d/freebsd-nq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

5244 zio pipeline callers should explicitly invoke next stage

Browse Source 
Reviewed by: Adam Leventhal <ahl@delphix.com>
Reviewed by: Alex Reece <alex.reece@delphix.com>
Reviewed by: Christopher Siden <christopher.siden@delphix.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Richard Elling <richard.elling@gmail.com>
Reviewed by: Dan McDonald <danmcd@omniti.com>
Reviewed by: Steven Hartland <killing@multiplay.co.uk>
Approved by: Gordon Ross <gwr@nexenta.com>
Author: George Wilson <george.wilson@delphix.com>

illumos/illumos-gate@738f37bc3d
This commit is contained in:
Xin LI 2014-11-08 06:34:37 +00:00
parent f41c6565a5
commit 6c426f3bd4
9 changed files with 52 additions and 31 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
lib/libzpool/common/taskq.c
View File

@ -25,6 +25,7 @@
/*
* Copyright 2011 Nexenta Systems, Inc. All rights reserved.
* Copyright 2012 Garrett D'Amore <garrett@damore.org>. All rights reserved.
* Copyright (c) 2014 by Delphix. All rights reserved.
*/
#include <sys/zfs_context.h>
@ -33,8 +34,10 @@ int taskq_now;
taskq_t *system_taskq;
#define TASKQ_ACTIVE 0x00010000
#define TASKQ_NAMELEN 31
struct taskq {
char tq_name[TASKQ_NAMELEN + 1];
kmutex_t tq_lock;
krwlock_t tq_threadlock;
kcondvar_t tq_dispatch_cv;
@ -247,6 +250,7 @@ taskq_create(const char *name, int nthreads, pri_t pri,
cv_init(&tq->tq_dispatch_cv, NULL, CV_DEFAULT, NULL);
cv_init(&tq->tq_wait_cv, NULL, CV_DEFAULT, NULL);
cv_init(&tq->tq_maxalloc_cv, NULL, CV_DEFAULT, NULL);
(void) strncpy(tq->tq_name, name, TASKQ_NAMELEN + 1);
tq->tq_flags = flags | TASKQ_ACTIVE;
tq->tq_active = nthreads;
tq->tq_nthreads = nthreads;

2
uts/common/fs/zfs/sys/vdev_impl.h
View File

@ -60,7 +60,7 @@ typedef int vdev_open_func_t(vdev_t *vd, uint64_t *size, uint64_t *max_size,
uint64_t *ashift);
typedef void vdev_close_func_t(vdev_t *vd);
typedef uint64_t vdev_asize_func_t(vdev_t *vd, uint64_t psize);
typedef int vdev_io_start_func_t(zio_t *zio);
typedef void vdev_io_start_func_t(zio_t *zio);
typedef void vdev_io_done_func_t(zio_t *zio);
typedef void vdev_state_change_func_t(vdev_t *vd, int, int);
typedef void vdev_hold_func_t(vdev_t *vd);

3
uts/common/fs/zfs/sys/zio.h
View File

@ -151,9 +151,6 @@ typedef enum zio_priority {
ZIO_PRIORITY_NOW /* non-queued i/os (e.g. free) */
} zio_priority_t;
#define ZIO_PIPELINE_CONTINUE 0x100
#define ZIO_PIPELINE_STOP 0x101
enum zio_flag {
/*
* Flags inherited by gang, ddt, and vdev children,

15
uts/common/fs/zfs/vdev_disk.c
View File

@ -715,7 +715,7 @@ vdev_disk_ioctl_done(void *zio_arg, int error)
zio_interrupt(zio);
}
static int
static void
vdev_disk_io_start(zio_t *zio)
{
vdev_t *vd = zio->io_vd;
@ -731,14 +731,16 @@ vdev_disk_io_start(zio_t *zio)
*/
if (dvd == NULL || (dvd->vd_ldi_offline && dvd->vd_lh == NULL)) {
zio->io_error = ENXIO;
return (ZIO_PIPELINE_CONTINUE);
zio_interrupt(zio);
return;
}
if (zio->io_type == ZIO_TYPE_IOCTL) {
/* XXPOLICY */
if (!vdev_readable(vd)) {
zio->io_error = SET_ERROR(ENXIO);
return (ZIO_PIPELINE_CONTINUE);
zio_interrupt(zio);
return;
}
switch (zio->io_cmd) {
@ -769,7 +771,7 @@ vdev_disk_io_start(zio_t *zio)
* and will call vdev_disk_ioctl_done()
* upon completion.
*/
return (ZIO_PIPELINE_STOP);
return;
}
if (error == ENOTSUP || error == ENOTTY) {
@ -790,7 +792,8 @@ vdev_disk_io_start(zio_t *zio)
zio->io_error = SET_ERROR(ENOTSUP);
}
return (ZIO_PIPELINE_CONTINUE);
zio_execute(zio);
return;
}
vb = kmem_alloc(sizeof (vdev_buf_t), KM_SLEEP);
@ -811,8 +814,6 @@ vdev_disk_io_start(zio_t *zio)
/* ldi_strategy() will return non-zero only on programming errors */
VERIFY(ldi_strategy(dvd->vd_lh, bp) == 0);
return (ZIO_PIPELINE_STOP);
}
static void

12
uts/common/fs/zfs/vdev_file.c
View File

@ -20,7 +20,7 @@
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2013 by Delphix. All rights reserved.
* Copyright (c) 2011, 2014 by Delphix. All rights reserved.
*/
#include <sys/zfs_context.h>
@ -182,7 +182,7 @@ vdev_file_io_strategy(void *arg)
}
}
static int
static void
vdev_file_io_start(zio_t *zio)
{
vdev_t *vd = zio->io_vd;
@ -194,7 +194,8 @@ vdev_file_io_start(zio_t *zio)
/* XXPOLICY */
if (!vdev_readable(vd)) {
zio->io_error = SET_ERROR(ENXIO);
return (ZIO_PIPELINE_CONTINUE);
zio_interrupt(zio);
return;
}
switch (zio->io_cmd) {
@ -206,7 +207,8 @@ vdev_file_io_start(zio_t *zio)
zio->io_error = SET_ERROR(ENOTSUP);
}
return (ZIO_PIPELINE_CONTINUE);
zio_execute(zio);
return;
}
vb = kmem_alloc(sizeof (vdev_buf_t), KM_SLEEP);
@ -225,8 +227,6 @@ vdev_file_io_start(zio_t *zio)
VERIFY3U(taskq_dispatch(system_taskq, vdev_file_io_strategy, bp,
TQ_SLEEP), !=, 0);
return (ZIO_PIPELINE_STOP);
}
/* ARGSUSED */

9
uts/common/fs/zfs/vdev_mirror.c
View File

@ -24,7 +24,7 @@
*/
/*
* Copyright (c) 2013 by Delphix. All rights reserved.
* Copyright (c) 2012, 2014 by Delphix. All rights reserved.
*/
#include <sys/zfs_context.h>
@ -260,7 +260,7 @@ vdev_mirror_child_select(zio_t *zio)
return (-1);
}
static int
static void
vdev_mirror_io_start(zio_t *zio)
{
mirror_map_t *mm;
@ -285,7 +285,8 @@ vdev_mirror_io_start(zio_t *zio)
zio->io_type, zio->io_priority, 0,
vdev_mirror_scrub_done, mc));
}
return (ZIO_PIPELINE_CONTINUE);
zio_execute(zio);
return;
}
/*
* For normal reads just pick one child.
@ -311,7 +312,7 @@ vdev_mirror_io_start(zio_t *zio)
c++;
}
return (ZIO_PIPELINE_CONTINUE);
zio_execute(zio);
}
static int

6
uts/common/fs/zfs/vdev_missing.c
View File

@ -24,7 +24,7 @@
*/
/*
* Copyright (c) 2013 by Delphix. All rights reserved.
* Copyright (c) 2012, 2014 by Delphix. All rights reserved.
*/
/*
@ -66,11 +66,11 @@ vdev_missing_close(vdev_t *vd)
}
/* ARGSUSED */
static int
static void
vdev_missing_io_start(zio_t *zio)
{
zio->io_error = SET_ERROR(ENOTSUP);
return (ZIO_PIPELINE_CONTINUE);
zio_execute(zio);
}
/* ARGSUSED */

9
uts/common/fs/zfs/vdev_raidz.c
View File

@ -21,7 +21,7 @@
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2013 by Delphix. All rights reserved.
* Copyright (c) 2012, 2014 by Delphix. All rights reserved.
* Copyright (c) 2013, Joyent, Inc. All rights reserved.
*/
@ -1711,7 +1711,7 @@ vdev_raidz_child_done(zio_t *zio)
* vdevs have had errors, then create zio read operations to the parity
* columns' VDevs as well.
*/
static int
static void
vdev_raidz_io_start(zio_t *zio)
{
vdev_t *vd = zio->io_vd;
@ -1759,7 +1759,8 @@ vdev_raidz_io_start(zio_t *zio)
ZIO_FLAG_NODATA | ZIO_FLAG_OPTIONAL, NULL, NULL));
}
return (ZIO_PIPELINE_CONTINUE);
zio_execute(zio);
return;
}
ASSERT(zio->io_type == ZIO_TYPE_READ);
@ -1799,7 +1800,7 @@ vdev_raidz_io_start(zio_t *zio)
}
}
return (ZIO_PIPELINE_CONTINUE);
zio_execute(zio);
}

23
uts/common/fs/zfs/zio.c
View File

@ -63,6 +63,9 @@ kmem_cache_t *zio_data_buf_cache[SPA_MAXBLOCKSIZE >> SPA_MINBLOCKSHIFT];
extern vmem_t *zio_alloc_arena;
#endif
#define ZIO_PIPELINE_CONTINUE 0x100
#define ZIO_PIPELINE_STOP 0x101
/*
* The following actions directly effect the spa's sync-to-convergence logic.
* The values below define the sync pass when we start performing the action.
@ -2466,6 +2469,18 @@ zio_free_zil(spa_t *spa, uint64_t txg, blkptr_t *bp)
* Read and write to physical devices
* ==========================================================================
*/
/*
* Issue an I/O to the underlying vdev. Typically the issue pipeline
* stops after this stage and will resume upon I/O completion.
* However, there are instances where the vdev layer may need to
* continue the pipeline when an I/O was not issued. Since the I/O
* that was sent to the vdev layer might be different than the one
* currently active in the pipeline (see vdev_queue_io()), we explicitly
* force the underlying vdev layers to call either zio_execute() or
* zio_interrupt() to ensure that the pipeline continues with the correct I/O.
*/
static int
zio_vdev_io_start(zio_t *zio)
{
@ -2483,7 +2498,8 @@ zio_vdev_io_start(zio_t *zio)
/*
* The mirror_ops handle multiple DVAs in a single BP.
*/
return (vdev_mirror_ops.vdev_op_io_start(zio));
vdev_mirror_ops.vdev_op_io_start(zio);
return (ZIO_PIPELINE_STOP);
}
/*
@ -2491,7 +2507,7 @@ zio_vdev_io_start(zio_t *zio)
* can quickly react to certain workloads. In particular, we care
* about non-scrubbing, top-level reads and writes with the following
* characteristics:
* - synchronous writes of user data to non-slog devices
* - synchronous writes of user data to non-slog devices
* - any reads of user data
* When these conditions are met, adjust the timestamp of spa_last_io
* which allows the scan thread to adjust its workload accordingly.
@ -2577,7 +2593,8 @@ zio_vdev_io_start(zio_t *zio)
}
}
return (vd->vdev_ops->vdev_op_io_start(zio));
vd->vdev_ops->vdev_op_io_start(zio);
return (ZIO_PIPELINE_STOP);
}
static int