OPENSSL_LH_COMPFUNC(3ossl)

πŸ’– OPENSSL_LH_COMPFUNC πŸ’–

okay so bestie, we're diving into OPENSSL_LH_COMPFUNC(3ossl), which is part of OpenSSL, and it's the whole ✨dynamic hash table✨ saga. buckle up, grab a strawberry πŸ“, this one's a chunky one but it's honestly iconic once it clicks.


✨ NAME ✨

bestie, the entire vibe of this page is dynamic hash tables, and it goes by so many names. here's literally every one of them, no cap:

  • LHASH, LHASH_OF, DEFINE_LHASH_OF_EX, DEFINE_LHASH_OF
  • OPENSSL_LH_COMPFUNC, OPENSSL_LH_HASHFUNC, OPENSSL_LH_DOALL_FUNC, LHASH_DOALL_ARG_FN_TYPE
  • IMPLEMENT_LHASH_HASH_FN, IMPLEMENT_LHASH_COMP_FN
  • lh_TYPE_new, lh_TYPE_free, lh_TYPE_flush, lh_TYPE_insert, lh_TYPE_delete, lh_TYPE_retrieve, lh_TYPE_doall, lh_TYPE_doall_arg, lh_TYPE_num_items, lh_TYPE_get_down_load, lh_TYPE_set_down_load, lh_TYPE_error
  • OPENSSL_LH_new, OPENSSL_LH_free, OPENSSL_LH_flush, OPENSSL_LH_insert, OPENSSL_LH_delete, OPENSSL_LH_retrieve, OPENSSL_LH_doall, OPENSSL_LH_doall_arg, OPENSSL_LH_doall_arg_thunk, OPENSSL_LH_set_thunks, OPENSSL_LH_num_items, OPENSSL_LH_get_down_load, OPENSSL_LH_set_down_load, OPENSSL_LH_error

and the plot of the whole thing: dynamic hash table πŸ’…


🌈 SYNOPSIS 🌈

the way that we start with the include, always:

        #include <openssl/lhash.h>

        LHASH_OF(TYPE)

        DEFINE_LHASH_OF_EX(TYPE);

        LHASH_OF(TYPE) *lh_TYPE_new(OPENSSL_LH_HASHFUNC hash, OPENSSL_LH_COMPFUNC compare);
        void lh_TYPE_free(LHASH_OF(TYPE) *table);
        void lh_TYPE_flush(LHASH_OF(TYPE) *table);
        OPENSSL_LHASH *OPENSSL_LH_set_thunks(OPENSSL_LHASH *lh,
                                             OPENSSL_LH_HASHFUNCTHUNK hw,
                                             OPENSSL_LH_COMPFUNCTHUNK cw,
                                             OPENSSL_LH_DOALL_FUNC_THUNK daw,
                                             OPENSSL_LH_DOALL_FUNCARG_THUNK daaw)

        TYPE *lh_TYPE_insert(LHASH_OF(TYPE) *table, TYPE *data);
        TYPE *lh_TYPE_delete(LHASH_OF(TYPE) *table, TYPE *data);
        TYPE *lh_TYPE_retrieve(LHASH_OF(TYPE) *table, TYPE *data);

        void lh_TYPE_doall(LHASH_OF(TYPE) *table, OPENSSL_LH_DOALL_FUNC func);
        void lh_TYPE_doall_arg(LHASH_OF(TYPE) *table, OPENSSL_LH_DOALL_FUNCARG func,
                               TYPE *arg);
        void OPENSSL_LH_doall_arg_thunk(OPENSSL_LHASH *lh,
                                        OPENSSL_LH_DOALL_FUNCARG_THUNK daaw,
                                        OPENSSL_LH_DOALL_FUNCARG fn, void *arg)

        unsigned long lh_TYPE_num_items(OPENSSL_LHASH *lh);
        unsigned long lh_TYPE_get_down_load(OPENSSL_LHASH *lh);
        void lh_TYPE_set_down_load(OPENSSL_LHASH *lh, unsigned long dl);

        int lh_TYPE_error(LHASH_OF(TYPE) *table);

        typedef int (*OPENSSL_LH_COMPFUNC)(const void *, const void *);
        typedef unsigned long (*OPENSSL_LH_HASHFUNC)(const void *);
        typedef void (*OPENSSL_LH_DOALL_FUNC)(const void *);
        typedef void (*LHASH_DOALL_ARG_FN_TYPE)(const void *, const void *);

        OPENSSL_LHASH *OPENSSL_LH_new(OPENSSL_LH_HASHFUNC h, OPENSSL_LH_COMPFUNC c);
        void OPENSSL_LH_free(OPENSSL_LHASH *lh);
        void OPENSSL_LH_flush(OPENSSL_LHASH *lh);

        void *OPENSSL_LH_insert(OPENSSL_LHASH *lh, void *data);
        void *OPENSSL_LH_delete(OPENSSL_LHASH *lh, const void *data);
        void *OPENSSL_LH_retrieve(OPENSSL_LHASH *lh, const void *data);

        void OPENSSL_LH_doall(OPENSSL_LHASH *lh, OPENSSL_LH_DOALL_FUNC func);
        void OPENSSL_LH_doall_arg(OPENSSL_LHASH *lh, OPENSSL_LH_DOALL_FUNCARG func, void *arg);

        unsigned long OPENSSL_LH_num_items(OPENSSL_LHASH *lh);
        unsigned long OPENSSL_LH_get_down_load(OPENSSL_LHASH *lh);
        void OPENSSL_LH_set_down_load(OPENSSL_LHASH *lh, unsigned long dl);

        int OPENSSL_LH_error(OPENSSL_LHASH *lh);

        #define LH_LOAD_MULT   /* integer constant */

and then, right at the bottom, the deprecated one:

        DEFINE_LHASH_OF(TYPE);

πŸ’… DESCRIPTION πŸ’…

okay so here's the tea: this library implements type-checked dynamic hash tables. the entries in the table can be arbitrary structures β€” usually they're a key field plus a value field, the classic combo. and throughout this whole description, TYPE is just a placeholder standing in for any of the OpenSSL datatypes, like SSL_SESSION. think of it as the blank in a mad libs, but make it ✨cryptography✨.

to spin up a brand new type-checked dynamic hash table, you use DEFINE_LHASH_OF_EX(). back in the day DEFINE_LHASH_OF() was the one everybody used for this, but it's deprecated now, bestie, we don't do that anymore. DEFINE_LHASH_OF_EX() gives you all the functionality of DEFINE_LHASH_OF() except for certain deprecated statistics functions (peep OPENSSL_LH_stats(3) if you're curious).

lh_TYPE_new() creates a new LHASH_OF(TYPE) structure to hold arbitrary data entries, and it wants you to hand over the 'hash' and 'compare' callbacks that it'll use to organise everything. so:

  • the hash callback takes a pointer to a table entry as its argument and returns an unsigned long hash value for its key field. heads up tho: that hash value is normally truncated to a power of 2, so make sure your hash function returns well mixed low order bits β€” that part is fr fr important.
  • the compare callback takes two arguments (pointers to two hash table entries), and returns 0 if their keys are equal, nonzero otherwise.

if your hash table is gonna hold items of one particular type, and your hash and compare callbacks hash/compare exactly that type, then the IMPLEMENT_LHASH_HASH_FN and IMPLEMENT_LHASH_COMP_FN macros can build you callback wrappers with the prototypes that lh_TYPE_new() requires. it's giving effortless:

        /*
         * Implement the hash and compare functions; "stuff" can be any word.
         */
        static unsigned long stuff_hash(const TYPE *a)
        {
            ...
        }
        static int stuff_cmp(const TYPE *a, const TYPE *b)
        {
            ...
        }

        /*
         * Implement the wrapper functions.
         */
        static IMPLEMENT_LHASH_HASH_FN(stuff, TYPE)
        static IMPLEMENT_LHASH_COMP_FN(stuff, TYPE)

and if this type is gonna show up in a bunch of places, drop these macros into a common header file to declare the function wrappers:

        DECLARE_LHASH_HASH_FN(stuff, TYPE)
        DECLARE_LHASH_COMP_FN(stuff, TYPE)

then a hash table of TYPE objects gets created like this:

        LHASH_OF(TYPE) *htable;

        htable = B<lh_I<TYPE>_new>(LHASH_HASH_FN(stuff), LHASH_COMP_FN(stuff));

lh_TYPE_free() frees the LHASH_OF(TYPE) structure table. but careful β€” the allocated hash table entries themselves will not be freed. if you want those gone too, consider using lh_TYPE_doall() to deallocate whatever's still hanging out in there (more on that below πŸ‘‡). and if the argument is NULL, literally nothing happens, it just vibes.

lh_TYPE_flush() empties the LHASH_OF(TYPE) structure table. new entries can be added to the flushed table afterwards, it's not a permanent goodbye. but same deal as above: allocated hash table entries will not be freed, so consider lh_TYPE_doall() for cleanup duty (again, see below).

lh_TYPE_insert() inserts the structure pointed to by data into table. if there's already an entry with the same key, the old value gets replaced, no drama. and note that lh_TYPE_insert() stores pointers β€” the data are not copied. so don't go freeing that memory behind its back, bestie. πŸ™ˆ

lh_TYPE_delete() deletes an entry from table.

lh_TYPE_retrieve() looks up an entry in table. normally data is a structure with the key field(s) set, and the function hands you back a pointer to a fully populated structure. cute.

lh_TYPE_doall() will, for every single entry in the hash table, call func with the data item as its parameter. like so:

        /* Cleans up resources belonging to 'a' (this is implemented elsewhere) */
        void TYPE_cleanup_doall(TYPE *a);

        /* Implement a prototype-compatible wrapper for "TYPE_cleanup" */
        IMPLEMENT_LHASH_DOALL_FN(TYPE_cleanup, TYPE)

        /* Call "TYPE_cleanup" against all items in a hash table. */
        lh_TYPE_doall(hashtable, LHASH_DOALL_FN(TYPE_cleanup));

        /* Then the hash table itself can be deallocated */
        lh_TYPE_free(hashtable);

lh_TYPE_doall_arg() is the same as lh_TYPE_doall() except func gets called with arg as the second argument, and func should be of type LHASH_DOALL_ARG_FN(TYPE) β€” that's a callback prototype that receives both the table entry and an extra argument. and just like with lh_doall(), you can instead declare your callback with a prototype matching the types you're actually dealing with, then use the declare/implement macros to build compatible wrappers that cast variables before calling your type-specific callbacks. here's that whole thing demonstrated (printing every hash table entry to a BIO that the caller provides):

        /* Prints item 'a' to 'output_bio' (this is implemented elsewhere) */
        void TYPE_print_doall_arg(const TYPE *a, BIO *output_bio);

        /* Implement a prototype-compatible wrapper for "TYPE_print" */
        static IMPLEMENT_LHASH_DOALL_ARG_FN(TYPE, const TYPE, BIO)

        /* Print out the entire hashtable to a particular BIO */
        lh_TYPE_doall_arg(hashtable, LHASH_DOALL_ARG_FN(TYPE_print), BIO,
                          logging_bio);
⚠️ BIG WARNING BESTIE: it is by default not safe to use lh_TYPE_delete() inside a callback passed to lh_TYPE_doall() or lh_TYPE_doall_arg(). the reason is that deleting an item from the hash table may cause the hash table to be contracted to a smaller size and rehashed. lh_TYPE_doall() and lh_TYPE_doall_arg() are unsafe and will exhibit undefined behaviour under these conditions, because they assume the hash table size and bucket pointers don't change during the call. πŸ’€

if you really want to use lh_TYPE_doall() or lh_TYPE_doall_arg() together with lh_TYPE_delete(), then it is essential that you first call lh_TYPE_set_down_load() with a down_load argument of 0. that disables hash table contraction and guarantees it'll be safe to delete items from the hash table during a call to lh_TYPE_doall() or lh_TYPE_doall_arg(). 🫢

and it is never safe to call lh_TYPE_insert() during a call to lh_TYPE_doall() or lh_TYPE_doall_arg(). not once. not ever. periodt.

lh_TYPE_error() can be used to determine if an error occurred in the last operation.

lh_TYPE_num_items() returns the number of items in the hash table. iconic, we love a headcount.

lh_TYPE_get_down_load() and lh_TYPE_set_down_load() get and set the factor used to determine when the hash table is contracted. the factor is the load factor at or below which hash table contraction will occur, multiplied by LH_LOAD_MULT, where the load factor is the number of items divided by the number of nodes. setting this value to 0 disables hash table contraction. ✨ math ✨

OPENSSL_LH_new() is the same as lh_TYPE_new() except it is not type specific. so instead of returning an LHASH_OF(TYPE) value, it returns a void *. in the same way, the functions OPENSSL_LH_free(), OPENSSL_LH_flush(), OPENSSL_LH_insert(), OPENSSL_LH_delete(), OPENSSL_LH_retrieve(), OPENSSL_LH_doall(), OPENSSL_LH_doall_arg(), OPENSSL_LH_num_items(), OPENSSL_LH_get_down_load(), OPENSSL_LH_set_down_load() and OPENSSL_LH_error() are equivalent to the similarly named lh_TYPE functions, except they return or use a void * where the equivalent lh_TYPE function returns or uses a TYPE * or LHASH_OF(TYPE) *. the lh_TYPE functions are implemented as type checked wrappers around the OPENSSL_LH functions. and most applications should not call the OPENSSL_LH functions directly, bestie, stay in your type-checked lane.

OPENSSL_LH_set_thunks() and OPENSSL_LH_doall_arg_thunk(), while public by necessity, are actually internal functions and should not be used. that's it, that's the whole note. don't touch. πŸ™ˆ

✨ RETURN VALUES ✨

Okay bestie, here's the receipts section β€” who gives back what πŸ’…

  • lh_TYPE_new() and OPENSSL_LH_new() β€” hand you NULL if something went wrong πŸ™ˆ, otherwise a pointer to the shiny new LHASH structure. It's giving brand new car smell ✨
  • lh_TYPE_insert() and OPENSSL_LH_insert() β€” when a hash table entry gets replaced, these return the value that got replaced. You get NULL back on normal operation and on error. (Yeah, both. We'll get to that in BUGS, hold that thought 😭)
  • lh_TYPE_delete() and OPENSSL_LH_delete() β€” return the entry being deleted. If there's no such value in the hash table, you get NULL. Not me getting nothing back from a delete πŸ˜”
  • lh_TYPE_retrieve() and OPENSSL_LH_retrieve() β€” return the hash table entry if it's been found, NULL otherwise. Found it or flopped, no in-between πŸ¦‹
  • lh_TYPE_error() and OPENSSL_LH_error() β€” return 1 if an error occurred in the last operation, 0 otherwise. It's meaningful only after non-retrieve operations. So don't come for it after a retrieve, it doesn't know anything πŸ‘‘
  • lh_TYPE_free(), OPENSSL_LH_free(), lh_TYPE_flush(), OPENSSL_LH_flush(), lh_TYPE_doall(), OPENSSL_LH_doall(), lh_TYPE_doall_arg() and OPENSSL_LH_doall_arg() β€” return no values. Nothing. Zero. Empty handed but iconic for it πŸ’œ

🚨 NOTE 🚨

Buckle up, this is the part that actually matters πŸ”₯

🧡 Thread safety (or the lack thereof)

The LHASH code is not thread safe. Not even a little. No cap.

  • All updating operations, plus any calls to lh_TYPE_error() or OPENSSL_LH_error(), must be performed under a write lock. Periodt.
  • All retrieve operations should be performed under a read lock β€” unless accurate usage statistics are desired πŸ‘€ In which case, use a write lock for retrieve operations too.
  • For outputting the usage statistics, using the functions from OPENSSL_LH_stats(3), a read lock suffices. Slay, one thing that's easy πŸ’…

🧊 The whole "const" situation

The LHASH code regards table entries as constant data. As such, it internally represents lh_insert()'d items with a "const void *" pointer type. That's why callbacks like the ones used by lh_doall() and lh_doall_arg() declare their prototypes with "const" β€” even for the parameters that pass back the table items' data pointers. For consistency, user-provided data is "const" at all times as far as the LHASH code is concerned. πŸ’€

But β€” and this is the fun part β€” since callers are the ones providing these pointers, they can choose whether they too should be treating all such parameters as constant. Main character energy: it's your pointer, your rules πŸŽ€

Take this example: a hash table may be maintained by code that, for reasons of encapsulation, has only "const" access to the data being indexed in the hash table (i.e. it is returned as "const" from elsewhere in their code) β€” in this case the LHASH prototypes are appropriate as-is. Easy, no notes ✨

Conversely, if the caller is responsible for the life-time of the data in question, then they may well wish to make modifications to the table item passed back in the lh_doall() or lh_doall_arg() callbacks (see the "TYPE_cleanup" example above). If so, the caller can either:

  • cast the "const" away (if they're providing the raw callbacks themselves), or
  • use the macros to declare/implement the wrapper functions without "const" types.
Callers that only have "const" access to data they're indexing in a table, yet declare callbacks without constant types (or cast the "const" away themselves), are therefore creating their own risks/bugs without being encouraged to do so by the API. 🫢

On a related note β€” those auditing code should pay special attention to any instances of DECLARE/IMPLEMENT_LHASH_DOALL_[ARG_]_FN macros that provide types without any "const" qualifiers. Auditors, that's your vibe check πŸ‘‘


πŸ’€ BUGS πŸ’€

  • lh_TYPE_insert() and OPENSSL_LH_insert() return NULL both for success and error.

Yeah. Both. So if you're checking for NULL to detect failure... bestie, I'm so sorry 😭 That's the whole bug. That's it. It just does that.


🌈 SEE ALSO 🌈


πŸ“œ HISTORY πŸ“œ

  • In OpenSSL 1.0.0, the lhash interface was revamped for better type checking. Glow up era ✨
  • In OpenSSL 3.1, DEFINE_LHASH_OF_EX() was introduced and DEFINE_LHASH_OF() was deprecated. Out with the old πŸ¦‹
  • OPENSSL_LH_doall_arg_thunk() and OPENSSL_LH_set_thunks() were added in OpenSSL 3.3. Fresh drop πŸ”₯

Copyright 2000-2024 The OpenSSL Project Authors. All Rights Reserved. πŸ’–

Licensed under the Apache License 2.0 (the "License"). You may not use this file except in compliance with the License. You can obtain a copy in the file LICENSE in the source distribution or at <https://www.openssl.org/source/license.html>.


And that's the whole page, bestie πŸŽ€ This was OPENSSL_LH_COMPFUNC(3ossl) from OpenSSL 3.5.9, dated 2026-09-29 β€” go forth, lock your writes, and stay iconic πŸ’…βœ¨

πŸ“¦ where this page comes from