Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 9 Feb 2016 10:31:16 +0000 (UTC)
From:      Guido Falsi <madpilot@FreeBSD.org>
To:        ports-committers@freebsd.org, svn-ports-all@freebsd.org, svn-ports-head@freebsd.org
Subject:   svn commit: r408548 - in head/net-mgmt/collectd5: . files
Message-ID:  <201602091031.u19AVG0u075892@repo.freebsd.org>

next in thread | raw e-mail | index | archive | help
Author: madpilot
Date: Tue Feb  9 10:31:16 2016
New Revision: 408548
URL: https://svnweb.freebsd.org/changeset/ports/408548

Log:
  Add back part of a patch removed by mistake.
  
  Allows disk module to report all disk statistics again.
  
  PR:		207049
  Submitted by:	me
  Reported by:	Pierre Guinoiseau <pierre at guinoiseau.eu>
  Approved by:	Krzysztof <ports at bsdserwis.com>

Modified:
  head/net-mgmt/collectd5/Makefile
  head/net-mgmt/collectd5/files/patch-src_disk.c

Modified: head/net-mgmt/collectd5/Makefile
==============================================================================
--- head/net-mgmt/collectd5/Makefile	Tue Feb  9 10:19:00 2016	(r408547)
+++ head/net-mgmt/collectd5/Makefile	Tue Feb  9 10:31:16 2016	(r408548)
@@ -3,6 +3,7 @@
 
 PORTNAME=	collectd
 PORTVERSION=	5.5.1
+PORTREVISION=	1
 CATEGORIES=	net-mgmt
 MASTER_SITES=	https://collectd.org/files/ \
 		http://collectd.org/files/

Modified: head/net-mgmt/collectd5/files/patch-src_disk.c
==============================================================================
--- head/net-mgmt/collectd5/files/patch-src_disk.c	Tue Feb  9 10:19:00 2016	(r408547)
+++ head/net-mgmt/collectd5/files/patch-src_disk.c	Tue Feb  9 10:31:16 2016	(r408548)
@@ -43,3 +43,118 @@
  #elif HAVE_LIBKSTAT
  	kstat_t *ksp_chain;
  
+@@ -505,6 +527,114 @@ static int disk_read (void)
+ 	IOObjectRelease (disk_list);
+ /* #endif HAVE_IOKIT_IOKITLIB_H */
+ 
++#elif KERNEL_FREEBSD
++	int retry, dirty;
++
++	void *snap = NULL;
++	struct devstat *snap_iter;
++
++	struct gident *geom_id;
++
++	const char *disk_name;
++	long double read_time, write_time;
++
++	for (retry = 0, dirty = 1; retry < 5 && dirty == 1; retry++) {
++		if (snap != NULL)
++			geom_stats_snapshot_free(snap);
++
++		/* Get a fresh copy of stats snapshot */
++		snap = geom_stats_snapshot_get();
++		if (snap == NULL) {
++			ERROR("disk plugin: geom_stats_snapshot_get() failed.");
++			return (-1);
++		}
++
++		/* Check if we have dirty read from this snapshot */
++		dirty = 0;
++		geom_stats_snapshot_reset(snap);
++		while ((snap_iter = geom_stats_snapshot_next(snap)) != NULL) {
++			if (snap_iter->id == NULL)
++				continue;
++			geom_id = geom_lookupid(&geom_tree, snap_iter->id);
++
++			/* New device? refresh GEOM tree */
++			if (geom_id == NULL) {
++				geom_deletetree(&geom_tree);
++				if (geom_gettree(&geom_tree) != 0) {
++					ERROR("disk plugin: geom_gettree() failed");
++					geom_stats_snapshot_free(snap);
++					return (-1);
++				}
++				geom_id = geom_lookupid(&geom_tree, snap_iter->id);
++			}
++			/*
++			 * This should be rare: the device come right before we take the
++			 * snapshot and went away right after it.  We will handle this
++			 * case later, so don't mark dirty but silently ignore it.
++			 */
++			if (geom_id == NULL)
++				continue;
++
++			/* Only collect PROVIDER data */
++			if (geom_id->lg_what != ISPROVIDER)
++				continue;
++
++			/* Only collect data when rank is 1 (physical devices) */
++			if (((struct gprovider *)(geom_id->lg_ptr))->lg_geom->lg_rank != 1)
++				continue;
++
++			/* Check if this is a dirty read quit for another try */
++			if (snap_iter->sequence0 != snap_iter->sequence1) {
++				dirty = 1;
++				break;
++			}
++		}
++	}
++
++	/* Reset iterator */
++	geom_stats_snapshot_reset(snap);
++	for (;;) {
++		snap_iter = geom_stats_snapshot_next(snap);
++		if (snap_iter == NULL)
++			break;
++
++		if (snap_iter->id == NULL)
++			continue;
++		geom_id = geom_lookupid(&geom_tree, snap_iter->id);
++		if (geom_id == NULL)
++			continue;
++		if (geom_id->lg_what != ISPROVIDER)
++			continue;
++		if (((struct gprovider *)(geom_id->lg_ptr))->lg_geom->lg_rank != 1)
++			continue;
++		/* Skip dirty reads, if present */
++		if (dirty && (snap_iter->sequence0 != snap_iter->sequence1))
++			continue;
++
++		disk_name = ((struct gprovider *)geom_id->lg_ptr)->lg_name;
++
++		if ((snap_iter->bytes[1] != 0) || (snap_iter->bytes[2] != 0)) {
++			disk_submit(disk_name, "disk_octets",
++					(derive_t)snap_iter->bytes[1],
++					(derive_t)snap_iter->bytes[2]);
++		}
++
++		if ((snap_iter->operations[1] != 0) || (snap_iter->operations[2] != 0)) {
++			disk_submit(disk_name, "disk_ops",
++					(derive_t)snap_iter->operations[1],
++					(derive_t)snap_iter->operations[2]);
++		}
++
++		read_time = devstat_compute_etime(&snap_iter->duration[DEVSTAT_READ], NULL);
++		write_time = devstat_compute_etime(&snap_iter->duration[DEVSTAT_WRITE], NULL);
++		if ((read_time != 0) || (write_time != 0)) {
++			disk_submit (disk_name, "disk_time",
++					(derive_t)(read_time*1000), (derive_t)(write_time*1000));
++		}
++	}
++	geom_stats_snapshot_free(snap);
++/* #endif KERNEL_FREEBSD */
++
+ #elif KERNEL_LINUX
+ 	FILE *fh;
+ 	char buffer[1024];



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201602091031.u19AVG0u075892>