For non-linear skb's there's a possibility that the kernel's Boyer-Moore text-search implementation may miss matches. There's a warning about this in the kernel source. Include that warning in the man-page. Signed-off-by: Jeremy Sowden <jeremy@xxxxxxxxxx> --- extensions/libxt_string.man | 15 +++++++++++++++ 1 file changed, 15 insertions(+) diff --git a/extensions/libxt_string.man b/extensions/libxt_string.man index 5f1a993c57eb..34a8755ba14e 100644 --- a/extensions/libxt_string.man +++ b/extensions/libxt_string.man @@ -29,3 +29,18 @@ iptables \-A INPUT \-p tcp \-\-dport 80 \-m string \-\-algo bm \-\-string 'GET / # The hex string pattern can be used for non-printable characters, like |0D 0A| or |0D0A|. .br iptables \-p udp \-\-dport 53 \-m string \-\-algo bm \-\-from 40 \-\-to 57 \-\-hex\-string '|03|www|09|netfilter|03|org|00|' +.P +Note: Since Boyer-Moore (BM) performs searches for matchings from right to left +and the kernel may store a packet in multiple discontiguous blocks, it's still +possible that a match could be spread over multiple blocks, in that case this +algorithm won't find it. +.P +If you wish to ensure that such thing won't ever happen, use the +Knuth-Pratt-Morris (KMP) implementation instead. In conclusion, choose the +proper string search algorithm depending on your setting. +.P +Say you're using the textsearch infrastructure for filtering, NIDS or any +similar security focused purpose, then go KMP. Otherwise, if you really care +about performance, say you're classifying packets to apply Quality of Service +(QoS) policies, and you don't mind about possible matchings spread over multiple +fragments, then go BM. -- 2.39.2
