From: Serge Petrenko via Tarantool-patches <tarantool-patches@dev.tarantool.org>
To: Vladislav Shpilevoy <v.shpilevoy@tarantool.org>,
tarantool-patches@dev.tarantool.org, gorcunov@gmail.com
Subject: Re: [Tarantool-patches] [PATCH 06/16] cord_buf: introduce ownership management
Date: Mon, 22 Mar 2021 19:48:03 +0300 [thread overview]
Message-ID: <fc58cdb8-54a0-ec10-a8e3-e59345d0dfdd@tarantool.org> (raw)
In-Reply-To: <eeb753ae3c035238444a76af01afbfcc532794ec.1616200860.git.v.shpilevoy@tarantool.org>
20.03.2021 03:42, Vladislav Shpilevoy пишет:
> The global ibuf used for hot Lua and Lua C code didn't have
> ownership management. As a result, it could be reused in some
> unexpected ways during Lua GC via __gc handlers, even if it was
> currently in use in some code below the stack.
>
> The patch makes cord_ibuf_take() steal the global buffer from its
> global stash, and assign to the current fiber. cord_ibuf_put()
> puts it back to the stash, and detaches from the fiber. If yield
> happens before cord_ibuf_put(), the buffer is detached
> automatically.
>
> Fiber attach/detach is done via on_yield/on_stop triggers. The
> buffer is not supposed to survive a yield, so this allows to
> free/put the buffer back to the stash even if the owner didn't do
> that. For instance, if a Lua exception was raised before
> cord_ibuf_put() was called.
>
> This makes cord buffer being safe to use in any yield-free code,
> even if Lua GC might be started. And in non-Lua code as well.
>
> Part of #5632
> ---
> src/lib/core/cord_buf.c | 150 +++++++++++++++++++++++++++++++----
> src/lib/core/cord_buf.h | 6 +-
> test/app-tap/buffer.test.lua | 59 ++++++++++++++
> 3 files changed, 199 insertions(+), 16 deletions(-)
> create mode 100755 test/app-tap/buffer.test.lua
>
> diff --git a/src/lib/core/cord_buf.c b/src/lib/core/cord_buf.c
> index cac508c3d..9450d75bc 100644
> --- a/src/lib/core/cord_buf.c
> +++ b/src/lib/core/cord_buf.c
> @@ -5,6 +5,7 @@
> */
> #include "cord_buf.h"
> #include "fiber.h"
> +#include "trigger.h"
>
> #include "small/ibuf.h"
>
> @@ -13,35 +14,154 @@ enum {
> CORD_IBUF_START_CAPACITY = 16384,
> };
>
> -static struct ibuf *cord_buf_global = NULL;
> +/** Global buffer with automatic collection on fiber yield. */
> +struct cord_buf {
> + /** Base buffer. */
> + struct ibuf base;
> + /**
> + * Triggers on fiber stop/yield when the buffer is either destroyed or
> + * cached to the global stash for later reuse.
> + */
> + struct trigger on_stop;
> + struct trigger on_yield;
> +#if !NDEBUG
Shouldn't it be `#if !defined(NDEBUG)`? I'm not sure, just asking.
--
Serge Petrenko
next prev parent reply other threads:[~2021-03-22 16:48 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-03-20 0:42 [Tarantool-patches] [PATCH 00/16] Cord buffer, static alloc, and Lua GC bug Vladislav Shpilevoy via Tarantool-patches
2021-03-20 0:42 ` [Tarantool-patches] [PATCH 01/16] fio: don't use shared buffer in pread() Vladislav Shpilevoy via Tarantool-patches
2021-03-22 7:19 ` Cyrill Gorcunov via Tarantool-patches
2021-03-20 0:42 ` [Tarantool-patches] [PATCH 10/16] uri: replace static_alloc with ffi stash and ibuf Vladislav Shpilevoy via Tarantool-patches
2021-03-20 0:42 ` [Tarantool-patches] [PATCH 11/16] buffer: remove static_alloc() from Lua Vladislav Shpilevoy via Tarantool-patches
2021-03-20 0:42 ` [Tarantool-patches] [PATCH 12/16] lua: use lua_pushfstring() instead of tt_sprintf() Vladislav Shpilevoy via Tarantool-patches
2021-03-20 0:42 ` [Tarantool-patches] [PATCH 13/16] sio: rework sio_strfaddr() Vladislav Shpilevoy via Tarantool-patches
2021-03-20 0:42 ` [Tarantool-patches] [PATCH 14/16] sio: increase SERVICE_NAME_MAXLEN size Vladislav Shpilevoy via Tarantool-patches
2021-03-21 21:58 ` Cyrill Gorcunov via Tarantool-patches
2021-03-22 22:32 ` Vladislav Shpilevoy via Tarantool-patches
2021-03-23 6:56 ` Cyrill Gorcunov via Tarantool-patches
2021-03-20 0:42 ` [Tarantool-patches] [PATCH 15/16] sio: introduce and use sio_snprintf() Vladislav Shpilevoy via Tarantool-patches
2021-03-20 0:42 ` [Tarantool-patches] [PATCH 16/16] buffer: remove Lua registers Vladislav Shpilevoy via Tarantool-patches
2021-03-20 0:42 ` [Tarantool-patches] [PATCH 02/16] test: don't use IBUF_SHARED in the tests Vladislav Shpilevoy via Tarantool-patches
2021-03-22 7:35 ` Cyrill Gorcunov via Tarantool-patches
2021-03-20 0:42 ` [Tarantool-patches] [PATCH 03/16] tuple: pass global ibuf explicitly where possible Vladislav Shpilevoy via Tarantool-patches
2021-03-20 0:42 ` [Tarantool-patches] [PATCH 04/16] iconv: take errno before reseting the context Vladislav Shpilevoy via Tarantool-patches
2021-03-20 0:42 ` [Tarantool-patches] [PATCH 05/16] cord_buf: introduce cord_buf API Vladislav Shpilevoy via Tarantool-patches
2021-03-20 0:42 ` [Tarantool-patches] [PATCH 06/16] cord_buf: introduce ownership management Vladislav Shpilevoy via Tarantool-patches
2021-03-22 16:48 ` Serge Petrenko via Tarantool-patches [this message]
2021-03-22 22:32 ` Vladislav Shpilevoy via Tarantool-patches
2021-03-23 7:46 ` Serge Petrenko via Tarantool-patches
2021-03-20 0:42 ` [Tarantool-patches] [PATCH 07/16] buffer: implement ffi stash Vladislav Shpilevoy via Tarantool-patches
2021-03-23 0:29 ` Vladislav Shpilevoy via Tarantool-patches
2021-03-20 0:42 ` [Tarantool-patches] [PATCH 08/16] uuid: replace static_alloc with " Vladislav Shpilevoy via Tarantool-patches
2021-03-20 0:42 ` [Tarantool-patches] [PATCH 09/16] uuid: drop tt_uuid_str() from Lua Vladislav Shpilevoy via Tarantool-patches
2021-03-21 16:38 ` [Tarantool-patches] [PATCH 00/16] Cord buffer, static alloc, and Lua GC bug Vladislav Shpilevoy via Tarantool-patches
2021-03-22 7:52 ` Cyrill Gorcunov via Tarantool-patches
2021-03-22 7:56 ` Konstantin Osipov via Tarantool-patches
2021-03-22 17:17 ` Serge Petrenko via Tarantool-patches
2021-03-23 23:45 ` Vladislav Shpilevoy via Tarantool-patches
2021-03-24 13:28 ` Kirill Yukhin 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=fc58cdb8-54a0-ec10-a8e3-e59345d0dfdd@tarantool.org \
--to=tarantool-patches@dev.tarantool.org \
--cc=gorcunov@gmail.com \
--cc=sergepetrenko@tarantool.org \
--cc=v.shpilevoy@tarantool.org \
--subject='Re: [Tarantool-patches] [PATCH 06/16] cord_buf: introduce ownership management' \
/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