsxedc (Quit: Page closed)
06:59:52 Join pretty_function [0] (~sigBART@
06:59:53 Quit pretty_function (Client Quit)
07:05:10 Quit ps-auxw (Remote host closed the connection)
07:20:11 Quit [Saint] (Remote host closed the connection)
07:23:52 Join [Saint] [0] (~saint@rockbox/user/saint)
07:30:07 Quit [Saint] (Read error: Connection reset by peer)
07:30:19 Quit foolsh (Quit: foolsh)
07:31:14 Join [Saint] [0] (~saint@rockbox/user/saint)
07:37:27 Quit cmhobbs (Ping timeout: 272 seconds)
07:38:46***Saving seen data "./dancer.seen"
07:39:51 Quit kiwicam (Remote host closed the connection)
07:50:40 Quit saratoga (Ping timeout: 250 seconds)
07:51:32 Quit JdGordon (Ping timeout: 250 seconds)
07:57:10 Quit chrisjj (Ping timeout: 250 seconds)
09:20:59 Join onder`_ [0] (
09:21:50 Quit onder` (Ping timeout: 246 seconds)
09:21:56 Nick onder`_ is now known as onder` (
09:38:50***Saving seen data "./dancer.seen"
09:42:15 Quit kevku (Ping timeout: 245 seconds)
09:44:33 Join lorenzo92 [0] (
09:45:21 Join bertrik [0] (
09:45:21 Quit bertrik (Changing host)
09:45:21 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
09:55:16 Join Ward [0] (
09:55:39 Nick Ward is now known as Guest71608 (
09:57:27 Join einhirn [0] (
10:03:25 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
10:12:20 Quit y4n (Disconnected by services)
10:12:26 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
10:17:17 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
10:37:10 Join kevku [0] (~kevku@2a01:d0:ffff:34a::8:3)
10:40:35 Join einhirn [0] (
11:10:21 Quit bertrik (Remote host closed the connection)
11:20:28 Nick SuperBrainAk is now known as DormantBrain (~andy@2001:470:8:a61::5f92:59a1)
11:34:36 Join n1s [0] (~n1s@rockbox/developer/n1s)
11:37:02 Quit Guest71608 (Quit: Blarglarg)
11:38:53***Saving seen data "./dancer.seen"
12:12:27copperpamaury: Fuze+ battery bench with revision c4f2a46 from July 23, 2013 (LAME V0, -20 dB, ear buds attached):
12:12:34copper41h 17m :D
12:13:10 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
12:17:56copperso we lost over 10 hours of battery life
12:18:59copperFuze+ with a recent build, 30h 20m:
12:19:50copperI just started a battery benchmark on the Clip+ with the latest revision, lossyFLAC, -24 dB and ear buds attached (like my Classic)
12:20:16copperand after that's completed, I'll start over with the same year old revision as on the Classic
12:23:23 Quit kevku (Read error: Connection reset by peer)
12:34:26 Join kugel [0] (~kugel@rockbox/developer/kugel)
12:34:43 Join kevku [0] (~kevku@2a01:d0:ffff:34a::8:3)
12:47:08 Quit lorenzo92 (Quit: ChatZilla [Firefox 25.0.1/20131112160018])
13:31:32 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
13:32:58n1scopperm some sort of current monitoring will be a lot faster if you're trying to track down what caused the loss
13:33:05n1scopper: ^
13:33:53pamaurycopper: thanks
13:38:23pamauryn1s: if you have a solution we'll be happy to ear it
13:38:55***Saving seen data "./dancer.seen"
13:40:32 Join krabador [0] (~krabador_@unaffiliated/krabador)
13:40:53n1spamaury: saratoga set up some current monitoring for a sansa. I'm not sure of the exact setup but i think he used some rather nice lab equipment but a regular multimeter should be enough to see a ~30% increase in power draw i think
13:41:23pamauryI'm abroad, I don't have access to any equipement
13:41:41coppern1s: a multimeter connected to what?
13:42:45n1scopper: best would be in a battery lead. or if it runs without a battery, on the usb cable
13:42:59pamaurycopper: assuming you setup the power charger correctly, you can probably do it using a modified usb cable and a multimeter
13:43:36copperI'm afraid that's above my pay grade
13:43:54pamaurythe theory is simple: cut the usb cable, put a high precision resistor on the 5V+ line and mesurement the voltage drop
13:44:00n1spamaury: some pmus can give you the current draw too, have you checked if whatever is in the fuze+ can?
13:45:11pamauryyes, it cannot
13:45:15copperI have a cheap (and old) multimeter, but I don't have either a spare USB cable, nor any other means of doing that
13:45:33pamauryif it could we wouldn't be loosing that amount of time on benchmark
13:45:58pamauryor at least I'm not aware of any way to measure it
13:46:15copperI should probably try a test_codec
13:46:39coppererm, maybe not
13:46:41copperI don'tk now
13:47:54copperright now I'm really curious to see if my Clip+ shows a battery life regression too
13:48:31copperif that were the case, could we rule out and target specific commits?
13:48:49pamaurycopper: did you start the test ?
13:48:54copper(the case being, three different targets showing the regression: Clip+, Fuze+ and iPod Classic)
13:49:05copperyeah I started the Clip+ test an hour ago
13:50:10copperah crap
13:50:19copperI should have started the Clip+ benchmark earlier
13:50:28coppernow it's probably going to complete after I go to bed
13:50:49pamauryby the way, mine is finished too
13:52:06pamauryI need to do some additions, I had to stop and restart battery bench several times ^^
13:52:34pamauryso we have a peak battery life in july
13:52:35copperuh? no change from your previous benchmark?
13:53:21pamauryI didn't expected to reach 40h anyway because in january we were missing some of the power management stuff
13:53:41pamauryit's probably chance that the two are 30h
13:54:01copperI don't understand: didn't you get 30 hours with a recent build?
13:54:16pamauryyeah, and 30hour with a very old
13:54:34copperoh, so you didn't run one with the July revision?
13:54:47pamauryno, you did ;)
13:54:50copperah ok
13:55:38pamauryI had you bench the recent one to see if your batteries were comparable and since they are (~30h with HEAD), we can benh different revisions and get consistent results
13:55:45copperif I see the regression on the Clip+, I'll be able to run more benchmarks on that
13:56:02copperI don't really want to run more benchmarks on the Fuze+ and iPod Classic
13:56:12pamauryMy fuze+ is recharging, when it's done I will pick a revision between july and now
13:56:27pamauryAnd run another revision on my second fuze+, do a trichotomy
13:57:51pamauryah good, I *knew* I had bring two micro usb cables, I have been looking for the second one since the beginning of my stay
13:58:10copperthen maybe I could run a couple more benchmarks on my Fuze+: you could tell me what revision to test, while you're testing a different one
13:59:07pamauryas you wish, if you prefer not to run too many benh on your Fuze+ that's no problem
13:59:32pamauryas you wish, if you prefer not to run too many benh on your Fuze+ that's no problem
14:09:29 Quit nosa-j (Ping timeout: 272 seconds)
14:10:40 Join nosa-j [0] (~m00k@
14:30:04 Quit nosa-j (Ping timeout: 246 seconds)
14:41:40 Quit krabador (Quit: Sto andando via)
14:47:30 Join mortalis [0] (~mortalis@
14:48:56 Join nosa-j [0] (~m00k@
14:53:30 Quit kevku (Ping timeout: 245 seconds)
14:53:45 Join ender` [0] (
15:00:17 Quit nosa-j (Ping timeout: 246 seconds)
15:00:53copperpamaury: I'm about ready for another Fuze+ battery benchmark
15:02:05copperer, bbiab
15:08:57 Join nosa-j [0] (~m00k@
15:14:15 Quit nosa-j (Ping timeout: 264 seconds)
15:14:23 Join thomasjfox [0] (~thomasjfo@rockbox/developer/thomasjfox)
15:25:21pamaurycopper: so here is it: you benchmarked c4f2a46, HEAD is at fb8faa1, there are 237 commits in between. I suggest we split evenly in four:
15:25:21pamauryc4f2a46 < a2a2e14 < 6ac481e < 5cfb148 < fb8faa1
15:25:21pamauryI will benchmark a2a2e14 and 5cfb148, and you can do 6ac481e, ok ?
15:25:58 Join nosa-j [0] (~m00k@
15:27:27copperpamaury: deal
15:30:13 Quit nosa-j (Read error: Operation timed out)
15:30:33 Quit K1773R (Quit: /dev/null)
15:30:34n1spamaury: i read up alittel on the imx233 pmu and it seems to be possible to just tell it to run off of 5V no matter what so probably no need to physically alter the device to do the usb cable current measurment thing, perhaps some pmu config changes are needed though, i didn't check the code
15:30:49n1ss/alittel/a little/
15:31:50 Join ps-auxw [0] (~arneb@2001:470:c807:0:1532:4e5f:2ad3:4123)
15:32:08 Quit pamaury (Ping timeout: 246 seconds)
15:33:20 Join K1773R [0] (~K1773R@unaffiliated/k1773r)
15:33:59 Join nosa-j [0] (~m00k@
15:38:58***Saving seen data "./dancer.seen"
15:42:59 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
15:43:25pamauryn1s: yes I know, that's my plan but at the moment I'm abroad and I don't have the tools to build such a cable
15:45:46pamauryI'm wondering, as anyone thought about porting rockbox to qemu ? that way you still get a native target but running inside a box makes it easier to debug
15:46:21pamauryI think there was a GSoC about it, right ?
15:47:37pamauryah it was abandonned
15:47:49 Join matthiaskrgr [0] (matthiaskr@gateway/shell/panicbnc/x-yhiphryeepjsreql)
15:48:17 Part matthiaskrgr
15:51:14pamauryI see to recall there was a path on FS but I can't find it
15:52:43pamaurygevaerts: ^ do you know ?
15:57:25 Quit kugel (Ping timeout: 272 seconds)
16:06:37 Join kevku [0] (~kevku@2001:470:27:773:0:feed:c0f:fee)
16:13:37 Join cmhobbs [0] (
16:13:44 Quit cmhobbs (Changing host)
16:13:44 Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs)
16:24:46 Join wodz [0] (
16:26:00wodzpamaury: I tried running rb in jz-qemu long time ago but I lost interest.
16:26:41pamaurythat's mips ?
16:27:52pamauryMy plan would be to have an arm target with (possibly) dual role usb. I know qemu support usb host and I think implementing usb device with projects usb-vhci is doable
16:28:00wodzpamaury: mortalis checked and ZLP fix didn't fix problems on his box
16:29:36wodzpamaury: What is the purpose of dual role usb qemu target?
16:29:47pamauryplay with usb host ^^
16:30:05wodzoh, I thought you want to fix something :P
16:30:15pamauryyeah I need usb device to fix something
16:30:31pamaurywe know most of the breakage comes in when you plug usb
16:32:30pamauryI'm wondering what is best: 1) port qemu to some rockbox supported target (like imx233) 2) port rockbox so some qemu supported target
16:33:13wodzwait you want to couple two qemu instances one with host and second with device emulations?
16:33:39pamauryno, you have one qemu instance which acts as a virtual usb device
16:33:55pamaurythe problem is that you need to convince linux it's a real device :)
16:35:17pamaury(though coupling two qemu instances like this would be fun)
16:38:00wodzI think implementing rb supported target in qemu is easier after all. From my experience with jz emulation you have to dive into qemu internals anyway since most targets are emulated only in part needed to boot linux in default config
16:38:17wodzWe usually configure hardware differently.
16:38:57pamauryyeah that's true
16:39:27 Quit thomasjfox (Quit: Konversation terminated!)
16:42:08wodzpamaury: Comming back to rk27xx usb I am out of ideas. On my box current driver works reasonably stable, comparably to OF for sure.
16:44:27pamauryhow bad does it fail on mortalis's box ? Under normal circumstances or when spamming the device ?
16:45:02wodzLast time it failed in normal copying of the file
16:46:10wodzI copied a few times in row of varying sizes 10-900MB and all succeeded (checked md5sums)
16:46:47pamaurydid you try with windows ?
16:47:40wodzNo, but I can later today. I happens I have w7 box from my work
16:48:16pamauryI'm out of ideas too, I really don't know what it could be
16:48:58pamaurySince it's random that would suggest memory corruption or cache problem or non-atomic operations being interrupted by an irq, but I can't spot any in the current code
17:00:43 Quit mc2739 (Quit: leaving)
17:06:47copperhow do I get git to revert to the latest revision, after running "git checkout XXXXXXXXXXXXX"?
17:07:02copper"You are not currently on a branch. Please specify which branch you want to rebase against."
17:07:11copper"git pull −−rebase" doesn't work
17:07:22copperit wants me to run "git pull <remote> <branch>"
17:07:34mortaliswodz: pamaury: it fails with the same test as before - dd if=/dev/urandom of=/media/5AEC-DD59/test bs=1M count=1024
17:08:07n1scopper: doesn't git checkout HEAD do it?
17:09:52copperthat command completes successfully but doesn't print anything; and then, git pull −−rebase still doesn't work
17:10:05pamaurycopper: git reset −−hard origin/HEAD
17:10:13pamaurythis will erase all modifications you might have
17:10:27coppergit pull −−rebase still doesn't work
17:10:37copper"You are not currently on a branch"
17:10:56pamauryare you in a rebase or a merge ?
17:11:13copperno idea
17:11:39pamauryeven after the hard reset ? that's strange
17:11:49pamaurytry git rebase −−abort ?
17:11:59mortaliscopper: "git checkout master" probably is what you want
17:11:59copperNo rebase in progress?
17:12:14pamauryah yeah you are not in any branch
17:12:17coppermortalis: thank you, that did it
17:12:51copperok, so the procedure is: "git checkout XXXXXXXXX" to try an older revision, and then "git checkout master" to get back to the latest revision?
17:13:50pamaurythat one possible way
17:14:56wodzmortalis: 1) is HID turned off? 2) Have you tried regular file copy?
17:15:08wodzmortalis: 3) What system?
17:20:19mortaliswodz: 1. yes 2. no 3. What exactly do you want to know?
17:21:15wodzmortalis: win/linux/bsd
17:21:28 Join einhirn [0] (
17:22:02mortalissorry, the first answer was wrong, HID turned on
17:25:06 Join einhirn_ [0] (
17:26:01 Quit einhirn (Ping timeout: 265 seconds)
17:26:21wodzmortalis: Could you try with HID turned off?
17:26:43mortalisyes, i'm trying it now
17:28:00mortaliswodz: failed
17:28:45mortalisand device hang
17:29:52 Quit pamaury (Ping timeout: 252 seconds)
17:38:12 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
17:39:02***Saving seen data "./dancer.seen"
17:52:52 Quit mortalis (Ping timeout: 248 seconds)
17:54:13 Join mortalis [0] (~mortalis@
18:02:28mortaliswodz: ihifi 760 passed dd test
18:03:03wodzmortalis: thats interesting
18:04:36mortalisiirc on ihifi rk2705
18:04:43mortalis2706 on hifimans
18:05:18mortalismaybe that the reason of different behavior
18:06:17wodzmortalis: Who knows. AFAIK pamaury's rk dap is 2706A
18:06:53mortalisand now it failed
18:09:07 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
18:12:26 Join Narod [0] (
18:17:54pamaurycopper: bench started, I had to switch to 78c060b for the second one because the build was broken
18:23:45coppercurrently running 6ac481e
18:27:42 Quit wodz (Ping timeout: 246 seconds)
18:38:02 Join rela [0] (~x@pdpc/supporter/active/rela)
18:44:17 Quit rela (Ping timeout: 245 seconds)
18:50:27 Quit Narod ()
18:51:57 Quit cmhobbs (Ping timeout: 272 seconds)
18:59:15 Join Narod [0] (
19:03:16 Quit nosa-j (Ping timeout: 248 seconds)
19:05:48 Join nosa-j [0] (~m00k@
19:17:55 Join rela [0] (~x@pdpc/supporter/active/rela)
19:21:19 Nick DormantBrain is now known as SuperBrainAk (~andy@2001:470:8:a61::5f92:59a1)
19:37:42 Nick syscrash is now known as jvd (~syscrash@poipu/developer/syscrash)
19:39:05***Saving seen data "./dancer.seen"
19:40:07 Part jvd
19:45:01 Join wodz [0] (
20:05:11 Quit mortalis (Remote host closed the connection)
20:11:56 Join Tukeke [0] (~Tukeke@unaffiliated/hlvsv)
20:15:02 Quit einhirn_ (Quit: Miranda IM! Smaller, Faster, Easier.
20:16:14 Quit gevaerts (Ping timeout: 245 seconds)
20:16:20 Quit evilnick (Ping timeout: 246 seconds)
20:19:15*pamaury spots some dubious calculations in buflib
20:20:33pamauryah I see, the code doesn't what the doc says, nice
20:20:33 Join lorenzo92 [0] (
20:37:20copperthat would be too easy, wouldn't it?
20:38:00 Quit y4n (Quit: PÆNTS ØLF!)
20:42:44 Quit nosa-j (Ping timeout: 265 seconds)
20:43:25 Join nosa-j [0] (~m00k@
20:46:33 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
20:47:28 Join gevaerts [0] (~fg@rockbox/developer/gevaerts)
20:52:00 Quit nosa-j (Ping timeout: 272 seconds)
20:53:25 Join nosa-j [0] (~m00k@
20:58:10 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
20:58:35 Quit mc2739 (Client Quit)
21:02:26 Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs)
21:13:39 Quit gevaerts (Ping timeout: 264 seconds)
21:13:40 Join evilnick_ [0] (
21:13:44 Quit evilnick (Ping timeout: 245 seconds)
21:15:33lorenzo92is the recording level trigger screen active even when no encoders are available? i.e. can I see some level monitoring when speaking into the microphone?
21:15:36 Join gevaerts [0] (~fg@rockbox/developer/gevaerts)
21:17:09wodzlorenzo92: at least on coldfires you can
21:17:38lorenzo92thanks, because I have to debug audio in without having any storage yet ^^
21:24:28 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
21:25:00 Quit gevaerts (Disconnected by services)
21:25:04 Join gevaerts_ [0] (~fg@rockbox/developer/gevaerts)
21:25:34 Nick gevaerts_ is now known as gevaerts (~fg@rockbox/developer/gevaerts)
21:26:12 Quit evilnick_ (Ping timeout: 272 seconds)
21:26:35 Join kiwicam [0] (~quassel@
21:39:03 Quit gevaerts (Read error: Operation timed out)
21:39:09***Saving seen data "./dancer.seen"
21:39:13 Join evilnick_ [0] (
21:39:32 Join gevaerts [0] (~fg@rockbox/developer/gevaerts)
21:40:24 Quit evilnick (Ping timeout: 272 seconds)
21:41:22 Quit cmhobbs (Ping timeout: 245 seconds)
21:49:19 Quit gevaerts (Disconnected by services)
21:49:22 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
21:49:24 Join gevaerts_ [0] (~fg@rockbox/developer/gevaerts)
21:49:49 Nick gevaerts_ is now known as gevaerts (~fg@rockbox/developer/gevaerts)
21:50:00 Quit evilnick_ (Ping timeout: 246 seconds)
22:02:12 Join foolsh [0] (
22:03:21 Quit bluebrother (Disconnected by services)
22:03:26 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
22:07:00 Quit fs-bluebot (Ping timeout: 272 seconds)
22:08:34 Join fs-bluebot [0] (
22:24:11 Join shadows [0] (lucent@unaffiliated/shadows)
22:39:03 Quit Narod ()
22:46:30wodzhmm, why my commits didn't trigger buildround?
22:46:52pamaurysometimes the bot is late
22:47:43wodzbuild status also doesn't see the commits
22:48:28pamaurythere is something seriously wrong with gcc: if(a >= -b) {}, a is of type intptr_t and b of type size_t, gcc warns against signed vs unsignec comparison
22:48:40pamauryso now, someone, explains me how on earth is -b unsigned
22:49:12gevaertsMaybe it first optimises that to a<b?
22:49:20gevaertsWell, I hope not
22:49:36*gevaerts hopes gcc has had more coffee than he had today :)
22:50:23gevaertsAnyway, I'd say something that's always negative is just as unsigned as something that's always positive
22:51:35 Join Narod [0] (
22:51:35 Join Gallomimia [0] (~gallomimi@
22:51:52wodzerr, what? -(-3) vs -(3)
22:53:52wodzbtw. playing with rk27xx usb patch I sometimes (read as quite frequently) trigger backlight state corruption
22:53:58wodzon disconnect
22:55:05pamaurythe usb core uses an inherently unsafe dma on rk27xx, you cannot limit the amount of received data
22:56:37wodzheh, true
22:58:04 Join cmhobbs [0] (
22:58:04 Quit cmhobbs (Changing host)
22:58:04 Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs)
22:58:10wodzbut still under normal conditions corruption should be pretty rare
22:58:23shadowswhat is the suggested filesystem to use on large (128gb) microsd storage?
22:58:49wodzshadows: If you mean rockbox you have no choice. rb supports only fat
22:59:04shadowsah okay
23:00:21wodzpamaury: (signed vs unsigned) more interesting is if emitted assembly is correct.
23:01:21pamauryif it fails on such a trivial statement, I'm wondering whether we should really use it :)
23:02:14 Quit kiwicam (Remote host closed the connection)
23:03:03wodzpamaury: Any alternative? clang would be nice but f***sup interrupts handlers on arm
23:03:26wodznot speaking about support for SH and coldfire
23:04:33pamauryyou should try to add support for interrupt handlers on arm in llvm then ;)
23:04:45wodzpamaury: Do you have any comments about 'structural' changes to hwstub in g#667
23:04:47fs-bluebot3Gerrit review #667 at : 3hwstub rk27xx port WIP by Marcin Bukat (changes/67/667/3)
23:07:26pamaurynot sure what is the status but it appears clang may support it now
23:07:37pamaurylet me see
23:08:10wodzbtw. my dap just passed dd if=/dev/urandom of=/media/845B-E8E3/test bs=1M count=1024 with write speed 1.9MB/s
23:09:24pamaurywodz: you should remove the #if 0 code about HWSTUB_INFO_STMP in target.c
23:09:59pamauryif for some reason it may be useful to know details about the rk27xx revision, you can create your own request for target information
23:10:13pamaurylike HWSTUB_INFO_RK27
23:15:57wodzpamaury: for now I can think of returning known rk27xx variant (A or B). The code for this is in place but I don't know it is worth separate command for now
23:16:49pamauryI'm reviewed the patch, mostly it would be nicer if udelay() was moved to target.c and named target_udelay() to match target.h
23:17:01pamauryas you wish
23:18:17wodzmaybe make this request more general as HWSTUB_INFO_SOC or something and return target specific info (regardless of how it is structured)
23:18:18 Quit rela (Read error: Connection reset by peer)
23:18:59pamaurywodz: yeah that's would make sense since there is TARGET to know the target
23:19:13pamaurythat would be part of another commit preferably
23:21:21wodzpamaury: The patch is based on version where target_udelay() was not in target.h
23:21:46wodzpamaury: I need to sync to see what changed as well
23:21:47pamauryah, hum, then maybe I forgot to commit it ?
23:22:01wodzno, see it
23:23:45 Quit kevku (Quit: KVIrc 4.3.1 Aria
23:24:20pamauryapparently interrupt are supported by clang now
23:24:28pamauryas of october, 1
23:24:44wodzhwstub is good candidate to test :-)
23:27:01wodzpamaury: what about 'require dumper' in load.lua ?
23:28:01pamauryit's generic
23:28:20pamauryit dumps every register it can find provided you loaded a register description for your target
23:30:31pamaurywodz: and the corresponding llvm commit:
23:30:42pamauryI would never have been able to write this myself oO
23:32:24 Join kiwicam [0] (~quassel@
23:32:59wodzargh, sorry for gerrit spam
23:33:58 Quit Gallomimia (Read error: Connection reset by peer)
23:34:38pamauryhum, valgrind doesn't like our simulator
23:34:50pamauryit fails hard with Thread creation failed. Retryingmake_context(): Operation not permitted
23:36:13wodzdd test passed 3 times in row
23:36:54gevaertspamaury: maybe try it with sdl threads?
23:38:44pamauryyeah i'm recompiling
23:39:11***Saving seen data "./dancer.seen"
23:39:42 Quit cmhobbs (Ping timeout: 245 seconds)
23:39:59 Join Gallomimia [0] (~gallomimi@
23:40:17[Saint]pamaury: I'll admit I skipped the logs a little - but - if you're looking for an ARM qemu supported target to port Rb to - raspberrypi?
23:40:25[Saint]Costs $25USD
23:40:54pamauryno actually I will rather add support for the imx233 in qemu
23:41:01*[Saint] nods
23:41:19pamauryI found out that someone at olinuxino wrote some code, maybe I can pickup on this and finish it
23:42:26shadowshm rpi as rockbox target ? interesting.
23:45:59[Saint]As long as we _completely_ ignored the GPU, it should be perfectly possible.
23:47:12wodz[Saint]: you cannot ignore GPU on pi as it is used to bootstrap :P
23:47:59[Saint]Right, I worded that incorectly. As long as we ignored the GPU past anything the released videocore binary gives access to.
23:48:44gevaertsjust run raaa on it, and you don't have to worry about such things
23:48:51[Saint]It would be more suited as a hosted application target, but it /should/ be possible to run as a native OS on the pi.
23:49:12wodzpamaury: Do we really need udc_helper() in case of hwstub? It was used to notify rb usb stack about address and configuration. AFAIK it is not needed in hwemul
23:49:39pamaurynot it's not needed in hwstub
23:50:02[Saint]Raaa, for some reason, runs absolutely terribly on Raspbian at least.
23:50:24[Saint]I can't say I've looked into why.
23:50:31gevaertsAnything runs absolutely terribly on a raspberry pi, so that doesn't mean much
23:50:33pamaurygevaerts: sdl threads work
23:50:50[Saint]gevaerts: that's not /entirely/ true. ;)
23:51:09pamauryit's slow as hell though, I guess valgrind accounts just a little for that ;)
23:51:36n1s[Saint]: does it enable the asm code?
23:52:16pamauryunsurpringly my buflib rework fails miserably
23:53:34[Saint]n1s: I couldn't say - its been a long while since I looked into this. I recall it didn't like compiling inititally but I think that was more due to the spotty state of the raspbian repositories about a year ago. It required a small amount of hand holding. I haven't tried it recently.
23:54:18n1sit's likely to make a big difference for the codecs at least
23:55:10n1sbut otoh the thing should be faster than the gigabeast so that shouldn't matter too much
23:55:26[Saint]The raspi could make a damn nice target if I was even vaguely interested in fighting with HW acceleration via the GPU.
23:55:43*[Saint] wishes he had more motivation to try things.
23:57:17[Saint]~$25 "open" audio player with a kickass GPU and a shit-tonne of ARM optimizations gathered over the course of a decade.
23:57:24[Saint]Damn I wish I wasn't boring...
23:58:40pamauryexcept that supporting the kickass GPU probably involves more code that rockbox itself :)
23:58:50n1si think i agree with gevaerts that raaa should be the way to go, but maybe it needs some work

Previous day | Next day