oce(4): Don't directly access usespace

[brooks]

Summary

Replace direct stores to userspace address (never safe and broken on
modern CPUs) with a copyout. Use a static assert on the size to ensure
we don't overflow the field.

Sponsored by: DARPA

[brooks]

This driver should probably be on the list of potential drivers to cull in 15. Other than quite a few bug fixes to ioctl handling it hasn't seen any non-sweeping commits since 2016.

[emaste]

This driver should probably be on the list of potential drivers to cull in 15.

Probably - are you willing to add the man page deprecation notice and gone_in?

[brooks]

In D40519#922646, @emaste wrote:

This driver should probably be on the list of potential drivers to cull in 15.

Probably - are you willing to add the man page deprecation notice and gone_in?

D40531

[markj]

sys/dev/oce/oce_if.c

2296

Probably we want to check for an error here? And not do anything if rc is already not equal to 0.

[brooks]

• Fix error returns

[markj]

sys/dev/oce/oce_if.c

2282

Missing braces.

[brooks]

• Fix missing braces

[jhb]

sys/dev/oce/oce_if.c

2296

rc is known to be zero here I think. If rc was non-zero up at line 2275 (new) we jump to the error label. Similarly the new goto at 2282 means that rc is known to be zero here.

[brooks]

sys/dev/oce/oce_if.c

2296

The previous version didn't check the copyout return (I'd blindly copied the strncpy). I chose a slightly different strategy since falling through is ok here.