Re: [PATCH net] inet: frag: Always orphan skbs inside ip_defrag()

2016-01-28 Thread David Miller
From: Joe Stringer Date: Fri, 22 Jan 2016 15:49:12 -0800 > Later parts of the stack (including fragmentation) expect that there is > never a socket attached to frag in a frag_list, however this invariant > was not enforced on all defrag paths. This could lead to the >

Re: [PATCH net] inet: frag: Always orphan skbs inside ip_defrag()

2016-01-26 Thread Joe Stringer
On 25 January 2016 at 17:11, Joe Stringer wrote: > On 22 January 2016 at 17:22, Eric Dumazet wrote: >> On Fri, 2016-01-22 at 15:49 -0800, Joe Stringer wrote: >>> Later parts of the stack (including fragmentation) expect that there is >>> never a socket

Re: [PATCH net] inet: frag: Always orphan skbs inside ip_defrag()

2016-01-25 Thread Eric Dumazet
On Mon, 2016-01-25 at 17:11 -0800, Joe Stringer wrote: > Thanks, I can roll this into a v2 (or keep as a separate patch?). I > got sidetracked on the IPv6 side, some other issues are blocking me on > that but I intend to continue following up there as well. No, don't worry, I will submit this in

Re: [PATCH net] inet: frag: Always orphan skbs inside ip_defrag()

2016-01-25 Thread Joe Stringer
On 22 January 2016 at 17:22, Eric Dumazet wrote: > On Fri, 2016-01-22 at 15:49 -0800, Joe Stringer wrote: >> Later parts of the stack (including fragmentation) expect that there is >> never a socket attached to frag in a frag_list, however this invariant >> was not

[PATCH net] inet: frag: Always orphan skbs inside ip_defrag()

2016-01-22 Thread Joe Stringer
Later parts of the stack (including fragmentation) expect that there is never a socket attached to frag in a frag_list, however this invariant was not enforced on all defrag paths. This could lead to the BUG_ON(skb->sk) during ip_do_fragment(), as per the call stack at the end of this commit

Re: [PATCH net] inet: frag: Always orphan skbs inside ip_defrag()

2016-01-22 Thread Eric Dumazet
On Fri, 2016-01-22 at 15:49 -0800, Joe Stringer wrote: > Later parts of the stack (including fragmentation) expect that there is > never a socket attached to frag in a frag_list, however this invariant > was not enforced on all defrag paths. This could lead to the > BUG_ON(skb->sk) during