From: Cyrill Gorcunov via Tarantool-patches <tarantool-patches@dev.tarantool.org> To: tml <tarantool-patches@dev.tarantool.org> Cc: Vladislav Shpilevoy <v.shpilevoy@tarantool.org> Subject: [Tarantool-patches] [PATCH 1/2] gc: use wide integer for schedule counting Date: Wed, 10 Mar 2021 23:47:09 +0300 [thread overview] Message-ID: <20210310204710.186461-2-gorcunov@gmail.com> (raw) In-Reply-To: <20210310204710.186461-1-gorcunov@gmail.com> Currently we use unsined int for "cleanup" schedule counting, this is safe while this routine is not called too often. Still there is a chance to hit a number wrap on code modification because there is no strict rule on how to use this garbage collector. Lets use wide integers instead, we have only one gc instance and such approach eliminates potential problems in future (actually this should had been done from the beginning since the current gc code flow developed without wrapping in mind). In-scope-of #5806 Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com> --- src/box/gc.c | 6 +++--- src/box/gc.h | 2 +- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/src/box/gc.c b/src/box/gc.c index 1f8cc818d..86c229c34 100644 --- a/src/box/gc.c +++ b/src/box/gc.c @@ -239,15 +239,15 @@ gc_cleanup_fiber_f(va_list ap) { (void)ap; while (!fiber_is_cancelled()) { - int delta = gc.cleanup_scheduled - gc.cleanup_completed; + int64_t delta = gc.cleanup_scheduled - gc.cleanup_completed; if (delta == 0) { /* No pending garbage collection. */ fiber_sleep(TIMEOUT_INFINITY); continue; } - assert(delta > 0); gc_run_cleanup(); gc.cleanup_completed += delta; + assert(delta > 0 && gc.cleanup_completed > 0); fiber_cond_signal(&gc.cleanup_cond); } return 0; @@ -278,7 +278,7 @@ gc_schedule_cleanup(void) static void gc_wait_cleanup(void) { - unsigned scheduled = gc.cleanup_scheduled; + int64_t scheduled = gc.cleanup_scheduled; while (gc.cleanup_completed < scheduled) fiber_cond_wait(&gc.cleanup_cond); } diff --git a/src/box/gc.h b/src/box/gc.h index 829aaf479..2a568c5f9 100644 --- a/src/box/gc.h +++ b/src/box/gc.h @@ -146,7 +146,7 @@ struct gc_state { * sleep until @completed reaches the value of @scheduled * taken at that moment of time. */ - unsigned cleanup_completed, cleanup_scheduled; + int64_t cleanup_completed, cleanup_scheduled; /** * Set if there's a fiber making a checkpoint right now. */ -- 2.29.2
next prev parent reply other threads:[~2021-03-10 20:47 UTC|newest] Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top 2021-03-10 20:47 [Tarantool-patches] [PATCH 0/2] A few fixes for gc and xlog Cyrill Gorcunov via Tarantool-patches 2021-03-10 20:47 ` Cyrill Gorcunov via Tarantool-patches [this message] 2021-03-14 16:30 ` [Tarantool-patches] [PATCH 1/2] gc: use wide integer for schedule counting Vladislav Shpilevoy via Tarantool-patches 2021-03-14 18:24 ` Cyrill Gorcunov via Tarantool-patches 2021-03-10 20:47 ` [Tarantool-patches] [PATCH 2/2] xlog: do not sort sole entry Cyrill Gorcunov via Tarantool-patches 2021-03-12 17:11 ` [Tarantool-patches] [PATCH 0/2] A few fixes for gc and xlog Serge Petrenko via Tarantool-patches 2021-03-12 17:45 ` Cyrill Gorcunov via Tarantool-patches 2021-03-15 23:21 ` Vladislav Shpilevoy via Tarantool-patches 2021-03-16 13:29 ` 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=20210310204710.186461-2-gorcunov@gmail.com \ --to=tarantool-patches@dev.tarantool.org \ --cc=gorcunov@gmail.com \ --cc=v.shpilevoy@tarantool.org \ --subject='Re: [Tarantool-patches] [PATCH 1/2] gc: use wide integer for schedule counting' \ /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