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 2021-04-22

00:10:01 Join daswf852 [0] (~daswf852@unaffiliated/dwf)
00:17:31***Saving seen data "./dancer.seen"
01:00
01:00:05 Nick mendel_munkis_ is now known as mendel_munkis (~mendelmun@ool-43568247.dyn.optonline.net)
01:26:27 Join ZincAlloy [0] (~Adium@2a02:8108:943f:d824:6cf4:e88:37b3:4460)
01:30:55 Quit ZincAlloy (Ping timeout: 260 seconds)
01:36:53 Join ZincAlloy [0] (~Adium@2a02:8108:943f:d824:6cf4:e88:37b3:4460)
01:41:07 Quit ZincAlloy (Ping timeout: 250 seconds)
01:54:04 Join ZincAlloy [0] (~Adium@ip5f5abcae.dynamic.kabel-deutschland.de)
01:58:48 Quit ZincAlloy (Ping timeout: 265 seconds)
02:00
02:04:30 Join nihilazo [0] (e33ac355b6@198.108.76.81)
02:17:34***Saving seen data "./dancer.seen"
02:18:20 Join petur [0] (~petur@199.59.5.11)
02:18:20 Quit petur (Changing host)
02:18:20 Join petur [0] (~petur@rockbox/developer/petur)
02:26:10 Quit bluebrother (Remote host closed the connection)
02:33:42 Quit ecs (Remote host closed the connection)
02:33:58 Join ecs [0] (esawady@sourcehut/interns/ecs)
02:41:32 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
02:42:14 Join fs-bluebot [0] (~fs-bluebo@55d4bc24.access.ecotel.net)
02:50:53 Quit edhelas (Remote host closed the connection)
03:00
03:56:27 Join S|h|a|w|n [0] (~shawn156@unaffiliated/shawn156)
04:00
04:16:03 Quit S|h|a|w|n (Quit: Leaving)
04:17:37***Saving seen data "./dancer.seen"
05:00
05:07:22 Quit petur (Quit: Connection reset by beer)
05:36:47 Quit Acou_Bass (Ping timeout: 265 seconds)
05:49:37 Join Acou_Bass [0] (~Acou_Bass@cpc96070-bolt17-2-0-cust175.10-3.cable.virginm.net)
06:00
06:01:13 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
06:17:38***Saving seen data "./dancer.seen"
07:00
07:09:22speachyhuh, the ipod6g with the shutdown issue has an ssd that claims to support ATA power management.
07:11:30speachyidentifies itself as "CF card"
07:15:51speachyI suspect the fix is still to disable power management commands, but we'll need a different heuristic.
07:32:23speachythe backlight shouldn't block anything...
07:45:39braewoodsspeachy: it sounds more and more like these pretend CF cards don't properly implement ATA
07:46:00braewoodsi've never seen issues like this with true CF cards
07:46:55speachyno idea.
07:46:57braewoodsi wonder if we could use the same approach the LK uses
07:47:06speachyand that is?
07:47:25braewoodslists of devices known to not do something properly to disable stuff selectively
07:47:35braewoodsthey use hard-coded tables for crap like this
07:48:03speachy(I tossed out a build that completely disables power management in the ipod6g's ata driver to see if that will matter)
07:48:21braewoodsif the iFlash adapters identify themselves in the ATA info we could selectively enable workaronds
07:48:54braewoodsassuming no better solution is found
07:49:03speachythere's a _lot_ of stuff out there that uses the same ATA->SD chipset not made by iflash.
07:49:20braewoodsi know.
07:49:36braewoodsbut should we punish everyone by disabling PM?
07:49:39speachynothing consistent I can use to generate a heuristic yet
07:49:56braewoodsor here's an idea
07:50:04braewoodswould making it a menu option workaround be viable?
07:50:49braewoodsthat way we can let the user choose what they want to use
07:57:23braewoodsspeachy: oh, this user on forums. they're using the extensions I added to disk info?
07:59:08braewoods... wow a bit of an attitude this one has
07:59:10braewoodso.O
08:00
08:04:51speachyI mean, not much I can do when my ISP has a major outage. heh.
08:10:28speachyI'm loathe to add a menu option for this but.. might be prudent.
08:17:39***No seen item changed, no save performed.
08:17:56 Join Saijin_Naib [0] (~Saijin_Na@2603-7081-1d05-7230-5545-1de4-4b23-f8cd.res6.spectrum.com)
08:19:31 Quit Saijin_Naib_ (Ping timeout: 245 seconds)
08:31:53braewoodsspeachy: simplest option without adding a ton of dynamic code heuristics
08:32:10braewoodsmaking it user selectable, defaulting to whichever is better for stability
08:33:24 Join cockroach [0] (~blattodea@pdpc/supporter/active/cockroach)
08:35:21 Quit speachy (Ping timeout: 245 seconds)
08:47:44 Join speachy [0] (~speachy@209.2.65.77)
08:48:52speachyhmm.. another possibility
08:49:00speachywe don't actually ever issue a FLUSH CACHE command
08:56:02braewoodsspeachy: doesn't sound like something we need to do normally
08:56:05braewoodsbut
08:56:09speachywhen going to sleep?
08:56:14braewoodsyea probably a good idea
08:56:17speachyexactly
08:56:23braewoodsor shutdown
08:56:29braewoodswe can combine with a timeout
08:56:30speachyI think with spinning rust it's sort of implied
08:56:32braewoodsor wait period
08:56:35speachythat when you sleep you flush things
08:56:37 Join massiveH [0] (~massiveH@ool-18e4e82f.dyn.optonline.net)
08:57:07braewoodsdoesn't matter for any read buffers
08:57:07speachybut god knows what these wonky CF/SD adapters actually implement
08:57:16braewoodsone way to find out
08:57:24braewoodsthe scientific method :D
08:58:41braewoodsspeachy: wouldn't it be funny if the problem was due to us forgetting to perform the ATA equivalent of fflush?
09:00
09:01:10speachyhuh, typo in the ATA-8 specification.
09:01:53braewoodsfrom the spec it sounds like this should be safe to issue on all targets before we're preparing to shutdown or so
09:02:45speachyit was added in ATA-4
09:03:14braewoodsso if the device advertises support for it we should probably use it sometimes
09:03:49speachyanother possibility is to completely disable powersaving stuff for anything that identifies itself as a CF card.
09:03:59speachy(or CFA-compliant, more accurately..)
09:04:06 Quit massiveH (*.net *.split)
09:04:06 Quit Acou_Bass (*.net *.split)
09:04:06 Quit Rower (*.net *.split)
09:04:06 Quit kugel (*.net *.split)
09:04:18braewoodsvia that bit I added to disk info?
09:04:29speachydunno. Just going by the spec here
09:04:50braewoodsI added those checks mostly as some extra info for users.
09:04:54speachyall of the dumps I have from the SD->CF thingeys advertise themselves as such
09:05:04braewoodsyou mean as
09:05:21braewoodsCF compatible, not Fixed, Removeable?
09:05:33 Join massiveH [0] (~massiveH@ool-18e4e82f.dyn.optonline.net)
09:05:33 Join Acou_Bass [0] (~Acou_Bass@cpc96070-bolt17-2-0-cust175.10-3.cable.virginm.net)
09:05:33 Join Rower [0] (~Rower@78-73-72-39-no2340.tbcn.telia.com)
09:05:33 Join kugel [0] (~kugel@rockbox/developer/kugel)
09:05:39braewoodsthat's what real CF cards do, most of the time
09:05:41speachy0x848a in the first word of the identify info
09:05:49speachyalso bit 2 of word 83
09:05:56speachy(treated as equivalent in the spec)
09:06:02braewoodsI see.
09:06:17braewoodsI know some things cared about the fixed bit and such
09:06:40braewoodsmaybe we should look at what Linux does for CF cards connected to an IDE bus
09:06:51speachybut, heh, CFA also mandates SLEEP support. :D
09:07:01 Quit Rower (Ping timeout: 245 seconds)
09:07:11braewoodslet's not disable PM just yet
09:07:23braewoodslast resort in my book
09:07:23 Join Rower [0] (~Rower@78-73-72-39-no2340.tbcn.telia.com)
09:07:34speachy(this is the iflash-quad-on-ipod-video issue, not the lcd stuck-on on the 6g)
09:07:42braewoodsi know
09:07:48braewoodsATA should have nothing to do with the LCD issue
09:08:23speachyactually I suspect the lcd-stuck-on might be entirely due to LCD settings; maybe the user configured it to stay on permanently, and the shutdown code can't handle that.
09:08:38braewoodswell during shutdown we should turn it off anyway
09:09:01speachybut −− the lcd isn't shut down until after the disk is shut down.
09:09:02braewoodsassuming that's even the issue
09:09:11speachyand that can block forever
09:09:29braewoodswell at least until the battery dies
09:09:31braewoods:D
09:09:40speachyor our sun engulfs the planet
09:09:51braewoodslol
09:13:14speachythe ipod6g ata driver doesn't use CMD_SLEEP, it uses CMD_STANDBY_IMMEDIATE
09:13:59speachywithout checking to see if it's supported, I might add.
09:14:07braewoodswe could also be dealing with stupid hardware / firmware that won't flush unless ordered to
09:16:31speachymaybe that's why this problem doesn't seem to be reported on the 6g series.
09:16:58speachystandby_immediate causes a cache flush
09:17:16speachyinteresting.
09:20:49speachystandby isn't as deep as sleep
09:20:58speachyso sleep implies flushing caches too
09:24:07_bilgusbahus on the forums posted about a 3rd party fat32 formatter that appears to be just the ticket for these windows users http://ridgecrop.co.uk/index.htm?guiformat.htm
09:24:41_bilgusreading the writeup on the cli version http://ridgecrop.co.uk/index.htm?guiformat.htm
09:25:29_bilgushttp://ridgecrop.co.uk/fat32format.htm
09:25:33_bilgusrather
09:26:19speachyoh that's good
09:26:26_bilgusanyway I never realized the sector size would be so important in the size of the FAT table
09:27:14_bilgusI mean DUH
09:27:55speachy_bilgus: yeah, it is one of the primary variables. : )
09:30:05speachy... though at this point, actually implementing exfat is a good idea.
09:30:38speachya substantial undertaking to be sure, but at least now it's actually viable legally.
09:32:07_bilguswouldn't you have to have fat to leave the bootloader and then switch to exfat?
09:32:35speachydepends on the target.. but yeah
09:32:44speachythe bootloader would have to support it too
09:33:19_bilgusI've put partitions on a sdcard
09:33:35speachymost of our targets don't support multiple partitions.
09:34:21speachythough that's easy to change from the rb perspective
09:34:22_bilgusas long as it picks up the first or do they fail to boot?
09:35:05speachy(ie we don't turn on MULTIVOLUME for most targets)
09:35:27speachyI think the various OFs don't generally care about anything other than the 1st partition
09:35:58speachybut one of he points of supporitng exfat is to not require folks to have to reformat their cards
09:36:24_bilgusyeah they come formatted that way
09:41:55speachyI wonder if there's any power disadvantage to using STANDBY_IMMEDIATE instead of SLEEP on the older spinning rust systems
09:42:30speachy(eg the irivers and 1g-5g ipods)
09:42:48speachy(whjich I think exclusively used toshiba disks)
09:48:08braewoodsiriver H10 uses a hitachi
09:48:11braewoodsoriginally
09:48:33braewoodsand seagate for the 5G/6G variants
09:49:19speachy(I'm talking about the 1.8" ones, not the microdrive-based ones..)
09:49:44speachyhuh, the ATA driver used to use STANDBY_IMMEDIATE but that was yanked in favor of SLEEP... in 2002.
09:50:03braewoodsthe H10 20GB doesn't use a microdrive
09:50:14braewoodsbut ok
09:56:24speachy; sighs.
09:57:22braewoods?
09:57:26braewoodsfind something?
09:57:48speachyspec-reading doesn't help much when it's the implementation that's b0rked.
09:58:07braewoodsthere's the spec and then there's reality
09:58:30braewoodsspec is only good if people always implement it right
09:58:37speachya lot of these devices claim to implement powermgmt.
09:59:31speachymaybe the right thing to do is use STANDBY_IMMEDIATE instead of SLEEP on these devices.
09:59:52speachybut I'll have to wait to see what the results of the last few test builds turn out to be
10:00
10:00:25speachy(eg use SLEEP on spinning rust but STANDBY_IMMEDIATE on things that advertise as CF?)
10:00:42speachywould be interesting to figure out what the ipod firmware actually uses
10:08:27 Quit pamaury (Ping timeout: 260 seconds)
10:08:51 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
10:17:42***Saving seen data "./dancer.seen"
10:22:14 Quit massiveH (Quit: Leaving)
10:34:49_bilgusI'm having a hard time figuring out a good way to represent the keymaps in a form easily digestible lua and easily parsable for me
10:36:16_bilgusbeing that multiple contexts are capable of being remapped to the same user_ctx
10:37:25_bilgusi'm thinking they should all carry a tag to their parent and then I can collect them when I want to display or generate a keymap
10:38:44_bilgusfor saving and restoreing I should be able to just serialize the table and restore without care to the order
10:39:38_bilgusI still need a way for a user to recover from bad keymaps
10:40:20_bilguslike maybe make it apply only if no buttons are pressed
10:40:28_bilgus?
10:41:13_bilgusif select is held on startup keymap is not applied
10:41:39_bilguswouldn't work for touch targets
10:41:55_bilguswould they need key remapping?
10:41:58speachywell, touch targets don't need keymaps in teh traditional sense
10:42:43_bilgusso select or vol up or something not currently used disableskey remap with a splash
10:44:39_bilgusIf I rempapped my direction keys in the main menu well shame on me but how do I recover? go into OF nad delete the remap file that would be annoying
10:45:27_bilgusotherwise it could apply temporary and only apply on startup if you were able to go set it
10:45:47_bilgusbut that seems annoying from a experienced user perspective
10:56:13speachywell, mucking with keymaps is something that's inherently dangerous and might cause your player to eat your cat
10:56:48speachyso.. as long as there is a documented way of resetting to "stock" keymaps I'd not worry about it.
11:00
11:00:27_bilgusdeleting the keymap does that
11:00:55_bilgusbut i figured if you can screw it up on player you should be able to recover on player
11:22:11speachywell, we have a "Restore to default settings" keypress on boot, right?
11:40:41 Join MrZeus_ [0] (~MrZeus@2a02:c7f:a0aa:4400:486b:5309:1b8:cfe0)
12:00
12:17:44***Saving seen data "./dancer.seen"
12:47:03 Join ZincAlloy [0] (~Adium@ip5f5abcae.dynamic.kabel-deutschland.de)
12:55:31 Quit atsampson (Ping timeout: 260 seconds)
13:00
13:12:18 Join atsampson [0] (~ats@cartman.offog.org)
13:25:03 Join lebellium [0] (~lebellium@89-92-69-66.hfc.dyn.abo.bbox.fr)
13:34:07 Join MrZeus__ [0] (~MrZeus@194.37.96.153)
13:35:26speachyoh joy, here's a dump from an actual iflash-branded device, and it _doesn't_ advertise itself as CF.
13:37:31 Quit MrZeus_ (Ping timeout: 260 seconds)
13:48:33 Quit nihilazo (*.net *.split)
13:56:15_bilgusspeachy, oh do we that will work as long as I make sure it happens before the keymap is loaded?
14:00
14:00:03 Join nihilazo [0] (e33ac355b6@198.108.76.81)
14:08:23 Join chris_s [0] (02cdf09c@dslb-002-205-240-156.002.205.pools.vodafone-ip.de)
14:09:23speachydon't hold me to that but I seem to recall it being there
14:11:20chris_sSpeaking of the "Restore to default settings" keypress: That's actually (for whatever reason) the Hold slider on iPods. Which is not a great idea, because iPods automatically turn on when docked and/or connected to power. It was one of the first things that motivated me to look into compiling my own build of Rockbox (since it's trivial to change in
14:11:21chris_scode), because having the settings reset all the time was so irritating in practice.
14:11:51chris_sI think there's actually a patch in Flyspray by someone, although i haven't looked at it
14:12:36chris_sFS #13193 (I think it would be a better idea to simply use a different button for that on iPods though)
14:12:37fs-bluebothttps://www.rockbox.org/tracker/task/13193 Add option to disable settings reset on startup (patches, unconfirmed)
14:17:48***Saving seen data "./dancer.seen"
14:18:32 Quit chris_s (Quit: Connection closed)
14:18:52speachyyeah, I'd concur
14:38:11 Quit fs-bluebot (Ping timeout: 240 seconds)
14:40:02 Quit bluebrother (Ping timeout: 246 seconds)
14:46:33speachywell, no change. drat.
14:47:09speachydisabling power saving altogether didn't change things, adding an explicit flush didn't change things.
14:47:21speachybacking off DMA didn't help either.
14:47:27speachyso there's _another_ failure mode going on.
14:55:48 Quit pamaury (Remote host closed the connection)
14:58:44 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
15:00
15:06:23 Join chris_s [0] (02cdf09c@dslb-002-205-240-156.002.205.pools.vodafone-ip.de)
15:07:57chris_s@speachy: I wonder why Rockbox on the iPod 4G on the other hand doesn't have the same issues with the iFlash adapter (at least using UDMA1)
15:08:09speachyah, but which iflash adapter..
15:10:19chris_sthe Solo, admittedly – I've had the same issues doing USB file transfers with that one in Rockbox on an iPod video though.
15:11:18speachyso the same adapter, behaving differently depending on which one it's plugged into.. interesting.
15:11:44chris_sI wasn't able at the time to test whether it *only* has issues in USB mode (on iPod 5Gs).
15:12:24speachyIIRC starting with the 5G they went away from the 50-pin CF connector?
15:13:30speachyif so, I wonder if it ends up getting pinstrapped differently and acts slightly differently −− and if that might show up in the identify_info dump
15:13:45chris_sno idea, yeah they started using ZIF connectors on the 5G: https://www.iflash.xyz/store/hdd-ribbon/
15:13:47speachy(eg on these dumps it's not identifying itself as a CF adapter)
16:00
16:00:47speachygreat, no immediate way to tell the microdrives apart from the actual solid-state CF cards.
16:17:51***Saving seen data "./dancer.seen"
16:42:41 Quit chris_s (Quit: Ping timeout (120 seconds))
16:54:08*speachy rubs his forehead.
16:56:19speachyok, the latest test build seemed to work. It ignores the flag saying power management isn't supported, and issues STANDBY_IMMEDIATE.
16:56:33speachyoh and a flush cache first.
16:57:45 Quit MrZeus__ (Ping timeout: 260 seconds)
16:58:22speachyif I go forward with this, what impacts will it have on spinning rust?
17:00
17:11:20braewoods"Find out next time the adventures of speachy."
17:11:25braewoods:P
17:11:31speachysame irc time, same irc channel
17:12:17speachywell, guess I'm done for the day. cat just completely engulfed my test rig.
17:24:19 Quit amiconn (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.)
17:24:19 Quit pixelma (Quit: .)
17:25:05 Join fs-bluebot [0] (~fs-bluebo@55d4bc24.access.ecotel.net)
17:25:19 Join pixelma [0] (marianne@rockbox/staff/pixelma)
17:25:19 Join amiconn [0] (jens@rockbox/developer/amiconn)
17:25:30 Join daswf8527 [0] (~daswf852@unaffiliated/dwf)
17:26:50speachybraewoods: if you could try g#3552 on your iriver devices, I'd appreciate it
17:27:02 Join bluebrother^ [0] (~dom@rockbox/developer/bluebrother)
17:27:05speachypreferably both modded and umodded.
17:27:32 Quit daswf852 (Ping timeout: 240 seconds)
17:27:33 Nick daswf8527 is now known as daswf852 (~daswf852@unaffiliated/dwf)
17:29:46*speachy pokes bluebot again.
17:31:02 Join MrZeus__ [0] (~MrZeus@2a02:c7f:a0aa:4400:486b:5309:1b8:cfe0)
17:37:37braewoodsspeachy: I don't have the original hard drive anymore except for the H10 5GB
17:40:41speachyI need some confidence this doesn't cause regressions, so anything that can be tested will help
17:41:09speachychris_s: if you could try g#3552 as well..
17:41:27speachyall I can say so far is that my mini2g hasn't eaten its microdrive yet.
17:50:05 Quit mendel_munkis (Remote host closed the connection)
17:50:23 Join mendel_munkis [0] (~mendelmun@ool-43568247.dyn.optonline.net)
17:53:04 Quit lebellium (Quit: Leaving)
17:53:12 Join chris_s [0] (02cdf09c@dslb-002-205-240-156.002.205.pools.vodafone-ip.de)
17:54:36chris_ssure thing  :)
17:55:39 Join ac_laptop [0] (~ac_laptop@186.2.247.129)
18:00
18:15:09 Quit chris_s (Quit: Ping timeout (120 seconds))
18:17:53***Saving seen data "./dancer.seen"
18:34:26 Quit ZincAlloy (Quit: Leaving.)
18:36:17 Quit ufdm (Quit: Leaving)
18:42:45 Join ZincAlloy [0] (~Adium@2a02:8108:943f:d824:894f:8a9a:1d21:ce0e)
18:44:38 Join Soap [0] (~Soap@rockbox/staff/soap)
18:44:57 Quit ZincAlloy (Client Quit)
18:56:06 Join ufdm [0] (~ufdm@c-73-164-63-214.hsd1.mn.comcast.net)
19:00
19:03:39 Join chris_s [0] (02cdf09c@dslb-002-205-240-156.002.205.pools.vodafone-ip.de)
19:04:41 Quit Soap (Ping timeout: 240 seconds)
19:05:11chris_sno problems so far on the iPod 4G using iFlash Solo. I'll check the unmodded iPod Photo/color tomorrow, but that one wasn't super reliable using Rockbox's USB mode in the first place last time I checked
19:07:49chris_s(that could have also been an issue with MacOS though,the volume would unexpectedly disappear after transferring a limited amount of data iirc...)
19:08:24 Quit chris_s (Client Quit)
19:10:50 Join Soap [0] (~Soap@rockbox/staff/soap)
19:12:21 Join Soap_ [0] (~Soap@rockbox/staff/soap)
19:15:12 Quit Soap (Ping timeout: 240 seconds)
19:21:40 Join Soap [0] (~Soap@rockbox/staff/soap)
19:21:52 Quit Soap_ (Ping timeout: 240 seconds)
19:23:45 Join Soap_ [0] (~Soap@rockbox/staff/soap)
19:26:11 Quit Soap (Ping timeout: 240 seconds)
19:35:01 Join MrZeus_ [0] (~MrZeus@194.37.96.153)
19:35:55 Quit pamaury (Ping timeout: 260 seconds)
19:38:15 Quit MrZeus__ (Ping timeout: 260 seconds)
19:52:55fs-bluebotBuild Server message: New build round started. Revision 79d1b68fe2, 298 builds, 9 clients.
19:53:44 Join amachronic [0] (~amachroni@82.132.187.194)
19:57:29amachronicseems that USB is not working as well as it appeared on the M3K
20:00
20:02:21amachronicI assume the designware USB driver works reliably on the other ports,
20:02:45speachyit would seem so
20:02:53amachronicbut for some reason I'm getting hangs and lockups
20:03:07speachyplugged into a USB 2 or USB3 port?
20:03:12amachronicUSB3
20:03:21amachronichaven't tried USB2 yet
20:03:41speachymight be worth a try, if nothing else it mucks with the timing a bit
20:04:10amachronicIt seems there's multiple symptoms but I'm not sure what the real cause is.
20:04:14speachy(There's still a report of the X3 having some usb hiccups, but it seems to only affect USB2 host contollers)
20:04:26fs-bluebotBuild Server message: Build round completed after 691 seconds.
20:04:29fs-bluebotBuild Server message: Revision 79d1b68fe2 result: All green
20:04:38speachycache management differences between ARM and MIPS?
20:04:55amachronicit seems data corruption isn't the issue.
20:04:57amachronicoddly
20:05:31amachronicI was able to write 20 gigs, and read them back just fine, so I'd figure cache bugs would've manifested
20:05:35speachywell, the x1000 is considerably faster than any of the other dwc2 SoCs..
20:05:38amachronicit only affects the early connection setup.
20:05:51speachycould be a classic race condition
20:06:09amachronicit does seem to occur most frequently when HID + mass storage are enabled together.
20:06:22amachronicusing mass storage only seems to be less problematic
20:07:21 Quit cockroach (Quit: leaving)
20:08:28amachronicit appears that on occasion, Rockbox doesn't respond on the mass storage endpoint
20:08:53amachroniccausing the host to try resetting, and that causes a hard lockup.
20:09:39amachronicpressing HID keys tricks the host into thinking the device is alive,
20:10:21amachronicbut after a while rockbox panics with usb_dw_gonak_effective: failed!
20:16:01speachyI really need to get this port onto my M3K.
20:16:15speachyso I can actually help dig into this stuff
20:16:15amachronicwell, it's at least one command to install now
20:16:53speachyI hope this iflash stuff can finally be put to rest now
20:17:55***Saving seen data "./dancer.seen"
20:25:20 Join cockroach [0] (~blattodea@pdpc/supporter/active/cockroach)
20:40:13 Quit amachronic (Quit: amachronic)
20:50:23speachyI think the underlying issue with the iflash stuff is that we might be entering sleep _too_ aggressively, as in momentary pauses between usb transfers are enough to trigger it.
20:51:08speachyand that's exacerbated by the iflash's chipset/firmware's not-quite-ATA-compliance
21:00
21:01:32fs-bluebotBuild Server message: New build round started. Revision f968d6032a, 298 builds, 9 clients.
21:05:46braewoodsspeachy: does this change things for bootloaders? i suspect not as the bootloaders almost never spindown the disks
21:05:54braewoodsplus BLs typically just rea
21:05:56braewoodsread
21:06:01braewoodsonly do writes if in USB mode
21:07:04 Quit MrZeus_ (Ping timeout: 252 seconds)
21:07:22speachyyeah, probably won't affect the bootloaders at all unless they do writes
21:11:17speachyamachronic: any objection to moving the m3k from 'unusable' to merely 'unstable' ?
21:12:18speachy(it'll start showing up in nightly build list, get voice builds, and so forth)
21:12:51fs-bluebotBuild Server message: Build round completed after 679 seconds.
21:12:54fs-bluebotBuild Server message: Revision f968d6032a result: All green
21:35:37 Quit St3ak (Quit: Free ZNC ~ Powered by LunarBNC: https://LunarBNC.net)
21:36:42 Join St3ak [0] (~st3ak@st3ak3000.powered.by.lunarbnc.net)
21:37:37 Quit St3ak (Client Quit)
21:38:54 Join St3ak [0] (~st3ak@st3ak3000.powered.by.lunarbnc.net)
22:00
22:17:58***Saving seen data "./dancer.seen"
22:21:20 Quit Strife89 (Quit: No Ping reply in 180 seconds.)
22:22:45 Join Strife89 [0] (~quassel@adsl-74-250-152-173.ags.bellsouth.net)
22:27:05 Quit cockroach (Quit: leaving)
22:58:35 Quit Saijin_Naib (Ping timeout: 250 seconds)
23:00
23:06:58 Quit ac_laptop (Ping timeout: 252 seconds)

Previous day | Next day