From: Pablo Neira Ayuso <pablo@netfilter.org>
To: Thomas Haller <thaller@redhat.com>
Cc: NetFilter <netfilter-devel@vger.kernel.org>
Subject: Re: [PATCH nft 3/4] src: cache result of time() during parsing/output
Date: Mon, 28 Aug 2023 17:02:18 +0200 [thread overview]
Message-ID: <ZOy2+ln2nMnveET2@calendula> (raw)
In-Reply-To: <20230825132942.2733840-4-thaller@redhat.com>
On Fri, Aug 25, 2023 at 03:24:19PM +0200, Thomas Haller wrote:
> When we parse/output a larger set of data, we should only call time()
> once. With every call of time(), the value keeps ticking (and is subject
> to time reset). Previously, one parse/output operation will make
> decisions on potentially different timestamps.
>
> Add a cache to the parse/output context, and only fetch time() once
> per operation.
>
> Signed-off-by: Thomas Haller <thaller@redhat.com>
> ---
> include/datatype.h | 6 ++++++
> src/datatype.c | 16 ++++++++++++++++
> src/meta.c | 4 ++--
> 3 files changed, 24 insertions(+), 2 deletions(-)
>
> diff --git a/include/datatype.h b/include/datatype.h
> index 79d996edd348..abd093765703 100644
> --- a/include/datatype.h
> +++ b/include/datatype.h
> @@ -2,6 +2,7 @@
> #define NFTABLES_DATATYPE_H
>
> #include <json.h>
> +#include <time.h>
>
> /**
> * enum datatypes
> @@ -121,12 +122,17 @@ enum byteorder {
> struct expr;
>
> struct ops_cache {
> + time_t time;
> + bool has_time;
> };
>
> #define CTX_CACHE_INIT() \
> { \
> + .has_time = false, \
> }
>
> +extern time_t ops_cache_get_time(struct ops_cache *cache);
> +
> /**
> * enum datatype_flags
> *
> diff --git a/src/datatype.c b/src/datatype.c
> index dd6a5fbf5df8..933d832c4f4d 100644
> --- a/src/datatype.c
> +++ b/src/datatype.c
> @@ -35,6 +35,22 @@
>
> #include <netinet/ip_icmp.h>
>
> +time_t ops_cache_get_time(struct ops_cache *cache)
> +{
> + time_t t;
> +
> + if (!cache || !cache->has_time) {
> + t = time(NULL);
> + if (cache) {
> + cache->has_time = true;
> + cache->time = time(NULL);
> + }
> + } else
> + t = cache->time;
> +
> + return t;
> +}
> +
> static const struct datatype *datatypes[TYPE_MAX + 1] = {
> [TYPE_INVALID] = &invalid_type,
> [TYPE_VERDICT] = &verdict_type,
> diff --git a/src/meta.c b/src/meta.c
> index 4f383269d032..1d853b219fe6 100644
> --- a/src/meta.c
> +++ b/src/meta.c
> @@ -496,7 +496,7 @@ static void hour_type_print(const struct expr *expr, struct output_ctx *octx)
> time_t ts;
>
> /* Obtain current tm, so that we can add tm_gmtoff */
> - ts = time(NULL);
> + ts = ops_cache_get_time(octx->ops_cache);
Following the idea of adding a specific time cache, I'd suggest:
ts = nft_time_get(...);
or similar.
> if (ts != ((time_t) -1) && localtime_r(&ts, &cur_tm))
> seconds = (seconds + cur_tm.tm_gmtoff) % SECONDS_PER_DAY;
>
> @@ -534,7 +534,7 @@ static struct error_record *hour_type_parse(struct parse_ctx *ctx,
> result = 0;
>
> /* Obtain current tm, so that we can substract tm_gmtoff */
> - ts = time(NULL);
> + ts = ops_cache_get_time(ctx->ops_cache);
> if (ts != ((time_t) -1) && localtime_r(&ts, &cur_tm_data))
> cur_tm = &cur_tm_data;
> else
> --
> 2.41.0
>
next prev parent reply other threads:[~2023-08-28 15:03 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-08-25 13:24 [PATCH nft 0/4] add operation cache for timestamp Thomas Haller
2023-08-25 13:24 ` [PATCH nft 1/4] evaluate: add and use parse_ctx_init() helper method Thomas Haller
2023-08-25 13:24 ` [PATCH nft 2/4] src: add ops_cache struct for caching information during parsing Thomas Haller
2023-08-28 15:00 ` Pablo Neira Ayuso
2023-08-25 13:24 ` [PATCH nft 3/4] src: cache result of time() during parsing/output Thomas Haller
2023-08-28 15:02 ` Pablo Neira Ayuso [this message]
2023-08-25 13:24 ` [PATCH nft 4/4] src: cache GMT offset for current time " Thomas Haller
2023-08-29 15:38 ` [PATCH nft 0/4] add operation cache for timestamp Pablo Neira Ayuso
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=ZOy2+ln2nMnveET2@calendula \
--to=pablo@netfilter.org \
--cc=netfilter-devel@vger.kernel.org \
--cc=thaller@redhat.com \
/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).