freemangordon | so (again a wild guess), the interrupt line is enabled when it should not (a) or luf's device/cable is screwed | 00:00 |
---|---|---|
freemangordon | luf: did you try with different cable | 00:00 |
freemangordon | ? | 00:00 |
luf | No because it happens only 4 times. | 00:00 |
luf | I'm using the cable with the dev phone very often. | 00:00 |
luf | * I use ... | 00:02 |
freemangordon | Pali: http://logs.nslu2-linux.org/livelogs/maemo/maemo.20101013.txt | 00:02 |
freemangordon | search for bug | 00:02 |
luf | freemangordon: I'm not alone :D | 00:03 |
luf | freemangordon: did you see that after osso-gnomevfs-extra cleanup the browserd core dump doesn't arrive? | 00:04 |
freemangordon | luf: yes | 00:04 |
freemangordon | though I wonder what cleanup you are talking about | 00:05 |
freemangordon | :) | 00:05 |
luf | I'll upload to git when I finish. After that I'll add posibility to recreate bda -> path hash (in case of gnomevfs daemon failure). | 00:06 |
freemangordon | luf: did you get such oopses with KP51r1? | 00:07 |
luf | I find that there is dbus_message_unref prior working with the message data (gchar *). So maybe it was this. | 00:07 |
luf | freemangordon: It's the only 4 oopses I have ever seen with N900. | 00:07 |
freemangordon | aah, for sure | 00:07 |
freemangordon | luf: all of them with KP52, ain't? | 00:08 |
luf | Shit now I'm checking prod N900 and the error is there also. | 00:10 |
freemangordon | luf: what kernel? | 00:10 |
luf | KP50 | 00:11 |
freemangordon | luf: what machine/os do you connect to? | 00:11 |
luf | It' not HW problem with N900. | 00:11 |
freemangordon | sure | 00:11 |
luf | freemangordon: FC16 (my desktop). | 00:11 |
luf | But I have no idea if I used the same or different cable :( | 00:12 |
freemangordon | luf: however the kernel should deal with broken cabling :) | 00:13 |
freemangordon | luf: do you have some reliable way of reproducing that? | 00:13 |
luf | freemangordon: no :( | 00:14 |
luf | The first time on dev N900 I saw when I was very low on battery. But last time N900 has nearly full battery. | 00:14 |
luf | I'm using dev N900 several month (nearly day to day) and only 4 oopses. | 00:15 |
freemangordon | hmm, BUG_ON is missing from the upstream code | 00:20 |
freemangordon | Pali: ^^^ http://lxr.free-electrons.com/source/drivers/usb/musb/musb_gadget.c#L908 | 00:20 |
freemangordon | neither there is such a loop | 00:21 |
*** NIN101 has quit IRC | 00:25 | |
luf | Another question. I have a lot ot useless d(x) calls and I want to remove it using #define d(x) ... What is the right construct for it? | 00:27 |
freemangordon | luf: I am not sure I get your question | 00:28 |
luf | #define d(x) | 00:28 |
luf | Is it correct? | 00:29 |
freemangordon | yep | 00:29 |
luf | I see recomendation to use it this (safer way): #define d(x) {} | 00:30 |
freemangordon | should not matter | 00:31 |
luf | Right. I'll try :) | 00:31 |
freemangordon | luf: unfortunately I don;t think we'll be able to fix that bug without having a way to reproduce it :(. anyway, lets see if Pali will come up with something | 00:39 |
luf | freemangordon: I understand. | 00:40 |
*** Guest25139 has quit IRC | 00:40 | |
freemangordon | uploading microb-engine in cssu-devel | 00:41 |
luf | freemangordon: I pushed the cleanup into git. | 00:47 |
freemangordon | do you mind to upluad to -devel too? | 00:47 |
freemangordon | *upload | 00:47 |
luf | Not now. I'll finish one task first. | 00:48 |
freemangordon | luf: microb-engine is in the repo | 00:55 |
luf | freemangordon: gread job. Thanks. | 00:55 |
freemangordon | the fuck, I had forgotten how slow is microb ARM build :) | 01:00 |
luf | :D | 01:00 |
*** arcean has quit IRC | 01:04 | |
luf | freemangordon: tell us how long it takes. | 01:04 |
freemangordon | which one? | 01:04 |
luf | freemangordon: microb ARM build | 01:04 |
*** arcean has joined #maemo-ssu | 01:05 | |
freemangordon | oh, you got me wrong, I mean - how slow is microb when microb-engine is build with gcc 4.2.1 | 01:05 |
freemangordon | otherwise the build takes ~40 minutes | 01:06 |
luf | freemangordon: What's the difference (%) ? | 01:06 |
DocScrutinizer05 | did *anything* regarding internet connection and particularly DHCP change in CSSU (lately)? | 01:06 |
freemangordon | NFC, it is just rendering much slower than -thumb build | 01:06 |
freemangordon | DocScrutinizer05: afaik Pali has changed something, though not sure if it is in -testing | 01:07 |
luf | freemangordon, DocScrutinizer05: it's disabled by default. | 01:07 |
DocScrutinizer05 | I had more issues with broken connectivity during last 2 weeks orso than the 4 years before | 01:07 |
Pali | DocScrutinizer05, changes are only in usb0 | 01:08 |
Pali | and everything is disabled by default | 01:08 |
*** arcean has quit IRC | 01:08 | |
DocScrutinizer05 | given the low frequency I leave home, I could almost claim it's 100% reproducible | 01:09 |
DocScrutinizer05 | I'll keep an eye on it | 01:09 |
* DocScrutinizer05 frowns | 01:09 | |
kerio | don't frown! you'll get wrinkles! | 01:15 |
DocScrutinizer05 | leaving WLAN area, 2.5G doesn't connect. Switching to "dual", takes 20s to crank up 3G, still no connect. "No Internet". Connecting to GPRS, switching back to 2.5G, works. | 01:27 |
DocScrutinizer05 | Feels like ICD2 or whomever doesn't connect to GPRS anymore, though I got setting "connect to any internet available" | 01:28 |
DocScrutinizer05 | rather "automatic connect: Any connection" | 01:30 |
* ShadowJK would blame operator | 01:41 | |
* ShadowJK left home today, N900 connected to 2.5g | 01:42 | |
luf | freemangordon: ping | 01:47 |
*** wmarone has quit IRC | 01:47 | |
*** wmarone has joined #maemo-ssu | 01:48 | |
luf | freemangordon: libnspr4 libnss3 libnss3-certs - these packages were uploaded with microb-engine? | 01:51 |
DocScrutinizer05 | ShadowJK: hard to blame those 0.5th level help desk robots | 02:00 |
DocScrutinizer05 | ShadowJK: unless I have some hard facts to slap them with | 02:00 |
ShadowJK | Basically the modem hides their buggy network config | 02:02 |
DocScrutinizer05 | ShadowJK: but yeah, it kinda explains the behavior when I assume >>DHCP doesn't work for new lease on 2G, while handing over an existing lease from 3G works" | 02:02 |
ShadowJK | I suspect dhcp isn't really relevant | 02:04 |
ShadowJK | like, I've never seen a dhcp request or refresh in 3g-2g-3g transitions | 02:04 |
DocScrutinizer05 | who however is the culprit for DHCP not working on 2G, that's still up to anybody's guess. Particularly since during last years I found 2G recovering from such lockup when I reboot the device | 02:04 |
DocScrutinizer05 | and yes, it could as well be something around APN sign-in | 02:05 |
ShadowJK | dhclient or dhcp overall is probably not at all involved at any stage of a cellular connection | 02:05 |
DocScrutinizer05 | errr, how would ICD2 resp dnsmasq know about the device's IP then? | 02:06 |
ShadowJK | "dhcp not working" is hard to tell apart from "connection is not working at all" unless you've got access to both ends | 02:06 |
DocScrutinizer05 | after all I config my GPRS DHCP in connections-advanced | 02:07 |
DocScrutinizer05 | I *could* | 02:08 |
DocScrutinizer05 | "get IP automatically" and "get DNS automatically" | 02:09 |
ShadowJK | It's not dhcp | 02:10 |
DocScrutinizer05 | o.O | 02:10 |
ShadowJK | Since when has gsm/3g ever reused an existing protocol when they could invent a new protocol :P | 02:11 |
ShadowJK | dhcp needs an existing IP connection, I suspect the addresses get negotiated before IP packets start flowing | 02:12 |
ShadowJK | there's also no dhcp on dialup internet, PPPoE and PPPoA +dsl either | 02:13 |
luf | freemangordon: I announced also your upload of microb-engine to cssu-devel on TMO. | 02:32 |
*** luf has quit IRC | 02:33 | |
DocScrutinizer05 | ShadowJK: I'd say you got a point there | 03:01 |
*** arcean has joined #maemo-ssu | 03:52 | |
*** arcean has quit IRC | 03:56 | |
*** kolp_ has joined #maemo-ssu | 04:14 | |
*** kolp has quit IRC | 04:17 | |
*** kolp_ has quit IRC | 04:44 | |
*** nox- has quit IRC | 05:02 | |
*** snx has joined #maemo-ssu | 05:16 | |
*** Jaded has joined #maemo-ssu | 05:21 | |
*** Jaded has quit IRC | 05:21 | |
*** Jaded has joined #maemo-ssu | 05:21 | |
*** Jade has quit IRC | 05:24 | |
*** amiconn_ has joined #maemo-ssu | 05:36 | |
*** amiconn has quit IRC | 05:36 | |
*** amiconn_ is now known as amiconn | 05:36 | |
*** DocScrutinizer05 has quit IRC | 06:03 | |
*** DocScrutinizer05 has joined #maemo-ssu | 06:03 | |
*** snx has quit IRC | 06:05 | |
*** Pali has quit IRC | 06:34 | |
*** snx has joined #maemo-ssu | 07:24 | |
*** Jaded is now known as jade | 07:26 | |
*** snx has quit IRC | 07:37 | |
*** luf has joined #maemo-ssu | 08:02 | |
*** Martix has quit IRC | 08:17 | |
freemangordon | luf: can we do anything about the missing thumbnails? to stop trying to get them for example? | 09:14 |
freemangordon | luf: this is for obexftp | 09:16 |
luf | freemangordon: I understand. I also have it in my todo. | 09:22 |
freemangordon | luf: another inconvenience (though I think nothing can be done) - the connection to the devices is tried in order BT returns them | 09:24 |
freemangordon | this leads that 5th device in the list (which is available) to be hardly accessible, because the 4 previous are not available | 09:25 |
freemangordon | I hope you get what I mean | 09:25 |
freemangordon | I wonder if async dbus calls with callbacks could help here | 09:26 |
luf | freemangordon: we talked about it last week. You said it's async ... | 09:26 |
freemangordon | luf: BT as a whole is async related to the other fs | 09:27 |
freemangordon | but device enumeration/browsing in BT is not AIUI | 09:28 |
luf | freemangordon: enumeration is fast enough. The problem is that it tries to preload the content one level down ... | 09:28 |
freemangordon | i.e. until there is no answer from device A (evenv if it is "device not found"), device B is not enumerated | 09:28 |
freemangordon | luf: exactly. however, it is not a major problem, so I think we could live with it | 09:29 |
freemangordon | if there is no some easy solution | 09:29 |
luf | freemangordon: I can add it to the todo however I have no idea how to do it. | 09:30 |
freemangordon | luf: I guess it is something in gw_... calls | 09:30 |
freemangordon | luf: it is up to you | 09:30 |
luf | freemangordon: Hmmm maybe some monitor ... (the problem is on FS side) | 09:31 |
freemangordon | yes | 09:31 |
luf | I have no problem with BLUEZ/DBUS but right now it's just function call from gnome-vfs-daemon and response (sync). | 09:31 |
freemangordon | could it be that the "serial.connect" call is causing it to stall? | 09:32 |
luf | freemangordon: however it'll have downside in periodical (?) checking of device status (online/offline). | 09:32 |
freemangordon | hmm, yep, afaik it blocks in "serial.connect" for the devices not in the range | 09:33 |
luf | freemangordon: you said that it's not a problem. However only adding some g_warning and checking logs can gives you the answer. | 09:33 |
freemangordon | luf: :nod: | 09:33 |
freemangordon | ok, i'll do it when have some time | 09:34 |
* freemangordon heads for the office | 09:34 | |
luf | I'll add todos: 1) try to change connecting in parallel; 2) try to handle BT devices using some monitor (but how to check availability periodically). | 09:36 |
luf | 1) just inside the one sync call from gnome-vfs-daemon 2) it'll be fully async | 09:36 |
luf | freemangordon: how long "freeze" do you see because of not reachable 4 devices? | 09:37 |
kerio | hm, is there an obexfs fuse file system? | 10:07 |
kerio | it would be neat | 10:07 |
luf | kerio: no idea | 10:09 |
*** Martix has joined #maemo-ssu | 10:18 | |
*** jade has quit IRC | 11:25 | |
*** jade has joined #maemo-ssu | 11:25 | |
*** jade has quit IRC | 11:25 | |
*** jade has joined #maemo-ssu | 11:25 | |
*** Martix has quit IRC | 11:42 | |
*** Martix has joined #maemo-ssu | 11:42 | |
*** Martix has quit IRC | 11:49 | |
*** andre__ has joined #maemo-ssu | 12:04 | |
*** andre__ has joined #maemo-ssu | 12:04 | |
luf | freemangordon: ping | 12:28 |
luf | freemangordon: never mind I had wrong idea | 12:31 |
*** gregoa has quit IRC | 12:37 | |
*** kolp has joined #maemo-ssu | 12:37 | |
*** gregoa has joined #maemo-ssu | 12:37 | |
*** Pali has joined #maemo-ssu | 12:52 | |
*** M4rtinK has joined #maemo-ssu | 13:08 | |
*** amiconn has quit IRC | 13:20 | |
*** amiconn has joined #maemo-ssu | 13:21 | |
*** arcean has joined #maemo-ssu | 13:21 | |
*** lizardo has joined #maemo-ssu | 13:40 | |
freemangordon | luf: I misexplained it - there is no "freeze", but a device in range which is 5th in the list, gives "unable to connect" error. I guess it is because something times out while bluez(or whoever) tries to connect to the 4 devices not in range | 13:45 |
freemangordon | otherwise UI is fully responsible all the time | 13:45 |
luf | freemangordon: can you post me what's in syslog? | 13:56 |
luf | freemangordon: I think this should be logged (the error why it's unable to connect). | 13:57 |
*** BCMM has joined #maemo-ssu | 13:59 | |
freemangordon | can't do it now, my second device is not with me here in the office | 14:05 |
freemangordon | will do later, when back home | 14:05 |
freemangordon | luf: ^^^ | 14:05 |
luf | freemangordon: ok | 14:16 |
*** joshgillies has quit IRC | 14:34 | |
*** arcean has quit IRC | 14:53 | |
*** wirr has joined #maemo-ssu | 15:02 | |
*** xmlich02 has joined #maemo-ssu | 15:08 | |
*** BCMM has quit IRC | 15:12 | |
*** luf has quit IRC | 16:00 | |
*** arcean has joined #maemo-ssu | 16:48 | |
*** MrPingu has joined #maemo-ssu | 16:59 | |
*** chainsawbike has quit IRC | 17:03 | |
*** chainsawbike has joined #maemo-ssu | 17:04 | |
*** chainsawbike has quit IRC | 17:16 | |
*** chainsawbike has joined #maemo-ssu | 17:18 | |
*** amiconn_ has joined #maemo-ssu | 18:05 | |
*** amiconn has quit IRC | 18:05 | |
*** amiconn_ is now known as amiconn | 18:05 | |
freemangordon | luf: http://pastebin.com/jUckH5qD | 18:05 |
*** toxaris has joined #maemo-ssu | 18:06 | |
*** toxaris has quit IRC | 18:28 | |
freemangordon | why Qt refuses to build? | 18:33 |
freemangordon | nice, if you have some escaped char in the build path ('+' for example) you can't build Qt :D | 18:48 |
*** MrPingu has quit IRC | 18:53 | |
*** NIN101 has joined #maemo-ssu | 18:55 | |
*** wumpwoast has joined #maemo-ssu | 19:02 | |
*** arcean has quit IRC | 19:47 | |
*** freemangordon has quit IRC | 19:51 | |
*** freemangordon has joined #maemo-ssu | 19:51 | |
*** arcean has joined #maemo-ssu | 20:08 | |
*** arcean_ has joined #maemo-ssu | 20:14 | |
*** arcean has quit IRC | 20:14 | |
*** MadeAllUp has joined #maemo-ssu | 20:19 | |
*** arcean_ is now known as arcean | 20:42 | |
*** MrPingu has joined #maemo-ssu | 21:03 | |
*** NIN101 has quit IRC | 21:04 | |
*** NIN102 has joined #maemo-ssu | 21:05 | |
*** Martix has joined #maemo-ssu | 21:20 | |
*** grummund has joined #maemo-ssu | 21:50 | |
*** dafox has joined #maemo-ssu | 22:14 | |
*** lizardo has quit IRC | 22:57 | |
*** LinuxCode has joined #maemo-ssu | 23:00 | |
*** wirr has left #maemo-ssu | 23:11 | |
*** MrPingu has quit IRC | 23:22 | |
*** NIN102 has quit IRC | 23:43 | |
*** xes has joined #maemo-ssu | 23:59 |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!