Tarantool development patches archive
 help / color / mirror / Atom feed
From: Vladislav Shpilevoy <v.shpilevoy@tarantool.org>
To: Alexander Turenko <alexander.turenko@tarantool.org>,
	Konstantin Osipov <kostja@tarantool.org>
Cc: tarantool-patches@freelists.org
Subject: [tarantool-patches] Re: [PATCH] Fix premature cdata collecting in luaT_pusherror()
Date: Fri, 7 Dec 2018 23:42:31 +0300	[thread overview]
Message-ID: <b1c653b1-d612-7593-15eb-448f84d3a09b@tarantool.org> (raw)
In-Reply-To: <f75b43c2bdf281b8bdbe79284c87023dbf328a61.1544206105.git.alexander.turenko@tarantool.org>



On 07/12/2018 21:15, Alexander Turenko wrote:
> This is follow up of 28c7e667aee9be8c3288597bcc179d9b4e7b4940 to fix
> luaT_pusherror() itself, not only luaT_error().
> 
> Fixes #1955 (again).
> 
> It is backport of 480868ff8f304ead2fd03e317902a1e8d41c6248 for 1.10.
> ---
> 
> https://github.com/tarantool/tarantool/issues/1955
> https://github.com/tarantool/tarantool/tree/Totktonada/gh-1955-fix-luaT_pusherror-1.10
> 
>   src/lua/utils.c | 20 ++++++++++----------
>   1 file changed, 10 insertions(+), 10 deletions(-)
> 
> diff --git a/src/lua/utils.c b/src/lua/utils.c
> index 27772a9d3..b05d16a73 100644
> --- a/src/lua/utils.c
> +++ b/src/lua/utils.c
> @@ -898,12 +898,20 @@ luaL_error_gc(struct lua_State *L)
>   void
>   luaT_pusherror(struct lua_State *L, struct error *e)
>   {
> +	/*
> +	 * gh-1955 luaT_pusherror allocates Lua objects, thus it

Sorry, but usually we do not mention issue numbers in comments.
Please, reword the comment so as to do not use explicit issue
number.

> +	 * may trigger GC. GC may invoke finalizers which are
> +	 * arbitrary Lua code, potentially invalidating last error
> +	 * object, hence error_ref below.
> +	 *
> +	 * It also important to reference the error first and only
> +	 * then set the finalizer.
> +	 */
> +	error_ref(e);
>   	assert(CTID_CONST_STRUCT_ERROR_REF != 0);
>   	struct error **ptr = (struct error **) luaL_pushcdata(L,
>   		CTID_CONST_STRUCT_ERROR_REF);
>   	*ptr = e;
> -	/* The order is important - first reference the error, then set gc */
> -	error_ref(e);
>   	lua_pushcfunction(L, luaL_error_gc);
>   	luaL_setcdatagc(L, -2);
>   }
> @@ -913,15 +921,7 @@ luaT_error(lua_State *L)
>   {
>   	struct error *e = diag_last_error(&fiber()->diag);
>   	assert(e != NULL);
> -	/*
> -	 * gh-1955 luaT_pusherror allocates Lua objects, thus it may trigger
> -	 * GC. GC may invoke finalizers which are arbitrary Lua code,
> -	 * potentially invalidating last error object, hence error_ref
> -	 * below.
> -	 */
> -	error_ref(e);
>   	luaT_pusherror(L, e);
> -	error_unref(e);
>   	lua_error(L);
>   	unreachable();
>   	return 0;
> 

  reply	other threads:[~2018-12-07 20:42 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-12-07 15:16 [tarantool-patches] " Alexander Turenko
2018-12-07 17:32 ` [tarantool-patches] " Konstantin Osipov
2018-12-07 18:15   ` [tarantool-patches] " Alexander Turenko
2018-12-07 20:42     ` Vladislav Shpilevoy [this message]
2018-12-07 22:24       ` [tarantool-patches] " Konstantin Osipov
2018-12-09 10:35 ` [tarantool-patches] " Vladimir Davydov

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=b1c653b1-d612-7593-15eb-448f84d3a09b@tarantool.org \
    --to=v.shpilevoy@tarantool.org \
    --cc=alexander.turenko@tarantool.org \
    --cc=kostja@tarantool.org \
    --cc=tarantool-patches@freelists.org \
    --subject='[tarantool-patches] Re: [PATCH] Fix premature cdata collecting in luaT_pusherror()' \
    /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