On Tue, Jul 3, 2018 at 3:47 PM, Logan Gunthorpe [off-list ref] wrote:

Ok, I'm at a bit of a loss... When I look at the assembly before and
after the patch, it looks pretty much the same. Additionally, the
function where the undefined exception occurs (caam_jr_interrupt())
doesn't make any use of wr_reg64 or rd_reg64 which is the only thing the
patch changes...

Also, it looks like the nonatomic headers are doing what they are
supposed to on this arch and generating two 32-bit ios and not a 64 bit one.

So I have no idea what's going on here... Are we sure this is the patch
causing the problem? Did you bisect?

Yes, I am sure that 46e4bf08f6388ba748 is the one causing the kernel boot issue.

If I revert 46e4bf08f6388ba748 in linux-next then I can boot mx6 just fine.

kernelci.org reports also confirm the same:

https://storage.kernelci.org/next/master/next-20180629/arm/imx_v6_v7_defconfig/lab-baylibre-seattle/boot-imx6q-wandboard.html

Thanks