10733088351

Committed 06 Sep 2024 05:03AM UTC coverage: 92.807% (-0.08%) from 92.888%

Build # 10733088351

Build Type

push

github

Committed by

Buristan

Commit Message

Limit number of string format elements to compile.

Reported by pwnhacker0x18.

(cherry picked from commit 4fc48c50f)

When compiling `string.format()` with a huge sequence of elements, it is
possible that too many KGC IRs underflow the IR buffer. This patch
limits the maximum number of `string.format()` elements to be compiled
to 100.

Sergey Kaplun:
* added the description and the test for the problem

Part of tarantool/tarantool#10199

Reviewed-by: Sergey Bronnikov <sergeyb@tarantool.org>
Reviewed-by: Maxim Kokryashkin <m.kokryashkin@tarantool.org>
Signed-off-by: Sergey Kaplun <skaplun@tarantool.org>
(cherry picked from commit d9e19aebc)

Run Details

5682 of 6026 branches covered (94.29%)

Branch coverage included in aggregate %.

2 of 2 new or added lines in 1 file covered. (100.0%)

20 existing lines in 4 files now uncovered.

21658 of 23433 relevant lines covered (92.43%)

2948743.58 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

95.69

/src/lj_str.c

/*
** String handling.
** Copyright (C) 2005-2017 Mike Pall. See Copyright Notice in luajit.h
*/

#define lj_str_c
#define LUA_CORE

#include "lj_obj.h"
#include "lj_gc.h"
#include "lj_err.h"
#include "lj_str.h"
#include "lj_char.h"

#if LUAJIT_USE_ASAN
/* These functions may read past a buffer end, that's ok. */
GCstr *lj_str_new(lua_State *L, const char *str, size_t lenx)
  __attribute__((no_sanitize_address));

int32_t LJ_FASTCALL lj_str_cmp(GCstr *a, GCstr *b)
  __attribute__((no_sanitize_address));

static LJ_AINLINE int str_fastcmp(const char *a, const char *b, MSize len)
  __attribute__((no_sanitize_address));
#endif

/* -- String helpers ------------------------------------------------------ */

/* Ordered compare of strings. Assumes string data is 4-byte aligned. */
int32_t LJ_FASTCALL lj_str_cmp(GCstr *a, GCstr *b)
{
  MSize i, n = a->len > b->len ? b->len : a->len;
  for (i = 0; i < n; i += 4) {
    /* Note: innocuous access up to end of string + 3. */
    uint32_t va = *(const uint32_t *)(strdata(a)+i);
    uint32_t vb = *(const uint32_t *)(strdata(b)+i);
    if (va != vb) {
#if LJ_LE
      va = lj_bswap(va); vb = lj_bswap(vb);
#endif
      i -= n;
      if ((int32_t)i >= -3) {
        va >>= 32+(i<<3); vb >>= 32+(i<<3);
        if (va == vb) break;
      }
      return va < vb ? -1 : 1;
    }
  }
  return (int32_t)(a->len - b->len);
}

/* Fast string data comparison. Caveat: unaligned access to 1st string! */
static LJ_AINLINE int str_fastcmp(const char *a, const char *b, MSize len)
{
  MSize i = 0;
  lj_assertX(len > 0, "fast string compare with zero length");
  lj_assertX((((uintptr_t)a+len-1) & (LJ_PAGESIZE-1)) <= LJ_PAGESIZE-4,
             "fast string compare crossing page boundary");
  do {  /* Note: innocuous access up to end of string + 3. */
    uint32_t v = lj_getu32(a+i) ^ *(const uint32_t *)(b+i);
    if (v) {
      i -= len;
#if LJ_LE
      return (int32_t)i >= -3 ? (v << (32+(i<<3))) : 1;
#else
      return (int32_t)i >= -3 ? (v >> (32+(i<<3))) : 1;
#endif
    }
    i += 4;
  } while (i < len);
  return 0;
}

/* Find fixed string p inside string s. */
const char *lj_str_find(const char *s, const char *p, MSize slen, MSize plen)
{
  if (plen <= slen) {
    if (plen == 0) {
      return s;
    } else {
      int c = *(const uint8_t *)p++;
      plen--; slen -= plen;
      while (slen) {
        const char *q = (const char *)memchr(s, c, slen);
        if (!q) break;
        if (memcmp(q+1, p, plen) == 0) return q;
        q++; slen -= (MSize)(q-s); s = q;
      }
    }
  }
  return NULL;
}

/* Check whether a string has a pattern matching character. */
int lj_str_haspattern(GCstr *s)
{
  const char *p = strdata(s), *q = p + s->len;
  while (p < q) {
    int c = *(const uint8_t *)p++;
    if (lj_char_ispunct(c) && strchr("^$*+?.([%-", c))
      return 1;  /* Found a pattern matching char. */
  }
  return 0;  /* No pattern matching chars found. */
}

/* -- String interning ---------------------------------------------------- */

/* Resize the string hash table (grow and shrink). */
void lj_str_resize(lua_State *L, MSize newmask)
{
  global_State *g = G(L);
  GCRef *newhash;
  MSize i;
  if (g->gc.state == GCSsweepstring || newmask >= LJ_MAX_STRTAB-1)
    return;  /* No resizing during GC traversal or if already too big. */
  newhash = lj_mem_newvec(L, newmask+1, GCRef);
  memset(newhash, 0, (newmask+1)*sizeof(GCRef));
  for (i = g->strmask; i != ~(MSize)0; i--) {  /* Rehash old table. */
    GCobj *p = gcref(g->strhash[i]);
    while (p) {  /* Follow each hash chain and reinsert all strings. */
      MSize h = gco2str(p)->hash & newmask;
      GCobj *next = gcnext(p);
      /* NOBARRIER: The string table is a GC root. */
      setgcrefr(p->gch.nextgc, newhash[h]);
      setgcref(newhash[h], p);
      p = next;
    }
  }
  lj_mem_freevec(g, g->strhash, g->strmask+1, GCRef);
  g->strmask = newmask;
  g->strhash = newhash;
}

#if LUAJIT_SMART_STRINGS
static LJ_AINLINE uint32_t
lj_fullhash(const uint8_t *v, MSize len)
{
  MSize a = 0, b = 0;
  MSize c = 0xcafedead;
  MSize d = 0xdeadbeef;
  MSize h = len;
  lj_assertX(len >= 12, "full hash calculation for too short (%d) string", len);
  for(; len>8; len-=8, v+=8) {
    a ^= lj_getu32(v);
    b ^= lj_getu32(v+4);
    c += a;
    d += b;
    a = lj_rol(a, 5) - d;
    b = lj_rol(b, 7) - c;
    c = lj_rol(c, 24) ^ a;
    d = lj_rol(d, 1) ^ b;
  }
  a ^= lj_getu32(v+len-8);
  b ^= lj_getu32(v+len-4);
  c += b; c -= lj_rol(a, 9);
  d += a; d -= lj_rol(b, 18);
  h -= lj_rol(a^b,7);
  h += c; h += lj_rol(d,13);
  d ^= c;  d -= lj_rol(c,25);
  h ^= d; h -= lj_rol(d,16);
  c ^= h; c -= lj_rol(h,4);
  d ^= c;  d -= lj_rol(c,14);
  h ^= d; h -= lj_rol(d,24);
  return h;
}
#endif

/* Intern a string and return string object. */
GCstr *lj_str_new(lua_State *L, const char *str, size_t lenx)
{
  global_State *g;
  GCstr *s;
  GCobj *o;
  MSize len = (MSize)lenx;
  uint8_t strflags = 0;
#if LUAJIT_SMART_STRINGS
  unsigned collisions = 0;
#endif
  if (lenx >= LJ_MAX_STR)
    lj_err_msg(L, LJ_ERR_STROV);
  g = G(L);
  if (len == 0)
    return &g->strempty;
  /* Compute string hash. Constants taken from lookup3 hash by Bob Jenkins. */
  MSize h = lua_hash(str, len);
  /* Check if the string has already been interned. */
  o = gcref(g->strhash[h & g->strmask]);
#if LUAJIT_SMART_STRINGS
/*
** The default "fast" string hash function samples only a few positions
** in a string, the remaining bytes don't affect the function's result.
** The function performs well for short strings; however long strings
** can yield extremely high collision rates.
**
** An adaptive schema was implemented. Two hash functions are used
** simultaneously. A bucket is picked based on the output of the fast
** hash function. If an item is to be inserted in a collision chain
** longer than a certain threshold, another bucket is picked based on
** the stronger hash function. Since two hash functions are used
** simultaneously, insert should consider two buckets. The second bucket
** is often NOT considered thanks to the bloom filter. The filter is
** rebuilt during GC cycle.
**
** Parameters below were tuned on a set of benchmarks. Max_collisions is
** also backed by theory: the expected maximum length of a collision
** chain in a hash table with the fill factor of 1.0 is
** O(log(N)/log(log(N))), assuming uniformly distributed random keys.
** The upper bound for N=65,000 is 10, hence 40 is a clear indication of
** an anomaly.
**/
#define max_collisions 40
#define inc_collision_soft() (collisions++)
/* If different strings yield the same hash sum, grow counter faster. */
#define inc_collision_hard() (collisions+=1+(len>>4), 1)
#else
#define inc_collision_hard() (1)
#define inc_collision_soft()
#endif
  if (LJ_LIKELY((((uintptr_t)str+len-1) & (LJ_PAGESIZE-1)) <= LJ_PAGESIZE-4)) {
    while (o != NULL) {
      GCstr *sx = gco2str(o);
      if (sx->hash == h && sx->len == len && inc_collision_hard() &&
                      str_fastcmp(str, strdata(sx), len) == 0) {
        /* Resurrect if dead. Can only happen with fixstring() (keywords). */
        if (isdead(g, o)) flipwhite(o);
        g->strhash_hit++;
        return sx;  /* Return existing string. */
      }
      o = gcnext(o);
      inc_collision_soft();
    }
  } else {  /* Slow path: end of string is too close to a page boundary. */
    while (o != NULL) {
      GCstr *sx = gco2str(o);
      if (sx->hash == h && sx->len == len && inc_collision_hard() &&
                      memcmp(str, strdata(sx), len) == 0) {
        /* Resurrect if dead. Can only happen with fixstring() (keywords). */
        if (isdead(g, o)) flipwhite(o);
        g->strhash_hit++;
        return sx;  /* Return existing string. */
      }
      o = gcnext(o);
      inc_collision_soft();
    }
  }
#if LUAJIT_SMART_STRINGS
  /* "Fast" hash function consumes all bytes of a string <= 12 bytes. */
  if (len > 12) {
    /*
    ** The bloom filter is keyed with the high 12 bits of the fast
    ** hash sum. The filter is rebuilt during GC cycle. It's beneficial
    ** to have these bits readily available and avoid hash sum
    ** recalculation during GC. High 6 bits are included in the "full"
    ** hash sum, and bits 19-25 are stored in s->strflags.
    **/
    int search_fullh =
       bloomtest(g->strbloom.cur[0], h>>(sizeof(h)*8- 6)) != 0 &&
       bloomtest(g->strbloom.cur[1], h>>(sizeof(h)*8-12)) != 0;
    if (LJ_UNLIKELY(search_fullh || collisions > max_collisions)) {
      MSize fh = lj_fullhash((const uint8_t*)str, len);
#define high6mask ((~(MSize)0)<<(sizeof(MSize)*8-6))
      fh = (fh >> 6) | (h & high6mask);
      if (search_fullh) {
        /* Recheck if the string has already been interned with "harder" hash. */
        o = gcref(g->strhash[fh & g->strmask]);
        if (LJ_LIKELY((((uintptr_t)str+len-1) & (LJ_PAGESIZE-1)) <= LJ_PAGESIZE-4)) {
          while (o != NULL) {
            GCstr *sx = gco2str(o);
            if (sx->hash == fh && sx->len == len && str_fastcmp(str, strdata(sx), len) == 0) {
              /* Resurrect if dead. Can only happen with fixstring() (keywords). */
              if (isdead(g, o)) flipwhite(o);
              g->strhash_hit++;
              return sx;  /* Return existing string. */
            }
            o = gcnext(o);
          }
        } else {  /* Slow path: end of string is too close to a page boundary. */
          while (o != NULL) {
            GCstr *sx = gco2str(o);
            if (sx->hash == fh && sx->len == len && memcmp(str, strdata(sx), len) == 0) {
              /* Resurrect if dead. Can only happen with fixstring() (keywords). */
              if (isdead(g, o)) flipwhite(o);
              g->strhash_hit++;
              return sx;  /* Return existing string. */
            }
            o = gcnext(o);
          }
        }
      }
      if (collisions > max_collisions) {
        strflags = 0xc0 | ((h>>(sizeof(h)*8-12))&0x3f);
        bloomset(g->strbloom.cur[0], h>>(sizeof(h)*8- 6));
        bloomset(g->strbloom.cur[1], h>>(sizeof(h)*8-12));
        bloomset(g->strbloom.next[0], h>>(sizeof(h)*8- 6));
        bloomset(g->strbloom.next[1], h>>(sizeof(h)*8-12));
        h = fh;
      }
    }
  }
#endif
  g->strhash_miss++;
  /* Nope, create a new string. */
  s = lj_mem_newt(L, sizeof(GCstr)+len+1, GCstr);
  newwhite(g, s);
  s->gct = ~LJ_TSTR;
  s->len = len;
  s->hash = h;
  s->reserved = 0;
  s->strflags = strflags;
  memcpy(strdatawr(s), str, len);
  strdatawr(s)[len] = '\0';  /* Zero-terminate string. */
  /* Add it to string hash table. */
  h &= g->strmask;
  s->nextgc = g->strhash[h];
  /* NOBARRIER: The string table is a GC root. */
  setgcref(g->strhash[h], obj2gco(s));
  if (g->strnum++ > g->strmask)  /* Allow a 100% load factor. */
    lj_str_resize(L, (g->strmask<<1)+1);  /* Grow string table. */
  return s;  /* Return newly interned string. */
}

void LJ_FASTCALL lj_str_free(global_State *g, GCstr *s)
{
  g->strnum--;
  lj_mem_free(g, s, sizestring(s));
}


1	/*
2	** String handling.
3	** Copyright (C) 2005-2017 Mike Pall. See Copyright Notice in luajit.h
4	*/
5
6	#define lj_str_c
7	#define LUA_CORE
8
9	#include "lj_obj.h"
10	#include "lj_gc.h"
11	#include "lj_err.h"
12	#include "lj_str.h"
13	#include "lj_char.h"
14
15	#if LUAJIT_USE_ASAN
16	/* These functions may read past a buffer end, that's ok. */
17	GCstr lj_str_new(lua_State L, const char *str, size_t lenx)
18	__attribute__((no_sanitize_address));
19
20	int32_t LJ_FASTCALL lj_str_cmp(GCstr a, GCstr b)
21	__attribute__((no_sanitize_address));
22
23	static LJ_AINLINE int str_fastcmp(const char a, const char b, MSize len)
24	__attribute__((no_sanitize_address));
25	#endif
26
27	/* -- String helpers ------------------------------------------------------ */
28
29	/* Ordered compare of strings. Assumes string data is 4-byte aligned. */
30	int32_t LJ_FASTCALL lj_str_cmp(GCstr a, GCstr b)	103,370✔
31	{
32	MSize i, n = a->len > b->len ? b->len : a->len;	103,370✔
33	for (i = 0; i < n; i += 4) {	109,870✔
34	/* Note: innocuous access up to end of string + 3. */
35	uint32_t va = (const uint32_t )(strdata(a)+i);	103,472✔
36	uint32_t vb = (const uint32_t )(strdata(b)+i);	103,472✔
37	if (va != vb) {	103,472✔
38	#if LJ_LE
39	va = lj_bswap(va); vb = lj_bswap(vb);	96,972✔
40	#endif
41	i -= n;	96,972✔
42	if ((int32_t)i >= -3) {	96,972✔
43	va >>= 32+(i<<3); vb >>= 32+(i<<3);	96,789✔
44	if (va == vb) break;	96,789✔
45	}
46	return va < vb ? -1 : 1;	96,508✔
47	}
48	}
49	return (int32_t)(a->len - b->len);	6,862✔
50	}
51
52	/* Fast string data comparison. Caveat: unaligned access to 1st string! */
53	static LJ_AINLINE int str_fastcmp(const char a, const char b, MSize len)	8,800,875✔
54	{
55	MSize i = 0;	8,800,875✔
56	lj_assertX(len > 0, "fast string compare with zero length");	16,909,824✔
57	lj_assertX((((uintptr_t)a+len-1) & (LJ_PAGESIZE-1)) <= LJ_PAGESIZE-4,	16,909,824✔
58	"fast string compare crossing page boundary");
59	do { /* Note: innocuous access up to end of string + 3. */	16,909,824✔
60	uint32_t v = lj_getu32(a+i) ^ (const uint32_t )(b+i);	16,909,824✔
61	if (v) {	16,909,824✔
62	i -= len;	3,214,138✔
63	#if LJ_LE
64	return (int32_t)i >= -3 ? (v << (32+(i<<3))) : 1;	3,214,138✔
65	#else
66	return (int32_t)i >= -3 ? (v >> (32+(i<<3))) : 1;
67	#endif
68	}
69	i += 4;	13,695,686✔
70	} while (i < len);	13,695,686✔
71	return 0;
72	}
73
74	/* Find fixed string p inside string s. */
75	const char lj_str_find(const char s, const char *p, MSize slen, MSize plen)	2,676✔
76	{
77	if (plen <= slen) {	2,676✔
78	if (plen == 0) {	2,599✔
79	return s;
80	} else {
81	int c = (const uint8_t )p++;	2,596✔
82	plen--; slen -= plen;	2,596✔
83	while (slen) {	9,022✔
84	const char q = (const char )memchr(s, c, slen);	8,921✔
85	if (!q) break;	8,921✔
86	if (memcmp(q+1, p, plen) == 0) return q;	6,704✔
87	q++; slen -= (MSize)(q-s); s = q;	6,426✔
88	}
89	}
90	}
91	return NULL;
92	}
93
94	/* Check whether a string has a pattern matching character. */
95	int lj_str_haspattern(GCstr *s)	2,674✔
96	{
97	const char p = strdata(s), q = p + s->len;	2,674✔
98	while (p < q) {	5,993✔
99	int c = (const uint8_t )p++;	4,652✔
100	if (lj_char_ispunct(c) && strchr("^$*+?.([%-", c))	4,652✔
101	return 1; /* Found a pattern matching char. */
102	}
103	return 0; /* No pattern matching chars found. */
104	}
105
106	/* -- String interning ---------------------------------------------------- */
107
108	/* Resize the string hash table (grow and shrink). */
109	void lj_str_resize(lua_State *L, MSize newmask)	961✔
110	{
111	global_State *g = G(L);	961✔
112	GCRef *newhash;	961✔
113	MSize i;	961✔
114	if (g->gc.state == GCSsweepstring \|\| newmask >= LJ_MAX_STRTAB-1)	961✔
115	return; /* No resizing during GC traversal or if already too big. */
116	newhash = lj_mem_newvec(L, newmask+1, GCRef);	961✔
117	memset(newhash, 0, (newmask+1)*sizeof(GCRef));	961✔
118	for (i = g->strmask; i != ~(MSize)0; i--) { /* Rehash old table. */	2,438,849✔
119	GCobj *p = gcref(g->strhash[i]);	2,437,888✔
120	while (p) { /* Follow each hash chain and reinsert all strings. */	3,534,114✔
121	MSize h = gco2str(p)->hash & newmask;	1,096,226✔
122	GCobj *next = gcnext(p);	1,096,226✔
123	/* NOBARRIER: The string table is a GC root. */
124	setgcrefr(p->gch.nextgc, newhash[h]);	1,096,226✔
125	setgcref(newhash[h], p);	1,096,226✔
126	p = next;	1,096,226✔
127	}
128	}
129	lj_mem_freevec(g, g->strhash, g->strmask+1, GCRef);	961✔
130	g->strmask = newmask;	961✔
131	g->strhash = newhash;	961✔
132	}
133
134	#if LUAJIT_SMART_STRINGS
135	static LJ_AINLINE uint32_t
136	lj_fullhash(const uint8_t *v, MSize len)
137	{
138	MSize a = 0, b = 0;
139	MSize c = 0xcafedead;
140	MSize d = 0xdeadbeef;
141	MSize h = len;	3,054,315✔
142	lj_assertX(len >= 12, "full hash calculation for too short (%d) string", len);
143	for(; len>8; len-=8, v+=8) {	3,054,315✔
144	a ^= lj_getu32(v);	3,010,216✔
145	b ^= lj_getu32(v+4);	3,010,216✔
146	c += a;	3,010,216✔
147	d += b;	3,010,216✔
148	a = lj_rol(a, 5) - d;	3,010,216✔
149	b = lj_rol(b, 7) - c;	3,010,216✔
150	c = lj_rol(c, 24) ^ a;	3,010,216✔
151	d = lj_rol(d, 1) ^ b;	3,010,216✔
152	}
153	a ^= lj_getu32(v+len-8);	44,099✔
154	b ^= lj_getu32(v+len-4);	44,099✔
155	c += b; c -= lj_rol(a, 9);	44,099✔
156	d += a; d -= lj_rol(b, 18);	44,099✔
157	h -= lj_rol(a^b,7);	44,099✔
158	h += c; h += lj_rol(d,13);	44,099✔
159	d ^= c; d -= lj_rol(c,25);	44,099✔
160	h ^= d; h -= lj_rol(d,16);	44,099✔
161	c ^= h; c -= lj_rol(h,4);	44,099✔
162	d ^= c; d -= lj_rol(c,14);	44,099✔
163	h ^= d; h -= lj_rol(d,24);	44,099✔
164	return h;	44,099✔
165	}
166	#endif
167
168	/* Intern a string and return string object. */
169	GCstr lj_str_new(lua_State L, const char *str, size_t lenx)	9,533,623✔
170	{
171	global_State *g;	9,533,623✔
172	GCstr *s;	9,533,623✔
173	GCobj *o;	9,533,623✔
174	MSize len = (MSize)lenx;	9,533,623✔
175	uint8_t strflags = 0;	9,533,623✔
176	#if LUAJIT_SMART_STRINGS
177	unsigned collisions = 0;	9,533,623✔
178	#endif
179	if (lenx >= LJ_MAX_STR)	9,533,623✔
180	lj_err_msg(L, LJ_ERR_STROV);	×
181	g = G(L);	9,533,623✔
182	if (len == 0)	9,533,623✔
183	return &g->strempty;	5,583✔
184	/* Compute string hash. Constants taken from lookup3 hash by Bob Jenkins. */
185	MSize h = lua_hash(str, len);	9,528,040✔
186	/* Check if the string has already been interned. */
187	o = gcref(g->strhash[h & g->strmask]);	9,528,040✔
188	#if LUAJIT_SMART_STRINGS
189	/*
190	** The default "fast" string hash function samples only a few positions
191	** in a string, the remaining bytes don't affect the function's result.
192	** The function performs well for short strings; however long strings
193	** can yield extremely high collision rates.
194	**
195	** An adaptive schema was implemented. Two hash functions are used
196	** simultaneously. A bucket is picked based on the output of the fast
197	** hash function. If an item is to be inserted in a collision chain
198	** longer than a certain threshold, another bucket is picked based on
199	** the stronger hash function. Since two hash functions are used
200	** simultaneously, insert should consider two buckets. The second bucket
201	** is often NOT considered thanks to the bloom filter. The filter is
202	** rebuilt during GC cycle.
203	**
204	** Parameters below were tuned on a set of benchmarks. Max_collisions is
205	** also backed by theory: the expected maximum length of a collision
206	** chain in a hash table with the fill factor of 1.0 is
207	** O(log(N)/log(log(N))), assuming uniformly distributed random keys.
208	** The upper bound for N=65,000 is 10, hence 40 is a clear indication of
209	** an anomaly.
210	**/
211	#define max_collisions 40
212	#define inc_collision_soft() (collisions++)
213	/* If different strings yield the same hash sum, grow counter faster. */
214	#define inc_collision_hard() (collisions+=1+(len>>4), 1)
215	#else
216	#define inc_collision_hard() (1)
217	#define inc_collision_soft()
218	#endif
219	if (LJ_LIKELY((((uintptr_t)str+len-1) & (LJ_PAGESIZE-1)) <= LJ_PAGESIZE-4)) {	9,528,040✔
220	while (o != NULL) {	13,823,142✔
221	GCstr *sx = gco2str(o);	12,897,664✔
222	if (sx->hash == h && sx->len == len && inc_collision_hard() &&	12,897,664✔
223	str_fastcmp(str, strdata(sx), len) == 0) {	17,382,517✔
224	/* Resurrect if dead. Can only happen with fixstring() (keywords). */
225	if (isdead(g, o)) flipwhite(o);	8,601,236✔
226	g->strhash_hit++;	8,601,236✔
227	return sx; /* Return existing string. */	8,601,236✔
228	}
229	o = gcnext(o);	4,296,428✔
230	inc_collision_soft();	4,296,428✔
231	}
232	} else { /* Slow path: end of string is too close to a page boundary. */
233	while (o != NULL) {	2,160✔
234	GCstr *sx = gco2str(o);	2,041✔
235	if (sx->hash == h && sx->len == len && inc_collision_hard() &&	2,041✔
236	memcmp(str, strdata(sx), len) == 0) {	1,221✔
237	/* Resurrect if dead. Can only happen with fixstring() (keywords). */
238	if (isdead(g, o)) flipwhite(o);	1,207✔
239	g->strhash_hit++;	1,207✔
240	return sx; /* Return existing string. */	1,207✔
241	}
242	o = gcnext(o);	834✔
243	inc_collision_soft();	834✔
244	}
245	}
246	#if LUAJIT_SMART_STRINGS
247	/* "Fast" hash function consumes all bytes of a string <= 12 bytes. */
248	if (len > 12) {	925,597✔
249	/*
250	** The bloom filter is keyed with the high 12 bits of the fast
251	** hash sum. The filter is rebuilt during GC cycle. It's beneficial
252	** to have these bits readily available and avoid hash sum
253	** recalculation during GC. High 6 bits are included in the "full"
254	** hash sum, and bits 19-25 are stored in s->strflags.
255	**/
256	int search_fullh =	441,720✔
257	bloomtest(g->strbloom.cur[0], h>>(sizeof(h)*8- 6)) != 0 &&	220,860✔
258	bloomtest(g->strbloom.cur[1], h>>(sizeof(h)*8-12)) != 0;	53,393✔
259	if (LJ_UNLIKELY(search_fullh \|\| collisions > max_collisions)) {	220,860✔
260	MSize fh = lj_fullhash((const uint8_t*)str, len);	44,099✔
261	#define high6mask ((~(MSize)0)<<(sizeof(MSize)*8-6))
262	fh = (fh >> 6) \| (h & high6mask);	44,099✔
263	if (search_fullh) {	44,099✔
264	/* Recheck if the string has already been interned with "harder" hash. */
265	o = gcref(g->strhash[fh & g->strmask]);	42,783✔
266	if (LJ_LIKELY((((uintptr_t)str+len-1) & (LJ_PAGESIZE-1)) <= LJ_PAGESIZE-4)) {	42,783✔
267	while (o != NULL) {	57,484✔
268	GCstr *sx = gco2str(o);	34,296✔
269	if (sx->hash == fh && sx->len == len && str_fastcmp(str, strdata(sx), len) == 0) {	53,890✔
270	/* Resurrect if dead. Can only happen with fixstring() (keywords). */
271	if (isdead(g, o)) flipwhite(o);	19,594✔
272	g->strhash_hit++;	19,594✔
273	return sx; /* Return existing string. */	19,594✔
274	}
275	o = gcnext(o);	14,702✔
276	}
277	} else { /* Slow path: end of string is too close to a page boundary. */
278	while (o != NULL) {	1✔
UNCOV 279	GCstr *sx = gco2str(o);	×
UNCOV 280	if (sx->hash == fh && sx->len == len && memcmp(str, strdata(sx), len) == 0) {	×
281	/* Resurrect if dead. Can only happen with fixstring() (keywords). */
282	if (isdead(g, o)) flipwhite(o);	×
283	g->strhash_hit++;	×
284	return sx; /* Return existing string. */	×
285	}
UNCOV 286	o = gcnext(o);	×
287	}
288	}
289	}
290	if (collisions > max_collisions) {	24,505✔
291	strflags = 0xc0 \| ((h>>(sizeof(h)*8-12))&0x3f);	21,399✔
292	bloomset(g->strbloom.cur[0], h>>(sizeof(h)*8- 6));	21,399✔
293	bloomset(g->strbloom.cur[1], h>>(sizeof(h)*8-12));	21,399✔
294	bloomset(g->strbloom.next[0], h>>(sizeof(h)*8- 6));	21,399✔
295	bloomset(g->strbloom.next[1], h>>(sizeof(h)*8-12));	21,399✔
296	h = fh;	21,399✔
297	}
298	}
299	}
300	#endif
301	g->strhash_miss++;	906,003✔
302	/* Nope, create a new string. */
303	s = lj_mem_newt(L, sizeof(GCstr)+len+1, GCstr);	906,003✔
304	newwhite(g, s);	906,003✔
305	s->gct = ~LJ_TSTR;	906,003✔
306	s->len = len;	906,003✔
307	s->hash = h;	906,003✔
308	s->reserved = 0;	906,003✔
309	s->strflags = strflags;	906,003✔
310	memcpy(strdatawr(s), str, len);	906,003✔
311	strdatawr(s)[len] = '\0'; /* Zero-terminate string. */	906,003✔
312	/* Add it to string hash table. */
313	h &= g->strmask;	906,003✔
314	s->nextgc = g->strhash[h];	906,003✔
315	/* NOBARRIER: The string table is a GC root. */
316	setgcref(g->strhash[h], obj2gco(s));	906,003✔
317	if (g->strnum++ > g->strmask) /* Allow a 100% load factor. */	906,003✔
318	lj_str_resize(L, (g->strmask<<1)+1); /* Grow string table. */	527✔
319	return s; /* Return newly interned string. */
320	}
321
322	void LJ_FASTCALL lj_str_free(global_State g, GCstr s)	901,728✔
323	{
324	g->strnum--;	901,728✔
325	lj_mem_free(g, s, sizestring(s));	901,728✔
326	}	901,728✔
327

tarantool / luajit / 10733088351

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous