diff options
Diffstat (limited to 'libglusterfs/src/glusterfs/syncop.h')
| -rw-r--r-- | libglusterfs/src/glusterfs/syncop.h | 637 |
1 files changed, 637 insertions, 0 deletions
diff --git a/libglusterfs/src/glusterfs/syncop.h b/libglusterfs/src/glusterfs/syncop.h new file mode 100644 index 00000000000..203abe92b57 --- /dev/null +++ b/libglusterfs/src/glusterfs/syncop.h @@ -0,0 +1,637 @@ +/* + Copyright (c) 2008-2012 Red Hat, Inc. <http://www.redhat.com> + This file is part of GlusterFS. + + This file is licensed to you under your choice of the GNU Lesser + General Public License, version 3 or any later version (LGPLv3 or + later), or the GNU General Public License, version 2 (GPLv2), in all + cases as published by the Free Software Foundation. +*/ + +#ifndef _SYNCOP_H +#define _SYNCOP_H + +#include "glusterfs/xlator.h" +#include <sys/time.h> +#include <pthread.h> +#include <ucontext.h> + +#define SYNCENV_PROC_MAX 16 +#define SYNCENV_PROC_MIN 2 +#define SYNCPROC_IDLE_TIME 600 + +/* + * Flags for syncopctx valid elements + */ +#define SYNCOPCTX_UID 0x00000001 +#define SYNCOPCTX_GID 0x00000002 +#define SYNCOPCTX_GROUPS 0x00000004 +#define SYNCOPCTX_PID 0x00000008 +#define SYNCOPCTX_LKOWNER 0x00000010 + +struct synctask; +struct syncproc; +struct syncenv; + +typedef int (*synctask_cbk_t)(int ret, call_frame_t *frame, void *opaque); + +typedef int (*synctask_fn_t)(void *opaque); + +typedef enum { + SYNCTASK_INIT = 0, + SYNCTASK_RUN, + SYNCTASK_SUSPEND, + SYNCTASK_WAIT, + SYNCTASK_DONE, + SYNCTASK_ZOMBIE, +} synctask_state_t; + +/* for one sequential execution of @syncfn */ +struct synctask { + struct list_head all_tasks; + struct syncenv *env; + xlator_t *xl; + call_frame_t *frame; + call_frame_t *opframe; + synctask_cbk_t synccbk; + synctask_fn_t syncfn; + synctask_state_t state; + void *opaque; + void *stack; + int woken; + int slept; + int ret; + + uid_t uid; + gid_t gid; + + ucontext_t ctx; + struct syncproc *proc; + + pthread_mutex_t mutex; /* for synchronous spawning of synctask */ + pthread_cond_t cond; + int done; + + struct list_head waitq; /* can wait only "once" at a time */ + char btbuf[GF_BACKTRACE_LEN]; +}; + +struct syncproc { + pthread_t processor; + ucontext_t sched; + struct syncenv *env; + struct synctask *current; +}; + +/* hosts the scheduler thread and framework for executing synctasks */ +struct syncenv { + struct syncproc proc[SYNCENV_PROC_MAX]; + int procs; + + struct list_head runq; + int runcount; + struct list_head waitq; + int waitcount; + + int procmin; + int procmax; + + pthread_mutex_t mutex; + pthread_cond_t cond; + + size_t stacksize; + + int destroy; /* FLAG to mark syncenv is in destroy mode + so that no more synctasks are accepted*/ +}; + +typedef enum { LOCK_NULL = 0, LOCK_TASK, LOCK_THREAD } lock_type_t; + +typedef enum { + SYNC_LOCK_DEFAULT = 0, + SYNC_LOCK_RECURSIVE, /*it allows recursive locking*/ +} lock_attr_t; + +struct synclock { + pthread_mutex_t guard; /* guard the remaining members, pair @cond */ + pthread_cond_t cond; /* waiting non-synctasks */ + struct list_head waitq; /* waiting synctasks */ + volatile int lock; /* true(non zero) or false(zero), lock status */ + lock_attr_t attr; + struct synctask *owner; /* NULL if current owner is not a synctask */ + pthread_t owner_tid; + lock_type_t type; +}; +typedef struct synclock synclock_t; + +struct syncbarrier { + gf_boolean_t initialized; /*Set on successful initialization*/ + pthread_mutex_t guard; /* guard the remaining members, pair @cond */ + pthread_cond_t cond; /* waiting non-synctasks */ + struct list_head waitq; /* waiting synctasks */ + int count; /* count the number of wakes */ + int waitfor; /* no. of wakes until which task can be in + waitq before being woken up. */ +}; +typedef struct syncbarrier syncbarrier_t; + +struct syncargs { + int op_ret; + int op_errno; + struct iatt iatt1; + struct iatt iatt2; + dict_t *xattr; + struct statvfs statvfs_buf; + struct iovec *vector; + int count; + struct iobref *iobref; + char *buffer; + dict_t *xdata; + struct gf_flock flock; + struct gf_lease lease; + dict_t *dict_out; + + /* some more _cbk needs */ + uuid_t uuid; + char *errstr; + dict_t *dict; + pthread_mutex_t lock_dict; + + syncbarrier_t barrier; + + /* do not touch */ + struct synctask *task; + pthread_mutex_t mutex; + pthread_cond_t cond; + int done; + + gf_dirent_t entries; + off_t offset; + + lock_migration_info_t locklist; +}; + +struct syncopctx { + unsigned int valid; /* valid flags for elements that are set */ + uid_t uid; + gid_t gid; + int grpsize; + int ngrps; + gid_t *groups; + pid_t pid; + gf_lkowner_t lk_owner; +}; + +#define __yawn(args) \ + do { \ + args->task = synctask_get(); \ + if (args->task) \ + break; \ + pthread_mutex_init(&args->mutex, NULL); \ + pthread_cond_init(&args->cond, NULL); \ + args->done = 0; \ + } while (0) + +#define __wake(args) \ + do { \ + if (args->task) { \ + synctask_wake(args->task); \ + } else { \ + pthread_mutex_lock(&args->mutex); \ + { \ + args->done = 1; \ + pthread_cond_signal(&args->cond); \ + } \ + pthread_mutex_unlock(&args->mutex); \ + } \ + } while (0) + +#define __yield(args) \ + do { \ + if (args->task) { \ + synctask_yield(args->task); \ + } else { \ + pthread_mutex_lock(&args->mutex); \ + { \ + while (!args->done) \ + pthread_cond_wait(&args->cond, &args->mutex); \ + } \ + pthread_mutex_unlock(&args->mutex); \ + pthread_mutex_destroy(&args->mutex); \ + pthread_cond_destroy(&args->cond); \ + } \ + } while (0) + +#define SYNCOP(subvol, stb, cbk, fn_op, params...) \ + do { \ + struct synctask *task = NULL; \ + call_frame_t *frame = NULL; \ + \ + task = synctask_get(); \ + stb->task = task; \ + if (task) \ + frame = task->opframe; \ + else \ + frame = syncop_create_frame(THIS); \ + \ + if (task) { \ + frame->root->uid = task->uid; \ + frame->root->gid = task->gid; \ + } \ + \ + __yawn(stb); \ + \ + frame->op = get_fop_index_from_fn(subvol, fn_op); \ + STACK_WIND_COOKIE(frame, cbk, (void *)stb, subvol, fn_op, params); \ + \ + __yield(stb); \ + if (task) \ + STACK_RESET(frame->root); \ + else \ + STACK_DESTROY(frame->root); \ + } while (0) + +/* + * syncop_xxx() calls are executed in two ways, one is inside a synctask where + * the executing function will do 'swapcontext' and the other is without + * synctask where the executing thread is made to wait using pthread_cond_wait. + * Executing thread may change when syncop_xxx() is executed inside a synctask. + * This leads to errno_location change i.e. errno may give errno of + * non-executing thread. So errno is not touched inside a synctask execution. + * All gfapi calls are executed using the second way of executing syncop_xxx() + * where the executing thread waits using pthread_cond_wait so it is ok to set + * errno in these cases. The following macro makes syncop_xxx() behave just + * like a system call, where -1 is returned and errno is set when a failure + * occurs. + */ +#define DECODE_SYNCOP_ERR(ret) \ + do { \ + if (ret < 0) { \ + errno = -ret; \ + ret = -1; \ + } else { \ + errno = 0; \ + } \ + } while (0) + +#define SYNCENV_DEFAULT_STACKSIZE (2 * 1024 * 1024) + +struct syncenv * +syncenv_new(size_t stacksize, int procmin, int procmax); +void +syncenv_destroy(struct syncenv *); +void +syncenv_scale(struct syncenv *env); + +int +synctask_new1(struct syncenv *, size_t stacksize, synctask_fn_t, synctask_cbk_t, + call_frame_t *frame, void *); +int +synctask_new(struct syncenv *, synctask_fn_t, synctask_cbk_t, + call_frame_t *frame, void *); +struct synctask * +synctask_create(struct syncenv *, size_t stacksize, synctask_fn_t, + synctask_cbk_t, call_frame_t *, void *); +int +synctask_join(struct synctask *task); +void +synctask_wake(struct synctask *task); +void +synctask_yield(struct synctask *task); +void +synctask_waitfor(struct synctask *task, int count); + +#define synctask_barrier_init(args) syncbarrier_init(&args->barrier) +#define synctask_barrier_wait(args, n) syncbarrier_wait(&args->barrier, n) +#define synctask_barrier_wake(args) syncbarrier_wake(&args->barrier) + +int +synctask_setid(struct synctask *task, uid_t uid, gid_t gid); +#define SYNCTASK_SETID(uid, gid) synctask_setid(synctask_get(), uid, gid); + +int +syncopctx_setfsuid(void *uid); +int +syncopctx_setfsgid(void *gid); +int +syncopctx_setfsgroups(int count, const void *groups); +int +syncopctx_setfspid(void *pid); +int +syncopctx_setfslkowner(gf_lkowner_t *lk_owner); + +static inline call_frame_t * +syncop_create_frame(xlator_t *this) +{ + call_frame_t *frame = NULL; + int ngrps = -1; + struct syncopctx *opctx = NULL; + + frame = create_frame(this, this->ctx->pool); + if (!frame) + return NULL; + + frame->root->type = GF_OP_TYPE_FOP; + opctx = syncopctx_getctx(); + + if (opctx && (opctx->valid & SYNCOPCTX_PID)) + frame->root->pid = opctx->pid; + else + frame->root->pid = getpid(); + + if (opctx && (opctx->valid & SYNCOPCTX_UID)) + frame->root->uid = opctx->uid; + else + frame->root->uid = geteuid(); + + if (opctx && (opctx->valid & SYNCOPCTX_GID)) + frame->root->gid = opctx->gid; + else + frame->root->gid = getegid(); + + if (opctx && (opctx->valid & SYNCOPCTX_GROUPS)) { + ngrps = opctx->ngrps; + + if (ngrps != 0 && opctx->groups != NULL) { + if (call_stack_alloc_groups(frame->root, ngrps) != 0) { + STACK_DESTROY(frame->root); + return NULL; + } + + memcpy(frame->root->groups, opctx->groups, (sizeof(gid_t) * ngrps)); + } + } else { + ngrps = getgroups(0, 0); + if (ngrps < 0) { + STACK_DESTROY(frame->root); + return NULL; + } + + if (call_stack_alloc_groups(frame->root, ngrps) != 0) { + STACK_DESTROY(frame->root); + return NULL; + } + + if (getgroups(ngrps, frame->root->groups) < 0) { + STACK_DESTROY(frame->root); + return NULL; + } + } + + if (opctx && (opctx->valid & SYNCOPCTX_LKOWNER)) + frame->root->lk_owner = opctx->lk_owner; + + return frame; +} + +int +synclock_init(synclock_t *lock, lock_attr_t attr); +int +synclock_destroy(synclock_t *lock); +int +synclock_lock(synclock_t *lock); +int +synclock_trylock(synclock_t *lock); +int +synclock_unlock(synclock_t *lock); + +int +syncbarrier_init(syncbarrier_t *barrier); +int +syncbarrier_wait(syncbarrier_t *barrier, int waitfor); +int +syncbarrier_wake(syncbarrier_t *barrier); +int +syncbarrier_destroy(syncbarrier_t *barrier); + +int +syncop_lookup(xlator_t *subvol, loc_t *loc, + /* out */ + struct iatt *iatt, struct iatt *parent, + /* xdata */ + dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_readdirp(xlator_t *subvol, fd_t *fd, size_t size, off_t off, + /* out */ + gf_dirent_t *entries, dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_readdir(xlator_t *subvol, fd_t *fd, size_t size, off_t off, + gf_dirent_t *entries, dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_opendir(xlator_t *subvol, loc_t *loc, fd_t *fd, dict_t *xdata_in, + dict_t **xdata_out); + +int +syncop_setattr(xlator_t *subvol, loc_t *loc, struct iatt *iatt, int valid, + /* out */ + struct iatt *preop, struct iatt *postop, dict_t *xdata_in, + dict_t **xdata_out); + +int +syncop_fsetattr(xlator_t *subvol, fd_t *fd, struct iatt *iatt, int valid, + /* out */ + struct iatt *preop, struct iatt *postop, dict_t *xdata_in, + dict_t **xdata_out); + +int +syncop_statfs(xlator_t *subvol, loc_t *loc, + /* out */ + struct statvfs *buf, dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_setxattr(xlator_t *subvol, loc_t *loc, dict_t *dict, int32_t flags, + dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_fsetxattr(xlator_t *subvol, fd_t *fd, dict_t *dict, int32_t flags, + dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_listxattr(xlator_t *subvol, loc_t *loc, dict_t **dict, dict_t *xdata_in, + dict_t **xdata_out); + +int +syncop_getxattr(xlator_t *xl, loc_t *loc, dict_t **dict, const char *key, + dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_fgetxattr(xlator_t *xl, fd_t *fd, dict_t **dict, const char *key, + dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_removexattr(xlator_t *subvol, loc_t *loc, const char *name, + dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_fremovexattr(xlator_t *subvol, fd_t *fd, const char *name, + dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_create(xlator_t *subvol, loc_t *loc, int32_t flags, mode_t mode, + fd_t *fd, struct iatt *iatt, dict_t *xdata_in, + dict_t **xdata_out); + +int +syncop_open(xlator_t *subvol, loc_t *loc, int32_t flags, fd_t *fd, + dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_close(fd_t *fd); + +int +syncop_write(xlator_t *subvol, fd_t *fd, const char *buf, int size, + off_t offset, struct iobref *iobref, uint32_t flags, + dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_writev(xlator_t *subvol, fd_t *fd, const struct iovec *vector, + int32_t count, off_t offset, struct iobref *iobref, + uint32_t flags, struct iatt *preiatt, struct iatt *postiatt, + dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_readv(xlator_t *subvol, fd_t *fd, size_t size, off_t off, uint32_t flags, + /* out */ + struct iovec **vector, int *count, struct iobref **iobref, + struct iatt *iatt, dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_ftruncate(xlator_t *subvol, fd_t *fd, off_t offset, struct iatt *preiatt, + struct iatt *postiatt, dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_truncate(xlator_t *subvol, loc_t *loc, off_t offset, dict_t *xdata_in, + dict_t **xdata_out); + +int +syncop_unlink(xlator_t *subvol, loc_t *loc, dict_t *xdata_in, + dict_t **xdata_out); + +int +syncop_rmdir(xlator_t *subvol, loc_t *loc, int flags, dict_t *xdata_in, + dict_t **xdata_out); + +int +syncop_fsync(xlator_t *subvol, fd_t *fd, int dataonly, struct iatt *preiatt, + struct iatt *postiatt, dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_flush(xlator_t *subvol, fd_t *fd, dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_fstat(xlator_t *subvol, fd_t *fd, struct iatt *stbuf, dict_t *xdata_in, + dict_t **xdata_out); + +int +syncop_stat(xlator_t *subvol, loc_t *loc, struct iatt *stbuf, dict_t *xdata_in, + dict_t **xdata_out); + +int +syncop_symlink(xlator_t *subvol, loc_t *loc, const char *newpath, + struct iatt *iatt, dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_readlink(xlator_t *subvol, loc_t *loc, char **buffer, size_t size, + dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_mknod(xlator_t *subvol, loc_t *loc, mode_t mode, dev_t rdev, + struct iatt *iatt, dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_mkdir(xlator_t *subvol, loc_t *loc, mode_t mode, struct iatt *iatt, + dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_link(xlator_t *subvol, loc_t *oldloc, loc_t *newloc, struct iatt *iatt, + dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_fsyncdir(xlator_t *subvol, fd_t *fd, int datasync, dict_t *xdata_in, + dict_t **xdata_out); + +int +syncop_access(xlator_t *subvol, loc_t *loc, int32_t mask, dict_t *xdata_in, + dict_t **xdata_out); + +int +syncop_fallocate(xlator_t *subvol, fd_t *fd, int32_t keep_size, off_t offset, + size_t len, dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_discard(xlator_t *subvol, fd_t *fd, off_t offset, size_t len, + dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_zerofill(xlator_t *subvol, fd_t *fd, off_t offset, off_t len, + dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_rename(xlator_t *subvol, loc_t *oldloc, loc_t *newloc, dict_t *xdata_in, + dict_t **xdata_out); + +int +syncop_lk(xlator_t *subvol, fd_t *fd, int cmd, struct gf_flock *flock, + dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_inodelk(xlator_t *subvol, const char *volume, loc_t *loc, int32_t cmd, + struct gf_flock *lock, dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_lease(xlator_t *subvol, loc_t *loc, struct gf_lease *lease, + dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_ipc(xlator_t *subvol, int op, dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_xattrop(xlator_t *subvol, loc_t *loc, gf_xattrop_flags_t flags, + dict_t *dict, dict_t *xdata_in, dict_t **dict_out, + dict_t **xdata_out); + +int +syncop_fxattrop(xlator_t *subvol, fd_t *fd, gf_xattrop_flags_t flags, + dict_t *dict, dict_t *xdata_in, dict_t **dict_out, + dict_t **xdata_out); + +int +syncop_seek(xlator_t *subvol, fd_t *fd, off_t offset, gf_seek_what_t what, + dict_t *xdata_in, off_t *off); + +int +syncop_getactivelk(xlator_t *subvol, loc_t *loc, + lock_migration_info_t *locklist, dict_t *xdata_in, + dict_t **xdata_out); + +int +syncop_setactivelk_cbk(call_frame_t *frame, void *cookie, xlator_t *this, + int32_t op_ret, int32_t op_errno, dict_t *xdata); + +int +syncop_setactivelk(xlator_t *subvol, loc_t *loc, + lock_migration_info_t *locklist, dict_t *xdata_in, + dict_t **xdata_out); + +int +syncop_put(xlator_t *subvol, loc_t *loc, mode_t mode, mode_t umask, + uint32_t flags, struct iovec *vector, int32_t count, off_t offset, + struct iobref *iobref, dict_t *xattr, struct iatt *iatt, + dict_t *xdata_in, dict_t **xdata_out); + +int +syncop_setactivelk_cbk(call_frame_t *frame, void *cookie, xlator_t *this, + int32_t op_ret, int32_t op_errno, dict_t *xdata); + +int +syncop_icreate(xlator_t *subvol, loc_t *loc, mode_t mode, dict_t *xdata_out); + +int +syncop_entrylk(xlator_t *subvol, const char *volume, loc_t *loc, + const char *basename, entrylk_cmd cmd, entrylk_type type, + dict_t *xdata_in, dict_t **xdata_out); + +#endif /* _SYNCOP_H */ |