**** BEGIN LOGGING AT Sun Nov 13 02:59:59 2016
Nov 13 07:31:41 <freemangordon> jonwil: hi!
Nov 13 07:31:50 <jonwil> hi
Nov 13 07:34:16 <freemangordon> jonwil: feeling adventurous? https://github.com/community-ssu/iphbd :)
Nov 13 07:34:43 <freemangordon> tested on x86_64, seems working
Nov 13 07:34:43 <jonwil> Yeah I saw that g
Nov 13 07:34:46 <jonwil> good job :)
Nov 13 07:35:01 <jonwil> Whats next? :)
Nov 13 07:35:04 <freemangordon> but without the kernel module, so
Nov 13 07:35:23 <freemangordon> well, this is all done in the context of porting ildon to gtk3
Nov 13 07:35:27 <freemangordon> *hildon
Nov 13 07:35:40 <freemangordon> so it depends on what will be needed next
Nov 13 07:36:48 <jonwil> libtime/clockd seem like useful targets (they are used all over the system by all sorts of things)
Nov 13 07:37:28 <freemangordon> yeah, I was thinking about it
Nov 13 07:38:01 <jonwil> sysinfod/libsysinfo/etc may also be worthwhile
Nov 13 07:38:18 <freemangordon> yep
Nov 13 07:41:41 <jonwil> oh and libplayback
Nov 13 07:43:32 <freemangordon> hmm, only maemo kernel has iphb.ko, no such module in the other nokia kernels
Nov 13 07:55:33 <bencoh> is that supposed to wake apps in "sync" in order not to wake up cpu too much?
Nov 13 08:00:20 <freemangordon> yes
Nov 13 08:25:00 <jonwil> Still getting nowhere with my entry for the maemo coding contest :(
Nov 13 10:02:47 <parazyd> freemangordon: isn't iphb a netfilter thing?
Nov 13 10:04:49 <parazyd> freemangordon: i have the source for it if needed btw
Nov 13 10:33:56 <jonwil> source to what? The source for the kernel module is in the standard Maemo kernel source, the source to libiphb was originally released by Nokia in a Git repo and the source for iphbd was reverse engineered by freemangordon (and along with libiphb now lives in https://github.com/community-ssu/iphbd)
Nov 13 10:36:17 <jonwil> Its always good when another entry in the table on http://wiki.maemo.org/Fremantle_closed_packages turns to green because someone reverse engineered another package :)
Nov 13 10:37:38 <jonwil> I just wish there was more I could reverse engineer
Nov 13 10:37:46 <jonwil> and help turn table entries green via FOSS clones
Nov 13 10:38:29 <bencoh> :)
Nov 13 10:40:54 <jonwil> I think there are many entries in that table that will never turn green though
Nov 13 10:41:46 <jonwil> csd-* for example
Nov 13 10:44:28 <jonwil> or pulseaudio-module-nokia-*
Nov 13 10:44:37 <jonwil> well pulseaudio-module-nokia-voice
Nov 13 10:44:44 <jonwil> or rtcom-call-ui
Nov 13 10:44:54 <jonwil> or tablet-browser-*
Nov 13 10:46:32 <parazyd> well in due time :)
Nov 13 10:46:37 <parazyd> things are moving well
Nov 13 10:47:20 <jonwil> I dont think there is anyone out there with both the skills AND the time/interest to reverse engineer the really tricky stuff (pulseaudio-module-nokia-voice being right up there)
Nov 13 10:48:20 <jonwil> Short of a miracle occurring and the code falling out of the sky or something, I doubt we will ever have a fully reverse engineered clone of pulseaudio-module-nokia-voice
Nov 13 10:48:35 <jonwil> Which is a pitty since its one of the key pieces needed to get Maemo working on the Neo900...
Nov 13 10:49:04 <bencoh> pretty sure it can do without
Nov 13 10:49:21 <jonwil> Trust me, it cant
Nov 13 10:50:26 <jonwil> I have been deep inside pulseuaudio-module-nokia-voice and the Maemo audio system in general and I can tell you its going to be needed (or at least some of what it does is needed) if we want usable audio on the thing with Maemo
Nov 13 11:14:54 * jonwil wishes he could make progress on his entry for the Maemo coding contest... :(
Nov 13 11:37:12 <Pali> freemangordon: iphb.ko for dkms is ready: https://github.com/community-ssu/iphb-dkms
Nov 13 12:06:47 <freemangordon> Pali: I ported the driver for newer kernel, unfortunately it crashes both 3.16 and 4.4 :(
Nov 13 12:07:20 <freemangordon> Pali: ali, I think dkms should be included in iphb
Nov 13 12:07:26 <freemangordon> *also
Nov 13 12:07:42 <freemangordon> however, lets ckech those oopses
Nov 13 12:07:47 <freemangordon> *check
Nov 13 12:08:02 <Pali> did you tried that my pushed version?
Nov 13 12:08:10 <freemangordon> no
Nov 13 12:08:14 <Pali> so use it
Nov 13 12:08:18 <Pali> I ported some functions
Nov 13 12:08:22 <freemangordon> hmm
Nov 13 12:08:25 <freemangordon> ok
Nov 13 12:08:35 <Pali> there are #ifdefs for 3.13 and 4.1 kernels
Nov 13 12:08:54 <freemangordon> same here
Nov 13 12:09:07 <freemangordon> but, you use the one from 2.6.32
Nov 13 12:09:58 <Pali> yes, from 2.6.32
Nov 13 12:10:06 <Pali> but there are only minimal changes
Nov 13 12:10:16 <Pali> plus it containd kernel-power null deref fix
Nov 13 12:11:07 <freemangordon> yes, I have that fix here as well
Nov 13 12:11:15 <freemangordon> and yes, the changes are minimal
Nov 13 12:11:28 <freemangordon> Pali: gimme some time to gather the oops backtrace
Nov 13 12:11:33 <Pali> ok
Nov 13 12:30:00 <jonwil> I wish my skills allowed me to contribute more to the N900 community.
Nov 13 12:30:11 <jon_y> port USB type-C drivers to maeomo :)
Nov 13 12:30:14 <jonwil> But what I have contributed to date certainly isn't useless
Nov 13 12:30:28 <jonwil> And I dont know the first thing about USB or writing kernel drivers
Nov 13 12:30:41 <jon_y> eh, its just plain C code
Nov 13 12:31:06 <jon_y> you just need to trace the code
Nov 13 12:31:17 <jon_y> and see how it works underneath
Nov 13 12:31:30 <jonwil> I know enough about kernel drivers to know that I dont know enough about kernel drivers to even think about trying to write one
Nov 13 12:31:35 <jonwil> Or port one to a different kernel
Nov 13 12:31:40 <jonwil> Or even try to understand what one is doing
Nov 13 12:32:10 <jonwil> Its hard enough to understand when userspace bits are talking to kernel drivers (like all the stuff I had to do for the MCE reverse engineering work)
Nov 13 12:32:27 <jon_y> char device node?
Nov 13 12:33:02 <jonwil> anyhow, anything involving kernel hacking is not on my radar screen
Nov 13 12:34:56 <jon_y> what other pointless porting work can be done :)
Nov 13 12:35:08 <jon_y> SATA drivers for maemo :)
Nov 13 12:36:46 <jonwil> Not interested in anything kernel related or driver related or hardware related
Nov 13 12:37:10 <jonwil> I keep looking at http://wiki.maemo.org/Fremantle_closed_packages and wishing there was something left on that list that I could reverse engineer
Nov 13 12:37:24 <jonwil> but all the good stuff is too hard :(
Nov 13 12:38:41 <jon_y> NFC drivers for Maemo!
Nov 13 12:38:52 <jon_y> not that the drivers will be useful
Nov 13 12:38:55 <jonwil> stop talking about hardware or kernel stuff :P
Nov 13 12:39:20 <jon_y> k, port JACK audio to maemo :)
Nov 13 12:39:37 <jonwil> stop taking about useless ports as well :P
Nov 13 12:39:39 <Pali> jonwil: alsaped?
Nov 13 12:39:52 <jon_y> hmm, not sure if Maemo would benefit from zram
Nov 13 12:39:53 <jonwil> looked into that, its too complex for me to tackle
Nov 13 12:40:12 <KotCzarny> isnt zram almost free with mainline?
Nov 13 12:40:27 <jon_y> does it work with powerkernel?
Nov 13 12:42:52 <jon_y> I wonder how hard is it to hack the N900 firmware
Nov 13 12:43:13 <KotCzarny> because ugly closed binary code
Nov 13 12:45:50 <freemangordon> Pali: hmm, actually my ko seems to work fine with 3.16
Nov 13 12:45:52 <jon_y> hmm does Linphone work on the N900?
Nov 13 12:46:09 <jon_y> sip with SRTP/ZRTP encrypted channels
Nov 13 12:46:12 <freemangordon> the oopses wew because I put some msleep() calls
Nov 13 12:46:16 <jon_y> might be too heavy for N900
Nov 13 12:46:17 <freemangordon> *were
Nov 13 12:46:48 <freemangordon> Pali: BTW, I don;t think iphb from harmattan wil work with maemo's iphbd
Nov 13 12:46:56 <Pali> really?
Nov 13 12:47:15 <Pali> ipbh.ko from 2.6.32 kernel will not work with maemo iphbd?
Nov 13 12:47:25 <freemangordon> I think so
Nov 13 12:47:39 <freemangordon> that timer, why it is there?
Nov 13 12:47:57 <freemangordon> iiuc it is iphbd that have to flush keepalives
Nov 13 12:48:01 <freemangordon> *has
Nov 13 12:54:38 <brolin_empey> joga: auenf: KotCzarny: IE11 on Win10 continues playing audio after Windows turns off the computer’s displays after an inactivity timeout.  So yes, the unwanted behaviour about which I asked was caused by Edge, not by Windows.
Nov 13 12:55:19 <KotCzarny> what sane person uses win10?
Nov 13 12:56:36 <freemangordon> Pali: do you have kernel newer than 3.16 to test iphb.ko with?
Nov 13 12:56:37 <brolin_empey> KotCzarny: I use WIA or TWAIN, not SANE. ;-)
Nov 13 12:56:51 <Pali> I'm using 3.13
Nov 13 12:56:55 <freemangordon> hmm
Nov 13 12:56:57 <Pali> LTS ubuntu 12.04
Nov 13 12:57:19 <freemangordon> well, I meant some VM
Nov 13 12:58:02 <freemangordon> as it has crashed my 4.4 so it is better to not play in native OS
Nov 13 13:00:53 <Pali> ok
Nov 13 13:02:19 <jon_y> do you guys play with bleeding edge kernels?
Nov 13 13:04:37 <freemangordon> Pali: https://github.com/community-ssu/iphbd/tree/master/kmod
Nov 13 13:09:15 <Pali> https://gitorious.org/meego-middleware/dsme/commit/735a5f058e082ff63e3cf65579def48b157b6705
Nov 13 13:09:21 <Pali> maybe this commit can be interesting...
Nov 13 13:09:28 <Pali> first public version of iphb
Nov 13 13:10:08 <Pali> anyway, also after certman update, I cannot connect to gmail imap server due to unknown certificate with bad signature
Nov 13 13:10:13 <Pali> jonwil: ^^^
Nov 13 13:10:59 <freemangordon> Pali: I used part of the code there, but it is not the same as the one in Fremantle
Nov 13 13:11:06 <Pali> ok
Nov 13 13:11:14 <freemangordon> that one uses epoll, maemo uses poll
Nov 13 13:17:39 <freemangordon> Pali: kernel 4.4 http://pastebin.com/u3dcJZLA
Nov 13 13:34:38 <Pali> freemangordon: gdb iphb.ko
Nov 13 13:34:43 <Pali> and look at flush_keepalives+0x143/0x220
Nov 13 13:37:08 <freemangordon> Pali: no symbols :(
Nov 13 13:37:18 <freemangordon> how to build it with debug info?
Nov 13 13:38:21 <freemangordon> hmm, found it how
Nov 13 13:39:33 <Pali> freemangordon: KCFLAGS=-g
Nov 13 13:39:42 <Pali> $ make -C /usr/src/linux-headers-3.13.0-101-generic/ M=$PWD KCFLAGS=-g
Nov 13 13:40:45 <Pali> just recompile and run:
Nov 13 13:40:46 <Pali> $ gdb iphb.ko
Nov 13 13:40:46 <Pali> disassemble /m flush_keepalives+0x143/0x220
Nov 13 13:44:15 <freemangordon> Pali: I modified Makefile to include debug symbols, unfortunately, there is no such address (0x143) in the disassembly
Nov 13 13:44:48 <freemangordon> will try with ida
Nov 13 13:47:46 <freemangordon> Pali: https://github.com/community-ssu/iphbd/blob/master/kmod/iphb.c#L174
Nov 13 13:47:51 <freemangordon> as I expected :)
Nov 13 13:49:05 <freemangordon> seems I can't store  nf_hook_state, I guess I have to keep the data, lemme try
Nov 13 13:58:14 <freemangordon> looks better that way :)
Nov 13 13:59:57 <freemangordon> Pali: what do you think? https://github.com/community-ssu/iphbd/commit/5c02e268587b50a1039e7d9e1dd65543efd3769d
Nov 13 14:00:43 <freemangordon> crashes no more my ubuntu 14.04 with kernel 4.4
Nov 13 14:13:21 <xes> do we have any kind of draft / project for an emergency mode for our N900, something like a boot option to boot the N900 and start only the phone-ui instead of the whole maemo desktop... It would be interesting since if you need to place an emergency phone call and the N900 is turned off sometimes it may needs more than 5 minutes to be ready
Nov 13 14:25:28 <KotCzarny> xes: neat idea, could be work as a recovery os too
Nov 13 14:25:34 <KotCzarny> *could work
Nov 13 14:40:31 <KotCzarny> case of dying wlans ta doo doom
Nov 13 14:40:39 <KotCzarny> opps. wrong chan
Nov 13 15:01:53 <freemangordon1> Pali: what about merging iphb-dkms with iphbd?
Nov 13 15:02:15 <Pali> iphb-dkms is not needed in cssu/maemo kernel
Nov 13 15:02:25 <freemangordon> is there a reason to keep ko split from the daemon?
Nov 13 15:02:27 <Pali> so I guess it would be better to have them separated
Nov 13 15:02:45 <freemangordon> well, I guess we can check during buildtime
Nov 13 15:02:45 <Pali> we do not want iphb-dkms in cssu repository
Nov 13 15:03:18 <freemangordon> and not install dkms/ko part if building on <3.x kernel
Nov 13 15:03:22 <Pali> you cannot (easily) change list of deb packages produced by debian source package
Nov 13 15:03:56 <Pali> list of deb packages is static and not changed by building process (debian/rules)...
Nov 13 15:04:01 <freemangordon> well, but then iphbd should depend on iphb-dkms, so isn't that the same?
Nov 13 15:04:24 <Pali> iphbd should depends on some virtual/meta package
Nov 13 15:04:40 <freemangordon> hmm, right
Nov 13 15:04:40 <Pali> which is provided by both iphb-dkms and also by stock kernel
Nov 13 15:04:57 <Pali> but that is hard in maemo, because stock kernel does not do that
Nov 13 15:05:11 <Pali> I would rather do not depends on any such package...
Nov 13 15:05:26 <freemangordon> what about kernel-feature?
Nov 13 15:05:35 <Pali> once we will have something working on debian/devuan/... we can add recomended dependency
Nov 13 15:05:50 <freemangordon> kernel-feature-something || iphb-dkms
Nov 13 15:06:05 <freemangordon> ok
Nov 13 15:06:22 <freemangordon> going to move iphb from the daemon to iphb-dkms
Nov 13 15:07:50 <freemangordon> Pali: shell i include -g in the Makefile?
Nov 13 15:08:09 <Pali> no
Nov 13 15:08:11 <freemangordon> ok
Nov 13 15:08:13 <Pali> that is big hack
Nov 13 15:08:23 <freemangordon> ok
Nov 13 15:14:11 <freemangordon> Pali: shall I keep the timed flush?
Nov 13 15:14:18 <freemangordon> from the harm kernel?
Nov 13 15:23:40 <Pali> maybe keep it for now
Nov 13 15:23:45 <freemangordon> ok
Nov 13 15:23:53 <Pali> when we find out that is cause really problems, we can delete it
Nov 13 15:23:58 <freemangordon> ok
Nov 13 15:24:49 <freemangordon> but I'll remove(#if 0) the read, as fremantle iphbd doesn;t read on the kernel fd. 
Nov 13 15:25:14 <Pali> I think no need...
Nov 13 15:25:36 <freemangordon> there is, see https://github.com/community-ssu/iphb-dkms/blob/master/iphb.c#L246
Nov 13 15:26:00 <Pali> and it is problem?
Nov 13 15:26:09 <Pali> I think not, just it is not called...
Nov 13 15:26:17 <freemangordon> in fremantle this is done in iphbd_poll
Nov 13 15:27:00 <Pali> ah, right
Nov 13 15:27:16 <Pali> it is really in iphbd_poll and harmattan kernel removed it
Nov 13 15:27:24 <freemangordon> :nod:
Nov 13 15:28:02 <freemangordon> I guess we can keep both
Nov 13 15:28:36 <Pali> ok
Nov 13 15:28:42 <freemangordon> but better #if the read code
Nov 13 15:28:53 <Pali> read code seems ok
Nov 13 15:28:59 <Pali> if it is not used, then nothing is called
Nov 13 15:29:16 <freemangordon> ok, will keep it
Nov 13 16:09:30 <freemangordon> Pali: https://github.com/community-ssu/iphb-dkms/commit/37d8be993bf23ebcd4a72f5cd128f587181b327f
Nov 13 16:09:58 <Pali> why UTS_RELEASE?
Nov 13 16:10:04 <Pali> I removed that crap...
Nov 13 16:10:09 <freemangordon> :)
Nov 13 16:10:16 <freemangordon> well, this is in fremantle
Nov 13 16:11:01 <freemangordon> Pali: feel free to remove it again, I wouldn;t mind
Nov 13 16:11:11 <Pali> #if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 29)
Nov 13 16:11:15 <Pali> thats look wrong...
Nov 13 16:11:46 <freemangordon> I couldn;t find at which kernel version that header was moved, thus...
Nov 13 16:12:08 <Pali> so you break it on my kernel too :-)
Nov 13 16:12:22 <freemangordon> wait, why?
Nov 13 16:12:44 <Pali> beacuse THIS_MODULE is needed
Nov 13 16:13:17 <freemangordon> aah, you mean for the hooks
Nov 13 16:13:41 <freemangordon> yeah, sorry, copy/paste error, going to fix it
Nov 13 16:14:47 <Pali> -				const struct net_device *indev,
Nov 13 16:14:47 <Pali> +				const struct net_device *in,
Nov 13 16:14:50 <Pali> reason?
Nov 13 16:15:08 <Pali> it is hard to read that diff
Nov 13 16:15:17 <Pali> couple of noop changes
Nov 13 16:15:35 <freemangordon> for name consistency
Nov 13 16:17:00 <Pali> #include <generated/utsrelease.h>
Nov 13 16:17:00 <Pali> ??
Nov 13 16:17:09 <Pali> this looks really bad
Nov 13 16:17:17 <freemangordon> needed for UTS_RELEASE ;)
Nov 13 16:17:49 <freemangordon> hmm, seems owner disappears in 4.4
Nov 13 16:17:52 <Pali> linux/vermagic.h if you really need it
Nov 13 16:18:26 <Pali> ideally do not mix 10 different things into one commit
Nov 13 16:18:38 <freemangordon> Pali: well, I don;t need it, it is just that the code was like that, feel free to remove it
Nov 13 16:18:48 <freemangordon> Pali: 
Nov 13 16:18:48 <Pali> I already removed it
Nov 13 16:18:55 <Pali> but you reverted my remove :-)
Nov 13 16:19:01 <freemangordon> it is not LKML :p
Nov 13 16:19:14 <freemangordon> ok, will reintroduce it
Nov 13 16:19:26 <Pali> well.... I can fix it
Nov 13 16:20:23 <freemangordon> Pali: lemme first push THIS_MODULE fix
Nov 13 16:22:39 <Pali> that is incomplete too... okfn usage, etc...
Nov 13 16:23:37 <freemangordon> ready
Nov 13 16:23:42 <freemangordon> which line?
Nov 13 16:23:59 <freemangordon> okfn usage should be correct
Nov 13 16:24:07 <freemangordon> I tested on 3.16 and 4.4
Nov 13 16:24:27 <Pali> I will fix it...
Nov 13 16:24:33 <freemangordon> which one?
Nov 13 16:24:43 <Pali> will create new branch and fix it here
Nov 13 16:25:35 <freemangordon> Pali: please, explain what are you going to fix. also, why new branch?
Nov 13 16:25:53 <Pali> to prevent push conflicts
Nov 13 16:25:54 <Pali> in master
Nov 13 16:26:00 <Pali> fix - okfn usage
Nov 13 16:26:13 <freemangordon> I pushed everything I have
Nov 13 17:12:09 <Pali> freemangordon: https://github.com/community-ssu/iphb-dkms/commits/pali
Nov 13 17:12:24 <Pali> I check that kernel versions against git kernel history
Nov 13 17:13:01 <Pali> this should be correct
Nov 13 17:31:55 <freemangordon> Pali: why not merge it with master?
Nov 13 17:32:30 <Pali> because now there are conflics :D
Nov 13 17:32:52 <Pali> I started earlier as you pushed changes
Nov 13 17:33:07 <Pali> I can (only) force push it into master...
Nov 13 17:34:00 <freemangordon> whatever is needed, I don;t see why we need branch "pali" :)
Nov 13 17:34:13 <Pali> recheck if it is working
Nov 13 17:34:22 <Pali> and I can force push it to master
Nov 13 19:08:33 <DocScrutinizer05> xes: FIVE minutes?
Nov 13 19:08:37 <DocScrutinizer05> never seen that
Nov 13 19:09:18 <DocScrutinizer05> when you got device lock, you can do 911 calls instead of enetring the unlock code. This is about as early as it gets
Nov 13 19:11:11 <DocScrutinizer05> xes: an alternative was to send "ATD911;" to modem (via ISI - *UGH* hardly feasible) from uBoot menu
Nov 13 20:37:12 <freemangordon> Pali: make KERNELRELEASE=4.4.0-47-generic -C /lib/modules/4.4.0-47-generic/build M=/var/lib/dkms/iphb/1.0/build....(bad exit status: 2)
Nov 13 20:45:00 <Pali> what is in log?
Nov 13 20:45:44 <freemangordon>  /var/lib/dkms/iphb/1.0/build/iphb.c:396:24: error: ‘const struct nf_hook_state’ has no member named ‘net’
Nov 13 20:48:35 <freemangordon> Pali: DKMSKernelVersion: 4.2.0-38-generic
Nov 13 20:50:12 <freemangordon> http://lxr.free-electrons.com/source/include/linux/netfilter.h?v=4.2 - no net there
Nov 13 20:50:26 <freemangordon> it appears in 4.4
Nov 13 20:50:48 <freemangordon> Pali: are you going to fix it or I to fix it?
Nov 13 20:52:58 <freemangordon> hmm, something is weird here
Nov 13 20:53:08 <xes> DocScrutinizer05: how much time do your N900 takes to go from power off to have the dialer ready to place a call?
Nov 13 20:54:00 <DocScrutinizer05> like maybe 90s, I never even tried since I don't know when the dialer is "ready"
Nov 13 20:54:02 <jonwil> hi
Nov 13 20:54:25 <DocScrutinizer05> might be 60
Nov 13 21:00:39 <DocScrutinizer05> yeah, desktop is up after 45s, probably usable after 60 - though on this particular device all icons (cache) are broken
Nov 13 21:02:16 <freemangordon> Pali: sorry, I gave you wrong log, this one is correct http://pastebin.com/j4Yv3Kwc
Nov 13 21:02:26 <freemangordon> anyway, I am going to fix that
Nov 13 21:03:24 <xes> DocScrutinizer05: my device is probably overloaded... but i can open the phone ui after 4 minutes.. then the system becomes reactive after > 5 minutes
Nov 13 21:03:58 <DocScrutinizer05> that's weird
Nov 13 21:11:56 <freemangordon> Pali: https://github.com/community-ssu/iphb-dkms/commit/3bc377199a3df5be22d39f667f666a3d26ee3a79
Nov 13 21:12:14 <freemangordon> lets see if it still works
Nov 13 21:32:18 <freemangordon> Pali: seems to work ok, please push to master
Nov 13 23:18:08 <Pali> freemangordon: git push --force origin 3bc3771:master was done
**** ENDING LOGGING AT Mon Nov 14 03:00:00 2016