**** BEGIN LOGGING AT Mon Jun 18 02:59:58 2012
Jun 18 05:27:52 <CIA-76> SHR: 03shr-devel 07buildhistory * r57b97a8c652d 10/packages/ (15 files in 15 dirs): packages: Build 201206180642 of shr 20120618 for machine om-gta02 on opmbuild
Jun 18 09:26:15 <mrmoku> hi GNUtoo-desktop 
Jun 18 09:57:50 <CIA-76> SHR: 03shr-devel 07buildhistory * rb40d3eb35b6d 10/packages/ (398 files in 398 dirs): packages: Build 201206180829 of shr 20120618 for machine om-gta04 on opmbuild
Jun 18 10:29:20 <CIA-76> SHR: 03shr-devel 07buildhistory * r6b6404cde2d1 10/images/crespo/eglibc/chroot-image/installed-package-sizes.txt: images: Build 201206181223 of shr 20120618 for machine crespo on opmbuild
Jun 18 13:10:13 <CIA-76> SHR: 03Martin.Jansa 07shr-chroot * r6cfc75156d4b 10/ (220 files in 28 dirs): system upgrade
Jun 18 14:22:39 <PaulFertser> This is one of the ways it fails on resume for me, my uSD getting remounted r/o http://paste.debian.net/175086/
Jun 18 14:39:50 <CIA-76> SHR: 03shr-devel 07buildhistory * r472b75bc7c89 10/packages/ (165 files in 165 dirs): packages: Build 201206181525 of shr 20120618 for machine om-gta02 on opmbuild
Jun 18 15:28:20 <rtp> PaulFertser: very stupid guess: mmc unsafe resume not enabled ? /o\
Jun 18 15:37:47 <CIA-76> SHR: 03shr-devel 07buildhistory * rfeebbb800a5a 10/packages/ (165 files in 165 dirs): packages: Build 201206181650 of shr 20120618 for machine nokia900 on opmbuild
Jun 18 16:06:15 <CIA-76> SHR: 03shr-devel 07buildhistory * rb6d8b02f996b 10/packages/om_gta04-oe-linux-gnueabi/shr-version/ (latest shr-version/latest): packages: Build 201206181747 of shr 20120618 for machine om-gta04 on opmbuild
Jun 18 16:35:02 <CIA-76> SHR: 03shr-devel 07buildhistory * r89779da67720 10/images/crespo/eglibc/chroot-image/ (4 files): images: Build 201206181815 of shr 20120618 for machine crespo on opmbuild
Jun 18 16:35:12 <CIA-76> SHR: 03shr-devel 07buildhistory * r99c4ff9cd03b 10/packages/crespo-oe-linux-gnueabi/shr-version/ (latest shr-version/latest): packages: Build 201206181815 of shr 20120618 for machine crespo on opmbuild
Jun 18 16:47:08 <CIA-76> freesmartphone.org: 03morphis 07morphis/hfp * rd0ee12d00f5d 10cornucopia/fsogsmd/src/lib/bluetooth.vala: fsogsmd: lib: bluetooth: handle service connect and disconnect
Jun 18 16:47:08 <CIA-76> freesmartphone.org: 03morphis 07morphis/hfp * r23fc6aa84b1c 10cornucopia/fsogsmd/src/plugins/dbus_service/plugin.vala: fsogsmd: dbus_service: be verbose when registering/unregistering a modem
Jun 18 16:47:09 <CIA-76> freesmartphone.org: 03morphis 07morphis/hfp * rc88ba45dc1bf 10cornucopia/libfsotransport/fsotransport/ (Makefile.am unix.vala): libfsotransport: a new unix filedescriptor based transport type
Jun 18 16:47:10 <CIA-76> freesmartphone.org: 03morphis 07morphis/hfp * r093743ba7ea5 10cornucopia/fsogsmd/src/lib/modem.vala: fsogsmd: lib: modem: add async modifier for powerOn/powerOff methods
Jun 18 16:47:10 <CIA-76> freesmartphone.org: 03morphis 07morphis/hfp * r8260eb33ddd5 10cornucopia/libfsotransport/fsotransport/commandqueue.vala: libfsotransport: don't assert when transport is already open but warn the user
Jun 18 16:47:11 <CIA-76> freesmartphone.org: 03morphis 07morphis/hfp * r0e83aa1f5020 10cornucopia/libfsotransport/fsotransport/unix.vala: libfsotransport: unix: we have a different open behavior for this type of transport
Jun 18 16:47:11 <CIA-76> freesmartphone.org: 03morphis 07morphis/hfp * r256adac63548 10cornucopia/fsogsmd/src/lib/at/atcommandqueue.vala: fsogsmd: lib: accept an AtCommandQueue instead of a AtChannel for a AtCommandSequence
Jun 18 16:47:12 <CIA-76> freesmartphone.org: 03morphis 07morphis/hfp * r5c786384b080 10cornucopia/fsogsmd/src/plugins/modem_hfp_hf/ (Makefile.am channel.vala modem.vala): fsogsmd: modem_hfp_hf: implement transport and channel setup procedure
Jun 18 16:52:42 <DocScrutinizer_> http://www.stericsson.com/press_releases/PR_Samsung_U8500.jsp
Jun 18 17:01:03 <CIA-76> SHR: 03shr-devel 07buildhistory * r33a41e90aa32 10/packages/palmpre-oe-linux-gnueabi/shr-version/ (latest shr-version/latest): packages: Build 201206181843 of shr 20120618 for machine palmpre on opmbuild
Jun 18 17:26:36 <PaulFertser> rtp: heh :) it works most of the time, moreover, that's the stock radekp's kernel.
Jun 18 17:26:49 <PaulFertser> That other folks are using with QtMoko
Jun 18 17:27:47 <CIA-76> SHR: 03shr-devel 07buildhistory * rdf61017027b7 10/packages/palmpre2-oe-linux-gnueabi/shr-version/ (latest shr-version/latest): packages: Build 201206181909 of shr 20120618 for machine palmpre2 on opmbuild
Jun 18 17:29:23 <PaulFertser> rtp: btw, hi, how're you doing?
Jun 18 17:31:57 <PaulFertser> DocScrutinizer: is that related to your work anyhow? I thought you're busy with LTE basebands.
Jun 18 17:34:05 <PaulFertser> BTW, is that U8500 modem controllable with free software?
Jun 18 17:35:50 <Alex[sp3dev]> I state here that uboot is utter crap and I hate USB
Jun 18 17:38:33 <PaulFertser> Alex[sp3dev]: what do you think about barebox? After dealing with u-boot i found barebox quite enjoyable.
Jun 18 17:38:40 <PaulFertser> And usb is a pain indeed.
Jun 18 17:39:56 <Alex[sp3dev]> PaulFertser: I have not really tried barebox yet but here is what I hate about uboot: ifdeffery and lack of menu-based config, no usb update mechanism and two incompatible usb stacks (gadget and udc), hush shell (as if C is not enough)..
Jun 18 17:41:44 <Alex[sp3dev]> huh. barebox has DFU? wonderful. But I don't feel like porting all omap4 support to it
Jun 18 17:53:46 <DocScrutinizer51> hi PaulFertser 
Jun 18 17:54:55 <DocScrutinizer51> PaulFertser: well, LTE is current project, since we want to sell chips for devices yet to build
Jun 18 17:55:13 <DocScrutinizer51> 8500 is bread+butter
Jun 18 17:56:06 <DocScrutinizer51> it's in no way special, just like always in industry
Jun 18 18:07:50 <PaulFertser> Alex[sp3dev]: it has menuconfig though. And considerably less ifdef hackery :)
Jun 18 18:09:53 <PaulFertser> DocScrutinizer51: do you think there is a chance those android devices would be suitable for running a free software only telephony stack?
Jun 18 18:10:16 <DocScrutinizer05> why not? It's AT over CAIF
Jun 18 18:10:24 <DocScrutinizer05> afaik
Jun 18 18:10:48 <DocScrutinizer05> would need to check if CAIF-SHM or CAIF-HSI
Jun 18 18:26:24 <PaulFertser> DocScrutinizer51: oh, nice. Though sometimes partially documented AT set is nasty enough too.
Jun 18 18:27:48 <DocScrutinizer05> most commands are standard, some maybe not. The modem binary shouldn't be encrypted and plain ARM though. And if nothing else helps, I may give an educated guess ;-)
Jun 18 18:29:25 <DocScrutinizer05> (though dunno how long)
Jun 18 18:30:10 <DocScrutinizer05> unlike OM I'm not supposed to play master of src here ;-D
Jun 18 19:29:01 <GNUtoo-desktop> hi DocScrutinizer 
Jun 18 19:32:22 <Alex[sp3dev]_> whoa I have a shitload of dmesg and rild strace. lets see if it has something useful
Jun 18 19:33:07 <Alex[sp3dev]_> GNUtoo-desktop: btw, I have now ANDROID! images working in uboot. so only fastboot is left and for that omap usb driver will probably need porting to gadget structure
Jun 18 19:33:21 <GNUtoo-desktop> ok, do you have gralloc working?
Jun 18 19:34:06 <Alex[sp3dev]_> GNUtoo-desktop: no, need to finish bootloader side first. I will look at gralloc next week, after the exams
Jun 18 19:34:16 <GNUtoo-desktop> ok
Jun 18 19:34:22 <GNUtoo-desktop> mirage335, he just said that:
Jun 18 19:34:26 <GNUtoo-desktop> <Alex[sp3dev]_> GNUtoo-desktop: no, need to finish bootloader side first. I will look at gralloc next week, after the exams
Jun 18 19:34:37 <mirage335> Oh.
Jun 18 19:34:41 <GNUtoo-desktop> Alex[sp3dev]_, mirage335 is the user with galaxy nexus
Jun 18 19:35:00 <Alex[sp3dev]_> GNUtoo-desktop: ok. is he going to hack anything onto it?
Jun 18 19:35:20 <GNUtoo-desktop> he can compile and test stuff
Jun 18 19:35:30 <GNUtoo-desktop> he wants replicant on it
Jun 18 19:35:39 <mirage335> Alex[sp3dev]_: Yes, I already have CM9, CWM and such on it. I have been trying to get the FOSS libgralloc to work on it for a while so replicant can support the Galaxy Nexus.
Jun 18 19:36:07 <Alex[sp3dev]_> GNUtoo-desktop: here is the cleaned uboot branch https://gitorious.org/i9100-proper-linux/i9100-uboot/commits/tuna_clean .
Jun 18 19:36:24 <GNUtoo-desktop> ok
Jun 18 19:37:47 <mirage335> We do already have some free code that works, from the recovery firmware. Unfortunately, it does not use libgralloc like CM (which we are modding into replicant), and porting the code there is fairly challenging for me.
Jun 18 19:38:10 <Alex[sp3dev]_> mirage335: I have played with the library yesterday
Jun 18 19:38:43 <mirage335> Good. Have you seen the wavy lines yet?
Jun 18 19:39:09 <Alex[sp3dev]_> mirage335: according to how the image looks, I think the framebuffer might not be completely linear (like, [XXXXXXXX, YYYYYYYY] where X and Y belong to first and second buffer) but stripped like [XYXYXYXY...]
Jun 18 19:40:32 <Alex[sp3dev]_> probably the stride is wrong or something but my priority now is uboot. partially because I want to reuse fastboot code later on other devices (I actually want ubuntu on the galaxy s2, but that's another story)
Jun 18 19:40:34 <mirage335> Hmm, I would never have thought of that.
Jun 18 19:41:41 <mirage335> Alex[sp3dev]_: Ok, so the ioctl line in the code might need just a number change?
Jun 18 19:41:55 <mirage335> I could certainly that at random, if the number is small.
Jun 18 19:43:07 <Alex[sp3dev]_> mirage335: not the ioctl, but the framebuffer parameters like stride, probably. I tried adding some debug to the kernel driver but that didn't help much
Jun 18 19:44:05 <Alex[sp3dev]_> first of all, is fb really rgb and not yuv? maybe test via /dev/graphics/fb0?
Jun 18 19:44:49 <mirage335> TBH, I have no idea how to try that. I have not been successful in uploading valid frambuffer captures (from CM9) to /dev/graphics/fb0.
Jun 18 19:45:17 <mirage335> I did try all formats from the recovery firmware however, so if it was *just* the pixel format, I think I would have found that.
Jun 18 19:45:37 <mirage335> As in, I modified gralloc to use all the pixel formats the recovery firmware can use.
Jun 18 19:46:18 <Alex[sp3dev]_> I actually suppose the whole gralloc thingie may be wrong since pmem is deprecated.
Jun 18 19:46:28 <Alex[sp3dev]_> maybe we should disassemble the blob
Jun 18 19:46:51 <Alex[sp3dev]_> I'm just a bit busy this week with the exams and right now I'm exploring the modem traces
Jun 18 19:47:03 <mirage335> Unfortunately, you've lost me with that. I have no idea what tools would allow us to disassemble the blob on android, and I'm not familiar with android.
Jun 18 19:47:27 <mirage335> GNUtoo-desktop: Maybe you can make more out of that statement than me.
Jun 18 19:47:38 <Alex[sp3dev]_> mirage335: IDA Pro (at least the demo. or you could disassemble the disassembler and hack it) or just objdump -d
Jun 18 19:48:36 <mirage335> That's it? Sounds pretty difficult.
Jun 18 19:49:32 <mirage335> Sounds worse than trying to figure out what the recovery firmware is doing differently, in fact.
Jun 18 19:50:28 <mirage335> Alex[sp3dev]_: Also, to clarify, you're saying that pmem is used by gralloc, but not the recovery firmware at this point?
Jun 18 19:51:36 <Alex[sp3dev]_> mirage335: no, I've not checked. I just suspect that pmem is not used on omap and we should use ion memory manager or omap ioctls for memory allocation
Jun 18 19:52:26 <GNUtoo-desktop> indeed pmem is not in the tuna kernel
Jun 18 19:52:33 <GNUtoo-desktop> there is ion or something else tough
Jun 18 19:52:39 <mirage335> I see.
Jun 18 19:53:09 <mirage335> If it helps "grep -r -i pmem ." in "external/bootmenu/minu" returns nothing.
Jun 18 19:56:39 <Alex[sp3dev]_> hmm. why not just.. strace it..
Jun 18 19:57:13 <mirage335> Because I have no idea how that would be done when the phone is just booting up the recovery firmware.
Jun 18 19:59:46 <Alex[sp3dev]_> ok, made some traces, let me see if they contain anything useful
Jun 18 20:03:29 <Alex[sp3dev]_> made a trace of surfaceflinger, but that's not extremely useful, only fbioget, no fbioput
Jun 18 20:04:31 <GNUtoo-desktop> also the non-free gralloc uses non-free 3d graphics so that might be less usefull since the system is quite different
Jun 18 20:05:04 <GNUtoo-desktop> Alex[sp3dev]_, btw you got GNU/Linux on galaxy nexus too?
Jun 18 20:05:59 <Alex[sp3dev]_> GNUtoo-desktop: not yet, it's a work phone and the aim is to make sure the boot path does not contain rootkits. actually, proprietary libs are OK since we can probably limit what they can access via MMU
Jun 18 20:06:14 <GNUtoo-desktop> ok
Jun 18 20:07:29 <Alex[sp3dev]_> of course RIL, sound, GPS must be free. but when it comes to using GNU/Linux, I would agree to use PowerVR blobs. because I both like to practice opengl and because I'm not yet morally ready to rewriting an opengl driver
Jun 18 20:07:53 <Alex[sp3dev]_> ok, back to tracing
Jun 18 20:10:44 <mirage335> GNUtoo-desktop: Would that path be acceptable for Replicant? Because if we can limit what the blobs do, I don't see an absolute need to purge them either.
Jun 18 20:11:22 <Alex[sp3dev]_> mirage335: nope
Jun 18 20:11:36 <mirage335> ?
Jun 18 20:12:38 <Alex[sp3dev]_> mirage335: actually, it's not only about freedom (well, I would not trust free software running on proprietary hardware anyway) but about being able to reuse technology in other projects
Jun 18 20:13:17 <mirage335> Righto. I'm just saying I don't mind the PowerVR blobs too much as long as they can't do evil things.
Jun 18 20:13:46 <mirage335> And I am more suspicious of GPS/modem things myself.
Jun 18 20:13:57 <mirage335> Especially modem.
Jun 18 20:15:16 <Alex[sp3dev]_> mirage335: looks like on omap4, the BP does not have access to the AP RAM. still, we need to write the initialization code for fw loading. I have made some strace on my sgs2 and plan to analyze that
Jun 18 20:16:34 <mirage335> Ok.
Jun 18 20:17:17 <mirage335> Are you saying the SGS2 needs the same code to work?
Jun 18 20:18:27 <Alex[sp3dev]_> mirage335: yes, I think 95% of code will be the same
Jun 18 20:18:36 <Alex[sp3dev]_> either way, I will try to support both
Jun 18 20:19:04 <mirage335> Ok.
Jun 18 20:20:08 <mirage335> Honestly though, I am not following this conversation all that well. For one thing, I don't know why we need to trace the propietary driver when the recovery firmware already works.
Jun 18 20:20:59 <Alex[sp3dev]_> mirage335: that's a bit easier. but I'm trying to strace recovery
Jun 18 20:21:21 <mirage335> Ok.
Jun 18 20:38:12 <GNUtoo-desktop> mirage335, no blobs in replicant
Jun 18 20:39:26 <mirage335> GNUtoo-desktop: Ok. I know that's the final goal, but in the short term, I'm thinking we still have the modem blobs to take care of, as well as some applications, and that may deserve to get done sooner.
Jun 18 20:40:26 <GNUtoo-desktop> then at the end you forget the long term goal.....
Jun 18 20:41:17 <GNUtoo-desktop> you're free to do what you want personally
Jun 18 20:41:23 <GNUtoo-desktop> but replicant will never accept blobs
Jun 18 20:41:29 <mirage335> Ok.
Jun 18 21:01:45 <Alex[sp3dev]_> we may reuse gralloc from qcom for ion. as for fb, I'm still debugging
Jun 18 22:08:11 <GNUtoo-desktop> ok
**** ENDING LOGGING AT Tue Jun 19 02:59:58 2012