This improves the expressiveness of unprivileged BPF by inserting
speculation barriers instead of rejcting the programs.

The approach was presented at LPC'24:
  https://lpc.events/event/18/contributions/1954/ ("Mitigating
  Spectre-PHT using Speculation Barriers in Linux eBPF")
and RAID'24:
  https://arxiv.org/pdf/2405.00078 ("VeriFence: Lightweight and Precise
  Spectre Defenses for Untrusted Linux Kernel Extensions")

Goal of this RFC is to get feedback on the approach and the structuring
into commits.

TODOs to be fixed for final version:
* actually emit arm64 barrier
* fix unexpected_load_success from test_progs for "bpf: Fall back to nospec for sanitization-failures"
* use bpf-next as base commit

Luis Gerhorst (9):
  bpf/arm64: Unset bypass_spec_v4() instead of ignoring BPF_NOSPEC
  bpf: Refactor do_check() if/else into do_check_insn()
  bpf: Return EFAULT on misconfigurations
  bpf: Return EFAULT on internal errors
  bpf: Fall back to nospec if v1 verification fails
  bpf: Allow nospec-protected var-offset stack access
  bpf: Refactor push_stack to return error code
  bpf: Fall back to nospec for sanitization-failures
  bpf: Cut speculative path verification short

 arch/arm64/net/bpf_jit_comp.c                 |  10 +-
 include/linux/bpf.h                           |  14 +-
 include/linux/bpf_verifier.h                  |   3 +-
 kernel/bpf/core.c                             |  17 +-
 kernel/bpf/verifier.c                         | 832 ++++++++++--------
 .../selftests/bpf/progs/verifier_and.c        |   3 +-
 .../selftests/bpf/progs/verifier_bounds.c     |  30 +-
 .../selftests/bpf/progs/verifier_movsx.c      |   6 +-
 .../selftests/bpf/progs/verifier_unpriv.c     |   3 +-
 .../bpf/progs/verifier_value_ptr_arith.c      |  11 +-
 10 files changed, 520 insertions(+), 409 deletions(-)


base-commit: d082ecbc71e9e0bf49883ee4afd435a77a5101b6
-- 
2.48.1