I am trying to build kernel 3.5.7 on Yocto Dora. (using imx6qsabresd)
Everything builds without an error and I get a new uImage, as shown below. However, the Kernel hangs while starting up.
# I first clean before changing versions in order to clean up the 3.0.35 files...
bitbake -c cleansstate linux-imx
# change from 3.0.35 to 3.5.7 by adding the following line to your local.config
PREFERRED_VERSION_linux-imx_mx6 = "3.5.7"
# rebuild kernel
bitbake linux-imx
# rebuild image
bitbake core-image-minimal
=========================================
Output files...
tmp/deploy/images/imx6qsabresd/u-boot.imx -> u-boot-imx6qsabresd-v2013.10-r0.imx*
tmp/deploy/images/imx6qsabresd/u-boot-imx6qsabresd.imx -> u-boot-imx6qsabresd-v2013.10-r0.imx*
tmp/deploy/images/imx6qsabresd/u-boot-imx6qsabresd-v2013.10-r0.imx*
tmp/deploy/images/imx6qsabresd/uImage -> uImage--3.5.7-r0-imx6qsabresd-20131104212852.bin
tmp/deploy/images/imx6qsabresd/uImage--3.5.7-r0-imx6qsabresd-20131104212852.bin
tmp/deploy/images/imx6qsabresd/uImage--3.5.7-r0-imx6q-sabresd-20131104212852.dtb
tmp/deploy/images/imx6qsabresd/uImage-imx6qsabresd.bin -> uImage--3.5.7-r0-imx6qsabresd-20131104212852.bin
tmp/deploy/images/imx6qsabresd/uImage-imx6q-sabresd.dtb -> uImage--3.5.7-r0-imx6q-sabresd-20131104212852.dtb
=========================================
U-Boot 2013.10 (Nov 04 2013 - 09:49:56)
CPU: Freescale i.MX6Q rev1.2 at 792 MHz
Reset cause: POR
Board: MX6-SabreSD
DRAM: 1 GiB
MMC: FSL_SDHC: 0, FSL_SDHC: 1, FSL_SDHC: 2
*** Warning - bad CRC, using default environment
No panel detected: default to HDMI
Display: HDMI (1024x768)
In: serial
Out: serial
Err: serial
Net: FEC [PRIME]
Hit any key to stop autoboot: 0
mmc1 is current device
reading boot.scr
** Unable to read file boot.scr **
reading uImage
4686224 bytes read in 218 ms (20.5 MiB/s)
Booting from mmc ...
reading imx6q-sabresd.dtb
46128 bytes read in 20 ms (2.2 MiB/s)
## Booting kernel from Legacy Image at 12000000 ...
Image Name: Linux-3.5.7-1.0.0+g3285970
Image Type: ARM Linux Kernel Image (uncompressed)
Data Size: 4686160 Bytes = 4.5 MiB
Load Address: 10008000
Entry Point: 10008000
Verifying Checksum ... OK
## Flattened Device Tree blob at 18000000
Booting using the fdt blob at 0x18000000
Loading Kernel Image ... OK
Using Device Tree in place at 18000000, end 1800e42f
Starting kernel ...
(hang)
Hi Doug.
Did you manage to figure out what the problem is?
I have the same problem after I changed to the 3.10.9 kernel.
Thanks,
Niklas
I have not been able to get 3.10.9 to build because master does not build for me. I have not changed anything and I am tired of having to rebuild and compile from scratch. Lauren's response in the meta-freescale list worked for me and I was able to continue.
If you have not sent this issue into meta-freescale list, please do it.
I had the same issue with the boundary-devices build until I RTFM'd. I believe this(3.5.7) has all moved to dts/dtb, and you need to build the dtb and have it local, PLUS bootm needs a second address to locate the loaded dtb, and 2 more environment variables need to be set
1 2 3 | U-Boot > setenv fdt_high 0xffffffff U-Boot > setenv fdt_addr 0x13000000 U-Boot > saveenv |
Thanks John, I will check that out. In the mean time, here is a response from the [meta-freescale] group on the issue of the 3.5.7 kernel hanging.
They think we might be seeing issues with the new toolchain 4.8. There were some ARM errata workarounds in defconfig in 3.10 to fix issues with new gcc toolchain. Also the new eglibc has performance issues they are patching in the beta release. They suggest building on the 4.7 toolchain which they believe is the same one that 3.5.7 was tested against. Best to switch to 3.10 as that will go GA not 3.5.7 which was stopped to move to an LTS kernel. They think 3.10.9 is almost ready to be fully upstreamed into dora branch. So until it is merged, use the master branch.
Hi Doug, I gave you the response on the toolchain. Did it work going down to the 4.7 gcc version?
3.10.9 will be upstreamed into dora branch by next week. Only remaining upstreaming is in meta-fsl-demos for image recipes.
Yes , it did. Thanks.