- Use an arbitrary but reasonably large import size for kva on architectures

that don't support superpages.  This keeps the number of spans and internal
   fragmentation lower.
 - When the user asks for alignment from vmem_xalloc adjust the imported size
   by 2*align to be certain we can satisfy the allocation.  This comes at
   the expense of potential failures when the backend can't supply enough
   memory but could supply the requested size and alignment.

Sponsored by:	EMC / Isilon Storage Division

sys/kern/subr_vmem.c

@@ -758,6 +758,7 @@ vmem_add1(vmem_t *vm, vmem_addr_t addr, vmem_size_t size, int type)
 	bt_t *btfree;
 
 	MPASS(type == BT_TYPE_SPAN || type == BT_TYPE_SPAN_STATIC);
+	MPASS((size & vm->vm_quantum_mask) == 0);
 
 	btspan = bt_alloc(vm);
 	btspan->bt_type = type;
@@ -805,7 +806,7 @@ vmem_destroy1(vmem_t *vm)
 }
 
 static int
-vmem_import(vmem_t *vm, vmem_size_t size, int flags)
+vmem_import(vmem_t *vm, vmem_size_t size, vmem_size_t align, int flags)
 {
 	vmem_addr_t addr;
 	int error;
@@ -813,6 +814,12 @@ vmem_import(vmem_t *vm, vmem_size_t size, int flags)
 	if (vm->vm_importfn == NULL)
 		return EINVAL;
 
+	/*
+	 * To make sure we get a span that meets the alignment we double it
+	 * and add the size to the tail.  This slightly overestimates.
+	 */
+	if (align != vm->vm_quantum_mask + 1)
+		size = (align * 2) + size;
 	size = roundup(size, vm->vm_import_quantum);
 
 	/*
@@ -1157,7 +1164,7 @@ vmem_xalloc(vmem_t *vm, const vmem_size_t size0, vmem_size_t align,
 		 * imported region.  It is up to the user to specify the
 		 * import quantum such that it can satisfy any allocation.
 		 */
-		if (vmem_import(vm, size, flags) == 0)
+		if (vmem_import(vm, size, align, flags) == 0)
 			continue;
 
 		/*

sys/vm/vm_init.c

@@ -156,7 +156,8 @@ vm_mem_init(dummy)
 #if VM_NRESERVLEVEL > 0
 	    1 << (VM_LEVEL_0_ORDER + PAGE_SHIFT));
 #else
-	    PAGE_SIZE);
+	    /* On non-superpage architectures want large import sizes. */
+	    PAGE_SIZE * 1024);
 #endif
 
 	kmem_init_zero_region();