**** BEGIN LOGGING AT Tue Jul 12 02:59:57 2011 Jul 12 04:43:50 ogra_, you about? Jul 12 04:58:07 about arm assemble constraint, "USAT r0, #7, r5" => should I use "i" for value #7 ? seems not correct.. Jul 12 05:02:48 usat Rd, #sat, Rm , #sat should be 0-31 value for arm, what should be the constraint? "i" not the right Jul 12 06:54:40 morning Jul 12 09:08:36 hello everyone Jul 12 09:42:54 oh my Jul 12 09:42:56 http://www.fedora.org/ Jul 12 09:44:49 hello, i got a problem. i installed ubuntu-arm on a pandaboard using the netinstall package as "ubuntu server". then i installed xinit and startx worked, in VGA resolution with framebuffer, then i installed pvr-omap4 (without rebooting)and startx kept working in VGA resolution. today i rebooted my pandaboard and now i get this http://hpaste.org/48969 Jul 12 09:46:16 looks like you are missing bits on either the kernel or X side Jul 12 09:47:10 well Jul 12 09:47:26 i solved it with touch ~/.Xauthority ( O_O ) Jul 12 09:47:48 oh, so it works fine now ? Jul 12 09:48:01 * ogra_ hasnt used xinit in years Jul 12 09:48:22 right, you are ubuntu user :p Jul 12 09:48:40 ogra_: do you remember my problem with audio input and ubuntu headless? Jul 12 09:48:44 I solved it Jul 12 09:48:53 well, i only work with what we ship :) Jul 12 09:48:56 it works correctly on headless too Jul 12 09:48:56 well. "works fine" but i still get vga resolution Jul 12 09:49:12 and i'm not sure it's using pvr_omap4 instead of framebuffer Jul 12 09:49:17 smells like the driver isnt getting the correct EDID data Jul 12 09:49:29 Spider-Pork, how did you solve it ? Jul 12 09:50:06 well, I installed ubuntu headless Jul 12 09:50:12 then alsa and pulseaudio Jul 12 09:50:35 yeah, these bits are indeed missing in headless :) Jul 12 09:50:37 then applyed the two commands in the bugfix page Jul 12 09:50:53 HiFi and Record Jul 12 09:50:57 right Jul 12 09:51:29 then i booted into ubuntu netbook and stored in a file the amixer configurations with alsactl store Jul 12 09:51:47 erm Jul 12 09:52:01 then rebooted with ubuntu headless, copyed in the config file and restored it with alsactl resotre Jul 12 09:52:04 *restore Jul 12 09:52:05 that happens automatically and you wouldnt need to boot into netbook Jul 12 09:52:32 a simple alsactl store would have sufficed Jul 12 09:52:48 store what? Jul 12 09:52:48 but alsa calls that on every shutdown anyway Jul 12 09:53:09 on my headless there wasn't alsa configurations Jul 12 09:53:15 so theoretically a reboot should have done the same Jul 12 09:53:32 if you have alsa-utils installed at least Jul 12 09:53:50 i used the netbook alsa configurations (that is correctly working with audio line-in) Jul 12 09:54:03 no wait Jul 12 09:54:04 /var/lib/alsa/asound.state holds the mixer states between bootzs Jul 12 09:54:35 it is re-written on every shutdown with the currently applied mixer settings Jul 12 09:54:36 my ubuntu netbook is corretly working, my ubuntu headless wasn't Jul 12 09:54:49 right, then there was something still missing Jul 12 09:54:53 yep Jul 12 09:54:58 but you shouldnt need to copy data Jul 12 09:55:02 ah Jul 12 09:55:22 removing /var/lib/alsa/asound.state and calling alsactl store should get you what you need Jul 12 09:55:35 ok i'll try it :) Jul 12 09:55:38 i think it is mentioned in the bug by tobin (GrueMaster) Jul 12 09:55:59 i solved it with the power of the noobs Jul 12 09:56:09 well, it works for you :) Jul 12 09:56:24 use the power, noob Jul 12 09:56:24 would just be good to find a way that doesnt need a second machine ;) Jul 12 09:56:30 for people having the same issue Jul 12 09:56:34 yep Jul 12 09:57:03 i will test it Jul 12 09:57:19 if you found a way, please comment on the bug :) Jul 12 09:57:29 yep Jul 12 09:57:42 http://comments.gmane.org/gmane.comp.embedded.pandaboard/754 i'm looking at this guide, i installed pvr_omap4 and other pvr packages but i don't have neither omap_gpu module nor pvrsrvinit am i missing something? Jul 12 09:58:02 the community helped me, so i will do the same :) Jul 12 09:58:28 LPhas, did it compile the modules with dkms on next reboot after installing the package ? Jul 12 09:59:02 you should just install the graphics metapackage Jul 12 09:59:49 ubuntu-omap4-extras-graphics Jul 12 10:00:11 from https://launchpad.net/~tiomap-dev/+archive/release indeed Jul 12 10:00:15 ogra_, i installed it, but i did'nt manually compile nothing Jul 12 10:00:32 not you, the system should autocompile Jul 12 10:00:38 well, the package actually :) Jul 12 10:00:52 well it seems to me that it compiled stuff Jul 12 10:00:59 i recall it used kernel headers Jul 12 10:01:09 should i remove/reinstall it? Jul 12 10:01:27 ARGH Jul 12 10:01:33 who pointed you to that guide Jul 12 10:01:52 thats super broken Jul 12 10:02:13 just adding the ppa and installing the meatapackage will suffice Jul 12 10:02:53 if you did any of the manual steps in there its doubtful that your system still functions the way it was intended (with all the automatisms) Jul 12 10:03:22 google Jul 12 10:03:32 and no, i didn't do any manual step Jul 12 10:04:41 well, wait for robher clark or rsalveti to get up ... they are both pvr specialists Jul 12 10:04:47 err Jul 12 10:04:58 robher, sorry, that wasnt for you Jul 12 10:05:08 * ogra_ curses autocompletion Jul 12 10:05:13 i meant robclark Jul 12 10:05:32 ok thx Jul 12 10:05:43 you were also using headless, right Jul 12 10:05:45 ? Jul 12 10:06:06 headless==? Jul 12 10:06:11 server Jul 12 10:06:15 mh yeah Jul 12 10:06:26 oh, wait, and you are using oneiric ? Jul 12 10:06:30 ubuntu-desktop failed to install due to dependencies problems Jul 12 10:06:51 mmh where oneiric was like "development version" something like that? Jul 12 10:06:57 yes Jul 12 10:07:00 yep Jul 12 10:07:04 lsb_release -a Jul 12 10:07:08 that should tell you Jul 12 10:07:30 i dont think anyone toucvhed pvr for oneiric yet Jul 12 10:07:44 yep, oneiric Jul 12 10:07:55 so its a matter of luck if it works at all Jul 12 10:07:58 but hey, i would be more than happy to use a more stable version Jul 12 10:08:12 natty should just work Jul 12 10:08:16 mh Jul 12 10:08:29 but i need to install it from netinstall image Jul 12 10:08:34 cause i don't have a 4gb sdcard Jul 12 10:08:59 netinstall is still pretty young ... Jul 12 10:09:22 you could use headless and then install the netbook task Jul 12 10:10:33 uhm Jul 12 10:11:00 headless will fit on the 2gb sdcard? and the i'd have to move / on the usb stick Jul 12 10:11:02 manually Jul 12 10:11:23 headless should fit on 1GB Jul 12 10:11:49 at least if you dont use it much beyond doing the basic config before copying to USB Jul 12 10:13:25 heh, let's do some dd work to save the current system and have a try Jul 12 10:42:35 ogra_, I have a few questions for you when you have a moment Jul 12 12:07:16 Hi. Has anybody else experienced problems with the PowerVR SGX GLES2 driver on Natty lately? Jul 12 12:09:02 what kind of issues? panda or beagle? Jul 12 12:10:20 Pandaboard. It seems that the depth buffer is screwed up: http://noend-rpg.de/users/alemariusnexus/error.png vs. http://noend-rpg.de/users/alemariusnexus/correct.png Jul 12 12:11:49 I don't have problems on my desktop computer, and PVRFrame (the PowerVR GLES "emulator") doesn't show this error, too Jul 12 12:24:02 It might be a problem in my application. But since it appears on my Pandaboard only and since I can't imagine what kind of error in my application could cause this, I suspect it might be a driver problem (also, I had another problem which is very likely to be a driver bug in a Qt program) Jul 12 12:32:09 using the installer in headless img for omap4/pandaboard, everyhing is messy, is there a dimension of the serial console i've to use to see it correctly? Jul 12 12:34:02 like this http://imageshack.us/photo/my-images/197/screenshot3gv.png/ Jul 12 12:34:04 LPhas - a lot of sort of question mark characters? Jul 12 12:36:03 yeah, there are a lot of question mark charters, a lot of misplaced charters and generally the layout is broken Jul 12 12:41:12 meh, i'd really like to have a way to install a stable version with the netinstaller Jul 12 12:41:58 i mean, shouldn't be only matter of writing the proper apt configuration file? Jul 12 13:11:55 is anyone else seeing init @100% after heavy IO loads? 2.6.38-1309 #14 on panda? Jul 12 13:12:46 PS. running threaded IO Jul 12 13:32:05 Someone used ubuntu-arm on Mini6410? Jul 12 13:54:03 fredim, The processor on that board can't run Ubuntu. Debian should run nicely on that. Jul 12 13:55:28 Fixed OpenMPI upstream. 1.5.4 release imminent Jul 12 13:55:42 Also getting a source deb pulled together in debian... Jul 12 13:59:30 1.5.4! My, we're a bit out of date for that. Jul 12 14:00:21 persia: I haven't forgotten about that package btw, just been a bit busy Jul 12 14:01:48 lilstevie, Totally understood. Jul 12 14:04:01 lilstevie, Based on https://wiki.ubuntu.com/OneiricReleaseSchedule, we have until 11th August, although I'd like to try for a bit sooner, just to have time to be sure we make that deadline. Jul 12 14:04:56 Also, I'll not be around so much 22 July to 1 August (although someone else may also be able to sponsor), so my personal timing goals aremore complicated. Jul 12 14:05:38 persia : The reason we're so far behind, is that until 1.5.2, there was no ARM support Jul 12 14:05:54 and the ARM tarball was broken (missing files) until just now, since the ARM maintainer was building from SVN Jul 12 14:06:12 so now it's getting fixed, and it should all work nicely for 1.5.4. I'm testing daily Jul 12 14:06:43 persia: well if we need to get something done https://github.com/lilstevie/samsung-kernel-tab/commits/GT-P1000-gingerbread/ Jul 12 14:06:55 And the other architectures are all sufficiently well supported with 1.4.3? Jul 12 14:07:23 persia: but that does have big bugs Jul 12 14:07:23 Sort of Jul 12 14:07:31 it's considered the "stable" release Jul 12 14:07:32 I haven't sync'd for a bit Jul 12 14:07:54 but it's missing features that are becoming important to users of OpenMPI ... 1.5.3 is mainline in Sid right now Jul 12 14:08:14 so it's probably time to move Ubuntu server up to 1.5.3 as well, and make sure to provide an alternatives config Jul 12 14:08:24 lilstevie, If you haven't found time by next week, I'll try to find time for first packaging and upload of that, directly from git. If you find time, and we can publish a *working* kernel where you understand the packaging, that's considerably preferable. Jul 12 14:08:27 well, 1.5.4 rather Jul 12 14:08:45 persia: :) Jul 12 14:08:47 Martyn, Hrm? My rmadison tells me sid has 1.4.3-2.1 Jul 12 14:09:27 Hmmm .. I've been working with the guys to make the .dsc ... Jul 12 14:09:38 http://packages.qa.debian.org/o/openmpi.html Jul 12 14:09:38 maybe they are waiting for 1.5.4? Jul 12 14:09:49 yes, I see Jul 12 14:12:25 They're probably waiting, indeed. Jul 12 14:22:17 Mrrf. Jul 12 14:22:31 1.4.x series, though 'stable' is also ludicrously behind in features Jul 12 14:22:39 and 1.6.x won't be out for a long while.. Jul 12 14:22:49 this is the old odd/even kernel problem... Jul 12 14:23:34 Well, it's a stable/unstable thing. Jul 12 14:23:53 Lots of folks do that in ways that don't appear so obvious in version numbers. Jul 12 14:24:33 But there's no reason we can't put it in oneiric, if we had some volunteer who would continue to make sure it works for 18 months (hint) Jul 12 14:26:30 *chuckle* Yes, I kno Jul 12 14:26:38 I just need to finish packaging it up Jul 12 14:28:14 Sylvestre is also an Ubuntu dev, so you probably want to get it into Experimental, and sync from there if he agrees, just to avoid potential checksum mismatches later. Jul 12 16:01:49 any idea on why i get this http://imageshack.us/photo/my-images/36/screenshot4ag.png/ in installing ubuntu headless? Jul 12 16:02:20 could be some minicom settings? Jul 12 16:03:10 Not off hand. The settings look ok, from what I see on the bottom of the screen. Jul 12 16:04:02 Can you run screen on your system? If so, try this: screen 115200 Jul 12 16:04:11 I.e. screen /dev/ttyUSB0 115200 Jul 12 16:05:40 LPhas: Minicom is a lousy terminal emulator, nothing new there. As GrueMaster says, try screen. Jul 12 16:06:16 well, with screen i get only gibberish Jul 12 16:07:07 oh, no, ok it seems working right now Jul 12 16:07:42 lol what is that? wait i'll make a screenshot Jul 12 16:08:28 http://imageshack.us/photo/my-images/15/screenshot5hc.png/ this is with screen Jul 12 16:10:11 I have never seen that before. You have something strange going on, possibly with your serial port. Jul 12 16:12:03 that's quite strange because with minicom that part of the boot is fine Jul 12 16:12:20 and also yesterday the ubuntu netinstall was fine Jul 12 16:13:26 and yet, you were right Jul 12 16:13:44 i changed usb port for my serial/usb adapter and now is fine with screen O_O Jul 12 16:14:14 computer science is weird Jul 12 16:19:14 Could be a data flow issue. I have seen weird things happen with one hub vs a different one, both are USB 2.0 Jul 12 16:19:57 GrueMaster, true Jul 12 16:20:47 still, it's a pity that there isn't a netinstall for stable Jul 12 16:23:30 There may be a way to do it. I am going to test a method today (once I get all these new usb enclosures filled and online). Jul 12 16:58:28 phas@panda:~$ sudo Jul 12 16:58:28 sudo: must be setuid root Jul 12 16:58:32 that's strange Jul 12 16:59:32 well ok i probably lost some permission while moving on usb duh Jul 12 17:02:45 How did you do the move from SD to USB? Jul 12 17:04:00 sudo mv Jul 12 17:04:05 well Jul 12 17:04:07 sudo cp -r Jul 12 17:04:24 but i should have fixed Jul 12 17:04:38 i just activated root in /etc/shadow and fixed permissions on /usr/bin/sudo Jul 12 17:10:05 I usually do: sudo su;tar -cf - . | (cd ; tar -xvf -) Jul 12 17:10:43 It also has the benefit of one thread doing reads and the other doing writes. Jul 12 17:10:50 makes sense Jul 12 17:10:57 next time i'll try it Jul 12 17:11:13 or if permissions problems starts to get an issue Jul 12 17:14:08 actually the only problems seems to be that i get Jul 12 17:14:09 Processing triggers for man-db ... Jul 12 17:14:09 fopen: Permission denied Jul 12 17:14:15 on aptitude ugrade Jul 12 17:15:57 ouch, is this normal? http://hpaste.org/48974 Jul 12 17:16:36 The find error is normal. Jul 12 17:16:39 ok Jul 12 17:16:55 Not that it is right, just happens all the time. Jul 12 17:17:02 eheh Jul 12 17:17:47 Well, if you were expecting perfection, you should have used...hrm. Nevermind. Jul 12 17:17:50 :P Jul 12 17:18:01 i used everything else that run on pandaboard Jul 12 17:18:24 so what i expect here from ubuntu-arm is a miracle :p Jul 12 17:19:59 I'm trying, but am still stuck on that whole wine from water issue. Jul 12 17:20:12 Priorities. Jul 12 17:20:36 there are probably powders for that Jul 12 17:22:37 Just don't expect good things to happen if you spray your system with Miracle Grow. Jul 12 17:23:03 (sadly I have heard of people doing much stupider things). Jul 12 17:23:18 what's Miracle Grow? Jul 12 17:23:55 Some plant food sold in stores in the US to make plants grow faster. Jul 12 17:25:27 oh i see Jul 12 17:31:14 aaand ok, we are at the point where startx starts in framebuffer with VGA resolution Jul 12 17:31:46 Hrm. Can you install read-edid? Jul 12 17:32:06 read-edid? Jul 12 17:32:13 but i still haven't install pvr-omap4 Jul 12 17:32:17 i'm about to try Jul 12 17:32:25 Don't need it. Jul 12 17:32:29 uhm Jul 12 17:32:39 This will see what the edid data is coming from the monitor. Jul 12 17:33:02 where "monitor" is a television Jul 12 17:33:32 (starting to run out of monitors here) Jul 12 17:35:02 heh. I know the feeling. That's why I have an HDMI switch. Jul 12 17:36:15 heh, my monitor has some hdmi ports Jul 12 17:36:22 but it's painful to switch Jul 12 17:36:49 that's weird Jul 12 17:36:52 i installed read-edid Jul 12 17:37:00 and i have parse-edid as an executable Jul 12 17:37:04 but i can't find get-edid Jul 12 17:38:28 hmmm. Not finding edid on my system. Should be at /sys/devices/omapdss/display0/edid. Use "cat /sys/devices/omapdss/display0/edid | parse-edid" Jul 12 17:39:16 mmh i can't file an edid file in /sys/devices/omapdss/display0/ Jul 12 17:39:50 Hrm. Wonder if it got left out of the kernel. Shouldn't have been. Jul 12 17:40:41 mmh wait a second here Jul 12 17:41:00 # Booting kernel from Legacy Image at 80000000 .. Jul 12 17:41:12 is it normal that he says "Legacy Image"? Jul 12 17:41:23 yes. Jul 12 17:41:26 ik Jul 12 17:41:28 ok Jul 12 17:41:40 well, let's try another monitor/hdmi cable Jul 12 17:45:08 lol Jul 12 17:45:12 it was the usb cable Jul 12 17:45:23 usb cable? Jul 12 17:45:25 hdmi cable Jul 12 17:45:29 stupid hdmi cable Jul 12 17:45:35 oh. ok. Jul 12 17:45:41 that's weird too Jul 12 17:45:52 a hdmi cable that let you use only vga resolution Jul 12 17:46:44 now should i try to install pvr_omap4 from the ti repository right? Jul 12 17:46:52 It was probably not getting the edid signal. Jul 12 17:46:56 yep Jul 12 17:47:00 Go for it. Jul 12 17:47:08 should i use TI OMAP release PPA or TI OMAP trunk PPA Jul 12 17:48:09 Not sure. Probably trunk, but I don;'t know which has the latest stuff. Jul 12 17:55:01 "trunk" probably has newer stuff, but "release" probably has stuff TI thinks is release-worthy. Jul 12 17:55:25 For oneiric, "trunk" is probably best: when the underlying environment isn't stable, expecting stable add-ons is dreaming. Jul 12 18:02:42 GrueMaster, i think i've got it, gg Jul 12 18:03:00 cool Jul 12 18:03:10 just this two errors bothers me Jul 12 18:03:17 (EE) AIGLX: reverting to software rendering Jul 12 18:03:26 (EE) Keyboard: Couldn't open mtdev device Jul 12 18:03:30 if they do mean something Jul 12 18:04:08 Unless it isn't working, I would ignore those. I think they are for multi-touch. Jul 12 18:04:17 Not sure. Jul 12 18:07:30 now the last omap4-related matter is to get omap-gstreamer work Jul 12 18:07:43 is it in the ti ppa that you know? Jul 12 18:10:08 There have been some gstreamers in the TI PPA. I'm not sure if any of them are compatible with oneiric. Jul 12 18:10:20 i'm not on oneiric Jul 12 18:10:23 i switchet to nanny Jul 12 18:11:07 Oh, heh. In that case, there's likely to be something. Jul 12 18:11:26 basically i look for the hardware decoding support for h264 Jul 12 18:11:31 If there isn't, the uploader of the maverick version is very likely to be lurking around in this channel, and may respond to a ping. Jul 12 18:11:48 i saw that there's a gstreamer_h264 Jul 12 18:12:06 and the usual gstreamer h264 plugin is in good Jul 12 18:12:06 This is probably what you want :) Jul 12 18:12:08 so it may be it Jul 12 18:54:22 thx guys, i think that it's tyme to bothers #gstreamers guys Jul 12 22:28:42 so I'm seeing init consume 100% cpu on panda while/after running IO and threaded IO tests (PTS) Jul 12 22:29:16 Installing phoronix now on 2 systems (third is still imaging from netinst). Jul 12 22:30:21 mahmoh, Is there some spinning job spewing to /var/log/syslog? When you strace init, what is it doing? Jul 12 22:31:44 haven't checked that but I'll add it to the list Jul 12 22:35:53 persia: i've run strace on it and it is looping, waiting for something Jul 12 22:36:43 persia: if you watch its process state, you can actually see it go to sleep between checks Jul 12 22:39:45 Well, this can certainly pose a problem. I have two panda systems with identical mac addresses. Sigh. Jul 12 22:40:07 And as a result, my dhcp server has assigned them the same IP address. Jul 12 22:40:43 set the ethaddr in uEnv.txt or boot.scr Jul 12 22:40:53 yep. Jul 12 22:41:28 actually, once you boot the linux kernel you should have unique macs Jul 12 22:41:42 Nope. Jul 12 22:41:56 Both systems have been powercycled since netinstall finished. Jul 12 22:42:09 unless both your boards came from the same die? Jul 12 22:42:16 u-boot is hardwired for the same mac. Jul 12 22:42:47 Definitely not same die. One is ea1, the other is a1. Jul 12 22:42:50 the u-boot mac is 00:02:03... yours should not be that once booted Jul 12 22:43:34 Hrm. Then something isn't right in my world. Jul 12 22:44:22 I can log into each independently via serial console, but that has limits (not to mention that my serial console is being upgraded). Jul 12 22:44:41 But they both show the same mac/ip. Jul 12 22:44:57 what's the mac? Jul 12 22:45:29 2e:40:70:f0:12:06 Jul 12 22:45:33 cmagina, Could you paste one or two repetitions of the strace? Jul 12 22:46:32 so mine both start with 2E:40:.... but diverge from there, odd Jul 12 22:46:43 persia: i wish i could, but apparently i failed to record it. on a good note, mahmoh has been able to preproduce this, so we can get one as soon as he hits it again Jul 12 22:47:39 cmagina: persia: I'll reproduce. If it appears to be a problem, where should the bug land? kernel, init, server-arm? Jul 12 22:48:15 I can't say without a bit more information. I'd guess upstart, but that's a completely uninformed guess at this point. Jul 12 22:49:03 The 'init' executable is provided by upstart, which ought only do job processing. It may be that some job is spinning, in which case the job provider is at fault. It may be that upstart is spinning internally, in which caes it's a bug in upstart. Jul 12 22:49:31 It may be that some kernel interface changed and something that is supposed to be a wait in userspace has converted to a poll (unlikely), which would make it a kernel issue. Jul 12 22:49:59 Most likely is that something else is spinning, and making N requests of init, which upstart is trying to service, making it look like upstart is at fault. Jul 12 22:51:42 and why do I see a bunch of "omapdss DISPC error: timeout waiting for EVSYNC", is that just an artifact of the panda-usb connectivity? Jul 12 22:53:01 that is the fb driver whining. Safe to ignore. Jul 12 22:53:11 looking through the logs i gathered from mahmoh recent repro, i noticed that the pts tests were hung Jul 12 22:53:14 It's bug #781318 Jul 12 22:53:15 Launchpad bug 781318 in linux-ti-omap4 "omapdss DISPC error: timeout waiting for EVSYNC" [Medium,Confirmed] https://launchpad.net/bugs/781318 Jul 12 22:53:34 [31200.956817] INFO: task tiotest:1615 blocked for more than 120 seconds. Jul 12 22:53:42 Or bug 758486 Jul 12 22:53:42 Launchpad bug 758486 in linux-ti-omap4 "omapdss DISPC error on Panda platform" [Medium,New] https://launchpad.net/bugs/758486 Jul 12 22:54:00 There's a chance that this could be driving odd stuff in udev, which could in turn be spinning upstart, but this is very unlikely, as more folk would see the problem. Jul 12 22:54:27 yeah but it's annoying on the server iamge Jul 12 22:54:29 iamge Jul 12 22:54:31 image Jul 12 22:54:46 GrueMaster, I thought 781318 was it being noisy as headless, and 758486 was about failure to resume from suspend cleanly. Jul 12 22:55:08 Read the bug reports. Jul 12 22:55:16 I did. Jul 12 22:55:53 This is a simple matter of someone filing a duplicate bug. Jul 12 22:56:10 Are they duplicates? The messages are different. Jul 12 22:56:20 rsalveti: GrueMaster: I've also confirmed I can load and run from a USB memstick (fatload usb 0 ..., root=/dev/sda2) Jul 12 22:56:58 rsalveti: ^ with your patch u-boot, I have to validate the package version [soon[ Jul 12 22:57:00 ] Jul 12 22:58:48 persia: Only slightly different messages, but they are caused by the same failing kernel FB code. I already did the leg work with rsalveti last cycle. Jul 12 22:58:53 rsalveti: GrueMaster: and root=USB-SATA (root=/dev/sdb2) booting from MMC/SD Jul 12 22:59:10 GrueMaster, Ah, that makes sense. Jul 12 23:00:10 It was supposed to be fixed with the edid update code that apparently never made it in. Jul 12 23:00:41 cmagina, That's probably just an artifact of the CPU spinning on something else. Unless a thread is asserting realtime, it shouldn't care if it doesn't get execution slices very often (although it's nice to warn the user in case it's all batch processing and the user needs to get more cycles or reduce the load) Jul 12 23:01:03 Ah, that's extra annoying. Jul 12 23:01:17 (and cooloney isn't idling yet :( ) Jul 12 23:01:46 persia: agreed, i'm still looking over the logs i did collect, hopefully there will be something of interest Jul 12 23:01:54 I thought paulo was the arm kernel guru now. Jul 12 23:02:16 (ppisati I think is his nick). Jul 12 23:02:21 Dunno. cooloney is still assigned both those bugs. If they ought be reassigned, I leave that between them. Jul 12 23:21:08 mahmoh: So in your USB testing above, was that with the kernel & initrd on USB or are they still on SD? Jul 12 23:21:53 both Jul 12 23:22:52 I loaded from MMC then booted to USB first, then tried loading from USB then booting to USB (sticks) Jul 12 23:22:56 Does that mean that we've reached a point that if we put the right u-boot in flash/MMC then we can boot from USB cleanly? Jul 12 23:22:58 Sweet. So in theory, we can just have x-loader (MLO) and u-boot.bin on SD, or even no SD and u-boot.bin pushed via usbboot. Jul 12 23:23:08 yes Jul 12 23:23:20 right Jul 12 23:23:25 * persia celebrates reaching parity with other architectures Jul 12 23:24:35 mahmoh, Does it still need to be FAT? Can we load off e.g. ext4? Jul 12 23:25:02 I wonder if elmo can utilize this in the buildd cluster? Jul 12 23:25:20 ooh, good question, I think it should be able to load off of ext2, I'll try it after my tests complete Jul 12 23:25:50 For extra points, load directly out of /boot on USB-SATA :) Jul 12 23:26:13 persia: that won't work :( drivers on;y support sticks Jul 12 23:26:18 only Jul 12 23:26:53 or so I'm led to believe Jul 12 23:27:21 mahmoh, Considering there's no way in the USB spec to distinguish sticks from other things that provide USB class-compliant storage facilities, I suspect you've been misinformed. Jul 12 23:27:44 (or perhaps I have) Jul 12 23:28:21 persia: tell that to the driver considering it only detects one USB storage device out of two (usb-stick & usb-sata) - guess which one ;) Jul 12 23:28:31 That said, the drivers might not be class-compliant, or be horridly buggy for certain sorts of storage characteristics. Jul 12 23:29:09 you always see USB-FDD/USB-HDD etc. in bios so there has to be some diff., maybe it's based on size? Jul 12 23:29:26 USB-FDD is something else. Jul 12 23:29:48 Rather than representing a mass-storage device, it represents a removable storage device. Jul 12 23:29:50 well that's supported though in u-boot, haven't tried though Jul 12 23:30:10 Do you have one? I have some dusty ones somewhere if it especially needs testing. Jul 12 23:30:23 Not sure if I have any floppies for it, mind you. Jul 12 23:30:35 u-boot "should" support usb-sata, but may not be enabled for panda. Jul 12 23:30:47 There's no "enable"ing to do for that. Jul 12 23:31:00 I do have one but do we really want to test it, except for curiosity's sake? Jul 12 23:31:04 USB MSC is USB MSC, and *doesn't* indicate whether it's one or another implementation. Jul 12 23:31:08 config flags? Jul 12 23:31:25 mahmoh, Up to you. It's a time/interest thing. Jul 12 23:31:47 when I'm super bored ... == never Jul 12 23:32:20 GrueMaster, Really? Please don't tell me that someone created configuration flags to turn on/off buggy behaviour. Jul 12 23:32:48 persia: no, that's never been done before, ever. Jul 12 23:32:54 well, considering usb support existis in u-boot since jaunty, but only recently for omap... Jul 12 23:32:59 * persia reviews version 1.3 of the spec, just to be sure Jul 12 23:33:26 * mahmoh has some land in FL to sell to persia Jul 12 23:33:32 GrueMaster, Oh, perhaps. But there really shouldn't be any way to turn on or off USB<->SATA gateways. Jul 12 23:34:05 * persia offers discount shares on a new alligator belt factory Jul 12 23:34:30 it's probably just an addressing space issue, can't support more than X gig so USB-SATA isn't an option? Jul 12 23:34:51 * GrueMaster offers to trade Oregon Coast property w/ ocean views for property in florida). Jul 12 23:34:51 Or rather, if the device is more than some size, it fails to load, and is skipped. Jul 12 23:35:09 That's buggy, but not configurable (and would work fine with a 2G USB PATA device) Jul 12 23:35:32 mahmoh: What filesystems does u-boot support? Could be as simple as no ext4 support. Jul 12 23:36:24 mahmoh, Could you verify the Subclass code for your two USB devices (working, not working)? Jul 12 23:36:42 (lsusb -v) Jul 12 23:37:31 I'm running ext3, so maybe Jul 12 23:37:53 GrueMaster: actually no, there's a fat fs and it doesn't detect the device Jul 12 23:38:48 ext3 is ext2+journal. Fully backwards compatible. Jul 12 23:38:51 bDeviceSubClass 0 for both Jul 12 23:39:32 Right. So if both are MSC, and both are bDeviceSubclass 0, then they are the same. Jul 12 23:39:44 Could be a simple "find 0-1 devices and return) thing. Jul 12 23:39:54 Is there anything other than vendor/device identifiers (name, number, etc.) that differs in the descriptors? Jul 12 23:40:00 iConfiguration 4 USB Mass Storage vs. iConfiguration 0 Jul 12 23:40:12 Which is which? Jul 12 23:40:26 bmAttributes 0xc0 vs bmAttributes 0x80 Jul 12 23:40:42 flash/works vs sata/not-recognized Jul 12 23:40:58 Where is that? Jul 12 23:41:37 iInterface 6 MSC Bulk-Only Transfer vs iInterface 0 Jul 12 23:41:49 in lsusb -v Jul 12 23:42:04 So, which is which? Are these all in the same order? Jul 12 23:42:08 why don't I just pastebin it all for you? Jul 12 23:42:15 May as well. Jul 12 23:42:36 One of those devices looks misimplemented, but I may be mistaken. Jul 12 23:42:58 http://pastebin.ubuntu.com/642971/ Jul 12 23:43:13 it's possible but they both work in linux Jul 12 23:43:32 the usb-sata was the cheapest I could find ;) Jul 12 23:44:26 Not surprising. Jul 12 23:44:39 It only implements a subset of MSC, specifically BBB. Jul 12 23:45:05 Side effect is that it would have poor performance with multiple parallel IO streams. Depending on the backing store, and caching, etc. this may not matter. Jul 12 23:45:27 Your USB stick supports CBI, which is a richer instruction set. Jul 12 23:46:03 I'm perfectly willing to believe that u-boot only supports CBI devices, but I'll assert that this is unrelated to the backing store. Jul 12 23:48:47 Your USB stick *should* have iConfiguration of 4 (USB Mass Storage), but in practice it doesn't matter, as it's defined at the interface level. Jul 12 23:52:16 Oh, heh, I am out of date. 1.3 says "Usage of CBI for any new design is discouraged.". Jul 12 23:52:45 This is probably worth a bug, as without BBB support, even USB "sticks" of the future may not work. Jul 12 23:54:39 persia: who's going to write it up? I'm sure everyone would love that support in u-boot, I know I would! Then you could have a real boot partition that really boots and updates easily! Jul 12 23:55:18 mahmoh, As you have hardware known to be recognised *AND* not-recognised, I suspect you're a natural candidate. Jul 12 23:55:31 mahmoh: cool, the u-boot binary from the package should work the same way as my binary Jul 12 23:55:31 Pass me the bug number and I'm happy to comment based on what I know of USB specs Jul 12 23:55:40 the patch is integrated and package uploaded Jul 12 23:55:59 rsalveti: how difficult would it be to get usb-sata support in u-boot? Jul 12 23:57:11 mahmoh: didn't work for you? Jul 12 23:57:23 I believe it should behave the same way as when using with a pen drive Jul 12 23:57:29 at least the usb protocol should be the same Jul 12 23:57:41 rsalveti: usb-flash-drive is recognized but usb-sata-drive is not Jul 12 23:58:03 I thought I read that it wasn't supported, only sticks Jul 12 23:58:11 oh, ok Jul 12 23:58:48 rsalveti: 31 Currently supported are USB Hubs, USB Keyboards, USB Floppys, USB Jul 12 23:58:48 32 flash sticks and USB network adaptors. http://git.linaro.org/gitweb?p=boot/u-boot-linaro-stable.git;a=blob;f=doc/README.usb;h=a8a4058de122d6c068d9cb91b218bdc89919c0b8;hb=HEAD Jul 13 00:00:03 mahmoh: I believe my sata disk did work Jul 13 00:00:10 using with usb Jul 13 00:00:18 let me check Jul 13 00:00:30 What! There's a static list in udev detailing which devices to support? Jul 13 00:01:35 rsalveti: if it does work then pls tell me which controller/oem so I can get a couple ;) Jul 13 00:02:08 mahmoh, Try inserting the vendor/device pair for your device into the u-boot code. That might be enough to make it work. Jul 13 00:02:37 that means I'll have to build it ;) Jul 13 00:03:02 mahmoh: run lsusb at ubuntu and paste me the output of your usb disk Jul 13 00:03:26 rsalveti, http://pastebin.ubuntu.com/642971/ Jul 13 00:03:36 mahmoh, Really, compilation isn't scary :) Jul 13 00:03:54 There are nice scripts that do everything for you. You only have to type `debuild -b` Jul 13 00:05:29 persia: mahmoh: http://paste.ubuntu.com/642981/ Jul 13 00:05:39 using a usb-sata disk with a powered usb hub Jul 13 00:05:45 persia: I know, I'm not scared, just don't want to get distracted Jul 13 00:06:32 ah, 'cause hubs are supported! that's cheating Jul 13 00:06:50 Ok, I just got to a point where I could reboot with usb-sata attached, and u-boot detects it. Jul 13 00:06:55 u-boot can't tell if the hubs are in the SoC, on the board, or external. Jul 13 00:07:05 you shouldn't connect a usb-sata disk directly, board doesn't have all the power to power it Jul 13 00:07:06 GrueMaster: you're using a hub? Jul 13 00:07:13 GrueMaster, Excellent. thanks for confirming this is a problem with mahmoh's device. Jul 13 00:07:14 no Jul 13 00:07:23 it's working on my board Jul 13 00:07:51 GrueMaster: what's the model/make? Jul 13 00:07:55 but don't know if u-boot is have the support to give more power to the device Jul 13 00:08:01 - Interfaces: 1 Self Powered 100mA Jul 13 00:08:07 ouch Jul 13 00:08:08 directly it'd consume 500mA Jul 13 00:08:17 let me try without a hub Jul 13 00:08:21 ok, powered hub it is Jul 13 00:08:23 Hornettek Rhino. Jul 13 00:08:57 Using external power supply (panda doesn't have enough power to support otherwise). Jul 13 00:08:58 gazoontite Jul 13 00:09:28 mahmoh: yeah, that's the problem Jul 13 00:09:33 ok Jul 13 00:09:47 http://www.hornettek.com/pcaccessory/index.php/25q-hdd-enclosure/usb-20/rhino Jul 13 00:09:52 I don't see anything in the u-boot USB code that would indicate it does power level negotiation, but I'm not that familiar with the codebase. Jul 13 00:09:52 rsalveti: but wait, it works under linux though? Jul 13 00:09:55 probably need to improve the usb protocol to do allow devices to request more power than the usual Jul 13 00:10:02 I bought 4 at Fry's. Jul 13 00:10:08 mahmoh: yeah, but because on linux the kernel allows it to take 500mA Jul 13 00:10:16 ok Jul 13 00:10:36 looks bulletproof Jul 13 00:12:21 so, it works, but you need to power it with a powered usb Jul 13 00:12:25 That would be the Defender model. http://www.youtube.com/watch?v=uOCVEVNU-PA Jul 13 00:12:41 It has a 5vdc plug. Jul 13 00:12:58 persia: I don't think it does Jul 13 00:13:28 rsalveti, http://www.usb.org/developers/devclass_docs/batt_charging_1_1.zip has the info on the new shiny for granting up to 1.8A Jul 13 00:14:14 persia: USB2.0 or 3.0? Makes a difference. Jul 13 00:14:59 GrueMaster, Battery Charging was introduced as part of 2.0, and I believe it was grandfathered into 3.0, although I'll admit that I haven't been paying as much attention to USB as I did 5 years ago. Jul 13 00:15:57 Reason I say that is that the 2.0 plug can only handle x amps electrically, and that may be pushing the envelope. Jul 13 00:16:18 There's been something like 5 revisions of the plugs. Jul 13 00:16:50 But if the host isn't implemented with a plug that can handle that amperage, it isn't likely to be able to negotiate the higher amperage. Jul 13 00:19:22 Mind you, some idiot might have wired up a port to appear to handle charging without sufficient support, but the legion of iPod owners complaining about burned devices would likely be a sufficient source of complaint that we oughtn't care. Jul 13 00:19:58 Section 3.5 says the standard A connector is capable of 1500mA. Jul 13 00:20:21 (not to pick on the iPod specifically, but most of the USB power hardware I see in shops that supports higher amperages is labeled "supports iPod", rather than "supports Battery Charging Detection") Jul 13 00:20:31 A charger can push more if it has a port that can handle the load. Jul 13 00:20:42 Right. Jul 13 00:20:53 But that ought be limited at a level u-boot can't fiddle. Jul 13 00:21:10 My nook color pulls 1500 if it has the custom microusb cable. Jul 13 00:21:20 u-boot would fiddle the device registers to increase the level, but isn't likely to be able to reach higher than the device supports. Jul 13 00:21:41 At any rate, I don't know what current the panda can push out the usb ports. Jul 13 00:22:17 Supposedly 2A of the input power is expected to be used by the USB ports. Mind you, that's hearsay: I've seen no documentation about this. Jul 13 00:22:37 But it doesn't appear to be enough to handle these drives even with a USB Y cable. The drives are 160GB rated at 0.56A Jul 13 00:22:40 I suspect that the ports are likely rated at 1.5A each. Jul 13 00:23:03 Remember, if the negotiation doesn't happen properly, even a Y cable will only provide 200mA Jul 13 00:23:41 Ah. So definitely a u-boot issue. Jul 13 00:24:07 (kernel as well). Jul 13 00:24:12 Looks that way. Looks like u-boot doesn't support current negotiation, so if the device needs more than 100mA, then it just won't work. Jul 13 00:24:42 Explains my earlier headache. Jul 13 00:24:50 linux does have some current negotiation, at least up to the 500mA level. That said, it may be that the specific USB driver used here isn't implementing it correctly, or something. Jul 13 00:25:18 check what is behind the ports Jul 13 00:25:40 I have no idea if linux can support battery charging: anecdotally folks don't seem to get fast-charge of their media players / phones / etc. from plugging them into Ubuntu, so I suspect it doesn't. Jul 13 00:25:47 usually you have something like an ISL6185XXC there which will be rated for a particular current and for a particular overcurrent protection Jul 13 00:26:23 NekoXP, Is that something software detectable? Jul 13 00:26:28 absolutely not Jul 13 00:26:34 but the pandaboard schematics are available right? Jul 13 00:26:35 That's what I thought. Jul 13 00:26:37 persia: I have not seen an issue charging my cell from my desktop, so I think the kernel has *some* logic in it. Jul 13 00:27:06 GrueMaster, Does it charge as fast as if you attach it to a 1.5A 5V supply? Jul 13 00:27:58 Not sure. never really measured it, but it appears to be fairly decent. And I don't think my phone can pull that much. Need to look at the spec on micro usb. Jul 13 00:28:10 NekoXP, So, while board designers might do all sorts of things, do you think there's any risk in attempting to comply with the current negotiation specs all the way up to 1.8A in u-boot/linux? Would it be safe to rely on the hardware to simply not do what was asked if it hit it's limites? Jul 13 00:28:13 Something to test with my nook though. Jul 13 00:28:50 I've pushed 2.0A over microUSB, but that required special HW on both ends. Jul 13 00:29:31 absolutely do not :D Jul 13 00:29:46 1.8A is way, way over the spec Jul 13 00:29:53 Which spec? Jul 13 00:30:46 According to Battery Charging Specification, it looks like a device can expect 5.25V at 1.8A assuming all negotiation is successful. Jul 13 00:31:06 ONLY if the USB controller and the stuff behind it complies with the battery charging specification Jul 13 00:31:22 which is a very complicated and kinda ass backwards spec Jul 13 00:31:44 Wouldn't a non-compliant implementation not respond to bit-banging to tell it to provide more than 1.5A (or, in practice, really 500mA)? Jul 13 00:32:01 (and yes, the spec is incredibly complicated) Jul 13 00:32:24 I would assume that the panda pmic is supplying port power Jul 13 00:32:29 so check if the pmic supports the BCS Jul 13 00:33:03 and check that it supports it from a *we are the host* POV, instead of the "we are the device and you are charging your pandaboard's battery via the OTG port" which I am absolutely sure it would support Jul 13 00:33:35 Oh, panda almost certainly doesn't support BCS. Jul 13 00:33:55 But in terms of adding support for current negotiation to upstream u-boot, I'm unsure how much we care about a specific implementation. Jul 13 00:37:28 it is entirely device specific Jul 13 00:37:35 but consider this Jul 13 00:37:49 if your system loaded and did not get past bootloader, why would you be plugging your iPhone into it Jul 13 00:38:32 Sure, but if the system loaded, and started the bootloader, I might want to load my OS off my USB optical drive.. Those usually require ~700mA to spin up. Jul 13 00:38:55 but the maximum in the USB spec is 500mA Jul 13 00:39:05 Unless one implements BCS. Jul 13 00:39:18 your CDROM drive will not comply with the BCS Jul 13 00:39:38 It will with a usb Y cable. Jul 13 00:39:46 500ma per port. Jul 13 00:40:01 Hrm. I've seen a number of one-plug optical drives floating around, which only work when attached to newer platforms. I had presumed they were doing something like that. Jul 13 00:40:22 GrueMaster, But that doesn't require BCS compliance: just regular current negotiation (which u-boot also doesn't have) Jul 13 00:40:41 a USB Y cable is not the battery charging specification Jul 13 00:42:52 Charging spec or not, the question remains of why I can't power a usb-sata drive with a usb Y cable. Charging my cell is very low on the list, but having working usb-sata is very high. Jul 13 00:43:19 (and I lost track of this thread when it diverged from that). Jul 13 00:43:34 GrueMaster, quite possibly because at the end of the day the port power provided will never reach the levels you want Jul 13 00:43:53 GrueMaster, There's a bug in u-boot. It doesn't do *any* current negotiation. As a result, your Y-cable is only providing 200mA to your HD. Jul 13 00:43:54 506mA? Jul 13 00:44:16 I'm curious if we want to implement BCS in u-boot. Jul 13 00:44:19 for instance if you're using Freescale's MC13892 PMIC on a device, using ther VUSB regulator to provide port power, it will supply 100mA and that's it Jul 13 00:44:26 I think NekoXP is suggesting that this probably isn't a good idea. Jul 13 00:44:48 Is it a u-boot or platform issue? If u-boot, it is fixable. If the platform can't handle the load due to design, that's a different issue. Jul 13 00:44:51 NekoXP, Even if the software supports current negotiation? The port can never do 500mA? Jul 13 00:45:21 So, simple answer - hardware limitation. Jul 13 00:45:28 GrueMaster, Both, but in the specific instance of the panda, there is support for 500mA. Jul 13 00:46:03 in the specific case of the board you're trying to make it work on you shouldn't need to NEGOTIATE for current Jul 13 00:46:27 persia: if the hw will support 500ma but u-boot doesn't support autoneg, that's a software issue. If the panda power regulator can't handle it, then it is hardware. Jul 13 00:46:27 What? Jul 13 00:46:28 you simply have to configure the host controller, configure the port, turn VBUS on and initialize the device Jul 13 00:46:46 NekoXP, But USB 2.0 specifies that the device is supposed to *ASK* if it wants more than 100mA. Jul 13 00:47:03 before you actually kick the device it cannot draw more than 100mA per spec anyway Jul 13 00:47:15 GrueMaster, For those definitions of "software" and "hardware", it's a software problem. Jul 13 00:47:17 the descriptors will say whether it wants more Jul 13 00:47:20 * GrueMaster will check back for an answer later. Jul 13 00:47:31 and if so the host can grant that by turning on the regulator Jul 13 00:47:37 there's not really any protocol for it Jul 13 00:47:43 NekoXP, Right, but what we've discovered is that u-boot isn't lifting the power even when the descriptors request more. Jul 13 00:48:59 On a more immediate note, I just discovered panda (u-boot, kernel, whatever) doesn't like the SD to be removed, even when rootfs is on usb. Jul 13 00:49:40 Very odd. Jul 13 00:49:49 GrueMaster, probably kernel Jul 13 00:49:57 GrueMaster, Did you boot kernel from SD, or from somewhere else? Jul 13 00:50:01 there is an option to make the device a little persistent Jul 13 00:50:11 i.e. assume it's there and if not, puke Jul 13 00:50:30 persia: Should be irrelevant as it isn't a mounted filesystem. Jul 13 00:50:53 persia, USB "current negotiation" is down to physics, mostly Jul 13 00:50:55 GrueMaster, I think linux still uses the boot source as a backing store, if it can. I might be mistaken. Jul 13 00:51:18 Works fine on babbage3. Jul 13 00:51:28 NekoXP, Except the bit where the OS reads the descriptors and tweaks the regulator to push to 500mA, unless I misunderstood something (in which case, please tell me what I should have read) Jul 13 00:51:36 GrueMaster, Very odd. Jul 13 00:53:29 a device will not draw more current than you can provide it, and it will not draw current unless you turn it on.. you plug in a USB stick, the USB host controller is providing capability of 500mA (or whatever) to the device 99.9% of the time. if this is controlled by some kind of regulator this voltage can be changed based on power management policy. The USB spec states that an unconfigured device may only draw 20mA suspended (which goes away the moment you Jul 13 00:53:30 talk to it) 100mA otherwise. If you read a descriptor that says 500mA then you may turn your regulator up to 500mA, but in actual fact, most systems just leave it at 500mA and you have no choice really Jul 13 00:53:42 if a device wants 1.8A then it will draw it if you supply it Jul 13 00:54:24 the real problem with magnetic media, cdrom drives and so on is that they have a ridiculous need for spinup current, which spikes and is probably way way over the maximum most USB regulators provide. Jul 13 00:54:40 as the device gets older, it actually needs more to spin up Jul 13 00:55:33 from experience, most board designers don't bother to actually have a configurable regulator Jul 13 00:55:39 it's on or it's off Jul 13 00:55:41 Aha. I think I understand then. So the panda apparently defaulting to 100mA unless told to act differently by software is a specific quirk of the implementation, and in general the practice is to pump 500mA while expecting 100mA draw unless something odd happens? Jul 13 00:55:44 it supports 500mA or not. Jul 13 00:55:57 yeah Jul 13 00:56:27 and the reason it doesn't work is your disk is about 6 months old, it has some wear and tear, and the switcher behind the usb port can't handle a load spike as fast as the disk needs it Jul 13 00:56:43 I see. So, if my understanding of the issue is correct, we'd need panda-specific code in u-boot that checked stuff and turned up the regulators if appropriate. Jul 13 00:56:58 possibly Jul 13 00:57:03 Mind you, u-boot needs to completely turn off USB before booting the OS, so just randomly cranking them probably isn't ideal. Jul 13 00:57:46 Do you happen to know that this sort of thing is not required for common i.MX51 and i.MX53 boards? Jul 13 00:59:07 on the efika, vbus controls whether power is on or off Jul 13 00:59:15 it's configured through the phy which some people consider weird Jul 13 00:59:22 we just turn the damn thing on Jul 13 00:59:28 and this always goes through a hub, which is always powered Jul 13 01:00:09 OK, so u-boot just scans USB, does it's thing, turns "off" the devices (but the hub is powered), and loads the OS? Jul 13 01:00:22 if you don't twiddle vbus, the hub gets 20mA and it fucking likes it. that's actually not enough to bring the hub up. Jul 13 01:02:57 Right. I think I understand. Thanks. Jul 13 01:07:53 oh, yeah Jul 13 01:08:08 the other problem with disks is they actually need 12V supplies Jul 13 01:08:10 and USB is 5V Jul 13 01:08:36 Depends on the disk. Lots of 2.5" and 1.8" disks work fine with 5V. Jul 13 01:08:39 so there's some stuff on the board to get a12V supply from a 5V one and they kind of take time to work, and it reduces the current, and of course as THOSE components age.. they suck :D Jul 13 01:10:09 depends on the disk, really Jul 13 01:11:10 Oh, indeed. Jul 13 01:28:49 I have yet to see a laptop sata drive that needs 12v. Jul 13 02:19:14 * jburkholder got his pandaboard **** ENDING LOGGING AT Wed Jul 13 02:59:56 2011