bpf: bpf_event_entry_gen's alloc needs to be in atomic context
authorDaniel Borkmann <daniel@iogearbox.net>
Fri, 15 Jul 2016 23:15:55 +0000 (01:15 +0200)
committerDavid S. Miller <davem@davemloft.net>
Sun, 17 Jul 2016 05:03:39 +0000 (22:03 -0700)
commit858d68f10238fdd1ebdd0096f912f063e97c6766
treea8ec4a97a18b1c5671f178bad48b84337ec81259
parentc5c4e45c4b79acb23f07e43dac1348e67b4ddf91
bpf: bpf_event_entry_gen's alloc needs to be in atomic context

Should have been obvious, only called from bpf() syscall via map_update_elem()
that calls bpf_fd_array_map_update_elem() under RCU read lock and thus this
must also be in GFP_ATOMIC, of course.

Fixes: 3b1efb196eee ("bpf, maps: flush own entries on perf map release")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
kernel/bpf/arraymap.c