**** BEGIN LOGGING AT Tue Jul 03 02:59:58 2012 Jul 03 12:07:45 paulk-desktop: hi Jul 03 12:07:54 hi Jul 03 12:08:15 I faced strange issues with RFS yesterday Jul 03 12:08:20 paulk-desktop: I've found 2 documents about gps (csr sirf start 4 gsd4t) Jul 03 12:08:29 oh, nice Jul 03 12:08:33 but you need flash to view them http://wenku.baidu.com/view/875911caa1c7aa00b52acb29.html### and http://wenku.baidu.com/view/ec038a03eff9aef8941e063c.html Jul 03 12:08:49 I think there is some SiRF support in gpsd Jul 03 12:09:21 yeah, but a brief glance at the library and the docs shows we'll have to write a firmware loader and cut the non-free arm firmware from the library Jul 03 12:09:27 paulk-desktop: so, what about RFS? Jul 03 12:09:52 it gets a 0x4000-sized data packet Jul 03 12:09:57 which seems strange Jul 03 12:10:24 that probably means you're off by one byte Jul 03 12:10:50 there is actually that much data sent Jul 03 12:11:00 on the other hand, NVDATA is 2MB and it makes sense to have huge blocks Jul 03 12:11:04 (I have that in the header in in read count) Jul 03 12:11:20 but then I get one with a size of -1 in the header Jul 03 12:11:26 so that's a problem Jul 03 12:11:29 so what's the problem? does it not work or what? Jul 03 12:11:39 it goes wild after that Jul 03 12:11:50 I got tired yesterday so I gave up after this Jul 03 12:11:52 do you have a trace? Jul 03 12:11:56 I'll make one Jul 03 12:12:09 I'm getting CM9 sources now Jul 03 12:12:12 I'll make another kernel Jul 03 12:12:20 with console service started as root Jul 03 12:13:20 then I'll try to figure things out Jul 03 12:13:39 bottomline: I'm not done yet as I expected Jul 03 13:19:54 oh, I think that I got it Jul 03 13:20:19 it sends the header once, giving a size of 0x4000 in the header but actually sends chunks of 4096bytes Jul 03 13:21:13 that's why next time I read only 0xff in the header: that's because it's not the header Jul 03 13:23:10 yep that's definitely it Jul 03 13:23:18 now I'm worried that I wrote a bunch of 0 to nv_data Jul 03 13:33:33 paulk-desktop: why are you worried? the default nvdata is all zeros. or have you trashed your nvdata.bin and need a backup? Jul 03 13:33:52 I have copied it somewhere as soon as I get the phone Jul 03 13:34:02 and it's not zeros everywhere Jul 03 13:34:05 there is the IMEI Jul 03 13:34:09 the simlock stored, etc Jul 03 13:34:54 yes, but the default nvdata (from radio firmware) is all zeroes.. either way, when I loaded nvdata.bin instead of nvdata from radio (modded the fwloader), the ril crashed all the time and the modem didn't reach 'online' state. need to look into it Jul 03 13:35:58 btw, most S2 ICS roms have a huge bug with the default player - the volume is raised to max when switching tracks. doesn't affect third-party players Jul 03 13:37:14 ok Jul 03 13:39:19 paulk-desktop: btw, is the first argument (the context with a handle) used in the ril client? its definition is different for omap (nexus), so it may be the reason gps/sound don't work Jul 03 13:40:19 oh, it is used. need to look into it Jul 03 13:40:47 ok Jul 03 13:52:07 I think it works now (RFS) Jul 03 14:08:22 Alex[sp3dev], btw there is "based on xmm6260 IPC code which is:" on the copyright lines Jul 03 14:08:40 I guess that was "based on crespo IPC code which is:" but then you s/crespo/xmm6260/ Jul 03 14:08:45 correct? Jul 03 14:09:12 paulk-desktop: probably Jul 03 14:09:54 paulk-desktop: so, what's the RFS status? Jul 03 14:10:03 working, doing the commit right now Jul 03 14:10:11 ok Jul 03 14:11:15 I'll look at it and comment. So we only have to impement loading NVDATA from /efs instead of radio.bin and fix nvdata path for galaxy nexus (make it settable together with ipc callbacks in a per-board manner) Jul 03 14:11:53 http://download.paulk.fr/replicant/galaxys2/0001-RFS-added-missing-send-recv-functions.patch Jul 03 14:13:55 I know well nv_data mechanism (I wrote most of it in libsamsung-ipc) so I think I'll do that next Jul 03 14:14:32 paulk-desktop: you can add the hook in the fwloader in the secure image loading routine. you may wish to look at the samsung ril in IDA Jul 03 14:14:44 paulk-desktop: I'll test your patch on nexus and commit Jul 03 14:15:42 ok Jul 03 14:24:58 paulk-desktop: you've forgotten to include rfs_hdr definition Jul 03 14:25:13 oops, it's in the header Jul 03 14:25:26 I can commit it, but just pastebin it Jul 03 14:26:31 updated: http://download.paulk.fr/replicant/galaxys2/0001-RFS-added-missing-send-recv-functions.patch Jul 03 14:41:36 paulk-desktop: crashes on nexus. I'm debugging but the whole do-while thingie seems wrong Jul 03 14:43:01 maybe reduce it to no loop, one read and assume it contaisn the header Jul 03 14:43:19 does it crash like segfault or what? Jul 03 14:43:37 maybe increase IPC_MAX_XFER a bit also Jul 03 14:45:31 why would the do-while thing be wrong also? Jul 03 14:46:07 paulk-desktop: ok, I'll look into it. I'm a bit distracted now by some chat Jul 03 14:46:19 alright :) Jul 03 15:18:45 paulk-desktop: btw, have you seen my uboot port to s2? it can boot a custom kernel/uboot script from the data partition Jul 03 15:19:11 that's great! I have heard that you did the ports, weeks ago Jul 03 15:19:24 the port* Jul 03 15:20:06 http://forum.xda-developers.com/showthread.php?t=1680898 it is flashed instead of kernel. so you have to flash kernel to the otherwise unused recovery partition.. since I don't have jtag for s2 or usb unbricker, I can't make a full port replacing samsung bootloader like on nexus Jul 03 15:20:58 ok, I'll be debugging nexus now. probably should be easy to fix Jul 03 15:22:20 ok Jul 03 15:23:09 paulk-desktop: looks to me, it was crashing because it was unable to find /efs. when I made a symlink and remounted, it booted Jul 03 15:23:22 oh! Jul 03 15:23:28 that's very likely Jul 03 15:23:57 because even though it's not loaded, when the modem asks to write on it via RFS requests, it will try to write there Jul 03 15:24:10 that means RFS works I guess Jul 03 15:24:24 have logs? Jul 03 15:25:25 paulk-desktop: yes, everything works. Will commit your patch and try making efs path overridable per-board Jul 03 15:25:36 do you want me to do it? Jul 03 15:25:46 (per-board efs) Jul 03 15:25:53 paulk-desktop: no, I want to get my hands dirty with code ;) Jul 03 15:25:59 okay :D Jul 03 15:26:03 paulk-desktop: I'd like you to test gprs and fix ussd Jul 03 15:26:11 will do about gprs Jul 03 15:26:17 since you know ipc better Jul 03 15:26:18 for ussd it's gonna be harder Jul 03 15:26:28 why Jul 03 15:26:38 basically it's broken for every device Jul 03 15:27:01 I'm not really sure about why -- it works for me with some SIM Jul 03 15:27:09 but on some operators, it doesn't Jul 03 15:27:36 I think it's an encoding issue: some operators want it GSM7 and some others ASCII Jul 03 15:28:15 paulk-desktop: pushed Jul 03 15:28:22 I should ask people who have nexus s to test and send me the log of the non-free RIL when doing USSD Jul 03 15:28:53 paulk-desktop: we have s2 and galaxy nexus. why can't we make ussd calls with non-free ril ourselves? Jul 03 15:29:11 because the logs won't help Jul 03 15:29:22 AFAIK only nexus s logs are verbose enough Jul 03 15:29:49 well.. strace is the way. and btw, I think it is possible to enable debugging in our ril Jul 03 15:30:34 in the galaxy s2 RIL? Jul 03 15:30:41 wow, it even displays baseband version ^_^. or is it not the ril feature? Jul 03 15:31:20 yep, the RIL reports baseband version Jul 03 15:32:32 so about USSD: I'll have to borrow my mom's SIM cause my operator doesn't have any UUSD code Jul 03 15:32:56 paulk-desktop: we had a discussion with Paul Fertser about error handling and didn't reach a conclusion. Do you think it is better to crash on low memory or check the return value of every malloc and shutdown gracefully? Jul 03 15:33:14 huh Jul 03 15:33:58 on the other hand, ril is restarted automatically, so that's not a problem Jul 03 15:34:01 I would say checking return value of every malloc Jul 03 15:34:18 yep, I usually restart it with killall Jul 03 15:34:30 paulk-desktop: then why is it not checked in libsamsung-ipc? I think it may be used outside android where ril is not reloaded on its own Jul 03 15:34:53 'cause we're lazy, probably Jul 03 15:35:12 so you don't mind if I add sanity checks everywhere when we're done/ Jul 03 15:35:13 ? Jul 03 15:35:26 sure, that's a good thing to do Jul 03 15:35:33 btw about return codes: Jul 03 15:35:36 0 means OK Jul 03 15:35:40 negative means error Jul 03 15:35:53 > 0 means OK with some indication (usually for read/write count) Jul 03 15:35:56 yes, but returning -1 on recv makes modemctl and probably ril crash Jul 03 15:36:20 I think samsung-ril will restart when it catches -1 Jul 03 15:36:32 it'll restart bootstrap, etc Jul 03 15:36:34 ok Jul 03 15:36:43 paulk-desktop: who mounts efs rw? Jul 03 15:37:05 who? Jul 03 15:37:17 paulk-desktop: ril or init.rc? Jul 03 15:37:29 not the ril Jul 03 15:37:37 must be init.rc Jul 03 15:38:16 ok. don't feel like editing it but I'll have to set up our own tree and rebuild everything anyway soon Jul 03 15:38:40 why? Jul 03 15:39:00 I mean editing for what? Jul 03 15:39:13 paulk-desktop: it is ro in galaxy nexus by default Jul 03 15:39:34 ah ok, I ignored that Jul 03 15:40:06 btw is the RIL running as root on galaxy s2? Jul 03 15:40:50 paulk-desktop: I would assume it is. Or the owners/permissions are set correctly Jul 03 15:40:57 radio 1860 1 10964 2660 ffffffff 400a4dd4 S /system/bin/rild Jul 03 15:40:59 it's not Jul 03 15:41:10 but on gingerbread, it probably is Jul 03 15:41:17 doesn't matter Jul 03 15:41:36 init.rc has good permissions set for /dev, /sys and /efs for radio-related files Jul 03 15:41:50 ok Jul 03 15:41:52 btw, any reason to build gingerbread? Jul 03 15:42:02 replicant will stick to GB Jul 03 15:42:06 in the near futur Jul 03 15:42:07 why Jul 03 15:42:16 because ICS requires hardware acceleration Jul 03 15:42:29 and devices are too slow with software GL on ICS Jul 03 15:42:40 while they're nearly good with GB Jul 03 15:43:02 though we'll try to work it out Jul 03 15:43:13 and eventually, we'll switch to ICS Jul 03 15:43:21 but no in the near future Jul 03 15:48:57 yes. I've tried with software GL right now. Unacceptable Jul 03 16:30:29 SHR: 03Martin.Jansa 07shr-chroot * r97c7cb8dbf72 10/ (13 files in 6 dirs): bitbake upgrade Jul 03 16:30:29 SHR: 03Martin.Jansa 07shr-chroot * r1802565da8b7 10/ (14 files in 8 dirs): bitbake: add patch for automatic svn upgrade to test it Jul 03 16:30:31 SHR: 03Martin.Jansa 07shr-chroot * r1f71d2b65a29 10/ (1086 files in 129 dirs): system upgrade Jul 03 17:22:54 SHR: 03shr-devel 07buildhistory * rb3bef9990d28 10/packages/ (16 files in 16 dirs): packages: Build 201207031852 of shr 20120703 for machine om-gta02 on opmbuild Jul 03 17:56:57 paulk-desktop: I'm desperate. I've written the code for efs path and it looks good to me, but I've been unable to trace why it crashes ;) Jul 03 17:57:46 what if you disable the RFS functions from rfs.c on samsung-ril? you'll at least see if the first RFS request is correctly handled Jul 03 17:59:15 SHR: 03shr-devel 07buildhistory * rb9712a8bcb5a 10/packages/ (16 files in 16 dirs): packages: Build 201207031937 of shr 20120703 for machine nokia900 on opmbuild Jul 03 18:00:12 paulk-desktop: I feel something is wrong with library linking. I will now test some things, should not be hard to find out Jul 03 18:07:15 paulk-desktop: here is the code in the rfs_path_hack branch. feels like client passes NULL check, but everything crashes after that. https://github.com/astarasikov/libsamsung-ipc/tree/rfs_path_hack and the log http://pastebin.com/PqVMQdQm Jul 03 18:07:52 i added stupid prints everywhere.. just study the code and say if you see something Jul 03 18:09:17 yes, it fails with both fs_ops commented and not Jul 03 18:10:43 your code seems all right Jul 03 18:10:56 any indication in the main buffer log? Jul 03 18:11:15 it would say if it's SEGFAULT, bad address, double-free, etc Jul 03 18:11:40 paulk-desktop: as soon as I comment it (change #if 1 to to 0), it works... My suggestion is that initializing char pointers with strings in c files is bad. or is it not? Jul 03 18:12:07 or, maybe NULL is not 0 here? hard to tell Jul 03 18:12:28 I have all sorts of crazy ideas.. But none of them worked ;) Jul 03 18:13:08 I would make a copy of the strings Jul 03 18:13:23 instead of returning the strings pointers Jul 03 18:13:56 like https://github.com/morphis/libsamsung-ipc/commit/63a3609348e04a1899b69a4f95e626408497f230 Jul 03 18:14:11 asprintf fixed strings and free later in the code Jul 03 18:14:28 that works everytime Jul 03 18:15:08 actually, the return at the end of function works (obviously, strings are put to .text section). ok, I'll try that and some other stuff. Renaming /factory to /efs could be easier. But now I need to know the reason Jul 03 18:15:54 this is bullshit Jul 03 18:16:32 I've added some prints. and here is the log Jul 03 18:16:33 http://pastebin.com/ARW4xPU9 Jul 03 18:17:04 it must not crash there. it is insane Jul 03 18:17:15 I guess it's printf("nv_data_path: %s\n", nv_data_path(client)); Jul 03 18:17:51 just copy the strings… Jul 03 18:18:35 that might be a scope issue Jul 03 18:18:44 the nv_data functions are called from samsung-ril Jul 03 18:18:52 no, you see, it prints the pointer to the client structure, but it crashes when printing the client member Jul 03 18:20:21 right Jul 03 18:20:34 I didn't quite look at the code that was upper the logs Jul 03 18:21:02 the only explanation I have is that ipc_client may be redefined with a different size somewhere Jul 03 18:21:30 did you modify anything in samsung-ril? Jul 03 18:21:34 no Jul 03 18:21:44 I have a crazy theory that android won't rebuild if you change its headers Jul 03 18:22:01 basically, you changed libsamsung-ipc headers Jul 03 18:22:07 but samsung-ril still know the old header Jul 03 18:22:17 and as you didn't rebuilt it, it doesn't have the new one Jul 03 18:22:21 touch ./* in samsung-ril Jul 03 18:22:24 may be that. I've even forced it to remove the binaries before building. I'll try to see if anything remains precompiled Jul 03 18:22:25 and rebuild it Jul 03 18:22:39 that sounds like it Jul 03 18:23:02 why is everything so fucked up in android. just everything Jul 03 18:23:09 :D Jul 03 18:24:22 though, you call client->fs_ops from a place where the correct header is included, else it would fail to build Jul 03 18:28:08 also, make sure you register maguro device Jul 03 18:29:26 SHR: 03shr-devel 07buildhistory * rf346523800fc 10/packages/om_gta04-oe-linux-gnueabi/shr-version/ (latest shr-version/latest): packages: Build 201207032015 of shr 20120703 for machine om-gta04 on opmbuild Jul 03 18:29:51 you, mortal, know nothing of how libsamsung-ipc works. I don't trust you anymore xD Jul 03 18:30:10 haha Jul 03 18:30:30 I was misguided by the fact ipc_client and ipc_device_desc look similiar. turns out I needed to copy the pointer in ipc_client_new() Jul 03 18:30:43 oh Jul 03 18:30:48 that makes sense Jul 03 18:30:50 the morale: always do a bzero after malloc Jul 03 18:31:51 we should have asked morphis, we would have know, we wrote that part Jul 03 18:33:33 well, the problem is probably that after memory for ipc_client was allocated, it was not zeroed out. so the pointer was non-zero and passed the check, but of course, we were receiving a segfault on dereferencing it Jul 03 18:34:21 that's strange anyway, it should've printed the pointer anyway. maybe a concurrent thread was interrupting the log printing and the library crashed at that point Jul 03 18:37:39 I'm pretty sure %s with a random pointer can segfault Jul 03 18:37:49 it was a % Jul 03 18:37:51 %x Jul 03 18:38:02 %s: client->fs_ops %x\n", client->fs_ops); Jul 03 18:38:07 both apparently Jul 03 18:38:21 apparently i'm stupid and forget to add __func__ all the time Jul 03 18:38:27 ^^ Jul 03 18:38:41 that seemed odd indeed Jul 03 18:39:49 ok, pushed Jul 03 18:40:12 now I need to rebuild nexus ramdisk :( Jul 03 18:40:40 or maybe I can shove a remount into some initscript in /system Jul 03 18:41:23 paulk-desktop: have you experimented with loading nvdata from /efs in fwloader? Jul 03 18:43:54 paulk-desktop: now I'm going to try fixing sound. Seems galaxy nexus and s2 are using an updated client pointer format so we'll have to make a new ril client for them. Should be just a couple of lines to fix, all the dlsym'd function names seem to be the same Jul 03 18:44:16 https://github.com/CyanogenMod/android_device_samsung_tuna/blob/ics/audio/ril_interface.h Jul 03 18:48:06 yay. didn't have to rebuild ramdisk, just put a script to init.d for remounting Jul 03 18:48:17 ok Jul 03 18:48:46 paulk-desktop: do we need to validate nvdata before loading it with fwloader? Jul 03 18:48:55 yes, there is a long process Jul 03 18:48:59 I knwo it though Jul 03 18:49:01 know* Jul 03 18:49:09 I'll do it right after I'm done eating Jul 03 18:49:16 (I'm off now for eating) Jul 03 18:49:18 ok, thank you. good appetite Jul 03 18:49:22 thanks Jul 03 18:59:32 SHR: 03shr-devel 07buildhistory * r1b69eda652c4 10/images/crespo/eglibc/chroot-image/ (build-id image-info.txt installed-package-sizes.txt): images: Build 201207032044 of shr 20120703 for machine crespo on opmbuild Jul 03 18:59:42 SHR: 03shr-devel 07buildhistory * r570830f468bf 10/packages/crespo-oe-linux-gnueabi/shr-version/ (latest shr-version/latest): packages: Build 201207032044 of shr 20120703 for machine crespo on opmbuild Jul 03 19:28:12 SHR: 03shr-devel 07buildhistory * r40741658f2fe 10/packages/palmpre-oe-linux-gnueabi/shr-version/ (latest shr-version/latest): packages: Build 201207032114 of shr 20120703 for machine palmpre on opmbuild Jul 03 19:30:01 morphis: hello, do you know the irc nic of Jarkko Lehtoranta ? Jul 03 19:32:53 nschle85: heyho Jul 03 19:32:56 who is that? Jul 03 19:34:33 morphis: jojojo, this is the guy who fixed huawei 173u-2 audio in ofono Jul 03 19:34:59 ok Jul 03 19:35:12 no I don't know his irc nickname Jul 03 19:35:17 maybe ask in #ofono? Jul 03 19:35:21 morphis: ok Jul 03 19:35:52 morphis: nobody answered since yesterdday Jul 03 19:36:47 this one http://comments.gmane.org/gmane.comp.handhelds.ofono/12317 ? Jul 03 19:37:40 nschle85: or this patch: https://bugs.meego.com/attachment.cgi?id=9702 Jul 03 19:37:50 morphis: its seems you are one of the main architects of fso, so if i propose patches, are you willing to integrate huawei usb dongles into fso ? Jul 03 19:38:49 nschle85: for sure Jul 03 19:39:11 if you respect my review comments :) Jul 03 19:39:58 morphis: i mean the first one you told me (http..comments..) Jul 03 19:40:24 ok Jul 03 19:40:39 nschle85: you're talking about this patch from him: https://bugs.meego.com/attachment.cgi?id=9702 Jul 03 19:41:31 morphis: ok its the same issue Jul 03 19:41:38 ok Jul 03 19:41:52 nschle85: do you already started with integrating huwai support in fsogsmd? Jul 03 19:42:02 morphis: no Jul 03 19:42:11 ok Jul 03 19:43:42 morphis: during last fsoshrcon12 is asked about what people think about that integration and people are not amused abou that Jul 03 19:44:01 who did you ask? Jul 03 19:44:12 and btw. whats your usecase for it? Jul 03 19:44:16 i was asking in the round Jul 03 19:44:21 ok Jul 03 19:44:26 I was there? Jul 03 19:44:32 may be Jul 03 19:44:57 my argument was to bring shr on the desktop Jul 03 19:45:21 ok, so you asked in context of SHR Jul 03 19:45:22 but nobody agreed Jul 03 19:45:52 * Alex[sp3dev] wishes someone integrated fso and dialer notifications into ubuntu UI instead of spreading EFL Jul 03 19:45:58 for SHR I can't speak but for FSO it's a wellcome contribution Jul 03 19:46:09 s/wellcome/welcome/ Jul 03 19:46:10 morphis meant: for SHR I can't speak but for FSO it's a welcome contribution Jul 03 19:47:01 for FSO on the desktop I see several use cases and so it would be really nice to support different usb sticks too Jul 03 19:47:02 but in my oppinion to have a cheep usb dongle acting as mobile phone will open fso and shr development for more people Jul 03 19:47:52 enabling more people to develop with FSO and SHR on the desktop is still no problem today Jul 03 19:48:01 you can do that with phonesim or via forwarding Jul 03 19:48:34 nschle85: I'm not trolling, but I personally think it is better to have a phone-oriented window manager and fso integrated to gnome instead of a EFL-based UI. Sorry, but I can't help you here now, I'm a bit busy. But I will try working on porting ubuntu to galaxy s2 when me and paulk finish the ril Jul 03 19:50:50 which makes me think it would make sense to try writing a plugin to use fso as a driver for ofono.. that would allow us to use meego on all our hardware Jul 03 19:51:28 Alex[sp3dev]: this i discussesd already with morphis :-) Jul 03 19:51:30 Alex[sp3dev]: I think you don't get this accepted upstream ever :) Jul 03 19:51:58 Alex[sp3dev], nschle85: that don't make sence Jul 03 19:52:20 if you want to do this you can take ofono as gsm itself and let fsogsmd out Jul 03 19:52:36 you can use FSO without fsogsmd Jul 03 19:52:57 there is nothing to win with a bridge from ofono to fsogsmd Jul 03 19:53:33 ofono is well supported today and already implements a lot more features than fsogsmd Jul 03 19:53:35 Alex[sp3dev]: i know that you are NOT trolling, but you can have a "phone" workink in your desktop environment and in your target environment Jul 03 19:53:35 well, the problem is that fsogsmd has plugins for samsung ipc, for htc ril and whatever and no one feels like rewriting them a thousand times. But I guess I'm one of probably <10 people that would like a destkop distro on the phone Jul 03 19:53:59 Alex[sp3dev]: yes :) Jul 03 19:54:46 paulk-desktop: ok, I'm sort of having problems with samsung ipc client. looks like for tuna, they're passing a pointer to socket fd instead of a structure. I'll go take a shower and continue hacking it Jul 03 19:55:23 I will not enter such discussions about possible ways to get a working distribution on phones here anymore, it takes me to much time I don't have today Jul 03 19:55:32 (just to say this) Jul 03 19:55:36 morphis: sorry :) Jul 03 19:56:14 SHR: 03shr-devel 07buildhistory * r40b024914aa7 10/packages/palmpre2-oe-linux-gnueabi/shr-version/ (latest shr-version/latest): packages: Build 201207032142 of shr 20120703 for machine palmpre2 on opmbuild Jul 03 19:57:15 :) Jul 03 19:57:38 Alex[sp3dev]: no ,i hoped that developmnent without having a samsung or GTA0? N900 will open a wider range (desktop) Jul 03 19:57:41 I will continue my work on getting FSO forwards and will try to respect your feature requests Jul 03 19:58:17 nschle85: I implemented initial support for HFP HF in the last weeks for fsogsmd Jul 03 19:58:29 ???? Jul 03 19:59:05 thing about the following: have your phone arround, connect it via bluetooth to your laptop and if an call arrives you can control it from your PC instead of using your phone Jul 03 19:59:36 HFP = handsfree profile Jul 03 19:59:45 normally used in cars Jul 03 20:00:35 morphis: you drive my crazy !!!!!! (sorry my keyboard is polling) Jul 03 20:02:13 nschle85: ? Jul 03 20:04:32 morphis: i think we should developers give an easy entry for development this ioncludes also hardware, so i guess using pandaboard and huawei stick may open a wider range for developers than currently supported hardware Jul 03 20:04:59 yes Jul 03 20:06:57 morphis: it does not mean that shr or fso must run on desktop but for developers its much easier to develop on Jul 03 20:07:53 nschle85: FSO itself already runs quite nice on the desktop (there are also debian packages for several SHR components available) Jul 03 20:08:27 if you just want to develop for FSO and SHR on the desktop the missing peace is some well written documentation Jul 03 20:08:33 or an easy to install SDK Jul 03 20:08:51 in my opinion it's not really about missing hardware support Jul 03 20:09:11 if you just want to develop and test on the desktop you can do this quite nice with phonesim Jul 03 20:10:53 nschle85: Qt and android provides such nice SDKs where you can develop on the desktop and test various phone aspects before doing this on the phone Jul 03 20:13:28 morphis: but i think we need more manpower for fso and shr . Jul 03 20:13:38 good point :) Jul 03 20:14:22 I don't have any plan how to get more people involved Jul 03 20:14:28 or just an idea Jul 03 20:14:45 morphis: so on application level we need a kind of cheep "standard" phone Jul 03 20:15:13 nschle85: who is "we"? Jul 03 20:15:39 so any kind of huawei modem may be a solution (we=shr+fso) Jul 03 20:16:37 I don't understand why a huawei modem ist the solution to get more people involved? Jul 03 20:17:17 morphis: its cheep and provides all phone functionallity Jul 03 20:17:40 you mean we need it for application development on the desktop? Jul 03 20:18:12 morphis: YES, i think now you understand what i want to tell Jul 03 20:18:38 you know phonesim? Jul 03 20:19:07 http://packages.debian.org/wheezy/ofono-phonesim Jul 03 20:19:13 morphis: why use a simulator ? Jul 03 20:19:23 why not? Jul 03 20:19:51 morphis: nobody wants to develop for simulators Jul 03 20:19:59 if it's just about application development it's easier to use and without paying any money Jul 03 20:20:16 you don't develop for simulators Jul 03 20:20:20 you develop with them Jul 03 20:20:28 morphis: ah, now I see why we won't ever reach a single point of view. I (and many other people) just want to run linux with all hardware on their phone. Developing the abstract phone stack without running it on a real hardware is too boring and academic Jul 03 20:21:04 Alex[sp3dev]: I want this too Jul 03 20:21:05 Alex[sp3dev]: i agree Jul 03 20:21:14 but I don't have the time do get this done anymore Jul 03 20:21:18 I spend to much time on it Jul 03 20:21:30 thats why I steped back and do the abstract phone stack thing ... :) Jul 03 20:22:04 morphis: you are not doing abstract things.. you are implementing Jul 03 20:22:08 morphis: I understand you. I've also spent a lot of time cleaning up kernel crap and trying to get it more standard. And you know how it goes - when you get your phone barely working, it's already too old and new ones are even more open Jul 03 20:23:49 SHR: 03shr-devel 07buildhistory * reb3116f7439d 10/images/om_gta02/eglibc/shr-image/ (5 files): images: Build 201207032158 of shr 20120703 for machine om-gta02 on opmbuild Jul 03 20:24:06 nschle85: :) Jul 03 20:24:23 Alex[sp3dev]: yes and thats why I using an android phone today and nothing more Jul 03 20:24:31 no experiments with it just using it Jul 03 20:24:53 if you're looking how the linux-on-phone-thing has moved over the past years it's getting better and better Jul 03 20:25:11 it's nothing you can get done tomorrow Jul 03 20:25:23 look at android, webOS, Meego, Tizen ... Jul 03 20:26:06 morphis: but in view of shr there is no cheap testing device Jul 03 20:26:10 the problem for me is that it is impossible to buy anything except android :( Jul 03 20:27:15 nschle85: abstract in a way of not targetting any platform directly Jul 03 20:27:30 quite honestly, I'd agree to use proprietary modem and opengl provided that alsa sound and v4l2 camera were opensource. still there are no devices except n9 and n900 Jul 03 20:27:56 ok, this is counter-productive. i'm back to ril hacking Jul 03 20:28:25 nschle85: but a testing device means a real phone Jul 03 20:29:02 Alex[sp3dev]: :) Jul 03 20:29:26 morphis: yes pandaboard or x86 system using hawei stick is a real phone Jul 03 20:29:35 a big phone, yes :) Jul 03 20:29:56 this to a phone what a getto blaster is to a ipod Jul 03 20:30:10 :) Jul 03 20:30:27 nschle85: but please go and do the huawei support in fsogsmd, should be not that complex Jul 03 20:30:30 morphis: yes, and ? Jul 03 20:30:40 you should be able to copy some stuff from ofono Jul 03 20:30:48 I have to leave, sorry Jul 03 20:30:56 morphis: bye Jul 03 20:31:11 nschle85: if you need any support while implemeting write me a mai Jul 03 20:31:12 l Jul 03 20:31:22 morphis: ill do Jul 03 20:31:23 nschle85: having a phone with a huawei open surely nice feature with asterisk Jul 03 20:31:53 misc: asterisk already supports huawei Jul 03 20:32:27 damn Jul 03 20:32:41 and i was still using my old bluetooth phone :) Jul 03 20:33:00 misc: but not all modems are supported Jul 03 20:33:55 Alex[sp3dev]: what am i doing wrong ? Jul 03 20:34:39 nschle85: what are you doing? ;) you guys are all doing good work and fsogsmd is nice, it even worked fine on my laptop Jul 03 20:35:50 Alex[sp3dev]: but if i start to support huawai modems and support desktop systems i get a storm of fire Jul 03 20:38:03 nschle85: just decide what you want - to get fsogsmd on desktop or what. I think porting SHR to desktop is a stupid idea. because ipks and OE sucks compared to debian/rpm or gentoo, and no one's going to use a non-mainstream distro on a desktop. packaging latest fso and dialer for debian/ubuntu is a sane idea and would require less work Jul 03 20:38:47 as for huawei, it makes sense to support them if these modems provide voice capabilities. otherwise a generic at modem is enough Jul 03 20:47:02 paulk-desktop: ping Jul 03 20:47:13 Alex[sp3dev], pong Jul 03 20:47:43 paulk-desktop: ril is not creating the socket. I've added a debug print and I get: " srs_server_open: server_fd -1, errno 94, err Socket type not supported " Jul 03 20:48:08 I mean, the srs socket Jul 03 20:48:16 mhh Jul 03 20:48:27 that's odd, IIRC it's a simple unix socket Jul 03 20:49:10 yes, and the proprietary library uses the same params (1 for NAMESPACE_RESERVED and 0 for STREAM) Jul 03 20:49:34 maybe it's a permission issue somehow? Jul 03 20:50:13 on other phones, it runs as non-root and it works Jul 03 20:50:25 (at least on nexus s) Jul 03 20:50:40 but maybe you should try to run it as root to see Jul 03 20:50:52 you have to remove the call to the function to change user in rild code Jul 03 20:50:56 hardware/ril/rild Jul 03 20:51:13 SHR: 03shr-devel 07buildhistory * r53cb19db2f09 10/images/nokia900/eglibc/shr-image/ (5 files): images: Build 201207032226 of shr 20120703 for machine nokia900 on opmbuild Jul 03 20:51:33 I can copy-paste server_local_server code and add prints to it to see where it fails. Love debug printing Jul 03 21:03:36 data working btw Jul 03 21:03:41 pushed? Jul 03 21:03:54 not out of the box but it works after setting route and dns Jul 03 21:03:57 not yet Jul 03 21:04:04 efs thing working too Jul 03 21:04:25 efs as in nvdata loading in fwloader? Jul 03 21:05:09 yeah Jul 03 21:05:12 paulk-desktop: it would be nice to fix route/dns. or at least set it to 8.8.8.8 Jul 03 21:05:21 don't worry about it Jul 03 21:05:24 it should be an easy fix Jul 03 21:05:38