Re: [PATCH mm/bpf v2] mm: bpf: add find_vma_no_check() without lockdep_assert on mm->mmap_lock

[Alexei Starovoitov <alexei.starovoitov@gmail.com>]

On Wed, Sep 8, 2021 at 11:15 AM Andrew Morton <akpm@linux-foundation.org> wrote:
>
> On Wed, 8 Sep 2021 11:02:58 -0700 Alexei Starovoitov <alexei.starovoitov@gmail.com> wrote:
>
> > > Please describe the expected userspace-visible change from Peter's
> > > patch in full detail?
> >
> > User space expects build_id to be available. Peter patch simply removes
> > that feature.
>
> Are you sure?  He ends up with

More than sure :)
Just look at below.

> static void stack_map_get_build_id_offset(struct bpf_stack_build_id *id_offs,
>                                           u64 *ips, u32 trace_nr, bool user)
> {
>         int i;
>
>         /* cannot access current->mm, fall back to ips */
>         for (i = 0; i < trace_nr; i++) {
>                 id_offs[i].status = BPF_STACK_BUILD_ID_IP;
>                 id_offs[i].ip = ips[i];
>                 memset(id_offs[i].build_id, 0, BUILD_ID_SIZE_MAX);
>         }
>         return;
> }
>
> and you're saying that userspace won't like this because we didn't set
> BPF_STACK_BUILD_ID_VALID?

The patch forces the "fallback path" that in production is seen 0.001%
Meaning that user space doesn't see build_id any more. It sees IPs only.
The user space cannot correlate IPs to binaries. That's what build_id enabled.