try to fix RACCT_RSS accounting

[avg]

Summary

There could be a race between the vm daemon setting RACCT_RSS based
on the vm space and vmspace_exit (called from exit1) resetting RACCT_RSS
to zero. In that case we can get a zombie process with non-zero
RACCT_RSS. If the process is jailed, that may break accounting for the
jail. There could be other consequences.

Add an assert that the RSS is zero when the process is reaped.

[kib]

This is probably fine.

For unrelated reasons, I looked at vm_mmap_object() today and I think that the function would cause significant over-accounting in case the decides to return an error.

[avg]

In D9464#195648, @kib wrote:

For unrelated reasons, I looked at vm_mmap_object() today and I think that the function would cause significant over-accounting in case the decides to return an error.

That's a good observation. I haven't seen any problems with the resources that are accounted there, but the code certainly does not handle the error paths properly. But that problem is for another day.