From: Ronan Pigott Date: Sun, 25 Feb 2024 07:23:32 +0000 (-0700) Subject: resolved: reduce the maximum nsec3 iterations to 100 X-Git-Tag: v256-rc1~733^2 X-Git-Url: http://git.ipfire.org/gitweb.cgi?a=commitdiff_plain;h=eba291124bc11f03732d1fc468db3bfac069f9cb;p=thirdparty%2Fsystemd.git resolved: reduce the maximum nsec3 iterations to 100 According to RFC9267, the 2500 value is not helpful, and in fact it can be harmful to permit a large number of iterations. Combined with limits on the number of signature validations, I expect this will mitigate the impact of maliciously crafted domains designed to cause excessive cryptographic work. --- diff --git a/src/resolve/resolved-dns-dnssec.c b/src/resolve/resolved-dns-dnssec.c index 33094fe0fa5..a373b610b0e 100644 --- a/src/resolve/resolved-dns-dnssec.c +++ b/src/resolve/resolved-dns-dnssec.c @@ -28,8 +28,9 @@ DEFINE_TRIVIAL_CLEANUP_FUNC_FULL(EC_KEY*, EC_KEY_free, NULL); /* Permit a maximum clock skew of 1h 10min. This should be enough to deal with DST confusion */ #define SKEW_MAX (1*USEC_PER_HOUR + 10*USEC_PER_MINUTE) -/* Maximum number of NSEC3 iterations we'll do. RFC5155 says 2500 shall be the maximum useful value */ -#define NSEC3_ITERATIONS_MAX 2500 +/* Maximum number of NSEC3 iterations we'll do. RFC5155 says 2500 shall be the maximum useful value, but + * RFC9276 § 3.2 says that we should reduce the acceptable iteration count */ +#define NSEC3_ITERATIONS_MAX 100 /* * The DNSSEC Chain of trust: