On Thu, Sep 7, 2017 at 5:56 PM, Stefano Brivio <sbri...@redhat.com> wrote: > On Thu, 07 Sep 2017 18:52:02 -0600 > Subash Abhinov Kasiviswanathan <subas...@codeaurora.org> wrote: > >> We are seeing a possible use after free in ip6_dst_destroy. >> >> It appears as if memory of the __DST_METRICS_PTR(old) was freed in some >> path and allocated >> to ion driver. ion driver has also freed it. Finally the memory is freed >> by the >> fib gc and crashes since it is already deallocated. >> >> Target is running an ARM64 Android based 4.9 kernel. >> Issue was seen once on a regression rack (sorry, no reproducer). >> Any pointers to debug this is highly appreciated. >> >> [ 3489.470581] [<ffffff83c0a289c0>] object_err+0x4c/0x5c >> [ 3489.470586] [<ffffff83c0a2b284>] free_debug_processing+0x2e0/0x398 >> [ 3489.470589] [<ffffff83c0a2b63c>] __slab_free+0x300/0x3e0 >> [ 3489.470593] [<ffffff83c0a2bfc8>] kfree+0x28c/0x290 >> [ 3489.470601] [<ffffff83c16b9580>] >> __dst_destroy_metrics_generic+0x6c/0x78 >> [ 3489.470609] [<ffffff83c17d3408>] ip6_dst_destroy+0xb0/0xb4 > > Should be fixed by: > > commit ad65a2f05695aced349e308193c6e2a6b1d87112 > Author: Wei Wang <wei...@google.com> > Date: Sat Jun 17 10:42:35 2017 -0700 > > ipv6: call dst_hold_safe() properly
Obviously it should not. One is dst metric, the other is dst.
