mpv/ta/ta_talloc.h

164 lines
7.5 KiB
C
Raw Normal View History

/* Copyright (C) 2017 the mpv developers
*
* Permission to use, copy, modify, and/or distribute this software for any
Replace talloc There are multiple reasons to do this. One big reason is the license: talloc is LGPLv3+, which forces mpv to be licensed as GPLv3+. Another one is that our talloc copy contains modifications, which makes it essentially incompatible with upstream talloc (in particular, our version aborts on out of memory conditions - well, it wasn't my idea). Updating from upstream is also a bit involved - the talloc source is not really organized in a way to allow copying it into projects (and this isn't an intended use-case). Finally, talloc is kind of big and bloated. The replacement halves the amount of code - mainly because we didn't use all talloc features. It's even more extreme if you compare upstream talloc (~4700 lines) and the new allocator without talloc compat (~900 lines). The replacement provides all features we need. It also doesn't clash with talloc. (The talloc compatibility wrapper uses macros to avoid introducing linker-level symbols which could clash with libtalloc.) It also tries to lower the overhead (only 4 words opposed to 10 words in talloc for leaf nodes in release mode). Debugging features like leak reporting can be enabled at compile time and add somewhat more overhead. Though I'm not sure whether the overhead reduction was actually successful: allocations with children need an "extra" header, which adds plenty of overhead, and it turns out that almost half of all allocations have children. Maybe the implementation could be simplified and the extra header removed - even then, overhead would be lower than talloc's. Currently, debugging features can be entirely deactivated by defining NDEBUG - I'm not sure if anything defines this directly yet, though. Unlike in talloc, the leak reporting stuff is thread-safe. (That's also why it's far less elegant, and requires extra list pointers.) Comes with a compatibility layer, so no changes to mpv source code are needed. The idea is that we will pretend to be using talloc for a while, so that we can revert to our old talloc implementation at any time for debugging purposes. Some inspiration was taken from Mesa's ralloc: http://cgit.freedesktop.org/mesa/mesa/tree/src/glsl/ralloc.h This is another talloc replacement, but lacks some features we need (getting size of an allocation, debugging features, being able to access children in the dtor). There's some information in ta/README what will happen next and how the transition is expected to progress.
2013-10-12 23:17:45 +00:00
* purpose with or without fee is hereby granted, provided that the above
* copyright notice and this permission notice appear in all copies.
*
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
*/
#ifndef TA_TALLOC_H_
#define TA_TALLOC_H_
#include <string.h>
Replace talloc There are multiple reasons to do this. One big reason is the license: talloc is LGPLv3+, which forces mpv to be licensed as GPLv3+. Another one is that our talloc copy contains modifications, which makes it essentially incompatible with upstream talloc (in particular, our version aborts on out of memory conditions - well, it wasn't my idea). Updating from upstream is also a bit involved - the talloc source is not really organized in a way to allow copying it into projects (and this isn't an intended use-case). Finally, talloc is kind of big and bloated. The replacement halves the amount of code - mainly because we didn't use all talloc features. It's even more extreme if you compare upstream talloc (~4700 lines) and the new allocator without talloc compat (~900 lines). The replacement provides all features we need. It also doesn't clash with talloc. (The talloc compatibility wrapper uses macros to avoid introducing linker-level symbols which could clash with libtalloc.) It also tries to lower the overhead (only 4 words opposed to 10 words in talloc for leaf nodes in release mode). Debugging features like leak reporting can be enabled at compile time and add somewhat more overhead. Though I'm not sure whether the overhead reduction was actually successful: allocations with children need an "extra" header, which adds plenty of overhead, and it turns out that almost half of all allocations have children. Maybe the implementation could be simplified and the extra header removed - even then, overhead would be lower than talloc's. Currently, debugging features can be entirely deactivated by defining NDEBUG - I'm not sure if anything defines this directly yet, though. Unlike in talloc, the leak reporting stuff is thread-safe. (That's also why it's far less elegant, and requires extra list pointers.) Comes with a compatibility layer, so no changes to mpv source code are needed. The idea is that we will pretend to be using talloc for a while, so that we can revert to our old talloc implementation at any time for debugging purposes. Some inspiration was taken from Mesa's ralloc: http://cgit.freedesktop.org/mesa/mesa/tree/src/glsl/ralloc.h This is another talloc replacement, but lacks some features we need (getting size of an allocation, debugging features, being able to access children in the dtor). There's some information in ta/README what will happen next and how the transition is expected to progress.
2013-10-12 23:17:45 +00:00
#include "ta.h"
// Note: all talloc wrappers are wired to the "x" functions, which abort on OOM.
// libtalloc doesn't do that, but the mplayer2 internal copy of it did.
#define talloc ta_xnew
#define talloc_zero ta_xznew
#define talloc_array ta_xnew_array
#define talloc_zero_array ta_xznew_array
#define talloc_array_size ta_xnew_array_size
#define talloc_realloc ta_xrealloc
#define talloc_ptrtype ta_xnew_ptrtype
#define talloc_array_ptrtype ta_xnew_array_ptrtype
#define talloc_steal ta_xsteal
#define talloc_realloc_size ta_xrealloc_size
#define talloc_new ta_xnew_context
#define talloc_set_destructor ta_xset_destructor
#define talloc_parent ta_find_parent
#define talloc_enable_leak_report ta_enable_leak_report
#define talloc_size ta_xalloc_size
#define talloc_zero_size ta_xzalloc_size
#define talloc_get_size ta_get_size
#define talloc_free_children ta_free_children
#define talloc_free ta_free
#define talloc_dup ta_xdup
Replace talloc There are multiple reasons to do this. One big reason is the license: talloc is LGPLv3+, which forces mpv to be licensed as GPLv3+. Another one is that our talloc copy contains modifications, which makes it essentially incompatible with upstream talloc (in particular, our version aborts on out of memory conditions - well, it wasn't my idea). Updating from upstream is also a bit involved - the talloc source is not really organized in a way to allow copying it into projects (and this isn't an intended use-case). Finally, talloc is kind of big and bloated. The replacement halves the amount of code - mainly because we didn't use all talloc features. It's even more extreme if you compare upstream talloc (~4700 lines) and the new allocator without talloc compat (~900 lines). The replacement provides all features we need. It also doesn't clash with talloc. (The talloc compatibility wrapper uses macros to avoid introducing linker-level symbols which could clash with libtalloc.) It also tries to lower the overhead (only 4 words opposed to 10 words in talloc for leaf nodes in release mode). Debugging features like leak reporting can be enabled at compile time and add somewhat more overhead. Though I'm not sure whether the overhead reduction was actually successful: allocations with children need an "extra" header, which adds plenty of overhead, and it turns out that almost half of all allocations have children. Maybe the implementation could be simplified and the extra header removed - even then, overhead would be lower than talloc's. Currently, debugging features can be entirely deactivated by defining NDEBUG - I'm not sure if anything defines this directly yet, though. Unlike in talloc, the leak reporting stuff is thread-safe. (That's also why it's far less elegant, and requires extra list pointers.) Comes with a compatibility layer, so no changes to mpv source code are needed. The idea is that we will pretend to be using talloc for a while, so that we can revert to our old talloc implementation at any time for debugging purposes. Some inspiration was taken from Mesa's ralloc: http://cgit.freedesktop.org/mesa/mesa/tree/src/glsl/ralloc.h This is another talloc replacement, but lacks some features we need (getting size of an allocation, debugging features, being able to access children in the dtor). There's some information in ta/README what will happen next and how the transition is expected to progress.
2013-10-12 23:17:45 +00:00
#define talloc_memdup ta_xmemdup
#define talloc_strdup ta_xstrdup
#define talloc_strndup ta_xstrndup
#define talloc_asprintf ta_xasprintf
#define talloc_vasprintf ta_xvasprintf
// Don't define linker-level symbols, as that would clash with real libtalloc.
#define talloc_strdup_append ta_talloc_strdup_append
#define talloc_strdup_append_buffer ta_talloc_strdup_append_buffer
#define talloc_strndup_append ta_talloc_strndup_append
#define talloc_strndup_append_buffer ta_talloc_strndup_append_buffer
#define talloc_vasprintf_append ta_talloc_vasprintf_append
#define talloc_vasprintf_append_buffer ta_talloc_vasprintf_append_buffer
#define talloc_asprintf_append ta_talloc_asprintf_append
#define talloc_asprintf_append_buffer ta_talloc_asprintf_append_buffer
char *ta_talloc_strdup(void *t, const char *p);
char *ta_talloc_strdup_append(char *s, const char *a);
char *ta_talloc_strdup_append_buffer(char *s, const char *a);
char *ta_talloc_strndup(void *t, const char *p, size_t n);
char *ta_talloc_strndup_append(char *s, const char *a, size_t n);
char *ta_talloc_strndup_append_buffer(char *s, const char *a, size_t n);
char *ta_talloc_vasprintf_append(char *s, const char *fmt, va_list ap) TA_PRF(2, 0);
char *ta_talloc_vasprintf_append_buffer(char *s, const char *fmt, va_list ap) TA_PRF(2, 0);
char *ta_talloc_asprintf_append(char *s, const char *fmt, ...) TA_PRF(2, 3);
char *ta_talloc_asprintf_append_buffer(char *s, const char *fmt, ...) TA_PRF(2, 3);
2013-12-17 01:18:16 +00:00
// mpv specific stuff - should be made part of proper TA API
#define TA_FREEP(pctx) do {talloc_free(*(pctx)); *(pctx) = NULL;} while(0)
2013-12-17 01:18:16 +00:00
#define TA_EXPAND_ARGS(...) __VA_ARGS__
2019-12-28 20:02:51 +00:00
// Return number of allocated entries in typed array p[].
#define MP_TALLOC_AVAIL(p) (talloc_get_size(p) / sizeof((p)[0]))
2013-12-17 01:18:16 +00:00
2019-12-28 20:02:51 +00:00
// Resize array p so that p[count-1] is the last valid entry. ctx as ta parent.
#define MP_RESIZE_ARRAY(ctx, p, count) \
do { \
(p) = ta_xrealloc_size(ctx, p, \
ta_calc_array_size(sizeof((p)[0]), count)); \
} while (0)
2013-12-17 01:18:16 +00:00
2019-12-28 20:02:51 +00:00
// Resize array p so that p[nextidx] is accessible. Preallocate additional
// space to make appending more efficient, never shrink. ctx as ta parent.
2013-12-17 01:18:16 +00:00
#define MP_TARRAY_GROW(ctx, p, nextidx) \
do { \
size_t nextidx_ = (nextidx); \
if (nextidx_ >= MP_TALLOC_AVAIL(p)) \
MP_RESIZE_ARRAY(ctx, p, ta_calc_prealloc_elems(nextidx_)); \
2013-12-17 01:18:16 +00:00
} while (0)
2019-12-28 20:02:51 +00:00
// Append the last argument to array p (with count idxvar), basically:
// p[idxvar++] = ...; ctx as ta parent.
2013-12-17 01:18:16 +00:00
#define MP_TARRAY_APPEND(ctx, p, idxvar, ...) \
do { \
MP_TARRAY_GROW(ctx, p, idxvar); \
(p)[(idxvar)] = (TA_EXPAND_ARGS(__VA_ARGS__));\
(idxvar)++; \
} while (0)
2019-12-28 20:02:51 +00:00
// Insert the last argument at p[at] (array p with count idxvar), basically:
// for(idxvar-1 down to at) p[n+1] = p[n]; p[at] = ...; idxvar++;
// ctx as ta parent. Required: at >= 0 && at <= idxvar.
#define MP_TARRAY_INSERT_AT(ctx, p, idxvar, at, ...)\
do { \
size_t at_ = (at); \
assert(at_ <= (idxvar)); \
MP_TARRAY_GROW(ctx, p, idxvar); \
memmove((p) + at_ + 1, (p) + at_, \
((idxvar) - at_) * sizeof((p)[0])); \
(idxvar)++; \
(p)[at_] = (TA_EXPAND_ARGS(__VA_ARGS__)); \
} while (0)
2019-12-28 20:04:55 +00:00
// Given an array p with count idxvar, insert c elements at p[at], so that
// p[at] to p[at+c-1] can be accessed. The elements at p[at] and following
// are shifted up by c before insertion. The new entries are uninitialized.
// ctx as ta parent. Required: at >= 0 && at <= idxvar.
#define MP_TARRAY_INSERT_N_AT(ctx, p, idxvar, at, c)\
do { \
size_t at_ = (at); \
assert(at_ <= (idxvar)); \
size_t c_ = (c); \
MP_TARRAY_GROW(ctx, p, (idxvar) + c_); \
memmove((p) + at_ + c_, (p) + at_, \
((idxvar) - at_) * sizeof((p)[0])); \
(idxvar) += c_; \
} while (0)
2019-12-28 20:02:51 +00:00
// Remove p[at] from array p with count idxvar (inverse of MP_TARRAY_INSERT_AT()).
2013-12-17 01:18:16 +00:00
// Doesn't actually free any memory, or do any other talloc calls.
#define MP_TARRAY_REMOVE_AT(p, idxvar, at) \
do { \
size_t at_ = (at); \
assert(at_ <= (idxvar)); \
memmove((p) + at_, (p) + at_ + 1, \
((idxvar) - at_ - 1) * sizeof((p)[0])); \
(idxvar)--; \
} while (0)
// Returns whether or not there was any element to pop.
#define MP_TARRAY_POP(p, idxvar, out) \
((idxvar) > 0 \
? (*(out) = (p)[--(idxvar)], true) \
: false \
)
2013-12-17 01:18:16 +00:00
#define talloc_struct(ctx, type, ...) \
talloc_memdup(ctx, &(type) TA_EXPAND_ARGS(__VA_ARGS__), sizeof(type))
Replace talloc There are multiple reasons to do this. One big reason is the license: talloc is LGPLv3+, which forces mpv to be licensed as GPLv3+. Another one is that our talloc copy contains modifications, which makes it essentially incompatible with upstream talloc (in particular, our version aborts on out of memory conditions - well, it wasn't my idea). Updating from upstream is also a bit involved - the talloc source is not really organized in a way to allow copying it into projects (and this isn't an intended use-case). Finally, talloc is kind of big and bloated. The replacement halves the amount of code - mainly because we didn't use all talloc features. It's even more extreme if you compare upstream talloc (~4700 lines) and the new allocator without talloc compat (~900 lines). The replacement provides all features we need. It also doesn't clash with talloc. (The talloc compatibility wrapper uses macros to avoid introducing linker-level symbols which could clash with libtalloc.) It also tries to lower the overhead (only 4 words opposed to 10 words in talloc for leaf nodes in release mode). Debugging features like leak reporting can be enabled at compile time and add somewhat more overhead. Though I'm not sure whether the overhead reduction was actually successful: allocations with children need an "extra" header, which adds plenty of overhead, and it turns out that almost half of all allocations have children. Maybe the implementation could be simplified and the extra header removed - even then, overhead would be lower than talloc's. Currently, debugging features can be entirely deactivated by defining NDEBUG - I'm not sure if anything defines this directly yet, though. Unlike in talloc, the leak reporting stuff is thread-safe. (That's also why it's far less elegant, and requires extra list pointers.) Comes with a compatibility layer, so no changes to mpv source code are needed. The idea is that we will pretend to be using talloc for a while, so that we can revert to our old talloc implementation at any time for debugging purposes. Some inspiration was taken from Mesa's ralloc: http://cgit.freedesktop.org/mesa/mesa/tree/src/glsl/ralloc.h This is another talloc replacement, but lacks some features we need (getting size of an allocation, debugging features, being able to access children in the dtor). There's some information in ta/README what will happen next and how the transition is expected to progress.
2013-10-12 23:17:45 +00:00
#endif