It's easy enough to change it to a single function...

use strict;
use warnings;

use Text::Balanced qw( extract_bracketed );
use HTML::Entities qw( encode_entities );

my $default_allowed_tags = do {
    my @tags = qw(
        A ABBR ACRONYM B BIG CITE CODE DFN EM I KBD Q SAMP
        SMALL SPAN STRONG SUB SUP TT VAR
    );
    qr/${\( join "|", map quotemeta, @tags )}/;
};

sub parse_markup
{
    my ($text, $tags) = @_;
    $tags ||= $default_allowed_tags;
    
    my ($before, $match) = ($text =~ m{
        \A             # start of string
        (.*?)          # leading text ($before)
        (              # either...
            \<\!--      #    the start of a comment
            |           # or...
            $tags\<     #    a tag
        )
    }xsm) or do {
        my @return = split /\|/, $text;
        $return[0] = encode_entities($return[0]);
        return @return;
    };
    
    # strip $before from $text
    substr($text, 0, length($before)) = '';
    
    # If the first thing that needed to be handled was a comment
    if ($match eq '<!--') {
        # Strip it out
        $text =~ s/\<\!--(.+?)--\>//g;
        
        # Handle the rest via recursion
        return join "", $before, parse_markup($text, $tags);
    }
            
    chop(my $found_tag = lc $match);
    substr($text, 0, length($found_tag)) = '';
    my ($got, $remainder) = extract_bracketed($text, q/<"'>/);
    $got = substr($got, 1, length($got) - 2);
    my ($markup, @attrs) = parse_markup($got, $tags);
    my ($more_markup, @more_attrs) = parse_markup($remainder, $tags);
    defined($_) or $_='' for $markup, $more_markup;
    
    join("",
        $before,
        (@attrs ? "<$found_tag @attrs>" : "<$found_tag>"),
        $markup,
        "</$found_tag>",
        $more_markup,
    ), @more_attrs;
}

print for parse_markup(<<'TEXT');
Anyone who watches the Syfy channel knows that on
Monday nights they aired three television series
I<A<EurSUP<e>ka|href="Movies_by_series.pl?series=EWA#EUReKA">|class="t
+itle">,
I<A<Warehouse & 13|href="Movies_by_series.pl?series=EWA#Warehouse_13">
+>,
and I<A<Alphas|href="Movies_by_series.pl?series=EWA#Alphas">>.
Some might not be aware that these three series have formed a crossove
+r
cosmology which I call A<EWA|href="Movies_by_series.pl?series=EWA">
<!-- This is a long string. -->
TEXT
[download]

But personally I prefer the OO version because it means I can have a $parser object, which I can pass around. Code that needs to process some markup gets given the markup, and given the parser, and simply throws the markup at the parser. This makes it easy to switch in a different parser if required (say, one that generated stricter XHTML, or one that processed Markdown).