/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"

/* 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_set(struct u8_strset *set,
                            uint8_t id, const char *str)
{
  char *val;

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

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

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: 2026-08-31 06:47

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
28		/* random patterns for API verification */
29		#ifdef DEBUGBUILD
30	23.3k	#define CURL_U8_STRSET_MAGIC 0x7117e783
31		#endif
32
33		#define CURL_U8_STRSET_DEBUG 0
34
35	2.49k	#define CURL_SWAP(a, b) (((a) ^= (b)), ((b) ^= (a)), ((a) ^= (b)))
36
37		static const uint8_t u8_smask[] = {
38		0x00U,
39		0x01U,
40		0x03U,
41		0x07U,
42		0x0FU,
43		0x1FU,
44		0x3FU,
45		0x7FU,
46		0xFFU,
47		};
48
49	146k	#define CURL_U8_SET_SLOT_IDX(s, i) (uint8_t)((i) & u8_smask[(s)->slotbits])
50	98.9k	#define CURL_U8_SLOT_CNT(i) ((uint16_t)u8_smask[(i)] + 1)
51	98.1k	#define CURL_U8_SET_SLOT_CNT(s) CURL_U8_SLOT_CNT((s)->slotbits)
52
53		/* A hashset for tuples (id, string) using Robin Hood Hashing.
54		* <https://www.cs.cornell.edu/courses/JavaAndDS/files/hashing_RobinHood.pdf>
55		* The basic idea here to handle collisions by robbing "rich" entries and
56		* giving to the "poor":
57		* - We have an array: (id, string) are ideally placed at index "id % size".
58		* - If slot at index is already occupied, we have a collision.
59		* - A simple collision strategy would look at the next index, and the
60		* next until finding an empty slot.
61		* - The drawback is that this may lead to many checks on lookups, as it
62		* will need to also look at subsequent slots until it finds the match.
63		* The amount of lookups is the "probe sequence length" (psl) and this
64		* may vary greatly between entries.
65		* - Robin Hood Hashing balances the 'psl's of all entries more evenly:
66		* - psl == 0 means an entry is in exactly the right slot
67		* - psl == 1 means it is in the slot right after. psl == 2 is the slot
68		* after that, etc.
69		* - when inserting a new entry, track its psl. Finding a slot where
70		* the existing entry has a lower psl makes a swap. Put the new entry
71		* and its psl there, take the previous entry and its psl and find
72		* the next best slot for the previous entry. */
73		void Curl_u8_strset_init(struct u8_strset *set)
74	23.3k	{
75		#if defined(__GNUC__) && __GNUC__ >= 13
76		#pragma GCC diagnostic push
77		#pragma GCC diagnostic ignored "-Warray-bounds"
78		#endif
79	23.3k	memset(set, 0, sizeof(*set));
80		#if defined(__GNUC__) && __GNUC__ >= 13
81		#pragma GCC diagnostic pop
82		#endif
83	23.3k	set->data = set->sdata;
84	23.3k	set->ids = set->sids;
85	23.3k	set->psl = set->spsl;
86	23.3k	set->slotbits = CURL_U8_STRSET_START_BITS;
87	23.3k	set->count = 0;
88	23.3k	#ifdef DEBUGBUILD
89	23.3k	set->init = CURL_U8_STRSET_MAGIC;
90	23.3k	#endif
91	23.3k	}
92
93		void Curl_u8_strset_clear(struct u8_strset *set)
94	7.77k	{
95	7.77k	uint16_t i;
96	7.77k	DEBUGASSERT(set->init == CURL_U8_STRSET_MAGIC);
97	70.6k	for(i = 0; i < CURL_U8_SET_SLOT_CNT(set); ++i)
98	62.8k	curlx_safefree(set->data[i]);
99
100	7.77k	if(set->data != set->sdata)
101	58	curlx_safefree(set->data);
102	7.77k	Curl_u8_strset_init(set);
103	7.77k	}
104
105		static void u8_strset_addn(struct u8_strset set, uint8_t id, char val)
106	28.1k	{
107	28.1k	uint8_t i = CURL_U8_SET_SLOT_IDX(set, id);
108	28.1k	uint8_t psl = 0;
109	37.0k	while(set->data[i]) {
110	8.91k	if(psl > set->psl[i]) { /* SWAP */
111	1.24k	char *tmpdata;
112	1.24k	tmpdata = set->data[i];
113	1.24k	set->data[i] = val;
114	1.24k	val = tmpdata;
115	1.24k	CURL_SWAP(set->psl[i], psl);
116	1.24k	CURL_SWAP(set->ids[i], id);
117	1.24k	}
118	8.91k	i = CURL_U8_SET_SLOT_IDX(set, i + 1);
119	8.91k	++psl;
120	8.91k	}
121	28.1k	set->ids[i] = id;
122	28.1k	set->data[i] = val;
123	28.1k	set->psl[i] = psl;
124	28.1k	++set->count;
125	28.1k	}
126
127		static bool u8_strset_grow(struct u8_strset *set)
128	72	{
129	72	uint8_t i, *prev_ids, nslotbits;
130	72	uint16_t prev_slots;
131	72	char **prev_data;
132	72	size_t nslots;
133	72	void *d;
134
135	72	if(set->slotbits >= 8)
136	0	return FALSE;
137	72	nslotbits = (uint8_t)(set->slotbits + 1);
138		#if CURL_U8_STRSET_DEBUG
139		curl_mfprintf(stderr, "u8_strset_grow from %d to %d\n",
140		set->slotbits, nslotbits);
141		#endif
142	72	nslots = CURL_U8_SLOT_CNT(nslotbits);
143	72	d = curlx_calloc(1, (nslots * sizeof(char )) + (2 nslots));
144	72	if(!d)
145	0	return FALSE;
146
147	72	prev_data = set->data;
148	72	prev_ids = set->ids;
149	72	prev_slots = set->slotbits;
150		#if defined(__GNUC__) && __GNUC__ >= 13
151		#pragma GCC diagnostic push
152		#pragma GCC diagnostic ignored "-Wanalyzer-allocation-size"
153		#endif
154	72	set->data = (char **)d;
155		#if defined(__GNUC__) && __GNUC__ >= 13
156		#pragma GCC diagnostic pop
157		#endif
158	72	set->ids = (uint8_t )d + (nslots sizeof(char *));
159	72	set->psl = set->ids + nslots;
160	72	set->slotbits = nslotbits;
161	72	set->count = 0;
162		/* re-add previous entries */
163	760	for(i = 0; i < CURL_U8_SLOT_CNT(prev_slots); ++i) {
164	688	if(prev_data[i])
165	688	u8_strset_addn(set, prev_ids[i], prev_data[i]);
166	688	}
167	72	if(prev_data != set->sdata)
168	14	curlx_free(prev_data);
169	72	return TRUE;
170	72	}
171
172		static bool u8_strset_get_index(struct u8_strset *set,
173		uint8_t id, uint8_t *pindex)
174	88.4k	{
175	88.4k	uint8_t i = CURL_U8_SET_SLOT_IDX(set, id);
176	88.4k	uint8_t psl = 0;
177	109k	while(set->data[i] && (psl <= set->psl[i])) {
178	36.6k	if(set->ids[i] == id) {
179	16.1k	*pindex = i;
180		#if CURL_U8_STRSET_DEBUG
181		curl_mfprintf(stderr, "u8_strset_index %d=%s\n", id, set->data[i]);
182		#endif
183	16.1k	return TRUE;
184	16.1k	}
185	20.5k	i = CURL_U8_SET_SLOT_IDX(set, i + 1);
186	20.5k	++psl;
187	20.5k	}
188		#if CURL_U8_STRSET_DEBUG
189		curl_mfprintf(stderr, "u8_strset_index %d not found\n", id);
190		#endif
191	72.3k	*pindex = 0;
192	72.3k	return FALSE;
193	88.4k	}
194
195		uint16_t Curl_u8_strset_count(struct u8_strset *set)
196	0	{
197	0	return set->count;
198	0	}
199
200		const char Curl_u8_strset_get(struct u8_strset set, uint8_t id)
201	21.0k	{
202	21.0k	uint8_t i;
203	21.0k	DEBUGASSERT(set->init == CURL_U8_STRSET_MAGIC);
204	21.0k	if(u8_strset_get_index(set, id, &i))
205	15.7k	return set->data[i];
206	5.29k	return NULL;
207	21.0k	}
208
209		CURLcode Curl_u8_strset_setn(struct u8_strset *set,
210		uint8_t id, char *str)
211	28.1k	{
212	28.1k	uint8_t i;
213
214	28.1k	DEBUGASSERT(set->init == CURL_U8_STRSET_MAGIC);
215		#if CURL_U8_STRSET_DEBUG
216		curl_mfprintf(stderr, "u8_strset_setn %d=%s\n", id, str);
217		#endif
218	28.1k	if(!str) {
219	526	Curl_u8_strset_unset(set, id);
220	526	return CURLE_OK;
221	526	}
222
223	27.6k	if(u8_strset_get_index(set, id, &i)) {
224		/* `id` is in set, replace value */
225	174	curlx_free(set->data[i]);
226	174	set->data[i] = str;
227	174	return CURLE_OK;
228	174	}
229		/* `id` not in set yet, grow if full */
230	27.4k	if((set->count >= CURL_U8_SET_SLOT_CNT(set)) && !u8_strset_grow(set)) {
231	0	curlx_free(str);
232	0	return CURLE_OUT_OF_MEMORY;
233	0	}
234
235	27.4k	u8_strset_addn(set, id, str);
236	27.4k	return CURLE_OK;
237	27.4k	}
238
239
240		CURLcode Curl_u8_strset_set(struct u8_strset *set,
241		uint8_t id, const char *str)
242	27.0k	{
243	27.0k	char *val;
244
245	27.0k	DEBUGASSERT(set->init == CURL_U8_STRSET_MAGIC);
246	27.0k	if(!str) {
247	0	Curl_u8_strset_unset(set, id);
248	0	return CURLE_OK;
249	0	}
250
251	27.0k	val = curlx_strdup(str);
252	27.0k	if(!val)
253	0	return CURLE_OUT_OF_MEMORY;
254	27.0k	return Curl_u8_strset_setn(set, id, val);
255	27.0k	}
256
257		static void u8_strset_unset(struct u8_strset *set, uint8_t id, bool zero)
258	39.8k	{
259	39.8k	uint8_t i, j;
260
261	39.8k	DEBUGASSERT(set->init == CURL_U8_STRSET_MAGIC);
262	39.8k	if(u8_strset_get_index(set, id, &i)) {
263		/* `id` is in set */
264	216	if(zero)
265	190	curlx_strzero(set->data[i]);
266	216	curlx_safefree(set->data[i]);
267	216	set->ids[i] = set->psl[i] = 0;
268	216	--set->count;
269	216	j = CURL_U8_SET_SLOT_IDX(set, i + 1);
270		/* shift all entries with positive psl "down" */
271	524	while(set->data[j] && set->psl[j]) {
272	308	set->data[i] = set->data[j];
273	308	set->ids[i] = set->ids[j];
274	308	set->psl[i] = (uint8_t)(set->psl[j] - 1);
275	308	set->data[j] = NULL;
276	308	set->ids[j] = set->psl[j] = 0;
277	308	i = j;
278	308	j = CURL_U8_SET_SLOT_IDX(set, i + 1);
279	308	}
280	216	}
281	39.8k	}
282
283		void Curl_u8_strset_unset(struct u8_strset *set, uint8_t id)
284	944	{
285	944	u8_strset_unset(set, id, FALSE);
286	944	}
287
288		void Curl_u8_strset_unset0(struct u8_strset *set, uint8_t id)
289	38.8k	{
290	38.8k	u8_strset_unset(set, id, TRUE);
291	38.8k	}
292
293		CURLcode Curl_u8_strset_copy(struct u8_strset dest, struct u8_strset src)
294	0	{
295	0	CURLcode result = CURLE_OK;
296	0	uint16_t i;
297
298	0	DEBUGASSERT(src->init == CURL_U8_STRSET_MAGIC);
299	0	Curl_u8_strset_clear(dest);
300	0	for(i = 0; !result && (i < CURL_U8_SET_SLOT_CNT(src)); ++i) {
301	0	if(src->data[i])
302	0	result = Curl_u8_strset_set(dest, src->ids[i], src->data[i]);
303	0	}
304	0	return result;
305	0	}