**** BEGIN LOGGING AT Mon Apr 09 02:59:58 2012
Apr 09 06:17:06 <alabd> Good day all ,  wana ssh to shr  , says ssh: connect to host 192.168.7.2 port 22: No route to host
Apr 09 06:17:23 <alabd> seems usb connection is not stablished 
Apr 09 06:17:33 <alabd> why ?cable is healthy 
Apr 09 07:49:14 <pabs3> alabd: check ifconfig on both sides
Apr 09 07:57:30 <pabs3> alabd: but it sounds like your computer doesn't have one of these done: the driver loaded, an interface created or a route set on that interface
Apr 09 08:27:33 <alabd> pabs3: thanks , but it's not internet sharing , does it need a route set on that interface ?
Apr 09 08:42:16 <lindi-> alabd: you always need a route if you send packets to other systems
Apr 09 08:42:42 <alabd> ping should be ok without route or not ?
Apr 09 08:44:20 <lindi-> alabd: I thought you wanted to use ssh?
Apr 09 08:44:35 <alabd> yes but ping also is not ok 
Apr 09 09:44:41 <Weiss> pabs3: d'oh... fixed
Apr 09 09:50:53 <pabs3> Weiss: thanks
Apr 09 10:07:37 <CIA-45> SHR: 03lukasmaerdian 07meta-smartphone * rc7ee3fc6db1e 10/meta-openmoko/recipes-kernel/linux/linux-gta04_git.bb: meta-openmoko: linux-gta04: bump SRCREV to 3.2.14-gta04
Apr 09 11:37:27 <DocScrutinizer> damn, agni.om.org completely down again, it seems
Apr 09 11:40:05 <DocScrutinizer> last night I learnt weird stories about what's going on with OM infra
Apr 09 11:40:16 <DocScrutinizer> though with bright perspective
Apr 09 11:42:46 <DocScrutinizer> if you're interested in details, see http://en.qi-hardware.com/irclogs/qi-hardware_2012-04-09.log.html
Apr 09 13:27:01 <PaulFertser> DocScrutinizer: new sickstories everyday
Apr 09 13:27:05 <PaulFertser> Meh
Apr 09 13:27:09 <PaulFertser> DocScrutinizer: thanks for caring!
Apr 09 13:28:48 <DocScrutinizer> PaulFertser: yw
Apr 09 15:38:46 <alabd> How to trun off sleeping on SHR in terminal ?
Apr 09 15:44:56 <alabd> have mounted one of my partition on uSD , but can't umount it 
Apr 09 15:44:57 <alabd> bash-4.2# umount -f /dev/mmcblk0p2
Apr 09 15:44:58 <alabd> umount2: Device or resource busy
Apr 09 15:44:58 <alabd> umount: /home/root/debian: device is busy.
Apr 09 15:44:58 <alabd>         (In some cases useful info about processes that use
Apr 09 15:44:58 <alabd>          the device is found by lsof(8) or fuser(1))
Apr 09 15:44:59 <alabd> umount2: Device or resource busy
Apr 09 15:45:08 <alabd> while it not busy ,
Apr 09 15:45:15 <alabd> it's not busy ..
Apr 09 15:54:16 <alabd> wana say debian installer.sh to install debian on second partition on uSD , but in default installer wana install it on whole uSD space , how to say installer use second partition for installing debian ?
Apr 09 16:08:58 <Alex[sp3dev]> do an lsof /home/root/debian to see who uses it. anyway, you can do "sync; umount -l /dev/mmcblk0p2"
Apr 09 18:58:52 <GNUtoo> Alex[sp3dev], hi
Apr 09 18:59:19 <Alex[sp3dev]> GNUtoo: hi
Apr 09 19:00:00 <GNUtoo> Alex[sp3dev], I've some bad news on samsung devices:
Apr 09 19:00:28 <GNUtoo> long time ago me and paulk-laptop looked at the datasheet and schematics
Apr 09 19:00:36 <GNUtoo> and it seemed ok
Apr 09 19:00:57 <GNUtoo> but recently while looking at the modem kernel code I saw that it shares a ram bank between the modem and the CPU
Apr 09 19:01:44 <GNUtoo> so now I'm trying to remove the usage of that ram bank by the kernel
Apr 09 19:01:47 <GNUtoo> it's only 80M
Apr 09 19:01:56 <GNUtoo> so we should be able to do it without
Apr 09 19:02:07 <GNUtoo> I removed it in the kernel but then I've that error:
Apr 09 19:02:39 <GNUtoo> http://www.pastie.org/private/sswyvbltpwhzkl0picatyw
Apr 09 19:02:51 <GNUtoo> so I wonder what's the best way and more important how it works
Apr 09 19:03:08 <GNUtoo> I guess the bootloader expects the kernel at 0x30000000 right?
Apr 09 19:03:30 <GNUtoo> or is it configurable by the image(mkbootimg)
Apr 09 19:03:44 <GNUtoo> because basically here's the ram mapping
Apr 09 19:04:18 <GNUtoo> 0x30000000 is the bank to remove, there is only 80M used generally by the kernel
Apr 09 19:04:29 <GNUtoo> then at 0x40000000 there is 256M
Apr 09 19:04:48 <GNUtoo> and there is 128M at 0x50000000(righ after the 0x40000000)
Apr 09 19:05:10 <GNUtoo> so I removed the references to 0x30000000
Apr 09 19:05:17 <GNUtoo> and I had the error that you see 
Apr 09 19:05:32 <Alex[sp3dev]> why do you want to remove that?
Apr 09 19:06:11 <GNUtoo> I don't feel confiortable with the modem(running non-free code) having potentially access to 80M of the kernel memory
Apr 09 19:06:17 <GNUtoo> *confortable
Apr 09 19:06:26 <GNUtoo> and I think paulk-laptop agrees with me
Apr 09 19:07:25 <GNUtoo> we want to fix this issue instead of telling, once more to the people that invested in that samsung platform(developers and users) to get a gta04 instead
Apr 09 19:08:09 <GNUtoo> so my question is how booting between the kernel and the samsung bootloader works
Apr 09 19:08:14 <Alex[sp3dev]> if it can access that memory bank, i do not see why it cannot access the rest of memory. anyway, could you show your code to disable the bank? if the bootloader is loading initrd at a fixed address, one solution may be to use initramfs instead of initrd. this way, it will be part of kernel and get relocated together with kernel
Apr 09 19:08:30 <Alex[sp3dev]> i don't know what samsung bootloader does, unfortunately
Apr 09 19:08:41 <GNUtoo> ok
Apr 09 19:08:46 <GNUtoo> I'll show my code
Apr 09 19:09:11 <GNUtoo> http://www.pastie.org/private/a6wrxqamwc1x9gnvfl0w
Apr 09 19:09:39 <GNUtoo> basically I think the rest of the memory is on a different memory chip
Apr 09 19:09:42 <GNUtoo> mapped differently
Apr 09 19:09:47 <GNUtoo> or am I wrong?
Apr 09 19:10:02 <GNUtoo> basically I thought it was like that:
Apr 09 19:10:18 <GNUtoo> modem<-physical ram chip to be shared->AP
Apr 09 19:10:24 <GNUtoo> and then:
Apr 09 19:10:41 <GNUtoo> [other chip->AP
Apr 09 19:10:44 <GNUtoo> [other chip->AP
Apr 09 19:11:00 <GNUtoo> but I've no precise schematics
Apr 09 19:11:04 <GNUtoo> and you could be right
Apr 09 19:11:24 <Alex[sp3dev]> i actually think both arm cores have access to any memory location like on msm
Apr 09 19:11:31 <GNUtoo> ha
Apr 09 19:11:36 <GNUtoo> why do you think that
Apr 09 19:11:47 <GNUtoo> do you have hints that makes you think that?
Apr 09 19:12:05 <GNUtoo> paulk-laptop, ^^^
Apr 09 19:12:15 <Alex[sp3dev]> well, i'll try to look up on modern soc architecture and tell you more later. now try hardcoding memory address via mkbootimg or using initramfs
Apr 09 19:12:33 <GNUtoo> ok
Apr 09 19:13:14 <GNUtoo> I think the initrd is easy to relocate
Apr 09 19:13:18 <GNUtoo> I guess it worked
Apr 09 19:13:28 <GNUtoo> but my problem is the kernel itself it seems
Apr 09 19:14:33 <Alex[sp3dev]> GNUtoo: as i understand, 0x31000000 is not in kernel's mapped memory. therefore, kernel cannot find its virtual address and use it.
Apr 09 19:14:46 <GNUtoo> yes I know
Apr 09 19:14:57 <GNUtoo> I fixed it on my desktop but forgott to import the fix
Apr 09 19:15:16 <GNUtoo> rest the <2>kernel BUG at /home/gnutoo/embedded/android/replicant-2.3/kernel-crespo/mm/bootmem.c:377! issue
Apr 09 19:15:27 <GNUtoo> which remains even if I fix that
Apr 09 19:22:41 <GNUtoo> also some stuff like exceptions vectors etc...are in a fixed place I guess
Apr 09 19:22:44 <GNUtoo> I must look that up
Apr 09 19:23:07 <Alex[sp3dev]> well, no. on armv6 vectors were fixed at 0x0. armv7 can set them at any address
Apr 09 19:25:10 <GNUtoo> ok nice
Apr 09 19:25:30 <Alex[sp3dev]> we need a description of the memory controller and a block diagram. on msm, for example, there were security domains and you could specify which core can access what.
Apr 09 19:25:47 <GNUtoo> ok
Apr 09 19:27:26 <Alex[sp3dev]> stupid firefox is sooo slow and so cpu hungry.. but chromium broke down after update yesterday
Apr 09 19:29:49 <GNUtoo> hmmm you have browsers problems too
Apr 09 19:31:05 <Alex[sp3dev]> well, i have installed chromium 17 from debian yesterday and most google sites and bitbucket won't open
Apr 09 19:31:29 <GNUtoo> ah ok
Apr 09 19:32:36 <GNUtoo> if I relocate all at 0x40000000 I get that: Uncompressing Linux... done, booting the kernel.
Apr 09 19:32:39 <GNUtoo> and then nothing more
Apr 09 19:34:41 <Alex[sp3dev]> how do you relocate? via mkbootimg? can you try using initramfs and no initrd?
Apr 09 19:35:01 <GNUtoo> trough mkbootimg and trough the platform 
Apr 09 19:35:11 <GNUtoo> -#define PHYS_OFFSET            UL(0x30000000)
Apr 09 19:35:11 <GNUtoo> +#define PHYS_OFFSET            UL(0x40000000)
Apr 09 19:40:32 <GNUtoo> ah I forgott some stuff
Apr 09 19:42:10 <Alex[sp3dev]> i think you're going to run into trouble if the bootloader loads kernel at 0x30 somewhere and phys_offset is after this address
Apr 09 19:43:24 <GNUtoo> there is still CMDLINE_FORCE tough
Apr 09 19:44:34 <Alex[sp3dev]> basically, what happens now when you have phys_offset at 0x30000000 and first memory bank at 0x40000000 is that kernel doesn't know what kind of memory the area between 0x30 and 0x40 is and does not know how to read it. either you create an io device for this region, or do not use initrd at all. or check if sgs bootloader understands addresses in mkbootimg
Apr 09 19:45:26 <GNUtoo> initrd is at 0x41000000
Apr 09 19:47:51 <Alex[sp3dev]> according to your log, at 0x31000000
Apr 09 19:48:52 <GNUtoo> yes according to the old log
Apr 09 19:48:58 <GNUtoo> I think I lost serial
Apr 09 19:50:16 <GNUtoo> one pin broke
Apr 09 21:06:05 <alabd> Good night ,  wana say debian installer.sh to install debian on second partition on uSD , but in default installer wana install it on whole uSD space , how to say installer use second partition for installing debian ?
**** ENDING LOGGING AT Tue Apr 10 03:00:00 2012