Hej, What would it mean if this testcase failure occured to me? I have observing this issue reproducibly with ipset-6.3-genl, though I can't make heads-tails of it. Apparently the members are the same, albeit in different order (which should be irrelevant in hashes anyway), and less memory happens to be used in this occurrence. What immediately comes to mind about the different order is that the hash was perturbed, perhaps by an uninitialized variable or otherwise undefined operation. I know I can't expect you to dig around in the GENL patch, but maybe you are interested. The latest is at git://dev.medozas.de/ipset-6 genl . The fact that occurs only on ipv6 hash types is also worth mentioning. hash:ip6,port: Check listing: --- .foo 2011-04-11 04:27:28.384089921 +0200 +++ hash:ip6,port.t.list0 2011-04-04 00:22:28.398552604 +0200 @@ -1,11 +1,11 @@ Name: test Type: hash:ip,port Header: family inet6 hashsize 1024 maxelem 65536 timeout x -Size in memory: 17112 +Size in memory: 8848 References: 0 Members: -2:1::,tcp:128 timeout x 2:1::1,tcp:128 timeout x -2::,tcp:5 timeout x +2:1::,tcp:128 timeout x 2::1,tcp:5 timeout x +2::,tcp:5 timeout x FAILED Failed test: diff -uI 'Size in memory.*' .foo hash:ip6,port.t.list0 Other tests affected: hash:ip6,port,ip6 hash:net6 hash:net6,port hash:ip6,port,net6 -- To unsubscribe from this list: send the line "unsubscribe netfilter-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
