Re: [RFC PATCH v2 00/26] perf tools: Support uBPF script

From: Alexei Starovoitov
Date: Sun Jun 26 2016 - 16:48:12 EST


On Sun, Jun 26, 2016 at 11:20:52AM +0000, He Kuang wrote:
> This patchset is based on Wang Nan's v1:
> http://thread.gmane.org/gmane.linux.kernel/2203717/focus=2203707
>
> """
> This patch set allows to perf invoke some user space BPF scripts on
> some point. uBPF scripts and kernel BPF scripts reside in one BPF
> object. They communicate with each other with BPF maps. uBPF
> scripts can invoke helper functions provided by perf.
>
> At least following new features can be achieved based on uBPF
> support:
>
> 1) Report statistical result:
>
> Like DTrace, perf print statistical report before quit. No need
> to extract data using 'perf report'. Statistical method is
> controled by user.
>
> 2) Control perf's behavior:
>
> Dynamically adjust period of different events. Policy is defined
> by user.
> """
>
> and modified by following the reviewers' suggestions.
>
> v1-v2:
>
> - Split bpf vm part out of kernel/bpf/core.c and link to it instead
> of using ubpf library(Suggested by Alexei Starovoitov). And add
> runtime bounds check just like ubpf library does.

hmm. I don't think I suggested to hack bpf/core.c into separate file
and compile it for userspace...
Also I think the prior experience taught us that sharing code between
kernel and user space will have lots of headaches long term.
I think it makes more sense to use bcc approach. Just have c+py
or c+lua or c+c. llvm has x86 backend too. If you integrate
clang/llvm (bcc approach) you can compile different functions with
different backends... if you don't want to embed the compiler,
have two .c files. Compile one for bpf target and another for native.