df: ceil() the percentage value

[jlduran]

Summary

"The percentage value shall be expressed as a positive integer, with any
fractional result causing it to be rounded [up] to the next highest
integer."

• https://pubs.opengroup.org/onlinepubs/9799919799/utilities/df.html

Use a hand-rolled ceil() equivalent (as long as it doesn't overflow), to
avoid using math.h in our percent() function.

Fix the tests accordingly, they now match the percentage values on
NetBSD.

Also, use the same percent() function for the calculation of used
inodes.

Update the percentage values in the manual page as well.

Test Plan

$ kyua test
t_df:hflag  ->  passed  [0.009s]
t_df:normal  ->  passed  [0.008s]

[jlduran]

• Round up the used inodes percentage as well. Although not required by POSIX, it would be odd to just have one percentage rounded up.

[des]

bin/df/df.c

541

Why pass a double if the value is an integer?

544

This is a direct quote from the standard; please add quotation marks and a citation.

560–561

Why pass a double if the value is an integer?

[des]

bin/df/df.c

544

(or replace the quote with a paraphrase)

[jlduran]

Address suggestions:

• Quote and cite the text from POSIX.
• Use an unsigned decimal number to express the percentage. Keep the current width of the field to avoid changing the spacing between columns.

[jlduran]

Update the manual page percentages as well.

[jlduran]

• Reduce duplication by abstracting the percent() function.