CVE-2026-68302

In the Linux kernel, the following vulnerability has been resolved:

amt: re-read skb header pointers after every pull

Several AMT receive and transmit paths cache a pointer into the skb head

(iphdr(), ipv6hdr(), eth_hdr() or the AMT message header) and then call

a helper that can reallocate that head before the cached pointer is used

again.  pskbmaypull(), ipmcmaypull(), ipv6mcmaypull(),

iptunnelpullheader(), ipmccheckigmp() and ipv6mccheckmld() can all

free the old head and move the data, so a pointer taken before the call

dangles afterwards and the later access is a use-after-free of the freed

head.

The affected sites are:

  amtrcv() caches iphdr() before amtparsetype() pulls, then reads

  iph->saddr.

  amtdevxmit() caches iphdr()/ipv6hdr() before ipmccheck_igmp()/

  ipv6mccheckmld() and pskbmay_pull(), then reads the group address.

  amtmulticastdatahandler() caches ethhdr() before pskbmaypull(),

  then writes the L2 header.

  amtmembershipquery_handler() caches the AMT header, the outer and

  inner ethhdr() and iphdr() before iptunnelpullheader() and several

  pulls, then reads and writes them.

  amtigmpv3reporthandler() and amtmldv2reporthandler() cache

  iphdr()/ipv6hdr() and the current group record and read the record

  count from the report header inside the record loop, across the

  *mcmay_pull() calls.

  amtupdatehandler() caches ip_hdr() and the AMT membership-update

  header before pskbmaypull(), iptunnelpullheader(),

  ipmccheck_igmp() and the report handler, then reads iph->daddr and

  amtmu->nonce / amtmu->response_mac.

Fix each site by either snapshotting the scalar that is used after the

pull before the first pull runs, or re-deriving the header pointer from

the skb after the last pull that can move the head.  Values that are

stable across the pull (source and group address, the response MAC and

nonce, the record count, the outer source MAC) are snapshotted; pointers

that are written through or read repeatedly are re-derived.