netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Daniel Borkmann <daniel@iogearbox.net>
To: ast@kernel.org
Cc: john.fastabend@gmail.com, andrii.nakryiko@gmail.com,
	netdev@vger.kernel.org, bpf@vger.kernel.org,
	Daniel Borkmann <daniel@iogearbox.net>
Subject: [PATCH bpf-next v2 0/8] Optimize BPF tail calls for direct jumps
Date: Fri, 22 Nov 2019 21:07:53 +0100	[thread overview]
Message-ID: <cover.1574452833.git.daniel@iogearbox.net> (raw)

This gets rid of indirect jumps for BPF tail calls whenever possible.
The series adds emission for *direct* jumps for tail call maps in order
to avoid the retpoline overhead from a493a87f38cf ("bpf, x64: implement
retpoline for tail call") for situations that allow for it, meaning,
for known constant keys at verification time which are used as index
into the tail call map. See patch 7/8 for more general details.

Thanks!

v1  -> v2:
  - added more test cases
  - u8 ip_stable -> bool (Andrii)
  - removed bpf_map_poke_{un,}lock and simplified the code (Andrii)
  - added break into prog_array_map_poke_untrack since there's just
    one prog (Andrii)
  - fixed typo: for for in commit msg (Andrii)
  - reworked __bpf_arch_text_poke (Andrii)
  - added subtests, and comment on tests themselves, NULL-NULL
    transistion (Andrii)
  - in constant map key tracking I've moved the map_poke_track callback
    to once we've finished creating the poke tab as otherwise concurrent
    access from tail call map would blow up (since we realloc the table)
rfc -> v1:
  - Applied Alexei's and Andrii's feeback from
    https://lore.kernel.org/bpf/cover.1573779287.git.daniel@iogearbox.net/T/#t

Daniel Borkmann (8):
  bpf, x86: generalize and extend bpf_arch_text_poke for direct jumps
  bpf: move bpf_free_used_maps into sleepable section
  bpf: move owner type,jited info into array auxiliary data
  bpf: add initial poke descriptor table for jit images
  bpf: add poke dependency tracking for prog array maps
  bpf: constant map key tracking for prog array pokes
  bpf, x86: emit patchable direct jump as tail call
  bpf, testing: add various tail call test cases

 arch/x86/net/bpf_jit_comp.c                   | 264 +++++++---
 include/linux/bpf.h                           |  60 ++-
 include/linux/bpf_verifier.h                  |   3 +-
 include/linux/filter.h                        |  10 +
 kernel/bpf/arraymap.c                         | 215 +++++++-
 kernel/bpf/core.c                             |  73 ++-
 kernel/bpf/map_in_map.c                       |   5 +-
 kernel/bpf/syscall.c                          |  56 +-
 kernel/bpf/verifier.c                         | 120 ++++-
 .../selftests/bpf/prog_tests/tailcalls.c      | 487 ++++++++++++++++++
 tools/testing/selftests/bpf/progs/tailcall1.c |  48 ++
 tools/testing/selftests/bpf/progs/tailcall2.c |  59 +++
 tools/testing/selftests/bpf/progs/tailcall3.c |  31 ++
 tools/testing/selftests/bpf/progs/tailcall4.c |  33 ++
 tools/testing/selftests/bpf/progs/tailcall5.c |  40 ++
 15 files changed, 1364 insertions(+), 140 deletions(-)
 create mode 100644 tools/testing/selftests/bpf/prog_tests/tailcalls.c
 create mode 100644 tools/testing/selftests/bpf/progs/tailcall1.c
 create mode 100644 tools/testing/selftests/bpf/progs/tailcall2.c
 create mode 100644 tools/testing/selftests/bpf/progs/tailcall3.c
 create mode 100644 tools/testing/selftests/bpf/progs/tailcall4.c
 create mode 100644 tools/testing/selftests/bpf/progs/tailcall5.c

-- 
2.21.0


             reply	other threads:[~2019-11-22 20:08 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-11-22 20:07 Daniel Borkmann [this message]
2019-11-22 20:07 ` [PATCH bpf-next v2 1/8] bpf, x86: generalize and extend bpf_arch_text_poke for direct jumps Daniel Borkmann
2019-11-22 20:07 ` [PATCH bpf-next v2 2/8] bpf: move bpf_free_used_maps into sleepable section Daniel Borkmann
2019-11-22 20:07 ` [PATCH bpf-next v2 3/8] bpf: move owner type,jited info into array auxiliary data Daniel Borkmann
2019-11-22 20:07 ` [PATCH bpf-next v2 4/8] bpf: add initial poke descriptor table for jit images Daniel Borkmann
2019-11-22 20:07 ` [PATCH bpf-next v2 5/8] bpf: add poke dependency tracking for prog array maps Daniel Borkmann
2019-11-22 22:55   ` Andrii Nakryiko
2019-11-22 23:06     ` Daniel Borkmann
2019-11-22 23:10       ` Andrii Nakryiko
2019-11-22 20:07 ` [PATCH bpf-next v2 6/8] bpf: constant map key tracking for prog array pokes Daniel Borkmann
2019-11-22 22:57   ` Andrii Nakryiko
2019-11-23 10:39   ` Jakub Sitnicki
2019-11-22 20:08 ` [PATCH bpf-next v2 7/8] bpf, x86: emit patchable direct jump as tail call Daniel Borkmann
2019-11-22 23:09   ` Andrii Nakryiko
2019-11-22 23:25     ` Daniel Borkmann
2019-11-23  2:28       ` Alexei Starovoitov
2019-11-23  5:00         ` Andrii Nakryiko
2019-11-23  6:18           ` Alexei Starovoitov
2019-11-23  9:24             ` Daniel Borkmann
2019-11-22 20:08 ` [PATCH bpf-next v2 8/8] bpf, testing: add various tail call test cases Daniel Borkmann
2019-11-22 23:14   ` Andrii Nakryiko
2019-11-23  2:22 ` [PATCH bpf-next v2 0/8] Optimize BPF tail calls for direct jumps Alexei Starovoitov

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=cover.1574452833.git.daniel@iogearbox.net \
    --to=daniel@iogearbox.net \
    --cc=andrii.nakryiko@gmail.com \
    --cc=ast@kernel.org \
    --cc=bpf@vger.kernel.org \
    --cc=john.fastabend@gmail.com \
    --cc=netdev@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).