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 2020-11-18

00:04:22braewoodsspeachy: i found something stupid while looking through MTP sources. apparently MTP responders are expected to allocate 3 usb endpoints but the interrupt one isn't actually used that i can tell. but you still need it to be detected as an MTP endpoint by clients like libmtp.
00:05:48braewoodsat least that tells us that MTP will only work on the same targets that also support HID
00:06:26braewoodsi'm probably going to develop MTP using PP devices
00:06:42braewoodsseems unlikely any coldfire will support it due to missing hardware or so?
00:09:26***Saving seen data "./dancer.seen"
01:15:52 Join karinka [0] (
01:18:09 Quit pixelma (Quit: .)
01:18:09 Quit amiconn (Quit: - Chat comfortably. Anywhere.)
01:19:19 Join amiconn [0] (jens@rockbox/developer/amiconn)
01:19:19 Join pixelma [0] (marianne@rockbox/staff/pixelma)
01:28:46 Quit karinka (Quit: karinka)
01:28:54 Join karinka [0] (
01:30:24 Quit karinka (Client Quit)
01:31:44 Join karinka [0] (
01:33:20 Quit karinka (Client Quit)
01:33:49 Join karinka [0] (
02:09:27***Saving seen data "./dancer.seen"
02:24:57 Join kugel [0] (
02:24:58 Quit kugel (Changing host)
02:24:58 Join kugel [0] (~kugel@rockbox/developer/kugel)
02:28:14 Quit St3ak (Ping timeout: 264 seconds)
02:34:24 Join St3ak [0] (
02:53:11 Join petur [0] (~petur@rockbox/developer/petur)
02:56:06 Quit karinka (Quit: karinka)
02:56:19 Join karinka [0] (
03:20:16 Quit karinka (Quit: karinka)
03:20:17 Join karinka [0] (
03:22:24 Join ender| [0] (~ender1@2a01:260:4094:1:6045:6fff:fedd:cbf3)
03:30:07 Quit karinka (Quit: karinka)
03:30:15 Join karinka [0] (
03:31:03 Quit ufdm (Read error: Connection reset by peer)
03:31:16 Join ufdm [0] (
03:34:27 Join prof_wolfff [0] (
03:38:36 Quit prof_wolfff (Ping timeout: 240 seconds)
03:39:01 Quit ufdm (Quit: Leaving)
04:09:29***Saving seen data "./dancer.seen"
04:33:57braewoodsspeachy: any idea what the wikipedia page on the H300 is talking about?
04:34:31braewoodsNorth American models can be modified to support USB OTG by means of a small internal soldering job, an external modified cable, or a USB transfer box. International firmware is also required.
04:34:45braewoodsNo specifics about what it means.
04:50:19 Join ubervison [0] (~ubervison@2a02:aa12:b106:1b80:4978:337a:24bd:4bbc)
06:09:31***No seen item changed, no save performed.
06:12:57 Quit CommunistWitchDr (Ping timeout: 272 seconds)
06:17:43 Quit Stanley00 (Remote host closed the connection)
06:19:25 Quit Moarc (Ping timeout: 264 seconds)
06:27:36braewoodsoh. so that's what's going on.
06:27:47braewoodsno-one implemented the usb otg driver for the H300.
06:28:16braewoodsi wonder why.
06:28:26braewoodswell that's one feature missing.
06:28:38braewoodsso it is incomplete in this sense
06:28:48braewoodsas it stands the media/host port is useless
06:29:58gevaertsA lot of work for not *that* much useful results
06:30:45 Join SovietShaman [0] (
06:31:28braewoodsgevaerts: i guess so.
06:31:38braewoodsi'll add it to a list of stuff i might try though.
06:31:48braewoodsbut it just means i can't use my h300s for testing usb stuff.
06:32:11braewoodsmain point of it would be MTP support
06:32:20braewoodsbut, chicken or the egg?
06:32:29braewoodsi'll get MTP support working on other OTG targets first
06:33:51gevaertsIf it *actually* does proper OTG. My impression has always been that in marketing "OTG" means "Host on a thing you'd normally think of as device". Not sure though
06:34:14gevaertsAlso, I *think* that one is actually restricted to full speed?
06:35:26braewoodsgevaerts: no idea. i'll test MTP on my unmodded H320.
06:35:34braewoodsif it is then it'll be super slow
06:35:48braewoodsthe chip it uses claims to support 2.0
06:36:09braewoodsjust i'd like to make it work if possible since the OF had it
06:36:20braewoodsbut i can see why it is absent
06:36:38braewoodswhen the first driver was implemented there was nothing to really use it yet in RB usb stack
06:36:57braewoodsbut now there is, the HID driver
06:37:09braewoodswhich was implemented in 2009 it appears
06:37:34gevaerts has some details
06:37:44gevaerts2.0 != high speed
06:37:54braewoodsgevaerts: i saw it earlier. i'll look into it later.
06:38:03gevaertsThe host and device
06:38:03gevaertscontrollers are compliant with Universal Serial Bus Specification Rev. 2.0, supporting
06:38:06gevaertsdata transfer at full-speed (12 Mbit/s) and low-speed (1.5 Mbit/s).
06:38:10braewoodsbut you're right. if that's all it can do
06:38:11gevaerts(from the chip data sheet)
06:38:17braewoodsnot much point
06:38:22braewoodsit's slower than the data port
06:39:11gevaertsI think OTG is one of those areas where speed and spec versions aren't as correlated as in the rest of USB, as in OTG was defined in the 2.0 era, not in the 1.1 era
06:39:42gevaertsIt's been a *long* time since I did USB...
06:39:54braewoodswow. it won't even do 2.0 speeds.
06:40:00braewoodsso why did iriver bother with it?
06:40:10braewoodsit's pretty useless if it can't do 2.0 speeds
06:40:28braewoodsi can see now why no one bothered
06:40:29gevaertsDidn't you need MTP for some DRM stuff?
06:40:36braewoodsyea, windows DRM.
06:40:46braewoodsall pointless now though.
06:40:53gevaertsMaybe they had plans involving that?
06:41:04braewoodsthey implemented it in the US firmware
06:41:07braewoodsbut nowhere else
06:41:34braewoodsif i get bored i might work on it but i can see now why it was low priority.
06:41:43braewoodsit's only useful for low bandwidth use cases.
06:42:26gevaertsOf course if you get the host side working you can attach HID devices to it, which would be useful. I mean, surely doom plays better with a proper keyboard :)
06:42:52braewoodsif you're going to do that you may as well just use a real PC
06:42:57braewoodsno seriously... lol
06:44:36braewoodsso the OTG on this is of limited interest
06:44:42braewoodsi can see now why it was never finished
06:47:22gevaertsI think petur was involved in that? Maybe he still remembers some of the context
06:47:57gevaertsIt was before my time though. I came in with USB on the PP502x targets
06:48:13braewoodsno idea
06:48:18braewoodsdoesn't matter for now
06:48:42braewoodsi just laugh at retaining DRM support on these anyway
06:48:52braewoodsthe service was shut down ages ago so it doesn't matter anymore
06:49:40 Join Moarc [0] (
06:50:21peturyeah, I looked at it, you can find pictures of one of the first devcons where I´m figuring out the connections ;)
06:51:00braewoodspetur: ah. i was just coming along to pick up the pieces to make the h300 a better port.
06:51:03braewoodsso to speak
06:51:34braewoodsit appears the h300 has 2 hardware revisions that make OTG work differently
06:51:36peturI have some sample driver code I got from the manufacturer even (mailed them at the time)
06:51:45braewoodsi see.
06:51:50braewoodsis that on the wiki?
06:52:07peturthe problem was 1) I was just getting back into embedded at that time and 2) I knew nothing of USB
06:52:21speachykarinka, thanks for the offer; the h2 is in reasonably decent shape, but auditing/fixing keymaps and writing the device-specific stuff for the manual needs to be done. There ar e a few platform-level quirks too.
06:53:08peturI have most sitting on my NAS drive, not sure what made it to the rockbox site...
06:53:14braewoodsanyway i was planning to look into adding OTG support for the H300 later on
06:53:22braewoodspetur: can you look for it sometime?
06:53:34braewoodsright now i'm focused on trying to implement MTP
06:53:42braewoodswhich i already have stuff i can test it with
06:53:54braewoodsand honestly OTG is of limited use without more stuff to use it really
06:56:42peturbraewoods: at the bottom has the sample code
06:57:32peturI´ll try to look if I have any other stuff tonight
06:58:06braewoodspetur: ok. good. thanks.
07:04:23braewoodsthough seems even with OTG rockbox is limited to slave role
07:04:49braewoodsthe stack isn't setup for host
07:05:04braewoodsjust OTG enables more slave roles?
07:05:12braewoodsseems to be the case
07:05:48speachyno reason why it can't be done; it's just nobody saw the point in implemting the master/host side of the stack (and the various protocols on top)
07:06:05braewoodsand what would be the use case?
07:06:16braewoodsrockbox is primarily targeted to a slave role
07:06:18braewoodsas it stands
07:06:18speachyplugging in external storage is the main one
07:06:34speachyusing an external DAC (all the vogue these days)
07:07:18speachythose features are already there "For free" for the hosted platforms.
07:08:08braewoodswell i'm honestly not interested in the complexity of developing a host stack...
07:08:17braewoodsfor native targets
07:08:56braewoodsi have to ask though. did any of the OFs for these devices with OTG ever support being connected as a host?
07:09:06braewoodsnot counting hosted
07:09:09braewoodsor Linux based ones
07:09:13speachynot that I recall.
07:09:33braewoodsas i thought. most of them seem to only take the slave role.
07:10:09braewoodsnot surprised honestly
07:10:22braewoodsthey were mainly intended as an extension of your PC, MAC, or other host
07:11:11braewoodsi have an idea for other slave roles
07:11:23braewoodsbut we may need a way to regulate how much we expose at once
07:11:50braewoodssome of these chips have a limited # of endpoints and i could see us exhausting that if we try to do too much at once
07:12:36braewoodsspeachy: i could see adding usb audio which would let the PC output to the audio ports that the RB device has
07:12:47braewoodspassthrough or so
07:13:05speachybraewoods: g#1009
07:13:05braewoodsbut not sure what else would be practical
07:13:07fs-bluebot_Gerrit review #1009 at : Add USB Audio 1.0 support (EXPERIMENTAL) by Amaury Pouly
07:15:10braewoodsi guess since many of them used OTG style usb connectors
07:15:22braewoodsdidn't use
07:16:12speachyhmm. I need to see what it will take to enable C++ in our toolchains.
07:16:37braewoodswhy would we need C++?
07:16:44braewoodsi thought rbutil was all that used it
07:17:59speachyprimarily for the hosted targets where c++ libraries are more common
07:18:32braewoodsi just know from using C++ that it has problems being used in embedded
07:18:52braewoodsthe automatic behaviors it has can cause a lot of headaches.
07:19:03speachyyeah, it's not ideal. but if it lets us reuse stable code, that's a net win.
07:19:28braewoodsi think arduino gets away with it because they throw away the entire standard library to use their own
07:19:42braewoodsso they can tweak what automatic behaviors get used in classes
07:19:49 Join Stanley00 [0] (~stanley00@unaffiliated/stanley00)
07:20:12braewoodsi think you'd have to do that in any case to develop with it in embedded
07:20:28braewoodssimply because you don't want it trying to be clever and screwing you over
07:22:44gevaertsHmmm, I'm vaguely remembering something about the Gigabeat F having host if you have a dock... That might be high speed.
07:23:26speachythe USB-IF really, really jumped the shark when it came to naming and figuring out what is supported.
07:23:43speachyUSB 2.0 was quasi-sane, but everything since then has been a dumpster fire
07:25:11speachyUSB "2.0" features are indepdenent of high speed; but every 2.0-capable USB interface I've seen actually supports high speed −− but typically requiring an external PHY to do it. Which often/usually isn't necessary.
07:28:29braewoodsmtp sure supports a lot of stuff we don't need
07:28:55braewoodslike misc. metadata options
07:29:04braewoodsmainly want MTP for file control and that's it
07:36:34peturwhich reminds me... I should be distributing some hardware to maintainers of iRiver: remotes, cases and an H10. Maybe also my h380
07:41:29braewoodspetur: there was an iriver h380?
07:42:25braewoodsthe units i got i had to buy on ebay. i used them to help produce a new set of bootloaders because they really needed them.
07:42:27peturyeah, a 340 with a bigger disk (installed myself)
07:42:33braewoodsI see.
07:42:55braewoodsI couldn't test on the H100 though since it's super rare
07:43:01braewoodsis that one of your units?
07:43:08braewoodsone of the H110 or so
07:43:21braewoodsi only have H120s and H320s
07:43:43peturyeah but I´m hanging on to it because it is my only device that can record off optical
07:44:11braewoodsif you have a H110 still i'd ask you to test the new bootloader on it
07:44:28braewoodsI tested the related bootloader, H120.
07:44:45braewoodsit works fine but i can't test the H100 one because the hardware is so hard to find
07:45:07braewoodssame with H320
07:45:42peturchecked, I have an h120
07:45:50braewoodsOk then nevermind.
07:45:57braewoodsI figure the H100 is just unpopular.
07:46:08braewoodscan't say i'm surprised. it was a smaller unit.
07:46:15braewoodsthe people who bought these wanted more space.
07:46:47braewoodsthe fact the H300 only came in 20+ GB units makes me think the H100's 10/15 GB didn't sell well.
07:46:59 Quit Stanley00 ()
07:47:49braewoodspetur: do you have any of the rare H300 remotes in your collection?
07:48:07braewoodsi read reports that they work but impossible to test them
07:48:07peturI have two: with and without display
07:48:23peturthey can go too
07:48:25braewoodsso you have both of them?
07:48:33braewoodshow rare.
07:48:45braewoodsif you feel up to it, i'd like to confirm that they work properly with modern RB builds
07:48:51braewoodsi don't know if i even live near you lol
07:49:35peturwell... I was part of rockbox dev team and focussed on iRiver, so... ;)
07:49:51braewoodsi'm mostly coming along later to pick up the pieces
07:49:56peturwhere are you located?
07:50:03braewoodsnorth america
07:50:11peturBelgium here
07:50:15braewoodsah, i figured.
07:50:40braewoodsyou can just test it. i don't think i'll need it. i only could find the more common H100 remotes though.
07:51:03gevaertsI'm pretty sure there's a wiki page that lists devices with who has one. Hopelessly outdated of course, but if I could remember what that page was, it might still help with tracking down a H100
07:51:38braewoodsit would help, gevaerts, but if it's this hard to find it may not be practical enough to matter
07:52:02braewoodsthe most common unit in the H100 series i see show up for sale is the H120
07:52:12braewoodsH140 is rarer but i still see it sometimes
07:52:15braewoodsH100 is the rarest
07:52:21braewoodsi've never seen one to date
07:53:57gevaertsNo H110 though
07:55:21braewoodsi did finally find an hdd6330 unit after a few months of searching
07:55:32braewoodsit was mislabeled as an hdd1630
07:55:32peturI see I forgot to add my H120
07:56:06braewoodsi found the hdd6330 interesting due to how it has so much RAM
07:57:21braewoodsanyway my bootloader project is complete it appears
07:57:42braewoodsat least until the next step can be realized
07:58:20braewoodspetur: i gotta thank whoever added the reset cookie to the crt0.S code
07:58:30braewoodsit allowed me to fix the H300 bootloader
07:58:38braewoodsit saved my unit from being bricked
07:59:26peturheh, I once bricked mine, gave it to Linus for reflashing (soldering) and got it back a few years later :)
07:59:57peturbut that was me in a hurry flashing the wrong file
08:07:38braewoodspetur: yea. i don't think anyone could easily undo it today.
08:07:43braewoodsthe tools are much rarer.
08:08:01braewoodsso it's good the bootloader has rescue code
08:08:20braewoodsit can boot the OF or the rockbox rom image
08:08:44braewoodsthough now that i think about it
08:08:58braewoodsthe rockbox rom image is useless for recovering from a bad flash
08:09:07braewoodssince you can't flash while booted from ROM
08:09:26braewoodsso still should use the OF
08:09:34***Saving seen data "./dancer.seen"
08:09:47speachyboot from rom, ROLO the ram binary, and then flash?
08:09:47braewoodsi guess you can use it to do rolo
08:10:31braewoodsthat's why i saw no point in adding support for the RAM image there
08:10:37gevaertsWhen flashing, always make sure to check both the file and the device
08:11:00braewoodsgevaerts: iriver_flash does most of that. the main thing it can't do is test if it still can boot.
08:11:13gevaertsOtherwise you end up like me, with a bricked bluetooth controller, because somehow that wasn't compatible with the experimental firmware I tried to send to this meizu device :)
08:11:43braewoodsi refactored iriver_flash and added some more sanity checks the original author never had
08:12:01braewoodslike checking the rockbox image to make sure it matches the expected model string
08:12:13braewoodsuseful for making sure they aren't mixing h100, h120, h300 builds up
08:12:23braewoodsthough unlikely it could save someone's bacon
08:13:19braewoodsthe bootloaders are already protected by the checksum table so it won't do much for that but it does validate the RAM/ROM images
08:15:16braewoodsspeachy: something else that occurred to me. MTP could be useful to exfilitrate debugging information over USB.
08:15:38braewoodsit is rather vague about resources so we could define some for exporting other info.
08:18:11braewoodsi'll see if i can find a mtp library that handles most of the hard work
08:18:26braewoodswhy reinvent the wheel if there's one that can be reworked a bit?
08:19:16braewoodstrouble is most are heavily tied to their intended environment
08:20:06gevaertsWell, that could work. We do have USB serial debugging though if you want that sort of thing today :)
08:20:17braewoodsjust an idea i had for now.
08:20:28braewoodsright now i just care about uh
08:20:36braewoodsgetting this to work
08:20:39braewoodsfor its main purpose
08:21:35gevaertsMain purposes are boring :)
08:22:03braewoodsstill it would be interesting to expose other things as MTP resources
08:22:15braewoodslike the current ROM image
08:22:19braewoodson some targets
08:23:02speachyimplement a CDC interface and stream debug data out that
08:23:28*gevaerts never installed debian on a laptop from a rockbox device with experimental USB code that exported a disk image!
08:24:21gevaertsspeachy: that's there though
08:24:42speachyoh? we have a CDC/serial-over-usb interface?
08:24:48braewoodsspeachy: usb_serial
08:24:53speachy(how did I miss that?)
08:24:56braewoodsi saw it when i was poking around
08:25:02 Quit mutax (Read error: Connection reset by peer)
08:25:45gevaertsAnd of course calls to that in firmware/logf.c
08:26:02gevaertsYou need USB_ENABLE_SERIAL
08:26:23speachyah, I see that now
08:28:53speachyI'm going to add that to the configure advanced options.
08:30:07gevaertsMaybe test if it still works first :)
08:31:02speachypfft, if it compiles that means it works, right?
08:31:31speachythough it obviously won't work on any of these newfangled hosted targets
08:31:36gevaertsYou make some excellent points there!
08:32:05 Join prof_wolfff [0] (
08:34:05braewoodsfrom reading the MTP spec it's probably possible to achieve async operations... but only if you support multiple sessions or so
08:34:41braewoodswe're probably only going to support single session due to limited resources
08:44:14 Join mutax [0] (
08:44:45speachywell, it builds on everything I have here
08:49:56speachywith one fix for hosted targets
08:53:05fs-bluebot_Build Server message: New build round started. Revision 7c87467, 293 builds, 9 clients.
09:04:40braewoodswe could start incorporating C11 stuff into RB if useful
09:04:46braewoodsnamely _Generic or _Static_assert
09:16:34fs-bluebot_Build Server message: Build round completed after 1410 seconds.
09:16:42fs-bluebot_Build Server message: Revision 7c87467 result: All green
09:24:30 Quit speachy (Quit: WeeChat 2.9)
09:25:41 Join massiveH [0] (
09:26:32 Join speachy [0] (~speachy@
09:32:15mendel_munkishow big of an issue are the changes in 3035?
09:33:36speachypretty minor I'd think.
09:33:57mendel_munkiscan I commit it as is?
09:33:58speachywell, the manual is now the last holdout for swcodec/lcd_bitmap stuff
09:34:27speachydon't see why not
09:35:19fs-bluebot_Build Server message: New build round started. Revision 0aa7028, 293 builds, 9 clients.
09:35:57mendel_munkisbtw the way gerrit handles timeing out is irritating.
09:38:23speachyI think the underlying issue is some sort of www server or tcp stack configuration I haven't quite figured out
09:51:46fs-bluebot_Build Server message: Build round completed after 987 seconds.
09:51:49fs-bluebot_Build Server message: Revision 0aa7028 result: All green
09:51:53fs-bluebot_Build Server message: New build round started. Revision 473aa61, 293 builds, 9 clients.
09:59:07mendel_munkishow do I make a new wiki page?
10:01:59speachytype in an unknown page (/wiki/SomeNewPage) or on an existing page, create a reference to the new page (basically any CamelCase word is treated as such) and then click on the link that gets created.
10:05:18braewoodsspeachy: seems rb has partial support for utf16/ucs2... but decode only
10:05:44braewoodsbut realistically we need it to support mtp fully
10:06:02braewoodsfat32 has no unicode?
10:06:14speachyfat32 doesn't care one way or another
10:06:38braewoodsit's easy to encode ASCII as utf16
10:06:44braewoodsjust not sure how we'd handle the rest
10:07:10fs-bluebot_Build Server message: Build round completed after 918 seconds.
10:07:15fs-bluebot_Build Server message: Revision 473aa61 result: All green
10:08:16speachywait, does mtf say "utf-16" or "ucs-2"?
10:08:27braewoodslet me look
10:08:42speachyutf16 <->utf8 is straightforward.
10:08:58braewoodsStrings in PTP (and thus MTP) consist of standard 2-byte Unicode characters as defined
10:08:59speachyucs-2 is 16-bit only, and thus is at best a subset of utf-16/utf-8
10:09:00braewoodsby ISO 10646.
10:09:15braewoodsthat's all it says
10:09:36***Saving seen data "./dancer.seen"
10:10:07braewoodsit's UCS
10:10:08speachysince it says "2-byte" that's nearly guaranteed to be UCS-2
10:10:43 Quit prof_wolfff (Ping timeout: 260 seconds)
10:10:59speachyutf16 was added later in an amendment.
10:11:41braewoodswe could probably get by with just BMP support
10:11:42speachyfwiw, it looks like at least the andtoid mtp code actually implements utf16.
10:16:37_bilgus__hmm that could be handy in lua
10:16:51braewoodsit appears rbunicode already does what i need
10:17:12braewoodsunless that's somehow wrong?
10:19:23speachyyeah, handling the BMP must suffice, because that's all that can be done with MTP.
10:19:41braewoodscould we reuse utf16 for ucs2
10:19:49braewoodssupport for that is in RB unicode already
10:20:12speachywe could; the question is what the MTP initiator supports.
10:20:29braewoodsi'll see what libmtp does
10:20:49speachy(I'd be surprised if anything widely deployed doesn't support UTF-16)
10:21:27braewoodslibmtp uses ucs2
10:21:33braewoodsvia iconv
10:21:51braewoodsso we'd need our own routine
10:22:01speachyI wound't personally bother at this point.
10:22:20braewoodsi'll look into it later but i'd copy it from somewhere else honestly
10:22:28braewoodsas long as the license is GPL compatible and documented
10:27:52braewoodsutf8encode / utf8decode works in terms of ucs2
10:27:56braewoodswe already got it
10:37:17mendel_munkisdoes anyone know if the IAUDIO X5 joystick does diagonals?
10:46:27braewoodsspeachy: i think i'm going to only go for MTP storage stuff but i have some ideas for how i will abstract it at all
10:46:58braewoodsi'm thinking of making MTP api internally to allow for different implementation of storage options
10:47:05braewoodsnot just the obvious one
10:47:18speachyokay, I managed to get approval to come into the $dayjob lab to perform "maintainence"
10:48:11speachywould be a good time to do upgrades on my server too, which means.. downtime.
10:58:35 Quit _bilgus__ (Remote host closed the connection)
10:59:41 Join _bilgus__ [0] (~bilgus@2605:a000:1301:89f6:71ab:2e9f:7f39:8176)
10:59:44_bilgus__mendel_munkis, no clue but you can add the button code to debug menu (if not already there) and see if it does multi press
11:02:49 Quit massiveH (Quit: Leaving)
11:10:46 Join Rower [0] (
11:19:19braewoodsthe H300's OTG chip has a driver for the linux kernel
11:19:22braewoodsthat may be useful
11:20:46mendel_munkis_bilgus__: I had to ask because I don't have one. I am trying to make a set of kemaps and knowing what key combos are available across targets is the first set. (but thanks for the advoce regardless)
11:22:03 Quit _bilgus__ (Remote host closed the connection)
11:22:35speachykey combos in general are... not something that can be assumed.
11:25:51 Join _bilgus [0] (~bilgus@2605:a000:1301:89f6:71ab:2e9f:7f39:8176)
11:30:55mendel_munkison targets with enough keys that can be worked around
12:09:37***Saving seen data "./dancer.seen"
12:34:03 Quit petur (Quit: Connection reset by beer)
12:39:29 Join lebellium [0] (
12:43:44karinkaI also have a iriver H20, it's kinda dead
12:43:58karinkaThe HDD works, just needs a new battery
12:44:32karinkaWhich should be coming soon
12:46:34braewoodskarinka: well i released an updated bootloader that enables some new stuff that used to be H100 exclusive.
12:47:11braewoodsit seems the h300 is only missing one thing to be complete at this point
12:47:14braewoodsOTG support
12:49:10braewoodskarinka: which means you can probably CF mod your h320 now if you wanted
12:52:51speachykarinka: so what issues with the H2 did you run into?
12:54:30 Quit lebellium (Quit: Leaving)
12:58:33karinkaStill new to IRC so bear with me
12:59:16karinkaIf you go from playing a song via a file, then switch to a song via the database, the song stops and I have to restart the player
12:59:24karinkaIt crashes
12:59:32braewoodswhat build?
12:59:54karinkaThe latest dev one i think
13:00:13 Join lebellium [0] (
13:00:14braewoodsi'll try it with mine
13:03:45speachyprobably the audio buffer underrunning and not recovering
13:04:13braewoodsnot happening on my h320
13:04:20braewoodskarinka: can you describe the exact process?
13:04:33speachyno, it's going to be a problem with the hiby-based hosted targets
13:04:53braewoodsit's not about the H320
13:04:55braewoodsi see
13:05:32karinkahang on ill put a fresh copy on it
13:06:25karinkagotta find my m3k first
13:18:58karinkaOkay I can't reproduce it with a fresh install, it must have something to do with my settings
13:19:23speachydatabase reinit?
13:20:00speachystill, this does seem to indicate that there's still a playback underrun scenario that's not recoverable
13:20:58speachy(and ratehr than try to find and hack in a workaround, I'd rather just push playback buffer refilling into a separate thread)
13:21:32braewoodskinda interesting. the OTG on the H300 sucks but... it at least has a chance of being supported
13:21:49braewoodsthe OTG for the other coldfire targets either don't exist or have a totally undocumented chip
13:21:56braewoodskarinka: it's limited to usb 1.1
13:22:03braewoodsso it's going to be slow
13:22:24karinkaah, well yeah its old
13:22:34braewoodsoddly enough some stuff around same era support 2.0
13:22:48braewoodsso... yea
13:23:07braewoodsthe h300 is a weird one. it supports OTG on a separate USB port.
13:23:14karinkado you have a good resource on how to do the CF mod for the H320?
13:23:40braewoodsit's not specific but it tells you what you need.
13:24:00braewoodsthe h320 like the h120 uses a 50 pin adapter
13:24:07braewoods50 pin hard drive 1.8"
13:24:07karinkaExcellent, exactly what I need
13:24:18braewoodsso you need an adapter
13:24:31braewoodsjust uh note, all the adapters i could find in recent days have a jumper
13:24:35braewoodsthat sticks up
13:24:43braewoodsthat will cause problems when you go to reassemble
13:24:56braewoodsand you need the jumper set or you have issues
13:24:58braewoodswith rockbox
13:25:05karinkaOkay, I'll have to look into it more
13:25:17braewoodsif the jumper is bridged, it sets master mode
13:25:19braewoodsotherwise slave mode
13:25:33braewoodsbest option is to cut the jumper and solder it so it will remain low profile
13:25:45braewoodsif that's not an option you can bend the jumper over and find some other way to permanently bridge it
13:26:12braewoodsbut it works fine
13:26:20braewoodsi've only CF modded my h120s but it works
13:26:34braewoodsas for batteries? the replacements all have really long cables so
13:26:39braewoodsyou need to get creative with it
13:26:50karinkaYeah, I haven't even opened it yet, but the battery is at like 3.9 volts or something, so yeah
13:26:54braewoodsfor my H320
13:27:02karinkaI have a battery on the way, should be easy enough
13:27:05braewoodsi flipped the battery over on its side
13:27:12braewoodsreaching the terminal is the hardest part
13:27:13karinkaI'm familiar with opening stuff up
13:27:27braewoodsi mean flipped battery over so the cable was underneath
13:27:34braewoodsprobably won't work too well if you do a CF mod though
13:27:40braewoodsit doesn't use as much space
13:27:51braewoodsso you'll need to think of some other way to keep it from moving around
13:27:54braewoodstape perhaps
13:28:24braewoodsbut that's just what i've observed from opening it
13:28:33braewoodsi found the H320 more difficult than the H12
13:28:39braewoodsso be careful with it
13:29:10braewoodsh320 is really in its case... you have to pry it out somehow
13:29:14braewoodsthe board
13:29:16braewoodsor so
13:29:23braewoodsmakes me a bit leery
13:29:32braewoodsh120 is kinda flat so
13:29:38braewoodseasy to disassemble and lift out
13:29:58braewoodsi like that i could pry it apart without having to stress the plastic
13:30:03braewoodsbut h320... ugh.
13:30:14braewoodsit's really in there
13:30:34karinkayeah sounds kinda like opening a phone to me lol
13:30:39braewoodsbe sure to pry off in the direction of the back
13:30:46braewoodsdon't try to pry the front off
13:30:55braewoodsthough it seems there's a seam there
13:31:04braewoodsyou may damage it if you try to force it open that early
13:31:22karinkaIs there a, like, a video or something showing it being done?
13:31:35karinkaOkay ill have a look
13:32:06braewoodsthis is the h120
13:32:09braewoodsmuch easier =p
13:32:40braewoodsit comes apart very easily
13:33:13karinkaYeah, ive tried to replace the battery in a Nexus 4 before, so I got all to tools
13:33:30karinka*tried i failed lol
13:33:41braewoodsi find the H120 caps out around 13 MB/s in sequential block write mode
13:33:45braewoodswith a CF card
13:33:55braewoodsprobably similar for the H320
13:34:18braewoodsmore realistic to get maybe 10 MB/s when you access the filesystem
13:34:47braewoodsyou can go for either a real CF card or use an adapter for SD card
13:35:18braewoodsin either case LBA48 is supported so you can install much larger drives
13:35:21braewoodsup to 2TB in theory
13:35:24karinkaah looks easy enough, its not like the battery is glued to the damn case like the Nexus
13:35:31karinkajust have to be careful
13:35:35speachyok, I think I have another missing buffer underrun recovery scenario covered.
13:35:36braewoodsit's taped down or held down
13:40:33karinkaOkay, that sounds fine
13:40:58karinkaGotta do some shit, but I'll definetly check out how to reproduce the H2 bug
13:45:37braewoodskarinka: be sure your bootloader is up to date before you do CF mod. the older ones had reports of issues.
13:45:44braewoodsplus the OF hates CF cards.
13:46:44 Quit karinka (Quit: karinka)
14:02:16 Join karinka [0] (
14:09:39***Saving seen data "./dancer.seen"
14:38:21braewoodsspeachy: for bulk transfers over USB, what's the largest payload transfer_complete callback will receive?
14:38:28braewoodsspeachy: the max packet size?
14:38:53speachyan individual USB packet is 512B max, but UMS block sizes can be up to 64K.
14:39:12speachygit diff
14:39:30*speachy smacks laggy sloppy focus
14:40:54fs-bluebot_Build Server message: New build round started. Revision 3027cea, 293 builds, 9 clients.
14:42:49speachykarinka: this build running should make underruns less likely and recover better when it does happen.
14:44:29speachy(and report the number of underruns we'ce seen on debug screen)
14:56:05 Join ufdm [0] (
15:01:47fs-bluebot_Build Server message: Build round completed after 1252 seconds.
15:01:53fs-bluebot_Build Server message: Revision 3027cea result: All green
15:02:43speachy (fresh out of the oven!)
15:08:35karinkaThanks for the link :)
15:09:35karinkado i need to update the bootloader aswell?
15:11:12speachythe bootloader hasn't been updated since October 17th. But if you're not using that version, I highly recommend it.
15:18:04karinkai think im using the latest one, but i'll check
15:18:47speachythe BL's tools menu has the version at the bottom, and it's also in System/Debug/HW Info
15:34:07 Join advcomp2019__ [0] (~advcomp20@unaffiliated/advcomp2019)
15:37:22 Quit advcomp2019_ (Ping timeout: 260 seconds)
15:47:49 Join prof_wolfff [0] (
16:09:40***Saving seen data "./dancer.seen"
16:33:32 Quit Rower ()
16:38:42 Join Rower [0] (
16:40:16 Join ac_laptop [0] (~ac_laptop@
16:46:13 Quit Rower ()
16:48:29 Join Rower [0] (
17:06:37 Join ulmutul [0] (
17:26:53 Nick amiconn is now known as Guest38565 (jens@rockbox/developer/amiconn)
17:26:54 Join amiconn_ [0] (jens@rockbox/developer/amiconn)
17:26:54 Quit Guest38565 (Killed ( (Nickname regained by services)))
17:26:54 Nick amiconn_ is now known as amiconn (jens@rockbox/developer/amiconn)
17:27:02 Quit XDjackieXD (Ping timeout: 260 seconds)
17:27:23 Join XDjackieXD [0] (
17:28:12 Quit braewoods (Ping timeout: 260 seconds)
17:28:45 Join braewoods [0] (~braewoods@learnprogramming/staff/braewoods)
17:30:06 Quit ubervison (Quit: Leaving)
17:35:44ulmutulbraewoods: some of the PP targets support OTG as host (for transfering pictures from a camera to the device). At least the Samsung YH925 and the Packard Bell Vibe500 are capable, and IIRC the HDD6330 can also do it.
17:36:08braewoodsulmutul: i see.
17:36:33braewoodsbut not very relevant today
17:36:59braewoodsi don't know why anyone would want their photos on their media player?
17:42:11ulmutulbank in the days SD cards were much smaller in capacity, and on a long trip you could backup your pictures to the player's HDD.
17:43:04ulmutulOnly the "big" players supported it (with 30 GB disk space), the smaller ones (e.g. Samsung YH820) didn't.
17:56:28 Quit kugel (Ping timeout: 260 seconds)
17:57:20 Join kugel [0] (
17:57:20 Quit kugel (Changing host)
17:57:20 Join kugel [0] (~kugel@rockbox/developer/kugel)
18:02:56 Quit ulmutul (Quit: Leaving)
18:09:42***Saving seen data "./dancer.seen"
18:12:25 Quit __builtin (Quit: No Ping reply in 180 seconds.)
18:13:04 Join __builtin [0] (~quassel@rockbox/developer/builtin)
18:19:16 Quit Rower (Ping timeout: 240 seconds)
18:35:20karinkai just recieved my battery for the h320
18:35:45karinkai cant seem to find the battery replacement page, just want to double check its the right one
18:35:50karinkacan anyone link?
18:38:55braewoodskarinka: there's no page, just a video.
18:39:25braewoodstake your pick
18:39:45braewoodsi recall you need to unscrew some side stuff first
18:39:49braewoodsphilips head
18:40:06braewoodsthen need to pry the back cover off
18:40:10braewoodsdon't mess with the front
18:42:00braewoodsfortunately you just need to pry the board up enough to reach the terminal.
18:42:03braewoodsit's on the side
18:42:05braewoodsof the board
18:44:12 Quit Acou_Bass (Ping timeout: 272 seconds)
18:44:37 Join Acou_Bass [0] (
19:05:01karinkaah okay, thanks
19:06:33 Quit karinka (Quit: karinka)
19:29:07braewoodsinteresting. the hdd6330 has a tv output. o.O
19:29:52braewoodsand the chip has a datasheet
19:30:41braewoodsthat might be interesting to try to support. RB output to a large screen.
19:38:53 Quit paulk-leonov (Ping timeout: 268 seconds)
19:38:56 Quit Acou_Bass (Ping timeout: 265 seconds)
19:41:29 Join paulk-leonov [0] (
19:43:05 Join Acou_Bass [0] (
19:53:22 Quit bluebrother (Disconnected by services)
19:53:26 Join bluebrother^ [0] (~dom@rockbox/developer/bluebrother)
19:54:57 Join fs-bluebot [0] (
19:57:49 Quit fs-bluebot_ (Ping timeout: 264 seconds)
20:00:02 Quit _bilgus (Ping timeout: 264 seconds)
20:02:15 Join _bilgus [0] (~bilgus@
20:09:20 Join karinka [0] (
20:09:46***Saving seen data "./dancer.seen"
20:39:42Strife89michael@kiruria:~/git/rockbox/tools$ make descramble.c \ make: Nothing to be done for 'descramble.c'.
20:40:03Strife89I could swear I've built this before
20:41:14Strife89I have binaries for mkboot and scramble
20:42:05Strife89(Trying to patch an iRiver FW with the new bootloader)
20:45:51Strife89Oh now I see, it just wants "make"
20:56:32karinkabattery replacement is done, just waiting for it to charge to see if its still alive
20:56:50karinkaits been sitting on my shelf for years, hopefully it was just the battery
20:57:11karinkait was reporting like, 3.93 volts or something when i tried it the other day
20:59:03karinkadoes the h320 have any indicators that its charging?
21:05:46Strife89It won't at first, if the battery is extrememly drained
21:06:04Strife89But normal behavior is to eventually turn on the screen and show a charging animation
21:06:52Strife89Just wrote 3027cead01 to the disk and rebooted my H320, and now it hangs at the main menu before drawing the menu items
21:07:48Strife89Naturally, this is not inspiring much confidence since I only upgraded my build in order to flash the new v8 bootloader :)
21:15:04 Join Misanthr- [0] (~Misanthro@
21:16:05 Quit Misanthropos (Ping timeout: 272 seconds)
21:16:05 Nick Misanthr- is now known as Misanthropos (~Misanthro@
21:36:25 Quit ac_laptop (Quit: WeeChat 2.9)
21:36:41 Join ac_laptop [0] (~ac_laptop@
22:09:47***Saving seen data "./dancer.seen"
22:17:37 Quit prof_wolfff (Ping timeout: 264 seconds)
22:26:23karinkajust confirmed its charging
22:30:32karinkaill be sure to update the bootloader once its fully charged
22:36:36 Quit ac_laptop (Ping timeout: 240 seconds)
23:13:16 Join ac_laptop [0] (~ac_laptop@
23:41:41karinkaokay to update the bootloader i need to patch the ofw?
23:48:35karinkafound the info i need i think
23:48:49karinkanot sure what ofw to patch
23:49:42 Quit mendel_munkis (Ping timeout: 256 seconds)
23:53:50 Join mendelmunkis [0] (
23:56:55 Join Stanley00 [0] (~stanley00@unaffiliated/stanley00)

Previous day | Next day