1 files changed, 93 insertions, 3 deletions

diff --git a/nuttx/mm/shm/shmat.c b/nuttx/mm/shm/shmat.c
index a76327e13..3a9bc4c08 100755..100644
--- a/nuttx/mm/shm/shmat.c
+++ b/nuttx/mm/shm/shmat.c
@@ -42,6 +42,12 @@
 #include <sys/shm.h>
 #include <errno.h>
 
+#include <nuttx/sched.h>
+#include <nuttx/arch.h>
+#include <nuttx/pgalloc.h>
+
+#include "shm/shm.h"
+
 #ifdef CONFIG_MM_SHM
 
 /****************************************************************************
@@ -94,7 +100,8 @@
  * Returned Value:
  *   Upon successful completion, shmat() will increment the value of
  *   shm_nattch in the data structure associated with the shared memory ID
- *   of the attached shared memory segment and return the segment's start address.
+ *   of the attached shared memory segment and return the segment's start
+ *   address.
  *
  *   Otherwise, the shared memory segment will not be attached, shmat() will
  *   return -1, and errno will be set to indicate the error.
@@ -119,8 +126,91 @@
 
 FAR void *shmat(int shmid, FAR const void *shmaddr, int shmflg)
 {
-#warning Not implemented
-  set_errno(ENOSYS);
+  FAR struct shm_region_s *region;
+  FAR struct task_group_s *group;
+  FAR struct tcb_s *tcb;
+  uintptr_t vaddr;
+  unsigned int npages;
+  int ret;
+
+  /* Get the region associated with the shmid */
+
+  DEBUGASSERT(shmid >= 0 && shmid < CONFIG_ARCH_SHM_MAXREGIONS);
+  region =  &g_shminfo.si_region[shmid];
+  DEBUGASSERT((region->sr_flags & SRFLAG_INUSE) != 0);
+
+  /* Get the TCB and group containing our virtual memory allocator */
+
+  tcb = sched_self();
+  DEBUGASSERT(tcb && tcb->group);
+  group = tcb->group;
+  DEBUGASSERT(group->tg_shm.gs_handle != NULL &&
+              group->tg_shm.gs_vaddr[shmid] == 0);
+
+  /* Get exclusive access to the region data structure */
+
+  ret = sem_wait(&region->sr_sem);
+  if (ret < 0)
+    {
+      shmdbg("sem_wait failed: %d\n", ret);
+      goto errout;
+    }
+
+  /* Set aside a virtual address space to span this physical region */
+
+  vaddr = (uintptr_t)gran_alloc(group->tg_shm.gs_handle,
+                                region->sr_ds.shm_segsz);
+  if (vaddr == 0)
+    {
+      shmdbg("gran_alloc() failed\n");
+      ret = -ENOMEM;
+      goto errout_with_semaphore;
+    }
+
+  /* Convert the region size to pages */
+
+  npages = MM_PGALIGNUP(region->sr_ds.shm_segsz);
+
+  /* Attach, i.e, map, on shared memory region to the user virtual address. */
+
+  ret = up_shmat(region->sr_pages, npages, vaddr);
+  if (ret < 0)
+    {
+      shmdbg("up_shmat() failed\n");
+      goto errout_with_vaddr;
+    }
+
+  /* Save the virtual address of the region.  We will need that in shmat()
+   * to do the reverse lookup:  Give the virtual address of the region to
+   * detach, we need to get the region table index.
+   */
+
+  group->tg_shm.gs_vaddr[shmid] = vaddr;
+
+  /* Increment the count of processes attached to this region */
+
+  region->sr_ds.shm_nattch++;
+
+  /* Save the process ID of the the last operation */
+
+  region->sr_ds.shm_lpid = tcb->pid;
+
+  /* Save the time of the last shmat() */
+
+  region->sr_ds.shm_atime = time(NULL);
+
+  /* Release our lock on the entry */
+
+  sem_post(&region->sr_sem);
+  return (FAR void *)vaddr;
+
+errout_with_vaddr:
+  gran_free(group->tg_shm.gs_handle, (FAR void *)vaddr,
+            region->sr_ds.shm_segsz);
+errout_with_semaphore:
+  sem_post(&region->sr_sem);
+  set_errno(-ret);
+errout:
   return (FAR void *)ERROR;
 }