I've recently pulled down imx_3.10.17_1.0.0_beta from git.freescale.com to try building it and I get this error message:
arch/arm/mach-imx/lpddr2_freq_imx6.S:365: Error: selected processor does not support ARM mode `dsb'
I'm using arm-2013.05-24-arm-none-linux-gnueabi-i686-pc-linux-gnu.tar.bz2 toolset,
and starting with imx_v6_v7_defconfig.
Anyone else seeing this?
Solved! Go to Solution.
The default config for 3.10.17 is the imx_v7_defconfig. Please try that one instead.
Regards,
Fabio Estevam
The default config for 3.10.17 is the imx_v7_defconfig. Please try that one instead.
Regards,
Fabio Estevam
How does that fix the bug in the Makefile though ? ;-)
Marek,
Yea, good question...
I was going to try the patch you sent, but then when Fabio's response that was an
easier first-try. It built clean, so I guess the different config caused the build to steer
around the snippet of code that caused my original error...
If I get some time, I'll dig deeper.
Regardless, thanks for replying.
Ed
Fabio,
I built this and I *almost* have it running on my hardware; based on some major hacks made to the device tree.
Can you tell me, is this being used much publicly yet or is this still bleeding-edge experimental for iMX6?
Ed
Try using ARMHF toolchain , that might just be it. I suspect your toolchain is ARMEL, right ?
Quite honestly, I've never heard the terms "ARMHF" and "ARMEL" till now.
Based on some googling, it appears that the primary difference is hard/soft floating point.
So 2 questions follow...
1. I'm using arm-none-linux-gnueabi-gcc (Sourcery CodeBench Lite 2013.05-24) 4.7.3,
how do I know if this is armhf or armel? Is there a way to tell through the commandline?
2. Since I'm working with iMX6, shouldn't I be ok to use ARMHF?
Try something like:
$ arm-linux-gnueabi-readelf -a `arm-linux-gnueabi-gcc -print-libgcc-file-name` | grep CPU_name | sort -u
Should tell you what your libgcc architecture is
Marek,
Thanks much, I'll give this a try later today.
Ed