/src/curl/lib/uint-hashset.c

Source
/***************************************************************************
 *                                  _   _ ____  _
 *  Project                     ___| | | |  _ \| |
 *                             / __| | | | |_) | |
 *                            | (__| |_| |  _ <| |___
 *                             \___|\___/|_| \_\_____|
 *
 * Copyright (C) Daniel Stenberg, <daniel@haxx.se>, et al.
 *
 * This software is licensed as described in the file COPYING, which
 * you should have received as part of this distribution. The terms
 * are also available at https://curl.se/docs/copyright.html.
 *
 * You may opt to use, copy, modify, merge, publish, distribute and/or sell
 * copies of the Software, and permit persons to whom the Software is
 * furnished to do so, under the terms of the COPYING file.
 *
 * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
 * KIND, either express or implied.
 *
 * SPDX-License-Identifier: curl
 *
 ***************************************************************************/
#include "curl_setup.h"

#include "uint-hashset.h"
#include "curlx/strdup.h"

/* random patterns for API verification */
#ifdef DEBUGBUILD
#define CURL_U8_STRSET_MAGIC 0x7117e783
#endif

#define CURL_U8_STRSET_DEBUG      0

#define CURL_SWAP(a, b) (((a) ^= (b)), ((b) ^= (a)), ((a) ^= (b)))

static const uint8_t u8_smask[] = {
  0x00U,
  0x01U,
  0x03U,
  0x07U,
  0x0FU,
  0x1FU,
  0x3FU,
  0x7FU,
  0xFFU,
};

#define CURL_U8_SET_SLOT_IDX(s, i)    (uint8_t)((i) & u8_smask[(s)->slotbits])
#define CURL_U8_SLOT_CNT(i)           ((uint16_t)u8_smask[(i)] + 1)
#define CURL_U8_SET_SLOT_CNT(s)       CURL_U8_SLOT_CNT((s)->slotbits)

/* A hashset for tuples (id, string) using Robin Hood Hashing.
 * <https://www.cs.cornell.edu/courses/JavaAndDS/files/hashing_RobinHood.pdf>
 * The basic idea here to handle collisions by robbing "rich" entries and
 * giving to the "poor":
 * - We have an array: (id, string) are ideally placed at index "id % size".
 * - If slot at index is already occupied, we have a collision.
 * - A simple collision strategy would look at the next index, and the
 *   next until finding an empty slot.
 * - The drawback is that this may lead to many checks on lookups, as it
 *   will need to also look at subsequent slots until it finds the match.
 *   The amount of lookups is the "probe sequence length" (psl) and this
 *   may vary greatly between entries.
 * - Robin Hood Hashing balances the 'psl's of all entries more evenly:
 *   - psl == 0 means an entry is in exactly the right slot
 *   - psl == 1 means it is in the slot right after. psl == 2 is the slot
 *     after that, etc.
 *   - when inserting a new entry, track its psl. Finding a slot where
 *     the existing entry has a lower psl makes a swap. Put the new entry
 *     and its psl there, take the previous entry and its psl and find
 *     the next best slot for the previous entry. */
void Curl_u8_strset_init(struct u8_strset *set)
{
#if defined(__GNUC__) && __GNUC__ >= 13
#pragma GCC diagnostic push
#pragma GCC diagnostic ignored "-Warray-bounds"
#endif
  memset(set, 0, sizeof(*set));
#if defined(__GNUC__) && __GNUC__ >= 13
#pragma GCC diagnostic pop
#endif
  set->data = set->sdata;
  set->ids = set->sids;
  set->psl = set->spsl;
  set->slotbits = CURL_U8_STRSET_START_BITS;
  set->count = 0;
#ifdef DEBUGBUILD
  set->init = CURL_U8_STRSET_MAGIC;
#endif
}

void Curl_u8_strset_clear(struct u8_strset *set)
{
  uint16_t i;
  DEBUGASSERT(set->init == CURL_U8_STRSET_MAGIC);
  for(i = 0; i < CURL_U8_SET_SLOT_CNT(set); ++i)
    curlx_safefree(set->data[i]);

  if(set->data != set->sdata)
    curlx_safefree(set->data);
  Curl_u8_strset_init(set);
}

static void u8_strset_addn(struct u8_strset *set, uint8_t id, char *val)
{
  uint8_t i = CURL_U8_SET_SLOT_IDX(set, id);
  uint8_t psl = 0;
  while(set->data[i]) {
    if(psl > set->psl[i]) { /* SWAP */
      char *tmpdata;
      tmpdata = set->data[i];
      set->data[i] = val;
      val = tmpdata;
      CURL_SWAP(set->psl[i], psl);
      CURL_SWAP(set->ids[i], id);
    }
    i = CURL_U8_SET_SLOT_IDX(set, i + 1);
    ++psl;
  }
  set->ids[i] = id;
  set->data[i] = val;
  set->psl[i] = psl;
  ++set->count;
}

static bool u8_strset_grow(struct u8_strset *set)
{
  uint8_t i, *prev_ids, nslotbits;
  uint16_t prev_slots;
  char **prev_data;
  size_t nslots;
  void *d;

  if(set->slotbits >= 8)
    return FALSE;
  nslotbits = (uint8_t)(set->slotbits + 1);
#if CURL_U8_STRSET_DEBUG
  curl_mfprintf(stderr, "u8_strset_grow from %d to %d\n",
                set->slotbits, nslotbits);
#endif
  nslots = CURL_U8_SLOT_CNT(nslotbits);
  d = curlx_calloc(1, (nslots * sizeof(char *)) + (2 * nslots));
  if(!d)
    return FALSE;

  prev_data = set->data;
  prev_ids = set->ids;
  prev_slots = set->slotbits;
#if defined(__GNUC__) && __GNUC__ >= 13
#pragma GCC diagnostic push
#pragma GCC diagnostic ignored "-Wanalyzer-allocation-size"
#endif
  set->data = (char **)d;
#if defined(__GNUC__) && __GNUC__ >= 13
#pragma GCC diagnostic pop
#endif
  set->ids = (uint8_t *)d + (nslots * sizeof(char *));
  set->psl = set->ids + nslots;
  set->slotbits = nslotbits;
  set->count = 0;
  /* re-add previous entries */
  for(i = 0; i < CURL_U8_SLOT_CNT(prev_slots); ++i) {
    if(prev_data[i])
      u8_strset_addn(set, prev_ids[i], prev_data[i]);
  }
  if(prev_data != set->sdata)
    curlx_free(prev_data);
  return TRUE;
}

static bool u8_strset_get_index(struct u8_strset *set,
                                uint8_t id, uint8_t *pindex)
{
  uint8_t i = CURL_U8_SET_SLOT_IDX(set, id);
  uint8_t psl = 0;
  while(set->data[i] && (psl <= set->psl[i])) {
    if(set->ids[i] == id) {
      *pindex = i;
#if CURL_U8_STRSET_DEBUG
      curl_mfprintf(stderr, "u8_strset_index %d=%s\n", id, set->data[i]);
#endif
      return TRUE;
    }
    i = CURL_U8_SET_SLOT_IDX(set, i + 1);
    ++psl;
  }
#if CURL_U8_STRSET_DEBUG
  curl_mfprintf(stderr, "u8_strset_index %d not found\n", id);
#endif
  *pindex = 0;
  return FALSE;
}

uint16_t Curl_u8_strset_count(struct u8_strset *set)
{
  return set->count;
}

const char *Curl_u8_strset_get(struct u8_strset *set, uint8_t id)
{
  uint8_t i;
  DEBUGASSERT(set->init == CURL_U8_STRSET_MAGIC);
  if(u8_strset_get_index(set, id, &i))
    return set->data[i];
  return NULL;
}

CURLcode Curl_u8_strset_setn(struct u8_strset *set,
                             uint8_t id, char *str)
{
  uint8_t i;

  DEBUGASSERT(set->init == CURL_U8_STRSET_MAGIC);
#if CURL_U8_STRSET_DEBUG
  curl_mfprintf(stderr, "u8_strset_setn %d=%s\n", id, str);
#endif
  if(!str) {
    Curl_u8_strset_unset(set, id);
    return CURLE_OK;
  }

  if(u8_strset_get_index(set, id, &i)) {
    /* `id` is in set, replace value */
    curlx_free(set->data[i]);
    set->data[i] = str;
    return CURLE_OK;
  }
  /* `id` not in set yet, grow if full */
  if((set->count >= CURL_U8_SET_SLOT_CNT(set)) && !u8_strset_grow(set)) {
    curlx_free(str);
    return CURLE_OUT_OF_MEMORY;
  }

  u8_strset_addn(set, id, str);
  return CURLE_OK;
}

CURLcode Curl_u8_strset_setx(struct u8_strset *set,
                             uint8_t id, const char *str, size_t slen)
{
  char *val;

  DEBUGASSERT(set->init == CURL_U8_STRSET_MAGIC);
  if(!str) {
    Curl_u8_strset_unset(set, id);
    return CURLE_OK;
  }

  val = curlx_memdup0(str, slen);
  if(!val)
    return CURLE_OUT_OF_MEMORY;
  return Curl_u8_strset_setn(set, id, val);
}

CURLcode Curl_u8_strset_set(struct u8_strset *set,
                            uint8_t id, const char *str)
{
  return Curl_u8_strset_setx(set, id, str, str ? strlen(str) : 0);
}

static void u8_strset_unset(struct u8_strset *set, uint8_t id, bool zero)
{
  uint8_t i, j;

  DEBUGASSERT(set->init == CURL_U8_STRSET_MAGIC);
  if(u8_strset_get_index(set, id, &i)) {
    /* `id` is in set */
    if(zero)
      curlx_strzero(set->data[i]);
    curlx_safefree(set->data[i]);
    set->ids[i] = set->psl[i] = 0;
    --set->count;
    j = CURL_U8_SET_SLOT_IDX(set, i + 1);
    /* shift all entries with positive psl "down" */
    while(set->data[j] && set->psl[j]) {
      set->data[i] = set->data[j];
      set->ids[i] = set->ids[j];
      set->psl[i] = (uint8_t)(set->psl[j] - 1);
      set->data[j] = NULL;
      set->ids[j] = set->psl[j] = 0;
      i = j;
      j = CURL_U8_SET_SLOT_IDX(set, i + 1);
    }
  }
}

void Curl_u8_strset_unset(struct u8_strset *set, uint8_t id)
{
  u8_strset_unset(set, id, FALSE);
}

void Curl_u8_strset_unset0(struct u8_strset *set, uint8_t id)
{
  u8_strset_unset(set, id, TRUE);
}

CURLcode Curl_u8_strset_copy(struct u8_strset *dest, struct u8_strset *src)
{
  CURLcode result = CURLE_OK;
  uint16_t i;

  DEBUGASSERT(src->init == CURL_U8_STRSET_MAGIC);
  Curl_u8_strset_clear(dest);
  for(i = 0; !result && (i < CURL_U8_SET_SLOT_CNT(src)); ++i) {
    if(src->data[i])
      result = Curl_u8_strset_set(dest, src->ids[i], src->data[i]);
  }
  return result;
}

Coverage Report

Created: 2026-09-04 07:16

Line	Count	Source
1		/***************************************************************************
2		* _ _ ____ _
3		* Project ___\| \| \| \| _ \\| \|
4		* / __\| \| \| \| \|_) \| \|
5		* \| (__\| \|_\| \| _ <\| \|___
6		* \___\|\___/\|_\| \_\_____\|
7		*
8		* Copyright (C) Daniel Stenberg, <daniel@haxx.se>, et al.
9		*
10		* This software is licensed as described in the file COPYING, which
11		* you should have received as part of this distribution. The terms
12		* are also available at https://curl.se/docs/copyright.html.
13		*
14		* You may opt to use, copy, modify, merge, publish, distribute and/or sell
15		* copies of the Software, and permit persons to whom the Software is
16		* furnished to do so, under the terms of the COPYING file.
17		*
18		* This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
19		* KIND, either express or implied.
20		*
21		* SPDX-License-Identifier: curl
22		*
23		***************************************************************************/
24		#include "curl_setup.h"
25
26		#include "uint-hashset.h"
27		#include "curlx/strdup.h"
28
29		/* random patterns for API verification */
30		#ifdef DEBUGBUILD
31	123k	#define CURL_U8_STRSET_MAGIC 0x7117e783
32		#endif
33
34		#define CURL_U8_STRSET_DEBUG 0
35
36	52.2k	#define CURL_SWAP(a, b) (((a) ^= (b)), ((b) ^= (a)), ((a) ^= (b)))
37
38		static const uint8_t u8_smask[] = {
39		0x00U,
40		0x01U,
41		0x03U,
42		0x07U,
43		0x0FU,
44		0x1FU,
45		0x3FU,
46		0x7FU,
47		0xFFU,
48		};
49
50	5.76M	#define CURL_U8_SET_SLOT_IDX(s, i) (uint8_t)((i) & u8_smask[(s)->slotbits])
51	509k	#define CURL_U8_SLOT_CNT(i) ((uint16_t)u8_smask[(i)] + 1)
52	506k	#define CURL_U8_SET_SLOT_CNT(s) CURL_U8_SLOT_CNT((s)->slotbits)
53
54		/* A hashset for tuples (id, string) using Robin Hood Hashing.
55		* <https://www.cs.cornell.edu/courses/JavaAndDS/files/hashing_RobinHood.pdf>
56		* The basic idea here to handle collisions by robbing "rich" entries and
57		* giving to the "poor":
58		* - We have an array: (id, string) are ideally placed at index "id % size".
59		* - If slot at index is already occupied, we have a collision.
60		* - A simple collision strategy would look at the next index, and the
61		* next until finding an empty slot.
62		* - The drawback is that this may lead to many checks on lookups, as it
63		* will need to also look at subsequent slots until it finds the match.
64		* The amount of lookups is the "probe sequence length" (psl) and this
65		* may vary greatly between entries.
66		* - Robin Hood Hashing balances the 'psl's of all entries more evenly:
67		* - psl == 0 means an entry is in exactly the right slot
68		* - psl == 1 means it is in the slot right after. psl == 2 is the slot
69		* after that, etc.
70		* - when inserting a new entry, track its psl. Finding a slot where
71		* the existing entry has a lower psl makes a swap. Put the new entry
72		* and its psl there, take the previous entry and its psl and find
73		* the next best slot for the previous entry. */
74		void Curl_u8_strset_init(struct u8_strset *set)
75	123k	{
76		#if defined(__GNUC__) && __GNUC__ >= 13
77		#pragma GCC diagnostic push
78		#pragma GCC diagnostic ignored "-Warray-bounds"
79		#endif
80	123k	memset(set, 0, sizeof(*set));
81		#if defined(__GNUC__) && __GNUC__ >= 13
82		#pragma GCC diagnostic pop
83		#endif
84	123k	set->data = set->sdata;
85	123k	set->ids = set->sids;
86	123k	set->psl = set->spsl;
87	123k	set->slotbits = CURL_U8_STRSET_START_BITS;
88	123k	set->count = 0;
89	123k	#ifdef DEBUGBUILD
90	123k	set->init = CURL_U8_STRSET_MAGIC;
91	123k	#endif
92	123k	}
93
94		void Curl_u8_strset_clear(struct u8_strset *set)
95	41.2k	{
96	41.2k	uint16_t i;
97	41.2k	DEBUGASSERT(set->init == CURL_U8_STRSET_MAGIC);
98	373k	for(i = 0; i < CURL_U8_SET_SLOT_CNT(set); ++i)
99	332k	curlx_safefree(set->data[i]);
100
101	41.2k	if(set->data != set->sdata)
102	294	curlx_safefree(set->data);
103	41.2k	Curl_u8_strset_init(set);
104	41.2k	}
105
106		static void u8_strset_addn(struct u8_strset set, uint8_t id, char val)
107	134k	{
108	134k	uint8_t i = CURL_U8_SET_SLOT_IDX(set, id);
109	134k	uint8_t psl = 0;
110	188k	while(set->data[i]) {
111	53.9k	if(psl > set->psl[i]) { /* SWAP */
112	26.1k	char *tmpdata;
113	26.1k	tmpdata = set->data[i];
114	26.1k	set->data[i] = val;
115	26.1k	val = tmpdata;
116	26.1k	CURL_SWAP(set->psl[i], psl);
117	26.1k	CURL_SWAP(set->ids[i], id);
118	26.1k	}
119	53.9k	i = CURL_U8_SET_SLOT_IDX(set, i + 1);
120	53.9k	++psl;
121	53.9k	}
122	134k	set->ids[i] = id;
123	134k	set->data[i] = val;
124	134k	set->psl[i] = psl;
125	134k	++set->count;
126	134k	}
127
128		static bool u8_strset_grow(struct u8_strset *set)
129	294	{
130	294	uint8_t i, *prev_ids, nslotbits;
131	294	uint16_t prev_slots;
132	294	char **prev_data;
133	294	size_t nslots;
134	294	void *d;
135
136	294	if(set->slotbits >= 8)
137	0	return FALSE;
138	294	nslotbits = (uint8_t)(set->slotbits + 1);
139		#if CURL_U8_STRSET_DEBUG
140		curl_mfprintf(stderr, "u8_strset_grow from %d to %d\n",
141		set->slotbits, nslotbits);
142		#endif
143	294	nslots = CURL_U8_SLOT_CNT(nslotbits);
144	294	d = curlx_calloc(1, (nslots * sizeof(char )) + (2 nslots));
145	294	if(!d)
146	0	return FALSE;
147
148	294	prev_data = set->data;
149	294	prev_ids = set->ids;
150	294	prev_slots = set->slotbits;
151		#if defined(__GNUC__) && __GNUC__ >= 13
152		#pragma GCC diagnostic push
153		#pragma GCC diagnostic ignored "-Wanalyzer-allocation-size"
154		#endif
155	294	set->data = (char **)d;
156		#if defined(__GNUC__) && __GNUC__ >= 13
157		#pragma GCC diagnostic pop
158		#endif
159	294	set->ids = (uint8_t )d + (nslots sizeof(char *));
160	294	set->psl = set->ids + nslots;
161	294	set->slotbits = nslotbits;
162	294	set->count = 0;
163		/* re-add previous entries */
164	2.64k	for(i = 0; i < CURL_U8_SLOT_CNT(prev_slots); ++i) {
165	2.35k	if(prev_data[i])
166	2.35k	u8_strset_addn(set, prev_ids[i], prev_data[i]);
167	2.35k	}
168	294	if(prev_data != set->sdata)
169	0	curlx_free(prev_data);
170	294	return TRUE;
171	294	}
172
173		static bool u8_strset_get_index(struct u8_strset *set,
174		uint8_t id, uint8_t *pindex)
175	2.82M	{
176	2.82M	uint8_t i = CURL_U8_SET_SLOT_IDX(set, id);
177	2.82M	uint8_t psl = 0;
178	5.58M	while(set->data[i] && (psl <= set->psl[i])) {
179	2.98M	if(set->ids[i] == id) {
180	221k	*pindex = i;
181		#if CURL_U8_STRSET_DEBUG
182		curl_mfprintf(stderr, "u8_strset_index %d=%s\n", id, set->data[i]);
183		#endif
184	221k	return TRUE;
185	221k	}
186	2.75M	i = CURL_U8_SET_SLOT_IDX(set, i + 1);
187	2.75M	++psl;
188	2.75M	}
189		#if CURL_U8_STRSET_DEBUG
190		curl_mfprintf(stderr, "u8_strset_index %d not found\n", id);
191		#endif
192	2.59M	*pindex = 0;
193	2.59M	return FALSE;
194	2.82M	}
195
196		uint16_t Curl_u8_strset_count(struct u8_strset *set)
197	0	{
198	0	return set->count;
199	0	}
200
201		const char Curl_u8_strset_get(struct u8_strset set, uint8_t id)
202	2.47M	{
203	2.47M	uint8_t i;
204	2.47M	DEBUGASSERT(set->init == CURL_U8_STRSET_MAGIC);
205	2.47M	if(u8_strset_get_index(set, id, &i))
206	219k	return set->data[i];
207	2.26M	return NULL;
208	2.47M	}
209
210		CURLcode Curl_u8_strset_setn(struct u8_strset *set,
211		uint8_t id, char *str)
212	134k	{
213	134k	uint8_t i;
214
215	134k	DEBUGASSERT(set->init == CURL_U8_STRSET_MAGIC);
216		#if CURL_U8_STRSET_DEBUG
217		curl_mfprintf(stderr, "u8_strset_setn %d=%s\n", id, str);
218		#endif
219	134k	if(!str) {
220	868	Curl_u8_strset_unset(set, id);
221	868	return CURLE_OK;
222	868	}
223
224	134k	if(u8_strset_get_index(set, id, &i)) {
225		/* `id` is in set, replace value */
226	1.72k	curlx_free(set->data[i]);
227	1.72k	set->data[i] = str;
228	1.72k	return CURLE_OK;
229	1.72k	}
230		/* `id` not in set yet, grow if full */
231	132k	if((set->count >= CURL_U8_SET_SLOT_CNT(set)) && !u8_strset_grow(set)) {
232	0	curlx_free(str);
233	0	return CURLE_OUT_OF_MEMORY;
234	0	}
235
236	132k	u8_strset_addn(set, id, str);
237	132k	return CURLE_OK;
238	132k	}
239
240		CURLcode Curl_u8_strset_setx(struct u8_strset *set,
241		uint8_t id, const char *str, size_t slen)
242	129k	{
243	129k	char *val;
244
245	129k	DEBUGASSERT(set->init == CURL_U8_STRSET_MAGIC);
246	129k	if(!str) {
247	0	Curl_u8_strset_unset(set, id);
248	0	return CURLE_OK;
249	0	}
250
251	129k	val = curlx_memdup0(str, slen);
252	129k	if(!val)
253	0	return CURLE_OUT_OF_MEMORY;
254	129k	return Curl_u8_strset_setn(set, id, val);
255	129k	}
256
257		CURLcode Curl_u8_strset_set(struct u8_strset *set,
258		uint8_t id, const char *str)
259	0	{
260	0	return Curl_u8_strset_setx(set, id, str, str ? strlen(str) : 0);
261	0	}
262
263		static void u8_strset_unset(struct u8_strset *set, uint8_t id, bool zero)
264	207k	{
265	207k	uint8_t i, j;
266
267	207k	DEBUGASSERT(set->init == CURL_U8_STRSET_MAGIC);
268	207k	if(u8_strset_get_index(set, id, &i)) {
269		/* `id` is in set */
270	509	if(zero)
271	448	curlx_strzero(set->data[i]);
272	509	curlx_safefree(set->data[i]);
273	509	set->ids[i] = set->psl[i] = 0;
274	509	--set->count;
275	509	j = CURL_U8_SET_SLOT_IDX(set, i + 1);
276		/* shift all entries with positive psl "down" */
277	617	while(set->data[j] && set->psl[j]) {
278	108	set->data[i] = set->data[j];
279	108	set->ids[i] = set->ids[j];
280	108	set->psl[i] = (uint8_t)(set->psl[j] - 1);
281	108	set->data[j] = NULL;
282	108	set->ids[j] = set->psl[j] = 0;
283	108	i = j;
284	108	j = CURL_U8_SET_SLOT_IDX(set, i + 1);
285	108	}
286	509	}
287	207k	}
288
289		void Curl_u8_strset_unset(struct u8_strset *set, uint8_t id)
290	868	{
291	868	u8_strset_unset(set, id, FALSE);
292	868	}
293
294		void Curl_u8_strset_unset0(struct u8_strset *set, uint8_t id)
295	206k	{
296	206k	u8_strset_unset(set, id, TRUE);
297	206k	}
298
299		CURLcode Curl_u8_strset_copy(struct u8_strset dest, struct u8_strset src)
300	0	{
301	0	CURLcode result = CURLE_OK;
302	0	uint16_t i;
303
304	0	DEBUGASSERT(src->init == CURL_U8_STRSET_MAGIC);
305	0	Curl_u8_strset_clear(dest);
306	0	for(i = 0; !result && (i < CURL_U8_SET_SLOT_CNT(src)); ++i) {
307	0	if(src->data[i])
308	0	result = Curl_u8_strset_set(dest, src->ids[i], src->data[i]);
309	0	}
310	0	return result;
311	0	}