Hi David,
The following pull-request contains BPF updates for your *net* tree.
The main changes are:
1) Fix a bug in the original fix to prevent out of bounds speculation when
multiple tail call maps from different branches or calls end up at the
same tail call helper invocation, from Daniel.
2) Two selftest fixes, one in reuseport_bpf_numa where test is skipped in
case of missing numa support and another one to update kernel config to
properly support xdp_meta.sh test, from Anders.
Please consider pulling these changes from:
git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf.git
Would be great if you have a chance to merge net into net-next after that.
The verifier fix would be needed later as a dependency in bpf-next for
upcomig work there. When you do the merge there's a trivial conflict on
BPF side with 849fa50662fb ("bpf/verifier: refine retval R0 state for
bpf_get_stack helper"): Resolution is to keep both functions, the
do_refine_retval_range() and record_func_map().
Thanks a lot!
----------------------------------------------------------------
The following changes since commit d775f26b295a0a303f7a73d7da46e04296484fe7:
cxgb4: fix offset in collecting TX rate limit info (2018-05-18 13:54:48 -0400)
are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf.git
for you to fetch changes up to c93552c443ebc63b14e26e46d2e76941c88e0d71:
bpf: properly enforce index mask to prevent out-of-bounds speculation
(2018-05-24 08:15:43 -0700)
----------------------------------------------------------------
Anders Roxell (2):
selftests: bpf: config: enable NET_SCH_INGRESS for xdp_meta.sh
selftests: net: reuseport_bpf_numa: don't fail if no numa support
Daniel Borkmann (1):
bpf: properly enforce index mask to prevent out-of-bounds speculation
include/linux/bpf_verifier.h | 2 +-
kernel/bpf/verifier.c | 86 ++++++++++++++++++------
tools/testing/selftests/bpf/config | 2 +
tools/testing/selftests/net/reuseport_bpf_numa.c | 4 +-
4 files changed, 70 insertions(+), 24 deletions(-)
