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 2020-11-14

00:00:02 Join [7] [0] (~quassel@rockbox/developer/TheSeven)
00:07:19***Saving seen data "./dancer.seen"
00:58:17braewoodsspeachy: first i'd ask if the port is worth saving
00:59:42braewoodssome of these are so old i kinda wonder a bit
01:00:06braewoodsnot to mention super rare
01:00:33braewoodsseems like the irivers are the most common coldfire targets
01:01:39braewoodslike the mpio units seem completely unavailable
01:02:31braewoodsand the only cowon i see for sale is the x5
01:02:43braewoodsthe other models seem to be really rare these days
01:03:49braewoodsanyway i'm going to do research on MTP in the near future, work on that after i finish the bootloader project
01:04:13braewoodsbut it appears i will have to write our own support code for it since all the MTP libraries of relevance are C++
01:04:23braewoodswe need a C port
01:05:20braewoodsi'll also need to figure out what ports can even use MTP... i assume you need something like OTG.
01:23:37braewoods_bilgus: ok, the bootloader i just build from master works again. i think it's fixed now for sure.
01:23:47braewoodsi'll test H120 next
01:24:08braewoodsjust want to confirm it working before i start tinkering
01:28:07braewoodsspeachy: though i think you've got some oddness in the new usb code.
01:28:36braewoodsspeachy: when i first connect from reset settings... i see a menu selection appear briefly and then disappear
01:28:41braewoodsit basically behaves as charge only
01:29:01braewoodsi go into the settings to select mass storage
01:29:19braewoodsthat fixes it though it was already set to that apparently
01:29:23braewoodsno idea what's going on there
01:29:28braewoodsthis is on the iriver h320 i have
01:29:31braewoodslatest master build
01:48:04braewoodsH120 still works on master
02:07:23***No seen item changed, no save performed.
02:07:59 Quit massiveH (Quit: Leaving)
02:16:21braewoodsthe H300 bootloader starts when usb is connected but not when DC is connected
02:16:40braewoodsthis code suggests it should start when DC is connected
02:17:55braewoodsspeachy: i noticed the original author had partially implemented a charging screen in the h300 bootloader which is not present in h100. do you think i should keep it and fix it up or rip it out? i can look into backporting this into H100 bootloader too.
02:31:11braewoodsfor my own sanity i think i'm just going to disable the check in iriver_flash for valid BL
02:31:25braewoodsPITA to keep updating it during development
02:38:37braewoodswow. found another bug.
02:38:50braewoodscharger_inserted() returns false whether i am connected to AC or not in the bootloader
02:47:26 Join prof_wolfff [0] (
02:59:42braewoodsseems i found the bug but let me confirm
03:05:39braewoodsspeachy: i found an issue at least in H300 bootloader for charger detection. it's because the power_thread_inputs variable is not initialized when the BL calls charger_inserted() from firmware/powermgmt.c
03:06:07braewoodsbut when i get the most recent data directly from power_input_status()
03:06:10braewoodsit works fine
03:06:21braewoodsso the code for initializing it is not being run
03:08:01braewoodsi can think of 2 solutions
03:08:24braewoodseither we patch the convenience functions to just use the function directly when compiling for the bootloader
03:08:46braewoodsor we need to initialize it somehow, perhaps when power_init is called
03:21:50braewoodson second though
03:21:52braewoodsi'll try this
03:30:12 Quit advcomp2019_ (Ping timeout: 260 seconds)
03:40:37 Quit prof_wolfff (Ping timeout: 260 seconds)
03:43:25braewoodsok the thread init doesn't work at all
03:43:44braewoodsi think we're going to need some custom code for the BLs
03:44:33braewoodsi tried launching the powermgmt_init but it doesn't seem to work for the iriver h300
03:44:39braewoodsin bootloader mode anyway
03:46:47braewoodsnot to mention using it makes the BL grow by 2K
03:57:14braewoods_bilgus: is there some way to override the stock charger_inserted() and charger_inserted() functions? i get linking errors, not surprisingly
03:57:32braewoodspower_input_present() function
03:58:07braewoodsi'm thinking of creating a version of these 2 that is basically a custom version of the detection code
03:58:22braewoodsto keep the size down of the BL and such
04:03:48 Join lebellium [0] (
04:06:51 Join advcomp2019 [0] (
04:06:51 Quit advcomp2019 (Changing host)
04:06:51 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
04:07:26***Saving seen data "./dancer.seen"
04:32:11 Join TheLemonMan [0] (~lemonboy@irssi/staff/TheLemonMan)
04:51:40 Quit S|h|a|w|n (Read error: Connection reset by peer)
05:14:46braewoodsok. implemented a proposed solution.
05:14:49braewoods g#2973
05:14:51fs-bluebot_Gerrit review #2973 at : rockbox: implement an alternative to detect_charger() for bootloaders by James Buren
05:16:31 Join advcomp2019_ [0] (
05:16:31 Quit advcomp2019_ (Changing host)
05:16:31 Join advcomp2019_ [0] (~advcomp20@unaffiliated/advcomp2019)
05:19:16 Quit advcomp2019 (Ping timeout: 258 seconds)
05:23:03 Join prof_wolfff [0] (
05:32:28braewoodsah. so that's why the unit turns on when AC is connected.
05:32:32braewoodsand has charging screen
05:32:38braewoodsthe OF of the H300 did that
05:33:42braewoodslet's find out what the H120 does.
05:40:51braewoodsok. nothing.
05:41:01braewoodsthe H120 is hardware managed evidently
05:41:10braewoodsso not much point to a charger screen
05:54:53 Quit t0mato (Quit: Ping timeout (120 seconds))
06:04:43 Join MrZeus [0] (MrZeus@gateway/vpn/mullvad/mrzeus)
06:07:28***Saving seen data "./dancer.seen"
06:10:49 Join t0mato [0] (t0mato@gateway/vpn/mullvad/t0mato)
06:21:37 Quit prof_wolfff (Ping timeout: 260 seconds)
06:29:30_bilgusbraewoods, i'm ok with that I think it might be cleaner to mark the function weak and override in the BL
06:36:10braewoods_bilgus: i already found a solution.
06:36:21braewoodsit's honestly not worth doing that for a single line function
06:36:23braewoodssee gerrit
06:36:36braewoodsunless you have a better idea?
06:36:46braewoodsi just implemented a feature on the H300 TODO list
06:36:56braewoodsusb charging now works in the bootloader
06:42:12_bilgusonly thing I don't like about it is that it kinda pollutes the namespace with a bootloader function I'd rtaher see it redefined for all bootloaders
06:43:11_bilgusI saw gerrit lol
06:43:27braewoodsi can see what i can do.
06:43:47braewoods_bilgus: can you review my other commit?
06:43:47_bilgusi'd probably just #ifdef BOOTLOADER
06:44:00braewoodsthing is i don't think all BLs need it
06:44:21braewoodsthere's some that manage to call powermgmt_init just fine evidently
06:44:27braewoodsbut not for coldfire
06:44:28_bilgusbut it won't hurt them and except for hosted I doubt the power thread is even running
06:45:06braewoodsi think that's incorrect. it is running in the regular native firmware
06:45:14braewoodsi know that
06:45:29braewoodsbecause nothing else can initialize the internal variable
06:45:44braewoodsbut eh
06:45:44_bilgusthe regular native fw?
06:45:55braewoodsi know that since the debug menu app has
06:46:03braewoodscalls to charger_inserted()
06:46:21_bilgusdo you mean the the actualy running RB firmware?
06:46:21braewoodsbut ok. i'll redo the commit.
06:46:26braewoods... YES
06:46:39braewoodsthe battery debug options use it
06:46:44braewoodsand it works correctly
06:46:51braewoodsthe power thread must be running
06:47:28braewoodssince the thread's function is the only thing that writes to the internal variable
06:47:49braewoodsthe issue is this can't run for w/e reason in the bootloader
06:47:53braewoodsso the convenience functions don't work
06:48:14braewoodsi first tried using powermgmt_init but no good
06:48:21braewoodsso ye
06:48:26braewoodsok gimme a sec
06:49:45_bilgusI'll look at the bootloaders and see how many others are using a power thread in the bootloader
06:50:37braewoodsok go ahead
06:50:43braewoodsit's probably best to wait a bit to find out
06:51:16braewoodsi just wodnered why the charging mode didn't trigger in h300 and it was because charger_inserted() wasn't working correctly
06:51:25braewoodsin bootloader mode
06:51:37braewoodsthe workaround i found was to bypass it in bootloader mode
06:52:09braewoodsit probably worked at one time but it doesn't now
06:52:14_bilgusso just ipod6g and imx233
06:52:14braewoodsquestion is how do we want to fix it?
06:52:30braewoodsi tried adding it to h300
06:52:33braewoodsbut it didn't help
06:52:57_bilguslet me see what they are doing with it idealy I'd like them to not use the thread
06:53:14braewoodsnot sure they really need it honestly
06:53:50_bilgusthey shouldn't after your change I think
06:54:35braewoodsi'll finish implementing it then
06:58:06 Quit Misanthropos (Ping timeout: 256 seconds)
07:02:25braewoods_bilgus: sent new one. g#2975
07:02:27fs-bluebot_Gerrit review #2975 at : rockbox: revise charger_inserted and power_input_present functions by James Buren
07:03:48_bilguson the ipod and imx they both need the thread to set up charging
07:04:06braewoodsso, undo the change?
07:04:26braewoodsor how shall we implement it?
07:04:30_bilgusits not going to hurt them to bypass
07:05:06braewoodsi'll revise my other patch
07:05:35_bilgusthey can keep calling the thread to do their init its not going to hurt to give a reacharound on the insert funcs
07:12:30braewoods_bilgus: ok. both are done. review if you got time?
07:21:04fs-bluebot_Build Server message: New build round started. Revision d5a2aeb, 293 builds, 9 clients.
07:21:10_bilguslooks good
07:22:44braewoodswell that's another one off the TODO list
07:23:49_bilgusSO the part of it probably used to work I wonder what got removed from that player that it no longer is able to do the thread
07:25:14_bilgusor was it that the power thread evolved and someone forgot it
07:26:15braewoodsi could trace it but i don't think it's worth it.
07:28:15_bilgusnah just rhetorical
07:28:58braewoodsok what else does the bootloader need
07:30:13braewoods"Add possibility to control USB hub setting"?
07:30:33braewoodsi can't see the value in that though i know the OF has an option for setting 2 different USB modes
07:30:43braewoodsonly one of them makes disk mode work in the OF
07:30:55braewoodswhat good is the other mode?
07:31:02braewoodsRB seems to use the one that works already
07:32:08braewoods"Option to remove the bootloader from the flash" ?
07:32:20braewoodsno point. that basically means returning to OF
07:32:33braewoodsif people want that they can use iriver_flash to reflash it then reflash from the OF
07:32:55braewoodslow value; already achieveable via another set of options
07:33:06braewoodslet the user do it if they want to go back to pure stock OF
07:35:39braewoodsnext up...
07:35:47fs-bluebot_Build Server message: Build round completed after 883 seconds.
07:35:51braewoodsbacklight turn off during usb mode it appears
07:36:00fs-bluebot_Build Server message: Revision d5a2aeb result: All green
07:36:36_bilgusI never like removing OF
07:36:48_bilgusI think we talked about this prior
07:37:21braewoodsiriver_flash allows for return to OF
07:37:29braewoodsbut it is mainly used for updating bootloader
07:38:06braewoodsright now i'm just exhausting the todo left the previous bootloader author left behind
07:38:19braewoodsthen i'll see what else i want to do
07:38:37braewoodsat the very least i would like to see if i can port over the H100 boot from rom stuff
07:41:08_bilgusbraewoods, I think we might have missed something thats would shrink the h300 bootloader
07:41:27braewoods_bilgus: what now?
07:41:41_bilguslet me double check but I think its still pulling in the power thread stuff
07:43:24braewoods_bilgus: doesn't seem likely. powermgmt_init is not present in the final elf binary.
07:43:49braewoodsit's currently sitting at 55k
07:44:00_bilgusnope its in the removed list
07:44:03braewoodsstill around 10k left to expand
07:44:43braewoodsi'll see what h120 does. i recall it turns off backlight in usb mode.
07:55:43 Join MrZeus_ [0] (MrZeus@gateway/vpn/mullvad/mrzeus)
07:56:24 Quit MrZeus (Ping timeout: 256 seconds)
07:59:28braewoodshm. this seems good. let's build and test it.
08:00:44_bilgusthe h300 can get the codepage_table removed I think odd that its getting pulled in
08:07:19braewoods_bilgus: ok. how would we do that?
08:07:30***Saving seen data "./dancer.seen"
08:07:34braewoodsi'm currently working on adding a small amount of code to implement a backlight timeout
08:07:46braewoodsit'll turn back on if buttons are pressed
08:08:44braewoodsthe loops already sleep for 1 second between updates in some manner
08:08:58braewoodsso it shuld be simple to define a timeout in terms of seconds past
08:16:36 Join sakax [0] (~r0b0t@unaffiliated/r0b0t)
08:27:47 Join prof_wolfff [0] (
08:27:58 Quit Moarc_ (Quit: i znowu NADMUCHAƁ BALONA)
08:28:56braewoodsok. this seems to be done.
08:29:46 Join Moarc [0] (
08:32:20 Join mendel_munkis_ [0] (
08:32:56braewoods g#2976
08:32:59fs-bluebot_Gerrit review #2976 at : h300: implement a backlight timeout function when in charge or disk mode by James Buren
08:34:49 Quit MrZeus_ (*.net *.split)
08:34:49 Quit mendel_munkis (*.net *.split)
08:34:49 Quit kugel (*.net *.split)
08:34:50 Quit vup (*.net *.split)
08:35:08 Join vup [0] (~~~~@
08:35:46braewoods_bilgus: 2 commits still waiting for merge.
08:36:05braewoodsanyway the timeout function seems to work just fine
08:36:14braewoodsthough perhaps it could be better optimized
08:36:20braewoodsi only found one issue that i have already fixed
08:36:33braewoodsi forgot to force the backlight back on when exitting them ode
08:37:14fs-bluebot_Build Server message: New build round started. Revision 5a22308, 293 builds, 9 clients.
08:37:43 Quit danielp3344 (*.net *.split)
08:37:43 Quit kadoban (*.net *.split)
08:37:43 Quit prg318 (*.net *.split)
08:37:43 Quit Strife89 (*.net *.split)
08:37:43 Quit fauweh (*.net *.split)
08:37:44 Join MrZeus_ [0] (MrZeus@gateway/vpn/mullvad/mrzeus)
08:37:44 Join kugel [0] (~kugel@rockbox/developer/kugel)
08:38:20_bilguswhy do we need to control the backlight timeout in bootloader?
08:39:41braewoods_bilgus: no idea but it was on the original author's TODO list in the wiki. see here:
08:39:46braewoods"H300: Shut off backlight properly in USB mode "
08:39:53braewoodsnot sure what else they could mean by that honestly.
08:40:09_bilgusI think that is a bad idea personally
08:40:39braewoodsi mean, we can just not implement it but i was trying to honor their last known intentions.
08:40:59 Join danielp3344 [0] (danielp334@gateway/shell/
08:40:59 Join kadoban [0] (kadobanmat@gateway/shell/
08:40:59 Join prg318 [0] (~prg@deadcodersociety/prg318)
08:40:59 Join Strife89 [0] (sid399903@gateway/web/
08:40:59 Join fauweh [0] (
08:41:16_bilgusif you are in bootloader its there for a specific reason either you can't get into fw or you are going around a bug in the fw
08:42:22braewoods_bilgus: i honestly don't care either way but...
08:42:29braewoodsi was just going off their notes.
08:42:39braewoodsi'll scrap it then.
08:44:15braewoodsok that's all on the TODO list then
08:44:18 Quit blbro[m] (Ping timeout: 246 seconds)
08:44:18 Quit kadoban (Ping timeout: 246 seconds)
08:45:10braewoodsjust one last feature
08:45:15braewoodsor so i think
08:45:24braewoodslet's compare the bootloaders now
08:47:47 Quit danielp3344 (Ping timeout: 246 seconds)
08:52:02fs-bluebot_Build Server message: Build round completed after 888 seconds.
08:52:03fs-bluebot_Build Server message: Revision 5a22308 result: All green
08:52:15_bilgusok as long as this works it should free 64k from the bootloaders
09:06:38 Join kadoban [0] (kadobanmat@gateway/shell/
09:08:52braewoods g#2977
09:08:54fs-bluebot_Gerrit review #2977 at : h1x0/h300: clean up the bootloader source a bit by James Buren
09:09:02braewoodssmall change to prepare for the next step i have in mind
09:10:04 Join ac_laptop [0] (~ac_laptop@
09:11:08_bilguswhen you are doing stuff like that just do another commit on top and submit both not sure why that last one didn't do both
09:11:28braewoodsNo idea.
09:12:06_bilgusalso sorry but can you separate the WS changes from the code changes
09:12:34braewoodsas you wish.
09:12:48_bilgustwo ways bot use a patch with only WS changes either make it the parent or do it as a child
09:13:07_bilguslong standing rule bitten me MANY times
09:13:34_bilgushaving to clean up my commits not in a code fail kinda way lol
09:15:43_bilgusok so shrinking codepage_table freed 1K in my bootloader but its 10k in the bin
09:15:54_bilgusoh yeah FFS clipplus is compressed
09:18:00braewoodsfirst one.
09:18:06_bilgusbraewoods, are your bootloader compressed or decompressed in any step?
09:18:13braewoods_bilgus: not to my knowledge.
09:18:21braewoodsthe main difference is optimizations
09:18:42braewoodsit's too big if you disable them entirely
09:19:57_bilgusok so this would truly help there let me make sure this work on the clip+ and then I'll pass it on to you should give you all kinds of room
09:20:53braewoodsi already got plenty for the foreseeable future
09:20:58braewoodsbut i guess it can't hurt.
09:21:17braewoodsthe H1x0 and H300 BLs use the last 64K of the ROM chip
09:21:26braewoodsso roughyl 65,536 bytes
09:21:33braewoodsit currently is around 55,000 bytes
09:22:14braewoodsstill the previous BLs they'd be replacing were around 45k
09:23:51 Join blbro[m] [0] (blbrostrat@gateway/shell/
09:25:34_bilgusappears to work fine its a simple one
09:29:29_bilgus g#2979
09:29:31fs-bluebot_Gerrit review #2979 at : Bootloaders −− Remove codepage_table by William Wilgus
09:29:47_bilguscould you try that on your bootloader?
09:30:27fs-bluebot_Build Server message: New build round started. Revision 37eabbf, 293 builds, 9 clients.
09:30:29braewoods_bilgus: sure. review my changes while i do that?
09:30:41_bilgusalready was
09:30:43braewoodsi rebased the oriignal commit
09:30:51braewoods g#2977
09:30:53fs-bluebot_Gerrit review #2977 at : h1x0/h300: clean up the bootloader source a bit by James Buren
09:31:18_bilgusI was gonna mention that but gerrit takes care of simple stuff
09:31:59braewoodssometimes i've had to do manual rebasing though
09:32:01braewoodsso eh
09:32:49_bilgusyeah it sucks when it fails
09:33:32braewoodsgimme a bit to test this
09:33:45 Join danielp3344 [0] (danielp334@gateway/shell/
09:34:05braewoodsnew size
09:34:13braewoodssaved about 3.2k
09:34:39_bilgusthats a smaller than the map too its 64k of bytes
09:34:59_bilgusdo you have a patching step as well?
09:35:08braewoodswhat's that?
09:35:22_bilgusor do you patch bootloader.iriver directly?
09:35:55_bilgusmeh for 3.2 k i'll leave it till there is a size issue
09:36:06braewoods_bilgus: the tools do it for me?
09:36:12 Quit Soap (Ping timeout: 256 seconds)
09:36:39_bilgusso there is compression for yours too just not as good as the clip+ apparently
09:37:25 Join Soap [0] (~Soap@rockbox/staff/soap)
09:37:27braewoods_bilgus: bootloader still boots
09:37:43_bilgusyeah they don't use it
09:38:06_bilgushosted might be one that would
09:38:24braewoodsmaybe just limit it to native targets for now
09:38:40_bilgusI'll put this in the abandoned pile till a time comes when we need room again'
09:38:56braewoodsi've still got 10k to spare
09:39:15braewoodsi don't see myself using anywhere near that just to port over the one remaining feature
09:39:20_bilgusclip+ gains 1k with it it has 1k to spare currently
09:39:49_bilgus1-1.5 forget the exact amount but its not much
09:40:01braewoodswhat else does the clip+ need?
09:40:07braewoodsisn't it effectively complete by now?
09:40:44braewoodsanyway i'm going to finish this up and try to cut a production bootloader candidate
09:40:58braewoodsnext week if this works out
09:41:38_bilgusno it is finished for now but its also not the only bootloader close to the edge like that
09:41:42braewoodsonly once i'm pretty sure they're ready
09:42:12_bilguswe will find more if given more coverage I think
09:42:40braewoodsthe iriver coldfire bootloaders still got plenty of room for common code expansion
09:43:12braewoodsanother thing that could help is toolchain upgrade
09:43:23braewoodsbut that should probably wait until after the next stable release
09:44:27speachywhat does the CP tabble actually do? translate filenames/user strings into display strings?
09:44:52fs-bluebot_Build Server message: Build round completed after 865 seconds.
09:44:53fs-bluebot_Build Server message: Revision 37eabbf result: All green
09:45:02speachyif so, I don't think there's anything in the hosted BLs that would care.
09:45:09fs-bluebot_Build Server message: New build round started. Revision aaa9a2c, 293 builds, 9 clients.
09:45:39speachyassuming the filenames the BL needs to hit are limited to ASCII. Which they are, in the case of RB.
09:47:38_bilgusspeachy pretty much
09:48:26_bilgusand my exact thought but since they are bootloaders I don't really want to mess with them unless its needed or profound
09:52:16braewoods_bilgus: thanks. now i'm going to port over the last remaining feature from the h1x0 bootloader
09:52:20braewoodsthe failsafe mode
10:01:22_bilgusmendel_munkis_, g#2290 might give you another avenue if removing apply glabal settings :P doesn't work
10:01:24fs-bluebot_Gerrit review #2290 at : synclist add method for setting selection color by William Wilgus
10:02:09fs-bluebot_Build Server message: Build round completed after 1020 seconds.
10:02:54fs-bluebot_Build Server message: Revision aaa9a2c result: All green
10:02:54_bilgusAll green...
10:07:34***Saving seen data "./dancer.seen"
10:24:59 Join efqw [0] (uid412670@gateway/web/
10:26:18speachythe remove global settings apply thing probably needs to go back in once the underlying issue is fixed (in 2290 maybe?)
10:28:57 Quit Acou_Bass (Ping timeout: 260 seconds)
10:33:15 Join Acou_Bass [0] (
10:40:02_bilgus2290 only enables the pieces I think it'd still need to be wired for plugins
10:40:41_bilgusor conversely set up in the quickscrren
10:43:32fs-bluebot_Build Server message: New build round started. Revision be1be79, 293 builds, 9 clients.
10:44:08 Join MrZeus [0] (MrZeus@gateway/vpn/mullvad/mrzeus)
10:47:01 Quit MrZeus_ (Ping timeout: 258 seconds)
10:57:23fs-bluebot_Build Server message: Build round completed after 831 seconds.
10:57:28fs-bluebot_Build Server message: Revision be1be79 result: All green
11:12:20 Quit sakax (Quit: Leaving)
11:19:33 Quit TheLemonMan (Quit: "It's now safe to turn off your computer.")
12:07:36***Saving seen data "./dancer.seen"
13:06:57braewoods_bilgus: i discovered a small issue with iriver_flash, not a real bug but something i'd like to correct
13:07:06braewoodsthe screen goes dark on the h300 during inactivity
13:07:10braewoodsand that may freak some people out
13:07:25braewoodsso how might i force it to stay on until iriver_flash exits?
13:15:14braewoodsok. g#2980 is the last major change i expect to make to the bootloaders for these.
13:15:16fs-bluebot_Gerrit review #2980 at : h1x0/h300: bring the two bootloaders more in sync by James Buren
13:46:26 Join amsomniac [0] (~nadya@2601:181:8300:4db::b45)
13:47:14amsomniacis there a way to enable debug info on the bootloader? I'm trying to figure out this ATA error
13:50:46braewoodsamsomniac: which bootloader?
14:03:25amsomniacThe ipod6g bootloader
14:04:22amsomniacIf people normally just modify the bootloader to get some more info that's cool too, I really like that the configure script lets me build a version of rockbox with more debugging info but until I can get it to boot I can't see it
14:05:11braewoodsso you can't even boot with the stable release?
14:05:15amsomniacI'm going to try putting DEBUG=1 in the Makefile
14:05:40braewoodslet me guess. you're using an ipod modded with some iflash thing?
14:05:42amsomniacNope, I'm starting to think it's a hardware problem, not a rockbox one−− I haven't been able to boot anything except the stock firmware
14:06:21braewoodsnot supported.,52560.0.html
14:07:28amsomniacthanks−− I might buy another SSD and then if that doesn't work give up
14:07:39***No seen item changed, no save performed.
14:08:06braewoodsi've been working to solve a similar issue with the iriver coldfire ports
14:09:01braewoodsthe OF they have tends to hate CF mods which are common for people upgrading theirs
14:09:19braewoodsand the existing RB bootloaders don't really like them either
14:09:30braewoodsbut it was fixed in master a long time ago
14:09:35braewoodsjust it never got released
14:22:01amsomniacthere must be a lot of configurations of mods and media players, it's amazing rockbox supports as much as it does.
14:23:48braewoodsthe iriver ones are probably one of the better non-apple ports
14:23:56braewoodsand also somewhat unique
14:24:30braewoodsthe h100 and h300 series both support common types of data/charging cables.
14:24:58braewoodsboth have the RB bootloader in ROM; same as the OF
14:25:07braewoodsso the OF is not stored on your storage device
14:25:09amsomniacNice. My sansa e200 and fuze are very happy with rockbox but I haven't used either in a while because I've lost their weird charging cables
14:25:26braewoodsthe h100 and h300 use barrel jacks
14:25:38braewoodsthough the h300 also supports usb charging
14:25:48braewoodsthe h300 also has color screen
14:25:55braewoodsh100 has a grayscale one
14:26:11braewoodsbut it also has LEDs indicating power or disk activity
14:26:22braewoodsthe h300 does not
14:26:26braewoodsjust kinda interesting
14:26:40braewoodsyou can mod them with CF cards
14:26:53braewoodsalmost any capacity
14:27:00braewoodsup to 2TB
14:27:03braewoodsif you're that inclined
14:27:42braewoodsi just use 64GB CF cards myself
14:27:54braewoodsthey make the player around 15% lighter
14:29:31amsomniacI'm glad they're still making CF cards in higher capacities, I thought they stopped for some reason.
14:30:17braewoodsmost of the ones i get are used
14:30:34braewoodseven if demand dries up though there's CF adapters to SD cards
14:30:47braewoodsthey're not too fast but the main bottleneck isn't the CF card
14:31:07braewoodse.g., i only get around 13 MB/s at maximum in my CF modded players
14:31:50braewoodsthey're the most cost effective option if you care more about price per GB
14:32:02braewoodsbut the write endurance is shit
14:32:13braewoodsso best only use if you don't plan to do a ton of writing
15:01:50 Quit _bilgus (Ping timeout: 264 seconds)
15:07:52 Quit amsomniac (Quit: Leaving.)
15:09:15 Quit amiconn (Quit: - Chat comfortably. Anywhere.)
15:09:15 Quit pixelma (Quit: .)
15:11:50 Join amiconn [0] (jens@rockbox/developer/amiconn)
15:11:51 Join pixelma [0] (marianne@rockbox/staff/pixelma)
15:13:21 Join _bilgus [0] (~bilgus@2605:a000:1301:89f6:1f3:380e:6252:f251)
15:40:44braewoodsspeachy: i've been researching mtp more. no real libraries that i can see. they're all tied to specific usb stacks.
15:41:07braewoodsmtp responders that is
15:41:38braewoodsspeachy: for what targets would it be useful anyway? from what i've seen it's mostly native targets with OTG support.
15:41:55braewoodsthe android ports already have MTP through android
15:42:37braewoodswhich leaves the Linux hosted ports but those can use something like umtp-responder
15:43:04braewoodswe need our own implementation for native
15:55:50braewoodsfor another day though
15:55:59braewoodsi still need to figure out how to make the usb communication work
16:03:49 Quit MrZeus (Ping timeout: 264 seconds)
16:06:39 Join MrZeus [0] (MrZeus@gateway/vpn/mullvad/mrzeus)
16:07:40***Saving seen data "./dancer.seen"
16:10:31 Join Misanthropos [0] (~Misanthro@
16:14:13 Join MrZeus_ [0] (MrZeus@gateway/vpn/mullvad/mrzeus)
16:17:01 Quit MrZeus (Ping timeout: 264 seconds)
16:38:19bluebrother^speachy: I'm currently reworking how build-info is handled. What do you think about changing the format of the daily / dev builds? We would need to keep the [bleeding] for compatibility, but could add a new [development] that uses pretty much the same format as [release] but have <player>=<revision>,<timestamp>
16:39:27bluebrother^and the same with a [daily] section for daily builds. We already support a [release-candidate] section with a format <player>=<version>,<url> so we could also add an optional ,<url> for [development] and [daily]
16:40:08bluebrother^that would simplify handling in Rockbox Utility a bit, since now I'm assuming that we always have daily / development builds (unless status = 0)
16:40:57bluebrother^but moving it to a similar format we have for [release] we could simply leave the ones out that don't have daily builds, and if it's not present we know the build doesn't exist.
16:46:04braewoodshm. i seem to have figured out how to tap into rockbox's usb stack
16:48:00braewoodspart one of MTP support... figuring out how to integrate into the usb stack
16:48:13_bilgusbraewoods, there is an override backlight function in the plugin.h list
16:52:14braewoods_bilgus: i figured it out. there's helpers to do what i needed.
16:52:49braewoodsalso started working on my WIP for MTP support
16:52:56braewoodsgot a skeleton drafted for future work
16:54:11braewoodsbut i see it being useful for safer file transfers
16:54:17braewoodshow much slower it'll be i don't know
16:54:39braewoodsthough since we use fat32 the argument about portability doesn't really fly
16:57:48speachybluebrother^: it's all generated by a couple of scripts in the www repo
16:58:42speachybluebrother^: is that a complete URL to the actual file, or a template of some sort?
17:00:19speachyand <revision> is only used for display purposes, so for [daily] we'd use a datestamp (eg '20201113') and [development] we'd use a shortened git id?
17:00:57speachybraewoods: pamaury's partially-completed MTP stuff is probably a better place to start for native targets
17:01:17braewoodsspeachy: i can reference it. we'll see where i get with it all.
17:01:53speachyon android base (pretty much only the dx50/dx90) we get it for free but on other hosted stuff I don't know if we can make any assumptions about it being there.
17:02:31braewoodsno idea but there are libraries we can use for Linux
17:02:45braewoodsprovided the target has a suitable kernel or so
17:03:13braewoodsthe easiest place to start though is native targets
17:03:43braewoods_bilgus: care to review my iriver_flash commits? they should be easy enough to review.
17:14:30speachybluebrother^: and do you only want the latest daily listed?
17:14:55speachy(I'm assuming you still want the [dailies] section with the timestamp / revision broken out?
17:48:02fs-bluebot_Build Server message: New build round started. Revision 33d42c2, 293 builds, 9 clients.
17:49:03 Quit Rower (Ping timeout: 260 seconds)
18:00:49fs-bluebot_Build Server message: Build round completed after 768 seconds.
18:00:51fs-bluebot_Build Server message: Revision 33d42c2 result: All green
18:07:42***Saving seen data "./dancer.seen"
18:20:44fs-bluebot_Build Server message: New build round started. Revision 7d9ac02, 293 builds, 9 clients.
18:28:58speachywould anyone mind terribly if I took down git+gerrit to do another upgrade?
18:30:42speachydowntime shouldn't be more than 10-15min.
18:33:08fs-bluebot_Build Server message: Build round completed after 744 seconds.
18:33:12fs-bluebot_Build Server message: Revision 7d9ac02 result: All green
18:33:13fs-bluebot_Build Server message: New build round started. Revision c10eea4, 293 builds, 9 clients.
18:43:59mendel_munkis_speachy: I am okay with that.
18:44:11 Nick mendel_munkis_ is now known as mendel_munkis (
18:44:14speachyonce the builders all finally go idle anyway.
18:44:16 Join amsomniac [0] (
18:45:10mendel_munkisnow I need to figure how to port a plugin off of memset(rb->lcd_framebuffer)
18:45:57fs-bluebot_Build Server message: Build round completed after 765 seconds.
18:45:58fs-bluebot_Build Server message: Revision c10eea4 result: All green
18:54:50_bilgusmendel have a look at the vp test program
18:55:09mendel_munkiswill do. thanks
19:06:05_bilgusthats probably closer to what you want just a raw fb ptr..
19:10:46speachyok, gerrit's going offline for a bit.
19:13:06 Quit amsomniac (Quit: Leaving.)
19:18:35 Quit lebellium (Quit: Leaving)
19:24:28speachyit's online again.
19:26:58mendel_munkishow is PLUGIN_BUFFER_SIZE determined at port time?
19:28:52mendel_munkisI'm looking at specs and I see the mrobe500 has P_B_S of 0x100000 with 256mbit of ram as opposed tothe fuze+ which has 0x80000 and 512.
19:32:31fs-bluebot_Build Server message: New build round started. Revision 3ef8455, 293 builds, 9 clients.
19:39:50speachylovely, broken.
19:41:27mendel_munkiswhat exactly is broken?
19:44:44speachyor rather, it's handing out a URL to the builders that's not working properly
19:45:20mendel_munkisby the way I like the page at
19:50:23fs-bluebot_Build Server message: New build round started. Revision c10eea4, 293 builds, 9 clients.
19:57:46 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
19:58:49 Join fs-bluebot [0] (
19:58:58speachyI'd appreciate it if folks could try out the new PolyGerrit UI
20:00:25 Quit bluebrother^ (Ping timeout: 240 seconds)
20:00:45 Quit fs-bluebot_ (Ping timeout: 240 seconds)
20:02:37fs-bluebotBuild Server message: Build round completed after 734 seconds.
20:02:38fs-bluebotBuild Server message: Revision c10eea4 result: All green
20:05:17_bilgusspeachy it gives 404
20:05:45speachyso it's not just my browser then. :/
20:06:17speachylooking at the www logs it's requesting clearly wrong URLs.
20:07:17speachyit's "working except for pulling up diffs
20:07:46***Saving seen data "./dancer.seen"
20:09:03 Quit prof_wolfff (Ping timeout: 260 seconds)
20:10:19speachyappears to be a problem with the reverse proxy.
20:18:57 Join lemon_jesus3 [0] (~lemon_jes@
20:20:45 Quit lemon_jesus (Ping timeout: 240 seconds)
20:20:45 Nick lemon_jesus3 is now known as lemon_jesus (~lemon_jes@
20:21:30_bilguslemon_jesus is looking for some help with the nand in the ipod 3g?
20:21:54lemon_jesusyes yes yes
20:23:11lemon_jesusI'm having a hard time figuring out the FIL
20:23:32__builtinyou want to ping [7], I think
20:24:15lemon_jesusis that what TheSeven goes by nowadays, or is that someone else?
20:24:25__builtinthat's TheSeven, yes
20:24:56__builtinI think he's on european time so you'll have to be patient
20:26:14lemon_jesusthat's cool, I've been picking away at understanding the interface. it's been fun but also infuriating!
20:26:57__builtinwhat exactly are you doing? the javascript code isn't encoding the URLs properly.
20:34:16lemon_jesus__builtin good question - the project started with me trying to double the internal storage size of the n3g by replacing the NAND. obviously this didn't work and now it's months later and I'm down this rabbit hole where I've been trying to understand how the S5L8702 NAND interface works and I figure as I investigate this problem, I can contribute
20:34:16lemon_jesusto a Rockbox n3g NAND driver.
20:36:36lemon_jesusI've primarily been doing some stuff in Ghidra to attempt to fully understand the interactions of the software with the S5L8702, but I'm stuck at understanding what these registers do. I'm at a point where I'm not even sure what to Google anymore, and the datasheets for other samsung/nand related chips don't really have analogs to what I'm seeing
20:36:37lemon_jesusin the code.
20:40:42__builtinwell, I wish you luck :)
20:40:49__builtinFTLs are notoriously difficult to RE
20:42:25fs-bluebotBuild Server message: New build round started. Revision 3ef8455, 293 builds, 9 clients.
20:42:44lemon_jesusheh yeah I've been getting that sense
20:43:07__builtinany chance you could share what you've got?
20:43:24__builtinI wouldn't mind poking around in ghidra
20:49:13mendel_munkiswell I switched to new gui and now I cant see anything
20:50:32 Quit ac_laptop (Ping timeout: 246 seconds)
21:02:34fs-bluebotBuild Server message: Build round completed after 1209 seconds.
21:02:35fs-bluebotBuild Server message: Revision 3ef8455 result: All green
21:03:35lemon_jesus__builtin yeah I'll send you what I got. I'll DM you.
21:05:16speachyFIXED IT. GAAH.
21:08:11speachy_bilgus: new UI should work properly finally
21:08:54speachy(apparently the old UI was more forgiving of a config error that was there before I took things over)
21:09:53_bilgusI can't resize the review panes and they no longer center in the screen
21:10:28speachy_bilgus: that's beyond my purview. I'm just glad it _works_ finally
21:10:55_bilgusok well the new ui sucks in the nicesest way possible lol
21:10:56speachythe next version drops the old UI entirely
21:11:23speachybut they do keep improving things so perhaps your complaint will become moot in the next upgrade.
21:11:24_bilgusI hope tey polished it then
21:11:52speachyIn this version (2.16) they officially deprecated the old UI, considering the new UI feature complete
21:12:14speachyspent _far_ too much time on this @#%@$ thing
21:13:36speachyheh, one thing I want to do is add a commit hook that will reject things if there's messed-up whitespace
21:13:49speachyor inappropriate tabs
21:14:02 Join massiveH [0] (
21:16:34_bilgusthat would be nice
21:19:16 Join ac_laptop [0] (~ac_laptop@
21:23:18 Quit ufdm (Read error: Connection reset by peer)
21:34:40 Join ufdm [0] (
21:42:40 Join MrZeus__ [0] (MrZeus@gateway/vpn/mullvad/mrzeus)
21:45:29 Quit MrZeus_ (Read error: Connection reset by peer)
22:03:05 Quit MrZeus__ (Ping timeout: 240 seconds)
22:04:59 Quit ufdm (Read error: Connection reset by peer)
22:05:02 Join ufdm_ [0] (
22:07:47***Saving seen data "./dancer.seen"
22:08:39 Quit ufdm_ (Client Quit)
22:08:51 Join ufdm [0] (
22:42:39 Quit TheEaterOfSouls (Remote host closed the connection)
22:46:18mendel_munkisit's not as bad as the old new UI
22:53:23speachy_bilgus: d'oh.
22:53:25speachygimme a few.
23:02:41speachy_bilgus: github access restored
23:03:36speachygoogle auth pending.
23:03:43speachy(they need to verify my changes)
23:05:31 Quit tchan (Quit: WeeChat 2.8)
23:13:23speachyoh ffs. Google mandates having a privacy policy
23:14:05_bilgusrabbit holes
23:14:15speachy...or rather, now they do
23:15:04_bilgusok so singing in allows me to set preferences that make it better but it still is off a bit but bearable
23:16:09_bilgushope the next one has a dark theme w/o resorting to css tricks
23:17:02_bilgusI like the show blame feature
23:34:49 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
23:58:37 Quit [7] (Ping timeout: 260 seconds)
23:59:00 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)

Previous day | Next day