**** BEGIN LOGGING AT Sun Nov 10 02:59:58 2013 Nov 10 03:21:04 what happened to git.freesmartphone.org? Nov 10 03:21:35 hmm, all of freesmartphone.org seems down Nov 10 03:30:02 anyone know who the admin is? Nov 10 07:18:30 Harald Nov 10 08:04:03 pabs3: ah, freesmartphone. That would be mickey Nov 10 09:46:52 is xf86-input-tslib or evdev used by SHR on gta02? Nov 10 09:48:40 o/ Nov 10 10:03:51 pabs3: no idea but I've tried both on gta02 and gta04 and saved some notes in my xmonad-on-smartphone repository if you're interested. Nov 10 10:05:27 zumbi: https://gitorious.org/xmonad-smartphone-config/xmonad-smartphone-config/source/aa36011d09c17d2fef0c40fd8d5451f27110fb59:README.org Nov 10 10:05:53 "Tslib driver supports advanced filtering" "Evdev driver works out of the box but needs calibration" Nov 10 10:07:36 PaulFertser: still using that setup? Nov 10 10:07:51 Yes, and there're some specific versions that work and some that do not, I think there're notes regarding that in the apt.conf pin files, probably outdated. Nov 10 10:08:02 pabs3: i'm still using my gta02 but gta04 is only for testing. Nov 10 10:08:17 I gave up on both Nov 10 10:13:59 pabs3: what are you using now then? Nov 10 10:14:18 pabs3: after I did the xmonad config I finally had GUI that I liked enough. Nov 10 10:40:30 * PaulFertser doesn't have a 'scope at home and so can't understand if openocd hotplugging doesn't work because of srst pulling (which shouldn't be happening with current config) or because of the way openocd talks to arm9 :/ Nov 10 10:53:41 PaulFertser: nothing, just a home phone and laptop Nov 10 11:03:38 And anyway something "fancy" is going on during the wakeup so samsung's jtag circuity is getting confused, leading to "Error: timeout waiting for SYSCOMP & DBGACK, last DBG_STATUS: 0" Nov 10 11:03:54 I wonder if using a larger timeout might help, unlikely though Nov 10 11:42:22 What dialer and SMS programs are available? Nov 10 11:42:34 The install.sh script promised me zhone but failed to deliver. Nov 10 11:47:29 It also wanted to install epdfview which seems to be unavailable. Nov 10 11:58:54 shurd: are you after using telephony on plain debian? Nov 10 11:59:04 Yeah. Nov 10 11:59:21 shurd: zhone is not feature-complete and is lacking. I've written my own Emacs mode that talks to FSO daemons... Nov 10 11:59:42 * shurd isn't an emacs guy. Nov 10 12:00:11 shurd: zhone is a python script, i'm not sure if it's working with the current fso version but it might be easy to fix. Nov 10 12:00:22 At this point, the features I want at "Can dial phone", "Can answer phone" and "can send/recieve SMS" Nov 10 12:00:29 s/at/are/ Nov 10 12:00:32 shurd meant: At this point, the feareures I want at "Can dial phone", "Can answer phone" and "can send/recieve SMS" Nov 10 12:01:00 Bah. Nov 10 12:01:04 s/ at /are/ Nov 10 12:01:15 * shurd looks at apt. Nov 10 12:01:24 Stop pointing out that me regexps are wrong. Nov 10 12:01:50 its a bot, ignore it Nov 10 12:01:59 Heh, I figured that out. Nov 10 12:02:16 Looks like freesmartphone.org is down. Nov 10 12:02:42 I guess that's a hint that I should be working on something else for now. Nov 10 12:02:44 It was up until today, sigh. Nov 10 12:03:06 shurd: zhone can dial and answer, and use messages but probably only on SIM, I do not remember. Nov 10 12:03:16 It's really simple so you can probably fix it yourself when you need it. Nov 10 12:03:20 * shurd nods. Nov 10 12:03:46 There doesn't seem to be a package for Debian... unless it's in the FSO repo? Nov 10 12:04:41 shurd: no idea, i self-compiled everything specific to gta02. Nov 10 12:05:04 Yeah, that's what it's looking like I want to do as well. Nov 10 12:06:41 openmoku-panel-plugin is using almost all my resources (RAM and CPU) and doesn't seem to be doing much. Nov 10 12:07:10 It sucks Nov 10 12:07:24 shurd: I use special xmonad-smartphone-config Nov 10 12:07:39 Ah... yeah, I was thinking something like that. Nov 10 12:08:10 shurd: https://gitorious.org/xmonad-smartphone-config/xmonad-smartphone-config/ Nov 10 12:08:19 But a dialer and SMS thing is first on my list. Nov 10 12:08:32 Perfect... Nov 10 12:08:39 thanks. Nov 10 12:08:52 What do you use for a stylus? Nov 10 12:09:10 shurd: welcome. It'd be a tough road getting a (semi-)usable device, be warned. Nov 10 12:09:13 I'm beginning to accept that I will need one. Nov 10 12:09:16 But I use as my only cellphone. Nov 10 12:09:29 Fingernail. No matter how short I cut them, it's still ok. Nov 10 12:09:39 Well, for me it'll be my backup phone for some time. Nov 10 12:10:19 Partly because my preferred provider is CDMA. Nov 10 12:10:33 And partly because... well... it doesn't phone yet. Nov 10 12:10:34 :-) Nov 10 12:11:15 It does seem that using SHR provides a working phone. Nov 10 12:11:32 So I can cannibalize bits of that I suppose. Nov 10 12:13:52 * PaulFertser sighs, 2.6.39 for gta02 thinks serial console should be disabled when gsm is powered, wtf? It was the case >5 years ago on gta01 only... Nov 10 12:14:12 There's a serial console? Nov 10 12:14:13 shurd: yes, if you can get e17's "elementary" libs working Nov 10 12:14:20 bitrot fun Nov 10 12:14:25 shurd: yes, on testpads or via flex cable Nov 10 12:14:31 Ah, not via USB then. Nov 10 12:14:58 The USB networking is... useless. Nov 10 12:15:08 Was hoping it could be switching to serial console mode. Nov 10 12:16:29 shurd: how is it useless? Nov 10 12:16:48 It's slower than WiFi. Nov 10 12:17:00 So it's only really useless where there is no WiFi AP. Nov 10 12:17:06 Which for me is effectively nowhere. Nov 10 12:17:20 gta02's wifi module sucks, not working properly with some APs. Nov 10 12:17:23 Only really useful rather. Nov 10 12:17:58 Also, it uses RNDIS rather than CDCE so doesn't work with FreeBSD currently. Nov 10 12:18:54 Or at least I haven't firgure out how to make it work. Nov 10 12:20:45 Which, I find the use of RNDIS rather mystifying given that there are open standards for the same thing. Nov 10 12:22:19 Looking into that is well down my bucket list though. Nov 10 12:36:44 Changelist says dzen2 from Debian does "mouse events" now... so I'll try using that first. Nov 10 12:37:08 As of april 2013. Nov 10 12:58:30 I'm using cdc-ether on gta02, not rndis. Nov 10 12:58:43 How do you configure that? Nov 10 13:04:20 shurd: under /etc/modprobe.d/ somewhere I think Nov 10 13:04:23 or /etc/modules Nov 10 13:04:46 Thanks. Nov 10 13:04:49 * shurd makes a note. Nov 10 13:05:16 Currently using g_ether which is RNDIS... at least by default. Nov 10 13:05:53 Don't see a non RNDIS option. Nov 10 13:06:12 So presumably it's a different module for CDCE Nov 10 13:07:05 g_cdc maybe... Nov 10 13:07:53 looks like it is g_ether: http://www.linux-usb.org/gadget/ Nov 10 13:08:40 zero idea how it chooses Nov 10 13:08:47 http://blackfin.uclinux.org/doku.php?id=linux-kernel:usb-gadget:ethernet Nov 10 13:09:06 Well, I can always peek at the source if it supports cdce. Nov 10 13:10:43 WiFi works fine once I disable power savings for now. Nov 10 13:11:13 check the output of lsusb it might be exporting both already Nov 10 13:11:26 Doesn't seem to, no. Nov 10 13:11:42 Anyway, I'm not sure I would use the feature even if I got it working. Nov 10 13:12:00 it is much more reliable and useful than wifi Nov 10 13:13:39 Well, WiFi is completely reliable for me now... Nov 10 13:14:07 I have g_ether and I am getting the usbnet,cdc_subset,cdc_ether modules loaded on my laptop but not rndis Nov 10 13:14:37 Possibly cdc_ether supports RNDIS? Nov 10 13:14:41 What's the device name? Nov 10 13:15:02 usb0 Nov 10 13:15:07 usb 4-1: New USB device found, idVendor=0525, idProduct=a4a2 Nov 10 13:15:07 usb 4-1: New USB device strings: Mfr=1, Product=2, SerialNumber=0 Nov 10 13:15:07 usb 4-1: Product: RNDIS/Ethernet Gadget Nov 10 13:15:07 usb 4-1: Manufacturer: Linux 2.6.39.4 with s3c2410_udc Nov 10 13:15:11 cdc_ether 4-1:1.0: eth1: register 'cdc_ether' at usb-0000:00:1d.2-1, CDC Ethernet Device, 00:1f:11:01:2c:9c Nov 10 13:15:15 On my host Nov 10 13:15:26 * shurd nods. Nov 10 13:15:29 Yeah, same here. Nov 10 13:16:23 Actually, the id and product seem different. Nov 10 13:16:54 those are determined by the kernel command-line here (sent by the bootloader) Nov 10 13:17:13 I'm using the Qi bootloader Nov 10 13:17:25 Ah, no, they're the same. Nov 10 13:17:48 idVendor = 0x0525 Nov 10 13:17:48 idProduct = 0xa4a2 Nov 10 13:17:48 bcdDevice = 0x0312 Nov 10 13:17:48 iManufacturer = 0x0001 Nov 10 13:17:48 iProduct = 0x0002 Nov 10 13:17:49 iSerialNumber = 0x0000 Nov 10 13:18:10 Anyway, if I ever become irritated with my WiFi, I'll take another look. Nov 10 13:28:10 Looks like Haskel (ie: xmonad) require gcc to actually compile anything... but there's no dependency. Nov 10 13:30:20 hm? Nov 10 13:30:52 When I tried launching xmonad per PaulFertser's configs, ghc failed because /usr/bin/gcc didn't exist. Nov 10 13:31:14 what was the error message? Nov 10 13:31:38 ghc: could not execute: /usr/bin/gcc Nov 10 13:31:52 huh Nov 10 13:33:41 09:26:21 Paprikachu: GHC emits assembler files. it uses GCC to turn the assembler files into object files Nov 10 13:33:44 http://irc-logs.googlecode.com/svn/trunk/logs/freenode/haskell/2012/2012-05-29 Nov 10 13:33:58 Yeah, I epxect it's just a missing dependency. Nov 10 13:34:09 But PaulFertser may want to update his system requirements. Nov 10 13:34:12 please report a bug Nov 10 13:34:24 * shurd nods. Nov 10 13:34:40 Once I'm sure this fixed the problem, I will. Nov 10 13:34:57 Only the highest quality bug reports from me. Nov 10 13:35:04 :) Nov 10 13:35:24 severity serious would be appropriate I think Nov 10 13:35:42 Well, the problem and solution are both obvious. Nov 10 13:35:57 It's nust not normal to install haskell before you install gcc. Nov 10 13:35:59 Just. Nov 10 13:36:02 Simple oversight. Nov 10 13:36:07 missing deps are severity serious in Debian Nov 10 13:36:12 Ah. Nov 10 13:36:17 * shurd puts on his serious face. Nov 10 13:36:22 :) Nov 10 13:37:10 * shurd is stuck in the "did this fail, or is it slow" limbo. Nov 10 13:39:42 Neither! Nov 10 13:39:44 Yay. Nov 10 13:39:45 :-) Nov 10 13:40:58 * shurd wonders if reportbug requires an MTA. Nov 10 13:41:53 Any sane *nix should have a working MTA :) ssmtp works ok. Nov 10 13:42:04 My phone does not have a working MTA. Nov 10 13:42:07 It is not sane. Nov 10 13:42:15 :)) Nov 10 13:43:26 My wife now officially hates my "new" phone. Nov 10 13:44:13 not shiny enough? Nov 10 13:44:22 Keeps me from getting the garage clean. Nov 10 13:44:28 :) Nov 10 13:44:54 She claims it is almost as bad as my radios. Nov 10 13:46:21 Hah! Nov 10 13:46:30 As a reportbug novice, I cannot choose a severity of serious. Nov 10 13:46:50 important, normal, minor, wishlist. Nov 10 13:46:55 modify it when you get into mutt or whatever mua Nov 10 13:48:29 * PaulFertser thought newer ghc uses llvm Nov 10 13:48:55 It looks like the LLVM that's installed is newer than is supported... Nov 10 13:48:58 And it gell back to gcc. Nov 10 13:52:25 Where did I get 2.6.39 from and why am I using it? It was some prebuilt deb package by some fine fellow, probably radekp? I need to know if the corresponding elf is available in somewhere so that I can load it in gdb (as I'm conected to a hanged gta02 via jtag atm and need to try to see where it's stuck). Nov 10 13:53:25 Is debug info (like the one in splitdebug) available for deb-built kernels usually? Nov 10 13:54:47 Linux kernels from ftp.debian.org do, not sure about others Nov 10 13:56:17 nevermind, hotplug attaching doesn't work anyway, i get data abort for reading mem :/ Nov 10 14:23:46 PaulFertser: attaching debug board to a running system does work Nov 10 14:24:37 lindi-: ok, good to know, thank you for the info, will try harder. Nov 10 14:24:38 PaulFertser: http://lindi.iki.fi/lindi/openmoko/linux-2.6-gta02/ does have debugging symbols Nov 10 14:24:54 lindi-: i'll self-compile myself anyway once I get close to really debugging this. Nov 10 14:25:28 PaulFertser: http://lists.openmoko.org/pipermail/openmoko-kernel/2010-November/011269.html Nov 10 14:25:31 lindi-: btw, big thanks for the debug board, I'm assuming it your device I'm using atm Nov 10 14:26:43 PaulFertser: I have enough of those already, good if you can make something useful with it Nov 10 14:26:59 I just got android phone and have spent a few days learning the system Nov 10 14:27:06 * PaulFertser haven't done anything useful yet alas :/ so far feeling sceptical about it Nov 10 14:27:23 lindi-: so, how does it feel, what are your first impressions? Nov 10 14:27:46 PaulFertser: making a backup of the pristine state seems to be very challenging Nov 10 14:27:57 PaulFertser: all backup instructions start with "install clockworkmod recovery" Nov 10 14:28:14 PaulFertser: and forget that installing it will overwrite the original recovery code which you'll never get back Nov 10 14:28:42 lindi-: android community is just like that, I'm not surprised Nov 10 14:29:08 PaulFertser: well, I did see a lot of confused people in the openmoko community too :) Nov 10 14:29:12 lindi-: if your device is supported by "ODIN" you can make a full backup with "heimdal" I heard. Nov 10 14:29:29 PaulFertser: heimdall does not support this usb product/vendor id Nov 10 14:29:54 PaulFertser: and I didn't manage to get odin to work last time I had access to a windows test laptop Nov 10 14:29:56 I do not understand why samsung android is supported and others are not. Nov 10 14:30:05 this is samsung :) Nov 10 14:30:42 So chances are you can beat heimdall into working (especially given odin should support this very device and is using the same protocol). Nov 10 14:31:19 PaulFertser: that's possible yes, i'd first like to see that it works with odin though Nov 10 14:31:24 PaulFertser: and then capture the usb traffic Nov 10 14:31:38 PaulFertser: then look at heimdall source code and figure out if it could work by just adding the usb product id Nov 10 14:34:48 lindi-: btw, if you are going to use openocd with ftdi adapters, I really recommend switching to the new ftdi driver, it's written from scratch by Andreas Fritiofson and is actively maintained, shows way better performance and stability. Nov 10 14:35:01 PaulFertser: is it in debian? Nov 10 14:35:50 lindi-: alas, no, Debian packaging for OpenOCD is lacking behind. I think it was released in 0.7.0 or something. I'm using git HEAD anyway, as releases in OpenOCD happen rarely and I prefer to be able to participate in development. Nov 10 14:36:22 the config files for this new drivers are in interface/ftdi Nov 10 14:37:04 It's ironic how I have difficulties reading CPU state currently (connecting to a running target) given how much I used and hacked on OpenOCD (but with different targets). Nov 10 14:37:37 PaulFertser: I did find openocd rather painful to use Nov 10 14:37:58 PaulFertser: I couldn't even figure out an easy way to dump the whole RAM with it (if the MMU is enabled) Nov 10 14:39:01 lindi-: hm, right, mdw has "phys" mode and "dump_image" doesn't. I think this would be an easy fix. Nov 10 14:39:27 PaulFertser: yes I made a patch for my own use but the code was quite a mess Nov 10 14:40:21 Most OpenOCD users do not deal with MMU as they're debugging microcontroller apps or bootloaders. Nov 10 14:45:31 Asserting trst stops cpu, deasserting resets, wtf... Nov 10 14:48:14 PaulFertser: seems dumping functionality has been removed from latest heimdall: https://github.com/Benjamin-Dobell/Heimdall/issues/97 Nov 10 14:48:30 PaulFertser: but that also mentions some forensics company that claims to know how to do it Nov 10 14:53:28 I see Nov 10 14:53:50 Thought the protocol was sane enough but meh. Android... **** ENDING LOGGING AT Mon Nov 11 02:59:58 2013