Bug Summary

File:dev/pci/drm/drm_mm.c
Warning:line 278, column 4
Value stored to 'first' is never read

Annotated Source Code

Press '?' to see keyboard shortcuts

clang -cc1 -cc1 -triple amd64-unknown-openbsd7.4 -analyze -disable-free -clear-ast-before-backend -disable-llvm-verifier -discard-value-names -main-file-name drm_mm.c -analyzer-checker=core -analyzer-checker=apiModeling -analyzer-checker=unix -analyzer-checker=deadcode -analyzer-checker=security.insecureAPI.UncheckedReturn -analyzer-checker=security.insecureAPI.getpw -analyzer-checker=security.insecureAPI.gets -analyzer-checker=security.insecureAPI.mktemp -analyzer-checker=security.insecureAPI.mkstemp -analyzer-checker=security.insecureAPI.vfork -analyzer-checker=nullability.NullPassedToNonnull -analyzer-checker=nullability.NullReturnedFromNonnull -analyzer-output plist -w -setup-static-analyzer -mrelocation-model static -mframe-pointer=all -relaxed-aliasing -ffp-contract=on -fno-rounding-math -mconstructor-aliases -ffreestanding -mcmodel=kernel -target-cpu x86-64 -target-feature +retpoline-indirect-calls -target-feature +retpoline-indirect-branches -target-feature -sse2 -target-feature -sse -target-feature -3dnow -target-feature -mmx -target-feature +save-args -target-feature +retpoline-external-thunk -disable-red-zone -no-implicit-float -tune-cpu generic -debugger-tuning=gdb -fcoverage-compilation-dir=/usr/src/sys/arch/amd64/compile/GENERIC.MP/obj -nostdsysteminc -nobuiltininc -resource-dir /usr/local/llvm16/lib/clang/16 -I /usr/src/sys -I /usr/src/sys/arch/amd64/compile/GENERIC.MP/obj -I /usr/src/sys/arch -I /usr/src/sys/dev/pci/drm/include -I /usr/src/sys/dev/pci/drm/include/uapi -I /usr/src/sys/dev/pci/drm/amd/include/asic_reg -I /usr/src/sys/dev/pci/drm/amd/include -I /usr/src/sys/dev/pci/drm/amd/amdgpu -I /usr/src/sys/dev/pci/drm/amd/display -I /usr/src/sys/dev/pci/drm/amd/display/include -I /usr/src/sys/dev/pci/drm/amd/display/dc -I /usr/src/sys/dev/pci/drm/amd/display/amdgpu_dm -I /usr/src/sys/dev/pci/drm/amd/pm/inc -I /usr/src/sys/dev/pci/drm/amd/pm/legacy-dpm -I /usr/src/sys/dev/pci/drm/amd/pm/swsmu -I /usr/src/sys/dev/pci/drm/amd/pm/swsmu/inc -I /usr/src/sys/dev/pci/drm/amd/pm/swsmu/smu11 -I /usr/src/sys/dev/pci/drm/amd/pm/swsmu/smu12 -I /usr/src/sys/dev/pci/drm/amd/pm/swsmu/smu13 -I /usr/src/sys/dev/pci/drm/amd/pm/powerplay/inc -I /usr/src/sys/dev/pci/drm/amd/pm/powerplay/hwmgr -I /usr/src/sys/dev/pci/drm/amd/pm/powerplay/smumgr -I /usr/src/sys/dev/pci/drm/amd/pm/swsmu/inc -I /usr/src/sys/dev/pci/drm/amd/pm/swsmu/inc/pmfw_if -I /usr/src/sys/dev/pci/drm/amd/display/dc/inc -I /usr/src/sys/dev/pci/drm/amd/display/dc/inc/hw -I /usr/src/sys/dev/pci/drm/amd/display/dc/clk_mgr -I /usr/src/sys/dev/pci/drm/amd/display/modules/inc -I /usr/src/sys/dev/pci/drm/amd/display/modules/hdcp -I /usr/src/sys/dev/pci/drm/amd/display/dmub/inc -I /usr/src/sys/dev/pci/drm/i915 -D DDB -D DIAGNOSTIC -D KTRACE -D ACCOUNTING -D KMEMSTATS -D PTRACE -D POOL_DEBUG -D CRYPTO -D SYSVMSG -D SYSVSEM -D SYSVSHM -D UVM_SWAP_ENCRYPT -D FFS -D FFS2 -D FFS_SOFTUPDATES -D UFS_DIRHASH -D QUOTA -D EXT2FS -D MFS -D NFSCLIENT -D NFSSERVER -D CD9660 -D UDF -D MSDOSFS -D FIFO -D FUSE -D SOCKET_SPLICE -D TCP_ECN -D TCP_SIGNATURE -D INET6 -D IPSEC -D PPP_BSDCOMP -D PPP_DEFLATE -D PIPEX -D MROUTING -D MPLS -D BOOT_CONFIG -D USER_PCICONF -D APERTURE -D MTRR -D NTFS -D SUSPEND -D HIBERNATE -D PCIVERBOSE -D USBVERBOSE -D WSDISPLAY_COMPAT_USL -D WSDISPLAY_COMPAT_RAWKBD -D WSDISPLAY_DEFAULTSCREENS=6 -D X86EMU -D ONEWIREVERBOSE -D MULTIPROCESSOR -D MAXUSERS=80 -D _KERNEL -O2 -Wno-pointer-sign -Wno-address-of-packed-member -Wno-constant-conversion -Wno-unused-but-set-variable -Wno-gnu-folding-constant -fdebug-compilation-dir=/usr/src/sys/arch/amd64/compile/GENERIC.MP/obj -ferror-limit 19 -fwrapv -D_RET_PROTECTOR -ret-protector -fcf-protection=branch -fgnuc-version=4.2.1 -vectorize-loops -vectorize-slp -fno-builtin-malloc -fno-builtin-calloc -fno-builtin-realloc -fno-builtin-valloc -fno-builtin-free -fno-builtin-strdup -fno-builtin-strndup -analyzer-output=html -faddrsig -o /home/ben/Projects/scan/2024-01-11-110808-61670-1 -x c /usr/src/sys/dev/pci/drm/drm_mm.c
1/**************************************************************************
2 *
3 * Copyright 2006 Tungsten Graphics, Inc., Bismarck, ND., USA.
4 * Copyright 2016 Intel Corporation
5 * All Rights Reserved.
6 *
7 * Permission is hereby granted, free of charge, to any person obtaining a
8 * copy of this software and associated documentation files (the
9 * "Software"), to deal in the Software without restriction, including
10 * without limitation the rights to use, copy, modify, merge, publish,
11 * distribute, sub license, and/or sell copies of the Software, and to
12 * permit persons to whom the Software is furnished to do so, subject to
13 * the following conditions:
14 *
15 * The above copyright notice and this permission notice (including the
16 * next paragraph) shall be included in all copies or substantial portions
17 * of the Software.
18 *
19 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
20 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
21 * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
22 * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
23 * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
24 * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
25 * USE OR OTHER DEALINGS IN THE SOFTWARE.
26 *
27 *
28 **************************************************************************/
29
30/*
31 * Generic simple memory manager implementation. Intended to be used as a base
32 * class implementation for more advanced memory managers.
33 *
34 * Note that the algorithm used is quite simple and there might be substantial
35 * performance gains if a smarter free list is implemented. Currently it is
36 * just an unordered stack of free regions. This could easily be improved if
37 * an RB-tree is used instead. At least if we expect heavy fragmentation.
38 *
39 * Aligned allocations can also see improvement.
40 *
41 * Authors:
42 * Thomas Hellström <thomas-at-tungstengraphics-dot-com>
43 */
44
45#include <linux/export.h>
46#include <linux/interval_tree_generic.h>
47#include <linux/seq_file.h>
48#include <linux/slab.h>
49#include <linux/stacktrace.h>
50
51#include <drm/drm_mm.h>
52
53/**
54 * DOC: Overview
55 *
56 * drm_mm provides a simple range allocator. The drivers are free to use the
57 * resource allocator from the linux core if it suits them, the upside of drm_mm
58 * is that it's in the DRM core. Which means that it's easier to extend for
59 * some of the crazier special purpose needs of gpus.
60 *
61 * The main data struct is &drm_mm, allocations are tracked in &drm_mm_node.
62 * Drivers are free to embed either of them into their own suitable
63 * datastructures. drm_mm itself will not do any memory allocations of its own,
64 * so if drivers choose not to embed nodes they need to still allocate them
65 * themselves.
66 *
67 * The range allocator also supports reservation of preallocated blocks. This is
68 * useful for taking over initial mode setting configurations from the firmware,
69 * where an object needs to be created which exactly matches the firmware's
70 * scanout target. As long as the range is still free it can be inserted anytime
71 * after the allocator is initialized, which helps with avoiding looped
72 * dependencies in the driver load sequence.
73 *
74 * drm_mm maintains a stack of most recently freed holes, which of all
75 * simplistic datastructures seems to be a fairly decent approach to clustering
76 * allocations and avoiding too much fragmentation. This means free space
77 * searches are O(num_holes). Given that all the fancy features drm_mm supports
78 * something better would be fairly complex and since gfx thrashing is a fairly
79 * steep cliff not a real concern. Removing a node again is O(1).
80 *
81 * drm_mm supports a few features: Alignment and range restrictions can be
82 * supplied. Furthermore every &drm_mm_node has a color value (which is just an
83 * opaque unsigned long) which in conjunction with a driver callback can be used
84 * to implement sophisticated placement restrictions. The i915 DRM driver uses
85 * this to implement guard pages between incompatible caching domains in the
86 * graphics TT.
87 *
88 * Two behaviors are supported for searching and allocating: bottom-up and
89 * top-down. The default is bottom-up. Top-down allocation can be used if the
90 * memory area has different restrictions, or just to reduce fragmentation.
91 *
92 * Finally iteration helpers to walk all nodes and all holes are provided as are
93 * some basic allocator dumpers for debugging.
94 *
95 * Note that this range allocator is not thread-safe, drivers need to protect
96 * modifications with their own locking. The idea behind this is that for a full
97 * memory manager additional data needs to be protected anyway, hence internal
98 * locking would be fully redundant.
99 */
100
101#ifdef CONFIG_DRM_DEBUG_MM
102#include <linux/stackdepot.h>
103
104#define STACKDEPTH 32
105#define BUFSZ 4096
106
107static noinline__attribute__((__noinline__)) void save_stack(struct drm_mm_node *node)
108{
109 unsigned long entries[STACKDEPTH];
110 unsigned int n;
111
112 n = stack_trace_save(entries, ARRAY_SIZE(entries)(sizeof((entries)) / sizeof((entries)[0])), 1);
113
114 /* May be called under spinlock, so avoid sleeping */
115 node->stack = stack_depot_save(entries, n, GFP_NOWAIT0x0002);
116}
117
118static void show_leaks(struct drm_mm *mm)
119{
120 struct drm_mm_node *node;
121 unsigned long *entries;
122 unsigned int nr_entries;
123 char *buf;
124
125 buf = kmalloc(BUFSZ, GFP_KERNEL(0x0001 | 0x0004));
126 if (!buf)
127 return;
128
129 list_for_each_entry(node, drm_mm_nodes(mm), node_list)for (node = ({ const __typeof( ((__typeof(*node) *)0)->node_list
) *__mptr = (((&(mm)->head_node.node_list))->next)
; (__typeof(*node) *)( (char *)__mptr - __builtin_offsetof(__typeof
(*node), node_list) );}); &node->node_list != ((&(
mm)->head_node.node_list)); node = ({ const __typeof( ((__typeof
(*node) *)0)->node_list ) *__mptr = (node->node_list.next
); (__typeof(*node) *)( (char *)__mptr - __builtin_offsetof(__typeof
(*node), node_list) );})) {
130 if (!node->stack) {
131 DRM_ERROR("node [%08llx + %08llx]: unknown owner\n",__drm_err("node [%08llx + %08llx]: unknown owner\n", node->
start, node->size)
132 node->start, node->size)__drm_err("node [%08llx + %08llx]: unknown owner\n", node->
start, node->size);
133 continue;
134 }
135
136 nr_entries = stack_depot_fetch(node->stack, &entries);
137 stack_trace_snprint(buf, BUFSZ, entries, nr_entries, 0);
138 DRM_ERROR("node [%08llx + %08llx]: inserted at\n%s",__drm_err("node [%08llx + %08llx]: inserted at\n%s", node->
start, node->size, buf)
139 node->start, node->size, buf)__drm_err("node [%08llx + %08llx]: inserted at\n%s", node->
start, node->size, buf);
140 }
141
142 kfree(buf);
143}
144
145#undef STACKDEPTH
146#undef BUFSZ
147#else
148static void save_stack(struct drm_mm_node *node) { }
149static void show_leaks(struct drm_mm *mm) { }
150#endif
151
152#define START(node)((node)->start) ((node)->start)
153#define LAST(node)((node)->start + (node)->size - 1) ((node)->start + (node)->size - 1)
154
155#ifdef __linux__
156INTERVAL_TREE_DEFINE(struct drm_mm_node, rb,
157 u64, __subtree_last,
158 START, LAST, static inline, drm_mm_interval_tree)
159#else
160static struct drm_mm_node *
161drm_mm_interval_tree_iter_first(const struct rb_root_cached *root,
162 uint64_t start, uint64_t last)
163{
164 struct drm_mm_node *node;
165 struct rb_node *rb;
166
167 for (rb = rb_first_cached(root)linux_root_RB_MINMAX((struct linux_root *)(&(root)->rb_root
), -1); rb; rb = rb_next(rb)linux_root_RB_NEXT((rb))) {
168 node = rb_entry(rb, typeof(*node), rb)({ const __typeof( ((typeof(*node) *)0)->rb ) *__mptr = (rb
); (typeof(*node) *)( (char *)__mptr - __builtin_offsetof(typeof
(*node), rb) );});
169 if (LAST(node)((node)->start + (node)->size - 1) >= start && START(node)((node)->start) <= last)
170 return node;
171 }
172 return NULL((void *)0);
173}
174
175static void
176drm_mm_interval_tree_remove(struct drm_mm_node *node,
177 struct rb_root_cached *root)
178{
179 rb_erase_cached(&node->rb, root)linux_root_RB_REMOVE((struct linux_root *)(&(root)->rb_root
), (&node->rb));
180}
181#endif
182
183struct drm_mm_node *
184__drm_mm_interval_first(const struct drm_mm *mm, u64 start, u64 last)
185{
186 return drm_mm_interval_tree_iter_first((struct rb_root_cached *)&mm->interval_tree,
187 start, last) ?: (struct drm_mm_node *)&mm->head_node;
188}
189EXPORT_SYMBOL(__drm_mm_interval_first);
190
191static void drm_mm_interval_tree_add_node(struct drm_mm_node *hole_node,
192 struct drm_mm_node *node)
193{
194 struct drm_mm *mm = hole_node->mm;
195 struct rb_node **link, *rb;
196 struct drm_mm_node *parent;
197 bool_Bool leftmost;
198
199 node->__subtree_last = LAST(node)((node)->start + (node)->size - 1);
200
201 if (drm_mm_node_allocated(hole_node)) {
202 rb = &hole_node->rb;
203 while (rb) {
204 parent = rb_entry(rb, struct drm_mm_node, rb)({ const __typeof( ((struct drm_mm_node *)0)->rb ) *__mptr
= (rb); (struct drm_mm_node *)( (char *)__mptr - __builtin_offsetof
(struct drm_mm_node, rb) );});
205 if (parent->__subtree_last >= node->__subtree_last)
206 break;
207
208 parent->__subtree_last = node->__subtree_last;
209 rb = rb_parent(rb)(rb)->__entry.rbe_parent;
210 }
211
212 rb = &hole_node->rb;
213 link = &hole_node->rb.rb_right__entry.rbe_right;
214 leftmost = false0;
215 } else {
216 rb = NULL((void *)0);
217 link = &mm->interval_tree.rb_root.rb_node;
218 leftmost = true1;
219 }
220
221 while (*link) {
222 rb = *link;
223 parent = rb_entry(rb, struct drm_mm_node, rb)({ const __typeof( ((struct drm_mm_node *)0)->rb ) *__mptr
= (rb); (struct drm_mm_node *)( (char *)__mptr - __builtin_offsetof
(struct drm_mm_node, rb) );});
224 if (parent->__subtree_last < node->__subtree_last)
225 parent->__subtree_last = node->__subtree_last;
226 if (node->start < parent->start) {
227 link = &parent->rb.rb_left__entry.rbe_left;
228 } else {
229 link = &parent->rb.rb_right__entry.rbe_right;
230 leftmost = false0;
231 }
232 }
233
234 rb_link_node(&node->rb, rb, link);
235#ifdef notyet
236 rb_insert_augmented_cached(&node->rb, &mm->interval_tree, leftmost,
237 &drm_mm_interval_tree_augment);
238#else
239 rb_insert_color_cached(&node->rb, &mm->interval_tree, leftmost)linux_root_RB_INSERT_COLOR((struct linux_root *)(&(&mm
->interval_tree)->rb_root), (&node->rb));
240#endif
241}
242
243#define DRM_RB_INSERT(root, member, expr)do { struct rb_node **link = &root.rb_node, *rb = ((void *
)0); u64 x = expr(node); while (*link) { rb = *link; if (x <
expr(({ const __typeof( ((struct drm_mm_node *)0)->member
) *__mptr = (rb); (struct drm_mm_node *)( (char *)__mptr - __builtin_offsetof
(struct drm_mm_node, member) );}))) link = &rb->__entry
.rbe_left; else link = &rb->__entry.rbe_right; } rb_link_node
(&node->member, rb, link); linux_root_RB_INSERT_COLOR(
(struct linux_root *)(&root), (&node->member)); } while
(0) do { \
244 struct rb_node **link = &root.rb_node, *rb = NULL((void *)0); \
245 u64 x = expr(node); \
246 while (*link) { \
247 rb = *link; \
248 if (x < expr(rb_entry(rb, struct drm_mm_node, member)({ const __typeof( ((struct drm_mm_node *)0)->member ) *__mptr
= (rb); (struct drm_mm_node *)( (char *)__mptr - __builtin_offsetof
(struct drm_mm_node, member) );}))) \
249 link = &rb->rb_left__entry.rbe_left; \
250 else \
251 link = &rb->rb_right__entry.rbe_right; \
252 } \
253 rb_link_node(&node->member, rb, link); \
254 rb_insert_color(&node->member, &root)linux_root_RB_INSERT_COLOR((struct linux_root *)(&root), (
&node->member)); \
255} while (0)
256
257#define HOLE_SIZE(NODE)((NODE)->hole_size) ((NODE)->hole_size)
258#define HOLE_ADDR(NODE)(__drm_mm_hole_node_start(NODE)) (__drm_mm_hole_node_start(NODE))
259
260static u64 rb_to_hole_size(struct rb_node *rb)
261{
262 return rb_entry(rb, struct drm_mm_node, rb_hole_size)({ const __typeof( ((struct drm_mm_node *)0)->rb_hole_size
) *__mptr = (rb); (struct drm_mm_node *)( (char *)__mptr - __builtin_offsetof
(struct drm_mm_node, rb_hole_size) );})->hole_size;
263}
264
265static void insert_hole_size(struct rb_root_cached *root,
266 struct drm_mm_node *node)
267{
268 struct rb_node **link = &root->rb_root.rb_node, *rb = NULL((void *)0);
269 u64 x = node->hole_size;
270 bool_Bool first = true1;
271
272 while (*link) {
273 rb = *link;
274 if (x > rb_to_hole_size(rb)) {
275 link = &rb->rb_left__entry.rbe_left;
276 } else {
277 link = &rb->rb_right__entry.rbe_right;
278 first = false0;
Value stored to 'first' is never read
279 }
280 }
281
282 rb_link_node(&node->rb_hole_size, rb, link);
283 rb_insert_color_cached(&node->rb_hole_size, root, first)linux_root_RB_INSERT_COLOR((struct linux_root *)(&(root)->
rb_root), (&node->rb_hole_size));
284}
285
286static void add_hole(struct drm_mm_node *node)
287{
288 struct drm_mm *mm = node->mm;
289
290 node->hole_size =
291 __drm_mm_hole_node_end(node) - __drm_mm_hole_node_start(node);
292 DRM_MM_BUG_ON(!drm_mm_hole_follows(node))((void)0);
293
294 insert_hole_size(&mm->holes_size, node);
295 DRM_RB_INSERT(mm->holes_addr, rb_hole_addr, HOLE_ADDR)do { struct rb_node **link = &mm->holes_addr.rb_node, *
rb = ((void *)0); u64 x = (__drm_mm_hole_node_start(node)); while
(*link) { rb = *link; if (x < (__drm_mm_hole_node_start((
{ const __typeof( ((struct drm_mm_node *)0)->rb_hole_addr )
*__mptr = (rb); (struct drm_mm_node *)( (char *)__mptr - __builtin_offsetof
(struct drm_mm_node, rb_hole_addr) );})))) link = &rb->
__entry.rbe_left; else link = &rb->__entry.rbe_right; }
rb_link_node(&node->rb_hole_addr, rb, link); linux_root_RB_INSERT_COLOR
((struct linux_root *)(&mm->holes_addr), (&node->
rb_hole_addr)); } while (0);
296
297 list_add(&node->hole_stack, &mm->hole_stack);
298}
299
300static void rm_hole(struct drm_mm_node *node)
301{
302 DRM_MM_BUG_ON(!drm_mm_hole_follows(node))((void)0);
303
304 list_del(&node->hole_stack);
305 rb_erase_cached(&node->rb_hole_size, &node->mm->holes_size)linux_root_RB_REMOVE((struct linux_root *)(&(&node->
mm->holes_size)->rb_root), (&node->rb_hole_size)
);
306 rb_erase(&node->rb_hole_addr, &node->mm->holes_addr)linux_root_RB_REMOVE((struct linux_root *)(&node->mm->
holes_addr), (&node->rb_hole_addr));
307 node->hole_size = 0;
308
309 DRM_MM_BUG_ON(drm_mm_hole_follows(node))((void)0);
310}
311
312static inline struct drm_mm_node *rb_hole_size_to_node(struct rb_node *rb)
313{
314 return rb_entry_safe(rb, struct drm_mm_node, rb_hole_size)(rb ? ({ const __typeof( ((struct drm_mm_node *)0)->rb_hole_size
) *__mptr = (rb); (struct drm_mm_node *)( (char *)__mptr - __builtin_offsetof
(struct drm_mm_node, rb_hole_size) );}) : ((void *)0));
315}
316
317static inline struct drm_mm_node *rb_hole_addr_to_node(struct rb_node *rb)
318{
319 return rb_entry_safe(rb, struct drm_mm_node, rb_hole_addr)(rb ? ({ const __typeof( ((struct drm_mm_node *)0)->rb_hole_addr
) *__mptr = (rb); (struct drm_mm_node *)( (char *)__mptr - __builtin_offsetof
(struct drm_mm_node, rb_hole_addr) );}) : ((void *)0));
320}
321
322static inline u64 rb_hole_size(struct rb_node *rb)
323{
324 return rb_entry(rb, struct drm_mm_node, rb_hole_size)({ const __typeof( ((struct drm_mm_node *)0)->rb_hole_size
) *__mptr = (rb); (struct drm_mm_node *)( (char *)__mptr - __builtin_offsetof
(struct drm_mm_node, rb_hole_size) );})->hole_size;
325}
326
327static struct drm_mm_node *best_hole(struct drm_mm *mm, u64 size)
328{
329 struct rb_node *rb = mm->holes_size.rb_root.rb_node;
330 struct drm_mm_node *best = NULL((void *)0);
331
332 do {
333 struct drm_mm_node *node =
334 rb_entry(rb, struct drm_mm_node, rb_hole_size)({ const __typeof( ((struct drm_mm_node *)0)->rb_hole_size
) *__mptr = (rb); (struct drm_mm_node *)( (char *)__mptr - __builtin_offsetof
(struct drm_mm_node, rb_hole_size) );});
335
336 if (size <= node->hole_size) {
337 best = node;
338 rb = rb->rb_right__entry.rbe_right;
339 } else {
340 rb = rb->rb_left__entry.rbe_left;
341 }
342 } while (rb);
343
344 return best;
345}
346
347static struct drm_mm_node *find_hole(struct drm_mm *mm, u64 addr)
348{
349 struct rb_node *rb = mm->holes_addr.rb_node;
350 struct drm_mm_node *node = NULL((void *)0);
351
352 while (rb) {
353 u64 hole_start;
354
355 node = rb_hole_addr_to_node(rb);
356 hole_start = __drm_mm_hole_node_start(node);
357
358 if (addr < hole_start)
359 rb = node->rb_hole_addr.rb_left__entry.rbe_left;
360 else if (addr > hole_start + node->hole_size)
361 rb = node->rb_hole_addr.rb_right__entry.rbe_right;
362 else
363 break;
364 }
365
366 return node;
367}
368
369static struct drm_mm_node *
370first_hole(struct drm_mm *mm,
371 u64 start, u64 end, u64 size,
372 enum drm_mm_insert_mode mode)
373{
374 switch (mode) {
375 default:
376 case DRM_MM_INSERT_BEST:
377 return best_hole(mm, size);
378
379 case DRM_MM_INSERT_LOW:
380 return find_hole(mm, start);
381
382 case DRM_MM_INSERT_HIGH:
383 return find_hole(mm, end);
384
385 case DRM_MM_INSERT_EVICT:
386 return list_first_entry_or_null(&mm->hole_stack,(list_empty(&mm->hole_stack) ? ((void *)0) : ({ const __typeof
( ((struct drm_mm_node *)0)->hole_stack ) *__mptr = ((&
mm->hole_stack)->next); (struct drm_mm_node *)( (char *
)__mptr - __builtin_offsetof(struct drm_mm_node, hole_stack) )
;}))
387 struct drm_mm_node,(list_empty(&mm->hole_stack) ? ((void *)0) : ({ const __typeof
( ((struct drm_mm_node *)0)->hole_stack ) *__mptr = ((&
mm->hole_stack)->next); (struct drm_mm_node *)( (char *
)__mptr - __builtin_offsetof(struct drm_mm_node, hole_stack) )
;}))
388 hole_stack)(list_empty(&mm->hole_stack) ? ((void *)0) : ({ const __typeof
( ((struct drm_mm_node *)0)->hole_stack ) *__mptr = ((&
mm->hole_stack)->next); (struct drm_mm_node *)( (char *
)__mptr - __builtin_offsetof(struct drm_mm_node, hole_stack) )
;}));
389 }
390}
391
392static struct drm_mm_node *
393next_hole(struct drm_mm *mm,
394 struct drm_mm_node *node,
395 enum drm_mm_insert_mode mode)
396{
397 switch (mode) {
398 default:
399 case DRM_MM_INSERT_BEST:
400 return rb_hole_size_to_node(rb_prev(&node->rb_hole_size)linux_root_RB_PREV((&node->rb_hole_size)));
401
402 case DRM_MM_INSERT_LOW:
403 return rb_hole_addr_to_node(rb_next(&node->rb_hole_addr)linux_root_RB_NEXT((&node->rb_hole_addr)));
404
405 case DRM_MM_INSERT_HIGH:
406 return rb_hole_addr_to_node(rb_prev(&node->rb_hole_addr)linux_root_RB_PREV((&node->rb_hole_addr)));
407
408 case DRM_MM_INSERT_EVICT:
409 node = list_next_entry(node, hole_stack)({ const __typeof( ((typeof(*(node)) *)0)->hole_stack ) *__mptr
= (((node)->hole_stack.next)); (typeof(*(node)) *)( (char
*)__mptr - __builtin_offsetof(typeof(*(node)), hole_stack) )
;});
410 return &node->hole_stack == &mm->hole_stack ? NULL((void *)0) : node;
411 }
412}
413
414/**
415 * drm_mm_reserve_node - insert an pre-initialized node
416 * @mm: drm_mm allocator to insert @node into
417 * @node: drm_mm_node to insert
418 *
419 * This functions inserts an already set-up &drm_mm_node into the allocator,
420 * meaning that start, size and color must be set by the caller. All other
421 * fields must be cleared to 0. This is useful to initialize the allocator with
422 * preallocated objects which must be set-up before the range allocator can be
423 * set-up, e.g. when taking over a firmware framebuffer.
424 *
425 * Returns:
426 * 0 on success, -ENOSPC if there's no hole where @node is.
427 */
428int drm_mm_reserve_node(struct drm_mm *mm, struct drm_mm_node *node)
429{
430 struct drm_mm_node *hole;
431 u64 hole_start, hole_end;
432 u64 adj_start, adj_end;
433 u64 end;
434
435 end = node->start + node->size;
436 if (unlikely(end <= node->start)__builtin_expect(!!(end <= node->start), 0))
437 return -ENOSPC28;
438
439 /* Find the relevant hole to add our node to */
440 hole = find_hole(mm, node->start);
441 if (!hole)
442 return -ENOSPC28;
443
444 adj_start = hole_start = __drm_mm_hole_node_start(hole);
445 adj_end = hole_end = hole_start + hole->hole_size;
446
447 if (mm->color_adjust)
448 mm->color_adjust(hole, node->color, &adj_start, &adj_end);
449
450 if (adj_start > node->start || adj_end < end)
451 return -ENOSPC28;
452
453 node->mm = mm;
454
455 __set_bit(DRM_MM_NODE_ALLOCATED_BIT0, &node->flags);
456 list_add(&node->node_list, &hole->node_list);
457 drm_mm_interval_tree_add_node(hole, node);
458 node->hole_size = 0;
459
460 rm_hole(hole);
461 if (node->start > hole_start)
462 add_hole(hole);
463 if (end < hole_end)
464 add_hole(node);
465
466 save_stack(node);
467 return 0;
468}
469EXPORT_SYMBOL(drm_mm_reserve_node);
470
471static u64 rb_to_hole_size_or_zero(struct rb_node *rb)
472{
473 return rb ? rb_to_hole_size(rb) : 0;
474}
475
476/**
477 * drm_mm_insert_node_in_range - ranged search for space and insert @node
478 * @mm: drm_mm to allocate from
479 * @node: preallocate node to insert
480 * @size: size of the allocation
481 * @alignment: alignment of the allocation
482 * @color: opaque tag value to use for this node
483 * @range_start: start of the allowed range for this node
484 * @range_end: end of the allowed range for this node
485 * @mode: fine-tune the allocation search and placement
486 *
487 * The preallocated @node must be cleared to 0.
488 *
489 * Returns:
490 * 0 on success, -ENOSPC if there's no suitable hole.
491 */
492int drm_mm_insert_node_in_range(struct drm_mm * const mm,
493 struct drm_mm_node * const node,
494 u64 size, u64 alignment,
495 unsigned long color,
496 u64 range_start, u64 range_end,
497 enum drm_mm_insert_mode mode)
498{
499 struct drm_mm_node *hole;
500 u64 remainder_mask;
501 bool_Bool once;
502
503 DRM_MM_BUG_ON(range_start > range_end)((void)0);
504
505 if (unlikely(size == 0 || range_end - range_start < size)__builtin_expect(!!(size == 0 || range_end - range_start <
size), 0))
506 return -ENOSPC28;
507
508 if (rb_to_hole_size_or_zero(rb_first_cached(&mm->holes_size)linux_root_RB_MINMAX((struct linux_root *)(&(&mm->
holes_size)->rb_root), -1)) < size)
509 return -ENOSPC28;
510
511 if (alignment <= 1)
512 alignment = 0;
513
514 once = mode & DRM_MM_INSERT_ONCE;
515 mode &= ~DRM_MM_INSERT_ONCE;
516
517 remainder_mask = is_power_of_2(alignment)(((alignment) != 0) && (((alignment) - 1) & (alignment
)) == 0) ? alignment - 1 : 0;
518 for (hole = first_hole(mm, range_start, range_end, size, mode);
519 hole;
520 hole = once ? NULL((void *)0) : next_hole(mm, hole, mode)) {
521 u64 hole_start = __drm_mm_hole_node_start(hole);
522 u64 hole_end = hole_start + hole->hole_size;
523 u64 adj_start, adj_end;
524 u64 col_start, col_end;
525
526 if (mode == DRM_MM_INSERT_LOW && hole_start >= range_end)
527 break;
528
529 if (mode == DRM_MM_INSERT_HIGH && hole_end <= range_start)
530 break;
531
532 col_start = hole_start;
533 col_end = hole_end;
534 if (mm->color_adjust)
535 mm->color_adjust(hole, color, &col_start, &col_end);
536
537 adj_start = max(col_start, range_start)(((col_start)>(range_start))?(col_start):(range_start));
538 adj_end = min(col_end, range_end)(((col_end)<(range_end))?(col_end):(range_end));
539
540 if (adj_end <= adj_start || adj_end - adj_start < size)
541 continue;
542
543 if (mode == DRM_MM_INSERT_HIGH)
544 adj_start = adj_end - size;
545
546 if (alignment) {
547 u64 rem;
548
549 if (likely(remainder_mask)__builtin_expect(!!(remainder_mask), 1))
550 rem = adj_start & remainder_mask;
551 else
552 div64_u64_rem(adj_start, alignment, &rem);
553 if (rem) {
554 adj_start -= rem;
555 if (mode != DRM_MM_INSERT_HIGH)
556 adj_start += alignment;
557
558 if (adj_start < max(col_start, range_start)(((col_start)>(range_start))?(col_start):(range_start)) ||
559 min(col_end, range_end)(((col_end)<(range_end))?(col_end):(range_end)) - adj_start < size)
560 continue;
561
562 if (adj_end <= adj_start ||
563 adj_end - adj_start < size)
564 continue;
565 }
566 }
567
568 node->mm = mm;
569 node->size = size;
570 node->start = adj_start;
571 node->color = color;
572 node->hole_size = 0;
573
574 __set_bit(DRM_MM_NODE_ALLOCATED_BIT0, &node->flags);
575 list_add(&node->node_list, &hole->node_list);
576 drm_mm_interval_tree_add_node(hole, node);
577
578 rm_hole(hole);
579 if (adj_start > hole_start)
580 add_hole(hole);
581 if (adj_start + size < hole_end)
582 add_hole(node);
583
584 save_stack(node);
585 return 0;
586 }
587
588 return -ENOSPC28;
589}
590EXPORT_SYMBOL(drm_mm_insert_node_in_range);
591
592static inline bool_Bool drm_mm_node_scanned_block(const struct drm_mm_node *node)
593{
594 return test_bit(DRM_MM_NODE_SCANNED_BIT1, &node->flags);
595}
596
597/**
598 * drm_mm_remove_node - Remove a memory node from the allocator.
599 * @node: drm_mm_node to remove
600 *
601 * This just removes a node from its drm_mm allocator. The node does not need to
602 * be cleared again before it can be re-inserted into this or any other drm_mm
603 * allocator. It is a bug to call this function on a unallocated node.
604 */
605void drm_mm_remove_node(struct drm_mm_node *node)
606{
607 struct drm_mm *mm = node->mm;
608 struct drm_mm_node *prev_node;
609
610 DRM_MM_BUG_ON(!drm_mm_node_allocated(node))((void)0);
611 DRM_MM_BUG_ON(drm_mm_node_scanned_block(node))((void)0);
612
613 prev_node = list_prev_entry(node, node_list)({ const __typeof( ((typeof(*(node)) *)0)->node_list ) *__mptr
= (((node)->node_list.prev)); (typeof(*(node)) *)( (char *
)__mptr - __builtin_offsetof(typeof(*(node)), node_list) );});
614
615 if (drm_mm_hole_follows(node))
616 rm_hole(node);
617
618 drm_mm_interval_tree_remove(node, &mm->interval_tree);
619 list_del(&node->node_list);
620
621 if (drm_mm_hole_follows(prev_node))
622 rm_hole(prev_node);
623 add_hole(prev_node);
624
625 clear_bit_unlock(DRM_MM_NODE_ALLOCATED_BIT0, &node->flags);
626}
627EXPORT_SYMBOL(drm_mm_remove_node);
628
629/**
630 * drm_mm_replace_node - move an allocation from @old to @new
631 * @old: drm_mm_node to remove from the allocator
632 * @new: drm_mm_node which should inherit @old's allocation
633 *
634 * This is useful for when drivers embed the drm_mm_node structure and hence
635 * can't move allocations by reassigning pointers. It's a combination of remove
636 * and insert with the guarantee that the allocation start will match.
637 */
638void drm_mm_replace_node(struct drm_mm_node *old, struct drm_mm_node *new)
639{
640 struct drm_mm *mm = old->mm;
641
642 DRM_MM_BUG_ON(!drm_mm_node_allocated(old))((void)0);
643
644 *new = *old;
645
646 __set_bit(DRM_MM_NODE_ALLOCATED_BIT0, &new->flags);
647 list_replace(&old->node_list, &new->node_list);
648 rb_replace_node_cached(&old->rb, &new->rb, &mm->interval_tree)rb_replace_node(&old->rb, &new->rb, &(&
mm->interval_tree)->rb_root);
649
650 if (drm_mm_hole_follows(old)) {
651 list_replace(&old->hole_stack, &new->hole_stack);
652 rb_replace_node_cached(&old->rb_hole_size,rb_replace_node(&old->rb_hole_size, &new->rb_hole_size
, &(&mm->holes_size)->rb_root)
653 &new->rb_hole_size,rb_replace_node(&old->rb_hole_size, &new->rb_hole_size
, &(&mm->holes_size)->rb_root)
654 &mm->holes_size)rb_replace_node(&old->rb_hole_size, &new->rb_hole_size
, &(&mm->holes_size)->rb_root);
655 rb_replace_node(&old->rb_hole_addr,
656 &new->rb_hole_addr,
657 &mm->holes_addr);
658 }
659
660 clear_bit_unlock(DRM_MM_NODE_ALLOCATED_BIT0, &old->flags);
661}
662EXPORT_SYMBOL(drm_mm_replace_node);
663
664/**
665 * DOC: lru scan roster
666 *
667 * Very often GPUs need to have continuous allocations for a given object. When
668 * evicting objects to make space for a new one it is therefore not most
669 * efficient when we simply start to select all objects from the tail of an LRU
670 * until there's a suitable hole: Especially for big objects or nodes that
671 * otherwise have special allocation constraints there's a good chance we evict
672 * lots of (smaller) objects unnecessarily.
673 *
674 * The DRM range allocator supports this use-case through the scanning
675 * interfaces. First a scan operation needs to be initialized with
676 * drm_mm_scan_init() or drm_mm_scan_init_with_range(). The driver adds
677 * objects to the roster, probably by walking an LRU list, but this can be
678 * freely implemented. Eviction candiates are added using
679 * drm_mm_scan_add_block() until a suitable hole is found or there are no
680 * further evictable objects. Eviction roster metadata is tracked in &struct
681 * drm_mm_scan.
682 *
683 * The driver must walk through all objects again in exactly the reverse
684 * order to restore the allocator state. Note that while the allocator is used
685 * in the scan mode no other operation is allowed.
686 *
687 * Finally the driver evicts all objects selected (drm_mm_scan_remove_block()
688 * reported true) in the scan, and any overlapping nodes after color adjustment
689 * (drm_mm_scan_color_evict()). Adding and removing an object is O(1), and
690 * since freeing a node is also O(1) the overall complexity is
691 * O(scanned_objects). So like the free stack which needs to be walked before a
692 * scan operation even begins this is linear in the number of objects. It
693 * doesn't seem to hurt too badly.
694 */
695
696/**
697 * drm_mm_scan_init_with_range - initialize range-restricted lru scanning
698 * @scan: scan state
699 * @mm: drm_mm to scan
700 * @size: size of the allocation
701 * @alignment: alignment of the allocation
702 * @color: opaque tag value to use for the allocation
703 * @start: start of the allowed range for the allocation
704 * @end: end of the allowed range for the allocation
705 * @mode: fine-tune the allocation search and placement
706 *
707 * This simply sets up the scanning routines with the parameters for the desired
708 * hole.
709 *
710 * Warning:
711 * As long as the scan list is non-empty, no other operations than
712 * adding/removing nodes to/from the scan list are allowed.
713 */
714void drm_mm_scan_init_with_range(struct drm_mm_scan *scan,
715 struct drm_mm *mm,
716 u64 size,
717 u64 alignment,
718 unsigned long color,
719 u64 start,
720 u64 end,
721 enum drm_mm_insert_mode mode)
722{
723 DRM_MM_BUG_ON(start >= end)((void)0);
724 DRM_MM_BUG_ON(!size || size > end - start)((void)0);
725 DRM_MM_BUG_ON(mm->scan_active)((void)0);
726
727 scan->mm = mm;
728
729 if (alignment <= 1)
730 alignment = 0;
731
732 scan->color = color;
733 scan->alignment = alignment;
734 scan->remainder_mask = is_power_of_2(alignment)(((alignment) != 0) && (((alignment) - 1) & (alignment
)) == 0) ? alignment - 1 : 0;
735 scan->size = size;
736 scan->mode = mode;
737
738 DRM_MM_BUG_ON(end <= start)((void)0);
739 scan->range_start = start;
740 scan->range_end = end;
741
742 scan->hit_start = U64_MAX0xffffffffffffffffULL;
743 scan->hit_end = 0;
744}
745EXPORT_SYMBOL(drm_mm_scan_init_with_range);
746
747/**
748 * drm_mm_scan_add_block - add a node to the scan list
749 * @scan: the active drm_mm scanner
750 * @node: drm_mm_node to add
751 *
752 * Add a node to the scan list that might be freed to make space for the desired
753 * hole.
754 *
755 * Returns:
756 * True if a hole has been found, false otherwise.
757 */
758bool_Bool drm_mm_scan_add_block(struct drm_mm_scan *scan,
759 struct drm_mm_node *node)
760{
761 struct drm_mm *mm = scan->mm;
762 struct drm_mm_node *hole;
763 u64 hole_start, hole_end;
764 u64 col_start, col_end;
765 u64 adj_start, adj_end;
766
767 DRM_MM_BUG_ON(node->mm != mm)((void)0);
768 DRM_MM_BUG_ON(!drm_mm_node_allocated(node))((void)0);
769 DRM_MM_BUG_ON(drm_mm_node_scanned_block(node))((void)0);
770 __set_bit(DRM_MM_NODE_SCANNED_BIT1, &node->flags);
771 mm->scan_active++;
772
773 /* Remove this block from the node_list so that we enlarge the hole
774 * (distance between the end of our previous node and the start of
775 * or next), without poisoning the link so that we can restore it
776 * later in drm_mm_scan_remove_block().
777 */
778 hole = list_prev_entry(node, node_list)({ const __typeof( ((typeof(*(node)) *)0)->node_list ) *__mptr
= (((node)->node_list.prev)); (typeof(*(node)) *)( (char *
)__mptr - __builtin_offsetof(typeof(*(node)), node_list) );});
779 DRM_MM_BUG_ON(list_next_entry(hole, node_list) != node)((void)0);
780 __list_del_entry(&node->node_list)list_del(&node->node_list);
781
782 hole_start = __drm_mm_hole_node_start(hole);
783 hole_end = __drm_mm_hole_node_end(hole);
784
785 col_start = hole_start;
786 col_end = hole_end;
787 if (mm->color_adjust)
788 mm->color_adjust(hole, scan->color, &col_start, &col_end);
789
790 adj_start = max(col_start, scan->range_start)(((col_start)>(scan->range_start))?(col_start):(scan->
range_start));
791 adj_end = min(col_end, scan->range_end)(((col_end)<(scan->range_end))?(col_end):(scan->range_end
));
792 if (adj_end <= adj_start || adj_end - adj_start < scan->size)
793 return false0;
794
795 if (scan->mode == DRM_MM_INSERT_HIGH)
796 adj_start = adj_end - scan->size;
797
798 if (scan->alignment) {
799 u64 rem;
800
801 if (likely(scan->remainder_mask)__builtin_expect(!!(scan->remainder_mask), 1))
802 rem = adj_start & scan->remainder_mask;
803 else
804 div64_u64_rem(adj_start, scan->alignment, &rem);
805 if (rem) {
806 adj_start -= rem;
807 if (scan->mode != DRM_MM_INSERT_HIGH)
808 adj_start += scan->alignment;
809 if (adj_start < max(col_start, scan->range_start)(((col_start)>(scan->range_start))?(col_start):(scan->
range_start)) ||
810 min(col_end, scan->range_end)(((col_end)<(scan->range_end))?(col_end):(scan->range_end
)) - adj_start < scan->size)
811 return false0;
812
813 if (adj_end <= adj_start ||
814 adj_end - adj_start < scan->size)
815 return false0;
816 }
817 }
818
819 scan->hit_start = adj_start;
820 scan->hit_end = adj_start + scan->size;
821
822 DRM_MM_BUG_ON(scan->hit_start >= scan->hit_end)((void)0);
823 DRM_MM_BUG_ON(scan->hit_start < hole_start)((void)0);
824 DRM_MM_BUG_ON(scan->hit_end > hole_end)((void)0);
825
826 return true1;
827}
828EXPORT_SYMBOL(drm_mm_scan_add_block);
829
830/**
831 * drm_mm_scan_remove_block - remove a node from the scan list
832 * @scan: the active drm_mm scanner
833 * @node: drm_mm_node to remove
834 *
835 * Nodes **must** be removed in exactly the reverse order from the scan list as
836 * they have been added (e.g. using list_add() as they are added and then
837 * list_for_each() over that eviction list to remove), otherwise the internal
838 * state of the memory manager will be corrupted.
839 *
840 * When the scan list is empty, the selected memory nodes can be freed. An
841 * immediately following drm_mm_insert_node_in_range_generic() or one of the
842 * simpler versions of that function with !DRM_MM_SEARCH_BEST will then return
843 * the just freed block (because it's at the top of the free_stack list).
844 *
845 * Returns:
846 * True if this block should be evicted, false otherwise. Will always
847 * return false when no hole has been found.
848 */
849bool_Bool drm_mm_scan_remove_block(struct drm_mm_scan *scan,
850 struct drm_mm_node *node)
851{
852 struct drm_mm_node *prev_node;
853
854 DRM_MM_BUG_ON(node->mm != scan->mm)((void)0);
855 DRM_MM_BUG_ON(!drm_mm_node_scanned_block(node))((void)0);
856 __clear_bit(DRM_MM_NODE_SCANNED_BIT1, &node->flags);
857
858 DRM_MM_BUG_ON(!node->mm->scan_active)((void)0);
859 node->mm->scan_active--;
860
861 /* During drm_mm_scan_add_block() we decoupled this node leaving
862 * its pointers intact. Now that the caller is walking back along
863 * the eviction list we can restore this block into its rightful
864 * place on the full node_list. To confirm that the caller is walking
865 * backwards correctly we check that prev_node->next == node->next,
866 * i.e. both believe the same node should be on the other side of the
867 * hole.
868 */
869 prev_node = list_prev_entry(node, node_list)({ const __typeof( ((typeof(*(node)) *)0)->node_list ) *__mptr
= (((node)->node_list.prev)); (typeof(*(node)) *)( (char *
)__mptr - __builtin_offsetof(typeof(*(node)), node_list) );});
870 DRM_MM_BUG_ON(list_next_entry(prev_node, node_list) !=((void)0)
871 list_next_entry(node, node_list))((void)0);
872 list_add(&node->node_list, &prev_node->node_list);
873
874 return (node->start + node->size > scan->hit_start &&
875 node->start < scan->hit_end);
876}
877EXPORT_SYMBOL(drm_mm_scan_remove_block);
878
879/**
880 * drm_mm_scan_color_evict - evict overlapping nodes on either side of hole
881 * @scan: drm_mm scan with target hole
882 *
883 * After completing an eviction scan and removing the selected nodes, we may
884 * need to remove a few more nodes from either side of the target hole if
885 * mm.color_adjust is being used.
886 *
887 * Returns:
888 * A node to evict, or NULL if there are no overlapping nodes.
889 */
890struct drm_mm_node *drm_mm_scan_color_evict(struct drm_mm_scan *scan)
891{
892 struct drm_mm *mm = scan->mm;
893 struct drm_mm_node *hole;
894 u64 hole_start, hole_end;
895
896 DRM_MM_BUG_ON(list_empty(&mm->hole_stack))((void)0);
897
898 if (!mm->color_adjust)
899 return NULL((void *)0);
900
901 /*
902 * The hole found during scanning should ideally be the first element
903 * in the hole_stack list, but due to side-effects in the driver it
904 * may not be.
905 */
906 list_for_each_entry(hole, &mm->hole_stack, hole_stack)for (hole = ({ const __typeof( ((__typeof(*hole) *)0)->hole_stack
) *__mptr = ((&mm->hole_stack)->next); (__typeof(*
hole) *)( (char *)__mptr - __builtin_offsetof(__typeof(*hole)
, hole_stack) );}); &hole->hole_stack != (&mm->
hole_stack); hole = ({ const __typeof( ((__typeof(*hole) *)0)
->hole_stack ) *__mptr = (hole->hole_stack.next); (__typeof
(*hole) *)( (char *)__mptr - __builtin_offsetof(__typeof(*hole
), hole_stack) );})) {
907 hole_start = __drm_mm_hole_node_start(hole);
908 hole_end = hole_start + hole->hole_size;
909
910 if (hole_start <= scan->hit_start &&
911 hole_end >= scan->hit_end)
912 break;
913 }
914
915 /* We should only be called after we found the hole previously */
916 DRM_MM_BUG_ON(&hole->hole_stack == &mm->hole_stack)((void)0);
917 if (unlikely(&hole->hole_stack == &mm->hole_stack)__builtin_expect(!!(&hole->hole_stack == &mm->hole_stack
), 0))
918 return NULL((void *)0);
919
920 DRM_MM_BUG_ON(hole_start > scan->hit_start)((void)0);
921 DRM_MM_BUG_ON(hole_end < scan->hit_end)((void)0);
922
923 mm->color_adjust(hole, scan->color, &hole_start, &hole_end);
924 if (hole_start > scan->hit_start)
925 return hole;
926 if (hole_end < scan->hit_end)
927 return list_next_entry(hole, node_list)({ const __typeof( ((typeof(*(hole)) *)0)->node_list ) *__mptr
= (((hole)->node_list.next)); (typeof(*(hole)) *)( (char *
)__mptr - __builtin_offsetof(typeof(*(hole)), node_list) );});
928
929 return NULL((void *)0);
930}
931EXPORT_SYMBOL(drm_mm_scan_color_evict);
932
933/**
934 * drm_mm_init - initialize a drm-mm allocator
935 * @mm: the drm_mm structure to initialize
936 * @start: start of the range managed by @mm
937 * @size: end of the range managed by @mm
938 *
939 * Note that @mm must be cleared to 0 before calling this function.
940 */
941void drm_mm_init(struct drm_mm *mm, u64 start, u64 size)
942{
943 DRM_MM_BUG_ON(start + size <= start)((void)0);
944
945 mm->color_adjust = NULL((void *)0);
946
947 INIT_LIST_HEAD(&mm->hole_stack);
948 mm->interval_tree = RB_ROOT_CACHED(struct rb_root_cached) { ((void *)0) };
949 mm->holes_size = RB_ROOT_CACHED(struct rb_root_cached) { ((void *)0) };
950 mm->holes_addr = RB_ROOT(struct rb_root) { ((void *)0) };
951
952 /* Clever trick to avoid a special case in the free hole tracking. */
953 INIT_LIST_HEAD(&mm->head_node.node_list);
954 mm->head_node.flags = 0;
955 mm->head_node.mm = mm;
956 mm->head_node.start = start + size;
957 mm->head_node.size = -size;
958 add_hole(&mm->head_node);
959
960 mm->scan_active = 0;
961}
962EXPORT_SYMBOL(drm_mm_init);
963
964/**
965 * drm_mm_takedown - clean up a drm_mm allocator
966 * @mm: drm_mm allocator to clean up
967 *
968 * Note that it is a bug to call this function on an allocator which is not
969 * clean.
970 */
971void drm_mm_takedown(struct drm_mm *mm)
972{
973 if (WARN(!drm_mm_clean(mm),({ int __ret = !!(!drm_mm_clean(mm)); if (__ret) printf("Memory manager not clean during takedown.\n"
); __builtin_expect(!!(__ret), 0); })
974 "Memory manager not clean during takedown.\n")({ int __ret = !!(!drm_mm_clean(mm)); if (__ret) printf("Memory manager not clean during takedown.\n"
); __builtin_expect(!!(__ret), 0); }))
975 show_leaks(mm);
976}
977EXPORT_SYMBOL(drm_mm_takedown);
978
979static u64 drm_mm_dump_hole(struct drm_printer *p, const struct drm_mm_node *entry)
980{
981 u64 start, size;
982
983 size = entry->hole_size;
984 if (size) {
985 start = drm_mm_hole_node_start(entry);
986 drm_printf(p, "%#018llx-%#018llx: %llu: free\n",
987 start, start + size, size);
988 }
989
990 return size;
991}
992/**
993 * drm_mm_print - print allocator state
994 * @mm: drm_mm allocator to print
995 * @p: DRM printer to use
996 */
997void drm_mm_print(const struct drm_mm *mm, struct drm_printer *p)
998{
999 const struct drm_mm_node *entry;
1000 u64 total_used = 0, total_free = 0, total = 0;
1001
1002 total_free += drm_mm_dump_hole(p, &mm->head_node);
1003
1004 drm_mm_for_each_node(entry, mm)for (entry = ({ const __typeof( ((__typeof(*entry) *)0)->node_list
) *__mptr = (((&(mm)->head_node.node_list))->next)
; (__typeof(*entry) *)( (char *)__mptr - __builtin_offsetof(__typeof
(*entry), node_list) );}); &entry->node_list != ((&
(mm)->head_node.node_list)); entry = ({ const __typeof( ((
__typeof(*entry) *)0)->node_list ) *__mptr = (entry->node_list
.next); (__typeof(*entry) *)( (char *)__mptr - __builtin_offsetof
(__typeof(*entry), node_list) );})) {
1005 drm_printf(p, "%#018llx-%#018llx: %llu: used\n", entry->start,
1006 entry->start + entry->size, entry->size);
1007 total_used += entry->size;
1008 total_free += drm_mm_dump_hole(p, entry);
1009 }
1010 total = total_free + total_used;
1011
1012 drm_printf(p, "total: %llu, used %llu free %llu\n", total,
1013 total_used, total_free);
1014}
1015EXPORT_SYMBOL(drm_mm_print);