Date: Thu, 19 Jun 2025 12:40:59 +0000 From: bugzilla-noreply@freebsd.org To: net@FreeBSD.org Subject: [Bug 287229] IP reassembly issue in FreeBSD 14.1 Message-ID: <bug-287229-7501-Xa6dsX86ow@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-287229-7501@https.bugs.freebsd.org/bugzilla/> References: <bug-287229-7501@https.bugs.freebsd.org/bugzilla/>
next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D287229 --- Comment #18 from Michael Tuexen <tuexen@freebsd.org> --- (In reply to Lucas Aubard from comment #15) The maxfragbucketsize is computed (in your cases) as follows: * We start with the kernel memory size, which you can observe via the sysctl variable vm.kmem_size. This is either set in /boot/loader.conf or it is a bit less then the memory of the machine. * The maximum memory used for mbufs maxmbufmem is computed as half of the kernel memory. * The maximum number of mbuf clusters is computed as nmbclusters =3D maxmbufmem / MCLBYTES / 4 where MCLBYTES is 2048. * The maximum number of fragments maxfrags is maxfrags =3D nmbclusters / 32. * The maximum number of fragments per queue is maxfragbucketsize =3D imax(maxfrags / (V_ipq_hashsize / 2), 1) where V_ipq_hashsize is 1024 when not changed in /boot/loader.conf kmem_size maxmbufmem nmbclusters maxfrags maxfragbucketsize 200 MB 100 MB 12800 400 1 500 MB 250 MB 32000 1000 1 1000 MB 500 MB 64000 2000 3 2000 MB 1000 MB 128000 4000 7 This explains the parameters you are observing. I was guessing the kmem_siz= e, so it might be a bit smaller or larger but it gives you an approximation of maxfrags and the values you observed for maxfragbucketsize. Please note that any host should avoid IP fragmentation. So, in my view, yo= ur tool tests the behavior under attack. We could increase maxfragbucketsize, = but then you could increase the number of threads you are using from 40 to 400. Then you will hit the maxfrags limit. Could you explain what attack you have in mind? Why are you testing with 40 packets in parallel and not with more or less? --=20 You are receiving this mail because: You are on the CC list for the bug.=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-287229-7501-Xa6dsX86ow>