**** BEGIN LOGGING AT Mon Apr 07 02:59:59 2014
Apr 07 07:07:13 <qt-x> hello
Apr 07 12:37:38 <pompomJuice> Hi. Can anyone please point out to me how to upgrade your kernel to a new one using opkg?
Apr 07 12:38:00 <pompomJuice> I have the latest oe angstrom distro, built and running on my ARM device
Apr 07 12:38:23 <pompomJuice> my problem is that I make changes to the kernel, but I cant install the new kernel using opkg
Apr 07 12:38:27 <pompomJuice> I have to build a new image
Apr 07 12:38:31 <XorA> pompomJuice: opkg update; opkg updrade
Apr 07 12:38:38 <pompomJuice> it does not seem to work
Apr 07 12:38:43 <pompomJuice> when I do a 
Apr 07 12:38:55 <pompomJuice> opkg install --force-reinstall kernel-vmlinux
Apr 07 12:38:58 <pompomJuice> that is the one
Apr 07 12:39:00 <pompomJuice> the other one
Apr 07 12:39:03 <bluelightning> pompomJuice: is the version (PV or PR) changing?
Apr 07 12:39:08 <pompomJuice> yes
Apr 07 12:39:12 <pompomJuice> every time I make a change to the kernel
Apr 07 12:39:32 <pompomJuice> I change the MACHINE_KERNEL_PR = "r50" in my machine
Apr 07 12:39:43 <pompomJuice> I bumpt the number
Apr 07 12:40:02 <pompomJuice> there are two packages that get generated:
Apr 07 12:40:37 <pompomJuice> Package: kernel-image-2.6.37 ->  kernel-image-2.6.37_2.6.37-psp04.04.00.01-r0.14_impedo_duo.ipk
Apr 07 12:40:46 <pompomJuice> that package contains the uImage
Apr 07 12:41:05 <pompomJuice> kernel-vmlinux contains the bigger one
Apr 07 12:41:11 <pompomJuice> for installing both has no effect
Apr 07 12:41:15 <pompomJuice> the files change
Apr 07 12:41:21 <pompomJuice> but the kernel does not
Apr 07 12:41:34 <pompomJuice> it is like uBoot finds its image some other place
Apr 07 12:42:00 <XorA> off a smaller boot partition?
Apr 07 12:42:03 <pompomJuice> yea
Apr 07 12:42:06 <pompomJuice> a dos partition
Apr 07 12:42:10 <XorA> mount that copy the files
Apr 07 12:42:19 <pompomJuice> if I do it manually it also does not work
Apr 07 12:42:23 <pompomJuice> TFTP boot works
Apr 07 12:42:34 <pompomJuice> then I can see the creation date on the kernel change
Apr 07 12:42:44 <XorA> is this BBB, is it using the correct device?
Apr 07 12:42:49 <pompomJuice> but I eed the modules with the kernel to come together
Apr 07 12:43:04 <pompomJuice> BBB?
Apr 07 12:43:38 <XorA> Beaglebone Black, which I guess its not with a kernel as old as 2.6.37
Apr 07 12:43:50 <pompomJuice> yea
Apr 07 12:43:55 <pompomJuice> I have a strange setup
Apr 07 12:44:05 <XorA> should put brain in gear before applying accelerator
Apr 07 12:44:08 <pompomJuice> Basically we have a ARM device that records stuff
Apr 07 12:44:33 <pompomJuice> to make this work we decided to install angstrom, but replace the kernel with a rowboat kernel so that we can run android ontop of it
Apr 07 12:44:43 <pompomJuice> this was 4 years ago
Apr 07 12:44:49 <pompomJuice> hence the old kernel
Apr 07 12:44:59 <XorA> you dont work for shotspotter?
Apr 07 12:45:05 <pompomJuice> no
Apr 07 12:45:16 <pompomJuice> ctlab
Apr 07 12:45:24 <pompomJuice> I desperately need help
Apr 07 12:45:33 <pompomJuice> I kind of adopted this mess 6 months ago
Apr 07 12:45:37 <pompomJuice> I have learned a lot
Apr 07 12:45:40 <pompomJuice> but things are going wrong
Apr 07 12:45:48 <XorA> well your u-boot environment should show where kernel comes from, then find that from within linux
Apr 07 12:45:49 <pompomJuice> I broke the image and I cant get it to work
Apr 07 12:46:18 <XorA> fatload mmc X:X 0x????????? uImage kind of lines
Apr 07 12:47:04 <pompomJuice> let me check
Apr 07 12:47:11 <pompomJuice> but XorA
Apr 07 12:47:23 <pompomJuice> lets start at what is causing me to rebuild the kernel in the first place
Apr 07 12:47:35 <pompomJuice> do you know about the linaro 4.8 ARM -mno-unaligned-access issue?
Apr 07 12:47:48 <XorA> I think I saw you on #linaro channel
Apr 07 12:47:57 <pompomJuice> when I moved from yocto 1.4 to yocto 1.5 my kernel crashed
Apr 07 12:48:03 <pompomJuice> after enabling earlypringk
Apr 07 12:48:07 <pompomJuice> earlyprintk
Apr 07 12:48:16 <pompomJuice> I saw that it was complainging about unaligned access
Apr 07 12:48:42 <pompomJuice> I fixed this with a patch in the kernel Makefile +KBUILD_CFLAGS   += -mno-unaligned-access
Apr 07 12:48:46 <XorA> ah right, it might be expecting u-boot to have set that up
Apr 07 12:48:57 <pompomJuice> exactly!
Apr 07 12:49:08 <pompomJuice> u-Boot also has that no-unaligned-access flag
Apr 07 12:49:24 <pompomJuice> but the kernel used to work without it
Apr 07 12:49:25 <XorA> earlyprintk is in the weird bit between bootloader and kernel
Apr 07 12:49:34 <pompomJuice> but with linaro containing gcc 4.8 something changed
Apr 07 12:49:50 <XorA> I think armv7 is assumed unaligned works as default on armv7
Apr 07 12:49:59 <pompomJuice> yea!
Apr 07 12:50:13 <pompomJuice> I don't know aything about these ARM chips
Apr 07 12:50:22 <pompomJuice> I have a Ti chip
Apr 07 12:50:30 <pompomJuice> and it is full of shit
Apr 07 12:50:53 <pompomJuice> don't understand why it's failing with the alignment errors if it should be able to handle it
Apr 07 12:51:02 <pompomJuice> I was hoping someone could explain that to me
Apr 07 12:51:24 <pompomJuice> because after setting the kernel to no-unaligned-access the rest of the system does not work
Apr 07 12:51:26 <XorA> pompomJuice: you need to set a flag and also the MMU needs to be running for unaligned to work AFAIK
Apr 07 12:51:35 <pompomJuice> in particular, my TI kernel modules
Apr 07 12:51:49 <XorA> you can just fix the file it oops in rather than whole kernel
Apr 07 12:52:05 <pompomJuice> makes sense
Apr 07 12:52:11 <pompomJuice> the faulre was here:
Apr 07 12:52:35 <pompomJuice> Virtual kernel memory layout:
Apr 07 12:52:36 <pompomJuice>     vector  : 0xffff0000 - 0xffff1000   (   4 kB)
Apr 07 12:52:36 <pompomJuice>     fixmap  : 0xfff00000 - 0xfffe0000   ( 896 kB)
Apr 07 12:52:36 <pompomJuice>     DMA     : 0xffc00000 - 0xffe00000   (   2 MB)
Apr 07 12:52:36 <pompomJuice>     vmalloc : 0xd8800000 - 0xf8000000   ( 504 MB)
Apr 07 12:52:36 <pompomJuice>     lowmem  : 0xc0000000 - 0xd8000000   ( 384 MB)
Apr 07 12:52:36 <pompomJuice>     modules : 0xbf000000 - 0xc0000000   (  16 MB)
Apr 07 12:52:37 <pompomJuice>       .init : 0xc0008000 - 0xc0035000   ( 180 kB)
Apr 07 12:52:37 <pompomJuice>       .text : 0xc0035000 - 0xc04d9000   (4752 kB)
Apr 07 12:52:38 <pompomJuice>       .data : 0xc04da000 - 0xc051b280   ( 261 kB)
Apr 07 12:52:39 <pompomJuice> SLUB: Genslabs=11, HWalign=64, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
Apr 07 12:52:49 <pompomJuice> when the boot spit out that part it failed with unaligned issues
Apr 07 12:53:08 <XorA> sounds like your better off just sticking with old toolchain though
Apr 07 12:53:15 <pompomJuice> yea
Apr 07 12:53:17 <XorA> those old TI kernels are very picky about toolchain
Apr 07 12:53:23 <pompomJuice> everything worked perfectly in yocto 1.4
Apr 07 12:53:35 <XorA> TI internally only ever use code sourcery
Apr 07 12:53:43 <pompomJuice> 1.5 I just cannot get the god damn TI SGX530 android kermel module driver to work
Apr 07 12:53:46 <pompomJuice> it just wont work
Apr 07 12:53:54 <pompomJuice> amagad
Apr 07 12:54:07 <pompomJuice> and I am on GNU linarop
Apr 07 12:59:14 <XorA> if the toolchain you used to use is still in meta-linaro you can just override the toolchain and roll back
Apr 07 13:05:08 <pompomJuice> yea
Apr 07 13:05:12 <pompomJuice> maybe I should do that
Apr 07 13:07:22 <XorA> sounds like you got you solution on #linaro :-)
Apr 07 13:13:35 <pompomJuice> yea
Apr 07 13:13:54 <pompomJuice> that gives me enough stuff to try more options
Apr 07 13:16:07 <pompomJuice> still don't understand why I cannot hotswap my kernel
Apr 07 13:16:24 <pompomJuice> I remember 6 months ago I had to do some kernel work also
Apr 07 13:16:38 <pompomJuice> then I just manual copied the uImage over to that fat partition
Apr 07 13:16:44 <pompomJuice> just would not work
Apr 07 13:16:54 <pompomJuice> it somehow finds the old kernel
Apr 07 13:17:13 <pompomJuice> inside linux my /boot is mounted on that fat32 partition so it is the correct place
Apr 07 13:17:16 <pompomJuice> I am stumped
Apr 07 13:18:12 <XorA> are you sure? double check, and if its not check that it doesn't need fscked
Apr 07 13:18:30 <XorA> in fact fscking it might be an idea anyway
Apr 07 13:20:49 <pompomJuice> ok...
Apr 07 13:20:55 <pompomJuice> I will try to make sure
Apr 07 15:47:41 <pompomJuice> Anyone here know how I can override default tjoons for certain recipies in angstrom?
Apr 07 15:47:52 <pompomJuice> I want to disable hard float for one of the packages
Apr 07 22:25:27 <cptG> hello! I'm trying to download then angstrom toolchain but I can't access any URLs that seem useful to me - http://www.angstrom-distribution.org/toolchains/ http://www.angstrom-distribution.org/building-angstrom/  - both don't work.
Apr 07 22:25:45 <cptG> Are there new URLs and the links are outdated or is the site broken?
Apr 07 22:29:47 <XorA> cptG: we lost out hosting, we have not recovered the site yet
Apr 07 22:30:11 <XorA> cptG: if you go to the base a-d.org site it will lead you to github where the setup-scripts are
Apr 07 22:30:58 <cptG> XorA: oh that's too bad.
Apr 07 22:31:11 <cptG> XorA: oh I see. Thanks for the hint!
**** ENDING LOGGING AT Tue Apr 08 02:59:58 2014