/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;
}

#ifdef UNITTESTS
/* @unittest 3211 */
uint16_t Curl_u8_strset_count(struct u8_strset *set)
{
  return set->count;
}
#endif

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);
}

/* @unittest 3211 */
UNITTEST CURLcode u8_strset_set(struct u8_strset *set,
                                uint8_t id, const char *str);
UNITTEST CURLcode 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 = u8_strset_set(dest, src->ids[i], src->data[i]);
  }
  return result;
}

Coverage Report

Created: 2026-09-14 07:05

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	242k	#define CURL_U8_STRSET_MAGIC 0x7117e783
32		#endif
33
34		#define CURL_U8_STRSET_DEBUG 0
35
36	189k	#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	10.9M	#define CURL_U8_SET_SLOT_IDX(s, i) (uint8_t)((i) & u8_smask[(s)->slotbits])
51	1.13M	#define CURL_U8_SLOT_CNT(i) ((uint16_t)u8_smask[(i)] + 1)
52	1.12M	#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	242k	{
76		#if defined(__GNUC__) && __GNUC__ >= 13
77		#pragma GCC diagnostic push
78		#pragma GCC diagnostic ignored "-Warray-bounds"
79		#endif
80	242k	memset(set, 0, sizeof(*set));
81		#if defined(__GNUC__) && __GNUC__ >= 13
82		#pragma GCC diagnostic pop
83		#endif
84	242k	set->data = set->sdata;
85	242k	set->ids = set->sids;
86	242k	set->psl = set->spsl;
87	242k	set->slotbits = CURL_U8_STRSET_START_BITS;
88	242k	set->count = 0;
89	242k	#ifdef DEBUGBUILD
90	242k	set->init = CURL_U8_STRSET_MAGIC;
91	242k	#endif
92	242k	}
93
94		void Curl_u8_strset_clear(struct u8_strset *set)
95	80.9k	{
96	80.9k	uint16_t i;
97	80.9k	DEBUGASSERT(set->init == CURL_U8_STRSET_MAGIC);
98	735k	for(i = 0; i < CURL_U8_SET_SLOT_CNT(set); ++i)
99	654k	curlx_safefree(set->data[i]);
100
101	80.9k	if(set->data != set->sdata)
102	876	curlx_safefree(set->data);
103	80.9k	Curl_u8_strset_init(set);
104	80.9k	}
105
106		static void u8_strset_addn(struct u8_strset set, uint8_t id, char val)
107	393k	{
108	393k	uint8_t i = CURL_U8_SET_SLOT_IDX(set, id);
109	393k	uint8_t psl = 0;
110	578k	while(set->data[i]) {
111	185k	if(psl > set->psl[i]) { /* SWAP */
112	94.5k	char *tmpdata;
113	94.5k	tmpdata = set->data[i];
114	94.5k	set->data[i] = val;
115	94.5k	val = tmpdata;
116	94.5k	CURL_SWAP(set->psl[i], psl);
117	94.5k	CURL_SWAP(set->ids[i], id);
118	94.5k	}
119	185k	i = CURL_U8_SET_SLOT_IDX(set, i + 1);
120	185k	++psl;
121	185k	}
122	393k	set->ids[i] = id;
123	393k	set->data[i] = val;
124	393k	set->psl[i] = psl;
125	393k	++set->count;
126	393k	}
127
128		static bool u8_strset_grow(struct u8_strset *set)
129	893	{
130	893	uint8_t i, *prev_ids, nslotbits;
131	893	uint16_t prev_slots;
132	893	char **prev_data;
133	893	size_t nslots;
134	893	void *d;
135
136	893	if(set->slotbits >= 8)
137	0	return FALSE;
138	893	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	893	nslots = CURL_U8_SLOT_CNT(nslotbits);
144	893	d = curlx_calloc(1, (nslots * sizeof(char )) + (2 nslots));
145	893	if(!d)
146	0	return FALSE;
147
148	893	prev_data = set->data;
149	893	prev_ids = set->ids;
150	893	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	893	set->data = (char **)d;
156		#if defined(__GNUC__) && __GNUC__ >= 13
157		#pragma GCC diagnostic pop
158		#endif
159	893	set->ids = (uint8_t )d + (nslots sizeof(char *));
160	893	set->psl = set->ids + nslots;
161	893	set->slotbits = nslotbits;
162	893	set->count = 0;
163		/* re-add previous entries */
164	8.17k	for(i = 0; i < CURL_U8_SLOT_CNT(prev_slots); ++i) {
165	7.28k	if(prev_data[i])
166	7.28k	u8_strset_addn(set, prev_ids[i], prev_data[i]);
167	7.28k	}
168	893	if(prev_data != set->sdata)
169	17	curlx_free(prev_data);
170	893	return TRUE;
171	893	}
172
173		static bool u8_strset_get_index(struct u8_strset *set,
174		uint8_t id, uint8_t *pindex)
175	5.30M	{
176	5.30M	uint8_t i = CURL_U8_SET_SLOT_IDX(set, id);
177	5.30M	uint8_t psl = 0;
178	10.3M	while(set->data[i] && (psl <= set->psl[i])) {
179	5.57M	if(set->ids[i] == id) {
180	535k	*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	535k	return TRUE;
185	535k	}
186	5.04M	i = CURL_U8_SET_SLOT_IDX(set, i + 1);
187	5.04M	++psl;
188	5.04M	}
189		#if CURL_U8_STRSET_DEBUG
190		curl_mfprintf(stderr, "u8_strset_index %d not found\n", id);
191		#endif
192	4.77M	*pindex = 0;
193	4.77M	return FALSE;
194	5.30M	}
195
196		#ifdef UNITTESTS
197		/* @unittest 3211 */
198		uint16_t Curl_u8_strset_count(struct u8_strset *set)
199	0	{
200	0	return set->count;
201	0	}
202		#endif
203
204		const char Curl_u8_strset_get(struct u8_strset set, uint8_t id)
205	4.49M	{
206	4.49M	uint8_t i;
207	4.49M	DEBUGASSERT(set->init == CURL_U8_STRSET_MAGIC);
208	4.49M	if(u8_strset_get_index(set, id, &i))
209	520k	return set->data[i];
210	3.97M	return NULL;
211	4.49M	}
212
213		CURLcode Curl_u8_strset_setn(struct u8_strset *set,
214		uint8_t id, char *str)
215	401k	{
216	401k	uint8_t i;
217
218	401k	DEBUGASSERT(set->init == CURL_U8_STRSET_MAGIC);
219		#if CURL_U8_STRSET_DEBUG
220		curl_mfprintf(stderr, "u8_strset_setn %d=%s\n", id, str);
221		#endif
222	401k	if(!str) {
223	3.47k	Curl_u8_strset_unset(set, id);
224	3.47k	return CURLE_OK;
225	3.47k	}
226
227	398k	if(u8_strset_get_index(set, id, &i)) {
228		/* `id` is in set, replace value */
229	12.1k	curlx_free(set->data[i]);
230	12.1k	set->data[i] = str;
231	12.1k	return CURLE_OK;
232	12.1k	}
233		/* `id` not in set yet, grow if full */
234	386k	if((set->count >= CURL_U8_SET_SLOT_CNT(set)) && !u8_strset_grow(set)) {
235	0	curlx_free(str);
236	0	return CURLE_OUT_OF_MEMORY;
237	0	}
238
239	386k	u8_strset_addn(set, id, str);
240	386k	return CURLE_OK;
241	386k	}
242
243		CURLcode Curl_u8_strset_setx(struct u8_strset *set,
244		uint8_t id, const char *str, size_t slen)
245	391k	{
246	391k	char *val;
247
248	391k	DEBUGASSERT(set->init == CURL_U8_STRSET_MAGIC);
249	391k	if(!str) {
250	0	Curl_u8_strset_unset(set, id);
251	0	return CURLE_OK;
252	0	}
253
254	391k	val = curlx_memdup0(str, slen);
255	391k	if(!val)
256	0	return CURLE_OUT_OF_MEMORY;
257	391k	return Curl_u8_strset_setn(set, id, val);
258	391k	}
259
260		/* @unittest 3211 */
261		UNITTEST CURLcode u8_strset_set(struct u8_strset *set,
262		uint8_t id, const char *str);
263		UNITTEST CURLcode u8_strset_set(struct u8_strset *set,
264		uint8_t id, const char *str)
265	0	{
266	0	return Curl_u8_strset_setx(set, id, str, str ? strlen(str) : 0);
267	0	}
268
269		static void u8_strset_unset(struct u8_strset *set, uint8_t id, bool zero)
270	408k	{
271	408k	uint8_t i, j;
272
273	408k	DEBUGASSERT(set->init == CURL_U8_STRSET_MAGIC);
274	408k	if(u8_strset_get_index(set, id, &i)) {
275		/* `id` is in set */
276	2.70k	if(zero)
277	2.06k	curlx_strzero(set->data[i]);
278	2.70k	curlx_safefree(set->data[i]);
279	2.70k	set->ids[i] = set->psl[i] = 0;
280	2.70k	--set->count;
281	2.70k	j = CURL_U8_SET_SLOT_IDX(set, i + 1);
282		/* shift all entries with positive psl "down" */
283	4.27k	while(set->data[j] && set->psl[j]) {
284	1.56k	set->data[i] = set->data[j];
285	1.56k	set->ids[i] = set->ids[j];
286	1.56k	set->psl[i] = (uint8_t)(set->psl[j] - 1);
287	1.56k	set->data[j] = NULL;
288	1.56k	set->ids[j] = set->psl[j] = 0;
289	1.56k	i = j;
290	1.56k	j = CURL_U8_SET_SLOT_IDX(set, i + 1);
291	1.56k	}
292	2.70k	}
293	408k	}
294
295		void Curl_u8_strset_unset(struct u8_strset *set, uint8_t id)
296	3.47k	{
297	3.47k	u8_strset_unset(set, id, FALSE);
298	3.47k	}
299
300		void Curl_u8_strset_unset0(struct u8_strset *set, uint8_t id)
301	404k	{
302	404k	u8_strset_unset(set, id, TRUE);
303	404k	}
304
305		CURLcode Curl_u8_strset_copy(struct u8_strset dest, struct u8_strset src)
306	0	{
307	0	CURLcode result = CURLE_OK;
308	0	uint16_t i;
309
310	0	DEBUGASSERT(src->init == CURL_U8_STRSET_MAGIC);
311	0	Curl_u8_strset_clear(dest);
312	0	for(i = 0; !result && (i < CURL_U8_SET_SLOT_CNT(src)); ++i) {
313	0	if(src->data[i])
314	0	result = u8_strset_set(dest, src->ids[i], src->data[i]);
315	0	}
316	0	return result;
317	0	}