bpf: introduce BPF_PROG_TEST_RUN command
authorAlexei Starovoitov <ast@fb.com>
Fri, 31 Mar 2017 04:45:38 +0000 (21:45 -0700)
committerDavid S. Miller <davem@davemloft.net>
Sat, 1 Apr 2017 19:45:57 +0000 (12:45 -0700)
commit1cf1cae963c2e6032aebe1637e995bc2f5d330f4
tree162a176108d3d01a164ddd48d58908da1660e068
parent98cd1552ea27e512c7e99e2aa76042a26e4fb25c
bpf: introduce BPF_PROG_TEST_RUN command

development and testing of networking bpf programs is quite cumbersome.
Despite availability of user space bpf interpreters the kernel is
the ultimate authority and execution environment.
Current test frameworks for TC include creation of netns, veth,
qdiscs and use of various packet generators just to test functionality
of a bpf program. XDP testing is even more complicated, since
qemu needs to be started with gro/gso disabled and precise queue
configuration, transferring of xdp program from host into guest,
attaching to virtio/eth0 and generating traffic from the host
while capturing the results from the guest.

Moreover analyzing performance bottlenecks in XDP program is
impossible in virtio environment, since cost of running the program
is tiny comparing to the overhead of virtio packet processing,
so performance testing can only be done on physical nic
with another server generating traffic.

Furthermore ongoing changes to user space control plane of production
applications cannot be run on the test servers leaving bpf programs
stubbed out for testing.

Last but not least, the upstream llvm changes are validated by the bpf
backend testsuite which has no ability to test the code generated.

To improve this situation introduce BPF_PROG_TEST_RUN command
to test and performance benchmark bpf programs.

Joint work with Daniel Borkmann.

Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
include/linux/bpf.h
include/uapi/linux/bpf.h
kernel/bpf/syscall.c
net/Makefile
net/bpf/Makefile [new file with mode: 0644]
net/bpf/test_run.c [new file with mode: 0644]
net/core/filter.c