**** BEGIN LOGGING AT Sat Aug 27 02:59:57 2011
Aug 27 06:51:20 <ka6sox> cryptk, lets walk first.
Aug 27 06:51:47 <ka6sox> get angstrom root_fs pivot rooted to.
Aug 27 06:51:48 <cryptk> oddly... the bootdevice setting in bootie points to mmc0p12... I would have thought it would be mmcblk0p12
Aug 27 06:52:28 <ka6sox> ya, I would have thought that too
Aug 27 06:52:53 <ka6sox> lets get userspace working first
Aug 27 06:53:13 <ka6sox> add any kernel modules to /lib/modules/* and use that first.
Aug 27 06:53:26 <cryptk> I already have angstrom working in a chroot
Aug 27 06:53:33 <ka6sox> riht.
Aug 27 06:53:34 <cryptk> I have had that working for a LONG time
Aug 27 06:53:49 <ka6sox> but to make it boot to that rootfs you need /lib/modules/*
Aug 27 06:54:21 <cryptk> my first goal is to just get bootie to boot a different kernel... such as booting the stock kernel even though F15 is installed
Aug 27 06:54:47 <ka6sox> often when I get something I need to dev linux on it comes with Angstrom kernel and I get the modules and then build a rootfs, and pivot root to it
Aug 27 06:54:59 <ka6sox> you need to work it backwards.
Aug 27 06:55:09 <ka6sox> makes it much easier.
Aug 27 06:56:35 <ka6sox> it doesn't HAVE to be a Angstrom kernel.
Aug 27 06:56:42 <ka6sox> and indeed should probably NOT be.
Aug 27 06:56:58 <ka6sox> since it won't know about device specific things for the TP.
Aug 27 06:57:42 <cryptk> well, I meant a kernel compiled for the purposes of running angstrom on a TP
Aug 27 06:57:53 <cryptk> since the palm kernel may not work
Aug 27 06:57:58 <cryptk> it may need it's own kernel
Aug 27 06:58:28 <ka6sox> do you have udev, devfs or what in your rootfs?
Aug 27 06:59:03 <ka6sox> thats biting off a LOT to make happne.
Aug 27 06:59:10 <ka6sox> break it down
Aug 27 06:59:16 <cryptk> udev
Aug 27 06:59:30 <cryptk> it is a beagleboard build of angstrom with e for a gui
Aug 27 06:59:55 <ka6sox> what devfs did you specify?
Aug 27 07:00:02 <ka6sox> they support a number
Aug 27 07:00:22 <ka6sox> well..since it works we don't need to mess with that
Aug 27 07:00:38 <ka6sox> use the stock kernel first.
Aug 27 07:00:45 <cryptk> yeah, the chroot is verified as fully functional
Aug 27 07:00:50 <ka6sox> lets remove as many variables as possible.
Aug 27 07:01:12 <cryptk> I have it located in /media/ext3fs/angstrom-e which is on the ext3fs partition created by metadoc
Aug 27 07:01:27 <ka6sox> okay
Aug 27 07:01:46 <ka6sox> let me think about how to safely do what we need to do.
Aug 27 07:02:08 <cryptk> I can (and for this likely will) adjust it to where the angstrom rootfs is directly in the ext3fs partition rather than in a folder
Aug 27 07:02:16 <ka6sox> cryptk, what we really need to do is unlink the pivot root
Aug 27 07:02:25 <ka6sox> point it back to /boot
Aug 27 07:02:36 <ka6sox> however
Aug 27 07:02:49 <ka6sox> I am not sure that novaterm is supported @ that point
Aug 27 07:03:05 <cryptk> it likely won't be
Aug 27 07:03:08 <ka6sox> ie we need a "console" so that you can pivot root to angstrom.
Aug 27 07:03:19 <cryptk> luckily, changing the env settings in bootie is not persistent
Aug 27 07:04:00 <cryptk> update for kindle app is out
Aug 27 07:04:06 <cryptk> no quick office update yet
Aug 27 07:04:08 <ka6sox> what we need is a kernel location + a initramfs location var
Aug 27 07:04:39 <ka6sox> and mod initramfs scripts
Aug 27 07:04:53 <ka6sox> so you can point to either webos or angstrom.
Aug 27 07:05:01 <ka6sox> that is how I bring it up.
Aug 27 07:05:12 <ka6sox> leaving the kernel the *same*
Aug 27 07:07:18 <ka6sox> cryptk, is there a initramfs arg?
Aug 27 07:07:44 <cryptk> http://www.webos-internals.org/wiki/Bootie#The_boot_environment_on_a_TouchPad_.283.0.2.29_follows:
Aug 27 07:07:49 <cryptk> that's everything there is
Aug 27 07:08:54 <cryptk> ok,  /media/ext3fs is now just a raw angstrom rootfs directly in the root of the LVM
Aug 27 07:09:01 <cryptk> so that part is done
Aug 27 07:09:08 <cryptk> now what would step 2 be...
Aug 27 07:09:26 <cryptk> and bootie can boot a uImage file that is a combination of a kernel and an intiramfs
Aug 27 07:10:05 <ka6sox> okay I see how it works now
Aug 27 07:10:35 <ka6sox> so we need to create a kernel with the initramfs compiled in
Aug 27 07:10:57 <ka6sox> then memboot thats
Aug 27 07:11:31 <ka6sox> this is what ubuntu chroot ought to do a great job to make this happen.
Aug 27 07:12:13 <ka6sox> lets see if the uboot tools are available in the armel port
Aug 27 07:12:47 <ka6sox> if they are then I'll compile it native and create an initramfs then uImage all that.
Aug 27 07:12:57 <cryptk> cool
Aug 27 07:14:26 <ka6sox> if they aren't I can still do it using my machine @ the office to do the mkuImage
Aug 27 07:14:52 <cryptk> I am assuming that the initram is what actually boots it from the LVM
Aug 27 07:15:27 <ka6sox> the kernel does not have the ability to boot from LVM (although 2.6.35 kernels can)
Aug 27 07:15:47 <cryptk> well, something tells it that the rootfs is in /dev/store/root
Aug 27 07:15:50 <ka6sox> so it goes to a ext2 partition it does know how to deal with.
Aug 27 07:16:00 <ka6sox> yes
Aug 27 07:16:04 <ka6sox> but /boot is ext2
Aug 27 07:16:09 <cryptk> yes
Aug 27 07:16:17 <cryptk> that isn't what I am talking about
Aug 27 07:16:18 <ka6sox> the kernel knows how to deal withthat.
Aug 27 07:16:26 <cryptk> the root line is just root=/dev/mmcblk0p13 rootwait ro
Aug 27 07:16:33 <ka6sox> correct
Aug 27 07:16:34 <cryptk> so something points it to the proper LVM for the rootfs
Aug 27 07:17:11 <ka6sox> yes, lvm is started and configured, then at the end pivot rooted to ..bin/busybox foo
Aug 27 07:17:24 <cryptk> yep
Aug 27 07:17:43 <ka6sox> that passes it to the lvm and the upstart scripts are run.
Aug 27 07:18:23 <ka6sox> what we want to do is pivot_root to angstrom instead.
Aug 27 07:18:25 <ka6sox> now...
Aug 27 07:18:38 <ka6sox> be aware that many things might break
Aug 27 07:19:02 <cryptk> I don't foresee this breaking anything that a doc won't fix though
Aug 27 07:19:03 <ka6sox> first thing I would be sure of is that we have usbnet configured with ssh.
Aug 27 07:19:31 <ka6sox> since we probably won't hae touch or much beyond fbdev output
Aug 27 07:19:39 <cryptk> yep
Aug 27 07:20:13 <ka6sox> so make sure you have an ssh daemon running against the usbnet device
Aug 27 07:20:32 <ka6sox> from inside the angstrom rootfs.
Aug 27 07:20:54 <ka6sox> btw..this is *exactly* how I did the last tablet I ported.
Aug 27 07:21:39 <cryptk> I should probably be working on this when I have actually had sleep
Aug 27 07:21:46 <ka6sox> right
Aug 27 07:21:47 <cryptk> which I am severely lacking atm
Aug 27 07:22:00 <ka6sox> I can do the kernel stuff ahead of time.
Aug 27 07:22:07 <ka6sox> lessee...
Aug 27 07:22:10 <cryptk> kk, you get your part ready
Aug 27 07:22:28 <cryptk> either sunday or monday I have tinker training with scout scheduled
Aug 27 07:22:39 <ka6sox> okay let me see if I can either get usb_console or usb_net working in the initramfs.
Aug 27 07:22:41 <cryptk> whichever day I have time on, we can get the rest of it in place
Aug 27 07:22:50 <ka6sox> kk.
Aug 27 07:23:16 <ka6sox> who knows...we may get recovery initramfs going.
Aug 27 07:23:32 <cryptk> so in theory... we /should/ be able to make the proper kernel and then memboot it to go to angstrom
Aug 27 07:23:34 <ka6sox> which would mean we can fix broken stuff.
Aug 27 07:23:38 <cryptk> and normal boot for webOS
Aug 27 07:23:54 <ka6sox> yes, let me see if I can get either console or net into initramfs.
Aug 27 07:24:00 <ka6sox> via USB
Aug 27 07:24:12 <ka6sox> depends on how they have it setup.
Aug 27 07:24:26 <ka6sox> net is more likely but not guaranteed.
Aug 27 07:24:31 <cryptk> since novacomd is already compiled for the right arch... there is a chance we can use the same binaries and get novaterm access... maybe...
Aug 27 07:25:02 <ka6sox> we could add that but I am trying for a totally GPL solution.
Aug 27 07:25:18 <ka6sox> meaning I can distribute it.
Aug 27 07:26:09 <cryptk> yes
Aug 27 07:26:33 <cryptk> I am thinking "down the road" possibly having a script for meta doc to build this
Aug 27 07:26:46 <cryptk> because meta-doc could copy novacomd out of the doctor and whatnot
Aug 27 07:27:03 <ka6sox> well...
Aug 27 07:28:03 <ka6sox> you would need the uboot tools installed on the machine doing the meta-doctor
Aug 27 07:28:09 <cryptk> yeah
Aug 27 07:28:17 <cryptk> well... the kernel would be distributable
Aug 27 07:28:25 <ka6sox> since the kernel is already packaged up.
Aug 27 07:28:42 <ka6sox> along with any initramfs
Aug 27 07:28:44 <cryptk> we could distribute an angstrom rootfs... which the meta-doc script could download, extract... unpack the doctor... yada yada yada
Aug 27 07:29:07 <cryptk> that would avoid the machine doing the meta-doc needing uboot tools
Aug 27 07:29:16 <ka6sox> that seems like a lotta work.
Aug 27 07:29:38 <cryptk> wouldn't be that hard, but that is WAY down the line
Aug 27 07:29:47 <ka6sox> yes, KISS pleaes.
Aug 27 07:29:58 <cryptk> yep
Aug 27 07:30:01 <ka6sox> lets get this working so I can memboot it first.
Aug 27 07:30:04 <cryptk> yep
Aug 27 07:30:12 <cryptk> that is step #1... get to where we can memboot it
Aug 27 07:30:25 <cryptk> step #2 is memboot it and have everything work
Aug 27 07:30:42 <ka6sox> I will *not* be using an angstrom kernel.
Aug 27 07:30:43 <cryptk> and by everything I mean full OS function... power button, vol buttons, touch screen
Aug 27 07:30:51 <cryptk> and maybe try and map the home button to something
Aug 27 07:30:51 <ka6sox> it knows NOTHING about the TP.
Aug 27 07:31:14 <cryptk> yeah, what I was thinking is use the Palm kernel, but we may have to compile it with some config flags that angstrom needs
Aug 27 07:31:22 <ka6sox> now you are talking about modding bootie
Aug 27 07:31:29 <cryptk> not sure what angstrom requires as far as config flags
Aug 27 07:31:34 <ka6sox> because all that takes place @ that level.
Aug 27 07:31:43 <cryptk> no I'm not
Aug 27 07:31:50 <cryptk> I'm talking about compiling a kernel and membooting it
Aug 27 07:32:07 <ka6sox> the buttons are all bootie (and maybe even before)
Aug 27 07:32:31 <cryptk> I would think that the button presses would be detectable by angstrom though
Aug 27 07:32:47 <cryptk> I would think that bootie passes them through somehow to the OS
Aug 27 07:33:00 <cryptk> otherwise how would Luna know to adjust the volume when you use the volume buttons
Aug 27 07:34:12 <ka6sox> bootie is LONG GONE when you reach the point where luna is even started.
Aug 27 07:34:18 <cryptk> exactly
Aug 27 07:34:28 <cryptk> and those buttons should be detectable through /dev/input
Aug 27 07:34:45 <cryptk> I am just talking about mapping those inputs to OS functions
Aug 27 07:34:50 <cryptk> as part of step #2
Aug 27 07:34:52 <ka6sox> during boot they are just addresses and registers to be read.
Aug 27 07:34:59 <cryptk> I know that ka6sox
Aug 27 07:35:03 <cryptk> I am talking about after that
Aug 27 07:35:04 <ka6sox> :D
Aug 27 07:35:12 <ka6sox> hmmm
Aug 27 07:35:14 <cryptk> step #1... get it membooted into angstrom
Aug 27 07:35:14 <ka6sox> welll..
Aug 27 07:35:25 <cryptk> step #2 get touchscreen and buttons working
Aug 27 07:35:27 <ka6sox> that is kind of an interesting idea
Aug 27 07:35:42 <ka6sox> iirc they are GPIOs
Aug 27 07:35:46 <ka6sox> soooo
Aug 27 07:35:55 <ka6sox> if you held down the right button at the right time.
Aug 27 07:36:05 <cryptk> that isn't what I am talking about
Aug 27 07:36:13 <cryptk> I am not talking about having the buttons work to select an OS
Aug 27 07:36:19 <ka6sox> you could make it decide which rootfs to run
Aug 27 07:36:21 <cryptk> I am talking about having them work inside angstrom
Aug 27 07:36:30 <cryptk> a.k.a. be able to adjust the volume with the volume buttons
Aug 27 07:36:44 <cryptk> AFTER angstrom has booted
Aug 27 07:36:47 <cryptk> that is step #2
Aug 27 07:37:03 <ka6sox> cryptk, I am trying to come up with a safe way to make sure I can make the thing boot *always*
Aug 27 07:37:32 <ka6sox> ie in the absence of any button pushes always boot the webOS partitons
Aug 27 07:37:55 <ka6sox> however...if you are pushing vol-dwn then pivot root to angstrom etc.
Aug 27 07:38:08 <cryptk> that would be step #3
Aug 27 07:38:11 <cryptk> true dual boot
Aug 27 07:38:34 <cryptk> step #1, get it to boot... step #2 make the OS fully functional... step #3 figure out a dual boot bootloader solution
Aug 27 07:38:44 <ka6sox> if I am going to do an initramfs I want it to be able to recover without any outside intervention.
Aug 27 07:39:29 <Jack87> fattire_, jacques did a number on one of his nooks :). But i dont think he or ka6sox had much time to play with it. So I will try to recover it soon
Aug 27 07:41:04 <ka6sox> ie not needing to have anything external at all to make it boot to webOS
Aug 27 07:41:57 <ka6sox> lesse...I need to find out if I can force a reboot somehow...
Aug 27 07:42:03 <ka6sox> since we can't battery pull
Aug 27 07:42:24 <Jack87> ka6sox, cryptk, dtzWill, when mapping keys we might be able to use two keys for a function such at Ctrl+shit for alt (arbitrary example)
Aug 27 07:42:44 <Jack87> ka6sox, hold home and power button for force reboot
Aug 27 07:42:52 <ka6sox> ctrl is not recognized on the external keyboard.
Aug 27 07:43:19 <Jack87> ka6sox, i know. (but that should change whether we change it or another webos update
Aug 27 07:43:23 <cryptk> ka6sox, tellbootie can force it
Aug 27 07:43:28 <cryptk> from within webos
Aug 27 07:43:42 <ka6sox> okay we *may* not get that far.
Aug 27 07:43:43 <cryptk> and from outside webos, you can likely use echo to the right spots in proc tto force a reboot
Aug 27 07:43:52 <cryptk> as in, from within angstrom
Aug 27 07:44:05 <ka6sox> okay there is a danger here.
Aug 27 07:44:08 <Jack87> cryptk, haha sometimes it might be frozen and not fully booted or what not... (Home+Power come to the rescue)
Aug 27 07:44:26 <ka6sox> we pivot_root to angstrom...and NOTHING HAPPENS.
Aug 27 07:44:40 <cryptk> you can force a reboot by holding home and power
Aug 27 07:44:43 <ka6sox> normally the proceedure @ that point is to power it off.
Aug 27 07:44:44 <cryptk> that go's straight to bootie
Aug 27 07:44:56 <cryptk> hold home+power for 10 seconds... it will reboot
Aug 27 07:45:12 <ka6sox> okay so that must hold off the WDT
Aug 27 07:45:17 <ka6sox> and make it time out
Aug 27 07:45:19 <cryptk> WDT?
Aug 27 07:45:25 <ka6sox> WatchDog timer
Aug 27 07:45:47 <cryptk> no, the watch dog is for if the device doesn't shut down in time
Aug 27 07:45:59 <cryptk> if you hold home+power for 10 seconds bootie just powercycles it
Aug 27 07:46:01 <ka6sox> huh?
Aug 27 07:46:16 <ka6sox> bootie is LONG gone.
Aug 27 07:46:24 <cryptk> no... no it is not...
Aug 27 07:46:55 <cryptk> unless they changed the way it works, there is a hardware killswitch basically that monitors those keys
Aug 27 07:47:01 <cryptk> that's what made the pwoer wiggle work
Aug 27 07:47:08 <cryptk> and what makes home+power work on the TP
Aug 27 07:47:18 <cryptk> it triggers a powercycle
Aug 27 07:47:33 <cryptk> btw, echo b > /proc/sysrq-trigger does work from within webos
Aug 27 07:47:44 <ka6sox> yes,
Aug 27 07:47:48 <cryptk> so it likely will from within angstrom as well as long as the kernel is compiled with sysrq enabled
Aug 27 07:47:49 <ka6sox> but thats not the same thing here.
Aug 27 07:47:54 <cryptk> I know it isn't
Aug 27 07:48:58 <ka6sox> I don't really care other than I just want to make sure it *always* work.s
Aug 27 07:49:21 <Jack87> cryptk, power wiggle == power + toggle mute switch?
Aug 27 07:49:24 <ka6sox> and what you are saying about bootie monitoring that is patently false.
Aug 27 07:49:39 <ka6sox> sorry..its HARDWARE that does this.
Aug 27 07:49:43 <ka6sox> has to be
Aug 27 07:50:32 <ka6sox> otherwise bootie could get into a funky mode and not reboot.
Aug 27 07:50:55 <ka6sox> aka...bootloader dev = bricking :D
Aug 27 07:51:22 <cryptk> Jack87, yes
Aug 27 07:51:27 <ka6sox> anyways..thats not the issue
Aug 27 07:51:40 <cryptk> ka6sox, yeah, I corrected myself
Aug 27 07:51:40 <ka6sox> if that *always* works then I am fine with it :D
Aug 27 07:51:45 <ka6sox> sorry.
Aug 27 07:51:50 <cryptk> <cryptk> unless they changed the way it works, there is a hardware killswitch basically that monitors those key
Aug 27 07:51:50 <Jack87> ka6sox, == bricking
Aug 27 07:51:59 <ka6sox> sorry!
Aug 27 07:52:02 <cryptk> np
Aug 27 07:52:08 <cryptk> I am running on very little sleep
Aug 27 07:52:08 <ka6sox> I should have read...my fault.
Aug 27 07:52:23 <cryptk> so all of that stuff I was saying do a s/bootie/hardware killswitch/g
Aug 27 07:52:26 <cryptk> lol
Aug 27 07:52:47 <ka6sox> okay
Aug 27 07:52:53 <cryptk> ok, and in /dev/input...
Aug 27 07:53:13 <cryptk> event0 = a GPIO key (likely home)
Aug 27 07:53:21 <cryptk> event1 = power button
Aug 27 07:53:32 <cryptk> event2 = probably volume rocker
Aug 27 07:53:40 <ka6sox> we can read those
Aug 27 07:53:44 <cryptk> yep
Aug 27 07:54:07 <ka6sox> for me the issue is to *always* be able to recover to webOS
Aug 27 07:55:01 <ka6sox> when I was taking pilot training....we learned that a properly trimmed airplane would correct itself if you let go of the controls.
Aug 27 07:55:19 <ka6sox> you could crank the controls till you made a mess
Aug 27 07:55:31 <ka6sox> let go...it got back to straight and level.
Aug 27 07:55:44 <ka6sox> this *has* to be like that.
Aug 27 07:56:27 <cryptk> yep, and it is
Aug 27 07:56:35 <cryptk> power+home=reboot...
Aug 27 07:56:39 <cryptk> it's hardware level
Aug 27 07:57:22 <ka6sox> cryptk, I am talking about not *requiring* all the fancy stuff like usb_console or usb_net
Aug 27 07:58:05 <ka6sox> cryptk, get fbcon running on fbdev for angstrom please
Aug 27 07:58:20 <cryptk> ok, now I am confused
Aug 27 07:58:29 <ka6sox> don't bet
Aug 27 07:58:41 <ka6sox> s/bet/be
Aug 27 07:58:50 <cryptk> first you are asking if the reboot is hardware level... then you are asking about fbcons
Aug 27 07:58:50 <ka6sox> its quite simple
Aug 27 07:58:56 <ka6sox> yep
Aug 27 07:59:05 <cryptk> those seem to be two different things to be
Aug 27 07:59:07 <cryptk> s/be/me/
Aug 27 07:59:15 <ka6sox> yes, I am working out my bit
Aug 27 07:59:17 <cryptk> ahh
Aug 27 07:59:31 <ka6sox> but...I am trying to keep it KISS
Aug 27 07:59:35 <cryptk> I can guarantee that I am going to be useless for this until I get some sleep though, lol
Aug 27 07:59:43 <ka6sox> understood
Aug 27 08:00:00 <ka6sox> I will just get it to boot the webOS partition first
Aug 27 08:00:04 <cryptk> git now I am just checking to see if I can get data from the /dev/input/event*'s (which I am sure I can... I just forget how to monitor them
Aug 27 08:00:27 <Jack87> cryptk, good to see you! and good night (if you hit the sack)
Aug 27 08:00:38 <ka6sox> in the absence of any external input the TP will *always* boot into webOS
Aug 27 08:00:50 <cryptk> yes
Aug 27 08:01:01 <cryptk> getting it to boot to angstrom by holding a button is WAY down the road
Aug 27 08:01:09 <cryptk> and not something we should be thinking about yet
Aug 27 08:01:10 <ka6sox> no, it isn't
Aug 27 08:01:21 <cryptk> lets get it to where we can memboot to it first
Aug 27 08:01:51 <ka6sox> that *will* work (to a certain extent)
Aug 27 08:02:06 <ka6sox> since the chroot works.
Aug 27 08:02:13 * Jack87 <3 memboot
Aug 27 08:02:38 <ka6sox> cryptk, I have to make a initramfs for it.
Aug 27 08:03:07 <ka6sox> in order to get it to boot the device
Aug 27 08:03:23 <ka6sox> and even pivot_root to the angstrom
Aug 27 08:03:48 <cryptk> yep, and that is what I want to get working first
Aug 27 08:03:53 <cryptk> it seems that is the simplest thing
Aug 27 08:03:58 <ka6sox> okay...
Aug 27 08:04:00 <cryptk> get a memboot verified working
Aug 27 08:04:05 <ka6sox> sure
Aug 27 08:04:17 <cryptk> once we have that, I can start working on the OS level stuff while you work on the kernel level stuff
Aug 27 08:04:36 <ka6sox> okay you work/sleep and I'll sort this stuff.
Aug 27 08:04:49 <cryptk> since once we can memboot to it, I will have a way to boot it (even though it isn't necessarily convenient)
Aug 27 08:04:57 <cryptk> but it will be good enough to get the OS config part going
Aug 27 08:05:20 <cryptk> I have 5 hours of work to go still...
Aug 27 08:05:43 <ka6sox> you realize that this will make the effort to get the *other* os running easier.
Aug 27 08:06:17 <cryptk> yep
Aug 27 08:06:25 <cryptk> I just want to beat them
Aug 27 08:07:27 <cryptk> the angstrom package manager is opkg right?
Aug 27 08:07:32 <ka6sox> ypu
Aug 27 08:07:35 <ka6sox> for latest
Aug 27 08:08:29 <ka6sox> you need to copy /dev/ into the chroot before we start this.
Aug 27 08:08:53 <cryptk> ok, event0 is the volume rocker
Aug 27 08:09:02 <ka6sox> because I am NOT going to be doing a bunch of mknods
Aug 27 08:09:24 <cryptk> and that is confirmed... event0 = vol rocker
Aug 27 08:09:38 <ka6sox> vol up or down?
Aug 27 08:10:10 <cryptk> event0 has 4 keys assigned to it amazingly enough...
Aug 27 08:10:28 <ka6sox> then thats the interrupt
Aug 27 08:10:47 <cryptk> KEY_VOLUMEDOWN, KEY_VOLUMEUP, KEY_REPLY and ?
Aug 27 08:10:49 <ka6sox> find the one that *only* responds to don
Aug 27 08:11:00 <ka6sox> dunno
Aug 27 08:11:09 <ka6sox> if not power then home
Aug 27 08:11:44 <cryptk> event0 is multiple gpio keys.... event1 is pmic8058_pwrkey and event2 is headset
Aug 27 08:12:08 <cryptk> KEY_REPLY looks to be the home button
Aug 27 08:12:50 <cryptk> tapping the power button hits KEY_END over event1
Aug 27 08:12:52 <ka6sox> Jack87, can you capture these into the wiki?
Aug 27 08:13:33 <cryptk> event0 they names are self-explanitory except KEY_REPLY is the home key
Aug 27 08:13:40 <ka6sox> okay we look for a KEY_VOLUMEDOWN event
Aug 27 08:14:02 * Jack87 is only passively monitoring while reading. will have to go over it. Only problem is a lot is over his head
Aug 27 08:14:20 <ka6sox> just do a dump, let me edit it.
Aug 27 08:14:30 <Jack87> oh i see
Aug 27 08:14:39 <Jack87> you just want to know what the mappings are in w iki for now
Aug 27 08:14:45 <cryptk> tapping power sends KEY_END over event1, holding it sends KEY_POWER
Aug 27 08:14:45 <ka6sox> yes
Aug 27 08:14:54 <Jack87> ya sure can do
Aug 27 08:15:08 <cryptk> root@CryptksTouchPad:/# evtest /dev/input/event0
Aug 27 08:15:08 <cryptk> Input driver version is 1.0.0
Aug 27 08:15:08 <cryptk> Input device ID: bus 0x19 vendor 0x1 product 0x1 version 0x100
Aug 27 08:15:08 <cryptk> Input device name: "gpio-keys"
Aug 27 08:15:08 <cryptk> Supported events:
Aug 27 08:15:09 <cryptk>   Event type 0 (Sync)
Aug 27 08:15:11 <cryptk>   Event type 1 (Key)
Aug 27 08:15:13 <cryptk>     Event code 114 (VolumeDown)
Aug 27 08:15:15 <cryptk>     Event code 115 (VolumeUp)
Aug 27 08:15:17 <cryptk>     Event code 232 (Reply)
Aug 27 08:15:21 <cryptk>     Event code 249 (?)
Aug 27 08:15:23 <cryptk>   Event type 5 (?)
Aug 27 08:15:25 <cryptk> root@CryptksTouchPad:/# evtest /dev/input/event1
Aug 27 08:15:27 <cryptk> Input driver version is 1.0.0
Aug 27 08:15:29 <cryptk> Input device ID: bus 0x0 vendor 0x0 product 0x0 version 0x0
Aug 27 08:15:31 <cryptk> Input device name: "pmic8058_pwrkey"
Aug 27 08:15:33 <cryptk> Supported events:
Aug 27 08:15:35 <cryptk>   Event type 0 (Sync)
Aug 27 08:15:37 <cryptk>   Event type 1 (Key)
Aug 27 08:15:39 <cryptk>     Event code 107 (End)
Aug 27 08:15:41 <cryptk>     Event code 116 (Power)
Aug 27 08:15:43 <cryptk> root@CryptksTouchPad:/# evtest /dev/input/event2
Aug 27 08:15:45 <cryptk> Input driver version is 1.0.0
Aug 27 08:15:47 <cryptk> Input device ID: bus 0x0 vendor 0x0 product 0x0 version 0x0
Aug 27 08:15:51 <cryptk> Input device name: "headset"
Aug 27 08:15:53 <cryptk> Supported events:
Aug 27 08:15:55 <cryptk>   Event type 0 (Sync)
Aug 27 08:15:57 <cryptk>   Event type 1 (Key)
Aug 27 08:15:59 <cryptk>     Event code 164 (PlayPause)
Aug 27 08:16:01 <cryptk>   Event type 5 (?)
Aug 27 08:16:03 <cryptk>     Event code 2 (?)
Aug 27 08:16:05 <cryptk>     Event code 4 (?)
Aug 27 08:16:07 <cryptk> there you go
Aug 27 08:16:15 <cryptk> drop this in the wiki somewhere
Aug 27 08:16:21 <ka6sox> Jack87, ^^
Aug 27 08:17:00 <ka6sox> this *may* not be available to me @ the time we have the boot happening
Aug 27 08:17:16 <ka6sox> I may need to read the /sysfs directly
Aug 27 08:17:47 <ka6sox> the event1 codes are i2c codes from the PMIC
Aug 27 08:18:00 <Jack87> ka6sox, working on making a page
Aug 27 08:18:24 <ka6sox> and event0 is on the CPU itself.
Aug 27 08:19:04 <ka6sox> event2 could be from the Audio Processor
Aug 27 08:19:10 <ka6sox> probably is.
Aug 27 08:19:31 <cryptk> yes it is
Aug 27 08:19:33 <cryptk> root@CryptksTouchPad:/# udevadm info --query=all --name=/dev/input/event0
Aug 27 08:19:33 <cryptk> P: /devices/platform/gpio-keys/input/input0/event0
Aug 27 08:19:33 <cryptk> E: UDEV_LOG=3
Aug 27 08:19:33 <cryptk> E: DEVPATH=/devices/platform/gpio-keys/input/input0/event0
Aug 27 08:19:33 <cryptk> E: MAJOR=13
Aug 27 08:19:34 <cryptk> E: MINOR=64
Aug 27 08:19:36 <cryptk> E: DEVNAME=input/event0
Aug 27 08:19:38 <cryptk> E: SUBSYSTEM=input
Aug 27 08:19:40 <cryptk> root@CryptksTouchPad:/# udevadm info --query=all --name=/dev/input/event1
Aug 27 08:19:42 <cryptk> P: /devices/i2c-6/6-0055/pm8058-pwrkey/input/input1/event1
Aug 27 08:19:44 <cryptk> E: UDEV_LOG=3
Aug 27 08:19:46 <cryptk> E: DEVPATH=/devices/i2c-6/6-0055/pm8058-pwrkey/input/input1/event1
Aug 27 08:19:48 <cryptk> E: MAJOR=13
Aug 27 08:19:52 <cryptk> E: MINOR=65
Aug 27 08:19:54 <cryptk> E: DEVNAME=input/event1
Aug 27 08:19:56 <cryptk> E: SUBSYSTEM=input
Aug 27 08:19:58 <cryptk> root@CryptksTouchPad:/# udevadm info --query=all --name=/dev/input/event2
Aug 27 08:20:00 <cryptk> P: /devices/platform/soc-audio.0/sound/card0/input2/event2
Aug 27 08:20:02 <cryptk> E: UDEV_LOG=3
Aug 27 08:20:04 <cryptk> E: DEVPATH=/devices/platform/soc-audio.0/sound/card0/input2/event2
Aug 27 08:20:06 <cryptk> E: MAJOR=13
Aug 27 08:20:08 <cryptk> E: MINOR=66
Aug 27 08:20:10 <cryptk> E: DEVNAME=input/event2
Aug 27 08:20:12 <cryptk> E: SUBSYSTEM=input
Aug 27 08:20:18 <cryptk> there's your sysfs locations
Aug 27 08:20:30 <ka6sox> ypu
Aug 27 08:20:32 <ka6sox> yup
Aug 27 08:20:38 <ka6sox> Jack87, capture those too :D
Aug 27 08:21:07 <cryptk> there is no udevadm or evtest in webOS... so all of this is coming from within the angstrom chroot
Aug 27 08:21:09 <ka6sox> major and minor are nods
Aug 27 08:21:18 <cryptk> but it is the same /dev /sys and /prob bind mounted in
Aug 27 08:21:21 <cryptk> so it's still valid
Aug 27 08:21:25 <ka6sox> right
Aug 27 08:21:30 <ka6sox> I am expecting that
Aug 27 08:21:34 <cryptk> s/prob/proc/
Aug 27 08:21:42 <ka6sox> I will have to do the mounts
Aug 27 08:21:43 <Jack87> yeppie
Aug 27 08:21:50 <ka6sox> but thats fine.
Aug 27 08:22:18 <cryptk> Jack87, make sure to include that for event1 a power button tap = KEY_END while a power button hold (for 1/2 second) = KEY_POWER
Aug 27 08:22:18 <ka6sox> cryptk, does the angstrom chroot have a initramfs?
Aug 27 08:22:36 <cryptk> there is a uImage file in it
Aug 27 08:22:40 <Jack87> cryptk, copy that
Aug 27 08:22:49 <ka6sox> cryptk, can you send that to me?
Aug 27 08:22:49 <cryptk> it is likely a kernel and initramfs combined
Aug 27 08:22:57 <ka6sox> ya, let me explode that.
Aug 27 08:23:15 <cryptk> Jack87, for event0 home button = KEY_REPLY
Aug 27 08:23:35 <cryptk> let me grab it for ya and email it off
Aug 27 08:23:40 <ka6sox> thanks.
Aug 27 08:23:55 <ka6sox> I just want to see what it expects
Aug 27 08:24:37 <ka6sox> so that I can make it happy.
Aug 27 08:25:11 * ka6sox wonders if early_fbcon is enabled in our kernels.
Aug 27 08:25:15 <cryptk> ka6sox, http://www.cryptkcoding.com/uImage-angstrom
Aug 27 08:25:21 <ka6sox> kk..thanks
Aug 27 08:25:48 <ka6sox> cryptk, got it
Aug 27 08:26:09 <cryptk> ka6sox, CONFIG_FRAMEBUFFER_CONSOLE=y
Aug 27 08:26:13 <cryptk> that's in F15 at least
Aug 27 08:26:35 <ka6sox> ya, there is another setting for EARLY_FBCON I think.
Aug 27 08:26:43 <ka6sox> we found it in the nook kernel.
Aug 27 08:27:23 <cryptk> not present in palm kernel config
Aug 27 08:27:29 <ka6sox> kk
Aug 27 08:27:34 <cryptk> at least not in the F15 config
Aug 27 08:27:47 <ka6sox> the defconfig probably doesnt' have it
Aug 27 08:27:57 <ka6sox> so its just *missing*
Aug 27 08:28:07 <cryptk> yep
Aug 27 08:28:10 <cryptk> and bootie has...
Aug 27 08:28:11 <cryptk> bootconsole = ttyS0,115200n8
Aug 27 08:28:19 <ka6sox> ya....RRRIGHT
Aug 27 08:28:28 <cryptk> lol
Aug 27 08:28:32 <ka6sox> (actually the pins are probably on the board.
Aug 27 08:28:34 <ka6sox> )
Aug 27 08:28:50 <ka6sox> and *maybe* even pads for a 232 chop
Aug 27 08:28:52 <ka6sox> chip
Aug 27 08:29:05 <cryptk> I doubt a 232 chip is actually there though
Aug 27 08:29:22 <ka6sox> doesnt' mean I can't figure it out when I salt.
Aug 27 08:29:25 <cryptk> yep
Aug 27 08:29:37 <cryptk> if you are careful when you salt, you can try soldering one on
Aug 27 08:29:55 <ka6sox> eventually I will need 2 hardwar dev things.
Aug 27 08:30:08 <ka6sox> 1 to salt and one to poke and prod while alive.
Aug 27 08:30:36 <ka6sox> okay this has been quite useful
Aug 27 08:30:44 <cryptk> partially salt... poke+prod... finish salt
Aug 27 08:30:56 <ka6sox> no
Aug 27 08:31:02 <ka6sox> you need a bare board to trace
Aug 27 08:31:13 <ka6sox> and then one to poke/prod
Aug 27 08:31:48 <ka6sox> Jack87, let me know where that page is please.
Aug 27 08:31:49 <Jack87> here is what i got so far.. addding second set in a sec http://www.webos-internals.org/wiki/HP_BT_Key_mapping
Aug 27 08:31:53 <ka6sox> whee!
Aug 27 08:32:36 <ka6sox> except this is the keys on the device..not the keyboard :D
Aug 27 08:32:49 <ka6sox> but that we can easily fix.
Aug 27 08:33:12 <cryptk> yeah, this isn't bluetooth stuff
Aug 27 08:33:15 <ka6sox> we should rename it TP_Buttons
Aug 27 08:33:20 <cryptk> this is the actual hardware buttons on the TP
Aug 27 08:33:43 <ka6sox> but the rest is awesome.
Aug 27 08:33:46 <cryptk> this is the info that I need to get the buttons working in angstrom
Aug 27 08:33:55 <cryptk> indeed
Aug 27 08:34:10 <ka6sox> and I will need to make it so I can switch which OS to boot.
Aug 27 08:34:22 <cryptk> that is gonna be a bootie hack of some sort
Aug 27 08:34:26 <Jack87> ka6sox, lol ya i figured that out after i was editing
Aug 27 08:34:29 <Jack87> will remake page
Aug 27 08:34:30 <ka6sox> cryptk, nope
Aug 27 08:34:35 <ka6sox> no bootie hack
Aug 27 08:34:40 <Jack87> cab we rename?
Aug 27 08:34:45 <cryptk> just a pivot_root hack?
Aug 27 08:34:51 <ka6sox> Move
Aug 27 08:34:51 <cryptk> yeah, we can rename it
Aug 27 08:34:55 <cryptk> you just move the page
Aug 27 08:35:03 <Jack87> just take out _BT
Aug 27 08:35:12 <ka6sox> TP_Buttons
Aug 27 08:35:27 <cryptk> done
Aug 27 08:35:27 <cryptk> http://www.webos-internals.org/wiki/HP_TP_Key_mapping
Aug 27 08:35:43 <Jack87> ka6sox, i like mapping being in there incase someone is using that as a keyboard
Aug 27 08:35:49 <Jack87> keyword*
Aug 27 08:35:54 <ka6sox> cryptk, it should be "Buttons" instead of KEY
Aug 27 08:36:18 <ka6sox> because for sure someone will think its they keyboard and say we forgot a bunch of STUFF>
Aug 27 08:36:26 <cryptk> they are reverenced as events and as keypads in /dev/input
Aug 27 08:36:37 <ka6sox> I know
Aug 27 08:36:39 <cryptk> and those people can piss off ;)
Aug 27 08:36:47 <cryptk> if they can't read... that's their fault
Aug 27 08:37:48 <ka6sox> cryptk, the plan is to create a initramfs that knows how to boot webOS *or* alternative OS.
Aug 27 08:38:03 <cryptk> ahh, nice
Aug 27 08:38:04 <Jack87> captcha is pissing me off
Aug 27 08:38:18 <ka6sox> cryptk, can you make him an admin?
Aug 27 08:38:23 <cryptk> so the initramfs checks the keys and passes along to the appropriate rootfs
Aug 27 08:38:41 <ka6sox> sets things up the way the rootfs needs them to be
Aug 27 08:38:46 <ka6sox> like mounts the right stuff.
Aug 27 08:38:52 <ka6sox> to get it booted. up
Aug 27 08:38:56 <cryptk> Jack87, where are you getting a captcha?
Aug 27 08:39:17 <Jack87> when editing page on wiki everytime saving
Aug 27 08:39:20 <cryptk> ahh
Aug 27 08:39:24 <cryptk> that should be fixed now
Aug 27 08:39:46 <Jack87> cryptk, thank you sir
Aug 27 08:39:52 <ka6sox> no button push...webOS
Aug 27 08:40:00 <ka6sox> button push...alternative OS
Aug 27 08:40:24 <cryptk> gotcha
Aug 27 08:40:39 <cryptk> would this initramfs pass along to the webos kernel and initramfs?
Aug 27 08:40:48 <cryptk> or would this be a complete replacement?
Aug 27 08:41:27 <ka6sox> this would be a complete replacement
Aug 27 08:41:48 <ka6sox> because they don't separate out initramfs from kernel in the uImage.
Aug 27 08:43:08 <ka6sox> let me get this booting first.
Aug 27 08:43:18 <ka6sox> ie booting webOS
Aug 27 08:43:28 <ka6sox> then I can get it to boot angstrom
Aug 27 08:43:39 <cryptk> kk
Aug 27 08:44:08 <Jack87> http://www.webos-internals.org/wiki/HP_TP_Key_mapping#sysfs_locations
Aug 27 08:44:28 <ka6sox> very nice..thanks
Aug 27 08:44:34 <cryptk> yep
Aug 27 08:44:43 <ka6sox> now it won't be lost in IRC HELL
Aug 27 08:44:45 <cryptk> all of the raw data is there... which is the important part
Aug 27 08:44:51 <cryptk> we can clean it up more later
Aug 27 08:45:14 <cryptk> that is the stuff that ka6sox will need for his dual boot solution... and what I will need to get those keys working in angstrom
Aug 27 08:46:35 <ka6sox> I will need to create a generic config to find the correct name for early_FBCON
Aug 27 08:46:45 <ka6sox> but we don't need that if I do this right.
Aug 27 08:48:35 <Jack87> i tried to make it devided as much as possible but as soon as we have everything sorted we can clean it and make it all pretty like
Aug 27 08:48:54 <ka6sox> yep...its great..I can find everything.
Aug 27 08:49:03 <Jack87> cryptk, is there an event when TP is placed on TS?
Aug 27 08:49:23 <cryptk> not that I have seen yet
Aug 27 08:49:31 <cryptk> that wouldn't be an input though
Aug 27 08:49:34 <ka6sox> that is probably in the "telecom" code
Aug 27 08:49:49 * Jack87 would like to know what triggers exhibition 
Aug 27 08:50:02 <ka6sox> which is probably monitoring the sysfs to find that event.
Aug 27 08:50:22 <ka6sox> via an i2c status message from the charger.
Aug 27 08:51:11 <ka6sox> the telecom code is a STUB almost in this box.
Aug 27 08:52:21 * Jack87 wonders if this will work http://www.buy.com/prod/bluetooth-keyboard-ideal-for-ipads-iphones-with-bluetooth-playstation/q/sellerid/9172234/loc/111/218200014.html
Aug 27 08:52:59 <ka6sox> Jack87, once I ID the actual charger chip (which I could do by looking @ sysfs entries) I can get the DS for it and we can probe it.
Aug 27 08:54:10 <Jack87> by the way like the ctrl key arrow keys are useless on BT keyboard as well
Aug 27 08:54:22 <ka6sox> Jack87, yup
Aug 27 08:54:35 <ka6sox> tested that one about 5days ago.
Aug 27 08:54:45 <ka6sox> its a bummer
Aug 27 08:54:53 <ka6sox> but we don't know where they are being eaten.
Aug 27 08:55:10 * Jack87 also thinks there should be a bring to forth key on the damn thing (perhaps down key while in card view)
Aug 27 08:55:21 <Jack87> close key would be nice to (up key?)
Aug 27 08:55:47 <Jack87> and shuffle via arrows
Aug 27 08:55:55 <cryptk> max8903b_chg.0
Aug 27 08:56:00 <ka6sox> that would be a hack.
Aug 27 08:56:08 <ka6sox> a SERIOUS HACK.
Aug 27 08:56:16 <ka6sox> cryptk, thank.s
Aug 27 08:56:24 <ka6sox> not gonna look yet
Aug 27 08:56:26 <cryptk> yep
Aug 27 08:56:36 <Jack87> a seriously useful hack
Aug 27 08:56:57 <ka6sox> okay back to hacking on the kernel stuff.
Aug 27 08:57:19 <ka6sox> if it acts like I think I could make this work with *any* kernel if we packaged it right.
Aug 27 08:57:47 <cryptk> yep
Aug 27 09:00:20 <ka6sox> because bootie changes the bootargs to use the initramfs
Aug 27 09:01:50 <ka6sox> with this hack I could probably have android booting before they do!
Aug 27 09:03:15 <ka6sox> not counting chickens yet though
Aug 27 09:19:02 * Jack87 is thinking about watching Tekken movie before bed. it looks really cheesy 
Aug 27 09:19:41 <cryptk> doit
Aug 27 09:26:37 <rrix> ka6sox: the irony in doing that would be so beautiful, lol
Aug 27 09:27:05 <rrix> I'm not sure if that would be more or less awesome than rwhitby buying one of those android testing models to flash it with the doctor, though :P
Aug 27 09:27:30 <rrix> wheee, my Qt lighthouse plugin is compiling properly; now we wonder if it actually works when it's done ...
Aug 27 09:27:54 <cryptk> well... we can always go the XDA route and just show a screenshot of android on the touchpad and say we got it working...
Aug 27 09:28:02 <cryptk> lol
Aug 27 09:28:08 <cryptk> but angstrom first
Aug 27 09:28:23 <cryptk> oooh.... I wanna get windows 3.11 working on the touchpad
Aug 27 09:28:29 <cryptk> because it's for WORKGROUPS!
Aug 27 09:29:32 <rrix> cryptk: wait for linux kernel 3.11 and it might be easier? :P
Aug 27 09:29:51 <cryptk> haha
Aug 27 09:32:41 * rrix is not looking forward to that cheasy slashdot post... lol
Aug 27 10:11:48 <ka6sox> cryptk, I need /boot/sbin/init or /sbin/init from your Angstrom chroot please
Aug 27 10:12:20 <ka6sox> and if you have an initramfs or initrd I could use that too please.
Aug 27 10:15:44 <cryptk> the only thing I have as far as initramfs or initrd is whatever is wrapped inside that uImage
Aug 27 10:15:51 <cryptk> I can get you the init file though
Aug 27 10:16:05 <ka6sox> thats good
Aug 27 10:16:22 <cryptk> /sbin/init is a symlink to /sbin/init.sysvinit
Aug 27 10:16:34 <ka6sox> that will tell me whether a real initrd is necessary.
Aug 27 10:16:42 <ka6sox> okay I'll take that
Aug 27 10:17:54 <cryptk> http://www.cryptkcoding.com/init.sysvinit
Aug 27 10:19:25 <cryptk> is that what you need ka6sox ?
Aug 27 10:19:31 <ka6sox> okay I need init.rc please.
Aug 27 10:19:35 <ka6sox> /etc/init.rc
Aug 27 10:19:55 <cryptk> there is no /etc/init.rc
Aug 27 10:20:18 <ka6sox> okay /etc/init.d/*.rc?
Aug 27 10:20:42 <cryptk> there is an init script for rc and for rcS
Aug 27 10:20:55 <ka6sox> kk...jas
Aug 27 10:21:50 <cryptk> http://www.cryptkcoding.com/rc
Aug 27 10:21:52 <cryptk> http://www.cryptkcoding.com/rcS
Aug 27 10:22:38 <ka6sox> is there rc.local?
Aug 27 10:22:57 <cryptk> nope
Aug 27 10:23:15 <cryptk> there is a /etc/default/rcS
Aug 27 10:23:35 <cryptk> and /etc/inputrc
Aug 27 10:24:14 <cryptk> do you want me to just put the whole rootfs on CC somewhere?
Aug 27 10:24:21 <cryptk> then you can grab whatever you need?
Aug 27 10:24:44 <ka6sox> ya, I could use that.
Aug 27 10:24:59 <ka6sox> do you feel like you could do the following?
Aug 27 10:25:22 <ka6sox> resize ext3fs to allow creation of another partition?
Aug 27 10:25:31 <ka6sox> create that partition
Aug 27 10:25:47 <ka6sox> move (or copy) the angstrom chroot there
Aug 27 10:25:56 <ka6sox> and (for now) mount it?
Aug 27 10:26:12 <cryptk> and what would that do for me?
Aug 27 10:26:24 <ka6sox> simplify my life
Aug 27 10:26:30 <cryptk> I already have angstrom sitting directly in /media/ext3fs/
Aug 27 10:26:49 <cryptk> so it is /media/ext3fs/{bin,sbin,proc,dev,sys....}
Aug 27 10:27:00 <cryptk> it isn't in a subfolder inside of it
Aug 27 10:27:25 <ka6sox> I don't think our kernels can handle ext3 FS for initial boots
Aug 27 10:27:32 <ka6sox> only ext2
Aug 27 10:27:54 <ka6sox> iirc we tried that earlier and found it didn't work.
Aug 27 10:27:56 <cryptk> this is the only thing I am using the ext3fs LVM for... I can just backup the rootfs and reformat it as ext2
Aug 27 10:28:26 <ka6sox> fine
Aug 27 10:28:38 <ka6sox> well.
Aug 27 10:28:45 <cryptk> or I can resize etc
Aug 27 10:29:00 <ka6sox> either way
Aug 27 10:29:00 <cryptk> make another one named ext2fs
Aug 27 10:29:07 <ka6sox> they both accomplish the same thing
Aug 27 10:29:23 <ka6sox> but leaving your ext3fs alone would be my pref (other than resizing)
Aug 27 10:29:43 <cryptk> kk
Aug 27 10:29:52 <cryptk> I will resize and make a second one named ext2fs
Aug 27 10:30:03 <ka6sox> then do a rsync
Aug 27 10:30:20 <ka6sox> because I do want to see if later we can get the kernel to boot a ext3
Aug 27 10:31:12 <ka6sox> usually when they boot to ext2 its because ext3 isn't gonna cut it (like why it can't boot to LVM)
Aug 27 10:31:17 <cryptk> ok, look in your home folder on CC
Aug 27 10:31:22 <ka6sox> kk
Aug 27 10:31:25 <cryptk> there is a folder named angstrom-rootfs
Aug 27 10:31:29 <cryptk> has the whole thing
Aug 27 10:31:44 <ka6sox> okay let me rsync that.
Aug 27 10:31:50 <ka6sox> oh.
Aug 27 10:31:53 <cryptk> gonna reboot the TP before doing the LV resize to clean up the mounts
Aug 27 10:32:06 <cryptk> it is extracted in there... and the tarball is still there as well
Aug 27 10:32:07 <ka6sox> can you tgz that and I'll just snag it?
Aug 27 10:32:11 <ka6sox> okay
Aug 27 10:32:14 <cryptk> so you can just grab the tarball if you want it locally
Aug 27 10:32:20 <ka6sox> lemme look
Aug 27 10:34:17 <ka6sox> the leap from Angstrom to Ubuntu is a short hop
Aug 27 10:34:24 <ka6sox> (from a boot perspective)
Aug 27 10:35:15 <cryptk> ok, 8GB ext3fs resized to 4GB
Aug 27 10:35:23 <cryptk> now to make a 4GB ext2fs one
Aug 27 10:35:38 <ka6sox> isn't that kind of MAX on ext2?
Aug 27 10:36:09 <ka6sox> or was that 2GB..can't remember.
Aug 27 10:36:56 <cryptk> you are thinking TB
Aug 27 10:37:11 <ka6sox> might be...its 3:33am
Aug 27 10:37:15 <cryptk> heh
Aug 27 10:37:43 <ka6sox> webOS uses /sbin/init which is a script
Aug 27 10:38:42 <ka6sox> Angstrom uses /sbin/init.sysvinit which is a binary
Aug 27 10:39:47 <ka6sox> Android uses something init.rc which is a script
Aug 27 10:40:31 <ka6sox> Ubuntu uses /sbin/sysvinit which is a binary
Aug 27 10:42:28 <cryptk> ka6sox, the rootfs for webOS on the TP is ext3
Aug 27 10:42:39 <cryptk> rootfs  /       ext3    ro,suid,dev,exec,auto,nouser,async,noatime      1       1
Aug 27 10:42:43 <ka6sox> yes, the realroot is
Aug 27 10:43:08 <ka6sox> the root which is in /dev/mmcblk0p13 is ext2
Aug 27 10:43:31 <cryptk> ahh
Aug 27 10:44:04 <ka6sox> hence why I wanted you, for this exercise to make angstrom live in a ext2
Aug 27 10:44:17 <cryptk> gotcha
Aug 27 10:44:26 <cryptk> well, I have it set up in an ext3 and an ext2
Aug 27 10:44:45 <cryptk> one is /dev/store/ext2fs the other is /dev/store/ext3fs
Aug 27 10:44:53 <ka6sox> you need to mod /etc/fstab in the ext2 to mount it right
Aug 27 10:45:03 <cryptk> mounted at /media/ext2fs and /media/ext3fs
Aug 27 10:45:06 <cryptk> yep
Aug 27 10:45:08 <cryptk> already done
Aug 27 10:45:32 <cryptk> /dev/mapper/store-ext2fs	/media/ext2fs	ext2	noatime	0	0
Aug 27 10:45:35 <ka6sox> also...you need lvm as well.
Aug 27 10:45:42 <ka6sox> lvm2
Aug 27 10:46:19 <ka6sox> if your image is missing that please recreate and add that in.
Aug 27 10:46:19 <cryptk> need it where?
Aug 27 10:46:28 <cryptk> oh, the angstrom image
Aug 27 10:46:30 <ka6sox> in the angstrom rootfs.
Aug 27 10:46:52 <ka6sox> the chroot doesnt' care
Aug 27 10:46:58 <ka6sox> but the real one will.
Aug 27 10:47:05 <cryptk> yeah... looks like I don't have it... lemme rebuild
Aug 27 10:47:13 <ka6sox> kk
Aug 27 10:47:45 <cryptk> well, since we are rebuilding...
Aug 27 10:47:50 <cryptk> does udev and sysvinit work for you?
Aug 27 10:48:12 <ka6sox> both do.
Aug 27 10:48:28 <ka6sox> and you need udev becuase of the kernel.
Aug 27 10:49:05 <cryptk> there is no option for lvm2 on narcissus
Aug 27 10:49:15 <ka6sox> ya there is...
Aug 27 10:49:17 <cryptk> can I just chroot into it and opkg install lvm2?
Aug 27 10:49:24 <ka6sox> I think its just lvm
Aug 27 10:49:52 <ka6sox> or lvm-tools can't remember which
Aug 27 10:49:59 <ka6sox> but for now...you can just opkg install
Aug 27 10:50:05 <cryptk> yeah, I will opkg install it
Aug 27 10:50:11 <cryptk> because the option definitely isn't ther eon narc
Aug 27 10:50:13 <ka6sox> do it to both.
Aug 27 10:50:39 <cryptk> will do
Aug 27 10:51:02 <cryptk> the package is named lvm2
Aug 27 10:51:10 <ka6sox> ys
Aug 27 10:51:12 <ka6sox> yes
Aug 27 10:51:26 <ka6sox> thats the debian/ubuntu name
Aug 27 10:51:42 <cryptk> ok, it's installed on both
Aug 27 10:51:46 <ka6sox> kk
Aug 27 10:51:53 <cryptk> do you want a new copy of the rootfs?
Aug 27 10:51:59 <ka6sox> yes please
Aug 27 10:52:08 <ka6sox> I will need to create a config file for it as well.
Aug 27 10:52:34 <ka6sox> I just need the name of the tgz please.
Aug 27 10:52:49 <cryptk> it will be the same name
Aug 27 10:53:01 <cryptk> it is the same as the last one, just has an opkg install lvm2 done on it
Aug 27 10:53:01 <ka6sox> kk
Aug 27 10:53:08 <cryptk> I will let you know once it is there
Aug 27 10:53:13 <ka6sox> okay
Aug 27 10:53:47 <ka6sox> I can't remember if Android can deal with LVM or not
Aug 27 10:54:04 <ka6sox> doesn't matter
Aug 27 10:55:02 <cryptk> ok it's there
Aug 27 10:55:07 <cryptk> MD5 is 6de48976bed6b59451c640313a6b910b
Aug 27 10:55:16 <ka6sox> thanks.
Aug 27 10:55:19 <ka6sox> lemme snag
Aug 27 10:55:58 <ka6sox> I hope this isn't armv5
Aug 27 10:57:20 <cryptk> root@CryptksTouchPad:/sbin# file init.sysvinit
Aug 27 10:57:20 <cryptk> init.sysvinit: ELF 32-bit LSB executable, ARM, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.16, stripped
Aug 27 10:57:48 <ka6sox> ARM not ARMEL?
Aug 27 10:57:49 <cryptk> Angstrom 2011.03
Aug 27 10:57:49 <cryptk> Built from branch: 2011.03-maintenance
Aug 27 10:57:51 <cryptk> Revision: 61a4492
Aug 27 10:57:53 <cryptk> Target system: arm-angstrom-linux-gnueabi
Aug 27 10:58:37 <ka6sox> okay I *hope* this isn't OABI
Aug 27 11:00:26 <Jack87> http://www.ebay.com/itm/HP-TouchPad-32GB-running-Android-/230664585087
Aug 27 11:00:36 <Jack87> really 1.3k?
Aug 27 11:00:47 <cryptk> root@CryptksTouchPad:~# file initctl
Aug 27 11:00:47 <cryptk> initctl: ELF 32-bit LSB executable, ARM, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.14, not stripped
Aug 27 11:00:53 <cryptk> that is the initctl from webos itself
Aug 27 11:01:16 <cryptk> seems to match the output from file init.sysvinit from the angstrom build
Aug 27 11:01:24 <cryptk> IIRC this is a beagleboard compile from narc
Aug 27 11:01:52 <ka6sox> not stripped...so gdb and strace would give us some good info
Aug 27 11:02:04 <ka6sox> no matter....
Aug 27 11:02:08 <ka6sox> we don't care
Aug 27 11:02:09 <cryptk> the not stripped file is the initctl from webOS
Aug 27 11:02:13 <ka6sox> ya
Aug 27 11:02:37 <cryptk> the angstrom ones are stripped
Aug 27 11:02:45 <ka6sox> saw that
Aug 27 11:02:59 <ka6sox> find powerd and check that one please?
Aug 27 11:04:15 <cryptk> root@CryptksTouchPad:~# file powerd
Aug 27 11:04:15 <cryptk> powerd: ELF 32-bit LSB executable, ARM, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.14, not stripped
Aug 27 11:04:18 <ka6sox> (webOS partition)
Aug 27 11:04:25 <ka6sox> w00 w00
Aug 27 11:04:32 <cryptk> there is no file command in webOS, so I have to copy it into the chroot... then chroot in... then hit it with file
Aug 27 11:04:56 <ka6sox> is it not in optware?
Aug 27 11:05:04 <cryptk> I don't have optware atm
Aug 27 11:05:09 <ka6sox> kk
Aug 27 11:05:14 <ka6sox> I don't either
Aug 27 11:05:22 <ka6sox> but I should.
Aug 27 11:05:22 <cryptk> I got rid of it in order to "re-purpose" the partition for this
Aug 27 11:05:44 <cryptk> and /media/ext3fs got a brand new filesystem when I resized it
Aug 27 11:05:54 <cryptk> I did it the lazy way... just resized the LVM and then re-did mkfs
Aug 27 11:09:41 <cryptk> ka6sox, anything else you need?
Aug 27 11:11:15 <ka6sox> I have the rootfs and am cruising thru it.
Aug 27 11:11:25 <ka6sox> I think I'm good for now.
Aug 27 11:12:06 <cryptk> sweet
Aug 27 11:12:25 <cryptk> once you have something ready for me to test with, just drop it in /home/cryptk on CC and ping me
Aug 27 11:12:45 <ka6sox> kk...look when you wake up later tonight
Aug 27 11:13:03 <cryptk> kk
Aug 27 11:13:08 <ka6sox> since I have your rootfs I will test here.
Aug 27 11:13:12 <ka6sox> then give it to you.
Aug 27 11:13:16 <cryptk> cool
Aug 27 11:14:19 <ka6sox> I am going to assume when you got this you didn't care about an initrd or initramfs right?
Aug 27 11:14:24 <cryptk> once we get it to where I can memboot and have angstrom come up, then I can work on getting SSH set up in it... then the OS config can start
Aug 27 11:14:32 <cryptk> nope, I sure didn't
Aug 27 11:15:07 <cryptk> IIRC it is beagleboard, task-base, udev, sysvinit with X11 and Enlightenment
Aug 27 11:15:08 <ka6sox> why don't you start a ssh daemon running against the usb net connection.
Aug 27 11:15:21 <ka6sox> since I know beagle supports that.
Aug 27 11:15:30 <cryptk> ka6sox, that is what I will do, once I can see that it does something
Aug 27 11:15:43 <ka6sox> it already does.
Aug 27 11:15:55 <cryptk> what do you mean it already does?
Aug 27 11:15:56 <ka6sox> you need to test that in the chroot first.
Aug 27 11:16:00 <cryptk> ahh
Aug 27 11:16:11 <cryptk> I will get that set up later after I sleep
Aug 27 11:16:23 <ka6sox> that way you have the configs setup and when it boots you have access.
Aug 27 11:16:47 <cryptk> my first boot I am just gonna have a script that runs at bootup that touches a file
Aug 27 11:17:00 <cryptk> memboot, wait a few minutes, reboot to webos and check for the file
Aug 27 11:17:03 <ka6sox> I will have already done that
Aug 27 11:17:07 <cryptk> ahh
Aug 27 11:17:20 <cryptk> well, I will get the SSH stuff setup after I get some sleep
Aug 27 11:17:28 <cryptk> I am in no shape to work with config files atm, lol
Aug 27 11:17:29 <ka6sox> since I don't wanna take the time to setup usbnet ssh
Aug 27 11:17:48 <cryptk> yeah, I will get it set up tomorrow and tar up a new rootfs for you after I test that
Aug 27 11:17:54 <ka6sox> kk
Aug 27 11:18:07 <cryptk> I /could/ set up a git repo for it...
Aug 27 11:18:09 <ka6sox> itmt I have a direction and enough to start testing.
Aug 27 11:18:15 <ka6sox> not yet
Aug 27 11:18:31 <cryptk> well, just to store the OS in... and have it so when I tag it, it will build a tarball
Aug 27 11:18:44 <cryptk> not a git repo on the WOSI git... have it on my git on CC
Aug 27 11:18:58 <cryptk> then once we are ready to announce it, move it to the WOSI git
Aug 27 11:19:37 <ka6sox> I'll just put the uImage on cc
Aug 27 11:19:42 <cryptk> kk
Aug 27 11:19:52 <cryptk> just drop it in my home directory
Aug 27 11:20:34 <ka6sox> will do
Aug 27 11:21:18 <ka6sox> the uImage will contain both the kernel(recovery) + an initramfs that will boot angstrom.
Aug 27 11:21:48 <cryptk> will boot angstrom from /dev/store/ext2fs right?
Aug 27 11:21:56 <ka6sox> yes
Aug 27 11:22:16 <ka6sox> it will pivot root out there.
Aug 27 11:22:53 <cryptk> cool
Aug 27 11:23:06 <cryptk> then we can start working simultaneously
Aug 27 11:23:50 <ka6sox> then I can test this on my ext3fs partition and see if it will boot out to there.
Aug 27 11:24:43 <ka6sox> with an initramfs it may.
Aug 27 11:24:53 <cryptk> yep
Aug 27 11:25:21 <ka6sox> because I can mount -t ext3 /dev/ext3fs /realroot
Aug 27 11:25:34 <ka6sox> and that will setup the pivot
Aug 27 11:26:21 <ka6sox> s/\dev\/ext3fs/\/dev\/mapper\/ext3fs
Aug 27 11:27:34 <cryptk> nice escaping
Aug 27 11:27:54 <ka6sox> :D
Aug 27 11:28:09 <ka6sox> okay what I need is the init script from /boot
Aug 27 11:28:39 <ka6sox> since both Angstrom and Ubuntu expect all the nods to be setup.
Aug 27 11:29:57 <ka6sox> Android sets them all up in the init.rc
Aug 27 11:30:02 <cryptk> in your home folder on CC
Aug 27 11:30:14 <cryptk> the /boot/sbin/init from webOS
Aug 27 11:30:31 <ka6sox> thanks
Aug 27 11:30:38 <ka6sox> I have it from my TP too :D
Aug 27 11:30:41 <cryptk> heh
Aug 27 11:30:53 <ka6sox> are you on 3.0.2?
Aug 27 11:31:35 <cryptk> yes
Aug 27 11:31:51 <cryptk> 3.0.2 build 68
Aug 27 11:32:04 <ka6sox> kk...good
Aug 27 11:32:45 <ka6sox> bleh...I'll have to chain the inits
Aug 27 11:34:46 <ka6sox> lessee what was that kernel bootarg line.....
Aug 27 11:35:29 <cryptk> bootargs = root=/dev/mmcblk0p13 rootwait ro
Aug 27 11:35:35 <cryptk> http://www.webos-internals.org/wiki/Bootie
Aug 27 11:36:30 <ka6sox> this will shortcircuit that.
Aug 27 11:36:43 <ka6sox> because the initramfs will take over.
Aug 27 11:36:47 <cryptk> yep
Aug 27 11:37:01 <cryptk> and I /think/ that is why I was not having any luck with overriding their values earlier
Aug 27 11:37:55 <ka6sox> I can barely get them to work right in uboot 1/2 the time.
Aug 27 11:38:00 <ka6sox> (the overrides)
Aug 27 11:38:28 <cryptk> heh
Aug 27 11:40:14 <ka6sox> this will be a 4 stage boot....
Aug 27 11:41:13 <cryptk> how many stages are there for the space shuttle launch?
Aug 27 11:41:22 <ka6sox> 3
Aug 27 11:41:31 <cryptk> yes! we beat the shuttle
Aug 27 11:44:13 <ka6sox> rrix, I have used android kernel to run linux apps.
Aug 27 11:44:27 <rrix> But have you used the linux kernel to run android? :)
Aug 27 11:44:34 <ka6sox> no
Aug 27 11:44:39 <ka6sox> I have not
Aug 27 11:45:51 <ka6sox> I have only ported *away* from Android.
Aug 27 11:46:07 * rrix still doesn't grok this android madness
Aug 27 11:46:12 <rrix> The platform is just ... ugh
Aug 27 11:46:44 <ka6sox> rrix, ya when porting away from Android I keep shaking my head when I look at it.
Aug 27 11:47:01 <rrix> There's just so much wrong with everything about android... :)
Aug 27 11:47:12 <rrix> Did they give up on trying to sync with mainline, btw?
Aug 27 11:47:26 <rrix> I haven't heard anything about that lately though I haven't been paying attention
Aug 27 11:47:36 <ka6sox> I think so...so its all TAKE and no GIVE.
Aug 27 11:47:49 <cryptk> brb
Aug 27 11:48:18 <rrix> ka6sox: That seems par the course with the Android community in general AFAICT
Aug 27 11:48:33 <rrix> from OHA all the way down to the homebrewers
Aug 27 11:48:49 <rrix> It's kinda depressing
Aug 27 11:49:12 <rrix> damnit, I rsynced this Qt stuff to my NAS >.<
Aug 27 11:49:13 <rrix> It's 5am
Aug 27 11:49:14 <ka6sox> its sad...
Aug 27 11:49:20 <rrix> Definitely
Aug 27 11:49:35 <rrix> .... gahhh, and my backup is derping
Aug 27 11:49:38 <ka6sox> its 5am and I am working on porting...why..
Aug 27 11:49:39 <rrix> silly chroot
Aug 27 11:49:55 <rrix> definitely just tried to backup /proc/kcore
Aug 27 11:50:08 <ka6sox> he he
Aug 27 11:50:17 <ka6sox> --exclude /proc
Aug 27 11:50:55 <rrix> well, my backup only grabs /home, but I have my chroot under /home/chroot, and everything is bind mounted in to it right now :(
Aug 27 11:51:27 <ka6sox> I have a plan, I don't like it but you can't produce a initramfs unless you have a working init
Aug 27 11:51:36 <ka6sox> strange as that seems.
Aug 27 11:53:58 <rrix> This has been me through all of you and cryptk's hacking tonight: http://rlv.zcache.com/poker_face_card-p137522142215148268tdn0_125.jpg
Aug 27 11:54:02 <rrix> Not my forte ;)
Aug 27 11:54:12 <cryptk> LOL
Aug 27 11:54:27 <rrix> *nod* that's cool *nod* ooooh
Aug 27 11:54:32 <cryptk> the initramfs stuff is strictly ka6sox 's territory
Aug 27 11:54:36 <cryptk> I'm not so good with that part
Aug 27 11:54:49 <cryptk> so I am just getting him whatever info he needs
Aug 27 11:54:56 <cryptk> once he has that part done, then I can get to work
Aug 27 11:55:04 <cryptk> I need the OS to boot in order to do my part
Aug 27 11:55:19 <rrix> The lowest I'm gonna get is doing some patching of the UX, prolly
Aug 27 11:55:22 <ka6sox> I'm just dressing things up for the party...making sure the tables are set and all the pieces are in place for the party to begin.
Aug 27 11:55:37 <rrix> Some folks in my hackerspace want kiosk-y tablets
Aug 27 11:55:38 <cryptk> yep
Aug 27 11:55:57 <rrix> heh
Aug 27 11:56:04 <rrix> You guys are doing awesome stuff in tandem, at any rate :)
Aug 27 11:56:53 <ka6sox> in reality...the /boot partition is exactly what I need to duplicate in the initramfs.
Aug 27 11:57:10 <ka6sox> and then a few tweaks..and we are golden.
Aug 27 11:57:52 <ka6sox> which mkinitramfs
Aug 27 11:58:01 <ka6sox> hmmm...not gonna work here.
Aug 27 11:58:15 <rrix> lol
Aug 27 11:58:24 <rrix> definitely 5
Aug 27 11:58:36 <rrix> god damnit, I'm backing up my dev stuff twice stupid chroot
Aug 27 11:58:40 * rrix excludes the whole fucking thing
Aug 27 11:59:43 * rrix is leaving on a weekend trip in a few hours and hasn't backed up in a month or so
Aug 27 11:59:50 <ka6sox> bzzzzt
Aug 27 11:59:59 <ka6sox> not present or accounted for.
Aug 27 12:00:26 <ka6sox> ipkg-opt -l mkinitramfs
Aug 27 12:00:42 <ka6sox> okay its 5am
Aug 27 12:01:09 <rrix> Good morning!
Aug 27 12:02:15 <ka6sox> rwhitby, I'm thinking about bypassing the /boot and going straight to pivot_rooting from the initramfs
Aug 27 12:02:58 <ka6sox> it will not change what is in /boot
Aug 27 12:03:23 <ka6sox> and a kernel reversion will make it boot normally again.
Aug 27 12:05:36 <ka6sox> that means that the only "mod" is checking the button, then mounting the appropriate realroot and pivot_rooting.
Aug 27 13:18:41 <cryptk> home and sleep time
Aug 27 13:18:46 <cryptk> nn all
Aug 27 17:19:57 <ka6sox> cat /devices/platform/gpio-keys/input/input0/event0
Aug 27 17:23:43 <rrix> root@ka6sox #
Aug 27 17:23:57 <ka6sox> he he
Aug 27 17:24:04 <ka6sox> what are you still doing up?
Aug 27 17:24:15 <ka6sox> (for that matter what am *I* doing up?
Aug 27 17:24:17 <ka6sox> )
Aug 27 17:24:19 <rrix> f if I know, lol.
Aug 27 17:24:32 <rrix> Waiting for my father to show up at my house so we can go on a roadtrip
Aug 27 17:24:53 <rrix> I was too lazy to scp my Qt binaries over to my device to test them last night, will probably do that today or something
Aug 27 17:24:58 <rrix> or sleep, since I didn't do that.
Aug 27 17:25:06 <ka6sox> I got a few hours.
Aug 27 17:25:14 <ka6sox> but I am trying to read that key.
Aug 27 17:25:31 <ka6sox> since the rest of the things won't be UP yet
Aug 27 17:25:33 <rrix> My roommate monopolized the dryer after I put my sheets in the washing machine so I couldn't sleep :(
Aug 27 17:25:46 <ka6sox> figures
Aug 27 17:25:57 <rrix> Such is life
Aug 27 17:26:18 <rrix> I cleaned my room and moved my furniture around so that my bed isn't against the hottest wall in the house now
Aug 27 17:26:27 <rrix> And I'm motivated to Do Stuff today
Aug 27 17:26:44 <rrix> If I can get the display part of this Qt lighthouse plugin working I'll be happy
Aug 27 17:26:47 <rrix> htg now
Aug 27 17:27:20 <rrix> have a good day ka6sox
Aug 27 17:28:19 <ka6sox> laters!
Aug 27 17:28:25 <ka6sox> have fun on the road trip
Aug 27 22:26:01 <ka6sox> Jack87|Away, ping?
Aug 27 22:40:52 <dtzWill> yay wosi-dev lives again
Aug 27 22:41:06 <dtzWill> was rather quiet in here ever since the tp madness
Aug 27 22:41:45 <dtzWill> ka6sox: any summary of what went on as i go through the backlog?
Aug 27 22:42:10 <dtzWill> loks like you guys had lots of fun with input devices
Aug 27 22:45:22 <rwhitby> dtzWill: how are we doing on keyboard mappings so we can put Xecutah/XServer/XTerm in public feed?
Aug 27 22:48:52 <cryptk> ka6sox, hows the initramfs magic coming along?
Aug 27 23:03:17 <ka6sox> dtzWill, ya, we need to ask for help on mapping :D
Aug 27 23:03:27 <ka6sox> and xrandr keeps coming up as a "question"
Aug 27 23:04:35 <ka6sox> cryptk|offline, I know *exactly* how to do it...
Aug 27 23:04:59 <ka6sox> I just need 2 more pieces of information
Aug 27 23:05:06 <dtzWill> ka6sox: heh, well i guess y'all are pushing for us to interface directly with unpinut or so to get key bindings directly?
Aug 27 23:05:21 <dtzWill> (or do we have other applications in mind?)
Aug 27 23:05:37 <ka6sox> uinput would help a bit...
Aug 27 23:05:37 <dtzWill> i lied when i said i was readnig backlog :P (not really, just got pulled away for something), doing so now.
Aug 27 23:05:44 <dtzWill> ka6sox: what information do you need?
Aug 27 23:05:49 <dtzWill> (and what are you envisioning?)
Aug 27 23:05:57 <ka6sox> might see if we can engage egaudet to help a bit
Aug 27 23:06:29 <ka6sox> dtzWill, can we stop the xrandr messages? (I know they are probably harmless but its disturbing the children)
Aug 27 23:07:22 <ka6sox> egaudet as he did some work with uinput a while back.
Aug 27 23:07:52 <ka6sox> what I'd like to do is prevent Luna from eating stuff if we can.
Aug 27 23:09:31 <ka6sox> are you relying on the SDL inputs?
Aug 27 23:10:48 <dtzWill> ka6sox: yep
Aug 27 23:11:10 <dtzWill> i imagine getting the inputs from lower level than luna is possible, and probably even decently do-able
Aug 27 23:11:20 <dtzWill> (if perhaps a good amount of discovery work)
Aug 27 23:11:27 <dtzWill> still have issue of luna seeing the same events
Aug 27 23:11:44 <ka6sox> some are harmless to us
Aug 27 23:11:48 <dtzWill> so we'd need to intercept/consume events somehow
Aug 27 23:11:51 <ka6sox> indeed most seem to be
Aug 27 23:12:01 <dtzWill> like if the user hits 'esc'--we want to get it and not have luna go into card mode or w/e
Aug 27 23:12:25 <ka6sox> then we need to try some "injection"
Aug 27 23:12:28 <ka6sox> into uinput
Aug 27 23:12:28 <dtzWill> also, wth OS/kernel are you guys going for?!
Aug 27 23:12:36 <dtzWill> well injection isn't it--we need more like interpositoin lol
Aug 27 23:12:47 <ka6sox> | tee blah
Aug 27 23:15:54 <dtzWill> cryptk|offline, ka6sox: sweet deals, and good luck!
Aug 27 23:16:15 <dtzWill> i'm rathre inclined to offer to help, but think i already am a bit saturated xD
Aug 27 23:16:50 <ka6sox> dtzWill, ya, I can see that...maybe some of the folks who seem to understand things in -internals can help you?
Aug 27 23:17:21 <dtzWill> well atm the only thing i have left really is xrandr
Aug 27 23:17:28 <dtzWill> escape is om nom'd by luna, i'm leaving that alone for now
Aug 27 23:18:50 <ka6sox> the hope I have is that we can tee off uinput and get those events before luna sucks them up
Aug 27 23:19:05 <ka6sox> for x-server (and stuff running inside it)
Aug 27 23:19:16 <ka6sox> we need that to be available.
Aug 27 23:19:40 <ka6sox> ctrl modifier/Alt modifier/ ESC
Aug 27 23:20:53 <dtzWill> ka6sox: well ctrl/alt already work--that would _all_ be for 'esc' afaict ;)
Aug 27 23:21:16 <dtzWill> maybe i can just make the keyboard dismiss button 'esc' or so lol
Aug 27 23:21:17 <dtzWill> call it a day
Aug 27 23:21:31 <dtzWill> oh well i guess bt keyboards still need an 'esc' button, darn :P
Aug 27 23:21:36 <ka6sox> yes
Aug 27 23:21:55 <ka6sox> maybe we map some other keyboard and use that one if the HP one is borked.
Aug 27 23:22:19 <dtzWill> ka6sox: the hp bt? I think it's not so much borked as luna eats the 'esc' key
Aug 27 23:23:27 <ka6sox> so we have to watch for that one sequence from uinput
Aug 27 23:24:03 <ka6sox> the BT keyboard's ctrl doesnt' work either
Aug 27 23:24:11 <ka6sox> I am not married to the HP keyboard.
Aug 27 23:24:23 <ka6sox> I *do* have other BT keyboards.
Aug 27 23:25:33 <dtzWill> ka6sox: the bluetooth's control key does work
Aug 27 23:25:45 <dtzWill> ka6sox: it works on mine, and reported to work on the apple keyboard as well as a generic bluetooth keyboard?
Aug 27 23:26:03 <ka6sox> these are things I would *love* to do but I simply lack time.
Aug 27 23:26:16 <ka6sox> I learned a LOT when I did the mappings the other day with xev
Aug 27 23:26:25 <dtzWill> the only key atm that doesn't work is escape--and to some extent 'shift', but that's not important.
Aug 27 23:26:39 <ka6sox> ya thats kind of "interesting"
Aug 27 23:26:43 <ka6sox> the shift issue
Aug 27 23:26:54 <ka6sox> the only key I absolutely need is esc
Aug 27 23:26:54 <dtzWill> oh, well that's my 'fault'
Aug 27 23:27:06 <dtzWill> i know exactly how to fix that for bt keyboards
Aug 27 23:27:10 <ka6sox> I'm not trying to assign blame.
Aug 27 23:27:16 <dtzWill> shrug, i'm just saying i understand the issue.
Aug 27 23:27:22 <dtzWill> well
Aug 27 23:27:25 <dtzWill> i think i do, should be careful O:)
Aug 27 23:29:26 <ka6sox> I need some help mapping events of the buttons...
Aug 27 23:29:49 <dtzWill> ka6sox: i don't mean to be difficult--what do you mena 'mapping'?
Aug 27 23:30:13 <dtzWill> what is trying to interpret them and what information mismatch are you hoping to build a map to bridge?
Aug 27 23:30:34 <ka6sox> oh, I mean finding either sysfs entries or events that we can tag to watch for changes.
Aug 27 23:32:03 <dtzWill> oh i see. if i understand the wiki/the backlog, we're still missing the exact way to get generic keyboard events?
Aug 27 23:32:29 <ka6sox> yes
Aug 27 23:32:31 <dtzWill> someone in -internals was playing with 'showkey'--maybe we can just look at the guy and see what it does
Aug 27 23:32:41 <dtzWill> might be too low-level, but it'd probably work just ifne.
Aug 27 23:32:52 * dtzWill likes making solutions that 'wokr' first and worrying about work /well/ as a second pass ;)
Aug 27 23:32:59 <ka6sox> I need the volume/power and home keys.
Aug 27 23:33:14 <ka6sox> since when I"m booting thats whats available.
Aug 27 23:34:45 <dtzWill> oh, of course.  no vkb obviously at first and no bt stack lolol
Aug 27 23:34:52 * dtzWill imagines morse-code based shell
Aug 27 23:35:07 <ka6sox> he he....
Aug 27 23:35:43 <ka6sox> uNiXpSyChO, has a good way to create a "menu" system but it relies on watching those key events.
Aug 27 23:45:23 <ka6sox> since they are the only thing we have during boot.
Aug 27 23:52:10 <dtzWill> ka6sox: you're right i need to get xrandr support immeidately so people stop complaning xD
Aug 27 23:52:33 <ka6sox> or...stub it out so it stops complaining.
Aug 27 23:52:43 <dtzWill> ka6sox: right. i thought i already had that, guess i'm misremembering.
Aug 27 23:57:50 <dtzWill> rwhitby: is there a reason i couldn't bump the HEADLESSAPP_VERSION to 0.2.0? that'd get them full-screened on the tp (and still be backwards-compat with phones), right?
Aug 28 02:11:52 <rwhitby> dtzWill: please do
Aug 28 02:12:21 <dtzWill> rwhitby: kk
Aug 28 02:13:17 <rwhitby> I think that we just need to release this stuff with "Keyboard mappings are incomplete, looking for X11 experts to help flesh them out" in the documentation.
Aug 28 02:22:32 <dtzWill> rwhitby: the only outstanding issue with keyboard mappings (afaik?) is that escape doesn't work
Aug 28 02:22:36 <dtzWill> and that's because luna eats it
Aug 28 02:22:36 <dtzWill> lol
Aug 28 02:22:37 <dtzWill> :(
Aug 28 02:22:52 <dtzWill> as for the transient errors where they don't get set
Aug 28 02:22:56 <dtzWill> i'm working to fix that rtfn
Aug 28 02:27:27 <ka6sox> hmmm...xecutah 0.8.4 won't install for me.
Aug 28 02:33:05 <ka6sox> curl keeps complaining
Aug 28 02:37:27 <dtzWill> rwhitby: iirc we needed the xecutah.sh scripts to run fast, right?
Aug 28 02:37:38 <dtzWill> i'm trying to design a solution to the xkb (and more generally lack of .xinitrc support)
Aug 28 02:38:05 <dtzWill> and i'd like to have xserver's xecutah sleep for a bit
Aug 28 02:41:23 <dtzWill> YAY worky
Aug 28 02:41:23 <dtzWill> haha
Aug 28 02:43:54 <dtzWill> oh i see, the sleeping in xecutah.sh means xecutah can't do anything until that times out
Aug 28 02:44:00 <ka6sox> okay xecutah worked after the 3rd try :D
Aug 28 02:47:41 <ka6sox> cryptk, it works.
Aug 28 02:47:51 <ka6sox> for booting webos
Aug 28 02:48:03 <rwhitby> dtzWill: we can put spinners in Xecutah until the command finishes
Aug 28 02:48:12 <cryptk> NICE!
Aug 28 02:48:26 <cryptk> now get me one that does the pivot_root and I can start getting angstrom configged
Aug 28 02:49:39 <ka6sox> now going to repoint it to at least try to boot up angstrom..I've added a few touches.
Aug 28 02:54:32 <dtzWill> wonderful, this is working nicely
Aug 28 02:54:35 <dtzWill> rwhitby: i think i can do without that
Aug 28 02:55:20 <ka6sox> cryptk, it already does the pivot_root as the final thing it does.
Aug 28 02:55:53 <ka6sox> currently its not setting up a different realroot (the one you want)
Aug 28 02:56:01 <cryptk> nice
Aug 28 02:56:10 <cryptk> so all you need to do is adjust where it pivot roots to
Aug 28 02:56:19 <ka6sox> yes, its 1 line
Aug 28 02:56:31 <cryptk> sweet, then adjust it and built it so I can membooterate ;)
Aug 28 02:56:32 <ka6sox> but I need the actual event to watch.
Aug 28 02:56:38 <cryptk> what event?
Aug 28 02:56:51 <cryptk> just get me one that does the pivot to the ext2fs LVM so I can start my work
Aug 28 02:56:51 <ka6sox> KEY_VOLUME_DOWN
Aug 28 02:56:59 <ka6sox> or event id 114
Aug 28 02:57:28 <cryptk> what do you need? the kernel event?
Aug 28 02:57:40 <ka6sox> sysfs entry
Aug 28 02:57:58 <cryptk> lemem track it down
Aug 28 02:58:02 <ka6sox> I can't depend on anything more advanced that that to be available@that point.
Aug 28 02:58:33 <cryptk>  /sys/devices/platform/gpio-keys/input/input0/event0
Aug 28 02:58:45 <ka6sox> nope
Aug 28 02:58:47 <ka6sox> not there
Aug 28 02:58:53 <ka6sox> thats a directory
**** ENDING LOGGING AT Sun Aug 28 02:59:57 2011