Tarantool development patches archive
 help / color / mirror / Atom feed
From: Maxim Kokryashkin via Tarantool-patches <tarantool-patches@dev.tarantool.org>
To: Sergey Kaplun <skaplun@tarantool.org>
Cc: tarantool-patches@dev.tarantool.org
Subject: Re: [Tarantool-patches] [PATCH luajit] Show name of NYI bytecode in -jv and -jdump.
Date: Mon, 1 Jul 2024 11:36:06 +0300	[thread overview]
Message-ID: <a3ez5msyes5xr3ecekmza7fnr4og7skzos4ohzucqahex5osaa@ndzhqj3zyt65> (raw)
In-Reply-To: <20240626083721.11596-1-skaplun@tarantool.org>

Hi, Sergey!
Thanks for the patch!
LGTM, except for two nits below.

On Wed, Jun 26, 2024 at 11:37:21AM GMT, Sergey Kaplun wrote:
> From: Mike Pall <mike>
>
> Suggested by Sergey Kaplun.
>
> (cherry picked from commit d2fe2a6d465a3e4c74c9876db94ae606f9c6983b)
>
> This patch replaces the number of NYI bytecodes that can't be compiled
I guess `numeric value` is a bit better than `number`, but feel free to
ignore.

> with their names in the `jit.dump()` and -jv outputs. Since the
> functionality is the same, only `jit.dump()` is tested as most popular.
>
> Sergey Kaplun:
> * added the description and the test for the feature
>
> Part of tarantool/tarantool#9924
> ---
>
> Branch: https://github.com/tarantool/luajit/tree/skaplun/lj-567-1176-print-nyi-names
> Related Issues:
> * https://github.com/tarantool/tarantool/issues/9924
> * https://github.com/LuaJIT/LuaJIT/pull/567
> * https://github.com/LuaJIT/LuaJIT/issues/1176
>
>  src/jit/dump.lua                              |  7 +-
>  src/jit/v.lua                                 |  9 +-
>  src/lj_traceerr.h                             |  2 +-
>  .../lj-567-1176-print-nyi-names.test.lua      | 82 +++++++++++++++++++
>  4 files changed, 96 insertions(+), 4 deletions(-)
>  create mode 100644 test/tarantool-tests/lj-567-1176-print-nyi-names.test.lua
>
> diff --git a/src/jit/dump.lua b/src/jit/dump.lua
> index 84fe0044..e7f05253 100644
> --- a/src/jit/dump.lua
> +++ b/src/jit/dump.lua
<snipped>

> diff --git a/src/jit/v.lua b/src/jit/v.lua
> index 934de985..ff4e9c7c 100644
> --- a/src/jit/v.lua
> +++ b/src/jit/v.lua
<snipped>

> diff --git a/src/lj_traceerr.h b/src/lj_traceerr.h
> index 1363c4f3..beecbd8c 100644
> --- a/src/lj_traceerr.h
> +++ b/src/lj_traceerr.h
<snipped>

> diff --git a/test/tarantool-tests/lj-567-1176-print-nyi-names.test.lua b/test/tarantool-tests/lj-567-1176-print-nyi-names.test.lua
> new file mode 100644
> index 00000000..38829d2c
> --- /dev/null
> +++ b/test/tarantool-tests/lj-567-1176-print-nyi-names.test.lua
> @@ -0,0 +1,82 @@
> +local tap = require('tap')
> +
> +-- Test dumping of names of NYI bytecodes to be compiled instead
> +-- of their numbers. See also:
> +-- * https://github.com/LuaJIT/LuaJIT/pull/567,
> +-- * https://github.com/LuaJIT/LuaJIT/issues/1176.
> +
> +local test = tap.test('lj-567-1176-print-nyi-names'):skipcond({
> +  ['Test requires JIT enabled'] = not jit.status(),
> +  ['Disabled on *BSD due to #4819'] = jit.os == 'BSD',
> +})
> +
> +local jparse = require('utils').jit.parse
> +
> +test:plan(3)
> +
> +local function reset_jit()
> +  -- Remove all previous traces.
> +  jit.off()
> +  jit.flush()
> +  jit.on()
> +  jit.opt.start('hotloop=1')
> +end
> +
> +local function nop() end
> +local function test_varg(...)
> +  for _ = 1, 4 do nop(...) end
> +end
> +
> +local function test_uv()
> +  do
> +    -- luacheck: ignore
> +    local uclo
> +    goto close_uv
> +    local function _()
> +      return uclo
> +    end
> +  end
> +  ::close_uv::
> +end
> +
> +-- We only need the abort reason in the test.
> +jparse.start('t')
> +reset_jit()
> +
> +test_varg()
> +
> +local _, aborted_traces = jparse.finish()
> +assert(aborted_traces and aborted_traces[1],
> +       'aborted trace with VARG is persisted')
> +
> +-- We tried to compile only one trace.
> +local reason = aborted_traces[1][1].abort_reason
> +test:like(reason, 'NYI: bytecode VARG', 'bytecode VARG name')
> +
> +jparse.start('t')
> +reset_jit()
> +
> +for _ = 1, 4 do
> +   local _ = function() end
> +end
Maybe it would be better to create a separate function for FNEW too,
just for the uniformity, but that's not a big deal. Feel free to ignore.

> +_, aborted_traces = jparse.finish()
> +assert(aborted_traces and aborted_traces[1],
> +       'aborted trace with FNEW is persisted')
> +
> +reason = aborted_traces[1][1].abort_reason
> +test:like(reason, 'NYI: bytecode FNEW', 'bytecode FNEW name')
> +
> +jparse.start('t')
> +reset_jit()
> +
> +test_uv()
> +test_uv()
> +
> +_, aborted_traces = jparse.finish()
> +assert(aborted_traces and aborted_traces[1],
> +       'aborted trace with UCLO is persisted')
> +
> +reason = aborted_traces[1][1].abort_reason
> +test:like(reason, 'NYI: bytecode UCLO', 'bytecode UCLO name')
> +
> +test:done(true)
> --
> 2.45.1
>

  reply	other threads:[~2024-07-01  8:36 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-06-26  8:37 Sergey Kaplun via Tarantool-patches
2024-07-01  8:36 ` Maxim Kokryashkin via Tarantool-patches [this message]
2024-07-01  9:33   ` Sergey Kaplun via Tarantool-patches
2024-07-04  6:38 ` Sergey Bronnikov via Tarantool-patches
2024-07-09  8:07 ` Sergey Kaplun via Tarantool-patches

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=a3ez5msyes5xr3ecekmza7fnr4og7skzos4ohzucqahex5osaa@ndzhqj3zyt65 \
    --to=tarantool-patches@dev.tarantool.org \
    --cc=m.kokryashkin@tarantool.org \
    --cc=skaplun@tarantool.org \
    --subject='Re: [Tarantool-patches] [PATCH luajit] Show name of NYI bytecode in -jv and -jdump.' \
    /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

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox