git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v2] Prefer sysconf(_SC_OPEN_MAX) over getrlimit(RLIMIT_NOFILE,...)
@ 2012-08-24  9:52 Joachim Schmitz
  2012-08-24 16:43 ` Junio C Hamano
  0 siblings, 1 reply; 3+ messages in thread
From: Joachim Schmitz @ 2012-08-24  9:52 UTC (permalink / raw)
  To: 'Junio C Hamano'; +Cc: git


Signed-off-by: Joachim Schmitz <jojo@schmitz-digital.de>
---
As discussed now as a small helper function rather than #ifdef/#endif in the primary flow of the code.
And hopefully without having screwed up whitespace and line breaks

 sha1_file.c | 22 +++++++++++++++-------
 1 file changed, 15 insertions(+), 7 deletions(-)

diff --git a/sha1_file.c b/sha1_file.c
index af5cfbd..427f9e6 100644
--- a/sha1_file.c
+++ b/sha1_file.c
@@ -731,6 +731,20 @@ void free_pack_by_name(const char *pack_name)
 	}
 }
 
+static unsigned int get_max_fd_limit(void)
+{
+#ifdef _SC_OPEN_MAX
+	return sysconf(_SC_OPEN_MAX);
+#else
+	struct rlimit lim;
+
+	if (getrlimit(RLIMIT_NOFILE, &lim))
+		die_errno("cannot get RLIMIT_NOFILE");
+
+	return lim.rlim_cur;
+#endif
+}
+
 /*
  * Do not call this directly as this leaks p->pack_fd on error return;
  * call open_packed_git() instead.
@@ -747,13 +761,7 @@ static int open_packed_git_1(struct packed_git *p)
 		return error("packfile %s index unavailable", p->pack_name);
 
 	if (!pack_max_fds) {
-		struct rlimit lim;
-		unsigned int max_fds;
-
-		if (getrlimit(RLIMIT_NOFILE, &lim))
-			die_errno("cannot get RLIMIT_NOFILE");
-
-		max_fds = lim.rlim_cur;
+		unsigned int max_fds = get_max_fd_limit();
 
 		/* Save 3 for stdin/stdout/stderr, 22 for work */
 		if (25 < max_fds)
-- 
1.7.12

^ permalink raw reply related	[flat|nested] 3+ messages in thread

* Re: [PATCH v2] Prefer sysconf(_SC_OPEN_MAX) over getrlimit(RLIMIT_NOFILE,...)
  2012-08-24  9:52 [PATCH v2] Prefer sysconf(_SC_OPEN_MAX) over getrlimit(RLIMIT_NOFILE,...) Joachim Schmitz
@ 2012-08-24 16:43 ` Junio C Hamano
  2012-08-24 18:36   ` Joachim Schmitz
  0 siblings, 1 reply; 3+ messages in thread
From: Junio C Hamano @ 2012-08-24 16:43 UTC (permalink / raw)
  To: Joachim Schmitz; +Cc: git

"Joachim Schmitz" <jojo@schmitz-digital.de> writes:

> Signed-off-by: Joachim Schmitz <jojo@schmitz-digital.de>
> ---
> As discussed now as a small helper function rather than #ifdef/#endif in the primary flow of the code.
> And hopefully without having screwed up whitespace and line breaks

The formatting looks fine.

Perhaps I am being overly paranoid, but I would prefer not to change
things for people who have been using getrlimit().  For them, if
they also have sysconf(_SC_OPEN_MAX), your code _ought to_ work, but
if it does not work for whatever reason (perhaps some platforms
claim to have both, but getrlimit() works and sysconf(_SC_OPEN_MAX)
is broken), it will given them an unnecessary regression.

So how about doing it this way instead?

-- >8 --
Subject: sha1_file.c: introduce get_max_fd_limit() helper

Not all platforms have getrlimit(), but there are other ways to see
the maximum number of files that a process can have open.  If
getrlimit() is unavailable, fall back to sysconf(_SC_OPEN_MAX) if
available, and use OPEN_MAX from <limits.h>.

Signed-off-by: Joachim Schmitz <jojo@schmitz-digital.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
---
 sha1_file.c | 26 +++++++++++++++++++-------
 1 file changed, 19 insertions(+), 7 deletions(-)

diff --git c/sha1_file.c w/sha1_file.c
index af5cfbd..9152974 100644
--- c/sha1_file.c
+++ w/sha1_file.c
@@ -731,6 +731,24 @@ void free_pack_by_name(const char *pack_name)
 	}
 }
 
+static unsigned int get_max_fd_limit(void)
+{
+#ifdef RLIMIT_NOFILE
+	struct rlimit lim;
+
+	if (getrlimit(RLIMIT_NOFILE, &lim))
+		die_errno("cannot get RLIMIT_NOFILE");
+
+	return lim.rlim_cur;
+#elif defined(_SC_OPEN_MAX)
+	return sysconf(_SC_OPEN_MAX);
+#elif defined(OPEN_MAX)
+	return OPEN_MAX;
+#else
+	return 1; /* see the caller ;-) */
+#endif
+}
+
 /*
  * Do not call this directly as this leaks p->pack_fd on error return;
  * call open_packed_git() instead.
@@ -747,13 +765,7 @@ static int open_packed_git_1(struct packed_git *p)
 		return error("packfile %s index unavailable", p->pack_name);
 
 	if (!pack_max_fds) {
-		struct rlimit lim;
-		unsigned int max_fds;
-
-		if (getrlimit(RLIMIT_NOFILE, &lim))
-			die_errno("cannot get RLIMIT_NOFILE");
-
-		max_fds = lim.rlim_cur;
+		unsigned int max_fds = get_max_fd_limit();
 
 		/* Save 3 for stdin/stdout/stderr, 22 for work */
 		if (25 < max_fds)

^ permalink raw reply related	[flat|nested] 3+ messages in thread

* RE: [PATCH v2] Prefer sysconf(_SC_OPEN_MAX) over getrlimit(RLIMIT_NOFILE,...)
  2012-08-24 16:43 ` Junio C Hamano
@ 2012-08-24 18:36   ` Joachim Schmitz
  0 siblings, 0 replies; 3+ messages in thread
From: Joachim Schmitz @ 2012-08-24 18:36 UTC (permalink / raw)
  To: 'Junio C Hamano'; +Cc: git

> From: Junio C Hamano [mailto:gitster@pobox.com]
> Sent: Friday, August 24, 2012 6:44 PM
> To: Joachim Schmitz
> Cc: git@vger.kernel.org
> Subject: Re: [PATCH v2] Prefer sysconf(_SC_OPEN_MAX) over getrlimit(RLIMIT_NOFILE,...)
> 
> "Joachim Schmitz" <jojo@schmitz-digital.de> writes:
> 
> > Signed-off-by: Joachim Schmitz <jojo@schmitz-digital.de>
> > ---
> > As discussed now as a small helper function rather than #ifdef/#endif in the primary flow of the code.
> > And hopefully without having screwed up whitespace and line breaks
> 
> The formatting looks fine.
> 
> Perhaps I am being overly paranoid, but I would prefer not to change
> things for people who have been using getrlimit().  For them, if
> they also have sysconf(_SC_OPEN_MAX), your code _ought to_ work, but
> if it does not work for whatever reason (perhaps some platforms
> claim to have both, but getrlimit() works and sysconf(_SC_OPEN_MAX)
> is broken), it will given them an unnecessary regression.

Sounds reasonable, so reasonable that I wonder why I didn't have that idea ;-)

> So how about doing it this way instead?
> 
> -- >8 --
> Subject: sha1_file.c: introduce get_max_fd_limit() helper
> 
> Not all platforms have getrlimit(), but there are other ways to see
> the maximum number of files that a process can have open.  If
> getrlimit() is unavailable, fall back to sysconf(_SC_OPEN_MAX) if
> available, and use OPEN_MAX from <limits.h>.
> 
> Signed-off-by: Joachim Schmitz <jojo@schmitz-digital.de>
> Signed-off-by: Junio C Hamano <gitster@pobox.com>
> ---
>  sha1_file.c | 26 +++++++++++++++++++-------
>  1 file changed, 19 insertions(+), 7 deletions(-)
> 
> diff --git c/sha1_file.c w/sha1_file.c
> index af5cfbd..9152974 100644
> --- c/sha1_file.c
> +++ w/sha1_file.c
> @@ -731,6 +731,24 @@ void free_pack_by_name(const char *pack_name)
>  	}
>  }
> 
> +static unsigned int get_max_fd_limit(void)
> +{
> +#ifdef RLIMIT_NOFILE
> +	struct rlimit lim;
> +
> +	if (getrlimit(RLIMIT_NOFILE, &lim))
> +		die_errno("cannot get RLIMIT_NOFILE");
> +
> +	return lim.rlim_cur;
> +#elif defined(_SC_OPEN_MAX)
> +	return sysconf(_SC_OPEN_MAX);
> +#elif defined(OPEN_MAX)
> +	return OPEN_MAX;
> +#else
> +	return 1; /* see the caller ;-) */
> +#endif
> +}
> +
>  /*
>   * Do not call this directly as this leaks p->pack_fd on error return;
>   * call open_packed_git() instead.
> @@ -747,13 +765,7 @@ static int open_packed_git_1(struct packed_git *p)
>  		return error("packfile %s index unavailable", p->pack_name);
> 
>  	if (!pack_max_fds) {
> -		struct rlimit lim;
> -		unsigned int max_fds;
> -
> -		if (getrlimit(RLIMIT_NOFILE, &lim))
> -			die_errno("cannot get RLIMIT_NOFILE");
> -
> -		max_fds = lim.rlim_cur;
> +		unsigned int max_fds = get_max_fd_limit();
> 
>  		/* Save 3 for stdin/stdout/stderr, 22 for work */
>  		if (25 < max_fds)

Looks good to me. 
Stupid newbie question: how would I revert my commit to my clone, to then add (and test) this one?

Bye, Jojo

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2012-08-24 18:36 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2012-08-24  9:52 [PATCH v2] Prefer sysconf(_SC_OPEN_MAX) over getrlimit(RLIMIT_NOFILE,...) Joachim Schmitz
2012-08-24 16:43 ` Junio C Hamano
2012-08-24 18:36   ` Joachim Schmitz

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).