freebsd-nq/sys/geom/vinum/geom_vinum_var.h

389 lines
14 KiB
C
Raw Normal View History

/*-
Import the gvinum work that have been done during and after Summer of Code 2007. The work have been under testing and fixing since then, and it is mature enough to be put into HEAD for further testing. A lot have changed in this time, and here are the most important: - Gvinum now uses one single workerthread instead of one thread for each volume and each plex. The reason for this is that the previous scheme was very complex, and was the cause of many of the bugs discovered in gvinum. Instead, gvinum now uses one worker thread with an event queue, quite similar to what used in gmirror. - The rebuild/grow/initialize/parity check routines no longer runs in separate threads, but are run as regular I/O requests with special flags. This made it easier to support mounted growing and parity rebuild. - Support for growing striped and raid5-plexes, meaning that one can extend the volumes for these plex types in addition to the concat type. Also works while the volume is mounted. - Implementation of many of the missing commands from the old vinum: attach/detach, start (was partially implemented), stop (was partially implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes with one plex of these organizations). - The parity check and rebuild no longer goes between userland/kernel, meaning that the gvinum command will not stay and wait forever for the rebuild to finish. You can instead watch the status with the list command. - Many problems with gvinum have been reported since 5.x, and some has been hard to fix due to the complicated architecture. Hopefully, it should be more stable and better handle edge cases that previously made gvinum crash. - Failed drives no longer disappears entirely, but now leave behind a dummy drive that makes sure the original state is not forgotten in case the system is rebooted between drive failures/swaps. - Update manpage to reflect new commands and extend it with some examples. Sponsored by: Google Summer of Code 2007 Mentored by: le Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
* Copyright (c) 2004, 2007 Lukas Ertl
* Copyright (c) 1997, 1998, 1999
* Nan Yang Computer Services Limited. All rights reserved.
*
* Parts copyright (c) 1997, 1998 Cybernet Corporation, NetMAX project.
* Parts written by Greg Lehey.
*
* This software is distributed under the so-called ``Berkeley
* License'': *
* 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.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by Nan Yang Computer
* Services Limited.
* 4. Neither the name of the Company nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* This software is provided ``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 company 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.
*
* $FreeBSD$
*/
#ifndef _GEOM_VINUM_VAR_H_
#define _GEOM_VINUM_VAR_H_
/*
* Slice header
*
* Vinum drives start with this structure:
*
*\ Sector
* |--------------------------------------|
* | PDP-11 memorial boot block | 0
* |--------------------------------------|
* | Disk label, maybe | 1
* |--------------------------------------|
* | Slice definition (vinum_hdr) | 8
* |--------------------------------------|
* | |
* | Configuration info, first copy | 9
* | |
* |--------------------------------------|
* | |
* | Configuration info, second copy | 9 + size of config
* | |
* |--------------------------------------|
*/
/* Sizes and offsets of our information. */
#define GV_HDR_OFFSET 4096 /* Offset of vinum header. */
#define GV_HDR_LEN 512 /* Size of vinum header. */
#define GV_CFG_OFFSET 4608 /* Offset of first config copy. */
#define GV_CFG_LEN 65536 /* Size of config copy. */
/* This is where the actual data starts. */
#define GV_DATA_START (GV_CFG_LEN * 2 + GV_CFG_OFFSET)
/* #define GV_DATA_START (GV_CFG_LEN * 2 + GV_HDR_LEN) */
#define GV_MAXDRIVENAME 32 /* Maximum length of a device name. */
#define GV_MAXSDNAME 64 /* Maximum length of a subdisk name. */
#define GV_MAXPLEXNAME 64 /* Maximum length of a plex name. */
#define GV_MAXVOLNAME 64 /* Maximum length of a volume name. */
/* Command line flags. */
#define GV_FLAG_R 0x01
#define GV_FLAG_S 0x02
#define GV_FLAG_V 0x04
#define GV_FLAG_VV 0x08
#define GV_FLAG_F 0x10
/* Object types. */
#define GV_TYPE_VOL 1
#define GV_TYPE_PLEX 2
#define GV_TYPE_SD 3
#define GV_TYPE_DRIVE 4
/* State changing flags. */
#define GV_SETSTATE_FORCE 0x1
#define GV_SETSTATE_CONFIG 0x2
/* Subdisk state bitmaps for plexes. */
#define GV_SD_DOWNSTATE 0x01 /* Subdisk is down. */
#define GV_SD_STALESTATE 0x02 /* Subdisk is stale. */
#define GV_SD_INITSTATE 0x04 /* Subdisk is initializing. */
#define GV_SD_UPSTATE 0x08 /* Subdisk is up. */
/* Synchronization/initialization request sizes. */
#define GV_MIN_SYNCSIZE 512
#define GV_MAX_SYNCSIZE MAXPHYS
#define GV_DFLT_SYNCSIZE 65536
/* Flags for BIOs, as they are processed within vinum. */
#define GV_BIO_GROW 0x01
#define GV_BIO_MALLOC 0x02
#define GV_BIO_ONHOLD 0x04
#define GV_BIO_SYNCREQ 0x08
Import the gvinum work that have been done during and after Summer of Code 2007. The work have been under testing and fixing since then, and it is mature enough to be put into HEAD for further testing. A lot have changed in this time, and here are the most important: - Gvinum now uses one single workerthread instead of one thread for each volume and each plex. The reason for this is that the previous scheme was very complex, and was the cause of many of the bugs discovered in gvinum. Instead, gvinum now uses one worker thread with an event queue, quite similar to what used in gmirror. - The rebuild/grow/initialize/parity check routines no longer runs in separate threads, but are run as regular I/O requests with special flags. This made it easier to support mounted growing and parity rebuild. - Support for growing striped and raid5-plexes, meaning that one can extend the volumes for these plex types in addition to the concat type. Also works while the volume is mounted. - Implementation of many of the missing commands from the old vinum: attach/detach, start (was partially implemented), stop (was partially implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes with one plex of these organizations). - The parity check and rebuild no longer goes between userland/kernel, meaning that the gvinum command will not stay and wait forever for the rebuild to finish. You can instead watch the status with the list command. - Many problems with gvinum have been reported since 5.x, and some has been hard to fix due to the complicated architecture. Hopefully, it should be more stable and better handle edge cases that previously made gvinum crash. - Failed drives no longer disappears entirely, but now leave behind a dummy drive that makes sure the original state is not forgotten in case the system is rebooted between drive failures/swaps. - Update manpage to reflect new commands and extend it with some examples. Sponsored by: Google Summer of Code 2007 Mentored by: le Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
#define GV_BIO_INIT 0x10
#define GV_BIO_REBUILD 0x20
2004-11-26 12:01:00 +00:00
#define GV_BIO_CHECK 0x40
#define GV_BIO_PARITY 0x80
Import the gvinum work that have been done during and after Summer of Code 2007. The work have been under testing and fixing since then, and it is mature enough to be put into HEAD for further testing. A lot have changed in this time, and here are the most important: - Gvinum now uses one single workerthread instead of one thread for each volume and each plex. The reason for this is that the previous scheme was very complex, and was the cause of many of the bugs discovered in gvinum. Instead, gvinum now uses one worker thread with an event queue, quite similar to what used in gmirror. - The rebuild/grow/initialize/parity check routines no longer runs in separate threads, but are run as regular I/O requests with special flags. This made it easier to support mounted growing and parity rebuild. - Support for growing striped and raid5-plexes, meaning that one can extend the volumes for these plex types in addition to the concat type. Also works while the volume is mounted. - Implementation of many of the missing commands from the old vinum: attach/detach, start (was partially implemented), stop (was partially implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes with one plex of these organizations). - The parity check and rebuild no longer goes between userland/kernel, meaning that the gvinum command will not stay and wait forever for the rebuild to finish. You can instead watch the status with the list command. - Many problems with gvinum have been reported since 5.x, and some has been hard to fix due to the complicated architecture. Hopefully, it should be more stable and better handle edge cases that previously made gvinum crash. - Failed drives no longer disappears entirely, but now leave behind a dummy drive that makes sure the original state is not forgotten in case the system is rebooted between drive failures/swaps. - Update manpage to reflect new commands and extend it with some examples. Sponsored by: Google Summer of Code 2007 Mentored by: le Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
#define GV_BIO_INTERNAL \
(GV_BIO_SYNCREQ | GV_BIO_INIT | GV_BIO_REBUILD | GV_BIO_CHECK | GV_BIO_GROW)
Import the gvinum work that have been done during and after Summer of Code 2007. The work have been under testing and fixing since then, and it is mature enough to be put into HEAD for further testing. A lot have changed in this time, and here are the most important: - Gvinum now uses one single workerthread instead of one thread for each volume and each plex. The reason for this is that the previous scheme was very complex, and was the cause of many of the bugs discovered in gvinum. Instead, gvinum now uses one worker thread with an event queue, quite similar to what used in gmirror. - The rebuild/grow/initialize/parity check routines no longer runs in separate threads, but are run as regular I/O requests with special flags. This made it easier to support mounted growing and parity rebuild. - Support for growing striped and raid5-plexes, meaning that one can extend the volumes for these plex types in addition to the concat type. Also works while the volume is mounted. - Implementation of many of the missing commands from the old vinum: attach/detach, start (was partially implemented), stop (was partially implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes with one plex of these organizations). - The parity check and rebuild no longer goes between userland/kernel, meaning that the gvinum command will not stay and wait forever for the rebuild to finish. You can instead watch the status with the list command. - Many problems with gvinum have been reported since 5.x, and some has been hard to fix due to the complicated architecture. Hopefully, it should be more stable and better handle edge cases that previously made gvinum crash. - Failed drives no longer disappears entirely, but now leave behind a dummy drive that makes sure the original state is not forgotten in case the system is rebooted between drive failures/swaps. - Update manpage to reflect new commands and extend it with some examples. Sponsored by: Google Summer of Code 2007 Mentored by: le Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
/* Error codes to be used within gvinum. */
#define GV_ERR_SETSTATE (-1) /* Error setting state. */
#define GV_ERR_BADSIZE (-2) /* Object has wrong size. */
#define GV_ERR_INVTYPE (-3) /* Invalid object type. */
#define GV_ERR_CREATE (-4) /* Error creating gvinum object. */
#define GV_ERR_ISBUSY (-5) /* Object is busy. */
#define GV_ERR_ISATTACHED (-6) /* Object is attached to another. */
#define GV_ERR_INVFLAG (-7) /* Invalid flag passed. */
#define GV_ERR_INVSTATE (-8) /* Invalid state. */
#define GV_ERR_NOTFOUND (-9) /* Object not found. */
#define GV_ERR_NAMETAKEN (-10) /* Object name is taken. */
#define GV_ERR_NOSPACE (-11) /* No space left on drive/subdisk. */
#define GV_ERR_BADOFFSET (-12) /* Invalid offset specified. */
#define GV_ERR_INVNAME (-13) /* Invalid object name. */
#define GV_ERR_PLEXORG (-14) /* Invalid plex organization. */
/*
* hostname is 256 bytes long, but we don't need to shlep multiple copies in
* vinum. We use the host name just to identify this system, and 32 bytes
* should be ample for that purpose.
*/
#define GV_HOSTNAME_LEN 32
struct gv_label {
char sysname[GV_HOSTNAME_LEN]; /* System name at creation time. */
char name[GV_MAXDRIVENAME]; /* Our name of the drive. */
struct timeval date_of_birth; /* The time it was created ... */
struct timeval last_update; /* ... and the time of last update. */
off_t drive_size; /* Total size incl. headers. */
};
/* The 'header' of each valid vinum drive. */
struct gv_hdr {
uint64_t magic;
#define GV_OLD_MAGIC 0x494E2056494E4F00LL
#define GV_OLD_NOMAGIC 0x4E4F2056494E4F00LL
#define GV_MAGIC 0x56494E554D2D3100LL
#define GV_NOMAGIC 0x56494E554D2D2D00LL
uint64_t config_length;
struct gv_label label;
};
/* A single freelist entry of a drive. */
struct gv_freelist {
off_t size; /* Size of this free slot. */
off_t offset; /* Offset on the drive. */
LIST_ENTRY(gv_freelist) freelist;
};
/*
* Since we share structures between userland and kernel, we need this helper
* struct instead of struct bio_queue_head and friends. Maybe I find a proper
* solution some day.
*/
struct gv_bioq {
struct bio *bp;
TAILQ_ENTRY(gv_bioq) queue;
};
Import the gvinum work that have been done during and after Summer of Code 2007. The work have been under testing and fixing since then, and it is mature enough to be put into HEAD for further testing. A lot have changed in this time, and here are the most important: - Gvinum now uses one single workerthread instead of one thread for each volume and each plex. The reason for this is that the previous scheme was very complex, and was the cause of many of the bugs discovered in gvinum. Instead, gvinum now uses one worker thread with an event queue, quite similar to what used in gmirror. - The rebuild/grow/initialize/parity check routines no longer runs in separate threads, but are run as regular I/O requests with special flags. This made it easier to support mounted growing and parity rebuild. - Support for growing striped and raid5-plexes, meaning that one can extend the volumes for these plex types in addition to the concat type. Also works while the volume is mounted. - Implementation of many of the missing commands from the old vinum: attach/detach, start (was partially implemented), stop (was partially implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes with one plex of these organizations). - The parity check and rebuild no longer goes between userland/kernel, meaning that the gvinum command will not stay and wait forever for the rebuild to finish. You can instead watch the status with the list command. - Many problems with gvinum have been reported since 5.x, and some has been hard to fix due to the complicated architecture. Hopefully, it should be more stable and better handle edge cases that previously made gvinum crash. - Failed drives no longer disappears entirely, but now leave behind a dummy drive that makes sure the original state is not forgotten in case the system is rebooted between drive failures/swaps. - Update manpage to reflect new commands and extend it with some examples. Sponsored by: Google Summer of Code 2007 Mentored by: le Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
#define GV_EVENT_DRIVE_TASTED 1
#define GV_EVENT_DRIVE_LOST 2
#define GV_EVENT_THREAD_EXIT 3
#define GV_EVENT_CREATE_DRIVE 4
#define GV_EVENT_CREATE_VOLUME 5
#define GV_EVENT_CREATE_PLEX 6
#define GV_EVENT_CREATE_SD 7
#define GV_EVENT_SAVE_CONFIG 8
#define GV_EVENT_RM_VOLUME 9
#define GV_EVENT_RM_PLEX 10
#define GV_EVENT_RM_SD 11
#define GV_EVENT_RM_DRIVE 12
#define GV_EVENT_SET_SD_STATE 13
#define GV_EVENT_SET_DRIVE_STATE 14
#define GV_EVENT_SET_VOL_STATE 15
#define GV_EVENT_SET_PLEX_STATE 16
#define GV_EVENT_RESET_CONFIG 17
#define GV_EVENT_PARITY_REBUILD 18
#define GV_EVENT_PARITY_CHECK 19
#define GV_EVENT_START_PLEX 20
#define GV_EVENT_START_VOLUME 21
#define GV_EVENT_ATTACH_PLEX 22
#define GV_EVENT_ATTACH_SD 23
#define GV_EVENT_DETACH_PLEX 24
#define GV_EVENT_DETACH_SD 25
#define GV_EVENT_RENAME_VOL 26
#define GV_EVENT_RENAME_PLEX 27
#define GV_EVENT_RENAME_SD 28
#define GV_EVENT_RENAME_DRIVE 29
#define GV_EVENT_MOVE_SD 30
#define GV_EVENT_SETUP_OBJECTS 31
#ifdef _KERNEL
struct gv_event {
int type;
void *arg1;
void *arg2;
intmax_t arg3;
intmax_t arg4;
TAILQ_ENTRY(gv_event) events;
};
/* This struct contains the main vinum config. */
struct gv_softc {
/* Linked lists of all objects in our setup. */
LIST_HEAD(,gv_drive) drives; /* All drives. */
LIST_HEAD(,gv_plex) plexes; /* All plexes. */
LIST_HEAD(,gv_sd) subdisks; /* All subdisks. */
LIST_HEAD(,gv_volume) volumes; /* All volumes. */
Import the gvinum work that have been done during and after Summer of Code 2007. The work have been under testing and fixing since then, and it is mature enough to be put into HEAD for further testing. A lot have changed in this time, and here are the most important: - Gvinum now uses one single workerthread instead of one thread for each volume and each plex. The reason for this is that the previous scheme was very complex, and was the cause of many of the bugs discovered in gvinum. Instead, gvinum now uses one worker thread with an event queue, quite similar to what used in gmirror. - The rebuild/grow/initialize/parity check routines no longer runs in separate threads, but are run as regular I/O requests with special flags. This made it easier to support mounted growing and parity rebuild. - Support for growing striped and raid5-plexes, meaning that one can extend the volumes for these plex types in addition to the concat type. Also works while the volume is mounted. - Implementation of many of the missing commands from the old vinum: attach/detach, start (was partially implemented), stop (was partially implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes with one plex of these organizations). - The parity check and rebuild no longer goes between userland/kernel, meaning that the gvinum command will not stay and wait forever for the rebuild to finish. You can instead watch the status with the list command. - Many problems with gvinum have been reported since 5.x, and some has been hard to fix due to the complicated architecture. Hopefully, it should be more stable and better handle edge cases that previously made gvinum crash. - Failed drives no longer disappears entirely, but now leave behind a dummy drive that makes sure the original state is not forgotten in case the system is rebooted between drive failures/swaps. - Update manpage to reflect new commands and extend it with some examples. Sponsored by: Google Summer of Code 2007 Mentored by: le Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
TAILQ_HEAD(,gv_event) equeue; /* Event queue. */
struct mtx equeue_mtx; /* Event queue lock. */
struct mtx bqueue_mtx; /* BIO queue lock. */
Import the gvinum work that have been done during and after Summer of Code 2007. The work have been under testing and fixing since then, and it is mature enough to be put into HEAD for further testing. A lot have changed in this time, and here are the most important: - Gvinum now uses one single workerthread instead of one thread for each volume and each plex. The reason for this is that the previous scheme was very complex, and was the cause of many of the bugs discovered in gvinum. Instead, gvinum now uses one worker thread with an event queue, quite similar to what used in gmirror. - The rebuild/grow/initialize/parity check routines no longer runs in separate threads, but are run as regular I/O requests with special flags. This made it easier to support mounted growing and parity rebuild. - Support for growing striped and raid5-plexes, meaning that one can extend the volumes for these plex types in addition to the concat type. Also works while the volume is mounted. - Implementation of many of the missing commands from the old vinum: attach/detach, start (was partially implemented), stop (was partially implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes with one plex of these organizations). - The parity check and rebuild no longer goes between userland/kernel, meaning that the gvinum command will not stay and wait forever for the rebuild to finish. You can instead watch the status with the list command. - Many problems with gvinum have been reported since 5.x, and some has been hard to fix due to the complicated architecture. Hopefully, it should be more stable and better handle edge cases that previously made gvinum crash. - Failed drives no longer disappears entirely, but now leave behind a dummy drive that makes sure the original state is not forgotten in case the system is rebooted between drive failures/swaps. - Update manpage to reflect new commands and extend it with some examples. Sponsored by: Google Summer of Code 2007 Mentored by: le Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
struct mtx config_mtx; /* Configuration lock. */
struct bio_queue_head *bqueue_down; /* BIO queue incoming
requests. */
struct bio_queue_head *bqueue_up; /* BIO queue for completed
requests. */
struct g_geom *geom; /* Pointer to our VINUM geom. */
};
#endif
/* softc for a drive. */
struct gv_drive {
char name[GV_MAXDRIVENAME]; /* The name of this drive. */
char device[GV_MAXDRIVENAME]; /* Associated device. */
int state; /* The state of this drive. */
#define GV_DRIVE_DOWN 0
#define GV_DRIVE_UP 1
off_t size; /* Size of this drive. */
off_t avail; /* Available space. */
int sdcount; /* Number of subdisks. */
int flags;
Import the gvinum work that have been done during and after Summer of Code 2007. The work have been under testing and fixing since then, and it is mature enough to be put into HEAD for further testing. A lot have changed in this time, and here are the most important: - Gvinum now uses one single workerthread instead of one thread for each volume and each plex. The reason for this is that the previous scheme was very complex, and was the cause of many of the bugs discovered in gvinum. Instead, gvinum now uses one worker thread with an event queue, quite similar to what used in gmirror. - The rebuild/grow/initialize/parity check routines no longer runs in separate threads, but are run as regular I/O requests with special flags. This made it easier to support mounted growing and parity rebuild. - Support for growing striped and raid5-plexes, meaning that one can extend the volumes for these plex types in addition to the concat type. Also works while the volume is mounted. - Implementation of many of the missing commands from the old vinum: attach/detach, start (was partially implemented), stop (was partially implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes with one plex of these organizations). - The parity check and rebuild no longer goes between userland/kernel, meaning that the gvinum command will not stay and wait forever for the rebuild to finish. You can instead watch the status with the list command. - Many problems with gvinum have been reported since 5.x, and some has been hard to fix due to the complicated architecture. Hopefully, it should be more stable and better handle edge cases that previously made gvinum crash. - Failed drives no longer disappears entirely, but now leave behind a dummy drive that makes sure the original state is not forgotten in case the system is rebooted between drive failures/swaps. - Update manpage to reflect new commands and extend it with some examples. Sponsored by: Google Summer of Code 2007 Mentored by: le Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
#define GV_DRIVE_REFERENCED 0x01 /* The drive isn't really existing,
but was referenced by a subdisk
during taste. */
struct gv_hdr *hdr; /* The drive header. */
Import the gvinum work that have been done during and after Summer of Code 2007. The work have been under testing and fixing since then, and it is mature enough to be put into HEAD for further testing. A lot have changed in this time, and here are the most important: - Gvinum now uses one single workerthread instead of one thread for each volume and each plex. The reason for this is that the previous scheme was very complex, and was the cause of many of the bugs discovered in gvinum. Instead, gvinum now uses one worker thread with an event queue, quite similar to what used in gmirror. - The rebuild/grow/initialize/parity check routines no longer runs in separate threads, but are run as regular I/O requests with special flags. This made it easier to support mounted growing and parity rebuild. - Support for growing striped and raid5-plexes, meaning that one can extend the volumes for these plex types in addition to the concat type. Also works while the volume is mounted. - Implementation of many of the missing commands from the old vinum: attach/detach, start (was partially implemented), stop (was partially implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes with one plex of these organizations). - The parity check and rebuild no longer goes between userland/kernel, meaning that the gvinum command will not stay and wait forever for the rebuild to finish. You can instead watch the status with the list command. - Many problems with gvinum have been reported since 5.x, and some has been hard to fix due to the complicated architecture. Hopefully, it should be more stable and better handle edge cases that previously made gvinum crash. - Failed drives no longer disappears entirely, but now leave behind a dummy drive that makes sure the original state is not forgotten in case the system is rebooted between drive failures/swaps. - Update manpage to reflect new commands and extend it with some examples. Sponsored by: Google Summer of Code 2007 Mentored by: le Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
struct g_consumer *consumer; /* Consumer attached to this drive. */
int freelist_entries; /* Count of freelist entries. */
LIST_HEAD(,gv_freelist) freelist; /* List of freelist entries. */
LIST_HEAD(,gv_sd) subdisks; /* Subdisks on this drive. */
LIST_ENTRY(gv_drive) drive; /* Entry in the vinum config. */
struct gv_softc *vinumconf; /* Pointer to the vinum conf. */
};
/* softc for a subdisk. */
struct gv_sd {
char name[GV_MAXSDNAME]; /* The name of this subdisk. */
off_t size; /* The size of this subdisk. */
off_t drive_offset; /* Offset in the underlying drive. */
off_t plex_offset; /* Offset in the associated plex. */
int state; /* The state of this subdisk. */
#define GV_SD_DOWN 0
#define GV_SD_STALE 1
#define GV_SD_INITIALIZING 2
#define GV_SD_REVIVING 3
#define GV_SD_UP 4
off_t initialized; /* Count of initialized bytes. */
int init_size; /* Initialization read/write size. */
int init_error; /* Flag error on initialization. */
int flags;
Import the gvinum work that have been done during and after Summer of Code 2007. The work have been under testing and fixing since then, and it is mature enough to be put into HEAD for further testing. A lot have changed in this time, and here are the most important: - Gvinum now uses one single workerthread instead of one thread for each volume and each plex. The reason for this is that the previous scheme was very complex, and was the cause of many of the bugs discovered in gvinum. Instead, gvinum now uses one worker thread with an event queue, quite similar to what used in gmirror. - The rebuild/grow/initialize/parity check routines no longer runs in separate threads, but are run as regular I/O requests with special flags. This made it easier to support mounted growing and parity rebuild. - Support for growing striped and raid5-plexes, meaning that one can extend the volumes for these plex types in addition to the concat type. Also works while the volume is mounted. - Implementation of many of the missing commands from the old vinum: attach/detach, start (was partially implemented), stop (was partially implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes with one plex of these organizations). - The parity check and rebuild no longer goes between userland/kernel, meaning that the gvinum command will not stay and wait forever for the rebuild to finish. You can instead watch the status with the list command. - Many problems with gvinum have been reported since 5.x, and some has been hard to fix due to the complicated architecture. Hopefully, it should be more stable and better handle edge cases that previously made gvinum crash. - Failed drives no longer disappears entirely, but now leave behind a dummy drive that makes sure the original state is not forgotten in case the system is rebooted between drive failures/swaps. - Update manpage to reflect new commands and extend it with some examples. Sponsored by: Google Summer of Code 2007 Mentored by: le Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
#define GV_SD_NEWBORN 0x01 /* Subdisk is created by user. */
#define GV_SD_TASTED 0x02 /* Subdisk is created during taste. */
#define GV_SD_CANGOUP 0x04 /* Subdisk can go up immediately. */
#define GV_SD_GROW 0x08 /* Subdisk is added to striped plex. */
char drive[GV_MAXDRIVENAME]; /* Name of underlying drive. */
char plex[GV_MAXPLEXNAME]; /* Name of associated plex. */
struct gv_drive *drive_sc; /* Pointer to underlying drive. */
struct gv_plex *plex_sc; /* Pointer to associated plex. */
LIST_ENTRY(gv_sd) from_drive; /* Subdisk list of underlying drive. */
LIST_ENTRY(gv_sd) in_plex; /* Subdisk list of associated plex. */
LIST_ENTRY(gv_sd) sd; /* Entry in the vinum config. */
struct gv_softc *vinumconf; /* Pointer to the vinum config. */
};
/* softc for a plex. */
struct gv_plex {
char name[GV_MAXPLEXNAME]; /* The name of the plex. */
off_t size; /* The size of the plex. */
int state; /* The plex state. */
#define GV_PLEX_DOWN 0
#define GV_PLEX_INITIALIZING 1
#define GV_PLEX_DEGRADED 2
Import the gvinum work that have been done during and after Summer of Code 2007. The work have been under testing and fixing since then, and it is mature enough to be put into HEAD for further testing. A lot have changed in this time, and here are the most important: - Gvinum now uses one single workerthread instead of one thread for each volume and each plex. The reason for this is that the previous scheme was very complex, and was the cause of many of the bugs discovered in gvinum. Instead, gvinum now uses one worker thread with an event queue, quite similar to what used in gmirror. - The rebuild/grow/initialize/parity check routines no longer runs in separate threads, but are run as regular I/O requests with special flags. This made it easier to support mounted growing and parity rebuild. - Support for growing striped and raid5-plexes, meaning that one can extend the volumes for these plex types in addition to the concat type. Also works while the volume is mounted. - Implementation of many of the missing commands from the old vinum: attach/detach, start (was partially implemented), stop (was partially implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes with one plex of these organizations). - The parity check and rebuild no longer goes between userland/kernel, meaning that the gvinum command will not stay and wait forever for the rebuild to finish. You can instead watch the status with the list command. - Many problems with gvinum have been reported since 5.x, and some has been hard to fix due to the complicated architecture. Hopefully, it should be more stable and better handle edge cases that previously made gvinum crash. - Failed drives no longer disappears entirely, but now leave behind a dummy drive that makes sure the original state is not forgotten in case the system is rebooted between drive failures/swaps. - Update manpage to reflect new commands and extend it with some examples. Sponsored by: Google Summer of Code 2007 Mentored by: le Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
#define GV_PLEX_GROWABLE 3
#define GV_PLEX_UP 4
int org; /* The plex organisation. */
#define GV_PLEX_DISORG 0
#define GV_PLEX_CONCAT 1
#define GV_PLEX_STRIPED 2
#define GV_PLEX_RAID5 4
int stripesize; /* The stripe size of the plex. */
char volume[GV_MAXVOLNAME]; /* Name of associated volume. */
struct gv_volume *vol_sc; /* Pointer to associated volume. */
Import the gvinum work that have been done during and after Summer of Code 2007. The work have been under testing and fixing since then, and it is mature enough to be put into HEAD for further testing. A lot have changed in this time, and here are the most important: - Gvinum now uses one single workerthread instead of one thread for each volume and each plex. The reason for this is that the previous scheme was very complex, and was the cause of many of the bugs discovered in gvinum. Instead, gvinum now uses one worker thread with an event queue, quite similar to what used in gmirror. - The rebuild/grow/initialize/parity check routines no longer runs in separate threads, but are run as regular I/O requests with special flags. This made it easier to support mounted growing and parity rebuild. - Support for growing striped and raid5-plexes, meaning that one can extend the volumes for these plex types in addition to the concat type. Also works while the volume is mounted. - Implementation of many of the missing commands from the old vinum: attach/detach, start (was partially implemented), stop (was partially implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes with one plex of these organizations). - The parity check and rebuild no longer goes between userland/kernel, meaning that the gvinum command will not stay and wait forever for the rebuild to finish. You can instead watch the status with the list command. - Many problems with gvinum have been reported since 5.x, and some has been hard to fix due to the complicated architecture. Hopefully, it should be more stable and better handle edge cases that previously made gvinum crash. - Failed drives no longer disappears entirely, but now leave behind a dummy drive that makes sure the original state is not forgotten in case the system is rebooted between drive failures/swaps. - Update manpage to reflect new commands and extend it with some examples. Sponsored by: Google Summer of Code 2007 Mentored by: le Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
int sddetached; /* Number of detached subdisks. */
int sdcount; /* Number of subdisks in this plex. */
int sddown; /* Number of subdisks that are down. */
int flags;
#define GV_PLEX_ADDED 0x01 /* Added to an existing volume. */
#define GV_PLEX_SYNCING 0x02 /* Plex is syncing from another plex. */
#define GV_PLEX_NEWBORN 0x20 /* The plex was just created. */
Import the gvinum work that have been done during and after Summer of Code 2007. The work have been under testing and fixing since then, and it is mature enough to be put into HEAD for further testing. A lot have changed in this time, and here are the most important: - Gvinum now uses one single workerthread instead of one thread for each volume and each plex. The reason for this is that the previous scheme was very complex, and was the cause of many of the bugs discovered in gvinum. Instead, gvinum now uses one worker thread with an event queue, quite similar to what used in gmirror. - The rebuild/grow/initialize/parity check routines no longer runs in separate threads, but are run as regular I/O requests with special flags. This made it easier to support mounted growing and parity rebuild. - Support for growing striped and raid5-plexes, meaning that one can extend the volumes for these plex types in addition to the concat type. Also works while the volume is mounted. - Implementation of many of the missing commands from the old vinum: attach/detach, start (was partially implemented), stop (was partially implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes with one plex of these organizations). - The parity check and rebuild no longer goes between userland/kernel, meaning that the gvinum command will not stay and wait forever for the rebuild to finish. You can instead watch the status with the list command. - Many problems with gvinum have been reported since 5.x, and some has been hard to fix due to the complicated architecture. Hopefully, it should be more stable and better handle edge cases that previously made gvinum crash. - Failed drives no longer disappears entirely, but now leave behind a dummy drive that makes sure the original state is not forgotten in case the system is rebooted between drive failures/swaps. - Update manpage to reflect new commands and extend it with some examples. Sponsored by: Google Summer of Code 2007 Mentored by: le Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
#define GV_PLEX_REBUILDING 0x40 /* The plex is rebuilding. */
#define GV_PLEX_GROWING 0x80 /* The plex is growing. */
off_t synced; /* Count of synced bytes. */
TAILQ_HEAD(,gv_raid5_packet) packets; /* RAID5 sub-requests. */
LIST_HEAD(,gv_sd) subdisks; /* List of attached subdisks. */
LIST_ENTRY(gv_plex) in_volume; /* Plex list of associated volume. */
LIST_ENTRY(gv_plex) plex; /* Entry in the vinum config. */
Import the gvinum work that have been done during and after Summer of Code 2007. The work have been under testing and fixing since then, and it is mature enough to be put into HEAD for further testing. A lot have changed in this time, and here are the most important: - Gvinum now uses one single workerthread instead of one thread for each volume and each plex. The reason for this is that the previous scheme was very complex, and was the cause of many of the bugs discovered in gvinum. Instead, gvinum now uses one worker thread with an event queue, quite similar to what used in gmirror. - The rebuild/grow/initialize/parity check routines no longer runs in separate threads, but are run as regular I/O requests with special flags. This made it easier to support mounted growing and parity rebuild. - Support for growing striped and raid5-plexes, meaning that one can extend the volumes for these plex types in addition to the concat type. Also works while the volume is mounted. - Implementation of many of the missing commands from the old vinum: attach/detach, start (was partially implemented), stop (was partially implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes with one plex of these organizations). - The parity check and rebuild no longer goes between userland/kernel, meaning that the gvinum command will not stay and wait forever for the rebuild to finish. You can instead watch the status with the list command. - Many problems with gvinum have been reported since 5.x, and some has been hard to fix due to the complicated architecture. Hopefully, it should be more stable and better handle edge cases that previously made gvinum crash. - Failed drives no longer disappears entirely, but now leave behind a dummy drive that makes sure the original state is not forgotten in case the system is rebooted between drive failures/swaps. - Update manpage to reflect new commands and extend it with some examples. Sponsored by: Google Summer of Code 2007 Mentored by: le Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
#ifdef _KERNEL
struct bio_queue_head *bqueue; /* BIO queue. */
struct bio_queue_head *wqueue; /* Waiting BIO queue. */
struct bio_queue_head *rqueue; /* Rebuild waiting BIO queue. */
#else
char *bpad, *wpad, *rpad; /* Padding for userland. */
#endif
struct gv_softc *vinumconf; /* Pointer to the vinum config. */
};
/* softc for a volume. */
struct gv_volume {
char name[GV_MAXVOLNAME]; /* The name of the volume. */
off_t size; /* The size of the volume. */
int plexcount; /* Number of plexes. */
int state; /* The state of the volume. */
#define GV_VOL_DOWN 0
#define GV_VOL_UP 1
int flags;
Import the gvinum work that have been done during and after Summer of Code 2007. The work have been under testing and fixing since then, and it is mature enough to be put into HEAD for further testing. A lot have changed in this time, and here are the most important: - Gvinum now uses one single workerthread instead of one thread for each volume and each plex. The reason for this is that the previous scheme was very complex, and was the cause of many of the bugs discovered in gvinum. Instead, gvinum now uses one worker thread with an event queue, quite similar to what used in gmirror. - The rebuild/grow/initialize/parity check routines no longer runs in separate threads, but are run as regular I/O requests with special flags. This made it easier to support mounted growing and parity rebuild. - Support for growing striped and raid5-plexes, meaning that one can extend the volumes for these plex types in addition to the concat type. Also works while the volume is mounted. - Implementation of many of the missing commands from the old vinum: attach/detach, start (was partially implemented), stop (was partially implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes with one plex of these organizations). - The parity check and rebuild no longer goes between userland/kernel, meaning that the gvinum command will not stay and wait forever for the rebuild to finish. You can instead watch the status with the list command. - Many problems with gvinum have been reported since 5.x, and some has been hard to fix due to the complicated architecture. Hopefully, it should be more stable and better handle edge cases that previously made gvinum crash. - Failed drives no longer disappears entirely, but now leave behind a dummy drive that makes sure the original state is not forgotten in case the system is rebooted between drive failures/swaps. - Update manpage to reflect new commands and extend it with some examples. Sponsored by: Google Summer of Code 2007 Mentored by: le Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
#define GV_VOL_NEWBORN 0x08 /* The volume was just created. */
Import the gvinum work that have been done during and after Summer of Code 2007. The work have been under testing and fixing since then, and it is mature enough to be put into HEAD for further testing. A lot have changed in this time, and here are the most important: - Gvinum now uses one single workerthread instead of one thread for each volume and each plex. The reason for this is that the previous scheme was very complex, and was the cause of many of the bugs discovered in gvinum. Instead, gvinum now uses one worker thread with an event queue, quite similar to what used in gmirror. - The rebuild/grow/initialize/parity check routines no longer runs in separate threads, but are run as regular I/O requests with special flags. This made it easier to support mounted growing and parity rebuild. - Support for growing striped and raid5-plexes, meaning that one can extend the volumes for these plex types in addition to the concat type. Also works while the volume is mounted. - Implementation of many of the missing commands from the old vinum: attach/detach, start (was partially implemented), stop (was partially implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes with one plex of these organizations). - The parity check and rebuild no longer goes between userland/kernel, meaning that the gvinum command will not stay and wait forever for the rebuild to finish. You can instead watch the status with the list command. - Many problems with gvinum have been reported since 5.x, and some has been hard to fix due to the complicated architecture. Hopefully, it should be more stable and better handle edge cases that previously made gvinum crash. - Failed drives no longer disappears entirely, but now leave behind a dummy drive that makes sure the original state is not forgotten in case the system is rebooted between drive failures/swaps. - Update manpage to reflect new commands and extend it with some examples. Sponsored by: Google Summer of Code 2007 Mentored by: le Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
LIST_HEAD(,gv_plex) plexes; /* List of attached plexes. */
LIST_ENTRY(gv_volume) volume; /* Entry in vinum config. */
struct g_provider *provider; /* Provider of this volume. */
#ifdef _KERNEL
struct bio_queue_head *wqueue; /* BIO delayed request queue. */
#else
Import the gvinum work that have been done during and after Summer of Code 2007. The work have been under testing and fixing since then, and it is mature enough to be put into HEAD for further testing. A lot have changed in this time, and here are the most important: - Gvinum now uses one single workerthread instead of one thread for each volume and each plex. The reason for this is that the previous scheme was very complex, and was the cause of many of the bugs discovered in gvinum. Instead, gvinum now uses one worker thread with an event queue, quite similar to what used in gmirror. - The rebuild/grow/initialize/parity check routines no longer runs in separate threads, but are run as regular I/O requests with special flags. This made it easier to support mounted growing and parity rebuild. - Support for growing striped and raid5-plexes, meaning that one can extend the volumes for these plex types in addition to the concat type. Also works while the volume is mounted. - Implementation of many of the missing commands from the old vinum: attach/detach, start (was partially implemented), stop (was partially implemented), concat, mirror, stripe, raid5 (shortcuts for creating volumes with one plex of these organizations). - The parity check and rebuild no longer goes between userland/kernel, meaning that the gvinum command will not stay and wait forever for the rebuild to finish. You can instead watch the status with the list command. - Many problems with gvinum have been reported since 5.x, and some has been hard to fix due to the complicated architecture. Hopefully, it should be more stable and better handle edge cases that previously made gvinum crash. - Failed drives no longer disappears entirely, but now leave behind a dummy drive that makes sure the original state is not forgotten in case the system is rebooted between drive failures/swaps. - Update manpage to reflect new commands and extend it with some examples. Sponsored by: Google Summer of Code 2007 Mentored by: le Tested by: Rick C. Petty <rick-freebsd2008 -at- kiwi-computer.com>
2009-03-28 17:20:08 +00:00
char *wpad; /* Padding for userland. */
#endif
struct gv_plex *last_read_plex;
struct gv_softc *vinumconf; /* Pointer to the vinum config. */
};
#endif /* !_GEOM_VINUM_VAR_H */