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

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

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

#rockbox log for 2014-11-14

00:00:15[Saint]it is _very_ strange that not even the fallback image provides a reliable USB connection.
00:00:38Cunctator_i doing another try, just to be sure...
00:01:10Cunctator_yeah, same. Removable Disk F: impossible to open
00:01:58copperreboot the PC
00:02:09Cunctator_? yeah?
00:02:56[Saint]Yeah. Windows is a strange creature. A reboot may well be advisable here.
00:02:59Cunctator_is this move [Saint] approved? :)
00:03:34 Quit Cunctator_ (Quit: Page closed)
00:04:28 Quit edhelas (Ping timeout: 250 seconds)
00:06:15 Quit fs-bluebot (Ping timeout: 240 seconds)
00:06:55 Join Cunctator [0] (4f2a005d@gateway/web/freenode/ip.
00:07:02Cunctatorand im back
00:07:06Cunctatorcomputer freshly rebooted
00:07:21Cunctatorshould I also reboot Rockbox, just to be sure?
00:09:15Cunctator(so before showing the main menu it says ; no rockbox installation found or 'incomplete')
00:09:31[Saint]thats expected.
00:10:12Cunctatoralso under System>ROckbox info it says : Version r31516-120101
00:10:13[Saint]Its just complaining that there's no user installed Rockbox image, which is totally fine.
00:10:20[Saint]As it will just load the fallback.
00:10:53Cunctatoralright, trying again with the USB cable now
00:11:19[Saint](which can't play media, as it has no codecs, and no plugins, it exists just to facilitate installation and a (supposedly) known stable USB connection)
00:11:58Cunctatornope same stuff. Oh joy.
00:12:20Cunctatordo you think that the fact that i still had that "iPod Controls" folder has anything to do with this all?
00:12:32Cunctatoryou seemd to think it was not normal
00:13:04[Saint]I'm not sure why it was there, but it definitely shouldn;t be there now. And the issue is still presenting itself.
00:13:50[Saint]This is all quite strange. I wonder, what driver is being used when the device is connected?
00:14:10Cunctatorwell yes if anything im worse off :)
00:14:13Cunctatorwhat do you mean?
00:14:36 Join fs-bluebot [0] (
00:14:49[Saint]What is it presenting itself as in the device manager, and what driver is used for the device.
00:17:18CunctatorOK says: portable device, Apple, location generic volume. Driver 6.3.9600.16384
00:17:38[Franklin]I don't think that's rockbox's driver
00:17:58[Saint]Lets uninstall that driver.
00:18:26Cunctatorand then?
00:18:31copperany chance my build with the phony USB ID would work?
00:19:19[Saint]eject and re-plug the device and see what it wants to do with it.
00:20:05[Franklin]ugh... I can't figure out how to get xworld to draw the screen pre-rotated
00:20:21[Franklin]so now it has to draw it, and then rotate it
00:20:39[Saint]since its going to be entirely irrelevant to you (in theory, if we get it working for you), you'll probably want to uninstall iTunes entirely as its of no use.
00:21:00[Franklin]it would save memory, and speed up the code
00:21:08[Franklin]but I can't figure it out
00:21:16[Saint]that's definitely one way of guaranteeing itunes isn't getting in the way.
00:21:25CunctatorOK on it
00:21:31[Saint](and, its a piece of shit ;))
00:22:21 Quit xorly (Ping timeout: 258 seconds)
00:24:07Cunctator(still uninstalling...)
00:24:30copper[Franklin]: needs an entry here
00:24:46[Franklin]copper: sure thing :)
00:25:31[Franklin]actually, maybe it should wait
00:25:40[Franklin]until it gets committed (if ever...)
00:26:19[Franklin]ok I got rotation working
00:27:27CunctatorOK, after the longest uninstall process ever, iTunes is out of the way. Need I reboot again?
00:27:46[Franklin]actually, no
00:27:59[Franklin]it should have terminated the process in the uninstall
00:28:06[Franklin]but still, reboot to remove the driver
00:28:12Cunctatorlets err on the side of caution. Back in a minute
00:28:19*[Saint] nods
00:29:03copperomg, I'm having flashbacks of Flashback
00:29:40copperboth 2D and 3D games
00:29:52[Franklin]aww but its reversed
00:31:16 Join Cunctator_ [0] (4f2a005d@gateway/web/freenode/ip.
00:31:24Cunctator_aaand, i'm back
00:31:27Cunctator_awaiting instructions
00:31:29[Saint]heh - I vaguely recall that.
00:32:34[Saint]Cunctator_: ok, well now the host OS should be treating it as the removable storage volume we'd like it to.
00:32:39 Quit Cunctator (Ping timeout: 246 seconds)
00:32:50Cunctator_lets see
00:33:07[Saint]It /should/ appear as 'iPod CLassic' or so, and the driver refer to Rockbox.
00:33:38 Quit bertrik (Remote host closed the connection)
00:33:48[Saint](i think...)
00:35:32Cunctator_no. Its still a Removable Disk (and cant be opened). The Device manager now says manufactuer Microsoft instead of Apple. and the driver number is 6.3.9600.16384 (which I dont remember if is the same as before or not)
00:36:26[Saint]this is...shennanigans.
00:36:50Cunctator_tell me about it...
00:37:10Cunctator_I'm super grateful to you, but I'm starting to lose hope
00:37:33[Saint]I really don't know how to go forward at tis point.
00:37:48[Saint]All the things that absolutely should work haven't.
00:38:24[Saint]So I'm not entirely sure what else to do other than guide you though restoring the original firmware so you have a functional device to use.
00:38:41***Saving seen data "./dancer.seen"
00:39:07Cunctator_that'd be better than nothing yes. I use Mediamonkey as a music manager, I can sync it with that and that's it. Until a stable release comes out (if ever?)
00:40:28[Saint]The type of debugging that this requires is beyond my ability and likely requires the oversight of TheSeven in #freemyipod-support (but he is super busy and getting a response from him may take some rather extended periods of idling in the channel as he's flat out with work and school)
00:41:11Cunctator_yeah...I'm not blaming anyone, dont worry. Just frustrated that of all people with working 7g classics I had to be the one :(
00:41:27[Saint]Ok. We'll want
00:41:59[Saint]That points to a very specific version of iTunes to restore the device with, as other versions are known to be problematic.
00:42:12Cunctator_crap. I have to re-download iTunes :(
00:42:21[Franklin]Cunctator_: sorry
00:42:36[Franklin]Cunctator_: you could always give it another try someday
00:42:49[Saint]I'm very sorry this has happened to you. But as I say, the level of debugging this would now require is far beyond my skill.
00:43:17Cunctator_crazy idea: what if I do the whole uninstall procedure and tomorrow try once again from scratch
00:43:29[Saint]It couldn't hurt.
00:43:44[Franklin]or do it again now!
00:43:52Cunctator_it doesn't? does it not "tire" the HD to do this over and over?
00:44:12Cunctator_well guys, it's nearly 1am over here, and I'll need to sleep rather soon :)
00:44:56copper → Flashback, Amiga - Part 1 - Continuing Classics - YouTube | Original URL:
00:45:09copper → Fade to Black (PC DOS) - YouTube | Original URL:
00:45:23[Saint]Cunctator_: Right. Well if you have any issue with the uninstallation, and restoration of the original firmware, whenever you get around to it, give me a yell.
00:45:30[Saint]Or, someone here.
00:45:32 Quit ZincAlloy1 (Quit: Leaving.)
00:46:24[Saint]And, again, I'm terribly sorry for this experience.
00:46:33Cunctator_OK, well, tomorrow morning I'll go back to the original firmware. Then I'll be away from home for a day or two, so I'll probably try again on Sunday
00:46:36Cunctator_nah, don't worry
00:47:10Cunctator_I mean, Rockbox is pretty cool, and I just wanted it to work. But worst case scenario, I'll just have an iPod
00:47:22Cunctator_just bought it because i heard it was getting discontinued
00:47:24[Saint]Ok. Well, feel free to try and get hold of me if you need to.
00:47:29Cunctator_and its hard to find players this big
00:47:33Cunctator_I will
00:47:49Cunctator_you've been...wait for it....a real Saint
00:49:04Cunctator_well, have a nice day/evening everyone. I'll be back in this corner of the internetz in a few days.
00:49:10Cunctator_and thanks again
00:51:02 Quit Cunctator_ (Quit: Page closed)
01:02:53 Quit petur (Quit: Leaving)
01:09:04 Quit Scr0mple (Read error: Connection reset by peer)
01:09:21 Join Scr0mple [0] (~Simon@
02:03:14[Franklin][Saint]: is there any way us lowly non-commiters can change the theme of the gerrit pages?
02:03:42[Franklin]the gerrit page seems really out-of-place when compared with the other pages
02:04:06[Franklin]It doesn't seem to be on www.git
02:37:51 Quit krabador (Quit: Sto andando via)
02:38:42***Saving seen data "./dancer.seen"
02:43:03 Quit bluebrother^ (Ping timeout: 264 seconds)
02:43:36 Quit fs-bluebot (Ping timeout: 272 seconds)
02:44:39 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
02:46:18 Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs)
02:50:02 Join fs-bluebot [0] (
02:55:03 Quit [Franklin] (Remote host closed the connection)
03:00:02 Quit AlexP (Remote host closed the connection)
03:02:21 Quit cmhobbs (Ping timeout: 245 seconds)
04:21:33 Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs)
04:31:17 Quit cmhobbs (Ping timeout: 272 seconds)
04:38:45 Quit pixelma (Disconnected by services)
04:38:45 Quit amiconn (Disconnected by services)
04:38:46***Saving seen data "./dancer.seen"
04:38:46 Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:38:46 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:38:48 Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
04:38:50 Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
04:44:22 Join bzed_ [0] (
04:46:05 Quit bzed (Ping timeout: 255 seconds)
04:46:05 Quit mazzoo (Ping timeout: 255 seconds)
04:46:10 Nick bzed_ is now known as bzed (
04:46:12 Join mazzoo [0] (
04:46:37 Quit byteframe_ (Ping timeout: 265 seconds)
04:47:33 Join byteframe_ [0] (~byteframe@unaffiliated/byteframe)
04:58:23 Quit pystar89 (Ping timeout: 244 seconds)
05:08:43 Quit bluebrother (Ping timeout: 244 seconds)
05:08:53 Quit fs-bluebot (Ping timeout: 264 seconds)
05:37:40 Quit [Saint] (Read error: Connection reset by peer)
05:38:33 Join [Saint] [0] (~saint@rockbox/staff/saint)
05:50:03 Quit TheSeven (Disconnected by services)
05:50:07 Join [7] [0] (~quassel@rockbox/developer/TheSeven)
05:59:23 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
05:59:42 Join Strife89 [0] (
06:10:43 Quit ruskie (Ping timeout: 265 seconds)
06:12:28 Quit prof_wolfff (Ping timeout: 264 seconds)
06:25:25 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
06:30:34 Quit Strife89 (Ping timeout: 250 seconds)
06:38:50***Saving seen data "./dancer.seen"
06:57:32 Join pystar89 [0] (
07:01:08 Quit dfkt (Read error: Connection reset by peer)
07:01:45 Join dfkt [0] (dfkt@unaffiliated/dfkt)
07:08:30 Join ruskie [0] (ruskie@sourcemage/mage/ruskie)
07:24:37 Join chrisb [0] (
07:29:01 Quit chrisb (Ping timeout: 245 seconds)
07:29:17 Join chrisb [0] (
07:40:48 Join xorly [0] (
08:09:47 Join ender` [0] (
08:27:04 Join prof_wolfff [0] (
08:29:15 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
08:30:44 Quit chrisb (Ping timeout: 244 seconds)
08:33:15yurikswish I knew what happened to USB
08:33:44yuriksI'm reduced to using the OF for the Clip+, it's the only way of getting usb access
08:34:21yuriks3.13 still randomly freezes or gives windows a "your USB device malfunctioned"
08:34:30copper3.13 is _old_
08:34:34yuriksI even bought a new clip+ thinking mine was flaky and still having the same problem
08:34:46yurikscopper: well, same thing with several dev versions (either nightly or compile myself)
08:34:58yuriksand it used to work just fine a year ago or so
08:35:22yuriksit's like there's a ghost haunting me or something, it's really the only explanation I have since I changed everything else :P
08:36:01yuriks(multiple cables, multiple USB ports, multiple PCs, multiple Clips, with or without a SD card)
08:36:18yuriksguess I should try 3.12 to see how it fares
08:36:31yurikshmm, true, both are on windows 8 now
08:37:21yuriksthough I *think* it was also working fine on windows 8 in the past, so it's probably 8.1 specifically
08:37:36yuriks(and I think they messed with the USB stack on 8.1 so that would make sense)
08:38:48yuriksbah, can't use 3.12 since I have opus, well worth a try anyway
08:38:52***Saving seen data "./dancer.seen"
08:43:06yuriksyeah, 3.12 is a no go
08:50:00 Join edhelas [0] (
09:04:55 Join mortalis [0] (~kvirc@
09:05:52 Quit pamaury (Ping timeout: 240 seconds)
09:10:21 Join petur [0] (
09:10:25 Quit petur (Changing host)
09:10:25 Join petur [0] (~petur@rockbox/developer/petur)
09:14:48 Quit edhelas (Ping timeout: 245 seconds)
09:22:18 Quit xorly (Ping timeout: 255 seconds)
09:25:10 Quit dfkt (Remote host closed the connection)
09:29:08 Join dfkt [0] (dfkt@unaffiliated/dfkt)
09:35:22 Quit akaWolf (Ping timeout: 240 seconds)
09:37:48 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
09:41:24 Join foolsh [0] (
09:49:02 Join edhelas [0] (~edhelas@
09:58:43 Join Cunctator [0] (4f2a005d@gateway/web/freenode/ip.
09:59:07Cunctatorhey [Saint] any chance you still around?
10:15:24copperCunctator: can you try this build, just in case?
10:21:18Cunctatorhey copper
10:21:29Cunctatorwell to be honest I'm a bit concerned about my iPod now
10:21:38Cunctatori restored it to Apple firmware
10:21:57Cunctatorbut now if I open it, it is full with folders and files named more or less like this : ₧ºo─█n╞
10:22:13Cunctatorlike, dozens of them
10:22:36Cunctatorthe name of the drive too is like "₧ºo─█n╞ (F:)"
10:22:44Cunctatori dont think this is a good sign
10:22:54gevaertsIt's not indeed.
10:25:25Cunctatorhave I managed to break a brand new ipod in 2 days? Awesome. Can I get a prize for that?
10:25:46gevaertsIf you have treated the ipod nicely (unplugged safely and so on) since restoring, I'd say it's getting more likely that there's some hardware issue somewhere
10:26:46gevaertsDoes the classic's OF have a diagnostics mode?
10:27:03*gevaerts wouldn't retry emcore/rockbox before checking things out properly with the OF
10:27:18Cunctatornot do i find out?
10:27:50 Quit mc2739 (Ping timeout: 272 seconds)
10:28:02gevaerts seems to have some hints
10:29:14 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
10:32:18Cunctatorumm, i have a large number of Reallocs (1384) but zero Pending Sectors
10:32:21Cunctatorwhatever that means
10:34:10gevaertsAre there tests you can run?
10:34:28*gevaerts doesn't have a classic, so he only knows about rumours and what he can quickly find...
10:34:39Cunctatorlooks like this is the only diagnostic screen i can get
10:35:53Cunctatoryeah it's OK you're already kind to help. I just simply cannot believe the HDD got fucked with 2 days of playing around
10:36:13Cunctatoreither I got the one faulty item out of thousands
10:36:30Cunctatoror I managed to screw up big time with the installation of Rockbox
10:36:37Cunctatorwhich, since it was working, I can't see how
10:37:07gevaertsI don't see how emcore or rockbox could have caused the sort of issues you're having
10:37:18gevaertsIt's either a RAM issue or a disk issue I'd say
10:37:47gevaertsSo I suspect you got unlucky and got the faulty item...
10:38:17CunctatorI guess this is what happens when you buy from ebay
10:38:37Cunctator(again, it was new, still in the original plastic wrap)
10:38:54***Saving seen data "./dancer.seen"
10:38:55gevaertsI'm not sure the wrap proves much
10:39:45Cunctatorwell, both that and the little protective plastic cover in which the ipod itself is stored. I mean it looked absolutely identical to a previous new ipod I bought in a store years ago
10:39:57Cunctatorif it was actually not new, they did an awesome job in making it look like
10:41:55gevaertsOf course, maybe it's as simple as it having been dropped in the wrong way once
10:42:05gevaertsPackaging can only do so much to protect things
10:42:47Cunctatornow whats particularly enraging is that i bought this when I heard they were stopping producing it
10:42:55Cunctatorbecause I really need a large storage space
10:43:08Cunctatornow im fucked. Not only this is broken, but I cannot even get a new one
10:43:45*gevaerts would talk to the seller first
10:45:40CunctatorI'm not sure the electronics shop in Singapore would give a damn...
10:46:19gevaertsYes, there is that. I'd still contact them though, you never know...
10:47:18Cunctator!!! hold on! I just restored once again through iTunes. The weirdly named folders are all gone, and the name of the drive is back to normal!
10:47:35gevaertsWill it stay that way though?
10:48:38gevaertsIf you keep having problems, try catching [7] (or TheSeven, his nick varies) either here or in #freemyipod to figure out if it's definitely the disk. If you know which specific part is bad, replacing that might be an option (although that would mean investing even more money)
10:49:45Cunctatoralright I'll try
10:49:48Cunctatorthanks a lot
10:50:11Cunctatorhaving some support helps delaying despair :)
11:11:24 Quit Cunctator (Ping timeout: 246 seconds)
11:11:50 Join Cunctator [0] (4f2a005d@gateway/web/freenode/ip.
11:15:21 Quit Provel (Ping timeout: 240 seconds)
11:16:21 Join Provel [0] (
12:29:15 Quit foolsh (Remote host closed the connection)
12:35:09 Join foolsh [0] (
12:38:58***Saving seen data "./dancer.seen"
12:38:59 Join akaWolf [0] (~akaWolf@unaffiliated/akawolf)
12:47:35 Join JdGordon_ [0] (
12:48:08 Quit JdGordon (Ping timeout: 245 seconds)
13:14:45 Quit kugel (Quit: leaving)
13:46:10 Join AlexP [0] (~alex@rockbox/staff/AlexP)
13:54:51 Quit shamus (Ping timeout: 240 seconds)
13:55:15 Join shamus [0] (
14:07:39 Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs)
14:13:43 Join wodz [0] (
14:20:03 Quit Cunctator (Ping timeout: 246 seconds)
14:21:57wodzpamaury: ping
14:25:41 Quit petur (Ping timeout: 245 seconds)
14:25:43wodzecho "0x40 0x80 0xe0 0x00"|Debug+Asserts/bin/llvm-mc -arch=mips -mcpu=mips32r2 -disassemble gives mtc0$zero, $gp, 0 which means that it can output symbolic register names
14:31:04 Quit foolsh (Ping timeout: 264 seconds)
14:32:09 Join petur [0] (~petur@rockbox/developer/petur)
14:32:58 Join foolsh [0] (
14:33:25 Quit wodz (Quit: Leaving)
14:39:00***Saving seen data "./dancer.seen"
15:09:45 Join Strife89 [0] (
15:18:03 Join einhirn [0] (
15:25:26 Join PurlingNayuki [0] (~yzflcyq@
15:26:19PurlingNayukiI just upgraded to Debian Jessie and rebuild the development environment using
15:26:46PurlingNayukiHowever it just tells me that libtool is missing, while libtool package is already installed.
15:26:55PurlingNayukiAny advice?
15:27:36 Join amayer [0] (
15:33:22PurlingNayukiSeems that solved. Besides libtool package, libtool-bin should also be installed.
15:41:57 Part PurlingNayuki
15:42:49 Part LinusN
16:08:11 Quit hob (Ping timeout: 245 seconds)
16:24:11pamaurywodz (logs): $zero and $gp are the only two exception in the 32-bit general purpose registers, all others are printed in the form $n
16:24:14 Quit Strife89 (Ping timeout: 250 seconds)
16:24:48pamaurywell not exactly, there are $sp, $fp and $ra too
16:29:19 Join n1s [0] (~n1s@rockbox/developer/n1s)
16:39:02***Saving seen data "./dancer.seen"
16:39:02 Quit mortalis (Quit: KVIrc 4.3.1 Aria
16:51:23pamaurywodz (logs): I have solved the encoding problem, it was my mistake, I misunderstood how the memory operaend was encoded by llvm
16:52:27 Join ZincAlloy [0] (
16:53:22pamaurywodz (logs): updated diff pamaury/7386b1765657f7ae7bac">
16:54:06 Quit petur (Quit: *plop*)
17:03:46 Quit edhelas (Quit: Quitte)
17:31:41 Quit byteframe_ (Quit: Deuces.)
17:32:16 Join byteframe [0] (~byteframe@unaffiliated/byteframe)
17:46:46 Join Cunctator [0] (5d2dd6db@gateway/web/freenode/ip.
17:55:45 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
17:57:43 Quit K1773R (Ping timeout: 245 seconds)
18:03:26 Join K1773R [0] (~K1773R@unaffiliated/k1773r)
18:09:58 Join einhirn [0] (
18:39:03***Saving seen data "./dancer.seen"
18:39:47 Join wodz [0] (
18:40:12wodzpamaury: Did you publish the patch on llvm ML or something?
18:40:21pamaurynot yet
18:40:48pamauryI guesss I need to implement assembly parsing for the instruction before
18:41:01pamauryand maybe more
18:41:16wodzpamaury: What was wrong with previous version?
18:41:24pamauryencoding was wrong
18:41:51 Quit Cunctator (Ping timeout: 246 seconds)
18:42:13pamaury(which explained your remark yesterday about decoding being correct but printed encoding being wrong)
18:42:15wodzah, this one
18:43:05 Join krabador [0] (~krabador_@unaffiliated/krabador)
18:44:44pamauryalso you read my comment about registers ?
18:53:14 Join chrisb [0] (
18:54:26 Quit wodz (Ping timeout: 245 seconds)
18:55:37 Join wodz [0] (
18:55:44wodzpamaury: yes
18:56:55pamauryI'll see if I can implement an asm variant with more common names
18:58:03 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
19:12:13 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
19:12:47 Join lebellium [0] (
19:23:17 Quit pamaury (Remote host closed the connection)
19:33:51 Quit wodz (Quit: Leaving)
19:55:42 Join rela [0] (~x@pdpc/supporter/active/rela)
19:57:23 Quit n1s (Quit: Ex-Chat)
20:07:09 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
20:11:06 Join xorly [0] (
20:11:19 Quit chrisb (Ping timeout: 244 seconds)
20:14:23 Join chrisb [0] (
20:25:09 Quit rela (Read error: Connection reset by peer)
20:29:16 Join wodz [0] (
20:29:52wodzpamaury: ping
20:33:05wodzpamaury: I have a problem with qeditor/hwstub. When I try to write something to register whole thing goes mad. qeditor stops to display reg values.
20:33:43pamaurylooks like hwstub crashes, is that with trunk or gerrit patch ?
20:34:07wodzits atj hwstub from HEAD
20:34:18pamauryok, wait a minutes, I'll try it
20:35:31wodzI tried setting bit26 in CMU_DEVCLKEN to enable GPIOs
20:39:07***Saving seen data "./dancer.seen"
20:41:08pamauryhum, adfuload is not happy with my command line
20:42:40wodzwith hwstub_shell I get libusb:warning [libusb_control_transfer] unrecognised status code 1
20:42:41wodzerror: [string "DEV.write32(0xB0010080, devrst)"]:1: fail to write32 @ 0xb0010080
20:42:58wodzadfuload -e -s1 hwstub.bin -s2 /dev/null
20:43:40pamauryadfuload has a bug when specifying usb vid, it lacks a i++ basically
20:44:12pamauryok qeditor running now
20:45:15wodzthat is strange, I swear I tested with other vid:pid :/
20:45:41pamaurywriting doesn't seem to have any effect here
20:45:50pamauryit doesn't have even change the register value :-/
20:48:05pamaurythere is something fishy
20:48:12wodzwriting to iram works
20:49:03pamauryI mean with qeditor
20:50:44wodzwriting to WDCTL also work
20:50:55pamauryI'm trying on CMU_DEVCLKEN
20:53:19pamaurydid you try with hwstub_shell ?
20:53:40wodzI posted error a few lines earlier
20:54:10wodzthe strange thing is that in lcm test binary I write to this register
20:55:16wodzI can't even write the same value I just read from this reg
20:55:56pamaurythe register is 32-bit wide ?
20:57:15pamauryhum, here writing doesn't have any effect
20:57:25pamaurywhat is the adresse of iram ?
20:58:58pamaurywriting to iram works in hwstub_shell, that's weird
21:03:25pamauryI don't understand
21:03:35wodzthe register should be 32bit (and in lcm test binary I am writing (1<<8) so it is not 8bit for sure
21:04:53pamauryI've added logf to hwstub and the write command is well received, address is correct
21:06:11pamauryi'll try to add a special case for 4-bytes write, see if maybe memcpy is broken
21:07:30 Quit krabador (Ping timeout: 272 seconds)
21:07:34pamaurynow writing fails :(
21:09:35pamauryis that fine to write 0x4000010 to 0xB0010080 (CMU.DEVCLKEN) ?
21:09:35 Quit chrisb (Ping timeout: 240 seconds)
21:11:40wodzin my datasheet bit4 is reserved
21:12:35pamauryok, do you have something easy to try ? like disabling backlight
21:14:26wodz <- with this hwstub still works so writing to this reg is possible
21:15:34pamaury> print(string.format("%x",
21:15:34DBUGEnqueued KICK pamaury
21:15:34pamaury> HW.PMU.CHG.write(0x10)
21:15:34pamaury> print(string.format("%x",
21:15:34***Alert Mode level 1
21:15:36wodzPMU_CTL &= ~PMU_CTL_BL_EN should disable backlight
21:15:36pamaury> DEV.print_log()
21:15:38pamauryhwstub 4.0.1
21:15:41pamauryargument: 0xFFFFFFFF
21:15:43pamaurywriting 4 bytes @ B0000008
21:15:45pamaurywrite value 10 to B0000008
21:15:48pamaurybut backlight goes black
21:16:49 Join krabador [0] (~krabador_@unaffiliated/krabador)
21:16:51pamauryand the read value (0x10) looks just wrong
21:17:15pamaurymaybe reading is broken ?!
21:17:32wodzI tried reading from iram and it works
21:17:51wodzboth read and write actually
21:18:34pamauryyeah but something is very wrong
21:18:54pamauryif I do a special case for read of 4 bytes, now I get:
21:18:54pamaury> print(string.format("%x",
21:19:49wodzbtw. I always forget what should I add to have this nice working? I need to provide desc file as an arg to shell but there is something more I can't remember
21:20:48pamaurymissing a file
21:21:00wodzcould it be that target buffer is unaligned and so memcpy() falls back to reading in bytes?
21:21:47pamauryI thought about it but apparently it is not the case
21:23:39wodzI should look for different memcpy() implementation to check
21:23:40pamauryyeah I think something is wrong with memcpy
21:23:55wodzhehe, it is copied from rb outright
21:24:06pamaurymaybe it doesn't do correct 32-bit accesses when it can ?
21:24:27pamauryhwstub really relies on memcpy to do smart stuff for register accesses to work as expected
21:25:13wodzI think the easiest thing is to check other memcpy() implementation for mips
21:25:27pamauryhwstub should probably have a default C implementation for reference
21:25:35***Alert Mode OFF
21:25:38pamauryI can do that in the mean time if you want
21:26:01wodzpamaury: there is reference implementation in rockbox as a failsafe
21:26:48pamauryindeed, I'll copy it
21:27:37pamauryhum, it won't do it, it does byte copy for small chunks
21:28:04 Join chrisb [0] (
21:28:15pamauryI'm wondering if hwstub shouldn't do a special case for 1/2/4 bytes access to make sure things work as expected
21:28:31pamaurydon't you think ?
21:28:45wodzIts hard to believe that memcpy() is broken. Thats very basic stuff.
21:29:01wodzMakes sense I'd say
21:29:36pamauryor maybe create new READ_ATOMIC and WRITE_ATOMIC requests...
21:31:07wodzpamaury: but this will improve things only when you restrict reads to aligned addresses so it will translate to native single instruction (ldr, ldrh, ldrb for arm; lw, lh, lb for mips)
21:31:35pamauryyeah but these are the only cases which make sense for register accesses
21:33:25wodzBut this should be separate special case, maybe separate request as you propose.
21:33:49wodzreading 4bytes from unaligned ram address is pretty valid usecase
21:35:48pamauryyes of course but then it's not atomic, so yeah separate request make sense
21:37:13 Quit krabador (Quit: Take the time.)
21:38:42pamaurydo you think it's safe to assume the compiler will not do nasty stuff or should this be target implemented in assembly ?
21:41:04 Join krabador [0] (~krabador@unaffiliated/krabador)
21:41:24wodzI wouldn't count on compiler
21:48:07wodzpamaury: btw. generic, slightly optimized memcpy in pure C
21:59:00 Nick TD-Linux is now known as TD-Middlebox (~Thomas@about/essy/indecisive/TD-Linux)
22:00:45wodzok, there is completely different mips memcpy() in bionic. It aggressively uses cache prefetch hovewer which is not so nice
22:04:04wodzha and it assumes 32bytes cacheline
22:15:21 Quit xorly (Ping timeout: 240 seconds)
22:32:30pamaurywodz: I'm implementing atomic read/writes, should be ready soon
22:39:08***Saving seen data "./dancer.seen"
22:46:31 Join franklin_ [0] (
22:46:33 Nick franklin_ is now known as [Franklin] (
22:47:09[Franklin]foolsh: G#1012 will probably go into the 100s of patch sets at this rate :)
22:47:19[Franklin]I mean, 77 patch sets now... lol
22:58:58 Quit krabador (Quit: Sto andando via)
23:00:07pamaurywodz: should I make hwstub's DEV.read8/16/32 atomic by default ?
23:00:16 Join krabador [0] (
23:00:26 Quit krabador (Changing host)
23:00:26 Join krabador [0] (~krabador@unaffiliated/krabador)
23:07:52wodzimo yes
23:09:21 Quit chrisb (Ping timeout: 240 seconds)
23:13:38 Quit yuriks (Remote host closed the connection)
23:15:55 Join yuriks [0] (~quassel@opentyrian/developer/yuriks)
23:16:53wodzpamaury: current mips memcpy() does byte access if size is less then 8 which explains the problem
23:18:02wodzsuch pattern looks common to all mips memcpy() implementations
23:18:06pamauryyeah it works now :) I'll upload it to gerrit in a minute
23:18:57 Join chrisb [0] (
23:20:56wodzthe funny thing is that all implementations I found seems to have common origin from some MIPS Technologies code
23:24:54pamaury g#1041
23:25:19[Franklin]fs-bluebot is missing :(
23:25:31pamaurywodz: I uploaded the thing to gerrit, it includes the adufload parsing bug and also adds atj to hwstub_shell
23:25:59wodzok, I'll check this but not during weekend most probably
23:26:07pamauryI didn't bother implementing target_{read,write}{8,16,32} in assembly yet, the C code seems to work fine
23:26:20pamauryand now it works in qeditor and hwstub_shell properly
23:29:42 Join kugel [0] (~kugel@rockbox/developer/kugel)
23:31:59 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
23:33:52wodzpamaury: this is the only implementation I found so far which does not fall back to byte copy for small sizes
23:36:42[Franklin]foolsh: could you try the latest patch set with rotation enabled and report the fps please?
23:40:39 Quit y4n (Quit: 6,000,000 ways to die — choose one.)
23:43:56 Quit amayer (Quit: Leaving)
23:52:00 Join Water255 [0] (~chatzilla@
23:52:33Water255Anyone else think the text view plugin should have an orientation option?
23:52:49[Franklin]Water255: why?
23:53:22Water255easier to read longer lines (don't have to track-back across the page as much).
23:53:45[Franklin]Water255: well, it's certainly possible
23:53:52[Franklin]in fact, I
23:54:03[Franklin]in fact, I'm working on framebuffer rotation code right now :)
23:55:12gevaertsThe easy solution is of course to get a player with a square screen :)
23:55:14[Franklin]Water255: keep in mind that this also means rotating the keymap
23:55:37[Franklin]which isn't hard, but it's not trivial either
23:56:18Water255i hear ya. FWIW i'd be fine with keymap not changing (though not prefer it).
23:57:21[Franklin]Water255: well, it's certainly possible
23:58:05[Franklin]but just because it's possible won't make it happen :)

Previous day | Next day