Date: Fri, 05 Sep 2025 09:53:29 +0000 From: bugzilla-noreply@freebsd.org To: bugs@FreeBSD.org Subject: [Bug 263229] virtio vtnet driver with rxcsum errors Message-ID: <bug-263229-227-8sAFAXbwPb@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-263229-227@https.bugs.freebsd.org/bugzilla/>
index | next in thread | previous in thread | raw e-mail
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=263229 --- Comment #16 from Michael Tuexen <tuexen@freebsd.org> --- (In reply to Danilo Egea Gondolfo from comment #15) Thanks for testing. I will put it up for review. It is possible to receive a SCTP/TCP/UDP packet, where the checksum is not yet computed. There is meta information available about this status. If you receive such packets via SCTP/TCP/UDP sockets, everything will be handled correctly. Please note that you can't receive TCP/UDP packets via raw sockets (the same applies to SCTP if the kernel module is loaded). The point here is that the dhclient doesn't use a UDP socket, but a bpf device. This passes the packet up as is without the metainformation. So the information that the checksum has not been computed is lost in this case. With the patch the checksum validation of received UDP packets accepts packets where the checksum field contains: (1) zero. (2) the correct checksum. (3) the correct pseudoheader checksum. (3) is new and corresponds to the case where the checksum is not computed yet. Does this makes things cleaner? -- You are receiving this mail because: You are the assignee for the bug.help
Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-263229-227-8sAFAXbwPb>