Date:   Mon, 27 Jan 2020 18:11:41 -0800
From:   Palmer Dabbelt <palmerdabbelt@...gle.com>
To:     Bjorn Topel <bjorn.topel@...il.com>
Cc:     daniel@...earbox.net, ast@...nel.org, zlim.lnx@...il.com,
        catalin.marinas@....com, will@...nel.org, kafai@...com,
        songliubraving@...com, yhs@...com, andriin@...com,
        shuah@...nel.org, Palmer Dabbelt <palmerdabbelt@...gle.com>,
        netdev@...r.kernel.org, bpf@...r.kernel.org,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
        linux-kselftest@...r.kernel.org,
        clang-built-linux@...glegroups.com, kernel-team@...roid.com
Subject: arm64: bpf: Elide some moves to a0 after calls

There's four patches here, but only one of them actually does anything.  The
first patch fixes a BPF selftests build failure on my machine and has already
been sent to the list separately.  The next three are just staged such that
there are some patches that avoid changing any functionality pulled out from
the whole point of those refactorings, with two cleanups and then the idea.

Maybe this is an odd thing to say in a cover letter, but I'm not actually sure
this patch set is a good idea.  The issue of extra moves after calls came up as
I was reviewing some unrelated performance optimizations to the RISC-V BPF JIT.
I figured I'd take a whack at performing the optimization in the context of the
arm64 port just to get a breath of fresh air, and I'm not convinced I like the
results.

That said, I think I would accept something like this for the RISC-V port
because we're already doing a multi-pass optimization for shrinking function
addresses so it's not as much extra complexity over there.  If we do that we
should probably start puling some of this code into the shared BPF compiler,
but we're also opening the doors to more complicated BPF JIT optimizations.
Given that the BPF JIT appears to have been designed explicitly to be
simple/fast as opposed to perform complex optimization, I'm not sure this is a
sane way to move forward.

I figured I'd send the patch set out as more of a question than anything else.
Specifically:

* How should I go about measuring the performance of these sort of
  optimizations?  I'd like to balance the time it takes to run the JIT with the
  time spent executing the program, but I don't have any feel for what real BPF
  programs look like or have any benchmark suite to run.  Is there something
  out there this should be benchmarked against?  (I'd also like to know that to
  run those benchmarks on the RISC-V port.)
* Is this the sort of thing that makes sense in a BPF JIT?  I guess I've just
  realized I turned "review this patch" into a way bigger rabbit hole than I
  really want to go down...

I worked on top of 5.4 for these, but trivially different versions of the
patches applied on Linus' master a few days ago when I tried.  LMK if those
aren't sane places to start from over here, I'm new to both arm64 and BPF so I
might be a bit lost.

[PATCH 1/4] selftests/bpf: Elide a check for LLVM versions that can't
[PATCH 2/4] arm64: bpf: Convert bpf2a64 to a function
[PATCH 3/4] arm64: bpf: Split the read and write halves of dst
[PATCH 4/4] arm64: bpf: Elide some moves to a0 after calls

Powered by blists - more mailing lists