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 2016-06-02

00:01:23gevaertsSo HZ=10, HZ=100 and HZ=1000 have the same write speed. I'd say that makes it rather unlikely that the thread setup is the main bloker
00:02:33gevaertsAlso, if the thread setup were the blocker (so not the USB hardware or driver), I'd say fuze+ should be faster than fuzev2 or the classic
00:02:39gevaertsAnd that's not the case
00:03:09gevaertsSo it follows that it's either the hardware or the driver
00:03:52*pamaury is still writing the code..
00:04:27 Quit paulk-collins (Quit: Leaving)
00:05:17pamaurygevaerts: in usb_storage.c, how does static struct buflib_callbacks dummy_ops; works ? is it garanteed that it will be initialised to 0 ?
00:05:22*pamaury does not remember this point of the spec
00:06:44gevaertspamaury: I'm not sure I understand the question
00:06:54gevaertsOh, C-wise?
00:07:04gevaertsYes, that's bss
00:07:10pamauryok thanks
00:07:29 Quit ender` (Quit: Washing your car to make it rain doesn't work.)
00:13:09 Quit smoke_fumus (Read error: Connection reset by peer)
00:16:29*pamaury has some untested code on the fuze+, now the libusb part...
00:25:31pamauryyeah the code does not work...
00:28:05 Quit edhelas (Ping timeout: 260 seconds)
00:32:28pamauryah, I miss a crucial detail
00:36:56*__builtin wonders if he could automatically rewrite times not in terms of HZ
00:43:41gevaerts__builtin: with what aim?
00:48:46pamauryhum, this is tricky, you can't quite loop, you need to make TDs active again
00:49:12__builtinto remove HZ-dependence
00:49:29__builtini.e. replace sleep(1) with sleep(HZ/100)
00:55:14pamauryyeah, the code works !
00:55:20 Join saratoga [0] (123e11e0@gateway/web/freenode/ip.
00:55:29saratogacan anyone recommend a clip zip replacement battery?
01:00:56__builtin g#1336
01:00:57fs-bluebotGerrit review #1336 at : rewrite some sleep() calls so they are not HZ-dependent by Franklin Wei
01:01:40pamauryhum, in fact it works unreliably
01:11:52 Quit ZincAlloy (Quit: Leaving.)
01:14:01***Saving seen data "./dancer.seen"
01:22:06pamaurygevaerts: it now works reliably
01:22:33pamaurythe VERY interesting observation is that I can't push more than 22MB/s with my current code
01:22:53pamaurywhich basically setups a 64Kb transfer and resetup the same transfer is the IRQ handler
01:22:59pamauryso basically you can't expect to be faster
01:23:54__builtingevaerts: can you try g#1336 with HZ=10000?
01:23:55fs-bluebotGerrit review #1336 at : rewrite some sleep() calls so they are not HZ-dependent by Franklin Wei
01:24:00 Join ungali [0] (ungali@unaffiliated/ungali)
01:24:16pamaurygevaerts: stress code on host: pamaury/6cea172644423f6bc7b5faab163ab712"> use with g#1337
01:24:18fs-bluebotGerrit review #1337 at : HACK: usb test by Amaury Pouly
01:25:54pamauryif I increase the number of TDs/QH to 16 and push 16*16Kb=256KB per transfer, I barely reach 30MB/s
01:27:37pamauryincreasing numbef of TD to 32 and push 512Kb merely increases this figure to 32MB/s
01:31:03saratogathis is for SD?
01:32:25 Quit ungali (Quit: ungali)
01:32:37 Join ungali [0] (ungali@unaffiliated/ungali)
01:33:12pamaurysaratoga: no, those are just raw usb transfers
01:33:38pamaury(we don't even make any use of the data/buffer, and I wrote a custom hack to stress the hardware as much as possible)
01:34:11 Join saratoga_ [0] (123e11e0@gateway/web/freenode/ip.
01:34:20saratoga_isn't that close to theretical limits for usb2?
01:36:15pamauryit's quite far from the theory but quite close to the pratical one for bulk
01:36:21pamaury(which would be around 35/40 iirc)
01:37:34 Quit saratoga (Ping timeout: 250 seconds)
01:38:51pamaurygevaerts: ah no scratch that, I forgot to remove some IOC bits, so in fact even with 4 TDs/QH, I can get ~32MB/s
01:40:19pamauryso clearly there is room between 22MB/s for storage with ramdisk and 32MB/s top, but UMS has a very inefficient protocol too
01:42:25saratoga_oh yeah, i guess theoretical would be 48 MB/s ignoring protocol overhead
01:42:44 Quit ungali (Quit: ungali)
01:44:54*pamaury goes to bed
01:45:05 Join JdGordon_ [0] (~jonno@rockbox/developer/JdGordon)
01:48:10 Quit JdGordon (Ping timeout: 250 seconds)
01:49:45 Quit pamaury (Ping timeout: 260 seconds)
01:53:05 Join [Saint] [0] (77e01fae@rockbox/staff/saint)
02:17:36 Join JanC_ [0] (~janc@lugwv/member/JanC)
02:18:04 Quit Guest65764 (Ping timeout: 250 seconds)
02:18:37 Nick JanC is now known as Guest60369 (~janc@lugwv/member/JanC)
02:18:37 Nick JanC_ is now known as JanC (~janc@lugwv/member/JanC)
02:20:16 Quit Guest60369 (Ping timeout: 272 seconds)
02:34:35 Join idonob [0] (
02:34:54 Join idonob_ [0] (
02:35:13 Quit idonob_ (Client Quit)
03:06:36 Quit amiconn (Ping timeout: 250 seconds)
03:06:36 Quit pixelma (Ping timeout: 250 seconds)
03:14:05***Saving seen data "./dancer.seen"
03:16:28 Join amiconn [0] (~amiconn@rockbox/developer/amiconn)
03:16:28 Join pixelma [0] (~pixelma@rockbox/staff/pixelma)
03:45:56 Join Strife89 [0] (
03:50:41 Quit PurlingNayuki (Remote host closed the connection)
04:04:38 Join PurlingNayuki [0] (
04:07:17 Quit krabador (Quit: Take The Time)
04:27:20 Quit PurlingNayuki (Remote host closed the connection)
04:47:37 Join PurlingNayuki [0] (
04:54:27 Quit PurlingNayuki (Ping timeout: 264 seconds)
05:11:36 Join PurlingNayuki [0] (
05:14:08***Saving seen data "./dancer.seen"
05:27:30 Join Stroboko1p [0] (
05:30:56 Quit Strobokopp (Ping timeout: 272 seconds)
05:48:05 Quit dfkt (Read error: Connection reset by peer)
05:49:25 Join dfkt [0] (~dfkt@unaffiliated/dfkt)
06:13:04saratoga_that RkNanoD is interesting, haven't seen an mp3 player yet based on a modern ARM cortex M series and a relatively advanced fab node
06:14:55saratoga_probably not enough RAM to do much with though
06:23:04 Quit alexweissman (Read error: Connection reset by peer)
06:23:40 Join alexweissman [0] (
07:14:04 Quit nlogex (Ping timeout: 250 seconds)
07:14:09***Saving seen data "./dancer.seen"
07:17:40alexbobpwhat if rockbox people kickstartered the dream rockbox mp3 player?
07:24:25 Join edhelas [0] (~edhelas@
07:37:07[Saint]You need a ridiculous amount of funds to create a product that could compete in any way with any existing solutions.
07:42:03[Saint]I don't know about anyone else but I sure don't have an inside track to a reliable/quality fabricator.
07:42:46[Saint]well, not any that could provide anything at a reasonable quality as well as quantity.
07:45:40[Saint]if you use a milled case, it is expensive and time consuming, and if you use a stamped or injected case you'll need several iterations to get the masters done - which is also time consuming and expensive especially at small scale.
07:45:47 Quit pixelma (Remote host closed the connection)
07:45:47 Quit amiconn (Remote host closed the connection)
07:47:06[Saint]there's a whole lot of areas of production of that type of product that small players end up sinking money into until they haemorrhage to death.
07:47:14duo8it's ok call the chinese
07:48:00 Join pixelma [0] (~pixelma@rockbox/staff/pixelma)
07:48:03 Join amiconn [0] (~amiconn@rockbox/developer/amiconn)
07:48:50[Saint]I was assuming no less. But "outsource fabrication to China" is no magic bullet.
07:51:18[Saint]I think the best the community could hope for with minimal optimism would be schemamtics for an open player that people could source and build themselves.
07:52:06duo8what would a good player need anw?
07:57:17[Saint]I think at the moment the really obvious choice if you wanted to cobble together an "open" (I use the term loosely due to semantics about the GPU firmware provided by Broadcom, even if there is a community fork to reverse it) player is a Raspberry Pi Zero.
07:57:52 Join nlogex [0] (
07:58:32[Saint]a tiny OLED from alibaba et al, and some momentary pushbuttons, and you've got yourself the makings of a DAP goind.
08:00:47[Saint]If you're ordering units in the value of around a hundred thousdand or more, and willing to sign and NDA, you can even get the RPF to fabricate custom board layouts and populations for their existing supported products.
08:01:47saratoga_I think a good player needs to sell a couple million units per year in order to be good
08:02:00saratoga_otherwise you can't get good components, can't pay for development, etc
08:02:13saratoga_niche will inevitably mean expensive and probably mean bad
08:02:25[Saint]sourcing becomes a bitch unless you're willing to commit to continuous flow.
08:02:49[Saint]banana pi found that out.
08:03:58[Saint]then they were all like 'to get more silicon, Broadcom says we need to but in ther hundreds of thousands of units, which will bankrupt us'.
08:04:05[Saint]so, lol. RIP.
08:04:17[Saint]*to buy
08:06:17[Saint]If only Niel Young was a Rockbox fan.
08:07:29 Join soadkombucha [0] (dennis@gateway/shell/
08:08:12soadkombuchaOK so I'm having an issue... Trying to install the (apparently new) rockbox port available for the Xuelin iHiFi 770c. It will install fine, but I keep getting some RKW or BKW file is missing upon boot
08:09:06soadkombuchamay have figured it out... may have simply downloaded the wrong file
08:09:10alexbobp[Saint]: and less of a pseudoscience fan?
08:15:24soadkombuchafnow I'm getting an "ATA error: -4" and can't even reflash
08:15:55 Join ender` [0] (
08:18:20[Saint]soadkombucha: I don't think that port has anything to do with us, per se.
08:19:00saratoga_is that one of those ingenic sd card players?
08:19:52 Join einhirn [0] (
08:20:28soadkombuchasaratoga_: No, it's by a company called Xuelin. Lol.
08:20:46[Saint]saratoga_: AH, I see, seems it's that Vortex guy.
08:20:47saratoga_oh wait, its rockchip, not ingenic
08:21:21[Saint]soadkombucha: for the moment at least Vortex's builds have exactly no direct connection with the ROckbox project.
08:21:29soadkombuchaIt was worth a shot
08:21:33soadkombuchabeen waiting for the port for ages
08:21:43[Saint]Sort it out with him, I guess. Sorry.
08:22:07soadkombuchaeh I can wait
08:22:10[Saint]He's kinda acting independently for reasons that aren't in any way obvious.
08:22:14soadkombuchaThe default firmware is just utter shite
08:22:21soadkombuchaIt's terrifyingly bad
08:22:21[Saint]So waiting may not be an option.
08:22:42[Saint]As far as I see there's been no effort from him to mainline his work at all.
08:22:53soadkombuchaThat seems bizarre
08:22:55[Saint]One assumes if he were concerned in doing so he would.
08:23:05*[Saint] shrugs
08:23:21soadkombuchaFor an open source project, why split like that
08:23:39[Saint]No idea whatsoever.
08:25:30saratoga_amazon reviewer for that player says that he gets less than 3.5 hours battery life on a charge
08:25:43saratoga_i wonder if thats just terrible firmware or if they screwed up the hardware some how
08:25:54soadkombuchai got it for like 85 bucks
08:26:16soadkombuchai've gottwn like 6-8hrs though
08:26:40[Saint]holy shit, it's huge.
08:26:48[Saint]and the screen...isn't.
08:26:49soadkombuchanot really
08:27:04soadkombuchait's maybe a little bigger than a first gen mini
08:28:02soadkombuchathe sound quality is great though; for a budget players the downfalls can be overlooked imo
08:28:25[Saint]I bet those capacitive buttons are real fun.
08:28:36saratoga_this broken sansa clip i got on ebay works fine
08:28:42saratoga_i wonder what the seller was referring to
08:28:49soadkombuchai have no issues.
08:28:51[Saint]huh, nice.
08:30:03saratoga_it also came with 3.13 installed
08:32:01soadkombuchawhat's the significance of that version
08:32:11soadkombuchaeasiest to install rockbox over?
08:45:14[Saint]release, no more.
08:45:29[Saint]no easier, nor harder.
09:00:54 Quit [Saint] (Ping timeout: 250 seconds)
09:02:57 Quit edhelas (Ping timeout: 276 seconds)
09:03:49 Join girafe [0] (
09:14:12***Saving seen data "./dancer.seen"
09:19:22 Join petur [0] (~petur@rockbox/developer/petur)
10:04:28 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
10:11:49pamaurysaratoga_: the sony nw-E390 is based on RKnanoD, it looks like an cool soc but 1MB of RAM is tight
10:15:03 Join elensil [0] (~edhelas@2001:1c02:1903:d800:7d76:39eb:1776:53bd)
10:25:26 Quit mc2739 (Ping timeout: 246 seconds)
10:30:28 Quit girafe (Read error: Connection reset by peer)
10:39:27 Quit pamaury (Ping timeout: 260 seconds)
10:50:07 Quit elensil (Ping timeout: 272 seconds)
10:51:12 Quit GeekShadow (Ping timeout: 244 seconds)
11:00:41 Join elensil [0] (~edhelas@2001:1c02:1903:d800:e999:c1e4:301a:1ea6)
11:14:13***Saving seen data "./dancer.seen"
11:18:50 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
11:52:14 Quit foolsh (Ping timeout: 260 seconds)
11:53:01 Join foolsh [0] (
11:53:58 Join maruk [0] (
11:54:40 Quit pamaury (Ping timeout: 240 seconds)
12:48:55 Quit PurlingNayuki (Quit: PurlingNayuki)
12:49:26 Join PurlingNayuki [0] (
12:50:20 Quit PurlingNayuki (Client Quit)
12:51:59 Join PurlingNayuki [0] (
13:14:17***Saving seen data "./dancer.seen"
13:39:37 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
13:54:06 Quit puckipedia (Ping timeout: 244 seconds)
13:55:22 Quit PurlingNayuki (Remote host closed the connection)
13:58:22 Quit kugel (Ping timeout: 260 seconds)
13:59:10 Join kugel [0] (~kugel@rockbox/developer/kugel)
13:59:11 Join puckipedia [0] (
13:59:45 Join pamaury_ [0] (~pamaury@rockbox/developer/pamaury)
13:59:58 Quit pamaury (Read error: Connection reset by peer)
14:06:53 Join krabador [0] (~krabador@unaffiliated/krabador)
14:11:12 Quit pamaury_ (Ping timeout: 260 seconds)
15:14:19***Saving seen data "./dancer.seen"
16:15:56 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
16:21:19 Join PurlingNayuki [0] (
16:25:23 Join dfkt [0] (~dfkt@unaffiliated/dfkt)
16:25:33 Quit foolsh (Ping timeout: 240 seconds)
16:32:52 Join foolsh [0] (~starchase@2601:241:c200:4bc5:4d44:bd78:e1e0:3671)
17:14:20***Saving seen data "./dancer.seen"
17:36:59 Quit krnlyng (Ping timeout: 260 seconds)
17:38:09 Quit Guest85177 (Ping timeout: 244 seconds)
17:40:22 Join Guest85177 [0] (
17:49:32 Join krnlyng [0] (
17:49:41 Quit maruk (Quit: Leaving.)
18:04:31 Join JdGordon [0] (
18:04:31 Quit JdGordon (Changing host)
18:04:31 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
18:07:39 Quit JdGordon_ (Ping timeout: 264 seconds)
18:28:26 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
18:30:05 Quit sLite (Quit: No Ping reply in 180 seconds.)
18:31:22 Join sLite [0] (
18:46:43 Join edhelas [0] (~edhelas@
18:57:41 Quit edhelas (Ping timeout: 276 seconds)
19:06:37 Join p3tur [0] (~petur@rockbox/developer/petur)
19:14:23***Saving seen data "./dancer.seen"
19:32:38 Quit krabador (Remote host closed the connection)
19:38:24 Join edhelas [0] (~edhelas@
19:39:55 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
19:50:51 Join GeekShadow [0] (~antoine@reactos/tester/GeekShadow)
19:55:32 Quit edhelas (Ping timeout: 276 seconds)
20:00:29bluebrothergevaerts: basically I want to get that done since a while but am still quite occupied
20:01:23bluebrotherI might get around the next couple of weeks, but can't promise anything
20:01:40bluebrotherbut if you want to get something in first I can postpone that a bit :)
20:09:01 Quit Guest85177 (Ping timeout: 244 seconds)
20:09:07gevaertsbluebrother: the thing is that we want to start using that block of usb ids that we got from openmoko, and it would be handy if a released rbutil version knew about those first
20:10:39bluebrotherdo we already have them assigned to the players?
20:11:12 Join Guest85177 [0] (
20:11:16gevaerts__builtin did an initial allocation on last night
20:11:25gevaertsRandomly, to avoid any arguments :)
20:12:50bluebrotheras long as we have agreed on that adding them to Rockbox Utility is no big deal. And might be another reason for me to finally get around and create a new release :)
20:15:47__builtingevaerts: randomly?
20:16:02__builtinthey're still in alpha order as far as I can tell
20:16:13gevaertsOh, they are indeed
20:16:22gevaertsI vote to leave them alone now to avoid confusion
20:17:02*__builtin thinks randomization is the best idea for now
20:17:17__builtinhas anything been done using the alpha-order ones?
20:18:49gevaertsNot *yet* until bluebrother is very fast
20:19:10__builtinok, I'll randomize them now
20:19:28__builtinalpha order really doesn't make sense if we plan on adding more
20:22:04 Quit pamaury (Ping timeout: 260 seconds)
20:33:54 Join girafe [0] (
20:39:00 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
21:11:44 Join krabador [0] (~krabador@unaffiliated/krabador)
21:11:51*pamaury does not quite understand this obssession for "not alpha order"
21:12:50__builtinwell, if we add a new USB target, it breaks everything and could confuse someone who looks at the first few and assumes it's alphabetized
21:14:27***Saving seen data "./dancer.seen"
21:17:01pamaurysoadkombucha: I am working on merging the work from Vortex, but he clearly told me he is not interested in maintaining the stuff within the projet, so it won't be merge until I am completely satisfied with my rework. Also my time is limited, I am doing my best
21:18:13soadkombuchapamaury: OK. Let me know if you need any help with testing
21:18:22soadkombuchaSince I have the device readily available
21:18:41soadkombuchaI couldn't get it to load the files at all, kept going into rockchip mode
21:18:52soadkombuchaand then at one point it threw an ATA: -4 error
21:18:59soadkombuchaso back on defaut firmware for the moment
21:26:44__builtinis it a bad idea to rely on the internal workings of the MENU_* macros?
21:27:47pamaury__builtin: possibly, depends how badly I guess
21:28:27__builtinor, is there anyone who plans on ever changing them?
21:28:43pamauryI doubt anyone will change them dramatically
21:32:19__builtinalso, do you think g#1336 is a good idea?
21:32:20fs-bluebotGerrit review #1336 at : rewrite some sleep() calls so they are not HZ-dependent by Franklin Wei
21:32:51__builtincould there be some code that relies on sleeping exactly one tick, no matter HZ's value?
21:34:12pamaury__builtin: I think most sleep(1) simply mean "I want to sleep a ~1ms but sleep() won't let me do that, so sleep the smallest amount you can"
21:34:42__builtinwell, at least it won't /break/ anything, right?
21:36:09pamauryin fact some sleep(1) may also use to 1) yield 2) not refresh too often (like in plugins). I think it's worth looking at of all them quickly and define if the rewrite makes sense
21:38:01__builtinok, thanks
21:39:29__builtinwould it be a better idea to rewrite them with yield() calls?
21:47:53 Quit petur (Quit: *plop*)
21:48:12 Nick p3tur is now known as petur (~petur@rockbox/developer/petur)
21:51:23 Join JdGordon_ [0] (~jonno@rockbox/developer/JdGordon)
21:54:08 Quit JdGordon (Ping timeout: 260 seconds)
21:55:14 Quit elensil (Ping timeout: 258 seconds)
21:58:16pamaury__builtin: no, some of them are not yield
22:10:42 Join elensil [0] (~edhelas@2001:1c02:1903:d800:59bf:a0a6:7855:635f)
22:37:30 Quit TheSeven (Ping timeout: 250 seconds)
22:37:56 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
22:42:59 Quit Guest85177 (Ping timeout: 244 seconds)
22:44:37 Join edhelas [0] (~edhelas@
22:44:51edhelaspamaury: here ?
22:45:04pamauryyes but busy
22:45:29 Join Guest85177 [0] (
23:14:31***Saving seen data "./dancer.seen"
23:36:46 Quit edhelas (Remote host closed the connection)

Previous day | Next day