On 2013-02-22 10:57, Vitaly Chipounov wrote: > Hi, > > On 21.02.2013 15:33, Jan Kiszka wrote: >> On 2013-02-15 12:00, Vitaly Chipounov wrote: >>> A socket may still have references to it in various queues >>> at the time it is freed, causing memory corruptions. >> Did you see it in practice? Or is this patch based on code review? What >> will happen if those queued mbufs find their ifq_so NULL? > > I have a packet trace that triggers this problem when it is injected > into the guest's NIC. > I am still not quite sure why this happens, but suspect that it could be > caused by malformed/partial TCP/IP streams.
OK. Is it shareable? We need to understand why these bufs still hang around - and exclude they leak or cause more problems elsewhere. Jan -- Siemens AG, Corporate Technology, CT RTC ITP SDP-DE Corporate Competence Center Embedded Linux