Extend the meaning of the CTLFLAG_TUN flag to automatically check if

there is an environment variable which shall initialize the SYSCTL
during early boot. This works for all SYSCTL types both statically and
dynamically created ones, except for the SYSCTL NODE type and SYSCTLs
which belong to VNETs. A new flag, CTLFLAG_NOFETCH, has been added to
be used in the case a tunable sysctl has a custom initialisation
function allowing the sysctl to still be marked as a tunable. The
kernel SYSCTL API is mostly the same, with a few exceptions for some
special operations like iterating childrens of a static/extern SYSCTL
node. This operation should probably be made into a factored out
common macro, hence some device drivers use this. The reason for
changing the SYSCTL API was the need for a SYSCTL parent OID pointer
and not only the SYSCTL parent OID list pointer in order to quickly
generate the sysctl path. The motivation behind this patch is to avoid
parameter loading cludges inside the OFED driver subsystem. Instead of
adding special code to the OFED driver subsystem to post-load tunables
into dynamically created sysctls, we generalize this in the kernel.

Other changes:
- Corrected a possibly incorrect sysctl name from "hw.cbb.intr_mask"
to "hw.pcic.intr_mask".
- Removed redundant TUNABLE statements throughout the kernel.
- Some minor code rewrites in connection to removing not needed
TUNABLE statements.
- Added a missing SYSCTL_DECL().
- Wrapped two very long lines.
- Avoid malloc()/free() inside sysctl string handling, in case it is
called to initialize a sysctl from a tunable, hence malloc()/free() is
not ready when sysctls from the sysctl dataset are registered.
- Bumped FreeBSD version to indicate SYSCTL API change.

MFC after:	2 weeks
Sponsored by:	Mellanox Technologies

4 files changed, 22 insertions, 53 deletions

diff --git a/sys/geom/raid/g_raid.c b/sys/geom/raid/g_raid.c
index 858cf4c..db9cf44 100644
--- a/sys/geom/raid/g_raid.c
+++ b/sys/geom/raid/g_raid.c
@@ -53,49 +53,38 @@ static MALLOC_DEFINE(M_RAID, "raid_data", "GEOM_RAID Data");
 SYSCTL_DECL(_kern_geom);
 SYSCTL_NODE(_kern_geom, OID_AUTO, raid, CTLFLAG_RW, 0, "GEOM_RAID stuff");
 int g_raid_enable = 1;
-TUNABLE_INT("kern.geom.raid.enable", &g_raid_enable);
-SYSCTL_INT(_kern_geom_raid, OID_AUTO, enable, CTLFLAG_RW,
+SYSCTL_INT(_kern_geom_raid, OID_AUTO, enable, CTLFLAG_RWTUN,
     &g_raid_enable, 0, "Enable on-disk metadata taste");
 u_int g_raid_aggressive_spare = 0;
-TUNABLE_INT("kern.geom.raid.aggressive_spare", &g_raid_aggressive_spare);
-SYSCTL_UINT(_kern_geom_raid, OID_AUTO, aggressive_spare, CTLFLAG_RW,
+SYSCTL_UINT(_kern_geom_raid, OID_AUTO, aggressive_spare, CTLFLAG_RWTUN,
     &g_raid_aggressive_spare, 0, "Use disks without metadata as spare");
 u_int g_raid_debug = 0;
-TUNABLE_INT("kern.geom.raid.debug", &g_raid_debug);
-SYSCTL_UINT(_kern_geom_raid, OID_AUTO, debug, CTLFLAG_RW, &g_raid_debug, 0,
+SYSCTL_UINT(_kern_geom_raid, OID_AUTO, debug, CTLFLAG_RWTUN, &g_raid_debug, 0,
     "Debug level");
 int g_raid_read_err_thresh = 10;
-TUNABLE_INT("kern.geom.raid.read_err_thresh", &g_raid_read_err_thresh);
-SYSCTL_UINT(_kern_geom_raid, OID_AUTO, read_err_thresh, CTLFLAG_RW,
+SYSCTL_UINT(_kern_geom_raid, OID_AUTO, read_err_thresh, CTLFLAG_RWTUN,
     &g_raid_read_err_thresh, 0,
     "Number of read errors equated to disk failure");
 u_int g_raid_start_timeout = 30;
-TUNABLE_INT("kern.geom.raid.start_timeout", &g_raid_start_timeout);
-SYSCTL_UINT(_kern_geom_raid, OID_AUTO, start_timeout, CTLFLAG_RW,
+SYSCTL_UINT(_kern_geom_raid, OID_AUTO, start_timeout, CTLFLAG_RWTUN,
     &g_raid_start_timeout, 0,
     "Time to wait for all array components");
 static u_int g_raid_clean_time = 5;
-TUNABLE_INT("kern.geom.raid.clean_time", &g_raid_clean_time);
-SYSCTL_UINT(_kern_geom_raid, OID_AUTO, clean_time, CTLFLAG_RW,
+SYSCTL_UINT(_kern_geom_raid, OID_AUTO, clean_time, CTLFLAG_RWTUN,
     &g_raid_clean_time, 0, "Mark volume as clean when idling");
 static u_int g_raid_disconnect_on_failure = 1;
-TUNABLE_INT("kern.geom.raid.disconnect_on_failure",
-    &g_raid_disconnect_on_failure);
-SYSCTL_UINT(_kern_geom_raid, OID_AUTO, disconnect_on_failure, CTLFLAG_RW,
+SYSCTL_UINT(_kern_geom_raid, OID_AUTO, disconnect_on_failure, CTLFLAG_RWTUN,
     &g_raid_disconnect_on_failure, 0, "Disconnect component on I/O failure.");
 static u_int g_raid_name_format = 0;
-TUNABLE_INT("kern.geom.raid.name_format", &g_raid_name_format);
-SYSCTL_UINT(_kern_geom_raid, OID_AUTO, name_format, CTLFLAG_RW,
+SYSCTL_UINT(_kern_geom_raid, OID_AUTO, name_format, CTLFLAG_RWTUN,
     &g_raid_name_format, 0, "Providers name format.");
 static u_int g_raid_idle_threshold = 1000000;
-TUNABLE_INT("kern.geom.raid.idle_threshold", &g_raid_idle_threshold);
-SYSCTL_UINT(_kern_geom_raid, OID_AUTO, idle_threshold, CTLFLAG_RW,
+SYSCTL_UINT(_kern_geom_raid, OID_AUTO, idle_threshold, CTLFLAG_RWTUN,
     &g_raid_idle_threshold, 1000000,
     "Time in microseconds to consider a volume idle.");
 static u_int ar_legacy_aliases = 1;
-SYSCTL_INT(_kern_geom_raid, OID_AUTO, legacy_aliases, CTLFLAG_RW,
+SYSCTL_INT(_kern_geom_raid, OID_AUTO, legacy_aliases, CTLFLAG_RWTUN,
            &ar_legacy_aliases, 0, "Create aliases named as the legacy ataraid style.");
-TUNABLE_INT("kern.geom_raid.legacy_aliases", &ar_legacy_aliases);
 
 
 #define	MSLEEP(rv, ident, mtx, priority, wmesg, timeout)	do {	\
diff --git a/sys/geom/raid/g_raid.h b/sys/geom/raid/g_raid.h
index 8a96fa9..f6aa6db 100644
--- a/sys/geom/raid/g_raid.h
+++ b/sys/geom/raid/g_raid.h
@@ -364,10 +364,8 @@ int g_raid_md_modevent(module_t, int, void *);
     SYSCTL_NODE(_kern_geom_raid, OID_AUTO, name, CTLFLAG_RD,	\
 	NULL, label " metadata module");			\
     SYSCTL_INT(_kern_geom_raid_##name, OID_AUTO, enable,	\
-	CTLFLAG_RW, &g_raid_md_##name##_class.mdc_enable, 0,	\
-	"Enable " label " metadata format taste");		\
-    TUNABLE_INT("kern.geom.raid." __XSTRING(name) ".enable",	\
-	&g_raid_md_##name##_class.mdc_enable)
+	CTLFLAG_RWTUN, &g_raid_md_##name##_class.mdc_enable, 0,	\
+	"Enable " label " metadata format taste")
 
 /*
  * KOBJ parent class of data transformation modules.
@@ -403,10 +401,8 @@ int g_raid_tr_modevent(module_t, int, void *);
     SYSCTL_NODE(_kern_geom_raid, OID_AUTO, name, CTLFLAG_RD,	\
 	NULL, label " transformation module");			\
     SYSCTL_INT(_kern_geom_raid_##name, OID_AUTO, enable,	\
-	CTLFLAG_RW, &g_raid_tr_##name##_class.trc_enable, 0,	\
-	"Enable " label " transformation module taste");	\
-    TUNABLE_INT("kern.geom.raid." __XSTRING(name) ".enable",	\
-	&g_raid_tr_##name##_class.trc_enable)
+	CTLFLAG_RWTUN, &g_raid_tr_##name##_class.trc_enable, 0,	\
+	"Enable " label " transformation module taste")
 
 const char * g_raid_volume_level2str(int level, int qual);
 int g_raid_volume_str2level(const char *str, int *level, int *qual);
diff --git a/sys/geom/raid/tr_raid1.c b/sys/geom/raid/tr_raid1.c
index 833655b..17c7a25 100644
--- a/sys/geom/raid/tr_raid1.c
+++ b/sys/geom/raid/tr_raid1.c
@@ -46,33 +46,25 @@ SYSCTL_DECL(_kern_geom_raid_raid1);
 
 #define RAID1_REBUILD_SLAB	(1 << 20) /* One transation in a rebuild */
 static int g_raid1_rebuild_slab = RAID1_REBUILD_SLAB;
-TUNABLE_INT("kern.geom.raid.raid1.rebuild_slab_size",
-    &g_raid1_rebuild_slab);
-SYSCTL_UINT(_kern_geom_raid_raid1, OID_AUTO, rebuild_slab_size, CTLFLAG_RW,
+SYSCTL_UINT(_kern_geom_raid_raid1, OID_AUTO, rebuild_slab_size, CTLFLAG_RWTUN,
     &g_raid1_rebuild_slab, 0,
     "Amount of the disk to rebuild each read/write cycle of the rebuild.");
 
 #define RAID1_REBUILD_FAIR_IO 20 /* use 1/x of the available I/O */
 static int g_raid1_rebuild_fair_io = RAID1_REBUILD_FAIR_IO;
-TUNABLE_INT("kern.geom.raid.raid1.rebuild_fair_io",
-    &g_raid1_rebuild_fair_io);
-SYSCTL_UINT(_kern_geom_raid_raid1, OID_AUTO, rebuild_fair_io, CTLFLAG_RW,
+SYSCTL_UINT(_kern_geom_raid_raid1, OID_AUTO, rebuild_fair_io, CTLFLAG_RWTUN,
     &g_raid1_rebuild_fair_io, 0,
     "Fraction of the I/O bandwidth to use when disk busy for rebuild.");
 
 #define RAID1_REBUILD_CLUSTER_IDLE 100
 static int g_raid1_rebuild_cluster_idle = RAID1_REBUILD_CLUSTER_IDLE;
-TUNABLE_INT("kern.geom.raid.raid1.rebuild_cluster_idle",
-    &g_raid1_rebuild_cluster_idle);
-SYSCTL_UINT(_kern_geom_raid_raid1, OID_AUTO, rebuild_cluster_idle, CTLFLAG_RW,
+SYSCTL_UINT(_kern_geom_raid_raid1, OID_AUTO, rebuild_cluster_idle, CTLFLAG_RWTUN,
     &g_raid1_rebuild_cluster_idle, 0,
     "Number of slabs to do each time we trigger a rebuild cycle");
 
 #define RAID1_REBUILD_META_UPDATE 1024 /* update meta data every 1GB or so */
 static int g_raid1_rebuild_meta_update = RAID1_REBUILD_META_UPDATE;
-TUNABLE_INT("kern.geom.raid.raid1.rebuild_meta_update",
-    &g_raid1_rebuild_meta_update);
-SYSCTL_UINT(_kern_geom_raid_raid1, OID_AUTO, rebuild_meta_update, CTLFLAG_RW,
+SYSCTL_UINT(_kern_geom_raid_raid1, OID_AUTO, rebuild_meta_update, CTLFLAG_RWTUN,
     &g_raid1_rebuild_meta_update, 0,
     "When to update the meta data.");
 
diff --git a/sys/geom/raid/tr_raid1e.c b/sys/geom/raid/tr_raid1e.c
index 404b9e6..b357742 100644
--- a/sys/geom/raid/tr_raid1e.c
+++ b/sys/geom/raid/tr_raid1e.c
@@ -48,33 +48,25 @@ SYSCTL_DECL(_kern_geom_raid_raid1e);
 
 #define RAID1E_REBUILD_SLAB	(1 << 20) /* One transation in a rebuild */
 static int g_raid1e_rebuild_slab = RAID1E_REBUILD_SLAB;
-TUNABLE_INT("kern.geom.raid.raid1e.rebuild_slab_size",
-    &g_raid1e_rebuild_slab);
-SYSCTL_UINT(_kern_geom_raid_raid1e, OID_AUTO, rebuild_slab_size, CTLFLAG_RW,
+SYSCTL_UINT(_kern_geom_raid_raid1e, OID_AUTO, rebuild_slab_size, CTLFLAG_RWTUN,
     &g_raid1e_rebuild_slab, 0,
     "Amount of the disk to rebuild each read/write cycle of the rebuild.");
 
 #define RAID1E_REBUILD_FAIR_IO 20 /* use 1/x of the available I/O */
 static int g_raid1e_rebuild_fair_io = RAID1E_REBUILD_FAIR_IO;
-TUNABLE_INT("kern.geom.raid.raid1e.rebuild_fair_io",
-    &g_raid1e_rebuild_fair_io);
-SYSCTL_UINT(_kern_geom_raid_raid1e, OID_AUTO, rebuild_fair_io, CTLFLAG_RW,
+SYSCTL_UINT(_kern_geom_raid_raid1e, OID_AUTO, rebuild_fair_io, CTLFLAG_RWTUN,
     &g_raid1e_rebuild_fair_io, 0,
     "Fraction of the I/O bandwidth to use when disk busy for rebuild.");
 
 #define RAID1E_REBUILD_CLUSTER_IDLE 100
 static int g_raid1e_rebuild_cluster_idle = RAID1E_REBUILD_CLUSTER_IDLE;
-TUNABLE_INT("kern.geom.raid.raid1e.rebuild_cluster_idle",
-    &g_raid1e_rebuild_cluster_idle);
-SYSCTL_UINT(_kern_geom_raid_raid1e, OID_AUTO, rebuild_cluster_idle, CTLFLAG_RW,
+SYSCTL_UINT(_kern_geom_raid_raid1e, OID_AUTO, rebuild_cluster_idle, CTLFLAG_RWTUN,
     &g_raid1e_rebuild_cluster_idle, 0,
     "Number of slabs to do each time we trigger a rebuild cycle");
 
 #define RAID1E_REBUILD_META_UPDATE 1024 /* update meta data every 1GB or so */
 static int g_raid1e_rebuild_meta_update = RAID1E_REBUILD_META_UPDATE;
-TUNABLE_INT("kern.geom.raid.raid1e.rebuild_meta_update",
-    &g_raid1e_rebuild_meta_update);
-SYSCTL_UINT(_kern_geom_raid_raid1e, OID_AUTO, rebuild_meta_update, CTLFLAG_RW,
+SYSCTL_UINT(_kern_geom_raid_raid1e, OID_AUTO, rebuild_meta_update, CTLFLAG_RWTUN,
     &g_raid1e_rebuild_meta_update, 0,
     "When to update the meta data.");