**** BEGIN LOGGING AT Mon Feb 22 02:59:58 2010 Feb 22 08:04:58 Morning. Feb 22 09:50:23 morning Feb 22 11:35:59 do any of the AOF Cupcake images include the recently reworked sound configs to try for better sound in Cupcake? Possibly a daily that we could try & report back on? Feb 22 12:41:54 hi is there any workaround for sound level while calling ? it is so low I cannot hear half the words spoken (I am using master though it looks like cupcake is affected too) Feb 22 12:43:06 also the suspend seems to not be low enough ... it last 5 hours without activities compared to at minimum 2 days for shr Feb 22 12:50:06 the suspend is fine for me Feb 22 12:50:24 sound depends on the specific FR Feb 22 12:51:05 mvzaanen, you mean it last more then 48 hours without a call ? Feb 22 12:51:23 hmmm well, at least more than 5 hours Feb 22 12:51:24 :D Feb 22 12:53:13 I have a friend with a samsung galaxy and it last less than a day too (with manufacturer provided image) though I am comparing against SHR (where it could last 4 days) not other android (anyway suspend time should be specific to phone and battery not OS) Feb 22 12:53:52 I believe android could ask the same level of suspend as SHR , I am wrong ? Feb 22 12:54:17 probably Feb 22 12:54:19 I don't know... Feb 22 12:54:37 to be honest, I haven't really compared SHR to AoF Feb 22 12:55:08 ok it was just to know if anybody compared or tweaked this area ... I ll take a look at the suspend asked for both of those Feb 22 12:57:47 about the sound level while calling , no workaround currently known ? I found talks about softkeyboards but was unable to trigger any volume keys (though sound level is fine if I enable speakers but it is not really practical in crowded places) Feb 22 12:58:10 people on the mailinglist are talking about different alsa config files Feb 22 12:58:35 ah, I've never even tried the speakers... that's an interesting option Feb 22 12:58:37 ok . that s an easy one so Feb 22 13:52:40 RussH, what did you mean? Feb 22 13:53:18 which sound rework? Feb 22 14:05:26 Serdar: the echo suppression patch and/or the new alsa configurations? Feb 22 14:41:38 hi all Feb 22 14:41:51 hey mattboll how are you doing? Feb 22 17:29:17 jancona, is system.img the same like out/target/product/freerunner/system? Feb 22 17:31:08 Serdar: yes Feb 22 17:46:50 hi pbaxter, others Feb 22 17:48:28 make -j 4 TARGET_PRODUCT=freerunner on cupcake give me some errors, trying to find why… Feb 22 17:54:00 erf, didn't export cc and cxx, seems better Feb 22 18:53:55 mattboll, have you solved the problem? Feb 22 19:08:19 mattboll, if you don't know what you need, go here: http://source.android.com/download Feb 22 19:14:25 pbaxter: it's compiled, after more than 1hour -_-' need to go lunch now, and I'll see after that what is goin on Feb 22 19:15:40 mattboll, so have you succeded? Feb 22 19:24:07 pbaxter: to compile yes Feb 22 19:25:35 but I don't know yet what it did unless compiling everything :/ Feb 22 19:26:50 e.g where are boot/ kernel.img md5sums qi.img ramdisk.img system.img uImage uImage.bin userdata.img Feb 22 19:27:03 or if they are done manually, then from where… Feb 22 19:40:20 I've got almost 2GB in out/target/product/freerunner Feb 22 19:41:11 found ramdisk.img ! yes ! :p Feb 22 19:42:04 found kernel… Feb 22 19:46:37 is there only a "demerden sie sich" process or is there some doc ? Feb 22 20:00:57 mattboll: From out/target/product/freerunner/ you just need kernel.img and system.img Feb 22 20:01:21 You can copy them to the installer SD or use dfu-util to flash them Feb 22 20:30:01 jancona: ok, so how about the aof "specific" install ? Feb 22 20:34:58 mattboll, but have you changed something now? Feb 22 20:39:55 pbaxter: not yet Feb 22 20:40:25 still trying to understand where are stuff… there is 3GB of sources ! Feb 22 20:44:52 mattboll, ok;) Feb 22 20:51:06 mattboll: what about the aof "specific" install ? Feb 22 21:00:22 jancona: where is the src >_< Feb 22 21:01:28 I'm completely new with android and also with freerunner :/ Feb 22 21:01:44 Try repo init -u git://gitorious.org/android-on-freerunner/freerunner_platform_manifest.git -b installer Feb 22 21:02:07 It's not currently buildable, but I hope to have it working soon Feb 22 21:02:22 jancona, but what are you and serdar willing to change in the installer? Feb 22 21:02:40 But you don't need the installer to install the image that you build Feb 22 21:03:32 pbaxter: I'm not sure. Why do you ask? Feb 22 21:04:03 jancona, serdar was always talking about chaning the installer in these last days Feb 22 21:04:07 jancona, serdar was always talking about chaning the installer in these last days Feb 22 21:04:10 so i'm asking Feb 22 21:04:53 yes, and asked me if I could do some stuff Feb 22 21:05:08 You should ask Serdar--I don't know what he wants to change Feb 22 21:05:12 install only images that are needed Feb 22 21:05:29 mattboll, thanks :) Feb 22 21:05:58 and do not flash qi each time Feb 22 21:06:04 I think that's a really simple think to change if we can get the installer code to build properly Feb 22 21:06:56 User data is on a separate partition, so if we don't flash that every time upgrades should trash user data Feb 22 21:07:44 With Qi install, we could probably have it detect that Qi exists and no reflash it Feb 22 21:08:48 why do you want to trash user data ? Feb 22 21:09:35 I want to NOT trash it Feb 22 21:09:50 The current installer (or flashing) does Feb 22 21:10:04 oh ok, that's better ^^ Feb 22 21:10:29 If I could type that would be better :-( Feb 22 21:12:08 if the installer is working fine, what else is needed to have cupcake 1.0-release ? Feb 22 21:20:20 Whatever other bugs we think should be fixed. Feb 22 21:20:44 which means ? Feb 22 21:21:18 I think GPRS doesn't work after wifi has been in use. Feb 22 21:21:50 Sound continues to be an issue, but I'm not sure there's one setting that will work for everyone Feb 22 21:25:02 not so much, it will be done before wednesday :D Feb 22 21:27:12 What's on Wednesday? Feb 22 21:28:15 nothing… unless the first release Feb 22 21:54:50 re Feb 22 21:56:38 jancona, what do you think of this: the installer should only install *.img files to the right place. Feb 22 21:57:02 if we put qi.img along with the installer it should flash it to u-boot. Feb 22 21:57:14 if we put kernel.img -> kernel Feb 22 21:57:28 system.img -> rootfs Feb 22 21:57:42 nothing -> message Feb 22 21:57:51 all together -> right place Feb 22 21:58:43 What do you mean by "the right place"? Feb 22 21:58:56 Doesn't it do this now? Feb 22 21:58:58 kernel, u-boot, rootfs... Feb 22 21:59:01 on nand Feb 22 21:59:22 yes but if flashes qi everytime and don't work with other distributions. Feb 22 21:59:45 and if we remove qi it does not work Feb 22 22:00:06 Android has a different partition layout than other distros Feb 22 22:01:19 We could probably only flash Qi if it's not already there Feb 22 22:01:20 you mean first fat and second ext? Feb 22 22:01:29 No, in NAND Feb 22 22:01:45 hm Feb 22 22:02:06 There is no "rootfs" partition with Android, there's system and userdata Feb 22 22:02:30 is the layout of any other distribution the same then? Feb 22 22:02:38 debian, gentoo, shr... Feb 22 22:03:14 but userdata is empty iirc. Feb 22 22:03:37 No, the standard FR partition layout has kernel and rootfs, but Andoid wants user data separate from system Feb 22 22:04:12 That's why Michael's (and the early Koolu) releases put userdata on SD Feb 22 22:04:48 hm Feb 22 22:06:23 then we should change the installer so we can use any kind of qi. or it should not reflash qi if it's flashed Feb 22 22:06:35 allready flashed Feb 22 22:07:14 That's probably do-able, but I'm not sure if it's difficult Feb 22 22:07:43 no idea, but we can remove qi from the installer, that should be simple Feb 22 22:08:48 but my installer isn't building properly Feb 22 22:12:02 case "$name" in Feb 22 22:12:02 u-boot|qi) Feb 22 22:12:02 flash_eraseall /dev/$device Feb 22 22:12:03 nandwrite -a -p /dev/$device /mnt/qi.img Feb 22 22:12:09 yes Feb 22 22:12:31 - u-boot|qi) Feb 22 22:12:32 - flash_eraseall /dev/$device Feb 22 22:12:32 - nandwrite -a -p /dev/$device /mnt/qi.img Feb 22 22:12:32 - ;; Feb 22 22:12:32 +# u-boot|qi) Feb 22 22:12:32 +# flash_eraseall /dev/$device Feb 22 22:12:34 +# nandwrite -a -p /dev/$device /mnt/qi.img Feb 22 22:12:36 +# ;; Feb 22 22:12:49 while read device name ; do Feb 22 22:12:50 case "$name" in Feb 22 22:12:50 - qi|kernel) Feb 22 22:12:50 + kernel) Feb 22 22:12:50 +# qi|kernel) Feb 22 22:12:54 flash_eraseall /dev/$device Feb 22 22:12:56 nandwrite -a -p /dev/$device /mnt/$name.img Feb 22 22:12:57 I did that earlier Feb 22 22:13:05 but the installer isn't building properly Feb 22 22:13:20 because of mtd-utils and busybox Feb 22 22:13:32 more busybox than mtd-utils I guess Feb 22 22:15:07 jancona said the same thing a little bit earlier ^^ Feb 22 22:16:05 btw, I didn't understand everything on the differences between fr other distros and android Feb 22 22:17:09 on shr (for example), you need a fat32 partition for kernel and an ext3 for some stuff Feb 22 22:17:37 (rootfs) Feb 22 22:18:23 actually I understood Feb 22 22:18:30 just needed to write it >_< Feb 22 22:18:56 qi is the boot loader right ? Feb 22 22:19:41 yes Feb 22 22:22:01 I love shell scripting Feb 22 22:22:02 for n in . . . . . . . . . . . . . . . . . . . . ; do log ; done Feb 22 22:22:04 >_< Feb 22 22:23:21 question Feb 22 22:23:43 is the aof qi able to run shr for example ? Feb 22 22:24:47 mattboll, no Feb 22 22:24:50 I guess not. Feb 22 22:24:52 they were saying it above Feb 22 22:25:34 So, we need a "normal" qi to install android, and a "hacked" qi to launch android, right ? Feb 22 22:25:55 mattboll, yes Feb 22 22:26:01 mattboll, if you see mounted partitions Feb 22 22:26:01 rootfs / rootfs ro 0 0 Feb 22 22:26:02 tmpfs /dev tmpfs rw,mode=755 0 0 Feb 22 22:26:02 devpts /dev/pts devpts rw,mode=600 0 0 Feb 22 22:26:02 proc /proc proc rw 0 0 Feb 22 22:26:02 sysfs /sys sysfs rw 0 0 Feb 22 22:26:04 tmpfs /sqlite_stmt_journals tmpfs rw,size=4096k 0 0 Feb 22 22:26:05 /dev/block/mtdblock6 /system jffs2 ro 0 0 Feb 22 22:26:07 /dev/block/mmcblk0p1 /sdcard vfat rw,nosuid,nodev,fmask=0000,dmask=0000,allow_utime=0022,codepage=cp437,iocharset=iso8859-1 0 0 Feb 22 22:26:11 /dev/block/mtdblock7 /data yaffs2 rw,nosuid,nodev 0 0 Feb 22 22:26:14 mtdblock6 is a partition for system Feb 22 22:26:29 and mtdblock7 for data Feb 22 22:26:39 in the past we used the SDcard for data partition Feb 22 22:26:44 now NAND memory is parted Feb 22 22:28:03 but, if it is always a standard qi then an aof qi, we will have to flash qi at each install because it is _needed_ each time, isn't it ? Feb 22 22:29:00 mattboll, it's not needed if you have installed the aof qi Feb 22 22:32:03 ok, sorry, understood, I thought that it was needed to have a standard qi to launch the install, but "boot from sdcard" in fr does not use qi to do this… Feb 22 22:33:45 mattboll, it use NOR Feb 22 22:34:46 nor ? Feb 22 22:35:19 mattboll, pressing AUX + Power you are into NOR menu Feb 22 22:35:25 and you boot from SD from NOR menu Feb 22 22:35:30 ok Feb 22 22:35:32 mattboll, i'm going out to smoke then to bed Feb 22 22:35:51 i'm tired and tomorrow have to got up at 6 Feb 22 22:35:52 am Feb 22 22:35:53 i'm going out to smoke too but bed later :p Feb 22 22:36:01 it's 11.35 pm Feb 22 22:36:03 i'm tired Feb 22 22:36:17 erf, I'm on hollydays, and it's also 23:35 ^^ Feb 22 22:36:28 wow Feb 22 22:36:31 mattboll, on shr one only needs an ext2/3 partition if using qi Feb 22 22:36:31 i'm at school Feb 22 22:36:51 fat32 is only to make uboot happy Feb 22 22:37:08 mattboll, last thing: i think serdar wanted you to check if aof qi is installed before flashing it again Feb 22 22:37:10 bye Feb 22 22:37:57 prahal: ok, i didn't notice that Feb 22 22:38:25 I've learned a lot of stuff tonight :p Feb 22 22:39:58 btw I just learned one thing .. That it is expected that /data is on internal flash ... so is it expected that android report not seeing the sd ? (if I do mount I also have /data on nand and /sdcard on vfat Feb 22 22:40:48 so the system partition on sd that is told to be required seems unused and useless Feb 22 22:41:56 do you mean the user data that is told to be required seems unused ? Feb 22 22:41:58 prahal, it was a TAB fail from mattboll Feb 22 22:42:11 he meant pbaxter I guess Feb 22 22:42:30 Serdar: I don't think so :/ Feb 22 22:42:39 btw Serdar, while you're here Feb 22 22:42:47 line 85 on init Feb 22 22:43:02 doesn't it check that there is not the right qi ? Feb 22 22:43:42 sorry I have a patched init, can you give me the complete line Feb 22 22:43:59 if [ $n -ne 4 ] ; then Feb 22 22:44:51 my problem is that when it tests qi, it also installs the kernel :/ Feb 22 22:45:33 I don't think that this tests if qi is installed Feb 22 22:46:21 or maybe it does. Feb 22 22:46:46 the problem is, we need a different qi than others. Feb 22 22:48:01 are lines 40 and 4 the sme? Feb 22 22:48:03 same Feb 22 22:48:55 4 is export $PATH and 40 is log . . . . . . Feb 22 22:49:09 err 40 and 41 Feb 22 22:49:32 in this case, yes they are Feb 22 22:49:58 line 4 is export PATH instead of export $PATH :/ Feb 22 22:50:54 - for u in uImage uImage.bin boot/uImage-GTA02.bin ; do Feb 22 22:50:57 - if test -e /mnt/$u ; then Feb 22 22:51:00 - nandwrite -a -p /dev/$device /mnt/$u Feb 22 22:51:02 + for u in uImage uImage.bin boot/uImage-GTA02.bin ; do Feb 22 22:51:05 + if test -e /mnt/$u ; then Feb 22 22:51:08 + nandwrite -a -p /dev/$device /mnt/uImage Feb 22 22:51:24 in this case, we could need only one uImage to cp to the sdcard… Feb 22 22:52:11 you will write /mnt/uImage 2 times? Feb 22 22:52:14 only the $u changed, sorry, bad patch, need a cigarett before thinking :p Feb 22 22:52:29 smoking is bad. Feb 22 22:54:20 sorry I was completely wrong, I thought I spotted why we needed 3 times the same kernel Feb 22 23:03:46 mattboll: The installer reboots after writing each partition Feb 22 23:04:44 and uses NOR uboot for the first boot then qi Feb 22 23:06:09 jancona, so we don't have no way but useing this qi version right? Feb 22 23:06:21 I don't know Feb 22 23:06:33 Try a different one and see :-) Feb 22 23:07:07 jancona: is the reboot done in nandwrite ? Feb 22 23:07:14 I'm not sure why Qi needs to know about the partition table Feb 22 23:07:34 I'm not sure Feb 22 23:07:48 May be Feb 22 23:08:03 It seems like a good thing to do after rewriting a partition Feb 22 23:08:27 strange :/ Feb 22 23:09:11 nope I isn't Feb 22 23:09:18 there are two reboots Feb 22 23:09:26 first one is after log Qi installed Feb 22 23:09:51 I'm leaving now Feb 22 23:10:02 drive careful Feb 22 23:10:04 be back later, bye! Feb 22 23:10:33 bye Feb 22 23:10:54 I see the reboot -f Feb 22 23:13:10 so qi is customized for android only ... because of the partition behing rewritten when flashing ... this makes no sense to me (the point behing qi is that we ought not to bother about the partition scheme as it loops through the partitions to find the kernel Feb 22 23:20:35 I don't think it is the reason why we need a customized qi Feb 22 23:21:20 but if you're right about qi, I don't understand why a customized one is needed Feb 22 23:22:48 can you build busybox? Feb 22 23:23:15 if you tell me what and where is it, I could try Feb 22 23:23:26 in the installer branch Feb 22 23:23:50 how did you get the installer? Feb 22 23:24:29 with repo init -u git://gitorious.org/android-on-freerunner/freerunner_platform_manifest.git -b installer Feb 22 23:24:42 okay Feb 22 23:24:47 there is busybox Feb 22 23:25:02 found it Feb 22 23:25:18 I give you my patch Feb 22 23:26:57 big patch ? do you want my mail ? Feb 22 23:27:10 nope Feb 22 23:28:39 it does not compile right now Feb 22 23:29:06 http://serdar-dere.net/~serdar/installer2.diff Feb 22 23:45:04 with make I've got no more and no less stuff than without the patch… what should I look for ? Feb 22 23:45:20 Image arch/arm/boot/uImage is ready Feb 22 23:50:27 may be I should try it now… **** ENDING LOGGING AT Tue Feb 23 02:59:58 2010