libeio/etp.c

/*
 * libetp implementation
 *
 * Copyright (c) 2007,2008,2009,2010,2011,2012,2013 Marc Alexander Lehmann <libetp@schmorp.de>
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without modifica-
 * tion, 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.
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED
 * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MER-
 * CHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO
 * EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPE-
 * CIAL, 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 OTH-
 * ERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
 * OF THE POSSIBILITY OF SUCH DAMAGE.
 *
 * Alternatively, the contents of this file may be used under the terms of
 * the GNU General Public License ("GPL") version 2 or any later version,
 * in which case the provisions of the GPL are applicable instead of
 * the above. If you wish to allow the use of your version of this file
 * only under the terms of the GPL and not to allow others to use your
 * version of this file under the BSD license, indicate your decision
 * by deleting the provisions above and replace them with the notice
 * and other provisions required by the GPL. If you do not delete the
 * provisions above, a recipient may use your version of this file under
 * either the BSD or the GPL.
 */

#ifndef ETP_API_DECL
# define ETP_API_DECL static
#endif

#ifndef ETP_PRI_MIN
# define ETP_PRI_MIN 0
# define ETP_PRI_MAX 0
#endif

#ifndef ETP_TYPE_QUIT
# define ETP_TYPE_QUIT 0
#endif

#ifndef ETP_TYPE_GROUP
# define ETP_TYPE_GROUP 1
#endif

#ifndef ETP_WANT_POLL
# define ETP_WANT_POLL(pool) pool->want_poll_cb (pool->userdata)
#endif
#ifndef ETP_DONE_POLL
# define ETP_DONE_POLL(pool) pool->done_poll_cb (pool->userdata)
#endif

#define ETP_NUM_PRI (ETP_PRI_MAX - ETP_PRI_MIN + 1)

#define ETP_TICKS ((1000000 + 1023) >> 10)

enum {
  ETP_FLAG_GROUPADD = 0x04, /* some request was added to the group */
  ETP_FLAG_DELAYED  = 0x08, /* groiup request has been delayed */
};

/* calculate time difference in ~1/ETP_TICKS of a second */
ecb_inline int
etp_tvdiff (struct timeval *tv1, struct timeval *tv2)
{
  return  (tv2->tv_sec  - tv1->tv_sec ) * ETP_TICKS
       + ((tv2->tv_usec - tv1->tv_usec) >> 10);
}

struct etp_tmpbuf
{
  void *ptr;
  int len;
};

static void *
etp_tmpbuf_get (struct etp_tmpbuf *buf, int len)
{
  if (buf->len < len)
    {
      free (buf->ptr);
      buf->ptr = malloc (buf->len = len);
    }

  return buf->ptr;
}

/*
 * a somewhat faster data structure might be nice, but
 * with 8 priorities this actually needs <20 insns
 * per shift, the most expensive operation.
 */
typedef struct
{
  ETP_REQ *qs[ETP_NUM_PRI], *qe[ETP_NUM_PRI]; /* qstart, qend */
  int size;
} etp_reqq;

struct etp_pool
{
   void *userdata;

   etp_reqq req_queue;
   etp_reqq res_queue;

   unsigned int started, idle, wanted;

   unsigned int max_poll_time;     /* pool->reslock */
   unsigned int max_poll_reqs;     /* pool->reslock */

   unsigned int nreqs;    /* pool->reqlock */
   unsigned int nready;   /* pool->reqlock */
   unsigned int npending; /* pool->reqlock */
   unsigned int max_idle;      /* maximum number of threads that can pool->idle indefinitely */
   unsigned int idle_timeout; /* number of seconds after which an pool->idle threads exit */

   void (*want_poll_cb) (void *userdata);
   void (*done_poll_cb) (void *userdata);
 
   xmutex_t wrklock;
   xmutex_t reslock;
   xmutex_t reqlock;
   xcond_t  reqwait;
};

typedef struct etp_pool *etp_pool;

typedef struct etp_worker
{
  etp_pool pool;

  struct etp_tmpbuf tmpbuf;

  /* locked by pool->wrklock */
  struct etp_worker *prev, *next;

  xthread_t tid;

#ifdef ETP_WORKER_COMMON
  ETP_WORKER_COMMON
#endif
} etp_worker;

static etp_worker wrk_first; /* NOT etp */

#define ETP_WORKER_LOCK(wrk)   X_LOCK   (pool->wrklock)
#define ETP_WORKER_UNLOCK(wrk) X_UNLOCK (pool->wrklock)

/* worker threads management */

static void
etp_worker_clear (etp_worker *wrk)
{
}

static void ecb_cold
etp_worker_free (etp_worker *wrk)
{
  free (wrk->tmpbuf.ptr);

  wrk->next->prev = wrk->prev;
  wrk->prev->next = wrk->next;

  free (wrk);
}

ETP_API_DECL unsigned int
etp_nreqs (etp_pool pool)
{
  int retval;
  if (WORDACCESS_UNSAFE) X_LOCK   (pool->reqlock);
  retval = pool->nreqs;
  if (WORDACCESS_UNSAFE) X_UNLOCK (pool->reqlock);
  return retval;
}

ETP_API_DECL unsigned int
etp_nready (etp_pool pool)
{
  unsigned int retval;

  if (WORDACCESS_UNSAFE) X_LOCK   (pool->reqlock);
  retval = pool->nready;
  if (WORDACCESS_UNSAFE) X_UNLOCK (pool->reqlock);

  return retval;
}

ETP_API_DECL unsigned int
etp_npending (etp_pool pool)
{
  unsigned int retval;

  if (WORDACCESS_UNSAFE) X_LOCK   (pool->reqlock);
  retval = pool->npending;
  if (WORDACCESS_UNSAFE) X_UNLOCK (pool->reqlock);

  return retval;
}

ETP_API_DECL unsigned int
etp_nthreads (etp_pool pool)
{
  unsigned int retval;

  if (WORDACCESS_UNSAFE) X_LOCK   (pool->reqlock);
  retval = pool->started;
  if (WORDACCESS_UNSAFE) X_UNLOCK (pool->reqlock);

  return retval;
}

static void ecb_noinline ecb_cold
reqq_init (etp_reqq *q)
{
  int pri;

  for (pri = 0; pri < ETP_NUM_PRI; ++pri)
    q->qs[pri] = q->qe[pri] = 0;

  q->size = 0;
}

static int ecb_noinline
reqq_push (etp_reqq *q, ETP_REQ *req)
{
  int pri = req->pri;
  req->next = 0;

  if (q->qe[pri])
    {
      q->qe[pri]->next = req;
      q->qe[pri] = req;
    }
  else
    q->qe[pri] = q->qs[pri] = req;

  return q->size++;
}

static ETP_REQ * ecb_noinline
reqq_shift (etp_reqq *q)
{
  int pri;

  if (!q->size)
    return 0;

  --q->size;

  for (pri = ETP_NUM_PRI; pri--; )
    {
      ETP_REQ *req = q->qs[pri];

      if (req)
        {
          if (!(q->qs[pri] = (ETP_REQ *)req->next))
            q->qe[pri] = 0;

          return req;
        }
    }

  abort ();
}

ETP_API_DECL int ecb_cold
etp_init (etp_pool pool, void (*want_poll)(void *userdata), void (*done_poll)(void *userdata))
{
  X_MUTEX_CREATE (pool->wrklock);
  X_MUTEX_CREATE (pool->reslock);
  X_MUTEX_CREATE (pool->reqlock);
  X_COND_CREATE  (pool->reqwait);

  reqq_init (&pool->req_queue);
  reqq_init (&pool->res_queue);

  wrk_first.next =
  wrk_first.prev = &wrk_first;

  pool->started  = 0;
  pool->idle     = 0;
  pool->nreqs    = 0;
  pool->nready   = 0;
  pool->npending = 0;
  pool->wanted   = 4;

  pool->max_idle = 4;      /* maximum number of threads that can pool->idle indefinitely */
  pool->idle_timeout = 10; /* number of seconds after which an pool->idle threads exit */

  pool->want_poll_cb = want_poll;
  pool->done_poll_cb = done_poll;

  return 0;
}

static void ecb_noinline ecb_cold
etp_proc_init (void)
{
#if HAVE_PRCTL_SET_NAME
  /* provide a more sensible "thread name" */
  char name[16 + 1];
  const int namelen = sizeof (name) - 1;
  int len;

  prctl (PR_GET_NAME, (unsigned long)name, 0, 0, 0);
  name [namelen] = 0;
  len = strlen (name);
  strcpy (name + (len <= namelen - 4 ? len : namelen - 4), "/eio");
  prctl (PR_SET_NAME, (unsigned long)name, 0, 0, 0);
#endif
}

X_THREAD_PROC (etp_proc)
{
  ETP_REQ *req;
  struct timespec ts;
  etp_worker *self = (etp_worker *)thr_arg;
  etp_pool pool = self->pool;

  etp_proc_init ();

  /* try to distribute timeouts somewhat evenly */
  ts.tv_nsec = ((unsigned long)self & 1023UL) * (1000000000UL / 1024UL);

  for (;;)
    {
      ts.tv_sec = 0;

      X_LOCK (pool->reqlock);

      for (;;)
        {
          req = reqq_shift (&pool->req_queue);

          if (req)
            break;

          if (ts.tv_sec == 1) /* no request, but timeout detected, let's quit */
            {
              X_UNLOCK (pool->reqlock);
              X_LOCK (pool->wrklock);
              --pool->started;
              X_UNLOCK (pool->wrklock);
              goto quit;
            }

          ++pool->idle;

          if (pool->idle <= pool->max_idle)
            /* we are allowed to pool->idle, so do so without any timeout */
            X_COND_WAIT (pool->reqwait, pool->reqlock);
          else
            {
              /* initialise timeout once */
              if (!ts.tv_sec)
                ts.tv_sec = time (0) + pool->idle_timeout;

              if (X_COND_TIMEDWAIT (pool->reqwait, pool->reqlock, ts) == ETIMEDOUT)
                ts.tv_sec = 1; /* assuming this is not a value computed above.,.. */
            }

          --pool->idle;
        }

      --pool->nready;

      X_UNLOCK (pool->reqlock);
     
      if (req->type == ETP_TYPE_QUIT)
        goto quit;

      ETP_EXECUTE (self, req);

      X_LOCK (pool->reslock);

      ++pool->npending;

      if (!reqq_push (&pool->res_queue, req))
        ETP_WANT_POLL (poll);

      etp_worker_clear (self);

      X_UNLOCK (pool->reslock);
    }

quit:
  free (req);

  X_LOCK (pool->wrklock);
  etp_worker_free (self);
  X_UNLOCK (pool->wrklock);

  return 0;
}

static void ecb_cold
etp_start_thread (etp_pool pool)
{
  etp_worker *wrk = calloc (1, sizeof (etp_worker));

  /*TODO*/
  assert (("unable to allocate worker thread data", wrk));

  wrk->pool = pool;

  X_LOCK (pool->wrklock);

  if (xthread_create (&wrk->tid, etp_proc, (void *)wrk))
    {
      wrk->prev = &wrk_first;
      wrk->next = wrk_first.next;
      wrk_first.next->prev = wrk;
      wrk_first.next = wrk;
      ++pool->started;
    }
  else
    free (wrk);

  X_UNLOCK (pool->wrklock);
}

static void
etp_maybe_start_thread (etp_pool pool)
{
  if (ecb_expect_true (etp_nthreads (pool) >= pool->wanted))
    return;
  
  /* todo: maybe use pool->idle here, but might be less exact */
  if (ecb_expect_true (0 <= (int)etp_nthreads (pool) + (int)etp_npending (pool) - (int)etp_nreqs (pool)))
    return;

  etp_start_thread (pool);
}

static void ecb_cold
etp_end_thread (etp_pool pool)
{
  ETP_REQ *req = calloc (1, sizeof (ETP_REQ)); /* will be freed by worker */

  req->type = ETP_TYPE_QUIT;
  req->pri  = ETP_PRI_MAX - ETP_PRI_MIN;

  X_LOCK (pool->reqlock);
  reqq_push (&pool->req_queue, req);
  X_COND_SIGNAL (pool->reqwait);
  X_UNLOCK (pool->reqlock);

  X_LOCK (pool->wrklock);
  --pool->started;
  X_UNLOCK (pool->wrklock);
}

ETP_API_DECL int
etp_poll (etp_pool pool)
{
  unsigned int maxreqs;
  unsigned int maxtime;
  struct timeval tv_start, tv_now;

  X_LOCK (pool->reslock);
  maxreqs = pool->max_poll_reqs;
  maxtime = pool->max_poll_time;
  X_UNLOCK (pool->reslock);

  if (maxtime)
    gettimeofday (&tv_start, 0);

  for (;;)
    {
      ETP_REQ *req;

      etp_maybe_start_thread (pool);

      X_LOCK (pool->reslock);
      req = reqq_shift (&pool->res_queue);

      if (req)
        {
          --pool->npending;

          if (!pool->res_queue.size)
            ETP_DONE_POLL (pool->userdata);
        }

      X_UNLOCK (pool->reslock);

      if (!req)
        return 0;

      X_LOCK (pool->reqlock);
      --pool->nreqs;
      X_UNLOCK (pool->reqlock);

      if (ecb_expect_false (req->type == ETP_TYPE_GROUP && req->size))
        {
          req->flags |= ETP_FLAG_DELAYED; /* mark request as delayed */
          continue;
        }
      else
        {
          int res = ETP_FINISH (req);
          if (ecb_expect_false (res))
            return res;
        }

      if (ecb_expect_false (maxreqs && !--maxreqs))
        break;

      if (maxtime)
        {
          gettimeofday (&tv_now, 0);

          if (etp_tvdiff (&tv_start, &tv_now) >= maxtime)
            break;
        }
    }

  errno = EAGAIN;
  return -1;
}

ETP_API_DECL void
etp_grp_cancel (etp_pool pool, ETP_REQ *grp);

ETP_API_DECL void
etp_cancel (etp_pool pool, ETP_REQ *req)
{
  req->cancelled = 1;

  etp_grp_cancel (pool, req);
}

ETP_API_DECL void
etp_grp_cancel (etp_pool pool, ETP_REQ *grp)
{
  for (grp = grp->grp_first; grp; grp = grp->grp_next)
    etp_cancel (pool, grp);
}

ETP_API_DECL void
etp_submit (etp_pool pool, ETP_REQ *req)
{
  req->pri -= ETP_PRI_MIN;

  if (ecb_expect_false (req->pri < ETP_PRI_MIN - ETP_PRI_MIN)) req->pri = ETP_PRI_MIN - ETP_PRI_MIN;
  if (ecb_expect_false (req->pri > ETP_PRI_MAX - ETP_PRI_MIN)) req->pri = ETP_PRI_MAX - ETP_PRI_MIN;

  if (ecb_expect_false (req->type == ETP_TYPE_GROUP))
    {
      /* I hope this is worth it :/ */
      X_LOCK (pool->reqlock);
      ++pool->nreqs;
      X_UNLOCK (pool->reqlock);

      X_LOCK (pool->reslock);

      ++pool->npending;

      if (!reqq_push (&pool->res_queue, req))
        ETP_WANT_POLL (pool);

      X_UNLOCK (pool->reslock);
    }
  else
    {
      X_LOCK (pool->reqlock);
      ++pool->nreqs;
      ++pool->nready;
      reqq_push (&pool->req_queue, req);
      X_COND_SIGNAL (pool->reqwait);
      X_UNLOCK (pool->reqlock);

      etp_maybe_start_thread (pool);
    }
}

ETP_API_DECL void ecb_cold
etp_set_max_poll_time (etp_pool pool, double nseconds)
{
  if (WORDACCESS_UNSAFE) X_LOCK   (pool->reslock);
  pool->max_poll_time = nseconds * ETP_TICKS;
  if (WORDACCESS_UNSAFE) X_UNLOCK (pool->reslock);
}

ETP_API_DECL void ecb_cold
etp_set_max_poll_reqs (etp_pool pool, unsigned int maxreqs)
{
  if (WORDACCESS_UNSAFE) X_LOCK   (pool->reslock);
  pool->max_poll_reqs = maxreqs;
  if (WORDACCESS_UNSAFE) X_UNLOCK (pool->reslock);
}

ETP_API_DECL void ecb_cold
etp_set_max_idle (etp_pool pool, unsigned int nthreads)
{
  if (WORDACCESS_UNSAFE) X_LOCK   (pool->reqlock);
  pool->max_idle = nthreads;
  if (WORDACCESS_UNSAFE) X_UNLOCK (pool->reqlock);
}

ETP_API_DECL void ecb_cold
etp_set_idle_timeout (etp_pool pool, unsigned int seconds)
{
  if (WORDACCESS_UNSAFE) X_LOCK   (pool->reqlock);
  pool->idle_timeout = seconds;
  if (WORDACCESS_UNSAFE) X_UNLOCK (pool->reqlock);
}

ETP_API_DECL void ecb_cold
etp_set_min_parallel (etp_pool pool, unsigned int nthreads)
{
  if (pool->wanted < nthreads)
    pool->wanted = nthreads;
}

ETP_API_DECL void ecb_cold
etp_set_max_parallel (etp_pool pool, unsigned int nthreads)
{
  if (pool->wanted > nthreads)
    pool->wanted = nthreads;

  while (pool->started > pool->wanted)
    etp_end_thread (pool);
}

Revision:	1.6
Committed:	Thu Jun 25 18:08:47 2015 UTC (8 years, 11 months ago) by root
Content type:	text/plain
Branch:	MAIN
Changes since 1.5:	+133 -120 lines
Log Message:	* empty log message *
#	User	Rev	Content
1	root	1.1	/*
2			* libetp implementation
3			*
4			* Copyright (c) 2007,2008,2009,2010,2011,2012,2013 Marc Alexander Lehmann <libetp@schmorp.de>
5			* All rights reserved.
6			*
7			* Redistribution and use in source and binary forms, with or without modifica-
8			* tion, are permitted provided that the following conditions are met:
9			*
10			* 1. Redistributions of source code must retain the above copyright notice,
11			* this list of conditions and the following disclaimer.
12			*
13			* 2. Redistributions in binary form must reproduce the above copyright
14			* notice, this list of conditions and the following disclaimer in the
15			* documentation and/or other materials provided with the distribution.
16			*
17			* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED
18			* WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MER-
19			* CHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO
20			* EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPE-
21			* CIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
22			* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
23			* OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
24			* WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTH-
25			* ERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
26			* OF THE POSSIBILITY OF SUCH DAMAGE.
27			*
28			* Alternatively, the contents of this file may be used under the terms of
29			* the GNU General Public License ("GPL") version 2 or any later version,
30			* in which case the provisions of the GPL are applicable instead of
31			* the above. If you wish to allow the use of your version of this file
32			* only under the terms of the GPL and not to allow others to use your
33			* version of this file under the BSD license, indicate your decision
34			* by deleting the provisions above and replace them with the notice
35			* and other provisions required by the GPL. If you do not delete the
36			* provisions above, a recipient may use your version of this file under
37			* either the BSD or the GPL.
38			*/
39
40			#ifndef ETP_API_DECL
41			# define ETP_API_DECL static
42			#endif
43
44			#ifndef ETP_PRI_MIN
45			# define ETP_PRI_MIN 0
46			# define ETP_PRI_MAX 0
47			#endif
48
49			#ifndef ETP_TYPE_QUIT
50			# define ETP_TYPE_QUIT 0
51			#endif
52
53			#ifndef ETP_TYPE_GROUP
54			# define ETP_TYPE_GROUP 1
55			#endif
56
57	root	1.6	#ifndef ETP_WANT_POLL
58			# define ETP_WANT_POLL(pool) pool->want_poll_cb (pool->userdata)
59			#endif
60			#ifndef ETP_DONE_POLL
61			# define ETP_DONE_POLL(pool) pool->done_poll_cb (pool->userdata)
62			#endif
63
64	root	1.1	#define ETP_NUM_PRI (ETP_PRI_MAX - ETP_PRI_MIN + 1)
65
66			#define ETP_TICKS ((1000000 + 1023) >> 10)
67
68	root	1.4	enum {
69			ETP_FLAG_GROUPADD = 0x04, /* some request was added to the group */
70			ETP_FLAG_DELAYED = 0x08, /* groiup request has been delayed */
71			};
72
73	root	1.1	/* calculate time difference in ~1/ETP_TICKS of a second */
74			ecb_inline int
75			etp_tvdiff (struct timeval tv1, struct timeval tv2)
76			{
77			return (tv2->tv_sec - tv1->tv_sec ) * ETP_TICKS
78			+ ((tv2->tv_usec - tv1->tv_usec) >> 10);
79			}
80
81	root	1.3	struct etp_tmpbuf
82			{
83			void *ptr;
84			int len;
85			};
86
87			static void *
88			etp_tmpbuf_get (struct etp_tmpbuf *buf, int len)
89			{
90			if (buf->len < len)
91			{
92			free (buf->ptr);
93			buf->ptr = malloc (buf->len = len);
94			}
95
96			return buf->ptr;
97			}
98
99	root	1.5	/*
100			* a somewhat faster data structure might be nice, but
101			* with 8 priorities this actually needs <20 insns
102			* per shift, the most expensive operation.
103			*/
104			typedef struct
105			{
106			ETP_REQ qs[ETP_NUM_PRI], qe[ETP_NUM_PRI]; /* qstart, qend */
107			int size;
108			} etp_reqq;
109
110			struct etp_pool
111			{
112	root	1.6	void *userdata;
113
114	root	1.5	etp_reqq req_queue;
115			etp_reqq res_queue;
116	root	1.6
117			unsigned int started, idle, wanted;
118
119			unsigned int max_poll_time; /* pool->reslock */
120			unsigned int max_poll_reqs; /* pool->reslock */
121
122			unsigned int nreqs; /* pool->reqlock */
123			unsigned int nready; /* pool->reqlock */
124			unsigned int npending; /* pool->reqlock */
125			unsigned int max_idle; /* maximum number of threads that can pool->idle indefinitely */
126			unsigned int idle_timeout; /* number of seconds after which an pool->idle threads exit */
127
128			void (want_poll_cb) (void userdata);
129			void (done_poll_cb) (void userdata);
130
131			xmutex_t wrklock;
132			xmutex_t reslock;
133			xmutex_t reqlock;
134			xcond_t reqwait;
135	root	1.5	};
136
137			typedef struct etp_pool *etp_pool;
138
139	root	1.1	typedef struct etp_worker
140			{
141	root	1.5	etp_pool pool;
142
143	root	1.3	struct etp_tmpbuf tmpbuf;
144	root	1.1
145	root	1.6	/* locked by pool->wrklock */
146	root	1.1	struct etp_worker prev, next;
147
148			xthread_t tid;
149
150			#ifdef ETP_WORKER_COMMON
151			ETP_WORKER_COMMON
152			#endif
153			} etp_worker;
154
155			static etp_worker wrk_first; /* NOT etp */
156
157	root	1.6	#define ETP_WORKER_LOCK(wrk) X_LOCK (pool->wrklock)
158			#define ETP_WORKER_UNLOCK(wrk) X_UNLOCK (pool->wrklock)
159	root	1.1
160			/* worker threads management */
161
162			static void
163			etp_worker_clear (etp_worker *wrk)
164			{
165			}
166
167			static void ecb_cold
168			etp_worker_free (etp_worker *wrk)
169			{
170			free (wrk->tmpbuf.ptr);
171
172			wrk->next->prev = wrk->prev;
173			wrk->prev->next = wrk->next;
174
175			free (wrk);
176			}
177
178			ETP_API_DECL unsigned int
179	root	1.5	etp_nreqs (etp_pool pool)
180	root	1.1	{
181			int retval;
182	root	1.6	if (WORDACCESS_UNSAFE) X_LOCK (pool->reqlock);
183			retval = pool->nreqs;
184			if (WORDACCESS_UNSAFE) X_UNLOCK (pool->reqlock);
185	root	1.1	return retval;
186			}
187
188			ETP_API_DECL unsigned int
189	root	1.5	etp_nready (etp_pool pool)
190	root	1.1	{
191			unsigned int retval;
192
193	root	1.6	if (WORDACCESS_UNSAFE) X_LOCK (pool->reqlock);
194			retval = pool->nready;
195			if (WORDACCESS_UNSAFE) X_UNLOCK (pool->reqlock);
196	root	1.1
197			return retval;
198			}
199
200			ETP_API_DECL unsigned int
201	root	1.5	etp_npending (etp_pool pool)
202	root	1.1	{
203			unsigned int retval;
204
205	root	1.6	if (WORDACCESS_UNSAFE) X_LOCK (pool->reqlock);
206			retval = pool->npending;
207			if (WORDACCESS_UNSAFE) X_UNLOCK (pool->reqlock);
208	root	1.1
209			return retval;
210			}
211
212			ETP_API_DECL unsigned int
213	root	1.5	etp_nthreads (etp_pool pool)
214	root	1.1	{
215			unsigned int retval;
216
217	root	1.6	if (WORDACCESS_UNSAFE) X_LOCK (pool->reqlock);
218			retval = pool->started;
219			if (WORDACCESS_UNSAFE) X_UNLOCK (pool->reqlock);
220	root	1.1
221			return retval;
222			}
223
224			static void ecb_noinline ecb_cold
225			reqq_init (etp_reqq *q)
226			{
227			int pri;
228
229			for (pri = 0; pri < ETP_NUM_PRI; ++pri)
230			q->qs[pri] = q->qe[pri] = 0;
231
232			q->size = 0;
233			}
234
235			static int ecb_noinline
236			reqq_push (etp_reqq q, ETP_REQ req)
237			{
238			int pri = req->pri;
239			req->next = 0;
240
241			if (q->qe[pri])
242			{
243			q->qe[pri]->next = req;
244			q->qe[pri] = req;
245			}
246			else
247			q->qe[pri] = q->qs[pri] = req;
248
249			return q->size++;
250			}
251
252			static ETP_REQ * ecb_noinline
253			reqq_shift (etp_reqq *q)
254			{
255			int pri;
256
257			if (!q->size)
258			return 0;
259
260			--q->size;
261
262			for (pri = ETP_NUM_PRI; pri--; )
263			{
264	root	1.2	ETP_REQ *req = q->qs[pri];
265	root	1.1
266			if (req)
267			{
268	root	1.2	if (!(q->qs[pri] = (ETP_REQ *)req->next))
269	root	1.1	q->qe[pri] = 0;
270
271			return req;
272			}
273			}
274
275			abort ();
276			}
277
278			ETP_API_DECL int ecb_cold
279	root	1.6	etp_init (etp_pool pool, void (want_poll)(void userdata), void (done_poll)(void userdata))
280	root	1.1	{
281	root	1.6	X_MUTEX_CREATE (pool->wrklock);
282			X_MUTEX_CREATE (pool->reslock);
283			X_MUTEX_CREATE (pool->reqlock);
284			X_COND_CREATE (pool->reqwait);
285	root	1.1
286	root	1.5	reqq_init (&pool->req_queue);
287			reqq_init (&pool->res_queue);
288	root	1.1
289			wrk_first.next =
290			wrk_first.prev = &wrk_first;
291
292	root	1.6	pool->started = 0;
293			pool->idle = 0;
294			pool->nreqs = 0;
295			pool->nready = 0;
296			pool->npending = 0;
297			pool->wanted = 4;
298
299			pool->max_idle = 4; /* maximum number of threads that can pool->idle indefinitely */
300			pool->idle_timeout = 10; /* number of seconds after which an pool->idle threads exit */
301	root	1.1
302	root	1.6	pool->want_poll_cb = want_poll;
303			pool->done_poll_cb = done_poll;
304	root	1.1
305			return 0;
306			}
307
308	root	1.4	static void ecb_noinline ecb_cold
309			etp_proc_init (void)
310			{
311			#if HAVE_PRCTL_SET_NAME
312			/* provide a more sensible "thread name" */
313			char name[16 + 1];
314			const int namelen = sizeof (name) - 1;
315			int len;
316
317			prctl (PR_GET_NAME, (unsigned long)name, 0, 0, 0);
318			name [namelen] = 0;
319			len = strlen (name);
320			strcpy (name + (len <= namelen - 4 ? len : namelen - 4), "/eio");
321			prctl (PR_SET_NAME, (unsigned long)name, 0, 0, 0);
322			#endif
323			}
324
325			X_THREAD_PROC (etp_proc)
326			{
327			ETP_REQ *req;
328			struct timespec ts;
329			etp_worker self = (etp_worker )thr_arg;
330	root	1.5	etp_pool pool = self->pool;
331	root	1.4
332			etp_proc_init ();
333
334			/* try to distribute timeouts somewhat evenly */
335			ts.tv_nsec = ((unsigned long)self & 1023UL) * (1000000000UL / 1024UL);
336
337			for (;;)
338			{
339			ts.tv_sec = 0;
340
341	root	1.6	X_LOCK (pool->reqlock);
342	root	1.4
343			for (;;)
344			{
345	root	1.5	req = reqq_shift (&pool->req_queue);
346	root	1.4
347			if (req)
348			break;
349
350			if (ts.tv_sec == 1) /* no request, but timeout detected, let's quit */
351			{
352	root	1.6	X_UNLOCK (pool->reqlock);
353			X_LOCK (pool->wrklock);
354			--pool->started;
355			X_UNLOCK (pool->wrklock);
356	root	1.4	goto quit;
357			}
358
359	root	1.6	++pool->idle;
360	root	1.4
361	root	1.6	if (pool->idle <= pool->max_idle)
362			/* we are allowed to pool->idle, so do so without any timeout */
363			X_COND_WAIT (pool->reqwait, pool->reqlock);
364	root	1.4	else
365			{
366			/* initialise timeout once */
367			if (!ts.tv_sec)
368	root	1.6	ts.tv_sec = time (0) + pool->idle_timeout;
369	root	1.4
370	root	1.6	if (X_COND_TIMEDWAIT (pool->reqwait, pool->reqlock, ts) == ETIMEDOUT)
371	root	1.4	ts.tv_sec = 1; /* assuming this is not a value computed above.,.. */
372			}
373
374	root	1.6	--pool->idle;
375	root	1.4	}
376
377	root	1.6	--pool->nready;
378	root	1.4
379	root	1.6	X_UNLOCK (pool->reqlock);
380	root	1.4
381			if (req->type == ETP_TYPE_QUIT)
382			goto quit;
383
384			ETP_EXECUTE (self, req);
385
386	root	1.6	X_LOCK (pool->reslock);
387	root	1.4
388	root	1.6	++pool->npending;
389	root	1.4
390	root	1.6	if (!reqq_push (&pool->res_queue, req))
391			ETP_WANT_POLL (poll);
392	root	1.4
393			etp_worker_clear (self);
394
395	root	1.6	X_UNLOCK (pool->reslock);
396	root	1.4	}
397
398			quit:
399			free (req);
400
401	root	1.6	X_LOCK (pool->wrklock);
402	root	1.4	etp_worker_free (self);
403	root	1.6	X_UNLOCK (pool->wrklock);
404	root	1.4
405			return 0;
406			}
407	root	1.1
408			static void ecb_cold
409	root	1.5	etp_start_thread (etp_pool pool)
410	root	1.1	{
411			etp_worker *wrk = calloc (1, sizeof (etp_worker));
412
413			/TODO/
414			assert (("unable to allocate worker thread data", wrk));
415
416	root	1.5	wrk->pool = pool;
417
418	root	1.6	X_LOCK (pool->wrklock);
419	root	1.1
420			if (xthread_create (&wrk->tid, etp_proc, (void *)wrk))
421			{
422			wrk->prev = &wrk_first;
423			wrk->next = wrk_first.next;
424			wrk_first.next->prev = wrk;
425			wrk_first.next = wrk;
426	root	1.6	++pool->started;
427	root	1.1	}
428			else
429			free (wrk);
430
431	root	1.6	X_UNLOCK (pool->wrklock);
432	root	1.1	}
433
434			static void
435	root	1.5	etp_maybe_start_thread (etp_pool pool)
436	root	1.1	{
437	root	1.6	if (ecb_expect_true (etp_nthreads (pool) >= pool->wanted))
438	root	1.1	return;
439
440	root	1.6	/* todo: maybe use pool->idle here, but might be less exact */
441	root	1.5	if (ecb_expect_true (0 <= (int)etp_nthreads (pool) + (int)etp_npending (pool) - (int)etp_nreqs (pool)))
442	root	1.1	return;
443
444	root	1.5	etp_start_thread (pool);
445	root	1.1	}
446
447			static void ecb_cold
448	root	1.5	etp_end_thread (etp_pool pool)
449	root	1.1	{
450	root	1.2	ETP_REQ req = calloc (1, sizeof (ETP_REQ)); / will be freed by worker */
451	root	1.1
452			req->type = ETP_TYPE_QUIT;
453			req->pri = ETP_PRI_MAX - ETP_PRI_MIN;
454
455	root	1.6	X_LOCK (pool->reqlock);
456	root	1.5	reqq_push (&pool->req_queue, req);
457	root	1.6	X_COND_SIGNAL (pool->reqwait);
458			X_UNLOCK (pool->reqlock);
459	root	1.1
460	root	1.6	X_LOCK (pool->wrklock);
461			--pool->started;
462			X_UNLOCK (pool->wrklock);
463	root	1.1	}
464
465			ETP_API_DECL int
466	root	1.5	etp_poll (etp_pool pool)
467	root	1.1	{
468			unsigned int maxreqs;
469			unsigned int maxtime;
470			struct timeval tv_start, tv_now;
471
472	root	1.6	X_LOCK (pool->reslock);
473			maxreqs = pool->max_poll_reqs;
474			maxtime = pool->max_poll_time;
475			X_UNLOCK (pool->reslock);
476	root	1.1
477			if (maxtime)
478			gettimeofday (&tv_start, 0);
479
480			for (;;)
481			{
482			ETP_REQ *req;
483
484	root	1.5	etp_maybe_start_thread (pool);
485	root	1.1
486	root	1.6	X_LOCK (pool->reslock);
487	root	1.5	req = reqq_shift (&pool->res_queue);
488	root	1.1
489			if (req)
490			{
491	root	1.6	--pool->npending;
492	root	1.1
493	root	1.6	if (!pool->res_queue.size)
494			ETP_DONE_POLL (pool->userdata);
495	root	1.1	}
496
497	root	1.6	X_UNLOCK (pool->reslock);
498	root	1.1
499			if (!req)
500			return 0;
501
502	root	1.6	X_LOCK (pool->reqlock);
503			--pool->nreqs;
504			X_UNLOCK (pool->reqlock);
505	root	1.1
506			if (ecb_expect_false (req->type == ETP_TYPE_GROUP && req->size))
507			{
508	root	1.4	req->flags \|= ETP_FLAG_DELAYED; /* mark request as delayed */
509	root	1.1	continue;
510			}
511			else
512			{
513			int res = ETP_FINISH (req);
514			if (ecb_expect_false (res))
515			return res;
516			}
517
518			if (ecb_expect_false (maxreqs && !--maxreqs))
519			break;
520
521			if (maxtime)
522			{
523			gettimeofday (&tv_now, 0);
524
525			if (etp_tvdiff (&tv_start, &tv_now) >= maxtime)
526			break;
527			}
528			}
529
530			errno = EAGAIN;
531			return -1;
532			}
533
534			ETP_API_DECL void
535	root	1.5	etp_grp_cancel (etp_pool pool, ETP_REQ *grp);
536	root	1.1
537			ETP_API_DECL void
538	root	1.5	etp_cancel (etp_pool pool, ETP_REQ *req)
539	root	1.1	{
540			req->cancelled = 1;
541
542	root	1.5	etp_grp_cancel (pool, req);
543	root	1.1	}
544
545			ETP_API_DECL void
546	root	1.5	etp_grp_cancel (etp_pool pool, ETP_REQ *grp)
547	root	1.1	{
548			for (grp = grp->grp_first; grp; grp = grp->grp_next)
549	root	1.5	etp_cancel (pool, grp);
550	root	1.1	}
551
552			ETP_API_DECL void
553	root	1.5	etp_submit (etp_pool pool, ETP_REQ *req)
554	root	1.1	{
555			req->pri -= ETP_PRI_MIN;
556
557			if (ecb_expect_false (req->pri < ETP_PRI_MIN - ETP_PRI_MIN)) req->pri = ETP_PRI_MIN - ETP_PRI_MIN;
558			if (ecb_expect_false (req->pri > ETP_PRI_MAX - ETP_PRI_MIN)) req->pri = ETP_PRI_MAX - ETP_PRI_MIN;
559
560			if (ecb_expect_false (req->type == ETP_TYPE_GROUP))
561			{
562			/* I hope this is worth it :/ */
563	root	1.6	X_LOCK (pool->reqlock);
564			++pool->nreqs;
565			X_UNLOCK (pool->reqlock);
566	root	1.1
567	root	1.6	X_LOCK (pool->reslock);
568	root	1.1
569	root	1.6	++pool->npending;
570	root	1.1
571	root	1.6	if (!reqq_push (&pool->res_queue, req))
572			ETP_WANT_POLL (pool);
573	root	1.1
574	root	1.6	X_UNLOCK (pool->reslock);
575	root	1.1	}
576			else
577			{
578	root	1.6	X_LOCK (pool->reqlock);
579			++pool->nreqs;
580			++pool->nready;
581	root	1.5	reqq_push (&pool->req_queue, req);
582	root	1.6	X_COND_SIGNAL (pool->reqwait);
583			X_UNLOCK (pool->reqlock);
584	root	1.1
585	root	1.5	etp_maybe_start_thread (pool);
586	root	1.1	}
587			}
588
589			ETP_API_DECL void ecb_cold
590	root	1.5	etp_set_max_poll_time (etp_pool pool, double nseconds)
591	root	1.1	{
592	root	1.6	if (WORDACCESS_UNSAFE) X_LOCK (pool->reslock);
593			pool->max_poll_time = nseconds * ETP_TICKS;
594			if (WORDACCESS_UNSAFE) X_UNLOCK (pool->reslock);
595	root	1.1	}
596
597			ETP_API_DECL void ecb_cold
598	root	1.5	etp_set_max_poll_reqs (etp_pool pool, unsigned int maxreqs)
599	root	1.1	{
600	root	1.6	if (WORDACCESS_UNSAFE) X_LOCK (pool->reslock);
601			pool->max_poll_reqs = maxreqs;
602			if (WORDACCESS_UNSAFE) X_UNLOCK (pool->reslock);
603	root	1.1	}
604
605			ETP_API_DECL void ecb_cold
606	root	1.5	etp_set_max_idle (etp_pool pool, unsigned int nthreads)
607	root	1.1	{
608	root	1.6	if (WORDACCESS_UNSAFE) X_LOCK (pool->reqlock);
609			pool->max_idle = nthreads;
610			if (WORDACCESS_UNSAFE) X_UNLOCK (pool->reqlock);
611	root	1.1	}
612
613			ETP_API_DECL void ecb_cold
614	root	1.5	etp_set_idle_timeout (etp_pool pool, unsigned int seconds)
615	root	1.1	{
616	root	1.6	if (WORDACCESS_UNSAFE) X_LOCK (pool->reqlock);
617			pool->idle_timeout = seconds;
618			if (WORDACCESS_UNSAFE) X_UNLOCK (pool->reqlock);
619	root	1.1	}
620
621			ETP_API_DECL void ecb_cold
622	root	1.5	etp_set_min_parallel (etp_pool pool, unsigned int nthreads)
623	root	1.1	{
624	root	1.6	if (pool->wanted < nthreads)
625			pool->wanted = nthreads;
626	root	1.1	}
627
628			ETP_API_DECL void ecb_cold
629	root	1.5	etp_set_max_parallel (etp_pool pool, unsigned int nthreads)
630	root	1.1	{
631	root	1.6	if (pool->wanted > nthreads)
632			pool->wanted = nthreads;
633	root	1.1
634	root	1.6	while (pool->started > pool->wanted)
635	root	1.5	etp_end_thread (pool);
636	root	1.1	}
637