net: Don't clobber the TTL for TCPv4 rx
authorBarret Rhoden <brho@cs.berkeley.edu>
Thu, 2 Feb 2017 20:08:30 +0000 (15:08 -0500)
committerBarret Rhoden <brho@cs.berkeley.edu>
Thu, 2 Feb 2017 20:08:30 +0000 (15:08 -0500)
commitdb16cd0c4ef480f510ac03446cd16e74efcb6928
treeaf7e04e8a059dded1695f4b042d810497fd27724
parent3e2947a5a0fe1f131085dcbfb3ae8a2e8d5b2ef7
net: Don't clobber the TTL for TCPv4 rx

We need to set it to 0 briefly for the xsum.  When we're done, change it
back.  Interestly enough, TCPv6 and UDP did this already.

You'd see this as packets with a TTL of 0 arriving at the guest VM.

Signed-off-by: Barret Rhoden <brho@cs.berkeley.edu>
kern/src/net/tcp.c