Nathan Lynch [off-list ref] writes:
Signed-off-by: Nathan Lynch <redacted>
---
Notes:
This could be considered a sequel to:
https://lore.kernel.org/linuxppc-dev/20210504030358.1715034-1-nathanl@linux.ibm.com/ (local)
which tried to address both the suboptimal delay behavior as well as naming
issues. The present change achieves the performance improvement and leaves
the matter of naming for another time.
I've incorporated Alexey's feedback from that series to avoid division when
calculating the arguments to usleep_range().
Oops, this part is redundant with information in the cover. Git should
discard it if/when applied, though.