* CDDL HEADER END
*/
/*
- * Copyright 2008 Sun Microsystems, Inc. All rights reserved.
- * Use is subject to license terms.
+ * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2016 Actifio, Inc. All rights reserved.
*/
#include <assert.h>
#include <stdlib.h>
#include <string.h>
#include <zlib.h>
+#include <libgen.h>
+#include <sys/signal.h>
#include <sys/spa.h>
#include <sys/stat.h>
#include <sys/processor.h>
#include <sys/zfs_context.h>
-#include <sys/zmod.h>
+#include <sys/rrwlock.h>
#include <sys/utsname.h>
+#include <sys/time.h>
+#include <sys/systeminfo.h>
+#include <zfs_fletcher.h>
+#include <sys/crypto/icp.h>
/*
* Emulation of kernel services in userland.
*/
+int aok;
uint64_t physmem;
vnode_t *rootdir = (vnode_t *)0xabcd1234;
-char hw_serial[11];
+char hw_serial[HW_HOSTID_LEN];
+struct utsname hw_utsname;
+vmem_t *zio_arena = NULL;
-struct utsname utsname = {
- "userland", "libzpool", "1", "1", "na"
-};
+/* If set, all blocks read will be copied to the specified directory. */
+char *vn_dumpdir = NULL;
+
+/* this only exists to have its address taken */
+struct proc p0;
/*
* =========================================================================
* threads
* =========================================================================
*/
-/*ARGSUSED*/
+
+pthread_cond_t kthread_cond = PTHREAD_COND_INITIALIZER;
+pthread_mutex_t kthread_lock = PTHREAD_MUTEX_INITIALIZER;
+pthread_key_t kthread_key;
+int kthread_nr = 0;
+
+void
+thread_init(void)
+{
+ kthread_t *kt;
+
+ VERIFY3S(pthread_key_create(&kthread_key, NULL), ==, 0);
+
+ /* Create entry for primary kthread */
+ kt = umem_zalloc(sizeof (kthread_t), UMEM_NOFAIL);
+ kt->t_tid = pthread_self();
+ kt->t_func = NULL;
+
+ VERIFY3S(pthread_setspecific(kthread_key, kt), ==, 0);
+
+ /* Only the main thread should be running at the moment */
+ ASSERT3S(kthread_nr, ==, 0);
+ kthread_nr = 1;
+}
+
+void
+thread_fini(void)
+{
+ kthread_t *kt = curthread;
+
+ ASSERT(pthread_equal(kt->t_tid, pthread_self()));
+ ASSERT3P(kt->t_func, ==, NULL);
+
+ umem_free(kt, sizeof (kthread_t));
+
+ /* Wait for all threads to exit via thread_exit() */
+ VERIFY3S(pthread_mutex_lock(&kthread_lock), ==, 0);
+
+ kthread_nr--; /* Main thread is exiting */
+
+ while (kthread_nr > 0)
+ VERIFY0(pthread_cond_wait(&kthread_cond, &kthread_lock));
+
+ ASSERT3S(kthread_nr, ==, 0);
+ VERIFY3S(pthread_mutex_unlock(&kthread_lock), ==, 0);
+
+ VERIFY3S(pthread_key_delete(kthread_key), ==, 0);
+}
+
+kthread_t *
+zk_thread_current(void)
+{
+ kthread_t *kt = pthread_getspecific(kthread_key);
+
+ ASSERT3P(kt, !=, NULL);
+
+ return (kt);
+}
+
+void *
+zk_thread_helper(void *arg)
+{
+ kthread_t *kt = (kthread_t *)arg;
+
+ VERIFY3S(pthread_setspecific(kthread_key, kt), ==, 0);
+
+ VERIFY3S(pthread_mutex_lock(&kthread_lock), ==, 0);
+ kthread_nr++;
+ VERIFY3S(pthread_mutex_unlock(&kthread_lock), ==, 0);
+ (void) setpriority(PRIO_PROCESS, 0, kt->t_pri);
+
+ kt->t_tid = pthread_self();
+ ((thread_func_arg_t)kt->t_func)(kt->t_arg);
+
+ /* Unreachable, thread must exit with thread_exit() */
+ abort();
+
+ return (NULL);
+}
+
kthread_t *
-zk_thread_create(void (*func)(), void *arg)
+zk_thread_create(caddr_t stk, size_t stksize, thread_func_t func, void *arg,
+ uint64_t len, proc_t *pp, int state, pri_t pri, int detachstate)
+{
+ kthread_t *kt;
+ pthread_attr_t attr;
+ char *stkstr;
+
+ ASSERT0(state & ~TS_RUN);
+ ASSERT0(len);
+
+ kt = umem_zalloc(sizeof (kthread_t), UMEM_NOFAIL);
+ kt->t_func = func;
+ kt->t_arg = arg;
+ kt->t_pri = pri;
+
+ VERIFY0(pthread_attr_init(&attr));
+ VERIFY0(pthread_attr_setdetachstate(&attr, detachstate));
+
+ /*
+ * We allow the default stack size in user space to be specified by
+ * setting the ZFS_STACK_SIZE environment variable. This allows us
+ * the convenience of observing and debugging stack overruns in
+ * user space. Explicitly specified stack sizes will be honored.
+ * The usage of ZFS_STACK_SIZE is discussed further in the
+ * ENVIRONMENT VARIABLES sections of the ztest(1) man page.
+ */
+ if (stksize == 0) {
+ stkstr = getenv("ZFS_STACK_SIZE");
+
+ if (stkstr == NULL)
+ stksize = TS_STACK_MAX;
+ else
+ stksize = MAX(atoi(stkstr), TS_STACK_MIN);
+ }
+
+ VERIFY3S(stksize, >, 0);
+ stksize = P2ROUNDUP(MAX(stksize, TS_STACK_MIN), PAGESIZE);
+ /*
+ * If this ever fails, it may be because the stack size is not a
+ * multiple of system page size.
+ */
+ VERIFY0(pthread_attr_setstacksize(&attr, stksize));
+ VERIFY0(pthread_attr_setguardsize(&attr, PAGESIZE));
+
+ VERIFY0(pthread_create(&kt->t_tid, &attr, &zk_thread_helper, kt));
+ VERIFY0(pthread_attr_destroy(&attr));
+
+ return (kt);
+}
+
+void
+zk_thread_exit(void)
{
- thread_t tid;
+ kthread_t *kt = curthread;
- VERIFY(thr_create(0, 0, (void *(*)(void *))func, arg, THR_DETACHED,
- &tid) == 0);
+ ASSERT(pthread_equal(kt->t_tid, pthread_self()));
- return ((void *)(uintptr_t)tid);
+ umem_free(kt, sizeof (kthread_t));
+
+ VERIFY0(pthread_mutex_lock(&kthread_lock));
+ kthread_nr--;
+ VERIFY0(pthread_mutex_unlock(&kthread_lock));
+
+ VERIFY0(pthread_cond_broadcast(&kthread_cond));
+ pthread_exit((void *)TS_MAGIC);
+}
+
+void
+zk_thread_join(kt_did_t tid)
+{
+ void *ret;
+
+ pthread_join((pthread_t)tid, &ret);
+ VERIFY3P(ret, ==, (void *)TS_MAGIC);
}
/*
*/
/*ARGSUSED*/
kstat_t *
-kstat_create(char *module, int instance, char *name, char *class,
- uchar_t type, ulong_t ndata, uchar_t ks_flag)
+kstat_create(const char *module, int instance, const char *name,
+ const char *class, uchar_t type, ulong_t ndata, uchar_t ks_flag)
{
return (NULL);
}
kstat_delete(kstat_t *ksp)
{}
+/*ARGSUSED*/
+void
+kstat_waitq_enter(kstat_io_t *kiop)
+{}
+
+/*ARGSUSED*/
+void
+kstat_waitq_exit(kstat_io_t *kiop)
+{}
+
+/*ARGSUSED*/
+void
+kstat_runq_enter(kstat_io_t *kiop)
+{}
+
+/*ARGSUSED*/
+void
+kstat_runq_exit(kstat_io_t *kiop)
+{}
+
+/*ARGSUSED*/
+void
+kstat_waitq_to_runq(kstat_io_t *kiop)
+{}
+
+/*ARGSUSED*/
+void
+kstat_runq_back_to_waitq(kstat_io_t *kiop)
+{}
+
+void
+kstat_set_raw_ops(kstat_t *ksp,
+ int (*headers)(char *buf, size_t size),
+ int (*data)(char *buf, size_t size, void *data),
+ void *(*addr)(kstat_t *ksp, loff_t index))
+{}
+
/*
* =========================================================================
* mutexes
* =========================================================================
*/
+
void
-zmutex_init(kmutex_t *mp)
+mutex_init(kmutex_t *mp, char *name, int type, void *cookie)
{
- mp->m_owner = NULL;
- mp->initialized = B_TRUE;
- (void) _mutex_init(&mp->m_lock, USYNC_THREAD, NULL);
+ ASSERT3S(type, ==, MUTEX_DEFAULT);
+ ASSERT3P(cookie, ==, NULL);
+ mp->m_owner = MTX_INIT;
+ mp->m_magic = MTX_MAGIC;
+ VERIFY3S(pthread_mutex_init(&mp->m_lock, NULL), ==, 0);
}
void
-zmutex_destroy(kmutex_t *mp)
+mutex_destroy(kmutex_t *mp)
{
- ASSERT(mp->initialized == B_TRUE);
- ASSERT(mp->m_owner == NULL);
- (void) _mutex_destroy(&(mp)->m_lock);
- mp->m_owner = (void *)-1UL;
- mp->initialized = B_FALSE;
+ ASSERT3U(mp->m_magic, ==, MTX_MAGIC);
+ ASSERT3P(mp->m_owner, ==, MTX_INIT);
+ ASSERT0(pthread_mutex_destroy(&(mp)->m_lock));
+ mp->m_owner = MTX_DEST;
+ mp->m_magic = 0;
}
void
mutex_enter(kmutex_t *mp)
{
- ASSERT(mp->initialized == B_TRUE);
- ASSERT(mp->m_owner != (void *)-1UL);
- ASSERT(mp->m_owner != curthread);
- VERIFY(mutex_lock(&mp->m_lock) == 0);
- ASSERT(mp->m_owner == NULL);
+ ASSERT3U(mp->m_magic, ==, MTX_MAGIC);
+ ASSERT3P(mp->m_owner, !=, MTX_DEST);
+ ASSERT3P(mp->m_owner, !=, curthread);
+ VERIFY3S(pthread_mutex_lock(&mp->m_lock), ==, 0);
+ ASSERT3P(mp->m_owner, ==, MTX_INIT);
mp->m_owner = curthread;
}
int
mutex_tryenter(kmutex_t *mp)
{
- ASSERT(mp->initialized == B_TRUE);
- ASSERT(mp->m_owner != (void *)-1UL);
- if (0 == mutex_trylock(&mp->m_lock)) {
- ASSERT(mp->m_owner == NULL);
+ int err;
+ ASSERT3U(mp->m_magic, ==, MTX_MAGIC);
+ ASSERT3P(mp->m_owner, !=, MTX_DEST);
+ if (0 == (err = pthread_mutex_trylock(&mp->m_lock))) {
+ ASSERT3P(mp->m_owner, ==, MTX_INIT);
mp->m_owner = curthread;
return (1);
} else {
+ VERIFY3S(err, ==, EBUSY);
return (0);
}
}
void
mutex_exit(kmutex_t *mp)
{
- ASSERT(mp->initialized == B_TRUE);
- ASSERT(mutex_owner(mp) == curthread);
- mp->m_owner = NULL;
- VERIFY(mutex_unlock(&mp->m_lock) == 0);
+ ASSERT3U(mp->m_magic, ==, MTX_MAGIC);
+ ASSERT3P(mutex_owner(mp), ==, curthread);
+ mp->m_owner = MTX_INIT;
+ VERIFY3S(pthread_mutex_unlock(&mp->m_lock), ==, 0);
}
void *
mutex_owner(kmutex_t *mp)
{
- ASSERT(mp->initialized == B_TRUE);
+ ASSERT3U(mp->m_magic, ==, MTX_MAGIC);
return (mp->m_owner);
}
+int
+mutex_held(kmutex_t *mp)
+{
+ return (mp->m_owner == curthread);
+}
+
/*
* =========================================================================
* rwlocks
* =========================================================================
*/
-/*ARGSUSED*/
+
void
rw_init(krwlock_t *rwlp, char *name, int type, void *arg)
{
- rwlock_init(&rwlp->rw_lock, USYNC_THREAD, NULL);
- rwlp->rw_owner = NULL;
- rwlp->initialized = B_TRUE;
+ ASSERT3S(type, ==, RW_DEFAULT);
+ ASSERT3P(arg, ==, NULL);
+ VERIFY3S(pthread_rwlock_init(&rwlp->rw_lock, NULL), ==, 0);
+ rwlp->rw_owner = RW_INIT;
+ rwlp->rw_wr_owner = RW_INIT;
+ rwlp->rw_readers = 0;
+ rwlp->rw_magic = RW_MAGIC;
}
void
rw_destroy(krwlock_t *rwlp)
{
- rwlock_destroy(&rwlp->rw_lock);
- rwlp->rw_owner = (void *)-1UL;
- rwlp->initialized = B_FALSE;
+ ASSERT3U(rwlp->rw_magic, ==, RW_MAGIC);
+ ASSERT(rwlp->rw_readers == 0 && rwlp->rw_wr_owner == RW_INIT);
+ VERIFY3S(pthread_rwlock_destroy(&rwlp->rw_lock), ==, 0);
+ rwlp->rw_magic = 0;
}
void
rw_enter(krwlock_t *rwlp, krw_t rw)
{
- ASSERT(!RW_LOCK_HELD(rwlp));
- ASSERT(rwlp->initialized == B_TRUE);
- ASSERT(rwlp->rw_owner != (void *)-1UL);
- ASSERT(rwlp->rw_owner != curthread);
+ ASSERT3U(rwlp->rw_magic, ==, RW_MAGIC);
+ ASSERT3P(rwlp->rw_owner, !=, curthread);
+ ASSERT3P(rwlp->rw_wr_owner, !=, curthread);
- if (rw == RW_READER)
- VERIFY(rw_rdlock(&rwlp->rw_lock) == 0);
- else
- VERIFY(rw_wrlock(&rwlp->rw_lock) == 0);
+ if (rw == RW_READER) {
+ VERIFY3S(pthread_rwlock_rdlock(&rwlp->rw_lock), ==, 0);
+ ASSERT3P(rwlp->rw_wr_owner, ==, RW_INIT);
+
+ atomic_inc_uint(&rwlp->rw_readers);
+ } else {
+ VERIFY3S(pthread_rwlock_wrlock(&rwlp->rw_lock), ==, 0);
+ ASSERT3P(rwlp->rw_wr_owner, ==, RW_INIT);
+ ASSERT3U(rwlp->rw_readers, ==, 0);
+
+ rwlp->rw_wr_owner = curthread;
+ }
rwlp->rw_owner = curthread;
}
void
rw_exit(krwlock_t *rwlp)
{
- ASSERT(rwlp->initialized == B_TRUE);
- ASSERT(rwlp->rw_owner != (void *)-1UL);
+ ASSERT3U(rwlp->rw_magic, ==, RW_MAGIC);
+ ASSERT(RW_LOCK_HELD(rwlp));
+
+ if (RW_READ_HELD(rwlp))
+ atomic_dec_uint(&rwlp->rw_readers);
+ else
+ rwlp->rw_wr_owner = RW_INIT;
- rwlp->rw_owner = NULL;
- VERIFY(rw_unlock(&rwlp->rw_lock) == 0);
+ rwlp->rw_owner = RW_INIT;
+ VERIFY3S(pthread_rwlock_unlock(&rwlp->rw_lock), ==, 0);
}
int
{
int rv;
- ASSERT(rwlp->initialized == B_TRUE);
- ASSERT(rwlp->rw_owner != (void *)-1UL);
+ ASSERT3U(rwlp->rw_magic, ==, RW_MAGIC);
if (rw == RW_READER)
- rv = rw_tryrdlock(&rwlp->rw_lock);
+ rv = pthread_rwlock_tryrdlock(&rwlp->rw_lock);
else
- rv = rw_trywrlock(&rwlp->rw_lock);
+ rv = pthread_rwlock_trywrlock(&rwlp->rw_lock);
if (rv == 0) {
+ ASSERT3P(rwlp->rw_wr_owner, ==, RW_INIT);
+
+ if (rw == RW_READER)
+ atomic_inc_uint(&rwlp->rw_readers);
+ else {
+ ASSERT3U(rwlp->rw_readers, ==, 0);
+ rwlp->rw_wr_owner = curthread;
+ }
+
rwlp->rw_owner = curthread;
return (1);
}
+ VERIFY3S(rv, ==, EBUSY);
+
return (0);
}
-/*ARGSUSED*/
int
rw_tryupgrade(krwlock_t *rwlp)
{
- ASSERT(rwlp->initialized == B_TRUE);
- ASSERT(rwlp->rw_owner != (void *)-1UL);
+ ASSERT3U(rwlp->rw_magic, ==, RW_MAGIC);
return (0);
}
* condition variables
* =========================================================================
*/
-/*ARGSUSED*/
+
void
cv_init(kcondvar_t *cv, char *name, int type, void *arg)
{
- VERIFY(cond_init(cv, type, NULL) == 0);
+ ASSERT3S(type, ==, CV_DEFAULT);
+ cv->cv_magic = CV_MAGIC;
+ VERIFY0(pthread_cond_init(&cv->cv, NULL));
}
void
cv_destroy(kcondvar_t *cv)
{
- VERIFY(cond_destroy(cv) == 0);
+ ASSERT3U(cv->cv_magic, ==, CV_MAGIC);
+ VERIFY0(pthread_cond_destroy(&cv->cv));
+ cv->cv_magic = 0;
}
void
cv_wait(kcondvar_t *cv, kmutex_t *mp)
{
- ASSERT(mutex_owner(mp) == curthread);
- mp->m_owner = NULL;
- int ret = cond_wait(cv, &mp->m_lock);
- VERIFY(ret == 0 || ret == EINTR);
+ ASSERT3U(cv->cv_magic, ==, CV_MAGIC);
+ ASSERT3P(mutex_owner(mp), ==, curthread);
+ mp->m_owner = MTX_INIT;
+ VERIFY0(pthread_cond_wait(&cv->cv, &mp->m_lock));
mp->m_owner = curthread;
}
cv_timedwait(kcondvar_t *cv, kmutex_t *mp, clock_t abstime)
{
int error;
- timestruc_t ts;
+ struct timeval tv;
+ struct timespec ts;
clock_t delta;
-top:
- delta = abstime - lbolt;
+ ASSERT3U(cv->cv_magic, ==, CV_MAGIC);
+
+ delta = abstime - ddi_get_lbolt();
if (delta <= 0)
return (-1);
- ts.tv_sec = delta / hz;
- ts.tv_nsec = (delta % hz) * (NANOSEC / hz);
+ VERIFY(gettimeofday(&tv, NULL) == 0);
- ASSERT(mutex_owner(mp) == curthread);
- mp->m_owner = NULL;
- error = cond_reltimedwait(cv, &mp->m_lock, &ts);
+ ts.tv_sec = tv.tv_sec + delta / hz;
+ ts.tv_nsec = tv.tv_usec * NSEC_PER_USEC + (delta % hz) * (NANOSEC / hz);
+ if (ts.tv_nsec >= NANOSEC) {
+ ts.tv_sec++;
+ ts.tv_nsec -= NANOSEC;
+ }
+
+ ASSERT3P(mutex_owner(mp), ==, curthread);
+ mp->m_owner = MTX_INIT;
+ error = pthread_cond_timedwait(&cv->cv, &mp->m_lock, &ts);
mp->m_owner = curthread;
- if (error == ETIME)
+ if (error == ETIMEDOUT)
+ return (-1);
+
+ VERIFY0(error);
+
+ return (1);
+}
+
+/*ARGSUSED*/
+clock_t
+cv_timedwait_hires(kcondvar_t *cv, kmutex_t *mp, hrtime_t tim, hrtime_t res,
+ int flag)
+{
+ int error;
+ struct timeval tv;
+ struct timespec ts;
+ hrtime_t delta;
+
+ ASSERT(flag == 0 || flag == CALLOUT_FLAG_ABSOLUTE);
+
+ delta = tim;
+ if (flag & CALLOUT_FLAG_ABSOLUTE)
+ delta -= gethrtime();
+
+ if (delta <= 0)
return (-1);
- if (error == EINTR)
- goto top;
+ VERIFY(gettimeofday(&tv, NULL) == 0);
+
+ ts.tv_sec = tv.tv_sec + delta / NANOSEC;
+ ts.tv_nsec = tv.tv_usec * NSEC_PER_USEC + (delta % NANOSEC);
+ if (ts.tv_nsec >= NANOSEC) {
+ ts.tv_sec++;
+ ts.tv_nsec -= NANOSEC;
+ }
- ASSERT(error == 0);
+ ASSERT(mutex_owner(mp) == curthread);
+ mp->m_owner = MTX_INIT;
+ error = pthread_cond_timedwait(&cv->cv, &mp->m_lock, &ts);
+ mp->m_owner = curthread;
+
+ if (error == ETIMEDOUT)
+ return (-1);
+
+ VERIFY0(error);
return (1);
}
void
cv_signal(kcondvar_t *cv)
{
- VERIFY(cond_signal(cv) == 0);
+ ASSERT3U(cv->cv_magic, ==, CV_MAGIC);
+ VERIFY0(pthread_cond_signal(&cv->cv));
}
void
cv_broadcast(kcondvar_t *cv)
{
- VERIFY(cond_broadcast(cv) == 0);
+ ASSERT3U(cv->cv_magic, ==, CV_MAGIC);
+ VERIFY0(pthread_cond_broadcast(&cv->cv));
}
/*
int
vn_open(char *path, int x1, int flags, int mode, vnode_t **vpp, int x2, int x3)
{
- int fd;
+ int fd = -1;
+ int dump_fd = -1;
vnode_t *vp;
- int old_umask;
- char realpath[MAXPATHLEN];
+ int old_umask = 0;
+ char *realpath;
struct stat64 st;
+ int err;
+
+ realpath = umem_alloc(MAXPATHLEN, UMEM_NOFAIL);
/*
* If we're accessing a real disk from userland, we need to use
* for its size. So -- gag -- we open the block device to get
* its size, and remember it for subsequent VOP_GETATTR().
*/
+#if defined(__sun__) || defined(__sun)
if (strncmp(path, "/dev/", 5) == 0) {
+#else
+ if (0) {
+#endif
char *dsk;
fd = open64(path, O_RDONLY);
- if (fd == -1)
- return (errno);
+ if (fd == -1) {
+ err = errno;
+ free(realpath);
+ return (err);
+ }
if (fstat64(fd, &st) == -1) {
+ err = errno;
close(fd);
- return (errno);
+ free(realpath);
+ return (err);
}
close(fd);
(void) sprintf(realpath, "%s", path);
dsk + 1);
} else {
(void) sprintf(realpath, "%s", path);
- if (!(flags & FCREAT) && stat64(realpath, &st) == -1)
- return (errno);
+ if (!(flags & FCREAT) && stat64(realpath, &st) == -1) {
+ err = errno;
+ free(realpath);
+ return (err);
+ }
+ }
+
+ if (!(flags & FCREAT) && S_ISBLK(st.st_mode)) {
+#ifdef __linux__
+ flags |= O_DIRECT;
+#endif
+ /* We shouldn't be writing to block devices in userspace */
+ VERIFY(!(flags & FWRITE));
}
if (flags & FCREAT)
* FREAD and FWRITE to the corresponding O_RDONLY, O_WRONLY, and O_RDWR.
*/
fd = open64(realpath, flags - FREAD, mode);
+ if (fd == -1) {
+ err = errno;
+ free(realpath);
+ return (err);
+ }
if (flags & FCREAT)
(void) umask(old_umask);
- if (fd == -1)
- return (errno);
+ if (vn_dumpdir != NULL) {
+ char *dumppath = umem_zalloc(MAXPATHLEN, UMEM_NOFAIL);
+ (void) snprintf(dumppath, MAXPATHLEN,
+ "%s/%s", vn_dumpdir, basename(realpath));
+ dump_fd = open64(dumppath, O_CREAT | O_WRONLY, 0666);
+ umem_free(dumppath, MAXPATHLEN);
+ if (dump_fd == -1) {
+ err = errno;
+ free(realpath);
+ close(fd);
+ return (err);
+ }
+ } else {
+ dump_fd = -1;
+ }
+
+ free(realpath);
- if (fstat64(fd, &st) == -1) {
+ if (fstat64_blk(fd, &st) == -1) {
+ err = errno;
close(fd);
- return (errno);
+ if (dump_fd != -1)
+ close(dump_fd);
+ return (err);
}
(void) fcntl(fd, F_SETFD, FD_CLOEXEC);
vp->v_fd = fd;
vp->v_size = st.st_size;
vp->v_path = spa_strdup(path);
+ vp->v_dump_fd = dump_fd;
return (0);
}
/*ARGSUSED*/
int
vn_rdwr(int uio, vnode_t *vp, void *addr, ssize_t len, offset_t offset,
- int x1, int x2, rlim64_t x3, void *x4, ssize_t *residp)
+ int x1, int x2, rlim64_t x3, void *x4, ssize_t *residp)
{
- ssize_t iolen, split;
+ ssize_t rc, done = 0, split;
if (uio == UIO_READ) {
- iolen = pread64(vp->v_fd, addr, len, offset);
+ rc = pread64(vp->v_fd, addr, len, offset);
+ if (vp->v_dump_fd != -1 && rc != -1) {
+ int status;
+ status = pwrite64(vp->v_dump_fd, addr, rc, offset);
+ ASSERT(status != -1);
+ }
} else {
/*
* To simulate partial disk writes, we split writes into two
* system calls so that the process can be killed in between.
*/
- split = (len > 0 ? rand() % len : 0);
- iolen = pwrite64(vp->v_fd, addr, split, offset);
- iolen += pwrite64(vp->v_fd, (char *)addr + split,
- len - split, offset + split);
+ int sectors = len >> SPA_MINBLOCKSHIFT;
+ split = (sectors > 0 ? rand() % sectors : 0) <<
+ SPA_MINBLOCKSHIFT;
+ rc = pwrite64(vp->v_fd, addr, split, offset);
+ if (rc != -1) {
+ done = rc;
+ rc = pwrite64(vp->v_fd, (char *)addr + split,
+ len - split, offset + split);
+ }
}
- if (iolen == -1)
+#ifdef __linux__
+ if (rc == -1 && errno == EINVAL) {
+ /*
+ * Under Linux, this most likely means an alignment issue
+ * (memory or disk) due to O_DIRECT, so we abort() in order to
+ * catch the offender.
+ */
+ abort();
+ }
+#endif
+ if (rc == -1)
return (errno);
+
+ done += rc;
+
if (residp)
- *residp = len - iolen;
- else if (iolen != len)
+ *residp = len - done;
+ else if (done != len)
return (EIO);
return (0);
}
vn_close(vnode_t *vp)
{
close(vp->v_fd);
+ if (vp->v_dump_fd != -1)
+ close(vp->v_dump_fd);
spa_strfree(vp->v_path);
umem_free(vp, sizeof (vnode_t));
}
-#ifdef ZFS_DEBUG
+/*
+ * At a minimum we need to update the size since vdev_reopen()
+ * will no longer call vn_openat().
+ */
+int
+fop_getattr(vnode_t *vp, vattr_t *vap)
+{
+ struct stat64 st;
+ int err;
+
+ if (fstat64_blk(vp->v_fd, &st) == -1) {
+ err = errno;
+ close(vp->v_fd);
+ return (err);
+ }
+
+ vap->va_size = st.st_size;
+ return (0);
+}
/*
* =========================================================================
*/
if (dprintf_find_string("on"))
dprintf_print_all = 1;
+
+ if (dprintf_string != NULL)
+ zfs_flags |= ZFS_DEBUG_DPRINTF;
}
/*
if (dprintf_find_string("pid"))
(void) printf("%d ", getpid());
if (dprintf_find_string("tid"))
- (void) printf("%u ", thr_self());
+ (void) printf("%u ", (uint_t)pthread_self());
if (dprintf_find_string("cpu"))
(void) printf("%u ", getcpuid());
if (dprintf_find_string("time"))
}
}
-#endif /* ZFS_DEBUG */
-
/*
* =========================================================================
* cmn_err() and panic()
int
kobj_read_file(struct _buf *file, char *buf, unsigned size, unsigned off)
{
- ssize_t resid;
+ ssize_t resid = 0;
- vn_rdwr(UIO_READ, (vnode_t *)file->_fd, buf, size, (offset_t)off,
- UIO_SYSSPACE, 0, 0, 0, &resid);
+ if (vn_rdwr(UIO_READ, (vnode_t *)file->_fd, buf, size, (offset_t)off,
+ UIO_SYSSPACE, 0, 0, 0, &resid) != 0)
+ return (-1);
return (size - resid);
}
void
delay(clock_t ticks)
{
- poll(0, 0, ticks * (1000 / hz));
+ (void) poll(0, 0, ticks * (1000 / hz));
}
/*
* High order bit is 31 (or 63 in _LP64 kernel).
*/
int
-highbit(ulong_t i)
+highbit64(uint64_t i)
{
register int h = 1;
+ if (i == 0)
+ return (0);
+ if (i & 0xffffffff00000000ULL) {
+ h += 32; i >>= 32;
+ }
+ if (i & 0xffff0000) {
+ h += 16; i >>= 16;
+ }
+ if (i & 0xff00) {
+ h += 8; i >>= 8;
+ }
+ if (i & 0xf0) {
+ h += 4; i >>= 4;
+ }
+ if (i & 0xc) {
+ h += 2; i >>= 2;
+ }
+ if (i & 0x2) {
+ h += 1;
+ }
+ return (h);
+}
+
+/*
+ * Find lowest one bit set.
+ * Returns bit number + 1 of lowest bit that is set, otherwise returns 0.
+ * This is basically a reimplementation of ffsll(), which is GNU specific.
+ */
+int
+lowbit64(uint64_t i)
+{
+ register int h = 64;
+ if (i == 0)
+ return (0);
+
+ if (i & 0x00000000ffffffffULL)
+ h -= 32;
+ else
+ i >>= 32;
+
+ if (i & 0x0000ffff)
+ h -= 16;
+ else
+ i >>= 16;
+
+ if (i & 0x00ff)
+ h -= 8;
+ else
+ i >>= 8;
+
+ if (i & 0x0f)
+ h -= 4;
+ else
+ i >>= 4;
+
+ if (i & 0x3)
+ h -= 2;
+ else
+ i >>= 2;
+
+ if (i & 0x1)
+ h -= 1;
+
+ return (h);
+}
+
+/*
+ * Find highest one bit set.
+ * Returns bit number + 1 of highest bit that is set, otherwise returns 0.
+ * High order bit is 31 (or 63 in _LP64 kernel).
+ */
+int
+highbit(ulong_t i)
+{
+register int h = 1;
+
if (i == 0)
return (0);
#ifdef _LP64
return (h);
}
+/*
+ * Find lowest one bit set.
+ * Returns bit number + 1 of lowest bit that is set, otherwise returns 0.
+ * Low order bit is 0.
+ */
+int
+lowbit(ulong_t i)
+{
+ register int h = 1;
+
+ if (i == 0)
+ return (0);
+
+#ifdef _LP64
+ if (!(i & 0xffffffff)) {
+ h += 32; i >>= 32;
+ }
+#endif
+ if (!(i & 0xffff)) {
+ h += 16; i >>= 16;
+ }
+ if (!(i & 0xff)) {
+ h += 8; i >>= 8;
+ }
+ if (!(i & 0xf)) {
+ h += 4; i >>= 4;
+ }
+ if (!(i & 0x3)) {
+ h += 2; i >>= 2;
+ }
+ if (!(i & 0x1)) {
+ h += 1;
+ }
+ return (h);
+}
+
static int random_fd = -1, urandom_fd = -1;
+void
+random_init(void)
+{
+ VERIFY((random_fd = open("/dev/random", O_RDONLY)) != -1);
+ VERIFY((urandom_fd = open("/dev/urandom", O_RDONLY)) != -1);
+}
+
+void
+random_fini(void)
+{
+ close(random_fd);
+ close(urandom_fd);
+
+ random_fd = -1;
+ urandom_fd = -1;
+}
+
static int
random_get_bytes_common(uint8_t *ptr, size_t len, int fd)
{
return (0);
}
+int
+ddi_strtoull(const char *str, char **nptr, int base, u_longlong_t *result)
+{
+ char *end;
+
+ *result = strtoull(str, &end, base);
+ if (*result == 0)
+ return (errno);
+ return (0);
+}
+
+utsname_t *
+utsname(void)
+{
+ return (&hw_utsname);
+}
+
/*
* =========================================================================
* kernel emulation setup & teardown
{
char errmsg[] = "out of memory -- generating core dump\n";
- write(fileno(stderr), errmsg, sizeof (errmsg));
+ (void) fprintf(stderr, "%s", errmsg);
abort();
return (0);
}
+#define HOSTID_MASK 0xffffffff
+
+static unsigned long
+get_spl_hostid(void)
+{
+ FILE *f;
+ unsigned long hostid;
+ char *env;
+
+ /*
+ * Allow the hostid to be subverted for testing.
+ */
+ env = getenv("ZFS_HOSTID");
+ if (env) {
+ hostid = strtoull(env, NULL, 0);
+ return (hostid & HOSTID_MASK);
+ }
+
+ f = fopen("/sys/module/spl/parameters/spl_hostid", "r");
+ if (!f)
+ return (0);
+
+ if (fscanf(f, "%lu", &hostid) != 1)
+ hostid = 0;
+
+ fclose(f);
+
+ return (hostid & HOSTID_MASK);
+}
+
+unsigned long
+get_system_hostid(void)
+{
+ unsigned long system_hostid = get_spl_hostid();
+ /*
+ * We do not use the library call gethostid() because
+ * it generates a hostid value that the kernel is
+ * unaware of, if the spl_hostid module parameter has not
+ * been set and there is no system hostid file (e.g.
+ * /etc/hostid). The kernel and userspace must agree.
+ * See comments above hostid_read() in the SPL.
+ */
+ if (system_hostid == 0) {
+ int fd, rc;
+ unsigned long hostid;
+ int hostid_size = 4; /* 4 bytes regardless of arch */
+
+ fd = open("/etc/hostid", O_RDONLY);
+ if (fd >= 0) {
+ rc = read(fd, &hostid, hostid_size);
+ if (rc > 0)
+ system_hostid = (hostid & HOSTID_MASK);
+ close(fd);
+ }
+ }
+ return (system_hostid);
+}
+
void
kernel_init(int mode)
{
+ extern uint_t rrw_tsd_key;
+
umem_nofail_callback(umem_out_of_memory);
physmem = sysconf(_SC_PHYS_PAGES);
dprintf("physmem = %llu pages (%.2f GB)\n", physmem,
(double)physmem * sysconf(_SC_PAGE_SIZE) / (1ULL << 30));
- snprintf(hw_serial, sizeof (hw_serial), "%ld", gethostid());
+ (void) snprintf(hw_serial, sizeof (hw_serial), "%ld",
+ (mode & FWRITE) ? get_system_hostid() : 0);
- VERIFY((random_fd = open("/dev/random", O_RDONLY)) != -1);
- VERIFY((urandom_fd = open("/dev/urandom", O_RDONLY)) != -1);
+ random_init();
+
+ VERIFY0(uname(&hw_utsname));
+ thread_init();
system_taskq_init();
+ icp_init();
spa_init(mode);
+
+ fletcher_4_init();
+
+ tsd_create(&rrw_tsd_key, rrw_tsd_destroy);
}
void
kernel_fini(void)
{
+ fletcher_4_fini();
spa_fini();
- close(random_fd);
- close(urandom_fd);
+ icp_fini();
+ system_taskq_fini();
+ thread_fini();
- random_fd = -1;
- urandom_fd = -1;
+ random_fini();
}
-int
-z_uncompress(void *dst, size_t *dstlen, const void *src, size_t srclen)
-{
- int ret;
- uLongf len = *dstlen;
-
- if ((ret = uncompress(dst, &len, src, srclen)) == Z_OK)
- *dstlen = (size_t)len;
-
- return (ret);
-}
-
-int
-z_compress_level(void *dst, size_t *dstlen, const void *src, size_t srclen,
- int level)
+uid_t
+crgetuid(cred_t *cr)
{
- int ret;
- uLongf len = *dstlen;
-
- if ((ret = compress2(dst, &len, src, srclen, level)) == Z_OK)
- *dstlen = (size_t)len;
-
- return (ret);
+ return (0);
}
uid_t
-crgetuid(cred_t *cr)
+crgetruid(cred_t *cr)
{
return (0);
}
return (0);
}
+int
+secpolicy_zfs(const cred_t *cr)
+{
+ return (0);
+}
+
ksiddomain_t *
ksid_lookupdomain(const char *dom)
{
spa_strfree(ksid->kd_name);
umem_free(ksid, sizeof (ksiddomain_t));
}
+
+char *
+kmem_vasprintf(const char *fmt, va_list adx)
+{
+ char *buf = NULL;
+ va_list adx_copy;
+
+ va_copy(adx_copy, adx);
+ VERIFY(vasprintf(&buf, fmt, adx_copy) != -1);
+ va_end(adx_copy);
+
+ return (buf);
+}
+
+char *
+kmem_asprintf(const char *fmt, ...)
+{
+ char *buf = NULL;
+ va_list adx;
+
+ va_start(adx, fmt);
+ VERIFY(vasprintf(&buf, fmt, adx) != -1);
+ va_end(adx);
+
+ return (buf);
+}
+
+/* ARGSUSED */
+int
+zfs_onexit_fd_hold(int fd, minor_t *minorp)
+{
+ *minorp = 0;
+ return (0);
+}
+
+/* ARGSUSED */
+void
+zfs_onexit_fd_rele(int fd)
+{
+}
+
+/* ARGSUSED */
+int
+zfs_onexit_add_cb(minor_t minor, void (*func)(void *), void *data,
+ uint64_t *action_handle)
+{
+ return (0);
+}
+
+/* ARGSUSED */
+int
+zfs_onexit_del_cb(minor_t minor, uint64_t action_handle, boolean_t fire)
+{
+ return (0);
+}
+
+/* ARGSUSED */
+int
+zfs_onexit_cb_data(minor_t minor, uint64_t action_handle, void **data)
+{
+ return (0);
+}
+
+fstrans_cookie_t
+spl_fstrans_mark(void)
+{
+ return ((fstrans_cookie_t)0);
+}
+
+void
+spl_fstrans_unmark(fstrans_cookie_t cookie)
+{
+}
+
+int
+__spl_pf_fstrans_check(void)
+{
+ return (0);
+}
+
+void *zvol_tag = "zvol_tag";
+
+void
+zvol_create_minors(spa_t *spa, const char *name, boolean_t async)
+{
+}
+
+void
+zvol_remove_minor(spa_t *spa, const char *name, boolean_t async)
+{
+}
+
+void
+zvol_remove_minors(spa_t *spa, const char *name, boolean_t async)
+{
+}
+
+void
+zvol_rename_minors(spa_t *spa, const char *oldname, const char *newname,
+ boolean_t async)
+{
+}
projects / mirror_zfs-debian.git / blobdiff