**** BEGIN LOGGING AT Mon Jan 04 02:59:58 2016 Jan 04 03:08:17 Wizzup: http://maemo.cloud-7.de/maemo5/fiasco+co/Readme_Ubuntu_Lucid_Desktop_SDK_Virtual_Image_Final.txt Jan 04 03:17:27 and http://maemo.cloud-7.de/maemo5/fiasco+co/Readme_Ubuntu_Lucid_Desktop_SDK_Virtual_Image_Final.7z_DIR.txt Jan 04 03:57:17 http://i.imgur.com/4pLE1jX.jpg Jan 04 07:33:51 Wizzup: Which model of ARM notebook computer do you have? Jan 04 08:55:24 DocScrutinizer05: thx Jan 04 08:55:33 brolin_empey: chromebook 1 and 2: snow and peach-pi Jan 04 08:56:02 http://linux-exynos.org/wiki/Samsung_Chromebook_XE303C12 and http://linux-exynos.org/wiki/Samsung_Chromebook_2_XE503C32 Jan 04 09:49:38 so it seems i'm not behind NAT on my mobile network Jan 04 09:49:55 do you pay for incoming traffic too? Jan 04 09:50:47 i don't know. i didn't even know there was no NAT Jan 04 09:54:24 well, external ip is usually a sign of it Jan 04 09:55:07 i've had others with external ip, but firewalled. this one is open Jan 04 09:56:59 no NAT as in you have your own IP address, or no NAT as in only PAT (which I think is actually usually considered NAT anyway)? Jan 04 09:57:40 it's a routable internet address Jan 04 09:58:09 routable as in all packets go to you? Jan 04 09:58:44 yes. i could ssh into my N900 using the address on gprs0 iface Jan 04 09:59:11 moin :) Jan 04 10:00:02 Interesting. Jan 04 10:07:34 mixed feelings Jan 04 10:08:37 sicelo: portscanner ahoy! Jan 04 10:10:24 i'm happy i have iptables properly configured since years ago, but now i need to examine this more closely before i get in big problems (if not already) Jan 04 10:10:57 yeah, external ip is a mixed blessing Jan 04 10:11:11 but as long you pay for incoming traffic, it can be costly Jan 04 10:11:45 because even with fw you still pay for data sent your way (think udp) Jan 04 10:14:15 i'll investigate. though i've had this net for about 6 months now, and have not really had mysterious data depletion Jan 04 10:14:49 lucky for you freenode has cloaking Jan 04 10:16:04 KotCzarny: I thought sometimes udp/icmp doesn't even count towards traffic 'costs' Jan 04 10:16:33 wizzup, nope, all data is counted, otherwise people would abuse it Jan 04 10:16:50 pretty sure my ICMP is not :) Jan 04 10:17:00 but i cant speak for every provider Jan 04 10:17:02 it's also not rate limited Jan 04 10:17:26 wizzup, setup icmp proxy at home and have free data? :> Jan 04 10:17:34 I already have free data Jan 04 10:17:40 but they still list how much I have used Jan 04 10:17:47 KotCzarny: I have 8KB/s internet Jan 04 10:17:51 :) Jan 04 10:17:56 'unlimited' Jan 04 10:18:13 Through EU, too :) Jan 04 10:18:16 Throughout* Jan 04 10:18:20 not bad Jan 04 10:21:30 i blocked icmp long ago (even via wifi). and, re:cloaking, yes, it's nice. i adjusted my irssi to wait before joining channels, although once in a while it "leaks" my IP, lol. then again, i'm not often logged in via N900 directly Jan 04 10:22:07 sicelo, for 100% safeness you can use bouncer at home Jan 04 10:22:19 or just ssh+irssi :) Jan 04 10:22:55 yes, i use Wizzup's solution .. remote server ;) Jan 04 10:23:19 Sicelo009N logs in once in a while (that's N900). Jan 04 10:23:37 He meant use ssh from the n900 Jan 04 10:23:39 I think Jan 04 10:23:44 or, that is what I meant Jan 04 10:23:49 he meant using your bnc at home Jan 04 10:24:24 and that's what i do most of the time (ssh & irssi). i don't find bouncers as user friendly (tried znc, psy) Jan 04 10:26:07 also, i don't have "IP" at "home" .. i'm living in company residence .. free internet via ethernet. it is filtered Jan 04 10:26:26 that's ... bad Jan 04 10:28:01 i don't mind it too much. unlimited fast internet would cost an arm and a leg in these parts. filtering is a much smaller price to pay Jan 04 10:28:14 filtering is philosophical problem Jan 04 10:28:18 I guess you can just use Tor. Jan 04 10:28:23 i do ;) Jan 04 10:28:28 or ssh proxy Jan 04 10:29:33 company residece :/ Jan 04 10:29:52 actually i proxy tor through ssh .. i can more easily explain ssh traffic than tor traffic if someone was to ask :p Jan 04 11:01:11 yeah, as if MB/s via ssh is easily explainable Jan 04 11:01:42 proxy :) Jan 04 11:04:43 'but you are supposed to use our *filtering* proxy' Jan 04 11:07:11 KotCzarny: didn't get your MB/s statement.. Jan 04 11:07:39 freemangordon_: is there a point to defining SW_CAMERA_LENS_COVER and others in event-switches.h ? Do the normal maemo headers lack them? Jan 04 11:07:50 My gentoo system has all the four that are being defined there Jan 04 11:13:50 sicelo: ssh is very low bandwidth, so its easy to distinguish normal ssh work from bulk transfers Jan 04 11:17:40 i am aware, can say sftp :p Jan 04 11:18:21 point is .. there's a bit more 'latitude' Jan 04 11:18:36 What exactly are they filtering anyhway? Jan 04 11:20:01 pr0n Jan 04 11:20:03 probably Jan 04 11:20:42 probably anything that would get the company in trouble Jan 04 11:21:38 not a lot actually .. pr0n, known malicious websites,proxies, gambling, and few other categories .. don't remember all of them Jan 04 11:22:06 pr0n is 'not a lot'? :> Jan 04 11:22:41 internet data is mostly pr0n and spam Jan 04 11:22:54 Sicelo: wikileaks? ;) Jan 04 11:23:10 Surprised they aren't blocking pirated content Jan 04 11:24:44 Also I do wonder what they are doing about encrypted web content (i.e. SSL) Jan 04 11:24:48 KotCzarny: i meant .. not a lot.. not many categories .. just the obvious ones Jan 04 11:25:46 jonwil: they do block known pirate stuff. sometimes even archive.org gets blocked .. not sure why (and this is the most common reason i enabled ssh proxy) Jan 04 11:26:01 yeah I suspected they would block pirated content Jan 04 11:26:06 maybe keyword matching Jan 04 11:26:26 don't seem they touch ssl @ home, but in the work network, they have ssl inspection. Jan 04 11:26:51 Corporate networks will usually have SSL proxying for work PCs Jan 04 11:27:02 with a special local root certificate installed for the SSL proxy Jan 04 11:27:10 yep. exactly that Jan 04 11:27:19 good that they aren't doing anything for SSL for the filtered home network Jan 04 11:27:45 Although I suspect they probably still filter by DNS or IP somewhere Jan 04 11:28:06 but all this doesn't trouble me as generally my browsing habits are "safe" haha. but the one day i got link from jonwil .. had to proxy that as i was unsure Jan 04 11:28:41 yep, the do dns filtering .. i made sure my dns queries go via the ssh proxy too (and tested it) :p Jan 04 11:28:59 that's why i know opera mobile on N900 is not very "safe" Jan 04 11:29:33 :> Jan 04 11:29:39 Security is something I care a lot about Jan 04 11:29:55 as far as proxying .. KotCzarny (re:opera) Jan 04 11:29:56 Which is why I am doing whatever I can to update the N900 and make it more secure Jan 04 11:30:05 :) Jan 04 11:30:11 jonwil: \m/ Jan 04 11:31:09 having updated kernel/iptables/libs(libc,ssl)/servers is usually enough Jan 04 11:32:42 jonwil++ Jan 04 11:33:43 Updating OpenSSL and the things that use it is high on the wishlist Jan 04 11:34:42 As is updating the set of root certificates on Maemo Jan 04 11:34:53 inside maemo-security-certman Jan 04 11:47:27 and also update nss inside microb Jan 04 11:47:50 Those 3 are the most important things to do for security Jan 04 11:50:48 currently waiting to see if Juhani Mäkelä can provide more info about how to update the certificates Jan 04 12:07:44 In the long run, everything should really be ported to Debian. Jan 04 12:07:59 or abandoned. Jan 04 12:10:40 Ideally yeah Jan 04 12:10:47 but for now we are stuck with maemo-security-certman Jan 04 12:13:01 and all the closed things that link to it like activesync for exchange, browser UI, wifi security daemon and various bits of the connectivity UI layer Jan 04 12:15:53 I don't care about Maemo enough to do most of it. Maemo is just more hackable (easy to swap components around—works relatively similar to other unix-like OSes) than others, so it's preferable to me. Jan 04 12:16:11 ok Jan 04 12:16:38 Luckily, others in here (#maemo), do Jan 04 12:16:54 My goal is to make Maemo more secure without sacrificing useful functionality Jan 04 12:17:24 I'd be happier running some custom xmonad configuration as a WM and some simple UI/database system for SMS/telephony. Jan 04 12:17:36 (on top of ofono or something) Jan 04 12:17:44 If it was up to me, I would say "screw ActiveSync" but aparently there are people on this earth who own a N900 and need to talk to an ActiveSync server Jan 04 12:17:46 well, do it Jan 04 12:17:52 * Maxdamantus hasn't figured out what there is of the last part yet. Jan 04 12:18:00 I find the maemo UI much more pleasant for mobile than any WM I have tried Jan 04 12:18:13 jonwil: there are? (re: activesync) Jan 04 12:18:21 aparently there are Jan 04 12:18:54 I would say screw activesync -atm- and worry about it later :) Jan 04 12:19:00 yes, i'm one of them. Jan 04 12:19:05 anyhow, zzz time :), later guys Jan 04 12:19:50 * Maxdamantus just sshes into his main computer to do email/IRC/etc Jan 04 12:20:22 EAS is currently the only well-integrated sync method still working on N900 Jan 04 12:20:57 discounting syncml via usb (pc suite) Jan 04 12:21:47 or bluetooth .. all of which are of course not so convenient Jan 04 12:22:47 Maxdamantus: EAS is not only email .. it can sync calendar & contacts. works very well on N900, trust me Jan 04 12:24:09 Wizzup: there are defined in linux/input.h, and are missing in 2.4.26 Jan 04 12:24:17 freemangordon_: 2.6.26 I presume Jan 04 12:24:44 2.6.28 :) Jan 04 12:24:55 eh, yeah Jan 04 12:27:03 (re:my network... torrents aren't blocked, but you'l! definitely be asked what's up) Jan 04 12:28:39 anyway, i don't generally torrent anything. Jan 04 16:13:36 Wizzup: OK, thank you. Jan 04 16:16:59 Maxdamantus: to be honest, the best part about maemo is probably the community/ecosystem around it :) Jan 04 17:17:42 Pali: russell applied the patches, just don;t ask me to which tree :). also, see https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=574aab1e02837927e3c94193eedf94128ad10b6d, going to cherrypick to see if it fixes the net oops Jan 04 17:38:06 freemangordon: ok, do not forget and just check that patches will be in 4.5-rc1 Jan 04 17:44:54 sure Jan 04 18:41:04 Pali: ke-recv still spits "ke_recv[2434]: get_usb_cable_udi:1725: coudn't find USB cable indicator" Jan 04 18:41:27 there is something broken in hal or ke-recv.. Jan 04 19:35:19 N900 doesn't have a selfie-stick .. if one was to make one, what's the best way to capture at that distance? :) Jan 04 19:36:08 does it have a timer? Jan 04 19:37:26 yes Jan 04 19:37:37 camera2-ui .. up to 20s Jan 04 19:38:12 but it would be nice to have some remote button/control Jan 04 19:40:36 do we really want a selfie-stick? :( Jan 04 19:41:14 haha, i don't (because i hate taking pics) Jan 04 19:42:07 Pali: yeah, hal Jan 04 19:42:56 merlin1991: : do we have maemo hal in gitorious backup? Jan 04 19:46:54 Sicelo: create a network input device with a 'camera button' ? :P Jan 04 19:47:30 :) Jan 04 19:48:33 i'll look to see how the others work Jan 04 19:50:11 i've seen people connect to earphone jack .. but i don't understand how that relates to camera Jan 04 19:51:05 Sicelo: listening to a hal/mce/whatever event and sending a dbus command, I guess Jan 04 19:51:12 or something close Jan 04 19:54:58 nice idea Jan 04 19:55:44 e.g. listen for headset button (which we already have in Maemo) Jan 04 20:01:49 freemangordon: no idea Jan 04 20:02:23 pitty that the n900 doesn't have an ir receiver Jan 04 20:02:33 it has bt Jan 04 20:02:38 should be enough Jan 04 20:02:48 ir is way simpler Jan 04 20:02:58 bt passes through obstacles Jan 04 20:03:12 so you can have it in your pocket etc Jan 04 20:03:30 I never did make selfies, so no idea if obstacles involved Jan 04 20:03:50 but ir could be set up in no time Jan 04 20:04:00 no paring.. Jan 04 20:04:02 although our N900 is quite a brick. i'm sure the weight would be felt Jan 04 20:04:43 sicelo, bt mini keyboard Jan 04 20:04:48 == instant button Jan 04 20:04:53 ;) Jan 04 20:05:27 couldnt be simpler i think Jan 04 20:05:33 that's also neat idea. camera-ui2 actually captures when you press C on keyboard Jan 04 20:05:44 only thing you need to set up is pairing devices Jan 04 20:05:52 no need to tweak/hack anything Jan 04 20:06:09 Finally, a use for bluetooth! Jan 04 20:06:13 :) Jan 04 20:06:24 bt mouse could work too Jan 04 20:06:45 assuming you block the laser and only rely on buttons Jan 04 20:07:42 that may start needing hacks .. how do you capture? Jan 04 20:07:52 press the button? Jan 04 20:08:17 i don't think that would work without some sort of mapping? Jan 04 20:08:17 (you need to position mouse pointer over the clicky area first, then put a piece of tape over the laser Jan 04 20:08:35 haha, nah.. that would be difficult Jan 04 20:08:39 camera-ui2 doesnt have a make-a-photo-on-click ? Jan 04 20:08:46 it does Jan 04 20:08:51 but bt keyboard is way easier Jan 04 20:08:59 yes, your idea is bed Jan 04 20:09:04 *best Jan 04 20:10:01 are there bt headsets with more buttons that can be mapped? Jan 04 20:10:26 what's wrong with the wired headset button ? :p Jan 04 20:10:32 not-simple-enough? Jan 04 20:10:33 merlin1991: who should I ask, xes? Jan 04 20:10:37 not-long-enough Jan 04 20:10:52 also you dont want cable in the photo, do you? Jan 04 20:11:14 it's on the stick ... Jan 04 20:11:16 it can be long enough .. but you need some dbus-foo as well to make it trigger Jan 04 20:11:38 sicelo, that dbus-foo is trivial Jan 04 20:11:57 you can even use oscp-dbus.py from oscp to just map the event to something Jan 04 20:12:10 for example faking 'c' key Jan 04 20:13:25 yes, of course. i did like the headset idea. i'm just saying that bt keyboard definitely is quickest/painless way (even for someone that knows zilch about dbus, etc.) Jan 04 20:13:45 freemangordon: https://github.com/community-ssu/hal Jan 04 20:14:54 my git repo from Jun 11 2014 Jan 04 20:19:19 Pali: thanks Jan 04 21:22:16 Pali: how do you think, where should we get usb cable status? bme or isp? Jan 04 21:22:21 isp1707 that is Jan 04 21:24:30 Pali: or we should add button.type = usb.cable to hald bme addon? Jan 04 21:24:50 how does it work thus far? Jan 04 21:25:16 bencoh: which one? Jan 04 21:25:17 from hal bme Jan 04 21:25:31 (hald-addon-bme) Jan 04 21:25:31 Pali: hal bme addon? Jan 04 21:25:34 ok Jan 04 21:25:41 I was thinking the same :) Jan 04 21:26:01 Pali: where is the code? Jan 04 21:26:10 in cssu? Jan 04 21:26:29 hald-addon-bme source code is on github cssu Jan 04 21:26:31 found it Jan 04 21:27:44 Pali: is that supposed to work with stock kernel? Jan 04 21:28:15 bme replacement hald-addon-bme has same API as nokia bme Jan 04 21:28:51 so if you use hald-addon-bme API it should work on both stock Maemo and bme-replacment Maemo Jan 04 21:29:15 Pali: I know, but is that addon supposed to work with KP and stock kernels? as we'll duplicate usb.cable Jan 04 21:29:40 ke-recv takes 1st found, but I wonder is someone else uses that Jan 04 21:29:44 no, bme replacment version work *only* with bme replacment (kp or upstream kernel) Jan 04 21:30:17 ok, bu KP exports /org/freedesktop/Hal/devices/usb_device_1d6b_2_musb_hdrc Jan 04 21:30:20 *but Jan 04 21:30:33 with button.type = 'usb.cable' (string) Jan 04 21:30:44 freemangordon: but I did not catch up where is problem... Jan 04 21:30:47 in ke-recv? Jan 04 21:30:49 I will add the same in bme Jan 04 21:31:01 so we'll have 2 udis with button.type = 'usb.cable' (string) Jan 04 21:31:08 ke-recv needs musb device! Jan 04 21:31:14 not bme! Jan 04 21:31:19 this is not musb device Jan 04 21:31:26 this is usb cable detection Jan 04 21:31:30 ok, so what is problem? Jan 04 21:31:40 (23,30,59) freemangordon: so we'll have 2 udis with button.type = 'usb.cable' (string) Jan 04 21:32:00 in which kernel? Jan 04 21:32:04 KP Jan 04 21:32:09 stock too? Jan 04 21:32:26 if used with bme-replacement, yes Jan 04 21:32:51 but I guess this is not possible Jan 04 21:33:59 Pali: I don't know it this will bring problems, just making a note Jan 04 21:34:00 ok, I'm lost... problem is in which combinations? stock+nokiabme; kp+nokiabme; kp+replacbme; upstream+replacbme Jan 04 21:34:12 kp+replacbme Jan 04 21:34:34 and which hal UDIs have that? Jan 04 21:35:08 now only /org/freedesktop/Hal/devices/usb_device_1d6b_2_musb_hdrc Jan 04 21:35:26 if I add it to bme, both bme and ^^^ will have it Jan 04 21:35:37 when the combination is kp+replacbme Jan 04 21:35:46 what if you add to bme? Jan 04 21:35:59 button.type = 'usb.cable' Jan 04 21:36:08 and why you want to add it? Jan 04 21:36:10 because ke-recv needs it Jan 04 21:36:20 ?? Jan 04 21:36:26 nokia bme does not add it Jan 04 21:36:31 libhal_manager_find_device_string_match(hal_ctx, "button.type", "usb.cable", &num_devices, &error); Jan 04 21:36:33 so ke-recv does not need it Jan 04 21:36:58 it is added by p, li { white-space: pre-wrap; } hald-addon-usb-cable Jan 04 21:37:13 hald-addon-usb-cable Jan 04 21:37:46 now I'm lost... can you explain me again what is problem? Jan 04 21:38:00 ke-recv needs to detect when usb cable is connected Jan 04 21:38:21 to show you the "mass storage" dialog Jan 04 21:38:24 ok? Jan 04 21:38:29 ok Jan 04 21:38:35 this is done by gadget driver Jan 04 21:38:36 musb Jan 04 21:38:42 musb kernel code Jan 04 21:38:45 not in upstream kernal Jan 04 21:38:51 *kernel Jan 04 21:38:59 hm? this should be there Jan 04 21:39:07 but it is not Jan 04 21:39:24 hm... maybe I removed something important from .config? Jan 04 21:39:29 no idea Jan 04 21:39:43 but it was the same before you cleaned it up Jan 04 21:40:14 so problem is that in ke-recv "usb_cable_udi" variable is empty? Jan 04 21:40:16 which udi you think should export that? Jan 04 21:40:34 that UDI which export musb state Jan 04 21:40:52 b_idle b_peripheral ... Jan 04 21:41:04 Pali: the problem is that ke-recv cannot find udi which exports "button.type"="usb.cable" Jan 04 21:41:30 can you share me again lshal? Jan 04 21:41:55 ~ # lshal | grep idle Jan 04 21:41:55 ~ # Jan 04 21:41:55 it has been said that # is pound, or hash, or sometimes the number sign. a tic-tac-toe board, or octothorpe Jan 04 21:41:56 sure Jan 04 21:42:32 $ cat /sys/devices/platform/musb_hdrc/mode Jan 04 21:42:39 is in stock kernel Jan 04 21:42:55 similar sysfs entry must be also on upstream kernel Jan 04 21:43:01 http://pastebin.com/RVb9Ruh5 Jan 04 21:43:15 and if exists it should be exported to HAL (if not, bug is in HAL) Jan 04 21:43:40 udi = '/org/freedesktop/Hal/devices/platform_musb_hdrc_0_auto' Jan 04 21:44:55 freemangordon: see HAL hald/linux/addons/addon-usb-cable.c Jan 04 21:45:35 what to look for there? Jan 04 21:45:39 sysfs paths? Jan 04 21:45:48 yes Jan 04 21:45:54 this is usb-cable module Jan 04 21:46:01 see 10-usb-cable.fdi Jan 04 21:46:09 so this module is responsible for exporting usb.cable right? Jan 04 21:46:10 I know Jan 04 21:46:14 no Jan 04 21:46:19 see ^^^ fdi Jan 04 21:46:56 HAL hald/linux/device.c Jan 04 21:46:57 Jan 04 21:47:03 search for "usb.cable" Jan 04 21:47:37 this code set button.type to usb.cable Jan 04 21:47:38 hmm Jan 04 21:47:50 if exists hal_util_get_string_from_file (sysfs_path, "../mode"); Jan 04 21:47:55 yes Jan 04 21:49:06 it should match in our case udi = '/org/freedesktop/Hal/devices/platform_musb_hdrc_0_auto' which has linux.sysfs_path = '/sys/devices/platform/68000000.ocp/480ab000.usb_otg_hs/musb-hdrc.0.auto' Jan 04 21:49:18 yes Jan 04 21:49:24 so check which sysfs node has entry "mode" Jan 04 21:49:32 it has, just checked Jan 04 21:49:41 find /sys -name mode Jan 04 21:49:43 what is result? Jan 04 21:49:58 /sys/devices/platform/68000000.ocp/480ab000.usb_otg_hs/musb-hdrc.0.auto # ls Jan 04 21:49:58 driver driver_override gadget modalias mode musb-hdrc.0.auto power srp subsystem uevent vbus Jan 04 21:50:37 oh Jan 04 21:50:43 it looks for "../mode" Jan 04 21:50:57 hmm Jan 04 21:51:02 will look at stock Maemo Jan 04 21:51:30 I have KP booted, will check Jan 04 21:51:48 /sys/devices/platform/musb_hdrc/mode Jan 04 21:52:22 ah right Jan 04 21:52:25 usb_device.linux.sysfs_path = '/sys/devices/platform/musb_hdrc/usb1' Jan 04 21:52:45 see for info.product Jan 04 21:52:51 host driver Jan 04 21:52:56 binded to usb1 Jan 04 21:53:08 yep Jan 04 21:53:14 you should see it in lsusb Jan 04 21:53:30 what do you see on upstream kernel in lsusb? Jan 04 21:53:41 nothing :) Jan 04 21:53:48 this looks like broken gadget support Jan 04 21:53:58 can you compile g_nokia as external module? Jan 04 21:54:05 (not linked into zImage) Jan 04 21:54:45 ok, gimme 5 minutes Jan 04 21:55:27 Pali: only g_nokia? Jan 04 21:55:42 yes Jan 04 21:55:51 or you can also g_mass_storage Jan 04 21:55:55 hmm, why is # CONFIG_USB_MUSB_GADGET is not set Jan 04 21:56:11 ah, p, li { white-space: pre-wrap; } CONFIG_USB_MUSB_DUAL_ROLE=y Jan 04 21:58:03 I know why Jan 04 21:58:12 isp1704 disconnect gadget at load Jan 04 21:58:26 hmm Jan 04 21:58:28 so if gadget is not loaded *after* isp704 we loose it Jan 04 21:58:57 and when gadget is compiled into zImage it cannot be initialized after isp1704 easily Jan 04 22:01:34 ok, rebooting Jan 04 22:07:15 Pali: [ 260.860717] g_nokia gadget: using musb-hdrc, OUT ep1out, IN ep1in Jan 04 22:07:26 still lsusb prints nothing Jan 04 22:07:42 hm... Jan 04 22:08:20 same when usb_f_mass_storage is loaded Jan 04 22:09:10 Pali: there is no /proc/bus/usb Jan 04 22:09:50 /proc/bus/usb does not exists Jan 04 22:10:05 ok Jan 04 22:10:06 use /sys/bus/usb or /dev/bus/usb Jan 04 22:12:54 Pali: there is /sys/devices/platform/68000000.ocp/480ab000.usb_otg_hs/musb-hdrc.0.auto/musb-hdrc.0.auto Jan 04 22:14:05 hm... Jan 04 22:14:12 ls -l -a /sys/devices/platform/68000000.ocp/480ab000.usb_otg_hs/musb-hdrc.0.auto/musb-hdrc.0.auto/../mode Jan 04 22:14:19 it exists! Jan 04 22:14:21 (in qemu) Jan 04 22:14:42 want me to boot a mainline kernel I had a bit ago? had basic gadgets working. Jan 04 22:14:52 yes Jan 04 22:15:08 hm... that HAL code first check for usb device Jan 04 22:15:12 hang on - will take a few minutes (I think the battery ran out) Jan 04 22:15:15 and then check for that ../mode Jan 04 22:15:24 * Oksanaa : moving bookmarks around + low battery shutdown -> loss of bookmarks being moved around? Jan 04 22:15:25 it exists here too Jan 04 22:15:31 freemangordon: do you stil need the backup? Or have you already solved? Jan 04 22:15:36 solved Jan 04 22:15:39 Wizzup: do you have working mainline kernel? Jan 04 22:15:44 freemangordon: ok Jan 04 22:15:50 Wizzup: which version? Jan 04 22:17:54 Pali: it is from a month or two ago. hang on. I made lots of notes Jan 04 22:18:16 older version is perfect Jan 04 22:18:26 4.1.0-rc4 is what I have Jan 04 22:18:34 (your branch) Jan 04 22:19:02 (it is not running maemo - though - but I had gadgets working in gentoo I think) Jan 04 22:20:21 what do you want me to do - it is booted Jan 04 22:20:35 output from lsusb Jan 04 22:20:40 (when gadget is active) Jan 04 22:20:52 gentoo900 ~ # lsusb Jan 04 22:20:52 unable to initialize libusb: -99 Jan 04 22:20:53 eh. Jan 04 22:21:00 sudo? Jan 04 22:21:04 no -as root Jan 04 22:21:08 :) Jan 04 22:21:16 http://sprunge.us/aIbO Jan 04 22:21:25 it is probed Jan 04 22:21:35 the same here Jan 04 22:22:52 interestingly enough Jan 04 22:22:58 the gadget does work Jan 04 22:23:56 I have /sys/bus/usb Jan 04 22:25:25 I'm confused - do I still need to do something - or not really Jan 04 22:26:03 Wizzup: I guess no Jan 04 22:26:26 It is weird that libusb doesn't want to initialise, though. but one problem at the time :) Jan 04 22:26:46 what is error -99? Jan 04 22:26:49 I tried to get host mode to work some time ago - but gave up at some point, I think because it wasn't possible to set the mode using the debugfs anymore or something - something was removed/changed Jan 04 22:26:53 good question. Jan 04 22:27:03 do you not get that error? Jan 04 22:27:12 as I said - this is an older version that what you are using, I think Jan 04 22:27:36 well, lsusb just outputs nothing here Jan 04 22:27:40