freebsd-dev/sys/geom/concat/g_concat.c

1273 lines
31 KiB
C
Raw Normal View History

/*-
* SPDX-License-Identifier: BSD-2-Clause-FreeBSD
*
* Copyright (c) 2004-2005 Pawel Jakub Dawidek <pjd@FreeBSD.org>
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
2006-02-01 12:06:01 +00:00
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHORS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/module.h>
#include <sys/lock.h>
#include <sys/mutex.h>
#include <sys/sx.h>
#include <sys/bio.h>
#include <sys/sbuf.h>
#include <sys/sysctl.h>
#include <sys/malloc.h>
#include <geom/geom.h>
#include <geom/geom_dbg.h>
#include <geom/concat/g_concat.h>
FEATURE(geom_concat, "GEOM concatenation support");
static MALLOC_DEFINE(M_CONCAT, "concat_data", "GEOM_CONCAT Data");
SYSCTL_DECL(_kern_geom);
static SYSCTL_NODE(_kern_geom, OID_AUTO, concat, CTLFLAG_RW | CTLFLAG_MPSAFE, 0,
"GEOM_CONCAT stuff");
static u_int g_concat_debug = 0;
SYSCTL_UINT(_kern_geom_concat, OID_AUTO, debug, CTLFLAG_RWTUN, &g_concat_debug, 0,
"Debug level");
static int g_concat_destroy(struct g_concat_softc *sc, boolean_t force);
static int g_concat_destroy_geom(struct gctl_req *req, struct g_class *mp,
struct g_geom *gp);
static g_taste_t g_concat_taste;
static g_ctl_req_t g_concat_config;
static g_dumpconf_t g_concat_dumpconf;
struct g_class g_concat_class = {
.name = G_CONCAT_CLASS_NAME,
.version = G_VERSION,
.ctlreq = g_concat_config,
.taste = g_concat_taste,
.destroy_geom = g_concat_destroy_geom
};
/*
* Greatest Common Divisor.
*/
static u_int
gcd(u_int a, u_int b)
{
u_int c;
while (b != 0) {
c = a;
a = b;
b = (c % b);
}
return (a);
}
/*
* Least Common Multiple.
*/
static u_int
lcm(u_int a, u_int b)
{
return ((a * b) / gcd(a, b));
}
/*
* Return the number of valid disks.
*/
static u_int
g_concat_nvalid(struct g_concat_softc *sc)
{
u_int no;
struct g_concat_disk *disk;
sx_assert(&sc->sc_disks_lock, SA_LOCKED);
no = 0;
TAILQ_FOREACH(disk, &sc->sc_disks, d_next) {
if (disk->d_consumer != NULL)
no++;
}
return (no);
}
static void
g_concat_remove_disk(struct g_concat_disk *disk)
{
struct g_consumer *cp;
struct g_concat_softc *sc;
g_topology_assert();
KASSERT(disk->d_consumer != NULL, ("Non-valid disk in %s.", __func__));
sc = disk->d_softc;
cp = disk->d_consumer;
if (!disk->d_removed) {
G_CONCAT_DEBUG(0, "Disk %s removed from %s.",
cp->provider->name, sc->sc_name);
disk->d_removed = 1;
}
if (sc->sc_provider != NULL) {
G_CONCAT_DEBUG(0, "Device %s deactivated.",
sc->sc_provider->name);
g_wither_provider(sc->sc_provider, ENXIO);
sc->sc_provider = NULL;
}
if (cp->acr > 0 || cp->acw > 0 || cp->ace > 0)
return;
disk->d_consumer = NULL;
g_detach(cp);
g_destroy_consumer(cp);
/* If there are no valid disks anymore, remove device. */
if (LIST_EMPTY(&sc->sc_geom->consumer))
g_concat_destroy(sc, 1);
}
static void
g_concat_orphan(struct g_consumer *cp)
{
struct g_concat_softc *sc;
struct g_concat_disk *disk;
struct g_geom *gp;
g_topology_assert();
gp = cp->geom;
sc = gp->softc;
if (sc == NULL)
return;
disk = cp->private;
if (disk == NULL) /* Possible? */
return;
g_concat_remove_disk(disk);
}
static int
g_concat_access(struct g_provider *pp, int dr, int dw, int de)
{
struct g_consumer *cp1, *cp2, *tmp;
struct g_concat_disk *disk;
struct g_geom *gp;
struct g_concat_softc *sc;
int error;
g_topology_assert();
gp = pp->geom;
sc = gp->softc;
/* On first open, grab an extra "exclusive" bit */
if (pp->acr == 0 && pp->acw == 0 && pp->ace == 0)
de++;
/* ... and let go of it on last close */
2004-07-18 06:54:29 +00:00
if ((pp->acr + dr) == 0 && (pp->acw + dw) == 0 && (pp->ace + de) == 0)
de--;
sx_slock(&sc->sc_disks_lock);
LIST_FOREACH_SAFE(cp1, &gp->consumer, consumer, tmp) {
error = g_access(cp1, dr, dw, de);
if (error != 0)
goto fail;
disk = cp1->private;
if (cp1->acr == 0 && cp1->acw == 0 && cp1->ace == 0 &&
disk->d_removed) {
g_concat_remove_disk(disk); /* May destroy geom. */
}
}
sx_sunlock(&sc->sc_disks_lock);
return (0);
fail:
sx_sunlock(&sc->sc_disks_lock);
LIST_FOREACH(cp2, &gp->consumer, consumer) {
if (cp1 == cp2)
break;
g_access(cp2, -dr, -dw, -de);
}
return (error);
}
static void
g_concat_candelete(struct bio *bp)
{
struct g_concat_softc *sc;
struct g_concat_disk *disk;
int val;
sc = bp->bio_to->geom->softc;
sx_assert(&sc->sc_disks_lock, SX_LOCKED);
TAILQ_FOREACH(disk, &sc->sc_disks, d_next) {
if (!disk->d_removed && disk->d_candelete)
break;
}
val = disk != NULL;
g_handleattr(bp, "GEOM::candelete", &val, sizeof(val));
}
static void
g_concat_kernel_dump(struct bio *bp)
{
struct g_concat_softc *sc;
struct g_concat_disk *disk;
struct bio *cbp;
struct g_kerneldump *gkd;
sc = bp->bio_to->geom->softc;
gkd = (struct g_kerneldump *)bp->bio_data;
TAILQ_FOREACH(disk, &sc->sc_disks, d_next) {
if (disk->d_start <= gkd->offset &&
disk->d_end > gkd->offset)
break;
}
if (disk == NULL) {
g_io_deliver(bp, EOPNOTSUPP);
return;
}
gkd->offset -= disk->d_start;
if (gkd->length > disk->d_end - disk->d_start - gkd->offset)
gkd->length = disk->d_end - disk->d_start - gkd->offset;
cbp = g_clone_bio(bp);
if (cbp == NULL) {
g_io_deliver(bp, ENOMEM);
return;
}
cbp->bio_done = g_std_done;
g_io_request(cbp, disk->d_consumer);
G_CONCAT_DEBUG(1, "Kernel dump will go to %s.",
disk->d_consumer->provider->name);
}
Merge GEOM direct dispatch changes from the projects/camlock branch. When safety requirements are met, it allows to avoid passing I/O requests to GEOM g_up/g_down thread, executing them directly in the caller context. That allows to avoid CPU bottlenecks in g_up/g_down threads, plus avoid several context switches per I/O. The defined now safety requirements are: - caller should not hold any locks and should be reenterable; - callee should not depend on GEOM dual-threaded concurency semantics; - on the way down, if request is unmapped while callee doesn't support it, the context should be sleepable; - kernel thread stack usage should be below 50%. To keep compatibility with GEOM classes not meeting above requirements new provider and consumer flags added: - G_CF_DIRECT_SEND -- consumer code meets caller requirements (request); - G_CF_DIRECT_RECEIVE -- consumer code meets callee requirements (done); - G_PF_DIRECT_SEND -- provider code meets caller requirements (done); - G_PF_DIRECT_RECEIVE -- provider code meets callee requirements (request). Capable GEOM class can set them, allowing direct dispatch in cases where it is safe. If any of requirements are not met, request is queued to g_up or g_down thread same as before. Such GEOM classes were reviewed and updated to support direct dispatch: CONCAT, DEV, DISK, GATE, MD, MIRROR, MULTIPATH, NOP, PART, RAID, STRIPE, VFS, ZERO, ZFS::VDEV, ZFS::ZVOL, all classes based on g_slice KPI (LABEL, MAP, FLASHMAP, etc). To declare direct completion capability disk(9) KPI got new flag equivalent to G_PF_DIRECT_SEND -- DISKFLAG_DIRECT_COMPLETION. da(4) and ada(4) disk drivers got it set now thanks to earlier CAM locking work. This change more then twice increases peak block storage performance on systems with manu CPUs, together with earlier CAM locking changes reaching more then 1 million IOPS (512 byte raw reads from 16 SATA SSDs on 4 HBAs to 256 user-level threads). Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-22 08:22:19 +00:00
static void
g_concat_done(struct bio *bp)
{
struct g_concat_softc *sc;
struct bio *pbp;
pbp = bp->bio_parent;
sc = pbp->bio_to->geom->softc;
mtx_lock(&sc->sc_completion_lock);
Merge GEOM direct dispatch changes from the projects/camlock branch. When safety requirements are met, it allows to avoid passing I/O requests to GEOM g_up/g_down thread, executing them directly in the caller context. That allows to avoid CPU bottlenecks in g_up/g_down threads, plus avoid several context switches per I/O. The defined now safety requirements are: - caller should not hold any locks and should be reenterable; - callee should not depend on GEOM dual-threaded concurency semantics; - on the way down, if request is unmapped while callee doesn't support it, the context should be sleepable; - kernel thread stack usage should be below 50%. To keep compatibility with GEOM classes not meeting above requirements new provider and consumer flags added: - G_CF_DIRECT_SEND -- consumer code meets caller requirements (request); - G_CF_DIRECT_RECEIVE -- consumer code meets callee requirements (done); - G_PF_DIRECT_SEND -- provider code meets caller requirements (done); - G_PF_DIRECT_RECEIVE -- provider code meets callee requirements (request). Capable GEOM class can set them, allowing direct dispatch in cases where it is safe. If any of requirements are not met, request is queued to g_up or g_down thread same as before. Such GEOM classes were reviewed and updated to support direct dispatch: CONCAT, DEV, DISK, GATE, MD, MIRROR, MULTIPATH, NOP, PART, RAID, STRIPE, VFS, ZERO, ZFS::VDEV, ZFS::ZVOL, all classes based on g_slice KPI (LABEL, MAP, FLASHMAP, etc). To declare direct completion capability disk(9) KPI got new flag equivalent to G_PF_DIRECT_SEND -- DISKFLAG_DIRECT_COMPLETION. da(4) and ada(4) disk drivers got it set now thanks to earlier CAM locking work. This change more then twice increases peak block storage performance on systems with manu CPUs, together with earlier CAM locking changes reaching more then 1 million IOPS (512 byte raw reads from 16 SATA SSDs on 4 HBAs to 256 user-level threads). Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-22 08:22:19 +00:00
if (pbp->bio_error == 0)
pbp->bio_error = bp->bio_error;
pbp->bio_completed += bp->bio_completed;
pbp->bio_inbed++;
if (pbp->bio_children == pbp->bio_inbed) {
mtx_unlock(&sc->sc_completion_lock);
Merge GEOM direct dispatch changes from the projects/camlock branch. When safety requirements are met, it allows to avoid passing I/O requests to GEOM g_up/g_down thread, executing them directly in the caller context. That allows to avoid CPU bottlenecks in g_up/g_down threads, plus avoid several context switches per I/O. The defined now safety requirements are: - caller should not hold any locks and should be reenterable; - callee should not depend on GEOM dual-threaded concurency semantics; - on the way down, if request is unmapped while callee doesn't support it, the context should be sleepable; - kernel thread stack usage should be below 50%. To keep compatibility with GEOM classes not meeting above requirements new provider and consumer flags added: - G_CF_DIRECT_SEND -- consumer code meets caller requirements (request); - G_CF_DIRECT_RECEIVE -- consumer code meets callee requirements (done); - G_PF_DIRECT_SEND -- provider code meets caller requirements (done); - G_PF_DIRECT_RECEIVE -- provider code meets callee requirements (request). Capable GEOM class can set them, allowing direct dispatch in cases where it is safe. If any of requirements are not met, request is queued to g_up or g_down thread same as before. Such GEOM classes were reviewed and updated to support direct dispatch: CONCAT, DEV, DISK, GATE, MD, MIRROR, MULTIPATH, NOP, PART, RAID, STRIPE, VFS, ZERO, ZFS::VDEV, ZFS::ZVOL, all classes based on g_slice KPI (LABEL, MAP, FLASHMAP, etc). To declare direct completion capability disk(9) KPI got new flag equivalent to G_PF_DIRECT_SEND -- DISKFLAG_DIRECT_COMPLETION. da(4) and ada(4) disk drivers got it set now thanks to earlier CAM locking work. This change more then twice increases peak block storage performance on systems with manu CPUs, together with earlier CAM locking changes reaching more then 1 million IOPS (512 byte raw reads from 16 SATA SSDs on 4 HBAs to 256 user-level threads). Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-22 08:22:19 +00:00
g_io_deliver(pbp, pbp->bio_error);
} else
mtx_unlock(&sc->sc_completion_lock);
Merge GEOM direct dispatch changes from the projects/camlock branch. When safety requirements are met, it allows to avoid passing I/O requests to GEOM g_up/g_down thread, executing them directly in the caller context. That allows to avoid CPU bottlenecks in g_up/g_down threads, plus avoid several context switches per I/O. The defined now safety requirements are: - caller should not hold any locks and should be reenterable; - callee should not depend on GEOM dual-threaded concurency semantics; - on the way down, if request is unmapped while callee doesn't support it, the context should be sleepable; - kernel thread stack usage should be below 50%. To keep compatibility with GEOM classes not meeting above requirements new provider and consumer flags added: - G_CF_DIRECT_SEND -- consumer code meets caller requirements (request); - G_CF_DIRECT_RECEIVE -- consumer code meets callee requirements (done); - G_PF_DIRECT_SEND -- provider code meets caller requirements (done); - G_PF_DIRECT_RECEIVE -- provider code meets callee requirements (request). Capable GEOM class can set them, allowing direct dispatch in cases where it is safe. If any of requirements are not met, request is queued to g_up or g_down thread same as before. Such GEOM classes were reviewed and updated to support direct dispatch: CONCAT, DEV, DISK, GATE, MD, MIRROR, MULTIPATH, NOP, PART, RAID, STRIPE, VFS, ZERO, ZFS::VDEV, ZFS::ZVOL, all classes based on g_slice KPI (LABEL, MAP, FLASHMAP, etc). To declare direct completion capability disk(9) KPI got new flag equivalent to G_PF_DIRECT_SEND -- DISKFLAG_DIRECT_COMPLETION. da(4) and ada(4) disk drivers got it set now thanks to earlier CAM locking work. This change more then twice increases peak block storage performance on systems with manu CPUs, together with earlier CAM locking changes reaching more then 1 million IOPS (512 byte raw reads from 16 SATA SSDs on 4 HBAs to 256 user-level threads). Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-22 08:22:19 +00:00
g_destroy_bio(bp);
}
/*
* Called for both BIO_FLUSH and BIO_SPEEDUP. Just pass the call down
*/
static void
g_concat_passdown(struct g_concat_softc *sc, struct bio *bp)
{
struct bio_queue_head queue;
struct g_consumer *cp;
struct bio *cbp;
struct g_concat_disk *disk;
sx_assert(&sc->sc_disks_lock, SX_LOCKED);
bioq_init(&queue);
TAILQ_FOREACH(disk, &sc->sc_disks, d_next) {
cbp = g_clone_bio(bp);
if (cbp == NULL) {
Merge GEOM direct dispatch changes from the projects/camlock branch. When safety requirements are met, it allows to avoid passing I/O requests to GEOM g_up/g_down thread, executing them directly in the caller context. That allows to avoid CPU bottlenecks in g_up/g_down threads, plus avoid several context switches per I/O. The defined now safety requirements are: - caller should not hold any locks and should be reenterable; - callee should not depend on GEOM dual-threaded concurency semantics; - on the way down, if request is unmapped while callee doesn't support it, the context should be sleepable; - kernel thread stack usage should be below 50%. To keep compatibility with GEOM classes not meeting above requirements new provider and consumer flags added: - G_CF_DIRECT_SEND -- consumer code meets caller requirements (request); - G_CF_DIRECT_RECEIVE -- consumer code meets callee requirements (done); - G_PF_DIRECT_SEND -- provider code meets caller requirements (done); - G_PF_DIRECT_RECEIVE -- provider code meets callee requirements (request). Capable GEOM class can set them, allowing direct dispatch in cases where it is safe. If any of requirements are not met, request is queued to g_up or g_down thread same as before. Such GEOM classes were reviewed and updated to support direct dispatch: CONCAT, DEV, DISK, GATE, MD, MIRROR, MULTIPATH, NOP, PART, RAID, STRIPE, VFS, ZERO, ZFS::VDEV, ZFS::ZVOL, all classes based on g_slice KPI (LABEL, MAP, FLASHMAP, etc). To declare direct completion capability disk(9) KPI got new flag equivalent to G_PF_DIRECT_SEND -- DISKFLAG_DIRECT_COMPLETION. da(4) and ada(4) disk drivers got it set now thanks to earlier CAM locking work. This change more then twice increases peak block storage performance on systems with manu CPUs, together with earlier CAM locking changes reaching more then 1 million IOPS (512 byte raw reads from 16 SATA SSDs on 4 HBAs to 256 user-level threads). Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-22 08:22:19 +00:00
while ((cbp = bioq_takefirst(&queue)) != NULL)
g_destroy_bio(cbp);
if (bp->bio_error == 0)
bp->bio_error = ENOMEM;
g_io_deliver(bp, bp->bio_error);
return;
}
bioq_insert_tail(&queue, cbp);
Merge GEOM direct dispatch changes from the projects/camlock branch. When safety requirements are met, it allows to avoid passing I/O requests to GEOM g_up/g_down thread, executing them directly in the caller context. That allows to avoid CPU bottlenecks in g_up/g_down threads, plus avoid several context switches per I/O. The defined now safety requirements are: - caller should not hold any locks and should be reenterable; - callee should not depend on GEOM dual-threaded concurency semantics; - on the way down, if request is unmapped while callee doesn't support it, the context should be sleepable; - kernel thread stack usage should be below 50%. To keep compatibility with GEOM classes not meeting above requirements new provider and consumer flags added: - G_CF_DIRECT_SEND -- consumer code meets caller requirements (request); - G_CF_DIRECT_RECEIVE -- consumer code meets callee requirements (done); - G_PF_DIRECT_SEND -- provider code meets caller requirements (done); - G_PF_DIRECT_RECEIVE -- provider code meets callee requirements (request). Capable GEOM class can set them, allowing direct dispatch in cases where it is safe. If any of requirements are not met, request is queued to g_up or g_down thread same as before. Such GEOM classes were reviewed and updated to support direct dispatch: CONCAT, DEV, DISK, GATE, MD, MIRROR, MULTIPATH, NOP, PART, RAID, STRIPE, VFS, ZERO, ZFS::VDEV, ZFS::ZVOL, all classes based on g_slice KPI (LABEL, MAP, FLASHMAP, etc). To declare direct completion capability disk(9) KPI got new flag equivalent to G_PF_DIRECT_SEND -- DISKFLAG_DIRECT_COMPLETION. da(4) and ada(4) disk drivers got it set now thanks to earlier CAM locking work. This change more then twice increases peak block storage performance on systems with manu CPUs, together with earlier CAM locking changes reaching more then 1 million IOPS (512 byte raw reads from 16 SATA SSDs on 4 HBAs to 256 user-level threads). Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-22 08:22:19 +00:00
cbp->bio_done = g_concat_done;
cbp->bio_caller1 = disk->d_consumer;
cbp->bio_to = disk->d_consumer->provider;
}
Merge GEOM direct dispatch changes from the projects/camlock branch. When safety requirements are met, it allows to avoid passing I/O requests to GEOM g_up/g_down thread, executing them directly in the caller context. That allows to avoid CPU bottlenecks in g_up/g_down threads, plus avoid several context switches per I/O. The defined now safety requirements are: - caller should not hold any locks and should be reenterable; - callee should not depend on GEOM dual-threaded concurency semantics; - on the way down, if request is unmapped while callee doesn't support it, the context should be sleepable; - kernel thread stack usage should be below 50%. To keep compatibility with GEOM classes not meeting above requirements new provider and consumer flags added: - G_CF_DIRECT_SEND -- consumer code meets caller requirements (request); - G_CF_DIRECT_RECEIVE -- consumer code meets callee requirements (done); - G_PF_DIRECT_SEND -- provider code meets caller requirements (done); - G_PF_DIRECT_RECEIVE -- provider code meets callee requirements (request). Capable GEOM class can set them, allowing direct dispatch in cases where it is safe. If any of requirements are not met, request is queued to g_up or g_down thread same as before. Such GEOM classes were reviewed and updated to support direct dispatch: CONCAT, DEV, DISK, GATE, MD, MIRROR, MULTIPATH, NOP, PART, RAID, STRIPE, VFS, ZERO, ZFS::VDEV, ZFS::ZVOL, all classes based on g_slice KPI (LABEL, MAP, FLASHMAP, etc). To declare direct completion capability disk(9) KPI got new flag equivalent to G_PF_DIRECT_SEND -- DISKFLAG_DIRECT_COMPLETION. da(4) and ada(4) disk drivers got it set now thanks to earlier CAM locking work. This change more then twice increases peak block storage performance on systems with manu CPUs, together with earlier CAM locking changes reaching more then 1 million IOPS (512 byte raw reads from 16 SATA SSDs on 4 HBAs to 256 user-level threads). Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-22 08:22:19 +00:00
while ((cbp = bioq_takefirst(&queue)) != NULL) {
G_CONCAT_LOGREQ(cbp, "Sending request.");
cp = cbp->bio_caller1;
cbp->bio_caller1 = NULL;
g_io_request(cbp, cp);
}
}
static void
g_concat_start(struct bio *bp)
{
struct bio_queue_head queue;
struct g_concat_softc *sc;
struct g_concat_disk *disk;
struct g_provider *pp;
off_t offset, end, length, off, len;
struct bio *cbp;
char *addr;
pp = bp->bio_to;
sc = pp->geom->softc;
/*
* If sc == NULL, provider's error should be set and g_concat_start()
* should not be called at all.
*/
KASSERT(sc != NULL,
("Provider's error should be set (error=%d)(device=%s).",
bp->bio_to->error, bp->bio_to->name));
G_CONCAT_LOGREQ(bp, "Request received.");
sx_slock(&sc->sc_disks_lock);
switch (bp->bio_cmd) {
case BIO_READ:
case BIO_WRITE:
case BIO_DELETE:
break;
case BIO_SPEEDUP:
case BIO_FLUSH:
g_concat_passdown(sc, bp);
goto end;
case BIO_GETATTR:
if (strcmp("GEOM::kerneldump", bp->bio_attribute) == 0) {
g_concat_kernel_dump(bp);
goto end;
} else if (strcmp("GEOM::candelete", bp->bio_attribute) == 0) {
g_concat_candelete(bp);
goto end;
}
/* To which provider it should be delivered? */
/* FALLTHROUGH */
default:
g_io_deliver(bp, EOPNOTSUPP);
goto end;
}
offset = bp->bio_offset;
length = bp->bio_length;
Merge GEOM direct dispatch changes from the projects/camlock branch. When safety requirements are met, it allows to avoid passing I/O requests to GEOM g_up/g_down thread, executing them directly in the caller context. That allows to avoid CPU bottlenecks in g_up/g_down threads, plus avoid several context switches per I/O. The defined now safety requirements are: - caller should not hold any locks and should be reenterable; - callee should not depend on GEOM dual-threaded concurency semantics; - on the way down, if request is unmapped while callee doesn't support it, the context should be sleepable; - kernel thread stack usage should be below 50%. To keep compatibility with GEOM classes not meeting above requirements new provider and consumer flags added: - G_CF_DIRECT_SEND -- consumer code meets caller requirements (request); - G_CF_DIRECT_RECEIVE -- consumer code meets callee requirements (done); - G_PF_DIRECT_SEND -- provider code meets caller requirements (done); - G_PF_DIRECT_RECEIVE -- provider code meets callee requirements (request). Capable GEOM class can set them, allowing direct dispatch in cases where it is safe. If any of requirements are not met, request is queued to g_up or g_down thread same as before. Such GEOM classes were reviewed and updated to support direct dispatch: CONCAT, DEV, DISK, GATE, MD, MIRROR, MULTIPATH, NOP, PART, RAID, STRIPE, VFS, ZERO, ZFS::VDEV, ZFS::ZVOL, all classes based on g_slice KPI (LABEL, MAP, FLASHMAP, etc). To declare direct completion capability disk(9) KPI got new flag equivalent to G_PF_DIRECT_SEND -- DISKFLAG_DIRECT_COMPLETION. da(4) and ada(4) disk drivers got it set now thanks to earlier CAM locking work. This change more then twice increases peak block storage performance on systems with manu CPUs, together with earlier CAM locking changes reaching more then 1 million IOPS (512 byte raw reads from 16 SATA SSDs on 4 HBAs to 256 user-level threads). Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-22 08:22:19 +00:00
if ((bp->bio_flags & BIO_UNMAPPED) != 0)
addr = NULL;
else
addr = bp->bio_data;
end = offset + length;
bioq_init(&queue);
TAILQ_FOREACH(disk, &sc->sc_disks, d_next) {
if (disk->d_end <= offset)
continue;
if (disk->d_start >= end)
break;
off = offset - disk->d_start;
len = MIN(length, disk->d_end - offset);
length -= len;
offset += len;
cbp = g_clone_bio(bp);
if (cbp == NULL) {
Merge GEOM direct dispatch changes from the projects/camlock branch. When safety requirements are met, it allows to avoid passing I/O requests to GEOM g_up/g_down thread, executing them directly in the caller context. That allows to avoid CPU bottlenecks in g_up/g_down threads, plus avoid several context switches per I/O. The defined now safety requirements are: - caller should not hold any locks and should be reenterable; - callee should not depend on GEOM dual-threaded concurency semantics; - on the way down, if request is unmapped while callee doesn't support it, the context should be sleepable; - kernel thread stack usage should be below 50%. To keep compatibility with GEOM classes not meeting above requirements new provider and consumer flags added: - G_CF_DIRECT_SEND -- consumer code meets caller requirements (request); - G_CF_DIRECT_RECEIVE -- consumer code meets callee requirements (done); - G_PF_DIRECT_SEND -- provider code meets caller requirements (done); - G_PF_DIRECT_RECEIVE -- provider code meets callee requirements (request). Capable GEOM class can set them, allowing direct dispatch in cases where it is safe. If any of requirements are not met, request is queued to g_up or g_down thread same as before. Such GEOM classes were reviewed and updated to support direct dispatch: CONCAT, DEV, DISK, GATE, MD, MIRROR, MULTIPATH, NOP, PART, RAID, STRIPE, VFS, ZERO, ZFS::VDEV, ZFS::ZVOL, all classes based on g_slice KPI (LABEL, MAP, FLASHMAP, etc). To declare direct completion capability disk(9) KPI got new flag equivalent to G_PF_DIRECT_SEND -- DISKFLAG_DIRECT_COMPLETION. da(4) and ada(4) disk drivers got it set now thanks to earlier CAM locking work. This change more then twice increases peak block storage performance on systems with manu CPUs, together with earlier CAM locking changes reaching more then 1 million IOPS (512 byte raw reads from 16 SATA SSDs on 4 HBAs to 256 user-level threads). Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-22 08:22:19 +00:00
while ((cbp = bioq_takefirst(&queue)) != NULL)
g_destroy_bio(cbp);
if (bp->bio_error == 0)
bp->bio_error = ENOMEM;
g_io_deliver(bp, bp->bio_error);
goto end;
}
bioq_insert_tail(&queue, cbp);
/*
* Fill in the component buf structure.
*/
Merge GEOM direct dispatch changes from the projects/camlock branch. When safety requirements are met, it allows to avoid passing I/O requests to GEOM g_up/g_down thread, executing them directly in the caller context. That allows to avoid CPU bottlenecks in g_up/g_down threads, plus avoid several context switches per I/O. The defined now safety requirements are: - caller should not hold any locks and should be reenterable; - callee should not depend on GEOM dual-threaded concurency semantics; - on the way down, if request is unmapped while callee doesn't support it, the context should be sleepable; - kernel thread stack usage should be below 50%. To keep compatibility with GEOM classes not meeting above requirements new provider and consumer flags added: - G_CF_DIRECT_SEND -- consumer code meets caller requirements (request); - G_CF_DIRECT_RECEIVE -- consumer code meets callee requirements (done); - G_PF_DIRECT_SEND -- provider code meets caller requirements (done); - G_PF_DIRECT_RECEIVE -- provider code meets callee requirements (request). Capable GEOM class can set them, allowing direct dispatch in cases where it is safe. If any of requirements are not met, request is queued to g_up or g_down thread same as before. Such GEOM classes were reviewed and updated to support direct dispatch: CONCAT, DEV, DISK, GATE, MD, MIRROR, MULTIPATH, NOP, PART, RAID, STRIPE, VFS, ZERO, ZFS::VDEV, ZFS::ZVOL, all classes based on g_slice KPI (LABEL, MAP, FLASHMAP, etc). To declare direct completion capability disk(9) KPI got new flag equivalent to G_PF_DIRECT_SEND -- DISKFLAG_DIRECT_COMPLETION. da(4) and ada(4) disk drivers got it set now thanks to earlier CAM locking work. This change more then twice increases peak block storage performance on systems with manu CPUs, together with earlier CAM locking changes reaching more then 1 million IOPS (512 byte raw reads from 16 SATA SSDs on 4 HBAs to 256 user-level threads). Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-22 08:22:19 +00:00
if (len == bp->bio_length)
cbp->bio_done = g_std_done;
else
cbp->bio_done = g_concat_done;
cbp->bio_offset = off;
cbp->bio_length = len;
Merge GEOM direct dispatch changes from the projects/camlock branch. When safety requirements are met, it allows to avoid passing I/O requests to GEOM g_up/g_down thread, executing them directly in the caller context. That allows to avoid CPU bottlenecks in g_up/g_down threads, plus avoid several context switches per I/O. The defined now safety requirements are: - caller should not hold any locks and should be reenterable; - callee should not depend on GEOM dual-threaded concurency semantics; - on the way down, if request is unmapped while callee doesn't support it, the context should be sleepable; - kernel thread stack usage should be below 50%. To keep compatibility with GEOM classes not meeting above requirements new provider and consumer flags added: - G_CF_DIRECT_SEND -- consumer code meets caller requirements (request); - G_CF_DIRECT_RECEIVE -- consumer code meets callee requirements (done); - G_PF_DIRECT_SEND -- provider code meets caller requirements (done); - G_PF_DIRECT_RECEIVE -- provider code meets callee requirements (request). Capable GEOM class can set them, allowing direct dispatch in cases where it is safe. If any of requirements are not met, request is queued to g_up or g_down thread same as before. Such GEOM classes were reviewed and updated to support direct dispatch: CONCAT, DEV, DISK, GATE, MD, MIRROR, MULTIPATH, NOP, PART, RAID, STRIPE, VFS, ZERO, ZFS::VDEV, ZFS::ZVOL, all classes based on g_slice KPI (LABEL, MAP, FLASHMAP, etc). To declare direct completion capability disk(9) KPI got new flag equivalent to G_PF_DIRECT_SEND -- DISKFLAG_DIRECT_COMPLETION. da(4) and ada(4) disk drivers got it set now thanks to earlier CAM locking work. This change more then twice increases peak block storage performance on systems with manu CPUs, together with earlier CAM locking changes reaching more then 1 million IOPS (512 byte raw reads from 16 SATA SSDs on 4 HBAs to 256 user-level threads). Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-22 08:22:19 +00:00
if ((bp->bio_flags & BIO_UNMAPPED) != 0) {
cbp->bio_ma_offset += (uintptr_t)addr;
cbp->bio_ma += cbp->bio_ma_offset / PAGE_SIZE;
cbp->bio_ma_offset %= PAGE_SIZE;
cbp->bio_ma_n = round_page(cbp->bio_ma_offset +
cbp->bio_length) / PAGE_SIZE;
} else
cbp->bio_data = addr;
addr += len;
cbp->bio_to = disk->d_consumer->provider;
cbp->bio_caller1 = disk;
if (length == 0)
break;
}
KASSERT(length == 0,
("Length is still greater than 0 (class=%s, name=%s).",
bp->bio_to->geom->class->name, bp->bio_to->geom->name));
Merge GEOM direct dispatch changes from the projects/camlock branch. When safety requirements are met, it allows to avoid passing I/O requests to GEOM g_up/g_down thread, executing them directly in the caller context. That allows to avoid CPU bottlenecks in g_up/g_down threads, plus avoid several context switches per I/O. The defined now safety requirements are: - caller should not hold any locks and should be reenterable; - callee should not depend on GEOM dual-threaded concurency semantics; - on the way down, if request is unmapped while callee doesn't support it, the context should be sleepable; - kernel thread stack usage should be below 50%. To keep compatibility with GEOM classes not meeting above requirements new provider and consumer flags added: - G_CF_DIRECT_SEND -- consumer code meets caller requirements (request); - G_CF_DIRECT_RECEIVE -- consumer code meets callee requirements (done); - G_PF_DIRECT_SEND -- provider code meets caller requirements (done); - G_PF_DIRECT_RECEIVE -- provider code meets callee requirements (request). Capable GEOM class can set them, allowing direct dispatch in cases where it is safe. If any of requirements are not met, request is queued to g_up or g_down thread same as before. Such GEOM classes were reviewed and updated to support direct dispatch: CONCAT, DEV, DISK, GATE, MD, MIRROR, MULTIPATH, NOP, PART, RAID, STRIPE, VFS, ZERO, ZFS::VDEV, ZFS::ZVOL, all classes based on g_slice KPI (LABEL, MAP, FLASHMAP, etc). To declare direct completion capability disk(9) KPI got new flag equivalent to G_PF_DIRECT_SEND -- DISKFLAG_DIRECT_COMPLETION. da(4) and ada(4) disk drivers got it set now thanks to earlier CAM locking work. This change more then twice increases peak block storage performance on systems with manu CPUs, together with earlier CAM locking changes reaching more then 1 million IOPS (512 byte raw reads from 16 SATA SSDs on 4 HBAs to 256 user-level threads). Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-22 08:22:19 +00:00
while ((cbp = bioq_takefirst(&queue)) != NULL) {
G_CONCAT_LOGREQ(cbp, "Sending request.");
disk = cbp->bio_caller1;
cbp->bio_caller1 = NULL;
g_io_request(cbp, disk->d_consumer);
}
end:
sx_sunlock(&sc->sc_disks_lock);
}
static void
g_concat_check_and_run(struct g_concat_softc *sc)
{
struct g_concat_disk *disk;
Merge GEOM direct dispatch changes from the projects/camlock branch. When safety requirements are met, it allows to avoid passing I/O requests to GEOM g_up/g_down thread, executing them directly in the caller context. That allows to avoid CPU bottlenecks in g_up/g_down threads, plus avoid several context switches per I/O. The defined now safety requirements are: - caller should not hold any locks and should be reenterable; - callee should not depend on GEOM dual-threaded concurency semantics; - on the way down, if request is unmapped while callee doesn't support it, the context should be sleepable; - kernel thread stack usage should be below 50%. To keep compatibility with GEOM classes not meeting above requirements new provider and consumer flags added: - G_CF_DIRECT_SEND -- consumer code meets caller requirements (request); - G_CF_DIRECT_RECEIVE -- consumer code meets callee requirements (done); - G_PF_DIRECT_SEND -- provider code meets caller requirements (done); - G_PF_DIRECT_RECEIVE -- provider code meets callee requirements (request). Capable GEOM class can set them, allowing direct dispatch in cases where it is safe. If any of requirements are not met, request is queued to g_up or g_down thread same as before. Such GEOM classes were reviewed and updated to support direct dispatch: CONCAT, DEV, DISK, GATE, MD, MIRROR, MULTIPATH, NOP, PART, RAID, STRIPE, VFS, ZERO, ZFS::VDEV, ZFS::ZVOL, all classes based on g_slice KPI (LABEL, MAP, FLASHMAP, etc). To declare direct completion capability disk(9) KPI got new flag equivalent to G_PF_DIRECT_SEND -- DISKFLAG_DIRECT_COMPLETION. da(4) and ada(4) disk drivers got it set now thanks to earlier CAM locking work. This change more then twice increases peak block storage performance on systems with manu CPUs, together with earlier CAM locking changes reaching more then 1 million IOPS (512 byte raw reads from 16 SATA SSDs on 4 HBAs to 256 user-level threads). Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-22 08:22:19 +00:00
struct g_provider *dp, *pp;
u_int sectorsize = 0;
off_t start;
int error;
g_topology_assert();
if (g_concat_nvalid(sc) != sc->sc_ndisks)
return;
pp = g_new_providerf(sc->sc_geom, "concat/%s", sc->sc_name);
Merge GEOM direct dispatch changes from the projects/camlock branch. When safety requirements are met, it allows to avoid passing I/O requests to GEOM g_up/g_down thread, executing them directly in the caller context. That allows to avoid CPU bottlenecks in g_up/g_down threads, plus avoid several context switches per I/O. The defined now safety requirements are: - caller should not hold any locks and should be reenterable; - callee should not depend on GEOM dual-threaded concurency semantics; - on the way down, if request is unmapped while callee doesn't support it, the context should be sleepable; - kernel thread stack usage should be below 50%. To keep compatibility with GEOM classes not meeting above requirements new provider and consumer flags added: - G_CF_DIRECT_SEND -- consumer code meets caller requirements (request); - G_CF_DIRECT_RECEIVE -- consumer code meets callee requirements (done); - G_PF_DIRECT_SEND -- provider code meets caller requirements (done); - G_PF_DIRECT_RECEIVE -- provider code meets callee requirements (request). Capable GEOM class can set them, allowing direct dispatch in cases where it is safe. If any of requirements are not met, request is queued to g_up or g_down thread same as before. Such GEOM classes were reviewed and updated to support direct dispatch: CONCAT, DEV, DISK, GATE, MD, MIRROR, MULTIPATH, NOP, PART, RAID, STRIPE, VFS, ZERO, ZFS::VDEV, ZFS::ZVOL, all classes based on g_slice KPI (LABEL, MAP, FLASHMAP, etc). To declare direct completion capability disk(9) KPI got new flag equivalent to G_PF_DIRECT_SEND -- DISKFLAG_DIRECT_COMPLETION. da(4) and ada(4) disk drivers got it set now thanks to earlier CAM locking work. This change more then twice increases peak block storage performance on systems with manu CPUs, together with earlier CAM locking changes reaching more then 1 million IOPS (512 byte raw reads from 16 SATA SSDs on 4 HBAs to 256 user-level threads). Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-22 08:22:19 +00:00
pp->flags |= G_PF_DIRECT_SEND | G_PF_DIRECT_RECEIVE |
G_PF_ACCEPT_UNMAPPED;
start = 0;
TAILQ_FOREACH(disk, &sc->sc_disks, d_next) {
Merge GEOM direct dispatch changes from the projects/camlock branch. When safety requirements are met, it allows to avoid passing I/O requests to GEOM g_up/g_down thread, executing them directly in the caller context. That allows to avoid CPU bottlenecks in g_up/g_down threads, plus avoid several context switches per I/O. The defined now safety requirements are: - caller should not hold any locks and should be reenterable; - callee should not depend on GEOM dual-threaded concurency semantics; - on the way down, if request is unmapped while callee doesn't support it, the context should be sleepable; - kernel thread stack usage should be below 50%. To keep compatibility with GEOM classes not meeting above requirements new provider and consumer flags added: - G_CF_DIRECT_SEND -- consumer code meets caller requirements (request); - G_CF_DIRECT_RECEIVE -- consumer code meets callee requirements (done); - G_PF_DIRECT_SEND -- provider code meets caller requirements (done); - G_PF_DIRECT_RECEIVE -- provider code meets callee requirements (request). Capable GEOM class can set them, allowing direct dispatch in cases where it is safe. If any of requirements are not met, request is queued to g_up or g_down thread same as before. Such GEOM classes were reviewed and updated to support direct dispatch: CONCAT, DEV, DISK, GATE, MD, MIRROR, MULTIPATH, NOP, PART, RAID, STRIPE, VFS, ZERO, ZFS::VDEV, ZFS::ZVOL, all classes based on g_slice KPI (LABEL, MAP, FLASHMAP, etc). To declare direct completion capability disk(9) KPI got new flag equivalent to G_PF_DIRECT_SEND -- DISKFLAG_DIRECT_COMPLETION. da(4) and ada(4) disk drivers got it set now thanks to earlier CAM locking work. This change more then twice increases peak block storage performance on systems with manu CPUs, together with earlier CAM locking changes reaching more then 1 million IOPS (512 byte raw reads from 16 SATA SSDs on 4 HBAs to 256 user-level threads). Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-22 08:22:19 +00:00
dp = disk->d_consumer->provider;
disk->d_start = start;
Merge GEOM direct dispatch changes from the projects/camlock branch. When safety requirements are met, it allows to avoid passing I/O requests to GEOM g_up/g_down thread, executing them directly in the caller context. That allows to avoid CPU bottlenecks in g_up/g_down threads, plus avoid several context switches per I/O. The defined now safety requirements are: - caller should not hold any locks and should be reenterable; - callee should not depend on GEOM dual-threaded concurency semantics; - on the way down, if request is unmapped while callee doesn't support it, the context should be sleepable; - kernel thread stack usage should be below 50%. To keep compatibility with GEOM classes not meeting above requirements new provider and consumer flags added: - G_CF_DIRECT_SEND -- consumer code meets caller requirements (request); - G_CF_DIRECT_RECEIVE -- consumer code meets callee requirements (done); - G_PF_DIRECT_SEND -- provider code meets caller requirements (done); - G_PF_DIRECT_RECEIVE -- provider code meets callee requirements (request). Capable GEOM class can set them, allowing direct dispatch in cases where it is safe. If any of requirements are not met, request is queued to g_up or g_down thread same as before. Such GEOM classes were reviewed and updated to support direct dispatch: CONCAT, DEV, DISK, GATE, MD, MIRROR, MULTIPATH, NOP, PART, RAID, STRIPE, VFS, ZERO, ZFS::VDEV, ZFS::ZVOL, all classes based on g_slice KPI (LABEL, MAP, FLASHMAP, etc). To declare direct completion capability disk(9) KPI got new flag equivalent to G_PF_DIRECT_SEND -- DISKFLAG_DIRECT_COMPLETION. da(4) and ada(4) disk drivers got it set now thanks to earlier CAM locking work. This change more then twice increases peak block storage performance on systems with manu CPUs, together with earlier CAM locking changes reaching more then 1 million IOPS (512 byte raw reads from 16 SATA SSDs on 4 HBAs to 256 user-level threads). Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-22 08:22:19 +00:00
disk->d_end = disk->d_start + dp->mediasize;
if (sc->sc_type == G_CONCAT_TYPE_AUTOMATIC)
Merge GEOM direct dispatch changes from the projects/camlock branch. When safety requirements are met, it allows to avoid passing I/O requests to GEOM g_up/g_down thread, executing them directly in the caller context. That allows to avoid CPU bottlenecks in g_up/g_down threads, plus avoid several context switches per I/O. The defined now safety requirements are: - caller should not hold any locks and should be reenterable; - callee should not depend on GEOM dual-threaded concurency semantics; - on the way down, if request is unmapped while callee doesn't support it, the context should be sleepable; - kernel thread stack usage should be below 50%. To keep compatibility with GEOM classes not meeting above requirements new provider and consumer flags added: - G_CF_DIRECT_SEND -- consumer code meets caller requirements (request); - G_CF_DIRECT_RECEIVE -- consumer code meets callee requirements (done); - G_PF_DIRECT_SEND -- provider code meets caller requirements (done); - G_PF_DIRECT_RECEIVE -- provider code meets callee requirements (request). Capable GEOM class can set them, allowing direct dispatch in cases where it is safe. If any of requirements are not met, request is queued to g_up or g_down thread same as before. Such GEOM classes were reviewed and updated to support direct dispatch: CONCAT, DEV, DISK, GATE, MD, MIRROR, MULTIPATH, NOP, PART, RAID, STRIPE, VFS, ZERO, ZFS::VDEV, ZFS::ZVOL, all classes based on g_slice KPI (LABEL, MAP, FLASHMAP, etc). To declare direct completion capability disk(9) KPI got new flag equivalent to G_PF_DIRECT_SEND -- DISKFLAG_DIRECT_COMPLETION. da(4) and ada(4) disk drivers got it set now thanks to earlier CAM locking work. This change more then twice increases peak block storage performance on systems with manu CPUs, together with earlier CAM locking changes reaching more then 1 million IOPS (512 byte raw reads from 16 SATA SSDs on 4 HBAs to 256 user-level threads). Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-22 08:22:19 +00:00
disk->d_end -= dp->sectorsize;
start = disk->d_end;
error = g_access(disk->d_consumer, 1, 0, 0);
if (error == 0) {
error = g_getattr("GEOM::candelete", disk->d_consumer,
&disk->d_candelete);
if (error != 0)
disk->d_candelete = 0;
(void)g_access(disk->d_consumer, -1, 0, 0);
} else
G_CONCAT_DEBUG(1, "Failed to access disk %s, error %d.",
dp->name, error);
if (disk == TAILQ_FIRST(&sc->sc_disks))
Merge GEOM direct dispatch changes from the projects/camlock branch. When safety requirements are met, it allows to avoid passing I/O requests to GEOM g_up/g_down thread, executing them directly in the caller context. That allows to avoid CPU bottlenecks in g_up/g_down threads, plus avoid several context switches per I/O. The defined now safety requirements are: - caller should not hold any locks and should be reenterable; - callee should not depend on GEOM dual-threaded concurency semantics; - on the way down, if request is unmapped while callee doesn't support it, the context should be sleepable; - kernel thread stack usage should be below 50%. To keep compatibility with GEOM classes not meeting above requirements new provider and consumer flags added: - G_CF_DIRECT_SEND -- consumer code meets caller requirements (request); - G_CF_DIRECT_RECEIVE -- consumer code meets callee requirements (done); - G_PF_DIRECT_SEND -- provider code meets caller requirements (done); - G_PF_DIRECT_RECEIVE -- provider code meets callee requirements (request). Capable GEOM class can set them, allowing direct dispatch in cases where it is safe. If any of requirements are not met, request is queued to g_up or g_down thread same as before. Such GEOM classes were reviewed and updated to support direct dispatch: CONCAT, DEV, DISK, GATE, MD, MIRROR, MULTIPATH, NOP, PART, RAID, STRIPE, VFS, ZERO, ZFS::VDEV, ZFS::ZVOL, all classes based on g_slice KPI (LABEL, MAP, FLASHMAP, etc). To declare direct completion capability disk(9) KPI got new flag equivalent to G_PF_DIRECT_SEND -- DISKFLAG_DIRECT_COMPLETION. da(4) and ada(4) disk drivers got it set now thanks to earlier CAM locking work. This change more then twice increases peak block storage performance on systems with manu CPUs, together with earlier CAM locking changes reaching more then 1 million IOPS (512 byte raw reads from 16 SATA SSDs on 4 HBAs to 256 user-level threads). Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-22 08:22:19 +00:00
sectorsize = dp->sectorsize;
else
sectorsize = lcm(sectorsize, dp->sectorsize);
/* A provider underneath us doesn't support unmapped */
if ((dp->flags & G_PF_ACCEPT_UNMAPPED) == 0) {
G_CONCAT_DEBUG(1, "Cancelling unmapped "
"because of %s.", dp->name);
pp->flags &= ~G_PF_ACCEPT_UNMAPPED;
}
}
pp->sectorsize = sectorsize;
/* We have sc->sc_disks[sc->sc_ndisks - 1].d_end in 'start'. */
pp->mediasize = start;
dp = TAILQ_FIRST(&sc->sc_disks)->d_consumer->provider;
pp->stripesize = dp->stripesize;
pp->stripeoffset = dp->stripeoffset;
sc->sc_provider = pp;
g_error_provider(pp, 0);
G_CONCAT_DEBUG(0, "Device %s activated.", sc->sc_provider->name);
}
static int
g_concat_read_metadata(struct g_consumer *cp, struct g_concat_metadata *md)
{
struct g_provider *pp;
u_char *buf;
int error;
g_topology_assert();
error = g_access(cp, 1, 0, 0);
if (error != 0)
return (error);
pp = cp->provider;
g_topology_unlock();
buf = g_read_data(cp, pp->mediasize - pp->sectorsize, pp->sectorsize,
&error);
g_topology_lock();
g_access(cp, -1, 0, 0);
if (buf == NULL)
return (error);
/* Decode metadata. */
concat_metadata_decode(buf, md);
g_free(buf);
return (0);
}
/*
* Add disk to given device.
*/
static int
g_concat_add_disk(struct g_concat_softc *sc, struct g_provider *pp, u_int no)
{
struct g_concat_disk *disk;
struct g_consumer *cp, *fcp;
struct g_geom *gp;
int error;
g_topology_assert();
sx_slock(&sc->sc_disks_lock);
/* Metadata corrupted? */
if (no >= sc->sc_ndisks) {
sx_sunlock(&sc->sc_disks_lock);
return (EINVAL);
}
for (disk = TAILQ_FIRST(&sc->sc_disks); no > 0; no--) {
disk = TAILQ_NEXT(disk, d_next);
}
/* Check if disk is not already attached. */
if (disk->d_consumer != NULL) {
sx_sunlock(&sc->sc_disks_lock);
return (EEXIST);
}
gp = sc->sc_geom;
fcp = LIST_FIRST(&gp->consumer);
cp = g_new_consumer(gp);
Merge GEOM direct dispatch changes from the projects/camlock branch. When safety requirements are met, it allows to avoid passing I/O requests to GEOM g_up/g_down thread, executing them directly in the caller context. That allows to avoid CPU bottlenecks in g_up/g_down threads, plus avoid several context switches per I/O. The defined now safety requirements are: - caller should not hold any locks and should be reenterable; - callee should not depend on GEOM dual-threaded concurency semantics; - on the way down, if request is unmapped while callee doesn't support it, the context should be sleepable; - kernel thread stack usage should be below 50%. To keep compatibility with GEOM classes not meeting above requirements new provider and consumer flags added: - G_CF_DIRECT_SEND -- consumer code meets caller requirements (request); - G_CF_DIRECT_RECEIVE -- consumer code meets callee requirements (done); - G_PF_DIRECT_SEND -- provider code meets caller requirements (done); - G_PF_DIRECT_RECEIVE -- provider code meets callee requirements (request). Capable GEOM class can set them, allowing direct dispatch in cases where it is safe. If any of requirements are not met, request is queued to g_up or g_down thread same as before. Such GEOM classes were reviewed and updated to support direct dispatch: CONCAT, DEV, DISK, GATE, MD, MIRROR, MULTIPATH, NOP, PART, RAID, STRIPE, VFS, ZERO, ZFS::VDEV, ZFS::ZVOL, all classes based on g_slice KPI (LABEL, MAP, FLASHMAP, etc). To declare direct completion capability disk(9) KPI got new flag equivalent to G_PF_DIRECT_SEND -- DISKFLAG_DIRECT_COMPLETION. da(4) and ada(4) disk drivers got it set now thanks to earlier CAM locking work. This change more then twice increases peak block storage performance on systems with manu CPUs, together with earlier CAM locking changes reaching more then 1 million IOPS (512 byte raw reads from 16 SATA SSDs on 4 HBAs to 256 user-level threads). Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-22 08:22:19 +00:00
cp->flags |= G_CF_DIRECT_SEND | G_CF_DIRECT_RECEIVE;
error = g_attach(cp, pp);
if (error != 0) {
sx_sunlock(&sc->sc_disks_lock);
g_destroy_consumer(cp);
return (error);
}
if (fcp != NULL && (fcp->acr > 0 || fcp->acw > 0 || fcp->ace > 0)) {
error = g_access(cp, fcp->acr, fcp->acw, fcp->ace);
if (error != 0) {
sx_sunlock(&sc->sc_disks_lock);
g_detach(cp);
g_destroy_consumer(cp);
return (error);
}
}
if (sc->sc_type == G_CONCAT_TYPE_AUTOMATIC) {
struct g_concat_metadata md;
// temporarily give up the lock to avoid lock order violation
// due to topology unlock in g_concat_read_metadata
sx_sunlock(&sc->sc_disks_lock);
/* Re-read metadata. */
error = g_concat_read_metadata(cp, &md);
sx_slock(&sc->sc_disks_lock);
if (error != 0)
goto fail;
if (strcmp(md.md_magic, G_CONCAT_MAGIC) != 0 ||
strcmp(md.md_name, sc->sc_name) != 0 ||
md.md_id != sc->sc_id) {
G_CONCAT_DEBUG(0, "Metadata on %s changed.", pp->name);
goto fail;
}
disk->d_hardcoded = md.md_provider[0] != '\0';
} else {
disk->d_hardcoded = false;
}
cp->private = disk;
disk->d_consumer = cp;
disk->d_softc = sc;
disk->d_start = 0; /* not yet */
disk->d_end = 0; /* not yet */
disk->d_removed = 0;
G_CONCAT_DEBUG(0, "Disk %s attached to %s.", pp->name, sc->sc_name);
g_concat_check_and_run(sc);
sx_sunlock(&sc->sc_disks_lock); // need lock for check_and_run
return (0);
fail:
sx_sunlock(&sc->sc_disks_lock);
if (fcp != NULL && (fcp->acr > 0 || fcp->acw > 0 || fcp->ace > 0))
g_access(cp, -fcp->acr, -fcp->acw, -fcp->ace);
g_detach(cp);
g_destroy_consumer(cp);
return (error);
}
static struct g_geom *
g_concat_create(struct g_class *mp, const struct g_concat_metadata *md,
u_int type)
{
struct g_concat_softc *sc;
struct g_concat_disk *disk;
struct g_geom *gp;
u_int no;
G_CONCAT_DEBUG(1, "Creating device %s (id=%u).", md->md_name,
md->md_id);
/* One disks is minimum. */
if (md->md_all < 1)
return (NULL);
/* Check for duplicate unit */
LIST_FOREACH(gp, &mp->geom, geom) {
sc = gp->softc;
if (sc != NULL && strcmp(sc->sc_name, md->md_name) == 0) {
G_CONCAT_DEBUG(0, "Device %s already configured.",
gp->name);
return (NULL);
}
}
gp = g_new_geomf(mp, "%s", md->md_name);
sc = malloc(sizeof(*sc), M_CONCAT, M_WAITOK | M_ZERO);
gp->start = g_concat_start;
gp->spoiled = g_concat_orphan;
gp->orphan = g_concat_orphan;
gp->access = g_concat_access;
gp->dumpconf = g_concat_dumpconf;
sc->sc_id = md->md_id;
sc->sc_ndisks = md->md_all;
TAILQ_INIT(&sc->sc_disks);
for (no = 0; no < sc->sc_ndisks; no++) {
disk = malloc(sizeof(*disk), M_CONCAT, M_WAITOK | M_ZERO);
TAILQ_INSERT_TAIL(&sc->sc_disks, disk, d_next);
}
sc->sc_type = type;
mtx_init(&sc->sc_completion_lock, "gconcat lock", NULL, MTX_DEF);
sx_init(&sc->sc_disks_lock, "gconcat append lock");
gp->softc = sc;
sc->sc_geom = gp;
sc->sc_provider = NULL;
G_CONCAT_DEBUG(0, "Device %s created (id=%u).", sc->sc_name, sc->sc_id);
return (gp);
}
static int
g_concat_destroy(struct g_concat_softc *sc, boolean_t force)
{
struct g_provider *pp;
struct g_consumer *cp, *cp1;
struct g_geom *gp;
struct g_concat_disk *disk;
g_topology_assert();
if (sc == NULL)
return (ENXIO);
pp = sc->sc_provider;
if (pp != NULL && (pp->acr != 0 || pp->acw != 0 || pp->ace != 0)) {
if (force) {
G_CONCAT_DEBUG(0, "Device %s is still open, so it "
"can't be definitely removed.", pp->name);
} else {
G_CONCAT_DEBUG(1,
"Device %s is still open (r%dw%de%d).", pp->name,
pp->acr, pp->acw, pp->ace);
return (EBUSY);
}
}
gp = sc->sc_geom;
LIST_FOREACH_SAFE(cp, &gp->consumer, consumer, cp1) {
g_concat_remove_disk(cp->private);
if (cp1 == NULL)
return (0); /* Recursion happened. */
}
if (!LIST_EMPTY(&gp->consumer))
return (EINPROGRESS);
gp->softc = NULL;
KASSERT(sc->sc_provider == NULL, ("Provider still exists? (device=%s)",
gp->name));
while ((disk = TAILQ_FIRST(&sc->sc_disks)) != NULL) {
TAILQ_REMOVE(&sc->sc_disks, disk, d_next);
free(disk, M_CONCAT);
}
mtx_destroy(&sc->sc_completion_lock);
sx_destroy(&sc->sc_disks_lock);
free(sc, M_CONCAT);
G_CONCAT_DEBUG(0, "Device %s destroyed.", gp->name);
g_wither_geom(gp, ENXIO);
return (0);
}
static int
g_concat_destroy_geom(struct gctl_req *req __unused,
struct g_class *mp __unused, struct g_geom *gp)
{
struct g_concat_softc *sc;
sc = gp->softc;
return (g_concat_destroy(sc, 0));
}
static struct g_geom *
g_concat_taste(struct g_class *mp, struct g_provider *pp, int flags __unused)
{
struct g_concat_metadata md;
struct g_concat_softc *sc;
struct g_consumer *cp;
struct g_geom *gp;
int error;
g_trace(G_T_TOPOLOGY, "%s(%s, %s)", __func__, mp->name, pp->name);
g_topology_assert();
/* Skip providers that are already open for writing. */
if (pp->acw > 0)
return (NULL);
G_CONCAT_DEBUG(3, "Tasting %s.", pp->name);
gp = g_new_geomf(mp, "concat:taste");
gp->start = g_concat_start;
gp->access = g_concat_access;
gp->orphan = g_concat_orphan;
cp = g_new_consumer(gp);
error = g_attach(cp, pp);
if (error == 0) {
error = g_concat_read_metadata(cp, &md);
g_detach(cp);
}
g_destroy_consumer(cp);
g_destroy_geom(gp);
if (error != 0)
return (NULL);
gp = NULL;
if (strcmp(md.md_magic, G_CONCAT_MAGIC) != 0)
return (NULL);
if (md.md_version > G_CONCAT_VERSION) {
printf("geom_concat.ko module is too old to handle %s.\n",
pp->name);
return (NULL);
}
/*
* Backward compatibility:
*/
/* There was no md_provider field in earlier versions of metadata. */
if (md.md_version < 3)
bzero(md.md_provider, sizeof(md.md_provider));
/* There was no md_provsize field in earlier versions of metadata. */
if (md.md_version < 4)
md.md_provsize = pp->mediasize;
if (md.md_provider[0] != '\0' &&
!g_compare_names(md.md_provider, pp->name))
return (NULL);
if (md.md_provsize != pp->mediasize)
return (NULL);
/*
* Let's check if device already exists.
*/
sc = NULL;
LIST_FOREACH(gp, &mp->geom, geom) {
sc = gp->softc;
if (sc == NULL)
continue;
if (sc->sc_type != G_CONCAT_TYPE_AUTOMATIC)
continue;
if (strcmp(md.md_name, sc->sc_name) != 0)
continue;
if (md.md_id != sc->sc_id)
continue;
break;
}
if (gp != NULL) {
G_CONCAT_DEBUG(1, "Adding disk %s to %s.", pp->name, gp->name);
error = g_concat_add_disk(sc, pp, md.md_no);
if (error != 0) {
G_CONCAT_DEBUG(0,
"Cannot add disk %s to %s (error=%d).", pp->name,
gp->name, error);
return (NULL);
}
} else {
gp = g_concat_create(mp, &md, G_CONCAT_TYPE_AUTOMATIC);
if (gp == NULL) {
G_CONCAT_DEBUG(0, "Cannot create device %s.",
md.md_name);
return (NULL);
}
sc = gp->softc;
G_CONCAT_DEBUG(1, "Adding disk %s to %s.", pp->name, gp->name);
error = g_concat_add_disk(sc, pp, md.md_no);
if (error != 0) {
G_CONCAT_DEBUG(0,
"Cannot add disk %s to %s (error=%d).", pp->name,
gp->name, error);
g_concat_destroy(sc, 1);
return (NULL);
}
}
return (gp);
}
static void
g_concat_ctl_create(struct gctl_req *req, struct g_class *mp)
{
u_int attached, no;
struct g_concat_metadata md;
struct g_provider *pp;
struct g_concat_softc *sc;
struct g_geom *gp;
struct sbuf *sb;
const char *name;
char param[16];
int *nargs;
g_topology_assert();
nargs = gctl_get_paraml(req, "nargs", sizeof(*nargs));
if (nargs == NULL) {
gctl_error(req, "No '%s' argument.", "nargs");
return;
}
if (*nargs < 2) {
gctl_error(req, "Too few arguments.");
return;
}
bzero(&md, sizeof(md));
strlcpy(md.md_magic, G_CONCAT_MAGIC, sizeof(md.md_magic));
md.md_version = G_CONCAT_VERSION;
name = gctl_get_asciiparam(req, "arg0");
if (name == NULL) {
gctl_error(req, "No 'arg%u' argument.", 0);
return;
}
strlcpy(md.md_name, name, sizeof(md.md_name));
md.md_id = arc4random();
md.md_no = 0;
md.md_all = *nargs - 1;
/* This field is not important here. */
md.md_provsize = 0;
/* Check all providers are valid */
for (no = 1; no < *nargs; no++) {
snprintf(param, sizeof(param), "arg%u", no);
pp = gctl_get_provider(req, param);
if (pp == NULL)
return;
}
gp = g_concat_create(mp, &md, G_CONCAT_TYPE_MANUAL);
if (gp == NULL) {
gctl_error(req, "Can't configure %s.", md.md_name);
return;
}
sc = gp->softc;
sb = sbuf_new_auto();
sbuf_printf(sb, "Can't attach disk(s) to %s:", gp->name);
for (attached = 0, no = 1; no < *nargs; no++) {
snprintf(param, sizeof(param), "arg%u", no);
pp = gctl_get_provider(req, param);
if (pp == NULL) {
name = gctl_get_asciiparam(req, param);
MPASS(name != NULL);
sbuf_printf(sb, " %s", name);
continue;
}
if (g_concat_add_disk(sc, pp, no - 1) != 0) {
G_CONCAT_DEBUG(1, "Disk %u (%s) not attached to %s.",
no, pp->name, gp->name);
sbuf_printf(sb, " %s", pp->name);
continue;
}
attached++;
}
sbuf_finish(sb);
if (md.md_all != attached) {
g_concat_destroy(gp->softc, 1);
gctl_error(req, "%s", sbuf_data(sb));
}
sbuf_delete(sb);
}
static struct g_concat_softc *
g_concat_find_device(struct g_class *mp, const char *name)
{
struct g_concat_softc *sc;
struct g_geom *gp;
if (strncmp(name, _PATH_DEV, strlen(_PATH_DEV)) == 0)
name += strlen(_PATH_DEV);
LIST_FOREACH(gp, &mp->geom, geom) {
sc = gp->softc;
if (sc == NULL)
continue;
if (strcmp(sc->sc_name, name) == 0)
return (sc);
}
return (NULL);
}
static void
g_concat_ctl_destroy(struct gctl_req *req, struct g_class *mp)
{
struct g_concat_softc *sc;
int *force, *nargs, error;
const char *name;
char param[16];
u_int i;
g_topology_assert();
nargs = gctl_get_paraml(req, "nargs", sizeof(*nargs));
if (nargs == NULL) {
gctl_error(req, "No '%s' argument.", "nargs");
return;
}
if (*nargs <= 0) {
gctl_error(req, "Missing device(s).");
return;
}
force = gctl_get_paraml(req, "force", sizeof(*force));
if (force == NULL) {
gctl_error(req, "No '%s' argument.", "force");
return;
}
for (i = 0; i < (u_int)*nargs; i++) {
snprintf(param, sizeof(param), "arg%u", i);
name = gctl_get_asciiparam(req, param);
if (name == NULL) {
gctl_error(req, "No 'arg%u' argument.", i);
return;
}
sc = g_concat_find_device(mp, name);
if (sc == NULL) {
gctl_error(req, "No such device: %s.", name);
return;
}
error = g_concat_destroy(sc, *force);
if (error != 0) {
gctl_error(req, "Cannot destroy device %s (error=%d).",
sc->sc_name, error);
return;
}
}
}
static struct g_concat_disk *
g_concat_find_disk(struct g_concat_softc *sc, const char *name)
{
struct g_concat_disk *disk;
sx_assert(&sc->sc_disks_lock, SX_LOCKED);
if (strncmp(name, "/dev/", 5) == 0)
name += 5;
TAILQ_FOREACH(disk, &sc->sc_disks, d_next) {
if (disk->d_consumer == NULL)
continue;
if (disk->d_consumer->provider == NULL)
continue;
if (strcmp(disk->d_consumer->provider->name, name) == 0)
return (disk);
}
return (NULL);
}
static void
g_concat_write_metadata(struct gctl_req *req, struct g_concat_softc *sc)
{
u_int no = 0;
struct g_concat_disk *disk;
struct g_concat_metadata md;
struct g_provider *pp;
u_char *sector;
int error;
bzero(&md, sizeof(md));
strlcpy(md.md_magic, G_CONCAT_MAGIC, sizeof(md.md_magic));
md.md_version = G_CONCAT_VERSION;
strlcpy(md.md_name, sc->sc_name, sizeof(md.md_name));
md.md_id = sc->sc_id;
md.md_all = sc->sc_ndisks;
TAILQ_FOREACH(disk, &sc->sc_disks, d_next) {
pp = disk->d_consumer->provider;
md.md_no = no;
if (disk->d_hardcoded)
strlcpy(md.md_provider, pp->name,
sizeof(md.md_provider));
md.md_provsize = disk->d_consumer->provider->mediasize;
sector = g_malloc(pp->sectorsize, M_WAITOK | M_ZERO);
concat_metadata_encode(&md, sector);
error = g_access(disk->d_consumer, 0, 1, 0);
if (error == 0) {
error = g_write_data(disk->d_consumer,
pp->mediasize - pp->sectorsize, sector,
pp->sectorsize);
(void)g_access(disk->d_consumer, 0, -1, 0);
}
g_free(sector);
if (error != 0)
gctl_error(req, "Cannot store metadata on %s: %d",
pp->name, error);
no++;
}
}
static void
g_concat_ctl_append(struct gctl_req *req, struct g_class *mp)
{
struct g_concat_softc *sc;
struct g_consumer *cp, *fcp;
struct g_provider *pp;
struct g_geom *gp;
const char *name, *cname;
struct g_concat_disk *disk;
int *nargs, *hardcode;
int error;
int disk_candelete;
g_topology_assert();
nargs = gctl_get_paraml(req, "nargs", sizeof(*nargs));
if (nargs == NULL) {
gctl_error(req, "No '%s' argument.", "nargs");
return;
}
if (*nargs != 2) {
gctl_error(req, "Invalid number of arguments.");
return;
}
hardcode = gctl_get_paraml(req, "hardcode", sizeof(*hardcode));
if (hardcode == NULL) {
gctl_error(req, "No '%s' argument.", "hardcode");
return;
}
cname = gctl_get_asciiparam(req, "arg0");
if (cname == NULL) {
gctl_error(req, "No 'arg%u' argument.", 0);
return;
}
sc = g_concat_find_device(mp, cname);
if (sc == NULL) {
gctl_error(req, "No such device: %s.", cname);
return;
}
if (sc->sc_provider == NULL) {
/*
* this won't race with g_concat_remove_disk as both
* are holding the topology lock
*/
gctl_error(req, "Device not active, can't append: %s.", cname);
return;
}
G_CONCAT_DEBUG(1, "Appending to %s:", cname);
sx_xlock(&sc->sc_disks_lock);
gp = sc->sc_geom;
fcp = LIST_FIRST(&gp->consumer);
name = gctl_get_asciiparam(req, "arg1");
if (name == NULL) {
gctl_error(req, "No 'arg%u' argument.", 1);
goto fail;
}
if (strncmp(name, "/dev/", strlen("/dev/")) == 0)
name += strlen("/dev/");
pp = g_provider_by_name(name);
if (pp == NULL) {
G_CONCAT_DEBUG(1, "Disk %s is invalid.", name);
gctl_error(req, "Disk %s is invalid.", name);
goto fail;
}
G_CONCAT_DEBUG(1, "Appending %s to this", name);
if (g_concat_find_disk(sc, name) != NULL) {
gctl_error(req, "Disk %s already appended.", name);
goto fail;
}
if ((sc->sc_provider->sectorsize % pp->sectorsize) != 0) {
gctl_error(req, "Providers sectorsize mismatch: %u vs %u",
sc->sc_provider->sectorsize, pp->sectorsize);
goto fail;
}
cp = g_new_consumer(gp);
cp->flags |= G_CF_DIRECT_SEND | G_CF_DIRECT_RECEIVE;
error = g_attach(cp, pp);
if (error != 0) {
g_destroy_consumer(cp);
gctl_error(req, "Cannot open device %s (error=%d).",
name, error);
goto fail;
}
error = g_access(cp, 1, 0, 0);
if (error == 0) {
error = g_getattr("GEOM::candelete", cp, &disk_candelete);
if (error != 0)
disk_candelete = 0;
(void)g_access(cp, -1, 0, 0);
} else
G_CONCAT_DEBUG(1, "Failed to access disk %s, error %d.", name, error);
/* invoke g_access exactly as deep as all the other members currently are */
if (fcp != NULL && (fcp->acr > 0 || fcp->acw > 0 || fcp->ace > 0)) {
error = g_access(cp, fcp->acr, fcp->acw, fcp->ace);
if (error != 0) {
g_detach(cp);
g_destroy_consumer(cp);
gctl_error(req, "Failed to access disk %s (error=%d).", name, error);
goto fail;
}
}
disk = malloc(sizeof(*disk), M_CONCAT, M_WAITOK | M_ZERO);
disk->d_consumer = cp;
disk->d_softc = sc;
disk->d_start = TAILQ_LAST(&sc->sc_disks, g_concat_disks)->d_end;
disk->d_end = disk->d_start + cp->provider->mediasize;
disk->d_candelete = disk_candelete;
disk->d_removed = 0;
disk->d_hardcoded = *hardcode;
cp->private = disk;
TAILQ_INSERT_TAIL(&sc->sc_disks, disk, d_next);
sc->sc_ndisks++;
if (sc->sc_type == G_CONCAT_TYPE_AUTOMATIC) {
/* last sector is for metadata */
disk->d_end -= cp->provider->sectorsize;
/* update metadata on all parts */
g_concat_write_metadata(req, sc);
}
g_resize_provider(sc->sc_provider, disk->d_end);
fail:
sx_xunlock(&sc->sc_disks_lock);
}
static void
g_concat_config(struct gctl_req *req, struct g_class *mp, const char *verb)
{
uint32_t *version;
g_topology_assert();
version = gctl_get_paraml(req, "version", sizeof(*version));
if (version == NULL) {
gctl_error(req, "No '%s' argument.", "version");
return;
}
if (*version != G_CONCAT_VERSION) {
gctl_error(req, "Userland and kernel parts are out of sync.");
return;
}
if (strcmp(verb, "create") == 0) {
g_concat_ctl_create(req, mp);
return;
} else if (strcmp(verb, "destroy") == 0 ||
strcmp(verb, "stop") == 0) {
g_concat_ctl_destroy(req, mp);
return;
} else if (strcmp(verb, "append") == 0) {
g_concat_ctl_append(req, mp);
return;
}
gctl_error(req, "Unknown verb.");
}
static void
g_concat_dumpconf(struct sbuf *sb, const char *indent, struct g_geom *gp,
struct g_consumer *cp, struct g_provider *pp)
{
struct g_concat_softc *sc;
g_topology_assert();
sc = gp->softc;
if (sc == NULL)
return;
sx_slock(&sc->sc_disks_lock);
if (pp != NULL) {
/* Nothing here. */
} else if (cp != NULL) {
struct g_concat_disk *disk;
disk = cp->private;
if (disk == NULL)
goto end;
sbuf_printf(sb, "%s<End>%jd</End>\n", indent,
(intmax_t)disk->d_end);
sbuf_printf(sb, "%s<Start>%jd</Start>\n", indent,
(intmax_t)disk->d_start);
} else {
sbuf_printf(sb, "%s<ID>%u</ID>\n", indent, (u_int)sc->sc_id);
sbuf_printf(sb, "%s<Type>", indent);
switch (sc->sc_type) {
case G_CONCAT_TYPE_AUTOMATIC:
sbuf_cat(sb, "AUTOMATIC");
break;
case G_CONCAT_TYPE_MANUAL:
sbuf_cat(sb, "MANUAL");
break;
default:
sbuf_cat(sb, "UNKNOWN");
break;
}
sbuf_cat(sb, "</Type>\n");
sbuf_printf(sb, "%s<Status>Total=%u, Online=%u</Status>\n",
indent, sc->sc_ndisks, g_concat_nvalid(sc));
sbuf_printf(sb, "%s<State>", indent);
if (sc->sc_provider != NULL && sc->sc_provider->error == 0)
sbuf_cat(sb, "UP");
else
sbuf_cat(sb, "DOWN");
sbuf_cat(sb, "</State>\n");
}
end:
sx_sunlock(&sc->sc_disks_lock);
}
DECLARE_GEOM_CLASS(g_concat_class, g_concat);
MODULE_VERSION(geom_concat, 0);