From: Alexei Starovoitov <alexei.starovoi...@gmail.com> Date: Fri, 30 Nov 2018 13:58:20 -0800
> On Thu, Nov 29, 2018 at 07:32:41PM -0800, David Miller wrote: >> +/* If BPF_F_ANY_ALIGNMENT is used in BPF_PROF_LOAD command, the >> + * verifier will allow any alignment whatsoever. This bypasses >> + * what CONFIG_EFFICIENT_UNALIGNED_ACCESS would cause it to do. > > I think majority of user space folks who read uapi/bpf.h have no idea > what that kernel config does. > Could you reword the comment here to say that this flag is only > effective on architectures and like sparc and mips that don't > have efficient unaligned access and ignored on x86/arm64 ? Ok. >> + if (!IS_ENABLED(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS) && >> + (attr->prog_flags & BPF_F_ANY_ALIGNMENT) && >> + !capable(CAP_SYS_ADMIN)) >> + return -EPERM; > > I guess we don't want to add: > if (IS_ENABLED(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS) && > (attr->prog_flags & BPF_F_ANY_ALIGNMENT)) > return -EINVAL; > > so that test_verifier.c can just unconditionally pass this flag > on all archs ? Right. >> @@ -247,7 +249,11 @@ int bpf_verify_program(enum bpf_prog_type type, const >> struct bpf_insn *insns, >> attr.log_level = log_level; >> log_buf[0] = 0; >> attr.kern_version = kern_version; >> - attr.prog_flags = strict_alignment ? BPF_F_STRICT_ALIGNMENT : 0; >> + if (strict_alignment) >> + prog_flags |= BPF_F_STRICT_ALIGNMENT; >> + if (any_alignment) >> + prog_flags |= BPF_F_ANY_ALIGNMENT; >> + attr.prog_flags = prog_flags; > > instead of adding another argument may be replace 'int strict_alignment' > with '__u32 prog_flags' ? > and future flags won't need tweaks to bpf_verify_program() api. Yeah the number of arguments are getting out of control, I'll do that.