Comment 18 for bug 1591293

I did test the kernel Joseph said:
linux-image-4.7.0-040700rc1-lowlatency_4.7.0-040700rc1.201606100619_amd64.deb

Bug is still there.

I did reset the PC as soon as kernel take too long to boot and began to inform that NCQ command was failing, with similar errors than with xenial kernel, such as:
equipment kernel: [ 64.970612] ata4.00: failed command: READ FPDMA QUEUED
equipment kernel: [ 64.970638] ata4.00: status: { DRDY }
equipment kernel: [ 64.994812] ata4: hard resetting link

After reset and reboot with libata.force=noncq nothing was logged to syslog. I didn't want to complete a fifteen minutes boot with a failing upstream kernel oriented to yakkety in my xenial installation, fearing it could break some bytes. I already spent too much time to install and configure xenial.
I hope this helps anyway.
BTW can't you just check if something has been done with upstream kernels to correct this behaviour?
Thanks for your job.