**** BEGIN LOGGING AT Sun Sep 02 02:59:59 2012 Sep 02 03:12:16 rebooting seems to have solved the g key issue Sep 02 07:07:14 mrmoku, ah, so not a ping, a pong. :) Alerady solved my issue, thanks. :) Sep 02 09:43:34 TAsn: hehe, yeah. And to not show the big latency I did ping...not pong ;) Sep 02 10:12:56 mrmoku, devious. ;P Sep 02 10:59:35 hi JaMa|Off Sep 02 10:59:56 the move to github is already effective ? Sep 02 11:06:53 yes Sep 02 11:33:27 GNUtoo, hello,what shall I do today? Sep 02 11:43:59 bbs Sep 02 11:58:50 ~nf Sep 02 11:58:50 The #openmoko-cdevel Newsflash Bulletin Board. (continued at ~NF2. For help see ~NF-help) - - - Recommended URLs and channels(chanlogs etc):see ~RL - - - NEWSFLASH [2012-03-28] server "agni.openmoko.org" got new hw, so no more downtime expected Sep 02 11:59:06 JaMa|Off, meta-smartphone is on github too? Sep 02 11:59:27 and one thing I notice, in static struct fb_ops msm_fb_ops {.fb_write = NULL} Sep 02 12:04:55 ayaka, hmmmm Sep 02 12:05:13 that would be really problematic Sep 02 12:05:21 maybe they use drm? Sep 02 12:05:26 like drm only? Sep 02 12:06:06 sorry, network problem Sep 02 12:11:25 np I'll look Sep 02 12:12:08 hmmm I don't have shooter_u sources here Sep 02 12:12:18 where are they already? Sep 02 12:12:36 I think "cat /dev/urandom > /dev/graphics/fb0" is done by user space in android Sep 02 12:12:42 can you use ssh? Sep 02 12:13:25 yes however it's slow as hell Sep 02 12:14:09 I see, I will try to use lxc, but can you wait some time(very long) or I paste the c file? Sep 02 12:15:28 yes pastebin the file Sep 02 12:15:39 but I can checkout the tgit Sep 02 12:15:41 *git Sep 02 12:15:46 with git remote add Sep 02 12:15:48 it will be fast Sep 02 12:17:27 sorry, I don't know how to use git server this time Sep 02 12:17:34 https://paste.debian.net/187187/ Sep 02 12:18:17 I notice "cat /dev/urandom > /dev/graphics/fb0" is different in my father's i9000 from mine in android Sep 02 12:19:10 his will be screen mess from top to below and mine only some screen mess happening randomly Sep 02 12:19:27 I think it is caused by user space program Sep 02 12:29:17 case MSMFB_SUSPEND_SW_REFRESHER: Sep 02 12:29:57 I guess you need that software refresher for X Sep 02 12:30:01 but that's after.... Sep 02 12:30:05 after we make it work Sep 02 12:31:19 ayaka, do you have DRM driver ? Sep 02 12:31:47 GNUtoo, yes I think I have, but I don't know whether it apply for msm Sep 02 12:31:47 because the msmfb driver you got seem very very bad Sep 02 12:31:52 ok Sep 02 12:32:02 it's another driver model....which is better Sep 02 12:32:04 bceause I can't compile msm_kgsl Sep 02 12:32:12 what's that? Sep 02 12:32:19 DRM is a driver model Sep 02 12:32:23 wait a moment Sep 02 12:32:28 which replaces framebuffer Sep 02 12:32:30 and v4l2 Sep 02 12:32:43 it offers a framebuffer abstraction compatibility Sep 02 12:33:54 DRM [=y] , but not Build a DRM interface for the MSM_KGSL driver Sep 02 12:34:06 it doesn't have any help, bad htc Sep 02 12:34:07 what is MSM_KSGL? Sep 02 12:34:11 ah ok Sep 02 12:34:15 I don't it either Sep 02 12:34:27 give me a link to the source of your kernel Sep 02 12:34:40 or I'll use the one from android project I gave you either Sep 02 12:34:46 can we force x to use drm, wait I momnet Sep 02 12:34:49 but I need some sources to think Sep 02 12:34:54 we can do that: Sep 02 12:35:02 DRM<->framebuffer<->Xorg Sep 02 12:35:16 with the framebuffer compatibility beeing in the kernel Sep 02 12:37:35 https://github.com/agrabren/android_kernel_htc_shooter Sep 02 12:37:40 is that source ok? Sep 02 12:37:46 does it support shooteru? Sep 02 12:37:51 http://dl4.htc.com/RomCode/Source_and_Binaries/shooteru-ics-crc-3.0.16-e733189.zip Sep 02 12:37:59 ah seem so Sep 02 12:38:03 GNUtoo: did you read my e-mail about repo move? Sep 02 12:38:05 I'll download the shooteru sources Sep 02 12:38:16 JaMa|Off, probably but then I forgott the details Sep 02 12:38:44 but I find htc realese the new version source, may I didn't use the correct version Sep 02 12:39:09 yes, I think the github one can Sep 02 12:39:39 try github kernel Sep 02 12:39:41 GNUtoo: please read at least this one http://lists.shr-project.org/pipermail/shr-devel/2012-August/004145.html Sep 02 12:39:56 ah right Sep 02 12:39:59 sorry JaMa|Off Sep 02 12:40:09 I read it already..... Sep 02 12:40:15 but I forget about the details Sep 02 12:40:42 thanks a lot!!!!! Sep 02 12:42:09 anyway I'll probably be offline some days Sep 02 12:42:15 because wifi cafee is closing Sep 02 12:47:13 GNUtoo: no problem Sep 02 12:47:23 GNUtoo, sorry, the github is slowly here, maybe you need wait a day to get my report, but shouteru has many Carrier and Region , they need different kernel sources, I don't what are the difference Sep 02 12:48:28 ok Sep 02 12:48:45 ayaka, then paste here that file: Sep 02 12:48:50 cat /proc/cmdline Sep 02 12:52:00 https://paste.debian.net/187192/ Sep 02 12:52:16 ayaka, https://github.com/agrabren/android_kernel_htc_shooter seem to have many many fixes Sep 02 12:52:43 I see Sep 02 12:53:05 hmmm with cat /proc/cmdline I guess that it selectes the right hardware: Sep 02 12:53:14 androidboot.hardware=shooteru Sep 02 12:53:48 but it's really slow to fetch: it didn't even start to count remote objects here Sep 02 12:53:58 and I think EU version can work, becuase that is what I use in debian, but the correct should be HTC-WWE Sep 02 12:55:07 ok Sep 02 12:55:16 it's really worth trying Sep 02 12:57:37 HTC-EU and htcAsia-WWE are the same file! Sep 02 12:57:56 I didn't notice bbefore Sep 02 12:59:24 does the framebuffer of src which you gave me support DRM? Sep 02 13:18:29 ayaka, warning: no common commits Sep 02 13:18:41 they're not good: they added the tarball + fixes Sep 02 13:18:51 GNUtoo, what is it? Sep 02 13:19:01 instead of extracting a patch of it Sep 02 13:19:11 basically they took the tarball Sep 02 13:19:16 did git add Sep 02 13:19:21 and fixed some stuff Sep 02 13:19:25 that's not the way to do it Sep 02 13:19:32 the way to do it is: Sep 02 13:19:34 vim Makefile Sep 02 13:19:37 oh, no the readme in zip is just told us how to compile it Sep 02 13:19:40 look at the kernel version Sep 02 13:19:49 diff with the tarball of the kernel version Sep 02 13:20:01 add it on top of the git of the same kernel version Sep 02 13:20:03 and push Sep 02 13:20:06 and then fix stuff Sep 02 13:20:12 that's the way to do it..... Sep 02 13:20:44 anyway we want their fixes Sep 02 13:21:41 anyway I've the sources now Sep 02 13:24:17 sorry, I can't understand, I only know git clone pull this time Sep 02 13:26:07 I got the github one, let me compile it Sep 02 13:35:54 ok Sep 02 13:37:25 GNUtoo, it seems that github's source has some error, I can't compile it Sep 02 13:37:37 what error? Sep 02 13:45:29 sorry, network problem, what I said just now? Sep 02 13:59:13 GNUtoo, ? Sep 02 13:59:33 GNUtoo, it seems that github's source has some error, I can't compile it Sep 02 13:59:33 what error? Sep 02 13:59:59 arch/arm/mach-msm/board-shooter.c:2553: error: unknown field 'exit_lpm_cb' specified in initializer Sep 02 14:00:06 and it seems no difference in msm_8x60/msm_fb.c Sep 02 14:00:27 then comment that part Sep 02 14:00:32 remove 'exit_lpm_cb' Sep 02 14:02:24 yes, I will Sep 02 14:03:54 oh, no another error happened, wait it finish Sep 02 14:07:52 ok Sep 02 14:08:12 mm/memcontrol.c:1251: error: conflicting types for 'mem_cgroup_isolate_pages' Sep 02 14:08:52 his source is full of mistakes, and there are in important part Sep 02 14:11:35 GNUtoo, but one thing, omap2 doesn't need write either Sep 02 14:11:48 ah ok Sep 02 14:11:49 strange Sep 02 14:11:52 let me look Sep 02 14:13:15 indeed Sep 02 14:13:16 /*.fb_write = omapfb_write,*/ Sep 02 14:13:34 but..... Sep 02 14:13:36 .fb_mmap = omapfb_mmap, Sep 02 14:14:09 beside that are callbacks for the framebuffer driver abstraction stuff Sep 02 14:14:27 .fb_mmap = msm_fb_mmap Sep 02 14:14:29 it's not stuff for the userspace, at least not directly Sep 02 14:14:30 ok Sep 02 14:14:48 also you could look in Documentation/ Sep 02 14:14:56 if they have something about the required callbacks Sep 02 14:16:12 I don't know, what will the files be named? actually, do we write to /dev/fb0 to be done by user space? Sep 02 14:18:36 and is the msm_fb for fb console or the whole framebuffer? Sep 02 14:29:35 GNUtoo, in linux device driver chapter 15, as it use mmap, I think the write is done in user space Sep 02 14:31:23 yes it's done by userspace Sep 02 14:31:34 but I didn't meant that Sep 02 14:32:11 I meant: Sep 02 14:32:26 userspace<->driver model<->framebuffer driver Sep 02 14:32:53 which mean that between the .mmap of the msmfb driver there is the framebuffer driver model Sep 02 14:32:57 like framebuffer core Sep 02 14:34:47 even what shall we do? Sep 02 14:34:54 s/even/then/ Sep 02 14:34:56 ayaka meant: then what shall we do? Sep 02 14:39:43 thank you apt Sep 02 14:40:01 apt seems a bot? Sep 02 14:42:42 ayaka, yes it's a bot Sep 02 14:42:45 bbl Sep 02 14:52:36 when you back , please call me Sep 02 15:09:23 sorry, my ISP is too bad Sep 02 15:35:55 GNUtoo, good bye, I have to sleep Sep 02 15:40:57 ok Sep 02 16:30:49 morphis, hi! Sep 02 16:31:03 can I ask you a bit about how GTA04 gets SMS messages? Sep 02 16:35:15 paulk-desktop: heyho Sep 02 16:35:30 paulk-desktop: yes you can but I will need some time to answer (busy with other things) Sep 02 16:35:39 sure, no problem at all Sep 02 16:35:40 thanks Sep 02 19:02:03 paulk-desktop: hi Sep 02 19:02:22 paulk-desktop: dunno if you have time to fix gprs on GB crespo, but I have a patch for a rssi segfault https://gitorious.org/~ast/replicant/asts-hardware_ril_samsung-ril/commit/541e79bfbae37e55672ed3ff05f9f37d92ba95ac Sep 02 19:02:23 hi Alex Sep 02 19:02:50 paulk-desktop: tbh, I think we need to take time and figure out all null-pointer thingy. or switch to ofono and then to meego :) Sep 02 19:03:24 yeah, what is about all that null-pointer stuff? Sep 02 19:03:39 I didn't ever see that happening on nexus s Sep 02 19:04:01 like responses that are supposed to have data but that don't send any on, for instance, galaxy nexus Sep 02 19:04:02 paulk-desktop: I wonder why we get segfaults in different places. Probably that comes somewhere from the ipc reading part Sep 02 19:04:19 segfaults are because I always assume data where there should be data Sep 02 19:04:44 and we should make a global commit that checks the pointer on all dispatch functions Sep 02 19:04:46 well, I'm the adept of the school of defensive programming so I add checks everywhere when I write code Sep 02 19:04:53 you're right Sep 02 19:05:08 though I didn't do it at the time Sep 02 19:05:14 ok, we'll get back to it later Sep 02 19:05:30 but the real question is to know if the modem send these requests without data or if we just miss it at some point Sep 02 19:05:35 for now, getting mer to run and porting sgs2 to an upstream kernel is more interesting Sep 02 19:05:43 okay! Sep 02 19:05:48 so you're trying mer? Sep 02 19:06:09 also, someone tried to port the nexus s to meego Sep 02 19:06:43 well, since galaxy nexus has omap4 and that has PVR blobs for X11, I can try mer. if it turns out to be usable, I'll have to think what to do about graphics. Though I can live with blobbed GL provided that RIL is free Sep 02 19:07:42 ok Sep 02 19:07:56 I'll try to make ICS work with free gralloc on the devices I have btw Sep 02 19:08:02 including the sgs2 Sep 02 19:08:54 paulk-desktop: Well, it will just work if you turn off GL in framework (as described in linaro) Sep 02 19:09:11 last time I checked it was flickering on nexus s Sep 02 19:09:17 (it was running though) Sep 02 19:09:22 with linaro patches **** ENDING LOGGING AT Mon Sep 03 03:00:00 2012