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 2020-11-22

00:09:21fs-bluebot_Build Server message: Build round completed after 728 seconds.
00:09:24fs-bluebot_Build Server message: Revision 2eb191a3f4 result: All green
00:10:41speachyhuh, looks like these improved build times are solely due to getting rid of the excessive logging in th upload cgi script
00:11:27_bilgus_wow quite a difference
00:12:06fs-bluebot_Build Server message: New build round started. Revision 6b3b4df6f6, 293 builds, 9 clients.
00:12:12speachy(resulted in a substantial upload speed difference from my colocated builder)
00:12:52speachythat and ccache resulting in nothing actually changing for most of the builds. :)
00:13:22speachyhopefully tonight's manual and voice builds will be problem-free
00:14:04_bilgus_any way to make it do full builds once a day or every n builds?
00:14:25_bilgus_like remove the cache
00:14:38speachywe could, but I don't see the benefit
00:14:50speachyccache has proven to be quite reliable
00:15:20_bilgus_I figure it might catch weird hard to find compat issues
00:15:42_bilgus_but I'll defer to your trust in it
00:16:58speachyI still want to push manual generation to the builders, to be done on every commit
00:17:41speachyvoices still are best done centrally, though rewriting teh script to use more than one CPU would help.
00:17:54_bilgus_what will that solve just lazy manual building current?
00:18:48_bilgus_I wish we could just supply code to do it at user level
00:18:57speachyReduction of special cases.
00:19:06speachy"to do it" ?
00:20:22_bilgus_like in rbutil on their hardware
00:20:46speachyup until recently, if you were't using english, rbutil was the only practical choice.
00:21:19_bilgus_oh it already allows that?
00:21:22speachybut the central voice thingey uses ./configure to do everything
00:22:22_bilgus_sorry rbutil is not a thing I use includes the languages and settings we use for the various TTS engines.
00:23:33speachymore are certainly possible but I only speak English and Bad English, so I'm not exactly a good judge of what sounds usable.
00:24:07fs-bluebot_Build Server message: Build round completed after 722 seconds.
00:24:10fs-bluebot_Build Server message: Revision 6b3b4df6f6 result: All green
00:24:22speachyif you want voiced filenames/etc too, rbutil is much simpler than the cmdline scripts
00:24:52_bilgus_braewoods, your things in
00:28:42 Join _bilgus [0] (
01:32:02 Quit massiveH (Quit: Leaving)
01:54:44***Saving seen data "./dancer.seen"
03:39:59 Join lebellium [0] (
03:54:45***No seen item changed, no save performed.
04:54:04 Join Oksana [0] (~Wikiwide@Maemo/community/ex-council/Wikiwide)
05:17:03 Quit borkitall (Quit: borkitall)
05:31:42bluebrothercurrent Rockbox Utility main screen:
05:32:33bluebrothervoice language is currently fixed to english, but right now does the same thing as we had before. Language selection support to follow.
05:54:47***Saving seen data "./dancer.seen"
06:07:38 Join johnb7 [0] (
06:09:51johnb7speachy As for the HifiWalker firmware: if youf feel they could have actually changed some hardware components, then I should not try to circumvent the name string ...
06:11:02johnb7Anyway, I posted questions on their homepage. Maybe they are more inclined to answer there.
06:17:16 Quit johnb7 (Ping timeout: 246 seconds)
06:34:18braewoods_bilgus_: thanks
07:06:59 Join johnb7 [0] (
07:32:10 Quit johnb7 (Ping timeout: 246 seconds)
07:54:49***Saving seen data "./dancer.seen"
08:22:54 Join ac_laptop [0] (~ac_laptop@
08:34:44speachyjohnb7: The hardware rev possibility is because the Eros Q II looks identical, except for a slightly thicker case.
08:34:54speachyand has incompatible guts.
08:35:22speachyand there's a possibility that the v1.3 is based on the newer Q II.
08:35:51speachy(the Q II is not based on a HibyOS platform, so the SW load wouldn't work anyway)
08:35:59 Join johnb7 [0] (
08:36:57 Join prof_wolfff [0] (
08:37:04johnb7I see
08:41:59speachy(but I doubt it, given that it tries and fails to do the update, which at least implies hibyos is in use)
08:42:11speachythe point is thought that we don't know what we don't know.
08:43:53 Quit johnb7 (Ping timeout: 260 seconds)
08:44:38fs-bluebot_Build Server message: New build round started. Revision 1b9eebb39d, 293 builds, 9 clients.
08:46:02bluebrotherspeachy: can you put the iriver v8 bootloaders to bootloader/iriver/8.0 ? Right now the bootloader installation doesn't work anymore since the checksums don't match :)
08:46:24speachybraewoods: ^^^
08:46:39speachyI'll put the files up when I have 'em.
08:47:02braewoodsspeachy: like i said before:">
08:47:19bluebrotherah, that's the problem :)
08:47:38braewoodsi put them there until they could be made official
08:48:14bluebrothercrap. Seems I lost my template for build-info. Didn't save the file :/
08:49:11speachythey're up.
08:50:22braewoodswe can drop fwpatcher at some point but for now i guess it needs to stay
08:50:50braewoodsat least until rbutil is more polished
08:50:56braewoodsabout how to handle this situation
08:51:28braewoodsregarding the script it uses for checksums...
08:51:42braewoodsi setup the local files needed to make it work
08:51:46braewoodswhat a pain that was
08:51:56braewoodsi'm tempted to archive that for future updates
08:54:42bluebrotheras in calculating those checksums?
08:55:12braewoodsbluebrother: yea. perl script in the fwpatcher directory is used to generate the checksum table compiled into it or rbutil.
08:55:13bluebrotherand what polishing you're missing apart from reading the before-after checksums from some file?
08:55:40bluebrotherthose aren't simple md5sums?
08:55:40braewoodsthat's it afaik. i was waiting for a solution to the hard-coded checksum issue to be implemented.
08:55:46braewoodsthey are md5sums
08:55:58bluebrotherah. So now the bootloaders will work again with current git.
08:56:01fs-bluebot_Build Server message: Build round completed after 682 seconds.
08:56:03fs-bluebot_Build Server message: Revision 1b9eebb39d result: All green
08:56:04fs-bluebot_Build Server message: New build round started. Revision e8f8df4ee0, 293 builds, 9 clients.
08:56:25braewoodsbut you can check the script to be sure or w/e
08:56:28bluebrotherI was thinking about finishing up some leftovers and then try to get Rockbox Utility 1.5.0 done.
08:56:47bluebrotherand leave that hardcoded checksums as well as the sudo issue for 1.6.0
08:56:49braewoodsbut i still had to setup the firmware archive the script needed
08:57:03braewoodsthat's what was the most tedious. haha
08:57:28braewoodswe can rewrite the script to generate the new checksum files when you get around toi t
08:58:04braewoodsthat's the only part i think it worth scavenging from fwpatcher
09:13:27fs-bluebot_Build Server message: Build round completed after 1043 seconds.
09:13:30fs-bluebot_Build Server message: Revision e8f8df4ee0 result: All green
09:43:21 Join johnb3 [0] (
09:45:19fs-bluebot_Build Server message: New build round started. Revision 25529e4fe0, 293 builds, 9 clients.
09:54:53***Saving seen data "./dancer.seen"
09:57:12fs-bluebot_Build Server message: Build round completed after 713 seconds.
09:57:15fs-bluebot_Build Server message: Revision 25529e4fe0 result: All green
10:06:18 Join fs-bluebot [0] (
10:06:25 Quit bluebrother (Disconnected by services)
10:06:30 Join bluebrother^ [0] (~dom@rockbox/developer/bluebrother)
10:07:48bluebrother^speachy: what / how I'd like to see things in build-info:
10:08:12 Quit fs-bluebot_ (Ping timeout: 272 seconds)
10:09:16speachybluebrother^: so for the [daily] section, daily builds use a datecode in their filename, not a shash.
10:09:24bluebrother^that's basically have daily / dev builds use the same format as for releases. Rockbox Utility will also allow an optional url appended, which, if present, will take precedence.
10:09:37bluebrother^daily could also stick to the date.
10:09:54speachyI think that's preferable, since you can't really tell at a gaance what date a given hash corresponds to
10:10:21speachy(plus that whole alphabetical sorting thing.. :)
10:10:49bluebrother^sorting? Who needs sorting? :P
10:12:07speachyI like the new [development] and [daily] sections. especially the dev stuff; as sometimes a build fails and there's no way to capture that a given target might not be on the same rev as the others.
10:12:24speachy(actually that's not captured on the bleeding edge main www site either)
10:13:13_bilguswoot finally have the toolchains installed again
10:13:15speachyI can get the [daily] knocked out this evening,
10:13:17 Quit vup (Quit: vup)
10:13:58 Join vup [0] (~~~~@
10:13:59speachyfixing [development] to reflect the true state of successful builds is another matter.
10:14:22speachyshould be able to get something usable by tonight.
10:15:16 Quit prof_wolfff (Ping timeout: 240 seconds)
10:15:42bluebrother^great. Right now having [daily] work is my next goal, so I can add support for daily builds.
10:15:57bluebrother^then people will be able to install our voices :)
10:16:58bluebrother^and once the rest is in place adjust things for the development stuff. Works right now (since it's been like that anyway), but I'd like to avoid introducing another special case for daily builds :)
10:17:47speachywith respect to the [bootloader] section, I think we should have explicit player = filename,hash
10:18:10speachyor somethign like that
10:18:55bluebrother^I'm fine with that too. Currently the filenames are hard coded in rbutil.ini, but getting that from the server might be a good idea. Less stuff hard coded in Rockbox Utility = good :)
10:19:04speachythink it's worth putting in a format string for the URLs too?
10:19:24bluebrother^but maybe use player = hash,filename so it uses the same order as the rest?
10:19:50bluebrother^if we have the URLs as part of the config we wouldn't need a format string here, right?
10:19:57speachyeg voices.dailyurl = https://d.r.o/daily/%target%/
10:20:15bluebrother^ah, for that part.
10:20:20bluebrother^forgot about this case :)
10:20:32speachy(I intend to put explicit URLs for everything)
10:21:50bluebrother^hmm. how about using release/voiceurl = ... instead?
10:22:22bluebrother^the thing is I'm trying to look up things by going <buildtype>/<player>, so that would make it more straightforward.
10:22:44speachyyou mean release/voiceurl = url_template ?
10:22:55 Quit johnb3 (Quit: Nettalk6 -
10:23:04speachyand daily/voiceurl = url_template
10:23:07speachyI like that
10:23:07bluebrother^and obviously the same for daily/voiceurl, etc.
10:24:13bluebrother^right now I have to go release/<player> for releases, and manually construct things for dev builds. That's why I'd like to have development/<player>. And in the same way we can add further info, like voiceurl
10:24:37bluebrother^you can check the template format I'm currently using in rbutil/rbutilqt/rbutil.ini
10:24:54bluebrother^just introduced a similar split there.
10:25:55bluebrother^hmm, why did I put a url for dev voices in there? We don't have that.
10:26:01bluebrother^but it's filtered in the UI anyway.
10:26:42bluebrother^though checking if we get a valid url might be better. In case that changes at some point in the future.
10:30:51 Join johnb7 [0] (
10:31:27_bilgusjohnb have you tried mb on the CLip+ was anything relatively recent?
10:31:52johnb7let me check
10:32:13_bilgusI can't reproduce with the old bootloader of the (self compiled new one)
10:32:50_bilgusalso speachy just to be sure we didn't upload a new bootloader to DRO?
10:34:40_bilgusyeah I just verified its the one from 2019 for sure maybe he compiled his own bootloader for some reason
10:35:38_bilgusI guess we could have broken something in the main FW in the past days I'll try a daily instead of whats in my tree
10:36:22johnb7I have it running on 7c498b9043-201030
10:36:37johnb7I will check with the dev build
10:36:43speachy_bilgus: the iriver h1xx/h3xx only
10:40:49johnb7works fine for me
10:41:37_bilgusjohnb thanks, I'll try rebuilding the bootloader from head just to be sure
10:41:56_bilgusI guess we will have to wait to hear from him
10:42:26_bilgusprobably just an error in the redirect file it is rather rigid
10:51:52braewoodsok. think i know what my next project will be.
10:52:03braewoodsputting mtp on hold for now since i can't test it right now
10:52:11braewoodsi'll look into the OTG chip on the H300
10:52:32braewoodsluckily there's linux kernel source for it so i can at least compare notes about it.
10:53:14johnb7_bilgus: I attached a sample redirect file to the post ...
10:59:02_bilgusthanks I'm still rebuilding just to be sure
10:59:05speachybluebrother^: oh, will the next version be https-enabled? (including templates)
11:09:02_bilgusok so something is afoot I can reproduce what he is reporting now just need to figure out why
11:13:47speachyso can anyone with a fuze+ look at/comment on fs#13020 ?
11:13:49fs-bluebot Hissing/Static on Sansa Fuze+ (bugs, unconfirmed)
11:17:58 Quit johnb7 (Ping timeout: 256 seconds)
11:20:38bluebrother^speachy: yes.
11:25:41speachybluebrother^: does rbutil care about source downloads too?
11:27:13braewoodsspeachy: something that occurred to me. most rockbox targets only have one USB hardware subsystem to deal with. but the H300 has 2 technically. the Cypress chip that deals with UMS paired with another chip for USB charging. but the OTG chip is separate from that. is it possible to let cypress to continue to manage UMS for the H300 while offloading all other USB to the OTG?
11:27:14speachybluebrother^: also:
11:28:00speachybraewoods: possible, sure, but it's going to take a decent amount of work to abstract away from the rockbox API.
11:28:14braewoodsyea, that's the other issue.
11:28:20braewoodsit's not enough to just make the chip work.
11:28:29braewoodswe also need a way to make it co-exist with the other one.
11:29:28braewoodsactually... i wonder...
11:29:49braewoodsi'll see what i can do but i think the cypress chip does most of the work for us. i don't recall H300 actually using the UMS code we have.
11:30:10braewoodsit may be possible to just disable our UMS on the H300 if that theory is right.
11:30:54braewoodseither way this won't matter to the new bootloaders
11:31:03braewoodsi wasn't planning to have OTG enabled when it is running
11:31:11braewoodsi plan to keep it disabled like it currently is
11:31:18braewoodsonly useful once RB is booted
11:33:33braewoodsthe linux kernel only implements half of the ISP1362 functionality. only so it can provide USB host functionality.
11:36:05speachybluebrother^: now has voiceurl and sourceurl templates too.
11:43:15speachybluebrother^: let me know if it looks sane (especially the template) and I'll implement [development] for the same way.
11:54:57***Saving seen data "./dancer.seen"
12:07:44 Join johnb7 [0] (
12:08:27 Quit Acou_Bass (Quit: ZNC 1.8.1 -
12:10:45 Join Acou_Bass [0] (
12:11:09 Join prof_wolfff [0] (
12:11:17speachybluebrother^: also, in the [release] section, is it kosher to add a url to the entries? eg archosfmrecorder=3.15,https://d.r.o/whatever
12:11:22thanevimSorry if this is terribly offtopic, but Rockbox may see a spike in new users as a result of getting shown on Techmoan's latest YouTube video
12:11:50speachy(will the old/existing releases use that URL or ignore it? if they use it, I guess it needs to be http not https?)
12:12:04thanevimJust a heads up, since he gets tens of thousands of views in each video, and this one specifically got 43k today alone
12:15:59_bilgus_thanevim, cool thanks :)
12:19:37bluebrother^speachy: that daily part looks good to me. Adding a manualurl= too would be good as well. And since we're adding all those a buildurl= would also make sense.
12:19:44speachyI'm actually adding that too.
12:19:49speachybuildurl pointing to what?
12:20:04bluebrother^the build
12:20:29bluebrother^pretty much what rbutil.ini has as build_url right now
12:20:54bluebrother^so we're consistent for all downloads we do :)
12:20:57speachyah, ok.
12:21:19bluebrother^could even leave out the individual urls then, unless we want to / need to specify different ones depending on the target.
12:21:24speachyI'm adding 'manualurl=/' and 'manualpdfurl="/.../bla.pdf'
12:22:15bluebrother^not sure if we should keep having a link to the html manual though.
12:22:45bluebrother^so previously I had three: pdf, zip, and html (online). Just changed that so I only have one URL, and use %FORMAT% to put in the extension.
12:23:08bluebrother^are we expecting those different versions to end up in different locations?
12:23:17speachyI think the html manual is only useful for online browsing; otherwise pdf is far better.
12:24:26bluebrother^there's one exception: screen readers. At least in the past the pdf wasn't friendly for those, which is one of the reasons why we even added the html version.
12:24:47bluebrother^being able to read it online is just an additional benefit :)
12:25:12bluebrother^so I've now also made a selection if you want to get the manual in pdf, zipped html, or extracted html
12:25:32speachyshould I use 'voiceurl' or 'voice_url' ?
12:25:33bluebrother^we had a link to the online manual before. I removed that one for now, not sure if we want to keep that.
12:26:12bluebrother^doesn't matter to me :)
12:26:49speachywell, I mean rbutil.ini uses the underscore
12:26:52bluebrother^I'd vote for consistency. Apart from that I don't care too much if its whateverurl or whatever_url :)
12:27:31bluebrother^yes, but I have to rework that anyway. You started without _. But using the same as in rbutil.ini could be useful.
12:27:49bluebrother^but then again, since rbutil.ini is internal to Rockbox Utility I could also easily change it there.
12:28:08speachynothing uses the infra side of things yet.
12:28:15speachyso that's easy to change
12:28:27bluebrother^right. So I guess you could simply do it as rbutil does.
12:31:25 Quit johnb7 (Ping timeout: 264 seconds)
12:31:50 Join TheLemonMan [0] (~lemonboy@irssi/staff/TheLemonMan)
12:31:59fs-bluebotBuild Server message: New build round started. Revision f598ef9c27, 293 builds, 9 clients.
12:32:01speachywhat should I use in the daily stuff? %REVISION% or someting else? (ie to map to the datecode)
12:33:14bluebrother^I'd use %VERSION%, and change all others to that as well. For a daily build the datecode is the version, for a release it's the release version number. Then I can internally treat them the same.
12:33:34bluebrother^rbutil.ini currently uses RELVERSION, but that's something I'd like to change.
12:34:03bluebrother^there's also a bit of an inconsistency with TARGET and MODEL. I need to clean that up as well.
12:34:12bluebrother^bit rot :)
12:35:56speachyso, um, tell me what you want it all to say, and I'll make it so on the build-info side
12:36:17speachydaily now uses %MODEL% and %REVISION%
12:36:32speachyrelease uses RELVERSION
12:36:44speachyand bleeding/development has no templates at all yet.
12:36:58speachyit'll have to wait until tonight though; I need to run now.
12:37:45bluebrother^just use VERSION for all of those. So daily REVISION -> VERSION, release RELVERSION -> VERSION. That will simplify things for me.
12:39:00speachyand with templates, we can ditch the per-item URLs?
12:39:19speachyok. I'd rather do templates in build-info anyway.
12:39:53bluebrother^we'd still have the option to add them on an as-needed base. Might prove useful for release candidates (which are the ones that was originally added)
12:40:14bluebrother^but if there is no per-item url we'll simply use the template.
12:40:27bluebrother^so hopefully best of both worlds :)
12:40:58speachyneed to fix/add a font package for daily builds.
12:42:32bluebrother^didn't we have a fonts pack for dailies?
12:42:39speachyno, just a symlink to the last release
12:42:49speachywhich is missing now.
12:42:56speachygotta scoot
12:42:57bluebrother^ah. Well, build-info could simply point to that file.
12:43:03fs-bluebotBuild Server message: Build round completed after 665 seconds.
12:43:10bluebrother^ok, cya
12:43:13fs-bluebotBuild Server message: Revision f598ef9c27 result: All green
12:43:14fs-bluebotBuild Server message: New build round started. Revision 64fe7e03a8, 293 builds, 9 clients.
12:45:51speachybluebrother^: current output:
12:46:37speachythe [daily] stuff will be updated tonight.
12:46:44speachy[development] is all that remains.
12:50:49_bilgus_cd ..
13:00:26fs-bluebotBuild Server message: Build round completed after 1032 seconds.
13:00:28fs-bluebotBuild Server message: Revision 64fe7e03a8 result: All green
13:02:06braewoodsinteresting. the h300 schematic is rather helpful.
13:02:31braewoodsi need to figure out how to access the ISP1362 pins to even begin to write this driver
13:07:16 Nick mendelmunkis is now known as mendel_munkis (
13:11:34_bilgusNope no issues with the clip+ bootloader at head or the clip+ firmware at head
13:12:12_bilgusI managed to fall into the trap of using linux text editors leaving a newline
13:14:25 Quit JanC (Remote host closed the connection)
13:14:48 Join JanC [0] (~janc@lugwv/member/JanC)
13:23:50braewoodson iriver h300... 2 DMA channels are already used for PCM...
13:23:56braewoods1 DMA channel is used by the LCD
13:24:02braewoodsthat leaves one unused
13:25:25braewoodsisp1362 can use up to 2 DMA channels...
13:25:44braewoods_bilgus: how do I know if DMA is possible for a chip?
13:26:20braewoodsthe ISP1362 datasheet mentions certain of its pins... but i don't know if it is even possible to use DMA with it.
13:26:28_bilgusnot my forte bud
13:26:46braewoodswho could I ask then?
13:27:10braewoodsi can always use PIO if i must but
13:27:13braewoodsfrom what i can tell
13:27:21braewoodsthe isp1362 interrupt lines are connected
13:27:26braewoodsbut i don't know if the DMA pins are
13:27:31_bilguswods or pamaury I imagine
13:29:29braewoodsjust looking at the driver code i realize there's a lot of time the code is spending twiddling its thumbs
13:29:41braewoodsso i'd like to find a way to let it do other stuff if possible
13:29:44braewoodswhile waiting on the cihp
13:30:31 Quit mendel_munkis (Remote host closed the connection)
13:30:49 Join mendel_munkis [0] (
13:34:12 Quit _bilgus (Remote host closed the connection)
13:34:21braewoodsi think it's save to say DMA is probably not an option
13:35:14braewoodsbut hm
13:36:05 Join _bilgus [0] (
13:45:03braewoods_bilgus: i think i know what i can do. i can't use the DMA pins of the chip but i can probably use the unused coldfire DMA channel to copy data from the chip when it signals that there's data to read.
13:45:16braewoodsor so
13:45:47braewoodsmay as well. if the dma channel hasn't found a use in these 15 years i don't think it would be a problem to use it now.
13:47:11 Join _bilgus__ [0] (~bilgus@
13:47:26 Quit _bilgus (Read error: Connection reset by peer)
13:51:47 Quit _bilgus__ (Ping timeout: 265 seconds)
13:54:59***Saving seen data "./dancer.seen"
14:11:22 Join johnb7 [0] (
14:21:27 Join _bilgus__ [0] (
14:24:23 Join livvy [0] (~livvy@gateway/tor-sasl/livvy)
14:37:16 Quit Acou_Bass (Quit: ZNC 1.8.1 -
14:42:59 Join Acou_Bass [0] (
14:43:03_bilgus__braewoods, what will that get you though? I guess you set upthe transfer and pool for completion later?
14:44:45braewoods_bilgus__: in theory but i'll look at it later.
14:44:56braewoodsi still need to get this working first
14:51:25 Quit johnb7 (Ping timeout: 246 seconds)
14:51:54yangWhich iriver model is this?
14:55:06 Join Soap [0] (~Soap@rockbox/staff/soap)
15:09:43 Quit livvy (Ping timeout: 240 seconds)
15:22:54 Quit fs-bluebot (*.net *.split)
15:22:54 Quit lebellium (*.net *.split)
15:22:54 Quit J_Darnley (*.net *.split)
15:22:54 Quit Ckat (*.net *.split)
15:24:10_bilgus__yang, iriver IPF?
15:25:02_bilgus__sorry IFP
15:25:21 Join fs-bluebot [0] (
15:25:21 Join lebellium [0] (
15:25:21 Join J_Darnley [0] (
15:25:21 Join Ckat [0] (~Ckat@xn--z7x.xn--6frz82g)
15:33:16 Quit Airwave (Quit: WeeChat 2.3)
15:41:51braewoodsspeachy: is USB_NUM_ENDPOINTS including or excluding the 2 control endpoints?
15:46:56braewoodsthe isp1362 technically has 16 endpoints but only 14 can be programmed
15:47:06braewoodsso not sure which to put
15:55:03***Saving seen data "./dancer.seen"
16:10:05 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
16:10:20 Join fs-bluebot_ [0] (
16:10:44 Quit fs-bluebot (Ping timeout: 260 seconds)
16:11:40 Quit bluebrother^ (Ping timeout: 256 seconds)
17:03:32 Quit lebellium (Quit: Leaving)
17:31:19 Quit kakaka (Remote host closed the connection)
17:31:53 Join kakaka [0] (~koniu@gateway/tor-sasl/koniu)
17:44:04 Quit TheLemonMan (Quit: "It's now safe to turn off your computer.")
17:55:06***Saving seen data "./dancer.seen"
17:55:24speachybraewoods: depends on how the chip is attached to the SoC. Is it on the main/parallel memory bus? Or is it on a serial bus ala SPI?
17:56:00braewoodsspeachy: you mean the isp1362?
17:56:18speachyif it's on the memory bus, you can use the SoC's built-in DMA controller to write data to the chip. Or perhaps the chip has its own busmastering capabilities.
17:56:20braewoodsit's connected directly to the GPIO pins in some places. the rest is accessed through special addresses.
17:56:31 Join amiconn_ [0] (jens@rockbox/developer/amiconn)
17:56:31 Nick amiconn is now known as Guest40560 (jens@rockbox/developer/amiconn)
17:56:31 Quit Guest40560 (Killed ( (Nickname regained by services)))
17:56:31 Nick amiconn_ is now known as amiconn (jens@rockbox/developer/amiconn)
17:56:38 Join pixelma_ [0] (marianne@rockbox/staff/pixelma)
17:56:38 Nick pixelma is now known as Guest61771 (marianne@rockbox/staff/pixelma)
17:56:38 Quit Guest61771 (Killed ( (Nickname regained by services)))
17:56:38 Nick pixelma_ is now known as pixelma (marianne@rockbox/staff/pixelma)
17:56:52braewoodsright now trying to figure out how to detect the presence of USB cable
17:57:11braewoodsi think i need to configure the interrupts
17:57:32braewoodsmy theory is that i can leave the chip asleep and have it wake up when a cable is detected
17:57:56speachyok, it has a 16-bit parallel bus. so it's presumably directly mapped into the SoC's address space somewhere.
17:58:01braewoodsit is
17:58:16braewoodsjust a lot of problems to solve.
17:58:31braewoodsi basically see needing to split where different usb code delegates to
17:58:57braewoodsthe USB data port on the h300 is connected to the 50606 chip that manages usb charging
17:59:04speachyso yeah, DMA is possible, assuming the dragonball has a spare memory-memory DMA channel you can use.
17:59:16braewoodsand the cypress chip which does USB UMS, all transparent to rockbox
17:59:20braewoodsfor the most part
17:59:51braewoodsso the issue is how to delegate the usb stack to the OTG chip and everything else to the hardwar
18:00:16braewoodswell one thing at a time
18:00:41braewoodsi expect i may end up adding conditional code to turn off part of the usb stack or reroute it
18:01:00braewoodslike i know i'll be disabling UMS in the usb stack since that's already hardware controlled
18:01:09braewoodson a different port no less
18:01:09 Join livvy [0] (~livvy@gateway/tor-sasl/livvy)
18:02:46 Join borkitall [0] (
18:02:51 Quit borkitall (Client Quit)
18:05:43 Quit livvy (Ping timeout: 240 seconds)
18:37:13speachyanyone recall what's needed to build a font package?
18:37:28 Join livvy [0] (~livvy@gateway/tor-sasl/livvy)
18:39:49 Join borkitall [0] (
18:46:23 Quit borkitall (Quit: borkitall)
18:56:23 Quit livvy (Ping timeout: 240 seconds)
19:00:40speachybluebrother: This should be complete:
19:01:32speachyHas everything you asked for except bootloader stuff, and font_url for daily&development builds.
19:09:36 Quit bluebrother (Disconnected by services)
19:09:41 Join bluebrother^ [0] (~dom@rockbox/developer/bluebrother)
19:10:22 Join fs-bluebot [0] (
19:12:49 Quit fs-bluebot_ (Ping timeout: 264 seconds)
19:17:49speachyok, now there will be daily fonts packages generated again. and font_url in [daily] and [development] point at it.
19:37:43braewoodswow. didn't expect to find this.
19:38:06braewoods"Interfacing the Philips™ ISP1362 USB OTG Controller to the MCF5249 ColdFire Microprocessor"
19:38:27braewoodsthis is helpful for at least understanding how to interface with the interrupts
19:38:59braewoodsoh, so that's why the gpio pins 5, 6, 7 were chosen? the GPIO interrupts are apparently bound to these pins.
19:40:26braewoodstime to find out if i can trigger the interrupts
19:55:09***Saving seen data "./dancer.seen"
20:18:42 Join borkitall [0] (
20:24:43borkitallhey guys, im trying to find where in the settings i can add silence between tracks
20:26:25borkitallim having a look in the iriver manual atm
20:26:35borkitallthe h320
20:29:25borkitalli remember seeing it somewhere, just dont know exactly where
20:41:38 Join DarkestEx [0] (
20:42:02DarkestExHi, I've stumbled over chatlogs talking about the Sony NW-A Series of Walkmans
20:42:12DarkestExWhat OS are they running and how's the firmware encrypted?
20:42:17DarkestExDoes anyone know?
20:47:22DarkestExBrowsing through the git repo, it appears that I have found the answer to the first question:
21:29:49DarkestExTo answer the second question, it actually runs Linux (that took quite a while to get there):
21:30:31 Part DarkestEx
21:54:22 Quit prof_wolfff (Ping timeout: 272 seconds)
21:55:12***Saving seen data "./dancer.seen"
22:09:56 Join prof_wolfff [0] (
22:31:40 Quit prof_wolfff (Ping timeout: 246 seconds)
22:35:55 Join Stanley00 [0] (~stanley00@unaffiliated/stanley00)
23:55:13***Saving seen data "./dancer.seen"

Previous day | Next day