**** BEGIN LOGGING AT Sun Aug 16 02:59:57 2009 Aug 16 04:12:41 openmoko: 03werner * r5459 10/trunk/eda/fped/ (10 files): Aug 16 04:12:41 openmoko: Columns and loops can now be entered in one step as var = val, val, ... Aug 16 04:12:41 openmoko: - moved definition of expr_result from expr.c to fpd.y Aug 16 04:12:41 openmoko: - new header file fpd.h with all the things fpd.l and fpd.y export Aug 16 04:12:41 openmoko: - edit_var already calls edit_nothing, so there's no need to call it before Aug 16 04:12:44 openmoko: - added rapid entry option for loops, variables, and columns: var = val, ... Aug 16 04:28:19 Hi... anyone knows if Debian install script is working know? Aug 16 05:05:20 openmoko: 03werner * r5460 10/trunk/eda/fped/ (8 files): Aug 16 05:05:20 openmoko: - when editing a value list, we need to return the values and free them in case Aug 16 05:05:20 openmoko: of successful validation Aug 16 05:05:20 openmoko: - free previous expression in set_col_values Aug 16 05:05:20 openmoko: - added rapid entry also for table values (fill the row) Aug 16 07:44:13 no WS afer one night's sleep with 2.6.29; that's a first, so I guess my patch had the desired effect. Aug 16 07:46:19 now I need a better measurement on how long this kernel lasts in suspend, but it seems the diff beween jbt in sleep vs deep stand by is not so big. Aug 16 07:54:23 how do I stop SHR from mounting all the partitions on my uSD card? Aug 16 07:58:04 /etc/udev/mount.blacklist Aug 16 08:00:02 * pabs3 disables /etc/udev/scripts/mount.sh instead Aug 16 08:12:59 removing mountpoints also works; didn't know udev could do/trigger the mounts, too. Aug 16 08:20:26 Kero: did you just disabled deep sleep mode for jbt and it fixed the WS issue for you? Aug 16 08:22:08 yeah. found a comment to that effect in the stable/2.6.24 kernel source. Aug 16 08:23:09 doing a `git lop -p` shows a lot of, let's say, confusion, about jbt. Aug 16 08:31:53 PaulFertser: o/ Aug 16 08:37:54 PaulFertser: thnx. removed gsm07muxd and fso-abyss started to work. not sure if mickeyterm will work... 8( Aug 16 08:39:27 Q-Master: why shouldn't it work? Aug 16 08:39:39 DocScrutinizer: yo Aug 16 08:40:18 PaulFertser: seems it depends on mickeyterm Aug 16 08:40:26 PaulFertser: seems it depends on gsm07muxd Aug 16 08:40:27 8) Aug 16 08:40:38 Q-Master: modern mickeyterm supports both, so np Aug 16 08:40:44 ok Aug 16 08:41:42 Q-Master: why should mickeyterm not work? Aug 16 08:42:06 aah, late :-/ Aug 16 08:42:14 nevermind then Aug 16 08:44:45 PaulFertser: I'll check more in the upcoming 24h :) Aug 16 08:45:34 and I'm open to other suggestions of changing the kernel; once done, cross compiling again is easy. Aug 16 09:06:20 too bad, GPRS connect+disconnect got somehing very upset. symptoms: settings app will wait infinitely for its modules to load and incoming call is not shown Aug 16 09:08:49 can't find the incoming call in frameworkd.log. hm. Aug 16 09:10:11 Kerba: known issue here Aug 16 09:10:32 DocScrutinizer: it worked prev time I tried :) Aug 16 09:10:40 frameworkd stalls for me on inbound call Aug 16 09:11:03 that is - during long call Aug 16 09:11:18 due to pppd timeout we suspect Aug 16 09:11:34 * Kero needs a way to run his restart_gsm script from GUI; as it restarts GUI atm. Aug 16 09:12:01 it wasn't very long. just connect+disconnect, actually. Aug 16 09:12:41 the call? Aug 16 09:12:54 interesting Aug 16 09:13:49 the GPRS connection was short. incoming call was entirely invisible. Aug 16 09:13:55 see if you can send any more mdbus -s commands (to frameworkd) Aug 16 09:14:08 anyone alive that knows where shr tells X to load illume? I want to use icewm instead and just placing a .Xsession file on ~/ does nothing Aug 16 09:14:23 oooh, that's definitely different failure pattern Aug 16 09:14:47 DocScrutinizer: I guess the haning "loading modules" in settings shows I could not? (restarted already) Aug 16 09:15:02 yes Aug 16 09:15:55 Kero: you should have saved the logs Aug 16 09:19:47 GSM.PDP.ContextStatus: 1 release {} / GSM.PDP.NetworkStatus: {'registration': 'unregistered'} and that last line is literally the last in the log before restart Aug 16 09:20:12 restarted framework and some other parts. did not reboot the phone ;) Aug 16 09:23:10 Kero: I guess you should tell FSO-crew about it. Preferrably mickeyl Aug 16 09:27:52 Crap, my gta01 battery is dead. Swollen to double size :( Aug 16 09:29:42 ouch Aug 16 09:30:45 Just had it plugged in via USB as usually Aug 16 09:31:08 I know there's a nokia battery that kind of works, but is there some better way? Aug 16 09:32:18 for gta01 the BL-5C isn't that much different Aug 16 09:33:10 CM: ehm Aug 16 09:34:16 CM: well, the best battery would be BL-6C for you. But... gta01 currently sets pcf50606 to control temperature of gta01 battery, bl-6c is incompatible with that setting. You'll have to charge the battery separately. Or i can prepare a custom ad-hoc patch fast or a proper patch slow to enable charging of bl-6c Aug 16 09:34:17 PaulFertser: possible result of restating MBC every 15min Aug 16 09:34:38 ok Aug 16 09:34:40 DocScrutinizer: the mbc restart was never implemented for pcf50606 i guess. Aug 16 09:35:01 no idea here Aug 16 09:35:37 Or, does someone want an extra gta01? Aug 16 09:36:19 I haven't spent enough time on it lately anyway. Aug 16 09:36:30 CM: where're you located? Aug 16 09:36:37 PaulFertser: Sweden, Stockholm Aug 16 09:36:53 You? Do you want it? Aug 16 09:37:05 I'll just mail it if you do Aug 16 09:38:05 PaulFertser: \o/ finally :D Aug 16 09:39:39 CM: PaulFertser is highly interested in getting a 01, to help other people with their 01 Aug 16 09:40:48 DocScrutinizer: I'll mail it to him :) Aug 16 09:41:13 I'm really glad to hear that great news Aug 16 09:43:39 PaulFertser: we urgently need an amendment to bat faq "never ever store a discharged bat for prolonged time. never leave bat inside device for weeks while device not in use" Aug 16 09:43:54 DocScrutinizer: right Aug 16 09:44:34 * DocScrutinizer checks his half a dozen devices :-P Aug 16 09:44:41 is that only for GTA01? Aug 16 09:45:40 nope, that holds true for *all* devices, no matter whether OM, Nokia, Samsung whatever Aug 16 09:47:03 though for 01 it's a particularly nasty situation due to power-down issues (might fail partially) Aug 16 10:51:38 how would I debug why uboot doesn't advertise DFU or a tty over USB? Aug 16 10:52:05 (the NAND uboot that is, the NOR one is fine) Aug 16 10:55:12 actually it doesn't advertise itself at all, even lsusb doesn't see it Aug 16 10:55:46 I'm not sure NAND is supposed to Aug 16 10:55:54 all that happens in NOR I think Aug 16 10:57:34 is there an environment option for it? or a way to get the NOR environment? Aug 16 10:58:05 you can probably read it, but you can't write to it Aug 16 10:58:38 sure, I wouldn't want to write to it ;) Aug 16 11:02:38 pabs3: nand u-boot should be visible over usb too. Aug 16 11:03:03 pabs3: you can read nor environment with "printenv" Aug 16 11:03:05 I just tried mine and didn't get it either Aug 16 11:03:19 nor works, nand doesnt Aug 16 11:03:37 oh duh Aug 16 11:03:43 nand on mine is qi Aug 16 11:03:48 ndnihil: probably one should select "activate usb console" in the menu or something. Aug 16 11:03:48 nevermind :) Aug 16 11:04:25 ndnihil: good boy ;) I'm using Qi from the beginning too :) Aug 16 11:04:39 I'm still a uboot fan Aug 16 11:05:00 gave qi a shot and haven't had a reason to switch back Aug 16 11:06:07 * pabs3 has multiple distros on uSD so needs uboot Aug 16 11:06:08 DocScrutinizer: btw, battery q&a already has a note how to properly store batteries. Aug 16 11:06:37 oooh, was about to type that next :-P Aug 16 11:06:38 hah, finally all .mid converted to .wav :) Aug 16 11:06:43 need to look more closely Aug 16 11:06:53 §? Aug 16 11:06:55 I can use my ringtones on the FR again :) Aug 16 11:07:22 DocScrutinizer: http://wiki.openmoko.org/wiki/Battery_Questions_and_Answers#I_have_several_compatible_batteries._What_are_the_storage_requirements_for_them.3F Aug 16 11:08:25 He, so anybody wants to tell me if his *bluetooth* keyboard or mouse works with modern distros if using updated instructions? Aug 16 11:08:59 PaulFertser: yo, good you pointed me to that just in time :-D Aug 16 11:13:24 DocScrutinizer: anyway should be clarified that inside the device nasty things (==deep discharge) can happen. Aug 16 11:14:31 http://wiki.openmoko.org/wiki/Battery_Questions_and_Answers#Battery_care_and_safety_guidelines Aug 16 11:15:41 PaulFertser: deep discharge in itself isn't that nasty. Not recharging for weeks after reaching that state is though Aug 16 11:15:54 my battery applet says 66% when the phone says battery is full :( Aug 16 11:16:07 blindcoder: yawn Aug 16 11:16:20 * blindcoder currently reading th ewiki page though ;) Aug 16 11:16:36 DocScrutinizer: s/Neo/gta02/ as we now know that gta01 won't restart charging Aug 16 11:16:42 blindcoder: please select theme-illume-shr when first time booting SHR freshly flashed device Aug 16 11:16:57 PaulFertser: true Aug 16 11:16:59 DocScrutinizer: I did. Aug 16 11:18:06 blindcoder: don't use HAL for battery readings, use internal method or disable hald. Aug 16 11:18:38 wtf Aug 16 11:18:38 my fr wont boot Aug 16 11:18:44 PaulFertser: ooops OT on community channel, no? Aug 16 11:19:16 wont boot without usb plugged in Aug 16 11:19:31 blindcoder: [2009-08-15 23:35:57] budfive_: switch to internal mode in hardware in advanced in batmeter in advanced in illume-wrench Aug 16 11:19:56 blindcoder: then restart X or reboot Aug 16 11:20:43 blindcoder: [2009-08-15 23:37:45] budfive_: or simply stop hald, as nobody seems to know whyt for we need it anyway Aug 16 11:21:12 I see, it was on auto-detect here Aug 16 11:21:52 92% now :-) Aug 16 11:21:52 openmoko: 03werner * r5461 10/trunk/eda/fped/ (13 files): (log message trimmed) Aug 16 11:21:52 openmoko: Added Postscript outout for measurements, plus assorted changes. Aug 16 11:21:52 openmoko: - postscript.c: added patterns backhatchpath and crosspath (for pad overlaid Aug 16 11:21:52 openmoko: with solder paste) Aug 16 11:21:52 openmoko: - postscript.c: added measurements Aug 16 11:21:53 openmoko: - postscript.c: prefixed style parameters with PS_ to avoid future conflicts Aug 16 11:21:55 openmoko: with gui_style.h Aug 16 11:21:58 * blindcoder .oO( if actually charging was that easy ) Aug 16 11:22:20 battery is dead Aug 16 11:22:22 for some reason Aug 16 11:22:27 it has been plugged in all night Aug 16 11:22:52 I guess it had just enough for it to boot with a boost from usb Aug 16 11:23:28 ndnihil: gta02 plugged in all night and the battery is dead? :-O Aug 16 11:23:33 yep Aug 16 11:24:10 ndnihil: plugged in to WHAT? Aug 16 11:24:37 workstation Aug 16 11:24:40 like it usually is Aug 16 11:24:47 HMMM Aug 16 11:25:10 PaulFertser: all the uboot usb stuff is the same between NOR & NAND. I notice this in dmesg with NAND uboot: [298891.664055] usb 3-1: device not accepting address 35, error -71 Aug 16 11:25:23 so I would blame my WS for powering down / suspending the USB host connector Aug 16 11:26:35 pabs3: strange. Are you running the latest u-boot in NAND? Aug 16 11:26:44 ndnihil: especially if FR goes to suspend, the host might decide to stop power to USB Aug 16 11:26:59 it doesn't normally suspend when plugged in Aug 16 11:27:32 ndnihil: but of course also a WS suspending itself may stop powering any USB device Aug 16 11:28:00 this thing has never suspended in its life Aug 16 11:28:11 PaulFertser: yes, tried several versions. currently u-boot-gta02v5-1.3.1+gitr650149a53dbdd48bf6dfef90930c8ab182adb512-r1.bin Aug 16 11:29:03 ndnihil: well anyway, CND, sorry Aug 16 11:29:16 I'm going to assume I locked it up playing with wireless again Aug 16 11:29:36 the phone that is Aug 16 11:29:49 that is the major cause of lockups for me Aug 16 11:30:23 that's also a possible reason for WS not to see anymore response from USB data level thus powering down USB-VBUS Aug 16 11:30:59 pabs3: you need to ask Rask about it, he says he cares about u-boot. Aug 16 11:31:25 ooh, really? Aug 16 11:31:28 great news Aug 16 11:32:10 so he might fix a few real nasties in uboot, as werner refuses to touch it Aug 16 11:32:22 He might Aug 16 11:33:53 PaulFertser: we have a fat bug wrt 256b-inodes Aug 16 11:34:07 (ext2 that is) Aug 16 11:34:29 AIUI Aug 16 11:36:29 ndnihil: (WIFI) yeah segfaults frequently for me as well :-( Aug 16 11:37:28 locks up the whole device for me Aug 16 11:37:32 have to pop the battery out Aug 16 11:37:51 not every time, but probably 2 out of 3 times I try to use WPA Aug 16 11:37:57 err s/segfaults/PANICs/ Aug 16 11:38:40 so yes, same her (though on WEP) Aug 16 11:39:01 open APs are no problem Aug 16 11:39:05 connects right up Aug 16 11:39:07 does its thing Aug 16 11:39:32 but mine is WPA2 with an obnoxiously long key, and it doesn't go anywhere unless you run VPN on top of that :) Aug 16 11:40:27 without VPN everything is blocked, except for all web traffic which gets redirected to a blinking page that makes fun of them for wasting time dicking around with my AP Aug 16 11:40:56 can't argue with paranoia ;) Aug 16 11:41:23 I just created my key by banging my head on the keyboard. it sure beat doing 'strings /dev/random' ;) Aug 16 11:42:06 mine is 55 random characters Aug 16 11:42:07 anyone can recommend an rss reader? Aug 16 11:42:31 Defiant-: for the FR or in general? Aug 16 11:42:44 blindcoder: for the freerunner...should be able to work offline Aug 16 11:43:04 then no, sorry. Aug 16 11:43:16 haven't got a data plan yet, so also no idea about that Aug 16 11:44:32 rssyl plugin for claws Aug 16 11:44:36 hi Aug 16 11:44:48 its in the repos Aug 16 11:45:45 I'd assume it works with the offline mode of claws itself Aug 16 11:46:05 ndnihil: hmm, does it support update rfom commandline, so I can do ssh gta02 "update-rss" Aug 16 11:46:11 *from Aug 16 11:46:28 dunno, never used it myself Aug 16 11:46:49 grep -i rss pkgs.txt Aug 16 11:46:50 :) Aug 16 11:47:19 (every time I opkg update, I opkg list > pkgs.txt, it's faster to grep that than it is to opkg search) Aug 16 11:52:44 Hi. I've been away for a month, vacation. And I've discovered that my SHR build is quite unstable. Missing messages etc. How is the current one? Aug 16 11:53:09 * DocScrutinizer wonders if opkg doesn't do exactly the same on "update" Aug 16 11:59:14 okay, maybe you'll massacre me one day or another for asking these questions, but still. Aug 16 11:59:19 DocScrutinizer: not conveniently in ~, and I figured it would be a bit more complex than a simple list of packages Aug 16 11:59:54 I got my FR from my employer, asking specifically for a V07. uBoot and /proc/cpuinfo say 06 though. Date code is march 2009. Aug 16 12:00:10 has the display of the revision just NOT changed or did the order go wrong after all? Aug 16 12:00:34 back up to the first part Aug 16 12:00:39 how did you get your employer to buy you a FR? Aug 16 12:00:46 ndnihil: asking? Aug 16 12:00:49 I had to buy mine myself Aug 16 12:00:58 they keep sending me blackberries Aug 16 12:01:09 I lost my blackberry Aug 16 12:02:10 I told them that I lost my Blackberry and if I should organise a new phone myself Aug 16 12:02:26 they said no, they'll take care of it and if I had any special wish for a phone Aug 16 12:02:59 nice Aug 16 12:03:02 I showed them the webpage of a reseller, said gta02v7 with a spare battery and a few days later it came with the mail Aug 16 12:03:08 as for how to tell revision Aug 16 12:03:14 its silkscreened on the PCB Aug 16 12:03:31 yeah, too bad my smallest torx is a t8 :( Aug 16 12:03:43 flatheads work Aug 16 12:03:45 can't open the screws Aug 16 12:03:48 just need the right size Aug 16 12:03:52 or grind one down to fit Aug 16 12:03:56 ndnihil: yes, proc/cpuinfo shows the old stuff Aug 16 12:04:04 openmoko: 03werner * r5462 10/trunk/eda/fped/ (6 files): Aug 16 12:04:04 openmoko: - adjust measurement bbox for projection (still not perfect, but better) Aug 16 12:04:04 openmoko: - we can now scroll-wheel through table rows Aug 16 12:04:04 openmoko: - added menu branch "View" with zoom and var/code display controls Aug 16 12:04:04 openmoko: - commented out is_min_of_next until its fate is decided, so that we don't get Aug 16 12:04:07 openmoko: a warning Aug 16 12:04:11 flathead? you know the german word for it? Aug 16 12:04:18 blindcoder: no need Aug 16 12:04:19 hrm, a flat bladed screwdriver Aug 16 12:04:21 blindcoder: A7 has components and empty footprints at base of uSD card. Thus you can tell apart from A6 easily Aug 16 12:04:24 blindcoder: let me show you another node Aug 16 12:04:54 blindcoder: cat /sys/devices/platform/neo1973-version.0/pcb Aug 16 12:05:10 I'd imagine the A7 is missing the cap across the pins of the uSD slot Aug 16 12:05:33 blindcoder: 0x101 for A7, 0x001 for A6 Aug 16 12:06:06 neat, didn't know about that one Aug 16 12:11:09 ndnihil: I don't have any MP A7 for actually checking, but afaik (and from PV devices A7) there *is* this cap Aug 16 12:12:36 oh really? I figured they would have fixed that Aug 16 12:12:40 PaulFertser: reads 0x101, thanks :) Aug 16 12:12:48 well, fixed/found a proper place for it Aug 16 12:13:15 ndnihil: what do you mean with "fixed" ?? the caps *are* the fix Aug 16 12:13:35 well yeah, but do they really belong soldered across the pins of the uSD carrier? Aug 16 12:14:10 actually there are 6 empty footprints and 3 caps on A7, wher there is nothing on A6 Aug 16 12:15:10 ndnihil: on A7 they aren't soldered across the uSD-holder contact pads Aug 16 12:15:37 ah, so they did find a place for it Aug 16 12:17:49 ndnihil: as I stated above, they are placed on the empty pcb real estate that can be found on A6 at base of uSD holder Aug 16 12:18:19 gotcha, I was just thinking that checking for the presence of the cap across the pins would be a way to differentiate between the a6 and a7 Aug 16 12:18:34 though I don't know if all a6's have them Aug 16 12:18:46 or if it is present in a5 and prior Aug 16 12:18:55 no, most A6 actually don't have Aug 16 12:19:13 afaik Aug 16 12:22:44 and anyway the A6-"fix" is actually useless, as it operates on sdclk only, and we switch that clock down while no access to uSD, this patch went to kernel long ago. While accessing uSD you got noise from datalines as well, so the A6-"fix" won't help there Aug 16 12:24:06 openmoko: 03werner * r5463 10/trunk/eda/fped/kicad.c: Aug 16 12:24:06 openmoko: - ctime(3) already provides a newline, so don't add an extra one to the .mod Aug 16 12:24:06 openmoko: file, possibly confusing KiCad Aug 16 12:28:54 Wasn't there a drive strength fix also - which will Aug 16 12:28:58 (help) Aug 16 12:29:29 * ndnihil solders a bunch of caps across the datalines Aug 16 12:29:57 openmoko: 03werner * r5464 10/trunk/gta02-core/ (Makefile gta02-core.pro modules/mkloe): Aug 16 12:29:57 openmoko: - modules/mkloe: script to generate a unified library from the .fpd files Aug 16 12:29:57 openmoko: - gta02-core.pro: removed stock libraries and added gta02-core.mod instead Aug 16 12:29:57 openmoko: - Makefile: run modules/mkloe on "make update" Aug 16 12:29:57 openmoko: - Makefile: new target "brd" to invoke pcbnew Aug 16 12:30:04 SpeedEvil: yup Aug 16 12:32:07 DocScrutinizer: my MP A6 has a cap acress sd pins. Aug 16 12:32:48 SpeedEvil: also the default value of drive strength is known to cause problems sometimes. Aug 16 12:33:17 with corruption? Aug 16 12:33:28 PaulFertser: yup, fab did a rework on a bunch of devices, until we realized the clock and drivestrength solution Aug 16 12:33:31 the new default you mean after drive strength reduction? Aug 16 12:33:47 SpeedEvil: yes Aug 16 12:34:26 iirc I recommended that back when, as removing a few caps in case they really hurt is much simpler than retrofitting them Aug 16 12:36:13 * PaulFertser wonders why the fuck do we all put up with cryptic DSDT acpi shit. Why no laptop vendor ever documents their hardware properly and instead provides buggy and weird DSDT without any comments or hints whatsoever? Aug 16 12:36:39 PaulFertser: the diffs in these caps being there or not may be the cause for some devices don't work with some cards Aug 16 12:37:14 DocScrutinizer: these caps == the one one the clock line in some A6s? Aug 16 12:37:22 yep Aug 16 12:37:44 rsp the few ones on A7 (if they actually are there for MP) Aug 16 12:38:35 DocScrutinizer: btw, why reducing clock is supposed to help GPS anyhow? What can we do now to understand if 8x slowdown when gps is powered is really justified or not? Aug 16 12:38:36 PaulFertser: works with windows/thatyear - shipit Aug 16 12:38:51 * SpeedEvil holds up a spectrum analyser. Aug 16 12:39:13 PaulFertser: I never completely understood the rationale behind that Aug 16 12:39:20 SpeedEvil: but quite some laptop manufacturers say they care about free software and ship with GNU/Linux preinstalled. Aug 16 12:39:22 Though to a limited extent - the signal reported by the GPS would be sane. Aug 16 12:39:59 compare signal reported with different GPS strengths. Aug 16 12:40:00 Also. Aug 16 12:40:36 A pretty good solution would be to decrease the signal level to GPS until it lost lock under various conditions. But this is tricky Aug 16 12:40:38 PaulFertser: is there more than Dell+Ubuntu and Lemote+Debian? Aug 16 12:40:40 DocScrutinizer: because as far as i understand it's drive strength that matters, be it 2MHz or 16Mhz is the same for ~2GHz signal. Aug 16 12:40:49 freesmartphone.org: 03mickey 07libeflvala * rbd9ceddd6878 10/vapi/eina.vapi: eina.vapi: catch up with API changes as of upstream rev 41187 Aug 16 12:41:08 pabs3: i thought yes, esp. in netbook market. Aug 16 12:41:30 A sharp rising (or falling) edge creates a broadband noise pulse - up to a bit over 1/risetime in frequency. Aug 16 12:41:33 PaulFertser: yes, especially taking into account it's a square wave anyway, no matter which clkfreq Aug 16 12:41:45 PaulFertser: oh, right. plus Android & Ubuntu in the newer arm netbooks Aug 16 12:41:52 If you have multiple edges at an evenly spaced period - some of the interference cancels. Aug 16 12:41:54 * pabs3 wasn't thinking of netbooks Aug 16 12:42:52 fsck 33°C Aug 16 12:43:00 openmoko: 03werner * r5465 10/trunk/gta02-core/docs/GETTING-STARTED: Aug 16 12:43:00 openmoko: - added GETTING-STARTED instructions for installing fped and for accessing Aug 16 12:43:00 openmoko: footprints Aug 16 12:43:15 * DocScrutinizer away for beergarden ;-) Aug 16 12:44:23 SpeedEvil: so you want to say that freq matters here? Aug 16 12:44:32 PaulFertser: how to permanently disable that clk/8 nuissance? Aug 16 12:44:43 DocScrutinizer: I'll trade you for 12°C Aug 16 12:45:11 DocScrutinizer: runtime parameter in /sys/module/glamo_mci/ or with kernel command line. Aug 16 12:45:37 freesmartphone.org: 03mickey 07libeflvala * r0b07a3e4bf89 10/vapi/elm.vapi: elm.vapi: add new API calls as of upstream rev 41721 Aug 16 12:45:50 shitty cmdline. hmm might consider adding an initscript line to boot Aug 16 12:46:10 PaulFertser: the spectrum of a repeating edge looks very different to a single edge. I'm unsure. There is of course the obvious fact that 1/8th as fast a clock is 1/8th as many clocks per unit time is 1/8th the total interfering energy Aug 16 12:46:19 I would need a FR to test with. Aug 16 12:47:45 SpeedEvil: (total energy) yes, sure. Aug 16 12:48:28 freesmartphone.org: 03mickey 07libeflvala * rfae4f37aed11 10/vapi/elm.vapi: elm.vapi: add additional API as of upstream rev 41746 Aug 16 12:51:49 (and time and energy to do the tests) Aug 16 12:58:42 openmoko: 03alvieboy * r5466 10/trunk/gta02-core/gps.sch: Fix footprint names Aug 16 12:59:16 freesmartphone.org: 03mickey 07cornucopia * r12a90f9bd0a5 10/tools/apm2/ (14 files in 2 dirs): Aug 16 12:59:16 freesmartphone.org: tools: add apm2 build skeleton. apm2 is the Vala implementation of fso-apm, Aug 16 12:59:16 freesmartphone.org: which in turn is the FSO compatibility utility for the apm legacy. Aug 16 13:01:11 grah, SHR doesn't like my laptop's USB until it gets a suspend/resume cycle Aug 16 13:08:08 freesmartphone.org: 03mickey 07libeflvala * r9c0055f8128b 10/vapi/elm.vapi: elm.vapi: add new Elm.ProgressBar object Aug 16 13:13:06 DocScrutinizer: my problem may be this https://docs.openmoko.org/trac/ticket/2298 but I'd have to run with debug output to confirm; running suspend test, though. so GPRS test is not going to be anytime soon. Aug 16 13:16:22 openmoko: 03alvieboy * r5467 10/trunk/gta02-core/ (docs/ecn/STATUS pmu.sch): Apply ECN0003 to ECN0007 Aug 16 13:18:35 freesmartphone.org: 03mickey 07cornucopia * r8f0d581efe7f 10/tools/apm2/src/main.vala: tools: apm2: call ousaged to implement -s option Aug 16 13:27:44 is it possible to have multiple USB gadgets loaded into the kernel at once? Aug 16 13:41:10 freesmartphone.org: 03mickey 07cornucopia * rb65a3f29ffa5 10/tools/apm2/src/main.vala: tools: apm: show power status Aug 16 13:42:12 pabs3: no, you need to unload one and use other Aug 16 13:43:01 :/ Aug 16 13:43:56 do we need hardware support for multigadget? Aug 16 13:44:02 like... more endpoints? Aug 16 13:45:03 SpeedEvil: can you give me a small EE hint? I've a small impulse power supply unit here (for siemens dect phone) it was supposedly subject to 380 AC and as a result a high voltage suppressor (?) blew and a pcb trace acted as a fuse. I removed the suppressor and shorted the trace with ~2A wire, do you think it can damage the equipment connected to it if it's exposed to high voltage again? Aug 16 13:45:15 mickeyl: i'm sorry, no idea. Probably mwester knows more. Aug 16 13:45:23 k Aug 16 13:51:07 PaulFertser: you probably can Aug 16 13:54:54 PaulFertser: but that's not that different to what it was before ;-) Aug 16 13:58:12 mickeyl: any luck with the logs I sent you about a week ago? Aug 16 13:58:21 DocScrutinizer: before the suppressor blew, shorted and pcb trace protected everything else so the supply is still functional. Since i'm not going to find any suppressor to fit there i'm afraid that without it can generate high voltage (>10V) on the output. I'm not exactly sure how that is possible (most probably it's not), but wanted to ask just in case. Aug 16 13:58:52 Kero: i did not have time to look at it yet, sorry. it's on my pile Aug 16 13:59:27 PaulFertser: the OVP suppressor isn't any waranty for device not producing OV Aug 16 14:00:29 DocScrutinizer: and yet it helped :) Aug 16 14:00:53 it blew, you don't know if it helped Aug 16 14:01:16 wrt stopping OV at output Aug 16 14:02:35 PaulFertser: see, for a OVP to blow, it actually needs OV at input. Aug 16 14:03:03 DocScrutinizer: measured it, 300 Ohm currently, so i guess it shorted the input, then pcb trace acted as a fuse. No chance for OV at output. Aug 16 14:03:26 so if device doesn't protect output, you got no protection from OVP at input Aug 16 14:03:32 DocScrutinizer: i say i guess the supply unit was subject to 380AC in the input. Aug 16 14:04:11 mickeyl: by Tue I can no longer check extra things, as the SIM will stop working. Aug 16 14:04:29 DocScrutinizer: just a primitive impulse power supply for an ordinary DECT phone. The power supply is ok after shorting the blown trace, so is the DECT phone. Aug 16 14:04:34 during the time OVP blew, device needed to cope with OV at input anyway Aug 16 14:05:25 maybe not as much as it would have been without that OVP Aug 16 14:05:56 as the fused trace acted as a R-divider with OVP Aug 16 14:06:10 DocScrutinizer: there's an L in series, i guess it helped too, so OVP safely shorted the input until the pcb "fuse" blew. Aug 16 14:08:03 my staement is: input OVP is to protect the device itself from surges. For output OVP you need other means. If a device is poorly built, you have that problem anyway. No matter if there's input OVP or not Aug 16 14:08:35 DocScrutinizer: ah, with that i fully agree :) Aug 16 14:08:53 if the device is "good" missing imput OVP prolly wont care Aug 16 14:11:46 1.2MB read uSD, without GPS as well as with GPS+slow-ratio=1. with slow-ratio=8 (default) it drops to half (~680) Aug 16 14:12:54 It definetely doesn't look like ordinary chinese crap. There's something that looks like a zener diod exactly on the output. Aug 16 14:13:13 DocScrutinizer: (speed) not surprising :D Aug 16 14:14:50 on starting dd on A7 at fullspeed, sat signal drops ~3..4dB Aug 16 14:15:02 negligable Aug 16 14:16:02 DocScrutinizer: i remember i tried to get some data about sd drive strength influence based on signal level as reported by the chip. The readings were floating too much anyway, so i failed to prove anything. Aug 16 14:16:06 so for sure I do *not* want slow-ratio 8 Aug 16 14:16:50 heh, GPS is US. Where you're livin? Aug 16 14:17:00 ;-) Aug 16 14:17:16 steady signal on 3..6 sat here Aug 16 14:18:31 wow even 1.9Mb this time Aug 16 14:19:48 I really consider it a BAD thing this ratio thing is done in kernel-space at all Aug 16 14:23:07 anyway, my official recommendation (at least for A7): "echo 1 >/sys/*/*/*/sd_slow_ratio" on boot-time Aug 16 14:23:44 Hm... What does it do again? In short? Aug 16 14:24:57 Faster reading? Or does it include writing as well? Aug 16 14:25:57 especially since the default 8 doesn't make *any* difference - still ~4dB drop, maybe even 5 ;-P Aug 16 14:27:46 azaghal: all uSD access Aug 16 14:29:03 DocScrutinizer: is your new apartment equipped with residual current device? Aug 16 14:29:28 err what's that? Aug 16 14:30:24 DocScrutinizer: http://en.wikipedia.org/wiki/Residual-current_device breaks circuits to protect from electric shock Aug 16 14:30:37 aah, sure Aug 16 14:30:55 mandatory (at least for bathroom) Aug 16 14:31:37 DocScrutinizer: WP says it's mandatory in Germany for all outlets. Aug 16 14:31:59 that's actually not true Aug 16 14:33:52 might be true for new installations, but for sure not for rework of old buildings like my appartment Aug 16 14:34:11 according to my electrician Aug 16 14:34:24 DocScrutinizer: my house was built around 1985 and the bathroom was connected through 1:1 transformer. But for an unknown reason our flat had the transformer in place but not connected at all, so no safety here. And the flats where it's connected had it emit quite audible noise :D Aug 16 14:35:05 hehehe, sounds like TPE 10th floor appartment Aug 16 14:35:57 though that transformer was to 110->220V for lights only. No idea what that nonsense is all about Aug 16 14:37:16 "However the most commonly used in the USA since the early 1980s are built into outlet receptacles. " :-O Aug 16 14:37:28 Never saw anything like that here. Aug 16 14:37:33 yeeha, settings segfaulted once again :-/ Aug 16 14:40:17 and fw_readenv still produces weird scary output Aug 16 14:40:58 need to swap to a device with clean u-boot mtdparts (hooray for uSD-/) Aug 16 14:44:41 nah, USB enumerates (as I booted from a obviously sane env setting for "boot from uSD", rather than my tweaked default bootcmd), so mtdparts seem to be ok Aug 16 14:45:34 and now back to cdevel before I get kicked for being totally OT here ;-D Aug 16 14:45:51 Any thoughts on why the SD isn't much faster? Aug 16 14:46:05 Surely it should approach the 7M/s of the glamo interface? Aug 16 14:46:13 Is this just 'glamo sucks' ? Aug 16 14:47:43 SpeedEvil: yes Aug 16 14:47:56 SpeedEvil: shares bus bandwidth with FB etc. Aug 16 14:48:21 SpeedEvil: and some other nastyness. .31 with cool new patches will bring some improvement but not much. Aug 16 14:48:28 Sure - but your above raw speed tests should surely approach 7M, not 1.3M? Aug 16 14:48:42 With FB idle Aug 16 14:49:15 SpeedEvil: with screen blanked larsc got +30% with _his_ patches. DocScrutinizer isn't using those yet. And still the maximum was ~4M iirc. Aug 16 14:50:13 Hmm. Aug 16 14:50:31 * SpeedEvil needs to get to doing stuff on the neo again. But first - more gardening. Aug 16 14:50:34 Wave. Aug 16 14:50:55 SpeedEvil: o/ Aug 16 14:53:12 Gardening... He's mowing the lawn probably :) Aug 16 15:11:43 Anyone used icecap+irssi2? Is it any better than bip+irssi? Aug 16 15:20:27 freesmartphone.org: 03mickey 07cornucopia * ra23eecf531dc 10/fsodeviced/src/plugins/kernel26_powersupply/plugin.vala: fsodeviced: kernel26_powersupply: synthesize coldplug, if we're on a mounted sysfs Aug 16 15:33:33 http://en.wikipedia.org/wiki/File:SanDisk_128MB_SD_Card_Insides.png :) Aug 16 15:35:50 SpeedEvil: :) Aug 16 15:44:35 freesmartphone.org: 03mickey 07cornucopia * rbc50f1627d32 10/fsousaged/src/plugins/ (3 files in 3 dirs): fsousaged: for some reason, my constructors were not called. use 'construct' to work around for now Aug 16 15:49:35 * DocScrutinizer considers to check wth irsii might be :-P Aug 16 15:49:59 irssi even Aug 16 16:07:55 <|Marco|> DocScrutinizer: ? Aug 16 16:08:32 just kidding, to let you know I never tested it Aug 16 16:32:28 No, I'm developing a new and innovative form of mapping. Aug 16 16:34:16