Re: [PATCH net-next] net: Initialize table in fib result

2015-09-16 Thread Fabio Estevam
On Wed, Sep 16, 2015 at 12:38 PM, David Ahern wrote: > The root cause is use of res.table uninitialized. > > Thanks to Nikolay for noticing the uninitialized use amongst the maze of > gotos. > > Fixes: b7503e0cdb5d ("net: Add FIB table id to rtable") > Reported-by: Sergey Senozhatsky > Reported-

Re: [PATCH net-next] net: Initialize table in fib result

2015-09-16 Thread David Ahern
On 9/16/15 9:56 AM, Nikolay Aleksandrov wrote: Just to have it documented: I don't think we need the second NULLing, but it doesn't hurt. I think we do. After the second one there is a goto to local_input which uses res.table. The second goto is reachable 'if !IN_DEV_FORWARD(in_dev)' in which

Re: [PATCH net-next] net: Initialize table in fib result

2015-09-16 Thread Nikolay Aleksandrov
On 09/16/2015 05:38 PM, David Ahern wrote: > Sergey, Richard and Fabio reported an oops in ip_route_input_noref. e.g., > from Richard: > > [0.877040] BUG: unable to handle kernel NULL pointer dereference at > 0056 > [0.877597] IP: [] ip_route_input_noref+0x1a2/0xb00 > [0.

[PATCH net-next] net: Initialize table in fib result

2015-09-16 Thread David Ahern
Sergey, Richard and Fabio reported an oops in ip_route_input_noref. e.g., from Richard: [0.877040] BUG: unable to handle kernel NULL pointer dereference at 0056 [0.877597] IP: [] ip_route_input_noref+0x1a2/0xb00 [0.877597] PGD 3fa14067 PUD 3fa6e067 PMD 0 [0.877597] Oo