Initiail implementation of lio_listio() and rethinking of signal logic

include/aio.h

@@ -134,8 +134,8 @@ struct aiocb
   off_t aio_offset;              /* File offset */
   size_t aio_nbytes;             /* Length of transfer */
   int aio_fildes;                /* File descriptor */
-  int aio_reqprio;               /* Request priority offset */
-  int aio_lio_opcode;            /* Operation to be performed */
+  int8_t aio_reqprio;            /* Request priority offset */
+  uint8_t aio_lio_opcode;        /* Operation to be performed */
 
   /* Non-standard, implementation-dependent data.  For portability reasons,
    * application code should never reference these elements.
@@ -144,6 +144,7 @@ struct aiocb
   struct work_s aio_work;        /* Used to defer I/O to the work thread */
   pid_t aio_pid;                 /* ID of client to be notify at completion */
   volatile ssize_t aio_result;   /* Support for aio_error() and aio_return() */
+  FAR void *aio_priv;            /* Used by signal handlers */
 };
 
 /****************************************************************************
@@ -170,11 +171,8 @@ ssize_t aio_return(FAR struct aiocb *aiocbp);
 int aio_suspend(FAR const struct aiocb *const list[], int nent,
                 FAR const struct timespec *timeout);
 int aio_write(FAR struct aiocb *aiocbp);
-
-#ifndef CONFIG_PTHREAD_DISABLE /* Depends on pthread support */
 int lio_listio(int mode, FAR struct aiocb *const list[], int nent,
                FAR struct sigevent *sig);
-#endif
 
 #undef EXTERN
 #ifdef __cplusplus

libc/aio/Make.defs

@@ -38,11 +38,7 @@ ifeq ($(CONFIG_LIBC_AIO),y)
 # Add the asynchronous I/O C files to the build
 
 CSRCS += aio_cancel.c aio_error.c aio_fsync.c aio_read.c aio_return.c
-CSRCS += aio_signal.c aio_suspend.c aio_write.c
-
-ifneq ($(CONFIG_PTHREAD_DISABLE),y)
-CSRCS += lio_listio.c
-endif
+CSRCS += aio_signal.c aio_suspend.c aio_write.c lio_listio.c
 
 # Add the asynchronous I/O directory to the build
 

libc/aio/aio_fsync.c

@@ -0,0 +1,227 @@
+/****************************************************************************
+ * libc/aio/aio_fsync.c
+ *
+ *   Copyright (C) 2014 Gregory Nutt. All rights reserved.
+ *   Author: Gregory Nutt <gnutt@nuttx.org>
+ *
+ * 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. Neither the name NuttX 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 BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "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
+ * COPYRIGHT OWNER 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.
+ *
+ ****************************************************************************/
+
+/****************************************************************************
+ * Included Files
+ ****************************************************************************/
+
+#include <nuttx/config.h>
+
+#include <unistd.h>
+#include <aio.h>
+#include <assert.h>
+#include <errno.h>
+
+#include <nuttx/wqueue.h>
+
+#include "lib_internal.h"
+#include "aio/aio.h"
+
+#ifndef CONFIG_LIBC_AIO
+
+/****************************************************************************
+ * Pre-processor Definitions
+ ****************************************************************************/
+/* Configuration ************************************************************/
+
+/****************************************************************************
+ * Private Types
+ ****************************************************************************/
+
+/****************************************************************************
+ * Private Variables
+ ****************************************************************************/
+
+/****************************************************************************
+ * Public Variables
+ ****************************************************************************/
+
+/****************************************************************************
+ * Private Functions
+ ****************************************************************************/
+
+/****************************************************************************
+ * Name: aio_fsync_worker
+ *
+ * Description:
+ *   This function executes on the worker thread and performs the
+ *   asynchronous I/O operation.
+ *
+ * Input Parameters:
+ *   arg - Worker argument.  In this case, a pointer to an instance of
+ *     struct aiocb cast to void *.
+ *
+ * Returned Value:
+ *   None
+ *
+ ****************************************************************************/
+
+static void aio_fsync_worker(FAR void *arg)
+{
+  FAR struct aiocb *aiocbp = (FAR struct aiocb *)arg;
+  DEBASSERT(arg);
+  int ret;
+
+  /* Perform the fsync using aio_fildes */
+
+  ret = fsync(aiocbp->aio_fildes);
+  if (ret < 0)
+    {
+      int errcode = get_errno();
+      fdbg("ERROR: fsync failed: %d\n", errode);
+      DEBUGASSERT(errcode > 0);
+      aicbp->result = -errcode;
+    }
+  else
+    {
+      aicbp->result = OK;
+    }
+
+  /* Signal the client */
+
+  (void)aio_signal(aiocbp);
+}
+
+/****************************************************************************
+ * Public Functions
+ ****************************************************************************/
+
+/****************************************************************************
+ * Name: aio_fsync
+ *
+ * Description:
+ *   The aio_fsync() function asynchronously forces all I/O operations
+ *   associated with the file indicated by the file descriptor aio_fildes
+ *   member of the aiocb structure referenced by the aiocbp argument and
+ *   queued at the time of the call to aio_fsync() to the synchronized
+ *   I/O completion state. The function call will return when the
+ *   synchronization request has been initiated or queued to the file or
+ *   device (even when the data cannot be synchronized immediately).
+ *
+ *   If op is O_DSYNC, all currently queued I/O operations will be
+ *   completed as if by a call to fdatasync(); that is, as defined for
+ *   synchronized I/O data integrity completion. If op is O_SYNC, all
+ *   currently queued I/O operations will be completed as if by a call to
+ *   fsync(); that is, as defined for synchronized I/O file integrity
+ *   completion. If the aio_fsync() function fails, or if the operation
+ *   queued by aio_fsync() fails, then, as for fsync() and fdatasync(),
+ *   outstanding I/O operations are not guaranteed to have been completed.
+ *   [See "POSIX Compliance" below]
+ *
+ *   If aio_fsync() succeeds, then it is only the I/O that was queued at
+ *   the time of the call to aio_fsync() that is guaranteed to be forced
+ *   to the relevant completion state. The completion of subsequent I/O
+ *   on the file descriptor is not guaranteed to be completed in a
+ *   synchronized fashion.
+ *
+ *   The aiocbp argument refers to an asynchronous I/O control block. The
+ *   aiocbp value may be used as an argument to aio_error() and aio_return()
+ *   in order to determine the error status and return status, respectively,
+ *   of the asynchronous operation while it is proceeding. When the request
+ *   is queued, the error status for the operation is [EINPROGRESS]. When
+ *   all data has been successfully transferred, the error status will be
+ *   reset to reflect the success or failure of the operation. If the
+ *   operation does not complete successfully, the error status for the
+ *   operation will be set to indicate the error. The aio_sigevent member
+ *   determines the asynchronous notification to occur when all operations
+ *   have achieved synchronized I/O completion. All other members of the
+ *   structure referenced by aiocbp are ignored. If the control block
+ *   referenced by aiocbp becomes an illegal address prior to asynchronous
+ *   I/O completion, then the behavior is undefined.
+ *
+ *   If the aio_fsync() function fails or aiocbp indicates an error condition,
+ *   data is not guaranteed to have been successfully transferred.
+ *
+ * Input Parameters:
+ *   op     - Should be either O_SYNC or O_DSYNC.  Ignored in this implementation.
+ *   aiocbp - A pointer to an instance of struct aiocb
+ *
+ * Returned Value:
+ *   The aio_fsync() function will return the value 0 if the I/O operation is 
+ *   successfully queued; otherwise, the function will return the value -1 and
+ *   set errno to indicate the error.
+ *
+ *   The aio_fsync() function will fail if:
+ *
+ *     EAGAIN - The requested asynchronous operation was not queued due to
+ *       temporary resource limitations.
+ *     EBADF - The aio_fildes member of the aiocb structure referenced by
+ *       the aiocbp argument is not a valid file descriptor open for writing.
+ *     EINVAL - This implementation does not support synchronized I/O for
+ *       this file.
+ *      EINVAL - A value of op other than O_DSYNC or O_SYNC was specified.
+ *
+ *   In the event that any of the queued I/O operations fail, aio_fsync()
+ *   will return the error condition defined for read() and write(). The
+ *   error is returned in the error status for the asynchronous fsync()
+ *   operation, which can be retrieved using aio_error().
+ *
+ * POSIX Compliance
+ * - NuttX does not currently make any distinction between O_DYSNC and O_SYNC.
+ *   Hence, the 'op' argument is ignored altogether.
+ * - Most errors required in the standard are not detected at this point.
+ *   There are no pre-queuing checks for the validity of the operation.
+ *
+ ****************************************************************************/
+
+int aio_fsync(int op, FAR struct aiocb *aiocbp)
+{
+  int ret;
+
+  DEBUGASSERT(op == O_SYNC /* || op == O_DSYNC */);
+  DEBUGASSERT(aiocbp);
+
+  /* The result -EINPROGRESS means that the transfer has not yet completed */
+
+  aiocbp->aio_result = -EINPROGRESS;
+
+  /* Save the ID of the calling, client thread */
+
+  aiocbp->aio_pid = getpid();
+
+  /* Defer the work to the worker thread */
+
+  ret = work_queue(AIO_QUEUE, &aiocbp->aio_work, aio_fsync_worker, aiocbp, 0);
+  if (ret < 0)
+    {
+      aio->aio_result = ret;
+      set_errno(ret);
+      return ERROR;
+    }
+
+  return OK;
+}
+
+#endif /* CONFIG_LIBC_AIO */

libc/aio/aio_signal.c

@@ -92,33 +92,54 @@
 
 int aio_signal(FAR struct aiocb *aiocbp)
 {
+  int errcode;
+  int status;
   int ret;
 
   DEBUGASSERT(aiocbp);
  
+  ret = OK; /* Assume success */
+
   /* Signal the client */
 
   if (aiocbp->aio_sigevent.sigev_notify == SIGEV_SIGNAL)
     {
 #ifdef CONFIG_CAN_PASS_STRUCTS
-      ret = sigqueue(aiocbp->aio_pid, aiocbp->aio_sigevent.sigev_signo,
-                     aiocbp->aio_sigevent.sigev_value);
+      status = sigqueue(aiocbp->aio_pid, aiocbp->aio_sigevent.sigev_signo,
+                        aiocbp->aio_sigevent.sigev_value);
 #else
-      ret = sigqueue(aiocbp->aio_pid, aiocbp->aio_sigevent.sigev_sign,
-                     aiocbp->aio_sigevent.sigev_value.sival_ptr);
+      status = sigqueue(aiocbp->aio_pid, aiocbp->aio_sigevent.sigev_sign,
+                        aiocbp->aio_sigevent.sigev_value.sival_ptr);
 #endif
+      if (ret < 0)
+        {
+          errcode = get_errno();
+          fdbg("ERROR: sigqueue failed: %d\n", errcode);
+          ret = ERROR;
+        }
     }
 
   /* Send the poll signal in any event in case the caller is waiting
    * on sig_suspend();
    */
 
-  else
+  status = kill(aiocbp->aio_pid, SIGPOLL);
+  if (status && ret == OK)
     {
-      ret = kill(aiocbp->aio_pid, SIGPOLL);
+      errcode = get_errno();
+      fdbg("ERROR: kill failed: %d\n", errcode);
+      ret = ERROR;
     }
 
-  return ret;
+  /* Make sure that errno is set correctly on return */
+
+  if (ret < 0)
+    {
+      set_errno(errcode);
+      return ERROR;
+    }
+
+  return OK;
 }
 
 #endif /* CONFIG_LIBC_AIO */

libc/aio/aio_suspend.c

@@ -125,11 +125,6 @@ int aio_suspend(FAR const struct aiocb *const list[], int nent,
 
   DEBUGASSERT(aiocbp);
 
-  /* Create the signal set */
-
-  sigemptyset(&set);
-  sigaddset(&set, SIGPOLL);
-
   /* Lock the scheduler so that no I/O events can complete on the worker
    * thread until we set our wait set up.  Pre-emption will, of course, be
    * re-enabled while we are waiting for the signal.
@@ -152,21 +147,19 @@ int aio_suspend(FAR const struct aiocb *const list[], int nent,
           sched_unlock();
           return OK;
         }
-
-      /* If the caller has a signal notification setup, add that to the
-       * signals we will wait for.
-       */
-
-      if (aiocbp->aio_sigevent.sigev_notify == SIGEV_SIGNAL)
-        {
-          sigaddset(&set, aiocbp->aio_sigevent.sigev_signo);
-        }
     }
 
-  /* Then wait.  sigtimedwait() will return the signal number that cause the
-   * error.  It will set errno appropriately for this function on errors.
+  /* Then wait for SIGPOLL.  On success sigtimedwait() will return the
+   * signal number that cause the error (SIGPOLL).  It will set errno
+   * appropriately for this function on errors.
+   *
+   * NOTE: If completion of the I/O causes other signals to be generated
+   * first, then this will wake up and return EINTR instead of success.
    */
 
+  sigemptyset(&set);
+  sigaddset(&set, SIGPOLL);
+
   ret = sigtimedwait(&set, NULL, &timeout)
   sched_unlock();
   return ret >= 0 ? OK : ERROR;