Re: [RFC PATCH v6 00/32] perf tools: filtering events using eBPF programs

[Alexei Starovoitov]

On 6/9/15 2:44 PM, Arnaldo Carvalho de Melo wrote:
> > btw we've been thinking how to make truly global programs
> > and maps, so that they can be used in 'perf probe' interface.
> > Right now in 'tc' we're using bpf_agent. It's a user space
> > demon that keeps prog_fd and map_fds and passes them to other
> > apps via scm_rights.
> > We've started experimenting with fuse for this purpose as well,
> > but that's similar user space demon and not very convenient
> > for perf to talk to such fuse.
> > The work in that area is still very rough, just fyi.
>
> Cool stuff, but please leave us without daemons, we've exorcised from
> day one ;-)

agreed. I've been looking at hacking debugfs, tracefs, sysfs, but
none of them fit well, since such infra needs to be generic and support
both tracing and networking needs. Looking at procfs at the moment.
Ideally I would like programs/maps to 'attach to a file', so that
they don't disappear when user app exits. Then opening such file
in the future would get user app a new FD that can be used to access
this global map or program. When the file is removed, the prog/map
would be unloaded.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/