**** BEGIN LOGGING AT Tue Sep 07 02:59:57 2010 Sep 07 07:12:29 ogra: morning! Sep 07 07:14:06 ogra: as I told you, I am trying the beta image, the installation went fine. everything seemed to be functional (i am using our kernel). then I rebooted, the splash screen is nice (good resolution) but when the desktop comes the resolution is screwed up totally. i remember rsalveti mentioning something similar. have you seen this already? Sep 07 07:33:56 cooloney: I have problems generating the header package with the 2.6.35 branch: I get only 1 header package (linux-headers-2.6.35-903-omap4), which depends on another one that is not generated (linux-ti-omap4-headers-2.6.35-903). On the 2.6.34 branch, there used to be 2 header packages. Sep 07 07:35:52 morning Sep 07 07:39:38 ogra: one more question: what are the right offset in the 10.10 pre installed images, if I want to loop mount the boot partition and the ext3 partition? Sep 07 07:40:50 sebjan: are you using cross compiling or schroot Sep 07 07:45:55 sebjan, are you building just armel, or i386 as well ... Sep 07 07:47:38 ogra: ok, I found for the offset. fdisk -ul on the .img file gives me the offset in cylinders... that works. Sep 07 07:47:46 cooloney, apw: I see that this issue is not just when I build. The packages listed above are on the official ppa. The headers package on the official ppa cannot be installed because of this broken dependency. Sep 07 07:48:53 apw: currently, i copied the header package name from debian.master Sep 07 07:48:57 which is linux-headers-PKGVER-ABINUM Sep 07 07:49:21 in debian.ti-omap4/control.stub.in Sep 07 07:49:25 cooloney, but did you ever get it generated Sep 07 07:49:29 when you tested Sep 07 07:49:51 yeah, it generated linux-headers-2.6.35-903-omap4 Sep 07 07:50:53 apw: i guess we need to change it as SRCPKGNAME-headers-PKGVER-ABINUM? Sep 07 07:51:18 not sure i'd expect that ... Sep 07 07:51:38 apw, how about our debian.master? Sep 07 07:51:46 do we got 2 header packages? Sep 07 07:51:59 i think i'd expect the linkage to be to linux-headers in both cases Sep 07 07:52:18 the primary header is linux-headers-....-omap4 right ? Sep 07 07:52:31 right, Sep 07 07:52:46 so the common header should be callled linux-headers-.... Sep 07 07:52:58 without the ti-omap Sep 07 07:53:02 according to the debian.ti-omap4/control.stub.in Sep 07 07:53:30 apw: i'm not sure about the name of that. do you know where can i found it? Sep 07 07:54:06 * apw will have a look at the packaging Sep 07 07:54:29 apw: thanks a lot. Sep 07 07:58:37 ndec, well, i have a working monitor and havent seen any issues, i guess you have to wait for ricardo Sep 07 07:59:22 apw, linux-headers-2.6.35-903-omap4 : Depends: linux-ti-omap4-headers-2.6.35-903 but it is not installable ... Sep 07 07:59:28 ogra: but my monitor is working too.. i was able to install and reboot the first time. everything was ok, i had the desktop and UNE full screen (full HD). then I rebooted and it no longer worked. Sep 07 07:59:30 apw, i have to check whats up there Sep 07 08:00:04 ndec, hmm ... Sep 07 08:00:12 * ogra wonders what that could be Sep 07 08:01:04 ndec, do you use any special boot.scr ? Sep 07 08:01:17 ogra, yeah cooloney mentioned it just a few mins ago ... i see noone has even build-tested these uploads ... Sep 07 08:01:44 ogra: I just added mem=1G and regenerated it. Sep 07 08:02:00 ndec, oh, i dont think that works Sep 07 08:02:22 ogra: yes, with the latest kernel it does Sep 07 08:02:23 ndec, mem=460M@0x80000000 mem=512M@0xA0000000 is what we got from sebjan Sep 07 08:02:30 ndec: did not panda required memory hole for dsp or something? Sep 07 08:02:36 like ogra shown Sep 07 08:02:44 (i think, not 100% sure where rsalveti got it from) Sep 07 08:02:48 ogra: i know about this. but this hole is only if you actually use the DSP. in my case I am not using it. Sep 07 08:03:37 well, the only thing i can imagine that changed is boot.scr Sep 07 08:04:11 ogra: i am restarting from scratch...let's see if I can reproduce. Sep 07 08:04:18 setenv bootargs quiet splash ro elevator=noop vram=32M mem=460M@0x80000000 mem=512M@0xA0000000 fixrtc Sep 07 08:04:25 thats our bootargs line atm Sep 07 08:13:44 cooloney, why would you set "suspend doesnt work" to wontfix ?? Sep 07 08:16:29 cooloney, apw, ogra : fyi, I just logged a defect for the broken headers package: 632235 Sep 07 08:16:49 sebjan, ah, thanks, i guess apw is already on it Sep 07 08:16:55 bug #632235 Sep 07 08:16:56 Launchpad bug 632235 in linux-ti-omap4 (Ubuntu) "2.6.35-903-omap4 header package cannot install - dependency broken (affects: 1) (heat: 8)" [Undecided,New] https://launchpad.net/bugs/632235 Sep 07 08:20:08 god arm builds are pitifully slow Sep 07 08:22:27 ogra: CONFIG_PM is not available on omap4 Sep 07 08:22:44 can we use suspend on it? Sep 07 08:23:24 does suspend mean anything on ARM, don't we just do clock/power management to near 0 consumption alll the time Sep 07 08:25:01 apw: i'm not sure about that PM can support near 0 consumption, but suspend is a menu option on Ubuntu, right? Sep 07 08:25:17 yep a menu option Sep 07 08:25:41 sebjan: did you guys try suspend/hibernation such things on panda? Sep 07 08:26:14 apw: suspend isn't frequently used Sep 07 08:26:45 amitk, i'd always assumed that PM on arm was essentially the same, that it could sleep everything but the ram but without explicit suspend Sep 07 08:26:54 correct Sep 07 08:27:15 ideally, i think Sep 07 08:27:46 if the PM framework (clock, regulators, idle call for system) is enabled for all drivers, that is what should happen Sep 07 08:27:57 well, all i want to be able to do is to click on the susped option in the menu and to properly power up again later Sep 07 08:28:28 ogra: i'm afraid that we can't do that on omap4 panda Sep 07 08:28:28 cooloney, there are devices based on the panda that you definitely want to be able to suspend Sep 07 08:28:36 which disabled CONFIG_PM Sep 07 08:28:40 cooloney, why ? Sep 07 08:28:50 we could do it on dove and babbage Sep 07 08:28:59 ogra_cmpc: whether or not that works will depend on if the drivers have even implemented the suspend/resume hooks Sep 07 08:29:02 yeah, we can do that Sep 07 08:29:10 and if i own a panda based netbook i definitely want to suspend it Sep 07 08:29:11 but from TI's patches, Sep 07 08:29:12 * amitk hasn't checked Sep 07 08:29:18 for omap4 Sep 07 08:29:19 they sugguest we disable that Sep 07 08:29:40 sebjan and ndec, right? for CONFIG_PM? ^^^ Sep 07 08:30:17 yes, we disabled the PM for this release as some drivers do not support it right now Sep 07 08:30:17 ogra_cmpc: yeah, i agree we need suspend. but maybe TI think CONFIG_PM is not ready for our omap4 Sep 07 08:32:35 sebjan: have you measured any differences in the system consumption (suspend vs. OFF) Sep 07 08:32:38 ? Sep 07 08:32:41 cooloney: ogra: amitk: PM is disabled for now in OMAP4, the code for this is being developed right now, that will be activated later, but it's too early for 10.10 Sep 07 08:32:45 sebjan: will that be available in the release of Sep Sep 07 08:32:59 cooloney: no PM in 10.10 Sep 07 08:33:05 ndec: ok, understood Sep 07 08:33:19 ndec: have you measured any differences in the system consumption (suspend vs. OFF states)? Sep 07 08:33:58 amitk: we don't have enough to share now. Sep 07 08:36:04 ogra: rsalveti: if I turn my display ON a few sec after boot, the resolution is fine. I think i have same problem with race condition and X start... Sep 07 08:38:25 ndec, hmm, that must be a DSS2 issue then Sep 07 08:38:54 i think i saw patches to prevent double activation on the kernel ML Sep 07 08:42:12 cooloney, ok, i'll de-milestone the bug then and adjust the title, but i wont close it so we can track the issue Sep 07 08:44:00 ogra_cmpc: np, that's the right way as we all know the story Sep 07 08:44:06 yup Sep 07 08:44:45 cooloney, and we have a PM team in linaro we can nag all the time if it doesnt work *g* Sep 07 08:44:54 * ogra_cmpc waves to amitk Sep 07 08:45:14 ogra_cmpc: great, thx Sep 07 08:47:53 * amitk waves abck Sep 07 08:47:55 back Sep 07 12:19:41 WOW Sep 07 12:19:55 * ogra is impressed by the panda running at !GHz and with 1GB Sep 07 12:20:00 *1GHz Sep 07 12:20:12 oem-config was done in under 10min Sep 07 12:24:14 ogra: 10mins from first poweron to useable desktop? Sep 07 12:31:21 amitk, rather 12 ... resizing takes about 2min, then reboot then 10min for oem-config Sep 07 12:31:34 the slow part is the SD card now Sep 07 12:31:55 the boot is impressive, after u-boot its only taking 10sec Sep 07 12:32:12 sadly its still 15-20sec for u-boot ... Sep 07 12:36:59 moop Sep 07 12:37:14 processor : 0 Sep 07 12:37:14 BogoMIPS : 2013.49 Sep 07 12:37:14 processor : 1 Sep 07 12:37:14 BogoMIPS : 1963.08 Sep 07 12:37:15 ha Sep 07 12:37:33 ogra@panda-ES2:~$ free Sep 07 12:37:33 total used free shared buffers cached Sep 07 12:37:33 Mem: 940580 497100 443480 0 18144 261452 Sep 07 12:37:45 soo nice :) Sep 07 12:52:18 * persia wonders if anyone feels like finding a way to drop u-boot and just boot into a kernel... Sep 07 12:52:49 10s to live environment would be good enough to justify turning computers off once in a while. Sep 07 12:53:23 persia: it is 10 _minutes_ Sep 07 12:54:52 "the boot is impressive, after u-boot its only taking 10sec" Sep 07 12:55:08 600s was for all of running oem-config Sep 07 12:55:21 right Sep 07 12:55:41 10 to ... say ... 15 mins is the whole install Sep 07 12:56:06 persia: aah, right. I was talking about the install, you were talking about boot Sep 07 12:56:07 booting is 10 sec as soon as the kernel is uncompressed Sep 07 12:56:34 Right, so we clearly have to not use Uboot: it makes it boot 150% slower Sep 07 12:56:48 heh Sep 07 12:57:04 kexecboot, etc... Sep 07 12:57:58 I'm guessing most of the 10s in u-boot is taken up loading uImage and uInitrd into memory Sep 07 14:04:00 amitk, and uncompressing them Sep 07 15:20:11 ogra: Hi, I just downloaded http://cdimage.ubuntu.com/ubuntu-server/ports/daily/current/maverick-server-armel+omap4.img and I think there is only one partition in there. Am I mistaking? Sep 07 15:20:42 server? Sep 07 15:21:07 i doubt server will work at all Sep 07 15:21:42 ogra: :) That must be why. I will rather look only at netbook. Thanks for pointing that. Sep 07 15:21:51 thats likely a debian-installer image, that indeed only has a single partition with a copy of the archive Sep 07 15:22:18 it will work, but wont be bootable without manual tinkering Sep 07 15:22:40 (after install i mean) Sep 07 15:24:12 ogra: Makes sense. I'll focus on http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/current/maverick-preinstalled-netbook-armel+omap4.img.gz Sep 07 15:24:20 yeah Sep 07 15:25:34 vstehle, that wont work either though, while the bootloader is es2, the es2 kernel isnt on the image yet due to a mistmatch of the linux-meta package Sep 07 15:25:58 i hope the kernel team will have that sorted until tomorrow Sep 07 15:26:20 (i know apw was looking into it, but dont know where it stands atm) Sep 07 15:26:48 ogra_cmpc, i think i have it sorted, last test build is going now Sep 07 15:26:56 perfect ! Sep 07 15:57:20 ogra_cmpc: No worry, I'll use another kernel to begin with. d/l is on-going... Sep 07 15:57:48 k Sep 07 15:58:16 make sure to keep the .img.gz file around that way you can use zsync next time Sep 07 15:58:39 speeds up significantly Sep 07 16:06:29 ogra_cmpc: Linaro meeting. Sep 07 19:16:29 does anyone know how can i read the cpsr register`? Sep 07 21:38:16 rsalveti: hi Sep 07 21:38:50 ndec, its a public holiday in .br Sep 07 21:39:07 ogra_cmpc: ah... didn't know ;-) Sep 07 21:39:22 ogra_cmpc: well it's close to midnight in germany though ... Sep 07 21:39:39 * ogra_cmpc neither until someone told me when i wondered where ricardo was all day :) Sep 07 21:39:58 yeah, just finished my last meeting Sep 07 22:44:24 anyone have a working rootfs that is functional on pandabaord? Sep 07 22:45:02 Which board version? Sep 07 22:45:18 8 layer Sep 07 22:45:31 ES2? Not yet. Sep 07 22:45:34 * orbarron tried using a blaze version but getting errors Sep 07 22:45:52 Need new kernel, uboot, and xloader. Sep 07 22:46:14 We should have it in the next daily image build. Sep 07 22:46:22 great... Sep 07 22:47:27 for some reason a am see the following error: http://pastebin.com/9bVZkhJx Sep 07 22:49:42 Apparently, not all of the hdmi patches have made it into the kernel. Not sure what is left (I've been out on holiday). **** ENDING LOGGING AT Wed Sep 08 02:59:59 2010