#Perl, 83 +1 = 84 bytes

Run with the -n flag.

$\="*
*";print$c="*"x($_+1);for$b(1..$_){@a=($")x$_;@a[$b-1,-$b]=(a,a);print@a}say$c

The literal newline saves 1 byte over \n or $/.

Readable:

$\="*\n*";
print$c="*"x($_+1);
for$b(1..$_){
    @a=($")x$_;
    @a[$b-1,-$b]=(a,a);
    print@a
}
say$c

The code prints the top line and saves it in $c, then prints a bunch of spaces with the appropriate slots replaced with as, then prints the top line again.

The assignment to the $\ variable tells the interpreter to print the contents (an asterisk, a newline, and another asterisk) after every print, but this does NOT occur after a say.

Perl, 83 +1 = 84 bytes

Run with the -n flag.

$\="*
*";print$c="*"x($_+1);for$b(1..$_){@a=($")x$_;@a[$b-1,-$b]=(a,a);print@a}say$c

The literal newline saves 1 byte over \n or $/.

Readable:

$\="*\n*";
print$c="*"x($_+1);
for$b(1..$_){
    @a=($")x$_;
    @a[$b-1,-$b]=(a,a);
    print@a
}
say$c

The code prints the top line and saves it in $c, then prints a bunch of spaces with the appropriate slots replaced with as, then prints the top line again.

The assignment to the $\ variable tells the interpreter to print the contents (an asterisk, a newline, and another asterisk) after every print, but this does NOT occur after a say.