drm: add simple DRM memory manager, and hash table
This adds the DRM hashtable and simple memory manager implementations from
Tungsten Graphics, this is NOT the new memory manager, this is a replacement
for the SIS and VIA memory managers.
Signed-off-by: Dave Airlie <airlied@linux.ie>
diff --git a/drivers/char/drm/drm_mm.c b/drivers/char/drm/drm_mm.c
new file mode 100644
index 0000000..c55ed45
--- /dev/null
+++ b/drivers/char/drm/drm_mm.c
@@ -0,0 +1,201 @@
+/**************************************************************************
+ *
+ * Copyright 2006 Tungsten Graphics, Inc., Bismarck, ND., USA.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ *
+ **************************************************************************/
+
+/*
+ * Generic simple memory manager implementation. Intended to be used as a base
+ * class implementation for more advanced memory managers.
+ *
+ * Note that the algorithm used is quite simple and there might be substantial
+ * performance gains if a smarter free list is implemented. Currently it is just an
+ * unordered stack of free regions. This could easily be improved if an RB-tree
+ * is used instead. At least if we expect heavy fragmentation.
+ *
+ * Aligned allocations can also see improvement.
+ *
+ * Authors:
+ * Thomas Hellström <thomas-at-tungstengraphics-dot-com>
+ */
+
+#include "drmP.h"
+
+drm_mm_node_t *drm_mm_get_block(drm_mm_node_t * parent,
+ unsigned long size, unsigned alignment)
+{
+
+ drm_mm_node_t *child;
+
+ if (alignment)
+ size += alignment - 1;
+
+ if (parent->size == size) {
+ list_del_init(&parent->fl_entry);
+ parent->free = FALSE;
+ return parent;
+ } else {
+ child = (drm_mm_node_t *) drm_alloc(sizeof(*child), DRM_MEM_MM);
+ if (!child)
+ return NULL;
+
+ INIT_LIST_HEAD(&child->ml_entry);
+ INIT_LIST_HEAD(&child->fl_entry);
+
+ child->free = FALSE;
+ child->size = size;
+ child->start = parent->start;
+
+ list_add_tail(&child->ml_entry, &parent->ml_entry);
+ parent->size -= size;
+ parent->start += size;
+ }
+ return child;
+}
+
+/*
+ * Put a block. Merge with the previous and / or next block if they are free.
+ * Otherwise add to the free stack.
+ */
+
+void drm_mm_put_block(drm_mm_t * mm, drm_mm_node_t * cur)
+{
+
+ drm_mm_node_t *list_root = &mm->root_node;
+ struct list_head *cur_head = &cur->ml_entry;
+ struct list_head *root_head = &list_root->ml_entry;
+ drm_mm_node_t *prev_node = NULL;
+ drm_mm_node_t *next_node;
+
+ int merged = FALSE;
+
+ if (cur_head->prev != root_head) {
+ prev_node = list_entry(cur_head->prev, drm_mm_node_t, ml_entry);
+ if (prev_node->free) {
+ prev_node->size += cur->size;
+ merged = TRUE;
+ }
+ }
+ if (cur_head->next != root_head) {
+ next_node = list_entry(cur_head->next, drm_mm_node_t, ml_entry);
+ if (next_node->free) {
+ if (merged) {
+ prev_node->size += next_node->size;
+ list_del(&next_node->ml_entry);
+ list_del(&next_node->fl_entry);
+ drm_free(next_node, sizeof(*next_node),
+ DRM_MEM_MM);
+ } else {
+ next_node->size += cur->size;
+ next_node->start = cur->start;
+ merged = TRUE;
+ }
+ }
+ }
+ if (!merged) {
+ cur->free = TRUE;
+ list_add(&cur->fl_entry, &list_root->fl_entry);
+ } else {
+ list_del(&cur->ml_entry);
+ drm_free(cur, sizeof(*cur), DRM_MEM_MM);
+ }
+}
+
+drm_mm_node_t *drm_mm_search_free(const drm_mm_t * mm,
+ unsigned long size,
+ unsigned alignment, int best_match)
+{
+ struct list_head *list;
+ const struct list_head *free_stack = &mm->root_node.fl_entry;
+ drm_mm_node_t *entry;
+ drm_mm_node_t *best;
+ unsigned long best_size;
+
+ best = NULL;
+ best_size = ~0UL;
+
+ if (alignment)
+ size += alignment - 1;
+
+ list_for_each(list, free_stack) {
+ entry = list_entry(list, drm_mm_node_t, fl_entry);
+ if (entry->size >= size) {
+ if (!best_match)
+ return entry;
+ if (size < best_size) {
+ best = entry;
+ best_size = entry->size;
+ }
+ }
+ }
+
+ return best;
+}
+
+int drm_mm_init(drm_mm_t * mm, unsigned long start, unsigned long size)
+{
+ drm_mm_node_t *child;
+
+ INIT_LIST_HEAD(&mm->root_node.ml_entry);
+ INIT_LIST_HEAD(&mm->root_node.fl_entry);
+ child = (drm_mm_node_t *) drm_alloc(sizeof(*child), DRM_MEM_MM);
+ if (!child)
+ return -ENOMEM;
+
+ INIT_LIST_HEAD(&child->ml_entry);
+ INIT_LIST_HEAD(&child->fl_entry);
+
+ child->start = start;
+ child->size = size;
+ child->free = TRUE;
+
+ list_add(&child->fl_entry, &mm->root_node.fl_entry);
+ list_add(&child->ml_entry, &mm->root_node.ml_entry);
+
+ return 0;
+}
+
+EXPORT_SYMBOL(drm_mm_init);
+
+void drm_mm_takedown(drm_mm_t * mm)
+{
+ struct list_head *bnode = mm->root_node.fl_entry.next;
+ drm_mm_node_t *entry;
+
+ entry = list_entry(bnode, drm_mm_node_t, fl_entry);
+
+ if (entry->ml_entry.next != &mm->root_node.ml_entry ||
+ entry->fl_entry.next != &mm->root_node.fl_entry) {
+ DRM_ERROR("Memory manager not clean. Delaying takedown\n");
+ return;
+ }
+
+ list_del(&entry->fl_entry);
+ list_del(&entry->ml_entry);
+
+ drm_free(entry, sizeof(*entry), DRM_MEM_MM);
+}
+
+EXPORT_SYMBOL(drm_mm_takedown);