Introduce tracking of bpf program stack depth in the verifier and use that
info to reduce bpf program stack consumption in the interpreter and x64 JIT.
Other JITs can take advantage of it as well in the future.
Most of the programs consume very little stack, so it's good optimization
in general and it's the first step toward bpf to bpf function calls.

Also use internal opcode for bpf_tail_call() marking to make clear
that jmp|call|x opcode is not uapi and may be used for actual
indirect call opcode in the future.

Alexei Starovoitov (9):
  bpf: free up BPF_JMP | BPF_CALL | BPF_X opcode
  bpf: split bpf core interpreter
  bpf: teach verifier to track stack depth
  bpf: reconcile bpf_tail_call and stack_depth
  bpf: track stack depth of classic bpf programs
  bpf: fix stack_depth usage by test_bpf.ko
  bpf: use different interpreter depending on required stack size
  bpf: change x86 JITed program stack layout
  bpf: take advantage of stack_depth tracking in x64 JIT

 arch/arm64/net/bpf_jit_comp.c     |  2 +-
 arch/powerpc/net/bpf_jit_comp64.c |  2 +-
 arch/s390/net/bpf_jit_comp.c      |  2 +-
 arch/sparc/net/bpf_jit_comp_64.c  |  2 +-
 arch/x86/net/bpf_jit.S            | 20 ++++++------
 arch/x86/net/bpf_jit_comp.c       | 65 +++++++++++++++++++++------------------
 include/linux/bpf.h               |  1 +
 include/linux/filter.h            |  3 ++
 kernel/bpf/core.c                 | 47 ++++++++++++++++++++++------
 kernel/bpf/verifier.c             | 13 ++++++--
 lib/test_bpf.c                    | 25 ++++++++++++++-
 net/core/filter.c                 | 36 +++++++++++++---------
 12 files changed, 147 insertions(+), 71 deletions(-)

-- 
2.9.3

Reply via email to