Prevent premature hsplit() calls, and only trigger REHASH after hsplit()

Triggering a hsplit due to long chain length allows an attacker
to create a carefully chosen set of keys which can cause the hash
to use 2 * (2**32) * sizeof(void *) bytes ram. AKA a DOS via memory
exhaustion. Doing so also takes non trivial time.

Eliminating this check, and only inspecting chain length after a
normal hsplit() (triggered when keys>buckets) prevents the attack
entirely, and makes such attacks relatively benign.

(cherry picked from commit f2a571dae7d70f7e3b59022834d8003ecd2df884)
(which was itself cherry picked (with changes) from commit f1220d61455253b170e81427c9d0357831ca0fac)

1 files changed, 16 insertions, 2 deletions

diff --git a/ext/Hash-Util-FieldHash/t/10_hash.t b/ext/Hash-Util-FieldHash/t/10_hash.t
index 29c2f4dafa..c266b6a45b 100644
--- a/ext/Hash-Util-FieldHash/t/10_hash.t
+++ b/ext/Hash-Util-FieldHash/t/10_hash.t
@@ -46,15 +46,29 @@ use constant START     => "a";
 
 # some initial hash data
 fieldhash my %h2;
-%h2 = map {$_ => 1} 'a'..'cc';
+my $counter= "a";
+$h2{$counter++}++ while $counter ne 'cd';
 
 ok (!Internals::HvREHASH(%h2), 
     "starting with pre-populated non-pathological hash (rehash flag if off)");
 
 my @keys = get_keys(\%h2);
+my $buckets= buckets(\%h2);
 $h2{$_}++ for @keys;
+$h2{$counter++}++ while buckets(\%h2) == $buckets; # force a split
 ok (Internals::HvREHASH(%h2), 
-    scalar(@keys) . " colliding into the same bucket keys are triggering rehash");
+    scalar(@keys) . " colliding into the same bucket keys are triggering rehash after split");
+
+# returns the number of buckets in a hash
+sub buckets {
+    my $hr = shift;
+    my $keys_buckets= scalar(%$hr);
+    if ($keys_buckets=~m!/([0-9]+)\z!) {
+        return 0+$1;
+    } else {
+        return 8;
+    }
+}
 
 sub get_keys {
     my $hr = shift;