freebsd-dev/sys/geom/vinum/geom_vinum_share.c

729 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 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.
*
*/
/* This file is shared between kernel and userland. */
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <sys/param.h>
#ifdef _KERNEL
#include <sys/malloc.h>
#include <sys/systm.h>
#include <geom/geom.h>
#define iswhite(c) (((c) == ' ') || ((c) == '\t'))
#else
#include <ctype.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#define iswhite isspace
#define g_free free
#endif /* _KERNEL */
#include <sys/mutex.h>
#include <sys/queue.h>
#include <geom/vinum/geom_vinum_var.h>
#include <geom/vinum/geom_vinum_share.h>
/*
* Take a blank separated list of tokens and turn it into a list of
* individual nul-delimited strings. Build a list of pointers at
* token, which must have enough space for the tokens. Return the
* number of tokens, or -1 on error (typically a missing string
* delimiter).
*/
int
gv_tokenize(char *cptr, char *token[], int maxtoken)
{
int tokennr; /* Index of this token. */
char delim; /* Delimiter for searching for the partner. */
for (tokennr = 0; tokennr < maxtoken;) {
/* Skip leading white space. */
while (iswhite(*cptr))
cptr++;
/* End of line. */
if ((*cptr == '\0') || (*cptr == '\n') || (*cptr == '#'))
return tokennr;
delim = *cptr;
token[tokennr] = cptr; /* Point to it. */
tokennr++; /* One more. */
/* Run off the end? */
if (tokennr == maxtoken)
return tokennr;
/* Quoted? */
if ((delim == '\'') || (delim == '"')) {
for (;;) {
cptr++;
/* Found the partner. */
if ((*cptr == delim) && (cptr[-1] != '\\')) {
cptr++;
/* Space after closing quote needed. */
if (!iswhite(*cptr))
return -1;
/* Delimit. */
*cptr++ = '\0';
/* End-of-line? */
} else if ((*cptr == '\0') || (*cptr == '\n'))
return -1;
}
/* Not quoted. */
} else {
while ((*cptr != '\0') &&
(!iswhite(*cptr)) &&
(*cptr != '\n'))
cptr++;
/* Not end-of-line; delimit and move to the next. */
if (*cptr != '\0')
*cptr++ = '\0';
}
}
/* Can't get here. */
return maxtoken;
}
/*
* Take a number with an optional scale factor and convert it to a number of
* bytes.
*
* The scale factors are:
*
* s sectors (of 512 bytes)
* b blocks (of 512 bytes). This unit is deprecated, because it's
* confusing, but maintained to avoid confusing Veritas users.
* k kilobytes (1024 bytes)
* m megabytes (of 1024 * 1024 bytes)
* g gigabytes (of 1024 * 1024 * 1024 bytes)
*
* XXX: need a way to signal error
*/
off_t
gv_sizespec(char *spec)
{
uint64_t size;
char *s;
int sign;
size = 0;
sign = 1;
if (spec != NULL) { /* we have a parameter */
s = spec;
if (*s == '-') { /* negative, */
sign = -1;
s++; /* skip */
}
/* It's numeric. */
if ((*s >= '0') && (*s <= '9')) {
/* It's numeric. */
while ((*s >= '0') && (*s <= '9'))
/* Convert it. */
size = size * 10 + *s++ - '0';
switch (*s) {
case '\0':
return size * sign;
case 'B':
case 'b':
case 'S':
case 's':
return size * sign * 512;
case 'K':
case 'k':
return size * sign * 1024;
case 'M':
case 'm':
return size * sign * 1024 * 1024;
case 'G':
case 'g':
return size * sign * 1024 * 1024 * 1024;
}
}
}
return (0);
}
const char *
gv_drivestate(int state)
{
switch (state) {
case GV_DRIVE_DOWN:
return "down";
case GV_DRIVE_UP:
return "up";
default:
return "??";
}
}
int
gv_drivestatei(char *buf)
{
if (!strcmp(buf, "up"))
return (GV_DRIVE_UP);
else
return (GV_DRIVE_DOWN);
}
/* Translate from a string to a subdisk state. */
int
gv_sdstatei(char *buf)
{
if (!strcmp(buf, "up"))
return (GV_SD_UP);
else if (!strcmp(buf, "reviving"))
return (GV_SD_REVIVING);
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
else if (!strcmp(buf, "initializing"))
return (GV_SD_INITIALIZING);
else if (!strcmp(buf, "stale"))
return (GV_SD_STALE);
else
return (GV_SD_DOWN);
}
/* Translate from a subdisk state to a string. */
const char *
gv_sdstate(int state)
{
switch (state) {
case GV_SD_INITIALIZING:
return "initializing";
case GV_SD_STALE:
return "stale";
case GV_SD_DOWN:
return "down";
case GV_SD_REVIVING:
return "reviving";
case GV_SD_UP:
return "up";
default:
return "??";
}
}
/* Translate from a string to a plex state. */
int
gv_plexstatei(char *buf)
{
if (!strcmp(buf, "up"))
return (GV_PLEX_UP);
else if (!strcmp(buf, "initializing"))
return (GV_PLEX_INITIALIZING);
else if (!strcmp(buf, "degraded"))
return (GV_PLEX_DEGRADED);
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
else if (!strcmp(buf, "growable"))
return (GV_PLEX_GROWABLE);
else
return (GV_PLEX_DOWN);
}
/* Translate from a plex state to a string. */
const char *
gv_plexstate(int state)
{
switch (state) {
case GV_PLEX_DOWN:
return "down";
case GV_PLEX_INITIALIZING:
return "initializing";
case GV_PLEX_DEGRADED:
return "degraded";
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
case GV_PLEX_GROWABLE:
return "growable";
case GV_PLEX_UP:
return "up";
default:
return "??";
}
}
/* Translate from a string to a plex organization. */
int
gv_plexorgi(char *buf)
{
if (!strcmp(buf, "concat"))
return (GV_PLEX_CONCAT);
else if (!strcmp(buf, "striped"))
return (GV_PLEX_STRIPED);
else if (!strcmp(buf, "raid5"))
return (GV_PLEX_RAID5);
else
return (GV_PLEX_DISORG);
}
int
gv_volstatei(char *buf)
{
if (!strcmp(buf, "up"))
return (GV_VOL_UP);
else
return (GV_VOL_DOWN);
}
const char *
gv_volstate(int state)
{
switch (state) {
case GV_VOL_UP:
return "up";
case GV_VOL_DOWN:
return "down";
default:
return "??";
}
}
/* Translate from a plex organization to a string. */
const char *
gv_plexorg(int org)
{
switch (org) {
case GV_PLEX_DISORG:
return "??";
case GV_PLEX_CONCAT:
return "concat";
case GV_PLEX_STRIPED:
return "striped";
case GV_PLEX_RAID5:
return "raid5";
default:
return "??";
}
}
const char *
gv_plexorg_short(int org)
{
switch (org) {
case GV_PLEX_DISORG:
return "??";
case GV_PLEX_CONCAT:
return "C";
case GV_PLEX_STRIPED:
return "S";
case GV_PLEX_RAID5:
return "R5";
default:
return "??";
}
}
struct gv_sd *
gv_alloc_sd(void)
{
struct gv_sd *s;
#ifdef _KERNEL
s = g_malloc(sizeof(struct gv_sd), M_NOWAIT);
#else
s = malloc(sizeof(struct gv_sd));
#endif
if (s == NULL)
return (NULL);
bzero(s, sizeof(struct gv_sd));
s->plex_offset = -1;
s->size = -1;
s->drive_offset = -1;
return (s);
}
struct gv_drive *
gv_alloc_drive(void)
{
struct gv_drive *d;
#ifdef _KERNEL
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
d = g_malloc(sizeof(struct gv_drive), M_NOWAIT);
#else
d = malloc(sizeof(struct gv_drive));
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
#endif
if (d == NULL)
return (NULL);
bzero(d, sizeof(struct gv_drive));
return (d);
}
struct gv_volume *
gv_alloc_volume(void)
{
struct gv_volume *v;
#ifdef _KERNEL
v = g_malloc(sizeof(struct gv_volume), M_NOWAIT);
#else
v = malloc(sizeof(struct gv_volume));
#endif
if (v == NULL)
return (NULL);
bzero(v, sizeof(struct gv_volume));
return (v);
}
struct gv_plex *
gv_alloc_plex(void)
{
struct gv_plex *p;
#ifdef _KERNEL
p = g_malloc(sizeof(struct gv_plex), M_NOWAIT);
#else
p = malloc(sizeof(struct gv_plex));
#endif
if (p == NULL)
return (NULL);
bzero(p, sizeof(struct gv_plex));
return (p);
}
/* Get a new drive object. */
struct gv_drive *
gv_new_drive(int max, char *token[])
{
struct gv_drive *d;
int j, errors;
char *ptr;
if (token[1] == NULL || *token[1] == '\0')
return (NULL);
d = gv_alloc_drive();
if (d == NULL)
return (NULL);
errors = 0;
for (j = 1; j < max; j++) {
if (!strcmp(token[j], "state")) {
j++;
if (j >= max) {
errors++;
break;
}
d->state = gv_drivestatei(token[j]);
} else if (!strcmp(token[j], "device")) {
j++;
if (j >= max) {
errors++;
break;
}
ptr = token[j];
if (strncmp(ptr, "/dev/", 5) == 0)
ptr += 5;
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
strlcpy(d->device, ptr, sizeof(d->device));
} else {
/* We assume this is the drive name. */
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
strlcpy(d->name, token[j], sizeof(d->name));
}
}
if (strlen(d->name) == 0 || strlen(d->device) == 0)
errors++;
if (errors) {
g_free(d);
return (NULL);
}
return (d);
}
/* Get a new volume object. */
struct gv_volume *
gv_new_volume(int max, char *token[])
{
struct gv_volume *v;
int j, errors;
if (token[1] == NULL || *token[1] == '\0')
return (NULL);
v = gv_alloc_volume();
if (v == NULL)
return (NULL);
errors = 0;
for (j = 1; j < max; j++) {
if (!strcmp(token[j], "state")) {
j++;
if (j >= max) {
errors++;
break;
}
v->state = gv_volstatei(token[j]);
} else {
/* We assume this is the volume name. */
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
strlcpy(v->name, token[j], sizeof(v->name));
}
}
if (strlen(v->name) == 0)
errors++;
if (errors) {
g_free(v);
return (NULL);
}
return (v);
}
/* Get a new plex object. */
struct gv_plex *
gv_new_plex(int max, char *token[])
{
struct gv_plex *p;
int j, errors;
if (token[1] == NULL || *token[1] == '\0')
return (NULL);
p = gv_alloc_plex();
if (p == NULL)
return (NULL);
errors = 0;
for (j = 1; j < max; j++) {
if (!strcmp(token[j], "name")) {
j++;
if (j >= max) {
errors++;
break;
}
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
strlcpy(p->name, token[j], sizeof(p->name));
} else if (!strcmp(token[j], "org")) {
j++;
if (j >= max) {
errors++;
break;
}
p->org = gv_plexorgi(token[j]);
if ((p->org == GV_PLEX_RAID5) ||
(p->org == GV_PLEX_STRIPED)) {
j++;
if (j >= max) {
errors++;
break;
}
p->stripesize = gv_sizespec(token[j]);
if (p->stripesize == 0) {
errors++;
break;
}
}
} else if (!strcmp(token[j], "state")) {
j++;
if (j >= max) {
errors++;
break;
}
p->state = gv_plexstatei(token[j]);
} else if (!strcmp(token[j], "vol") ||
!strcmp(token[j], "volume")) {
j++;
if (j >= max) {
errors++;
break;
}
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
strlcpy(p->volume, token[j], sizeof(p->volume));
} else {
errors++;
break;
}
}
if (errors) {
g_free(p);
return (NULL);
}
return (p);
}
/* Get a new subdisk object. */
struct gv_sd *
gv_new_sd(int max, char *token[])
{
struct gv_sd *s;
int j, errors;
if (token[1] == NULL || *token[1] == '\0')
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
return (NULL);
s = gv_alloc_sd();
if (s == NULL)
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
return (NULL);
errors = 0;
for (j = 1; j < max; j++) {
if (!strcmp(token[j], "name")) {
j++;
if (j >= max) {
errors++;
break;
}
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
strlcpy(s->name, token[j], sizeof(s->name));
} else if (!strcmp(token[j], "drive")) {
j++;
if (j >= max) {
errors++;
break;
}
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
strlcpy(s->drive, token[j], sizeof(s->drive));
} else if (!strcmp(token[j], "plex")) {
j++;
if (j >= max) {
errors++;
break;
}
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
strlcpy(s->plex, token[j], sizeof(s->plex));
} else if (!strcmp(token[j], "state")) {
j++;
if (j >= max) {
errors++;
break;
}
s->state = gv_sdstatei(token[j]);
} else if (!strcmp(token[j], "len") ||
!strcmp(token[j], "length")) {
j++;
if (j >= max) {
errors++;
break;
}
s->size = gv_sizespec(token[j]);
if (s->size <= 0)
s->size = -1;
} else if (!strcmp(token[j], "driveoffset")) {
j++;
if (j >= max) {
errors++;
break;
}
s->drive_offset = gv_sizespec(token[j]);
if (s->drive_offset != 0 &&
s->drive_offset < GV_DATA_START) {
errors++;
break;
}
} else if (!strcmp(token[j], "plexoffset")) {
j++;
if (j >= max) {
errors++;
break;
}
s->plex_offset = gv_sizespec(token[j]);
if (s->plex_offset < 0) {
errors++;
break;
}
} else {
errors++;
break;
}
}
if (strlen(s->drive) == 0)
errors++;
if (errors) {
g_free(s);
return (NULL);
}
return (s);
}
/*
* Take a size in bytes and return a pointer to a string which represents the
* size best. If lj is != 0, return left justified, otherwise in a fixed 10
* character field suitable for columnar printing.
*
* Note this uses a static string: it's only intended to be used immediately
* for printing.
*/
const char *
gv_roughlength(off_t bytes, int lj)
{
static char desc[16];
/* Gigabytes. */
if (bytes > (off_t)MEGABYTE * 10000)
snprintf(desc, sizeof(desc), lj ? "%jd GB" : "%10jd GB",
bytes / GIGABYTE);
/* Megabytes. */
else if (bytes > KILOBYTE * 10000)
snprintf(desc, sizeof(desc), lj ? "%jd MB" : "%10jd MB",
bytes / MEGABYTE);
/* Kilobytes. */
else if (bytes > 10000)
snprintf(desc, sizeof(desc), lj ? "%jd kB" : "%10jd kB",
bytes / KILOBYTE);
/* Bytes. */
else
snprintf(desc, sizeof(desc), lj ? "%jd B" : "%10jd B", bytes);
return (desc);
}