wbrawner/openssl
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
openssl/crypto/init.c

712 lines
20 KiB
C
Raw Normal View History

Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
/*
Copyright update of more files that have changed this year Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/5110)
2018-01-19 09:49:22 +00:00
* Copyright 2016-2018 The OpenSSL Project Authors. All Rights Reserved.
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
*
Following the license change, modify the boilerplates in crypto/ [skip ci] Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7827)
2018-12-06 13:03:01 +00:00
* Licensed under the Apache License 2.0 (the "License"). You may not use
Copyright consolidation 08/10 Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-05-17 18:51:34 +00:00
* 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
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
*/
Move e_os.h to be the very first include. cryptilib.h is the second. Reviewed-by: Andy Polyakov <appro@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4188)
2017-08-23 23:05:07 +00:00
#include "e_os.h"
Use "" not <> for internal/ includes Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4217)
2017-08-22 12:35:43 +00:00
#include "internal/cryptlib_int.h"
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
#include <openssl/err.h>
Use "" not <> for internal/ includes Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4217)
2017-08-22 12:35:43 +00:00
#include "internal/rand_int.h"
#include "internal/bio.h"
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
#include <openssl/evp.h>
Use "" not <> for internal/ includes Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4217)
2017-08-22 12:35:43 +00:00
#include "internal/evp_int.h"
#include "internal/conf.h"
#include "internal/async.h"
#include "internal/engine.h"
#include "internal/comp.h"
#include "internal/err.h"
#include "internal/err_int.h"
#include "internal/objects.h"
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
#include <stdlib.h>
Can't re-init after stop. Remoce DYANMIC once-init stuff. After the library is stopped, you can't restart it. Reviewed-by: Matt Caswell <matt@openssl.org>
2016-02-10 05:39:29 +00:00
#include <assert.h>
Use "" not <> for internal/ includes Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4217)
2017-08-22 12:35:43 +00:00
#include "internal/thread_once.h"
Include "internal/dso_conf.h" where needed and appropriate Reviewed-by: Andy Polyakov <appro@openssl.org> (Merged from https://github.com/openssl/openssl/pull/5733)
2018-03-23 00:05:23 +00:00
#include "internal/dso_conf.h"
Use "" not <> for internal/ includes Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/4217)
2017-08-22 12:35:43 +00:00
#include "internal/dso.h"
#include "internal/store.h"
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
#include <openssl/trace.h>
Can't re-init after stop. Remoce DYANMIC once-init stuff. After the library is stopped, you can't restart it. Reviewed-by: Matt Caswell <matt@openssl.org>
2016-02-10 05:39:29 +00:00
static int stopped = 0;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
Hide OPENSSL_INIT_SETTINGS. Make OPENSSL_INIT_SETTINGS an opaque structure. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-10 14:55:48 +00:00
typedef struct ossl_init_stop_st OPENSSL_INIT_STOP;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
struct ossl_init_stop_st {
void (*handler)(void);
OPENSSL_INIT_STOP *next;
};
static OPENSSL_INIT_STOP *stop_handlers = NULL;
Remove use of CRYPTO_LOCK_INIT in init code Swap the use of CRYPTO_LOCK_INIT in the init code to use the new threading API mechanism for locking. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-07 14:39:22 +00:00
static CRYPTO_RWLOCK *init_lock = NULL;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
static CRYPTO_ONCE base = CRYPTO_ONCE_STATIC_INIT;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
static int base_inited = 0;
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
DEFINE_RUN_ONCE_STATIC(ossl_init_base)
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
{
prevent endless recursion when trace API is used within OPENSSL_init_crypto() Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Yang <kaishen.yy@antfin.com> (Merged from https://github.com/openssl/openssl/pull/9567)
2019-08-10 14:51:12 +00:00
/* no need to init trace */
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "ossl_init_base: setting up stop handlers\n");
Add "random malloc failure" tooling Still needs to be documented, somehow/somewhere. The env var OPENSSL_MALLOC_FAILURES controls how often malloc/realloc should fail. It's a set of fields separated by semicolons. Each field is a count and optional percentage (separated by @) which defaults to 100. If count is zero then it lasts "forever." For example: 100;@25 means the first 100 allocations pass, then the rest have a 25% chance of failing until the program exits or crashes. If env var OPENSSL_MALLOC_FD parses as a positive integer, a record of all malloc "shouldfail" tests is written to that file descriptor. If a malloc will fail, and OPENSSL_NO_CRYPTO_MDEBUG_BACKTRACE is not set (platform specific), then a backtrace will be written to the descriptor when a malloc fails. This can be useful because a malloc may fail but not be checked, and problems will only occur later. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/1252)
2016-07-08 17:40:08 +00:00
#ifndef OPENSSL_NO_CRYPTO_MDEBUG
ossl_malloc_setup_failures();
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
#endif
Split thread intialisation and handling out of init.c We're going to need some of these functions in the FIPS module, but most of the rest of the code in init.c is not needed. Therefore we split it out. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9040)
2019-05-24 16:52:17 +00:00
Ensure the thread keys are always allocated in the same order Fixes: #5899 Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/5911)
2018-04-20 13:45:06 +00:00
if ((init_lock = CRYPTO_THREAD_lock_new()) == NULL)
goto err;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
OPENSSL_cpuid_setup();
Add a warning stipulating how things should be coded in ossl_init_base Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/1922)
2016-11-14 23:58:51 +00:00
Standardise the function naming conventions in initthread.c Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9040)
2019-06-17 14:16:36 +00:00
if (!ossl_init_thread())
Split thread intialisation and handling out of init.c We're going to need some of these functions in the FIPS module, but most of the rest of the code in init.c is not needed. Therefore we split it out. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9040)
2019-05-24 16:52:17 +00:00
return 0;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
base_inited = 1;
Ensure the thread keys are always allocated in the same order Fixes: #5899 Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/5911)
2018-04-20 13:45:06 +00:00
return 1;
err:
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "ossl_init_base failed!\n");
Ensure the thread keys are always allocated in the same order Fixes: #5899 Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/5911)
2018-04-20 13:45:06 +00:00
CRYPTO_THREAD_lock_free(init_lock);
init_lock = NULL;
Ensure that libcrypto and libssl do not unload until the process exits Because we use atexit() to cleanup after ourselves, this will cause a problem if we have been dynamically loaded and then unloaded again: the atexit() handler may no longer be there. Most modern atexit() implementations can handle this, however there are still difficulties if libssl gets unloaded before libcrypto, because of the atexit() callback that libcrypto makes to libssl. The most robust solution seems to be to ensure that libcrypto and libssl never unload. This is done by simply deliberately leaking a dlopen() reference to them. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-18 13:13:25 +00:00
Ensure the thread keys are always allocated in the same order Fixes: #5899 Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/5911)
2018-04-20 13:45:06 +00:00
return 0;
}
Implement OPENSSL_INIT_NO_ATEXIT Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-15 16:27:34 +00:00
static CRYPTO_ONCE register_atexit = CRYPTO_ONCE_STATIC_INIT;
Support _onexit() in preference to atexit() on Windows This enables cleanup to happen on DLL unload instead of at process exit. [extended tests] Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-16 17:26:23 +00:00
#if !defined(OPENSSL_SYS_UEFI) && defined(_WIN32)
static int win32atexit(void)
{
OPENSSL_cleanup();
return 0;
}
#endif
Implement OPENSSL_INIT_NO_ATEXIT Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-15 16:27:34 +00:00
DEFINE_RUN_ONCE_STATIC(ossl_init_register_atexit)
{
Support _onexit() in preference to atexit() on Windows This enables cleanup to happen on DLL unload instead of at process exit. [extended tests] Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-16 17:26:23 +00:00
#ifdef OPENSSL_INIT_DEBUG
Implement OPENSSL_INIT_NO_ATEXIT Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-15 16:27:34 +00:00
fprintf(stderr, "OPENSSL_INIT: ossl_init_register_atexit()\n");
Support _onexit() in preference to atexit() on Windows This enables cleanup to happen on DLL unload instead of at process exit. [extended tests] Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-16 17:26:23 +00:00
#endif
Implement OPENSSL_INIT_NO_ATEXIT Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-15 16:27:34 +00:00
#ifndef OPENSSL_SYS_UEFI
Support _onexit() in preference to atexit() on Windows This enables cleanup to happen on DLL unload instead of at process exit. [extended tests] Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-16 17:26:23 +00:00
# ifdef _WIN32
/* We use _onexit() in preference because it gets called on DLL unload */
if (_onexit(win32atexit) == NULL)
return 0;
# else
Implement OPENSSL_INIT_NO_ATEXIT Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-15 16:27:34 +00:00
if (atexit(OPENSSL_cleanup) != 0)
return 0;
Support _onexit() in preference to atexit() on Windows This enables cleanup to happen on DLL unload instead of at process exit. [extended tests] Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-16 17:26:23 +00:00
# endif
Implement OPENSSL_INIT_NO_ATEXIT Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-15 16:27:34 +00:00
#endif
return 1;
}
DEFINE_RUN_ONCE_STATIC_ALT(ossl_init_no_register_atexit,
ossl_init_register_atexit)
{
#ifdef OPENSSL_INIT_DEBUG
fprintf(stderr, "OPENSSL_INIT: ossl_init_no_register_atexit ok!\n");
#endif
/* Do nothing in this case */
return 1;
}
Ensure the thread keys are always allocated in the same order Fixes: #5899 Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/5911)
2018-04-20 13:45:06 +00:00
static CRYPTO_ONCE load_crypto_nodelete = CRYPTO_ONCE_STATIC_INIT;
DEFINE_RUN_ONCE_STATIC(ossl_init_load_crypto_nodelete)
{
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "ossl_init_load_crypto_nodelete()\n");
Rework DSO API conditions and configuration option 'no-dso' is meaningless, as it doesn't get any macro defined. Therefore, we remove all checks of OPENSSL_NO_DSO. However, there may be some odd platforms with no DSO scheme. For those, we generate the internal macro DSO_NONE aand use it. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/#8622)
2019-03-31 14:15:02 +00:00
#if !defined(OPENSSL_USE_NODELETE) \
Introduce a no-pinshared option This option prevents OpenSSL from pinning itself in memory. Fixes #7598 [extended tests] Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-16 14:05:14 +00:00
&& !defined(OPENSSL_NO_PINSHARED)
For the lack of GetModuleHandleEx(), we use DSO route for WinCE. Revert win32_pathbyaddr() which is used in DSO_dsobyaddr(). Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8596)
2019-03-27 08:30:47 +00:00
# if defined(DSO_WIN32) && !defined(_WIN32_WCE)
Implement GET_MODULE_HANDLE_EX_FLAG_PIN for windows Rather than leaking a reference, just call GetModuleHandleEx and pin the module on Windows. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-28 10:03:22 +00:00
{
HMODULE handle = NULL;
BOOL ret;
/* We don't use the DSO route for WIN32 because there is a better way */
ret = GetModuleHandleEx(GET_MODULE_HANDLE_EX_FLAG_FROM_ADDRESS
| GET_MODULE_HANDLE_EX_FLAG_PIN,
(void *)&base_inited, &handle);
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE1(INIT,
"ossl_init_load_crypto_nodelete: "
"obtained DSO reference? %s\n",
(ret == TRUE ? "No!" : "Yes."));
Implement GET_MODULE_HANDLE_EX_FLAG_PIN for windows Rather than leaking a reference, just call GetModuleHandleEx and pin the module on Windows. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-28 10:03:22 +00:00
return (ret == TRUE) ? 1 : 0;
}
Rework DSO API conditions and configuration option 'no-dso' is meaningless, as it doesn't get any macro defined. Therefore, we remove all checks of OPENSSL_NO_DSO. However, there may be some odd platforms with no DSO scheme. For those, we generate the internal macro DSO_NONE aand use it. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/#8622)
2019-03-31 14:15:02 +00:00
# elif !defined(DSO_NONE)
Ensure that libcrypto and libssl do not unload until the process exits Because we use atexit() to cleanup after ourselves, this will cause a problem if we have been dynamically loaded and then unloaded again: the atexit() handler may no longer be there. Most modern atexit() implementations can handle this, however there are still difficulties if libssl gets unloaded before libcrypto, because of the atexit() callback that libcrypto makes to libssl. The most robust solution seems to be to ensure that libcrypto and libssl never unload. This is done by simply deliberately leaking a dlopen() reference to them. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-18 13:13:25 +00:00
/*
* Deliberately leak a reference to ourselves. This will force the library
Don't leave stale errors on queue if DSO_dsobyaddr() fails The init code uses DSO_dsobyaddr() to leak a reference to ourselves to ensure we remain loaded until atexit() time. In some circumstances that can fail and leave stale errors on the error queue. Fixes #3372 Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3383)
2017-05-04 11:51:18 +00:00
* to remain loaded until the atexit() handler is run at process exit.
Ensure that libcrypto and libssl do not unload until the process exits Because we use atexit() to cleanup after ourselves, this will cause a problem if we have been dynamically loaded and then unloaded again: the atexit() handler may no longer be there. Most modern atexit() implementations can handle this, however there are still difficulties if libssl gets unloaded before libcrypto, because of the atexit() callback that libcrypto makes to libssl. The most robust solution seems to be to ensure that libcrypto and libssl never unload. This is done by simply deliberately leaking a dlopen() reference to them. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-18 13:13:25 +00:00
*/
{
Ensure the thread keys are always allocated in the same order Fixes: #5899 Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/5911)
2018-04-20 13:45:06 +00:00
DSO *dso;
void *err;
if (!err_shelve_state(&err))
return 0;
Ensure that libcrypto and libssl do not unload until the process exits Because we use atexit() to cleanup after ourselves, this will cause a problem if we have been dynamically loaded and then unloaded again: the atexit() handler may no longer be there. Most modern atexit() implementations can handle this, however there are still difficulties if libssl gets unloaded before libcrypto, because of the atexit() callback that libcrypto makes to libssl. The most robust solution seems to be to ensure that libcrypto and libssl never unload. This is done by simply deliberately leaking a dlopen() reference to them. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-18 13:13:25 +00:00
dso = DSO_dsobyaddr(&base_inited, DSO_FLAG_NO_UNLOAD_ON_FREE);
Add dladdr() for AIX Although it deviates from the actual prototype of DSO_dsobyaddr(), this is now ISO C compliant and gcc -Wpedantic accepts the code. Added DATA segment checking to catch ptrgl virtual addresses. Avoid memleaks with every AIX/dladdr() call. Removed debug-fprintf()s. Added test case for DSO_dsobyaddr(), which will eventually call dladdr(). Removed unecessary AIX ifdefs again. The implementation can only lookup function symbols, no data symbols. Added PIC-flag to aix*-cc build targets. As AIX is missing a dladdr() implementation it is currently uncertain our exit()-handlers can still be called when the application exits. After dlclose() the whole library might have been unloaded already. Signed-off-by: Matthias Kraft <makr@gmx.eu> Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Rich Salz <rsalz@openssl.org> (Merged from https://github.com/openssl/openssl/pull/5668)
2018-03-19 17:37:46 +00:00
/*
* In case of No!, it is uncertain our exit()-handlers can still be
* called. After dlclose() the whole library might have been unloaded
* already.
*/
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE1(INIT, "obtained DSO reference? %s\n",
(dso == NULL ? "No!" : "Yes."));
Ensure that libcrypto and libssl do not unload until the process exits Because we use atexit() to cleanup after ourselves, this will cause a problem if we have been dynamically loaded and then unloaded again: the atexit() handler may no longer be there. Most modern atexit() implementations can handle this, however there are still difficulties if libssl gets unloaded before libcrypto, because of the atexit() callback that libcrypto makes to libssl. The most robust solution seems to be to ensure that libcrypto and libssl never unload. This is done by simply deliberately leaking a dlopen() reference to them. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-18 13:13:25 +00:00
DSO_free(dso);
Ensure the thread keys are always allocated in the same order Fixes: #5899 Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/5911)
2018-04-20 13:45:06 +00:00
err_unshelve_state(err);
Ensure that libcrypto and libssl do not unload until the process exits Because we use atexit() to cleanup after ourselves, this will cause a problem if we have been dynamically loaded and then unloaded again: the atexit() handler may no longer be there. Most modern atexit() implementations can handle this, however there are still difficulties if libssl gets unloaded before libcrypto, because of the atexit() callback that libcrypto makes to libssl. The most robust solution seems to be to ensure that libcrypto and libssl never unload. This is done by simply deliberately leaking a dlopen() reference to them. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-18 13:13:25 +00:00
}
Implement GET_MODULE_HANDLE_EX_FLAG_PIN for windows Rather than leaking a reference, just call GetModuleHandleEx and pin the module on Windows. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-28 10:03:22 +00:00
# endif
Link using -znodelete Instead of deliberately leaking a reference to ourselves, use nodelete which does this more neatly. Only for Linux at the moment. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-18 14:11:57 +00:00
#endif
Ensure that libcrypto and libssl do not unload until the process exits Because we use atexit() to cleanup after ourselves, this will cause a problem if we have been dynamically loaded and then unloaded again: the atexit() handler may no longer be there. Most modern atexit() implementations can handle this, however there are still difficulties if libssl gets unloaded before libcrypto, because of the atexit() callback that libcrypto makes to libssl. The most robust solution seems to be to ensure that libcrypto and libssl never unload. This is done by simply deliberately leaking a dlopen() reference to them. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-18 13:13:25 +00:00
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
return 1;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
}
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
static CRYPTO_ONCE load_crypto_strings = CRYPTO_ONCE_STATIC_INIT;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
static int load_crypto_strings_inited = 0;
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
DEFINE_RUN_ONCE_STATIC(ossl_init_load_crypto_strings)
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
{
Check for errors allocating the error strings. Reviewed-by: Richard Levitte <levitte@openssl.org> GH: #1330
2016-07-12 13:50:06 +00:00
int ret = 1;
Add an OPENSSL_NO_AUTOERRINIT option This option disables automatic loading of the crypto/ssl error strings in order to keep statically linked executable file size down Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 09:39:07 +00:00
/*
* OPENSSL_NO_AUTOERRINIT is provided here to prevent at compile time
* pulling in all the error strings during static linking
*/
#if !defined(OPENSSL_NO_ERR) && !defined(OPENSSL_NO_AUTOERRINIT)
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "err_load_crypto_strings_int()\n");
Check for errors allocating the error strings. Reviewed-by: Richard Levitte <levitte@openssl.org> GH: #1330
2016-07-12 13:50:06 +00:00
ret = err_load_crypto_strings_int();
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
load_crypto_strings_inited = 1;
Fix a bundle of trailing spaces in several files Signed-off-by: Paul Yang <paulyang.inf@gmail.com> Reviewed-by: Kurt Roeckx <kurt@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> Reviewed-by: Rich Salz <rsalz@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3622)
2017-06-06 15:35:43 +00:00
#endif
Check for errors allocating the error strings. Reviewed-by: Richard Levitte <levitte@openssl.org> GH: #1330
2016-07-12 13:50:06 +00:00
return ret;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
}
Fix a RUN_ONCE bug We have a number of instances where there are multiple "init" functions for a single CRYPTO_ONCE variable, e.g. to load config automatically or to not load config automatically. Unfortunately the RUN_ONCE mechanism was not correctly giving the right return value where an alternative init function was being used. Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-20 15:32:55 +00:00
DEFINE_RUN_ONCE_STATIC_ALT(ossl_init_no_load_crypto_strings,
ossl_init_load_crypto_strings)
{
/* Do nothing in this case */
return 1;
}
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
static CRYPTO_ONCE add_all_ciphers = CRYPTO_ONCE_STATIC_INIT;
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
DEFINE_RUN_ONCE_STATIC(ossl_init_add_all_ciphers)
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
{
/*
* OPENSSL_NO_AUTOALGINIT is provided here to prevent at compile time
* pulling in all the ciphers during static linking
*/
#ifndef OPENSSL_NO_AUTOALGINIT
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "openssl_add_all_ciphers_int()\n");
Rename int_*() functions to *_int() There is a preference for suffixes to indicate that a function is internal rather than prefixes. Note: the suffix is only required to disambiguate internal functions and public symbols with the same name (but different case) Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-04-12 11:20:16 +00:00
openssl_add_all_ciphers_int();
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
#endif
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
return 1;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
}
Fix a RUN_ONCE bug We have a number of instances where there are multiple "init" functions for a single CRYPTO_ONCE variable, e.g. to load config automatically or to not load config automatically. Unfortunately the RUN_ONCE mechanism was not correctly giving the right return value where an alternative init function was being used. Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-20 15:32:55 +00:00
DEFINE_RUN_ONCE_STATIC_ALT(ossl_init_no_add_all_ciphers,
ossl_init_add_all_ciphers)
{
/* Do nothing */
return 1;
}
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
static CRYPTO_ONCE add_all_digests = CRYPTO_ONCE_STATIC_INIT;
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
DEFINE_RUN_ONCE_STATIC(ossl_init_add_all_digests)
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
{
/*
* OPENSSL_NO_AUTOALGINIT is provided here to prevent at compile time
* pulling in all the ciphers during static linking
*/
#ifndef OPENSSL_NO_AUTOALGINIT
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "openssl_add_all_digests()\n");
Rename int_*() functions to *_int() There is a preference for suffixes to indicate that a function is internal rather than prefixes. Note: the suffix is only required to disambiguate internal functions and public symbols with the same name (but different case) Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-04-12 11:20:16 +00:00
openssl_add_all_digests_int();
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
#endif
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
return 1;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
}
Fix a RUN_ONCE bug We have a number of instances where there are multiple "init" functions for a single CRYPTO_ONCE variable, e.g. to load config automatically or to not load config automatically. Unfortunately the RUN_ONCE mechanism was not correctly giving the right return value where an alternative init function was being used. Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-20 15:32:55 +00:00
DEFINE_RUN_ONCE_STATIC_ALT(ossl_init_no_add_all_digests,
ossl_init_add_all_digests)
{
/* Do nothing */
return 1;
}
Added EVP_KDF (similiar to the EVP_MAC) Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8808)
2019-04-22 07:18:56 +00:00
static CRYPTO_ONCE add_all_kdfs = CRYPTO_ONCE_STATIC_INIT;
DEFINE_RUN_ONCE_STATIC(ossl_init_add_all_kdfs)
{
/*
* OPENSSL_NO_AUTOALGINIT is provided here to prevent at compile time
* pulling in all the macs during static linking
*/
#ifndef OPENSSL_NO_AUTOALGINIT
OSSL_TRACE(INIT, "openssl_add_all_kdfs_int()\n");
openssl_add_all_kdfs_int();
#endif
return 1;
}
DEFINE_RUN_ONCE_STATIC_ALT(ossl_init_no_add_all_kdfs, ossl_init_add_all_kdfs)
{
/* Do nothing */
return 1;
}
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
static CRYPTO_ONCE config = CRYPTO_ONCE_STATIC_INIT;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
static int config_inited = 0;
More configurable crypto and ssl library initialization 1. In addition to overriding the default application name, one can now also override the configuration file name and flags passed to CONF_modules_load_file(). 2. By default we still keep going when configuration file processing fails. But, applications that want to be strict about initialization errors can now make explicit flag choices via non-null OPENSSL_INIT_SETTINGS that omit the CONF_MFLAGS_IGNORE_RETURN_CODES flag (which had so far been both undocumented and unused). 3. In OPENSSL_init_ssl() do not request OPENSSL_INIT_LOAD_CONFIG if the options already include OPENSSL_INIT_NO_LOAD_CONFIG. 4. Don't set up atexit() handlers when called with INIT_BASE_ONLY. Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7986)
2019-01-01 07:53:24 +00:00
static const OPENSSL_INIT_SETTINGS *conf_settings = NULL;
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
DEFINE_RUN_ONCE_STATIC(ossl_init_config)
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
{
More configurable crypto and ssl library initialization 1. In addition to overriding the default application name, one can now also override the configuration file name and flags passed to CONF_modules_load_file(). 2. By default we still keep going when configuration file processing fails. But, applications that want to be strict about initialization errors can now make explicit flag choices via non-null OPENSSL_INIT_SETTINGS that omit the CONF_MFLAGS_IGNORE_RETURN_CODES flag (which had so far been both undocumented and unused). 3. In OPENSSL_init_ssl() do not request OPENSSL_INIT_LOAD_CONFIG if the options already include OPENSSL_INIT_NO_LOAD_CONFIG. 4. Don't set up atexit() handlers when called with INIT_BASE_ONLY. Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7986)
2019-01-01 07:53:24 +00:00
int ret = openssl_config_int(conf_settings);
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
config_inited = 1;
More configurable crypto and ssl library initialization 1. In addition to overriding the default application name, one can now also override the configuration file name and flags passed to CONF_modules_load_file(). 2. By default we still keep going when configuration file processing fails. But, applications that want to be strict about initialization errors can now make explicit flag choices via non-null OPENSSL_INIT_SETTINGS that omit the CONF_MFLAGS_IGNORE_RETURN_CODES flag (which had so far been both undocumented and unused). 3. In OPENSSL_init_ssl() do not request OPENSSL_INIT_LOAD_CONFIG if the options already include OPENSSL_INIT_NO_LOAD_CONFIG. 4. Don't set up atexit() handlers when called with INIT_BASE_ONLY. Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7986)
2019-01-01 07:53:24 +00:00
return ret;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
}
Fix a RUN_ONCE bug We have a number of instances where there are multiple "init" functions for a single CRYPTO_ONCE variable, e.g. to load config automatically or to not load config automatically. Unfortunately the RUN_ONCE mechanism was not correctly giving the right return value where an alternative init function was being used. Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-20 15:32:55 +00:00
DEFINE_RUN_ONCE_STATIC_ALT(ossl_init_no_config, ossl_init_config)
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
{
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "openssl_no_config_int()\n");
Rename int_*() functions to *_int() There is a preference for suffixes to indicate that a function is internal rather than prefixes. Note: the suffix is only required to disambiguate internal functions and public symbols with the same name (but different case) Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-04-12 11:20:16 +00:00
openssl_no_config_int();
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
config_inited = 1;
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
return 1;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
}
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
static CRYPTO_ONCE async = CRYPTO_ONCE_STATIC_INIT;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
static int async_inited = 0;
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
DEFINE_RUN_ONCE_STATIC(ossl_init_async)
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
{
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "async_init()\n");
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
if (!async_init())
return 0;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
async_inited = 1;
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
return 1;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
}
#ifndef OPENSSL_NO_ENGINE
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
static CRYPTO_ONCE engine_openssl = CRYPTO_ONCE_STATIC_INIT;
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
DEFINE_RUN_ONCE_STATIC(ossl_init_engine_openssl)
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
{
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "engine_load_openssl_int()\n");
Rename int_*() functions to *_int() There is a preference for suffixes to indicate that a function is internal rather than prefixes. Note: the suffix is only required to disambiguate internal functions and public symbols with the same name (but different case) Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-04-12 11:20:16 +00:00
engine_load_openssl_int();
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
return 1;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
}
# ifndef OPENSSL_NO_RDRAND
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
static CRYPTO_ONCE engine_rdrand = CRYPTO_ONCE_STATIC_INIT;
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
DEFINE_RUN_ONCE_STATIC(ossl_init_engine_rdrand)
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
{
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "engine_load_rdrand_int()\n");
Rename int_*() functions to *_int() There is a preference for suffixes to indicate that a function is internal rather than prefixes. Note: the suffix is only required to disambiguate internal functions and public symbols with the same name (but different case) Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-04-12 11:20:16 +00:00
engine_load_rdrand_int();
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
return 1;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
}
# endif
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
static CRYPTO_ONCE engine_dynamic = CRYPTO_ONCE_STATIC_INIT;
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
DEFINE_RUN_ONCE_STATIC(ossl_init_engine_dynamic)
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
{
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "engine_load_dynamic_int()\n");
Rename int_*() functions to *_int() There is a preference for suffixes to indicate that a function is internal rather than prefixes. Note: the suffix is only required to disambiguate internal functions and public symbols with the same name (but different case) Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-04-12 11:20:16 +00:00
engine_load_dynamic_int();
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
return 1;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
}
# ifndef OPENSSL_NO_STATIC_ENGINE
e_devcrypto: make the /dev/crypto engine dynamic Engine has been moved from crypto/engine/eng_devcrypto.c to engines/e_devcrypto.c. Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7859)
2018-11-06 12:57:03 +00:00
# ifndef OPENSSL_NO_DEVCRYPTOENG
static CRYPTO_ONCE engine_devcrypto = CRYPTO_ONCE_STATIC_INIT;
DEFINE_RUN_ONCE_STATIC(ossl_init_engine_devcrypto)
{
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "engine_load_devcrypto_int()\n");
e_devcrypto: make the /dev/crypto engine dynamic Engine has been moved from crypto/engine/eng_devcrypto.c to engines/e_devcrypto.c. Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7859)
2018-11-06 12:57:03 +00:00
engine_load_devcrypto_int();
return 1;
}
# endif
Deprecate the "hw" configuration options, make "padlockeng" disablable The "hw" and "hw-.*" style options are historical artifacts, sprung from the time when ENGINE was first designed, with hardware crypto accelerators and HSMs in mind. Today, these options have largely lost their value, replaced by options such as "no-{foo}eng" and "no-engine". This completes the transition by making "hw" and "hw-.*" deprecated, but automatically translated into more modern variants of the same. In the process, we get rid of the last regular expression in Configure's @disablables, a feature that was ill supported anyway. Also, padlock now gets treated just as every other engine. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8380)
2019-03-01 12:18:11 +00:00
# if !defined(OPENSSL_NO_PADLOCKENG)
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
static CRYPTO_ONCE engine_padlock = CRYPTO_ONCE_STATIC_INIT;
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
DEFINE_RUN_ONCE_STATIC(ossl_init_engine_padlock)
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
{
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "engine_load_padlock_int()\n");
Rename int_*() functions to *_int() There is a preference for suffixes to indicate that a function is internal rather than prefixes. Note: the suffix is only required to disambiguate internal functions and public symbols with the same name (but different case) Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-04-12 11:20:16 +00:00
engine_load_padlock_int();
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
return 1;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
}
# endif
# if defined(OPENSSL_SYS_WIN32) && !defined(OPENSSL_NO_CAPIENG)
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
static CRYPTO_ONCE engine_capi = CRYPTO_ONCE_STATIC_INIT;
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
DEFINE_RUN_ONCE_STATIC(ossl_init_engine_capi)
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
{
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "engine_load_capi_int()\n");
Rename int_*() functions to *_int() There is a preference for suffixes to indicate that a function is internal rather than prefixes. Note: the suffix is only required to disambiguate internal functions and public symbols with the same name (but different case) Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-04-12 11:20:16 +00:00
engine_load_capi_int();
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
return 1;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
}
# endif
Rework based on feedback: 1. Cleaned up eventfd handling 2. Reworked socket setup code to allow other algorithms to be added in future 3. Fixed compile errors for static build 4. Added error to error stack in all cases of ALG_PERR/ALG_ERR 5. Called afalg_aes_128_cbc() from bind() to avoid race conditions 6. Used MAX_INFLIGHT define in io_getevents system call 7. Coding style fixes Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org>
2016-02-23 08:01:01 +00:00
# if !defined(OPENSSL_NO_AFALGENG)
Small typo OPENSSL_INIT_ONCE and OPENSSL_INIT_ONCE_STATIC_INIT are really CRYPTO_ONCE and CRYPTO_ONCE_STATIC_INIT. Reviewed-by: Rich Salz <rsalz@openssl.org>
2016-03-15 13:06:34 +00:00
static CRYPTO_ONCE engine_afalg = CRYPTO_ONCE_STATIC_INIT;
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
DEFINE_RUN_ONCE_STATIC(ossl_init_engine_afalg)
Rework based on feedback: 1. Cleaned up eventfd handling 2. Reworked socket setup code to allow other algorithms to be added in future 3. Fixed compile errors for static build 4. Added error to error stack in all cases of ALG_PERR/ALG_ERR 5. Called afalg_aes_128_cbc() from bind() to avoid race conditions 6. Used MAX_INFLIGHT define in io_getevents system call 7. Coding style fixes Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org>
2016-02-23 08:01:01 +00:00
{
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "engine_load_afalg_int()\n");
Rename int_*() functions to *_int() There is a preference for suffixes to indicate that a function is internal rather than prefixes. Note: the suffix is only required to disambiguate internal functions and public symbols with the same name (but different case) Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-04-12 11:20:16 +00:00
engine_load_afalg_int();
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
return 1;
Rework based on feedback: 1. Cleaned up eventfd handling 2. Reworked socket setup code to allow other algorithms to be added in future 3. Fixed compile errors for static build 4. Added error to error stack in all cases of ALG_PERR/ALG_ERR 5. Called afalg_aes_128_cbc() from bind() to avoid race conditions 6. Used MAX_INFLIGHT define in io_getevents system call 7. Coding style fixes Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org>
2016-02-23 08:01:01 +00:00
}
# endif
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
# endif
#endif
Fix no-comp build Reviewed-by: Matt Caswell <matt@openssl.org>
2016-03-11 20:13:19 +00:00
#ifndef OPENSSL_NO_COMP
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
static CRYPTO_ONCE zlib = CRYPTO_ONCE_STATIC_INIT;
Fix no-comp build Reviewed-by: Matt Caswell <matt@openssl.org>
2016-03-11 20:13:19 +00:00
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
static int zlib_inited = 0;
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
DEFINE_RUN_ONCE_STATIC(ossl_init_zlib)
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
{
/* Do nothing - we need to know about this for the later cleanup */
zlib_inited = 1;
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
return 1;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
}
Fix no-comp build Reviewed-by: Matt Caswell <matt@openssl.org>
2016-03-11 20:13:19 +00:00
#endif
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
Rename INIT funtions, deprecate old ones. Man, there were a lot of renamings :) Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 16:52:40 +00:00
void OPENSSL_cleanup(void)
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
{
OPENSSL_INIT_STOP *currhandler, *lasthandler;
Add some TODO notes into init.c We should be seeking to move the OPENSSL_init_crypto and OPENSSL_cleanup processing into OPENSSL_CTX instead. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8857)
2019-05-02 12:42:31 +00:00
/*
* TODO(3.0): This function needs looking at with a view to moving most/all
* of this into onfree handlers in OPENSSL_CTX.
*/
If we've not been inited don't deinit If you call an explicit deinit when we've not been inited then a seg fault can occur. We should check that we've been inited before attempting to deinit. Reviewed-by: Rich Salz <rsalz@openssl.org>
2016-02-10 09:47:51 +00:00
/* If we've not been inited then no need to deinit */
if (!base_inited)
return;
Can't re-init after stop. Remoce DYANMIC once-init stuff. After the library is stopped, you can't restart it. Reviewed-by: Matt Caswell <matt@openssl.org>
2016-02-10 05:39:29 +00:00
/* Might be explicitly called and also by atexit */
if (stopped)
return;
stopped = 1;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
/*
* Thread stop may not get automatically called by the thread library for
* the very last thread in some situations, so call it directly.
*/
Split thread intialisation and handling out of init.c We're going to need some of these functions in the FIPS module, but most of the rest of the code in init.c is not needed. Therefore we split it out. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9040)
2019-05-24 16:52:17 +00:00
OPENSSL_thread_stop();
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
currhandler = stop_handlers;
while (currhandler != NULL) {
currhandler->handler();
lasthandler = currhandler;
currhandler = currhandler->next;
OPENSSL_free(lasthandler);
}
stop_handlers = NULL;
Remove use of CRYPTO_LOCK_INIT in init code Swap the use of CRYPTO_LOCK_INIT in the init code to use the new threading API mechanism for locking. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-07 14:39:22 +00:00
CRYPTO_THREAD_lock_free(init_lock);
Restore clearing of init_lock after free The behavior of resetting the init_lock value to NULL after freeing it during OPENSSL_cleanup() was added as part of the global lock commits that were just reverted, but there is desire to retain this behavior for clarity. It is unclear that the library would actually remain usable in any form after OPENSSL_cleanup(), since the required re-initialization occurs under a CRYPTO_ONCE check that cannot be reset at cleanup time. That said, a NULL dereference is probably more friendly behavior in these treacherous waters than using freed memory would be. Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> (Merged from https://github.com/openssl/openssl/pull/5089)
2018-01-26 15:32:40 +00:00
init_lock = NULL;
Remove use of CRYPTO_LOCK_INIT in init code Swap the use of CRYPTO_LOCK_INIT in the init code to use the new threading API mechanism for locking. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-07 14:39:22 +00:00
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
/*
* We assume we are single-threaded for this function, i.e. no race
* conditions for the various "*_inited" vars below.
*/
Fix no-comp build Reviewed-by: Matt Caswell <matt@openssl.org>
2016-03-11 20:13:19 +00:00
#ifndef OPENSSL_NO_COMP
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
if (zlib_inited) {
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "OPENSSL_cleanup: comp_zlib_cleanup_int()\n");
Rename int_*() functions to *_int() There is a preference for suffixes to indicate that a function is internal rather than prefixes. Note: the suffix is only required to disambiguate internal functions and public symbols with the same name (but different case) Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-04-12 11:20:16 +00:00
comp_zlib_cleanup_int();
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
}
Fix no-comp build Reviewed-by: Matt Caswell <matt@openssl.org>
2016-03-11 20:13:19 +00:00
#endif
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
Ensure Async is deinited properly The global thread local keys were not being deinited properly in async. Reviewed-by: Rich Salz <rsalz@openssl.org>
2016-03-02 16:52:43 +00:00
if (async_inited) {
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "OPENSSL_cleanup: async_deinit()\n");
Ensure Async is deinited properly The global thread local keys were not being deinited properly in async. Reviewed-by: Rich Salz <rsalz@openssl.org>
2016-03-02 16:52:43 +00:00
async_deinit();
}
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
if (load_crypto_strings_inited) {
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "OPENSSL_cleanup: err_free_strings_int()\n");
Rename int_*() functions to *_int() There is a preference for suffixes to indicate that a function is internal rather than prefixes. Note: the suffix is only required to disambiguate internal functions and public symbols with the same name (but different case) Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-04-12 11:20:16 +00:00
err_free_strings_int();
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
}
Fix the init cleanup order There are internal dependencies between the various cleanup functions. This re-orders things to try and get that right. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-14 10:34:59 +00:00
/*
* Note that cleanup order is important:
Remove check_defer() The check_defer() function was used to ensure that EVP_cleanup() was always called before OBJ_cleanup(). The new cleanup code ensures this so it is no longer needed. Remove obj_cleanup() call in OID config module: it is not needed any more either. Reviewed-by: Matt Caswell <matt@openssl.org>
2016-04-13 11:11:59 +00:00
* - rand_cleanup_int could call an ENGINE's RAND cleanup function so
Rename int_*() functions to *_int() There is a preference for suffixes to indicate that a function is internal rather than prefixes. Note: the suffix is only required to disambiguate internal functions and public symbols with the same name (but different case) Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-04-12 11:20:16 +00:00
* must be called before engine_cleanup_int()
Fix the init cleanup order There are internal dependencies between the various cleanup functions. This re-orders things to try and get that right. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-14 10:34:59 +00:00
* - ENGINEs use CRYPTO_EX_DATA and therefore, must be cleaned up
Instead of global data store it in an OPENSSL_CTX Various core and property related code files used global data. We should store all of that in an OPENSSL_CTX instead. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8857)
2019-05-01 10:02:43 +00:00
* before the ex data handlers are wiped during default openssl_ctx deinit.
Rename int_*() functions to *_int() There is a preference for suffixes to indicate that a function is internal rather than prefixes. Note: the suffix is only required to disambiguate internal functions and public symbols with the same name (but different case) Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-04-12 11:20:16 +00:00
* - conf_modules_free_int() can end up in ENGINE code so must be called
* before engine_cleanup_int()
Remove check_defer() The check_defer() function was used to ensure that EVP_cleanup() was always called before OBJ_cleanup(). The new cleanup code ensures this so it is no longer needed. Remove obj_cleanup() call in OID config module: it is not needed any more either. Reviewed-by: Matt Caswell <matt@openssl.org>
2016-04-13 11:11:59 +00:00
* - ENGINEs and additional EVP algorithms might use added OIDs names so
* obj_cleanup_int() must be called last
Fix the init cleanup order There are internal dependencies between the various cleanup functions. This re-orders things to try and get that right. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-14 10:34:59 +00:00
*/
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "OPENSSL_cleanup: rand_cleanup_int()\n");
Rename int_*() functions to *_int() There is a preference for suffixes to indicate that a function is internal rather than prefixes. Note: the suffix is only required to disambiguate internal functions and public symbols with the same name (but different case) Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-04-12 11:20:16 +00:00
rand_cleanup_int();
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "OPENSSL_cleanup: conf_modules_free_int()\n");
Rename int_*() functions to *_int() There is a preference for suffixes to indicate that a function is internal rather than prefixes. Note: the suffix is only required to disambiguate internal functions and public symbols with the same name (but different case) Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-04-12 11:20:16 +00:00
conf_modules_free_int();
Instead of global data store it in an OPENSSL_CTX Various core and property related code files used global data. We should store all of that in an OPENSSL_CTX instead. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8857)
2019-05-01 10:02:43 +00:00
Call CONF_modules_free() before ENGINE_cleanup() in auto-deinit During auto de-init we were calling ENGINE_cleanup(), and then later CONF_modules_free(). However the latter function can end up calling engine code, which can lead to a use of the global_engine_lock after it has already been freed. Therefore we should swap the calling order of these two functions. Reviewed-by: Rich Salz <rsalz@openssl.org>
2016-03-11 09:52:52 +00:00
#ifndef OPENSSL_NO_ENGINE
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "OPENSSL_cleanup: engine_cleanup_int()\n");
Rename int_*() functions to *_int() There is a preference for suffixes to indicate that a function is internal rather than prefixes. Note: the suffix is only required to disambiguate internal functions and public symbols with the same name (but different case) Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-04-12 11:20:16 +00:00
engine_cleanup_int();
Call CONF_modules_free() before ENGINE_cleanup() in auto-deinit During auto de-init we were calling ENGINE_cleanup(), and then later CONF_modules_free(). However the latter function can end up calling engine code, which can lead to a use of the global_engine_lock after it has already been freed. Therefore we should swap the calling order of these two functions. Reviewed-by: Rich Salz <rsalz@openssl.org>
2016-03-11 09:52:52 +00:00
#endif
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "OPENSSL_cleanup: ossl_store_cleanup_int()\n");
Add the STORE module This STORE module adds the following functionality: - A function OSSL_STORE_open(), OSSL_STORE_load() and OSSL_STORE_close() that accesses a URI and helps loading the supported objects (PKEYs, CERTs and CRLs for the moment) from it. - An opaque type OSSL_STORE_INFO that holds information on each loaded object. - A few functions to retrieve desired data from a OSSL_STORE_INFO reference. - Functions to register and unregister loaders for different URI schemes. This enables dynamic addition of loaders from applications or from engines. Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3542)
2016-12-08 14:51:31 +00:00
ossl_store_cleanup_int();
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
Instead of global data store it in an OPENSSL_CTX Various core and property related code files used global data. We should store all of that in an OPENSSL_CTX instead. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8857)
2019-05-01 10:02:43 +00:00
OSSL_TRACE(INIT, "OPENSSL_cleanup: openssl_ctx_default_deinit()\n");
openssl_ctx_default_deinit();
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
Provide an ability to deregister thread stop handlers If a provider gets unloaded then any thread stop handlers that it had registered will be left hanging. We should clean them up before tearing down the provider. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9186)
2019-06-18 17:37:38 +00:00
ossl_cleanup_thread();
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "OPENSSL_cleanup: bio_cleanup()\n");
Ensure all locks are properly cleaned up Some locks were not being properly cleaned up during close down. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-11 21:53:18 +00:00
bio_cleanup();
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "OPENSSL_cleanup: evp_cleanup_int()\n");
Rename int_*() functions to *_int() There is a preference for suffixes to indicate that a function is internal rather than prefixes. Note: the suffix is only required to disambiguate internal functions and public symbols with the same name (but different case) Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-04-12 11:20:16 +00:00
evp_cleanup_int();
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "OPENSSL_cleanup: obj_cleanup_int()\n");
Rename int_*() functions to *_int() There is a preference for suffixes to indicate that a function is internal rather than prefixes. Note: the suffix is only required to disambiguate internal functions and public symbols with the same name (but different case) Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-04-12 11:20:16 +00:00
obj_cleanup_int();
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "OPENSSL_cleanup: err_int()\n");
Ensure all locks are properly cleaned up Some locks were not being properly cleaned up during close down. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-11 21:53:18 +00:00
err_cleanup();
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "OPENSSL_cleanup: CRYPTO_secure_malloc_done()\n");
OPENSSL_cleanup: cleanup secure memory If the global DRBGs are allocated on the secure heap, then calling CRYPTO_secure_malloc_done() inside main() will have no effect, unless OPENSSL_cleanup() has been called explicitely before that, because otherwise the DRBGs will still be allocated. So it is better to cleanup the secure heap automatically at the end of OPENSSL_cleanup(). Reviewed-by: Rich Salz <rsalz@openssl.org> (Merged from https://github.com/openssl/openssl/pull/5294)
2018-02-12 01:37:27 +00:00
CRYPTO_secure_malloc_done();
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE(INIT, "OPENSSL_cleanup: ossl_trace_cleanup()\n");
ossl_trace_cleanup();
If we've not been inited don't deinit If you call an explicit deinit when we've not been inited then a seg fault can occur. We should check that we've been inited before attempting to deinit. Reviewed-by: Rich Salz <rsalz@openssl.org>
2016-02-10 09:47:51 +00:00
base_inited = 0;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
}
/*
* If this function is called with a non NULL settings value then it must be
* called prior to any threads making calls to any OpenSSL functions,
* i.e. passing a non-null settings value is assumed to be single-threaded.
*/
The new init functions can now fail so shouldn't be void The new init functions can fail if the library has already been stopped. We should be able to indicate failure with a 0 return value. Reviewed-by: Rich Salz <rsalz@openssl.org>
2016-02-10 13:59:15 +00:00
int OPENSSL_init_crypto(uint64_t opts, const OPENSSL_INIT_SETTINGS *settings)
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
{
Add some TODO notes into init.c We should be seeking to move the OPENSSL_init_crypto and OPENSSL_cleanup processing into OPENSSL_CTX instead. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8857)
2019-05-02 12:42:31 +00:00
/*
* TODO(3.0): This function needs looking at with a view to moving most/all
* of this into OPENSSL_CTX.
*/
Attempt to log an error if init failed If init failed we'd like to set an error code to indicate that. But if init failed then when the error system tries to load its strings its going to fail again. We could get into an infinite loop. Therefore we just set a single error the first time around. After that no error is set. Reviewed-by: Rich Salz <rsalz@openssl.org>
2016-02-10 15:16:06 +00:00
if (stopped) {
Ensure the thread keys are always allocated in the same order Fixes: #5899 Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/5911)
2018-04-20 13:45:06 +00:00
if (!(opts & OPENSSL_INIT_BASE_ONLY))
CRYPTOerr(CRYPTO_F_OPENSSL_INIT_CRYPTO, ERR_R_INIT_FAIL);
The new init functions can now fail so shouldn't be void The new init functions can fail if the library has already been stopped. We should be able to indicate failure with a 0 return value. Reviewed-by: Rich Salz <rsalz@openssl.org>
2016-02-10 13:59:15 +00:00
return 0;
Attempt to log an error if init failed If init failed we'd like to set an error code to indicate that. But if init failed then when the error system tries to load its strings its going to fail again. We could get into an infinite loop. Therefore we just set a single error the first time around. After that no error is set. Reviewed-by: Rich Salz <rsalz@openssl.org>
2016-02-10 15:16:06 +00:00
}
Can't re-init after stop. Remoce DYANMIC once-init stuff. After the library is stopped, you can't restart it. Reviewed-by: Matt Caswell <matt@openssl.org>
2016-02-10 05:39:29 +00:00
More configurable crypto and ssl library initialization 1. In addition to overriding the default application name, one can now also override the configuration file name and flags passed to CONF_modules_load_file(). 2. By default we still keep going when configuration file processing fails. But, applications that want to be strict about initialization errors can now make explicit flag choices via non-null OPENSSL_INIT_SETTINGS that omit the CONF_MFLAGS_IGNORE_RETURN_CODES flag (which had so far been both undocumented and unused). 3. In OPENSSL_init_ssl() do not request OPENSSL_INIT_LOAD_CONFIG if the options already include OPENSSL_INIT_NO_LOAD_CONFIG. 4. Don't set up atexit() handlers when called with INIT_BASE_ONLY. Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7986)
2019-01-01 07:53:24 +00:00
/*
* When the caller specifies OPENSSL_INIT_BASE_ONLY, that should be the
* *only* option specified. With that option we return immediately after
* doing the requested limited initialization. Note that
* err_shelve_state() called by us via ossl_init_load_crypto_nodelete()
* re-enters OPENSSL_init_crypto() with OPENSSL_INIT_BASE_ONLY, but with
* base already initialized this is a harmless NOOP.
*
* If we remain the only caller of err_shelve_state() the recursion should
* perhaps be removed, but if in doubt, it can be left in place.
*/
Ensure the thread keys are always allocated in the same order Fixes: #5899 Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/5911)
2018-04-20 13:45:06 +00:00
if (!RUN_ONCE(&base, ossl_init_base))
return 0;
More configurable crypto and ssl library initialization 1. In addition to overriding the default application name, one can now also override the configuration file name and flags passed to CONF_modules_load_file(). 2. By default we still keep going when configuration file processing fails. But, applications that want to be strict about initialization errors can now make explicit flag choices via non-null OPENSSL_INIT_SETTINGS that omit the CONF_MFLAGS_IGNORE_RETURN_CODES flag (which had so far been both undocumented and unused). 3. In OPENSSL_init_ssl() do not request OPENSSL_INIT_LOAD_CONFIG if the options already include OPENSSL_INIT_NO_LOAD_CONFIG. 4. Don't set up atexit() handlers when called with INIT_BASE_ONLY. Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7986)
2019-01-01 07:53:24 +00:00
if (opts & OPENSSL_INIT_BASE_ONLY)
return 1;
/*
* Now we don't always set up exit handlers, the INIT_BASE_ONLY calls
* should not have the side-effect of setting up exit handlers, and
* therefore, this code block is below the INIT_BASE_ONLY-conditioned early
* return above.
*/
Implement OPENSSL_INIT_NO_ATEXIT Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-15 16:27:34 +00:00
if ((opts & OPENSSL_INIT_NO_ATEXIT) != 0) {
if (!RUN_ONCE_ALT(&register_atexit, ossl_init_no_register_atexit,
ossl_init_register_atexit))
return 0;
} else if (!RUN_ONCE(&register_atexit, ossl_init_register_atexit)) {
return 0;
}
More configurable crypto and ssl library initialization 1. In addition to overriding the default application name, one can now also override the configuration file name and flags passed to CONF_modules_load_file(). 2. By default we still keep going when configuration file processing fails. But, applications that want to be strict about initialization errors can now make explicit flag choices via non-null OPENSSL_INIT_SETTINGS that omit the CONF_MFLAGS_IGNORE_RETURN_CODES flag (which had so far been both undocumented and unused). 3. In OPENSSL_init_ssl() do not request OPENSSL_INIT_LOAD_CONFIG if the options already include OPENSSL_INIT_NO_LOAD_CONFIG. 4. Don't set up atexit() handlers when called with INIT_BASE_ONLY. Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7986)
2019-01-01 07:53:24 +00:00
if (!RUN_ONCE(&load_crypto_nodelete, ossl_init_load_crypto_nodelete))
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
return 0;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
if ((opts & OPENSSL_INIT_NO_LOAD_CRYPTO_STRINGS)
Fix a RUN_ONCE bug We have a number of instances where there are multiple "init" functions for a single CRYPTO_ONCE variable, e.g. to load config automatically or to not load config automatically. Unfortunately the RUN_ONCE mechanism was not correctly giving the right return value where an alternative init function was being used. Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-20 15:32:55 +00:00
&& !RUN_ONCE_ALT(&load_crypto_strings,
ossl_init_no_load_crypto_strings,
ossl_init_load_crypto_strings))
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
return 0;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
if ((opts & OPENSSL_INIT_LOAD_CRYPTO_STRINGS)
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
&& !RUN_ONCE(&load_crypto_strings, ossl_init_load_crypto_strings))
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
return 0;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
if ((opts & OPENSSL_INIT_NO_ADD_ALL_CIPHERS)
Fix a RUN_ONCE bug We have a number of instances where there are multiple "init" functions for a single CRYPTO_ONCE variable, e.g. to load config automatically or to not load config automatically. Unfortunately the RUN_ONCE mechanism was not correctly giving the right return value where an alternative init function was being used. Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-20 15:32:55 +00:00
&& !RUN_ONCE_ALT(&add_all_ciphers, ossl_init_no_add_all_ciphers,
ossl_init_add_all_ciphers))
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
return 0;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
if ((opts & OPENSSL_INIT_ADD_ALL_CIPHERS)
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
&& !RUN_ONCE(&add_all_ciphers, ossl_init_add_all_ciphers))
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
return 0;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
if ((opts & OPENSSL_INIT_NO_ADD_ALL_DIGESTS)
Fix a RUN_ONCE bug We have a number of instances where there are multiple "init" functions for a single CRYPTO_ONCE variable, e.g. to load config automatically or to not load config automatically. Unfortunately the RUN_ONCE mechanism was not correctly giving the right return value where an alternative init function was being used. Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-20 15:32:55 +00:00
&& !RUN_ONCE_ALT(&add_all_digests, ossl_init_no_add_all_digests,
ossl_init_add_all_digests))
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
return 0;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
if ((opts & OPENSSL_INIT_ADD_ALL_DIGESTS)
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
&& !RUN_ONCE(&add_all_digests, ossl_init_add_all_digests))
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
return 0;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
Added EVP_KDF (similiar to the EVP_MAC) Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8808)
2019-04-22 07:18:56 +00:00
if ((opts & OPENSSL_INIT_NO_ADD_ALL_KDFS)
&& !RUN_ONCE_ALT(&add_all_kdfs, ossl_init_no_add_all_kdfs,
ossl_init_add_all_kdfs))
return 0;
if ((opts & OPENSSL_INIT_ADD_ALL_KDFS)
&& !RUN_ONCE(&add_all_kdfs, ossl_init_add_all_kdfs))
return 0;
Fix atfork flag. Avoid double-negatives :) Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> (Merged from https://github.com/openssl/openssl/pull/3815)
2017-06-30 17:55:08 +00:00
if ((opts & OPENSSL_INIT_ATFORK)
Add fork handlers, based on pthread_atfork Only for Unix platforms Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3754)
2017-06-22 18:00:06 +00:00
&& !openssl_init_fork_handlers())
return 0;
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
if ((opts & OPENSSL_INIT_NO_LOAD_CONFIG)
Fix a RUN_ONCE bug We have a number of instances where there are multiple "init" functions for a single CRYPTO_ONCE variable, e.g. to load config automatically or to not load config automatically. Unfortunately the RUN_ONCE mechanism was not correctly giving the right return value where an alternative init function was being used. Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-20 15:32:55 +00:00
&& !RUN_ONCE_ALT(&config, ossl_init_no_config, ossl_init_config))
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
return 0;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
if (opts & OPENSSL_INIT_LOAD_CONFIG) {
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
int ret;
Remove use of CRYPTO_LOCK_INIT in init code Swap the use of CRYPTO_LOCK_INIT in the init code to use the new threading API mechanism for locking. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-07 14:39:22 +00:00
CRYPTO_THREAD_write_lock(init_lock);
More configurable crypto and ssl library initialization 1. In addition to overriding the default application name, one can now also override the configuration file name and flags passed to CONF_modules_load_file(). 2. By default we still keep going when configuration file processing fails. But, applications that want to be strict about initialization errors can now make explicit flag choices via non-null OPENSSL_INIT_SETTINGS that omit the CONF_MFLAGS_IGNORE_RETURN_CODES flag (which had so far been both undocumented and unused). 3. In OPENSSL_init_ssl() do not request OPENSSL_INIT_LOAD_CONFIG if the options already include OPENSSL_INIT_NO_LOAD_CONFIG. 4. Don't set up atexit() handlers when called with INIT_BASE_ONLY. Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7986)
2019-01-01 07:53:24 +00:00
conf_settings = settings;
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
ret = RUN_ONCE(&config, ossl_init_config);
More configurable crypto and ssl library initialization 1. In addition to overriding the default application name, one can now also override the configuration file name and flags passed to CONF_modules_load_file(). 2. By default we still keep going when configuration file processing fails. But, applications that want to be strict about initialization errors can now make explicit flag choices via non-null OPENSSL_INIT_SETTINGS that omit the CONF_MFLAGS_IGNORE_RETURN_CODES flag (which had so far been both undocumented and unused). 3. In OPENSSL_init_ssl() do not request OPENSSL_INIT_LOAD_CONFIG if the options already include OPENSSL_INIT_NO_LOAD_CONFIG. 4. Don't set up atexit() handlers when called with INIT_BASE_ONLY. Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7986)
2019-01-01 07:53:24 +00:00
conf_settings = NULL;
Remove use of CRYPTO_LOCK_INIT in init code Swap the use of CRYPTO_LOCK_INIT in the init code to use the new threading API mechanism for locking. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-07 14:39:22 +00:00
CRYPTO_THREAD_unlock(init_lock);
OPENSSL_init_crypto(): check config return code correctly It was assumed that the config functionality returned a boolean. However, it may return a negative number on error, so we need to take that into account. Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8679)
2019-04-05 08:28:32 +00:00
if (ret <= 0)
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
return 0;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
}
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
if ((opts & OPENSSL_INIT_ASYNC)
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
&& !RUN_ONCE(&async, ossl_init_async))
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
return 0;
Remove some OPENSSL_NO_ASYNC guards in init When config'd with "no-async" the ASYNC_NULL implementation is used, so async symbols still exist. We should still init the NULL implementation so that when we get the async ctx it is NULL rather than undefined. Reviewed-by: Rich Salz <rsalz@openssl.org>
2016-03-17 17:06:28 +00:00
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
#ifndef OPENSSL_NO_ENGINE
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
if ((opts & OPENSSL_INIT_ENGINE_OPENSSL)
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
&& !RUN_ONCE(&engine_openssl, ossl_init_engine_openssl))
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
return 0;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
# ifndef OPENSSL_NO_RDRAND
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
if ((opts & OPENSSL_INIT_ENGINE_RDRAND)
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
&& !RUN_ONCE(&engine_rdrand, ossl_init_engine_rdrand))
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
return 0;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
# endif
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
if ((opts & OPENSSL_INIT_ENGINE_DYNAMIC)
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
&& !RUN_ONCE(&engine_dynamic, ossl_init_engine_dynamic))
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
return 0;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
# ifndef OPENSSL_NO_STATIC_ENGINE
e_devcrypto: make the /dev/crypto engine dynamic Engine has been moved from crypto/engine/eng_devcrypto.c to engines/e_devcrypto.c. Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7859)
2018-11-06 12:57:03 +00:00
# ifndef OPENSSL_NO_DEVCRYPTOENG
if ((opts & OPENSSL_INIT_ENGINE_CRYPTODEV)
&& !RUN_ONCE(&engine_devcrypto, ossl_init_engine_devcrypto))
return 0;
# endif
Deprecate the "hw" configuration options, make "padlockeng" disablable The "hw" and "hw-.*" style options are historical artifacts, sprung from the time when ENGINE was first designed, with hardware crypto accelerators and HSMs in mind. Today, these options have largely lost their value, replaced by options such as "no-{foo}eng" and "no-engine". This completes the transition by making "hw" and "hw-.*" deprecated, but automatically translated into more modern variants of the same. In the process, we get rid of the last regular expression in Configure's @disablables, a feature that was ill supported anyway. Also, padlock now gets treated just as every other engine. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8380)
2019-03-01 12:18:11 +00:00
# if !defined(OPENSSL_NO_PADLOCKENG)
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
if ((opts & OPENSSL_INIT_ENGINE_PADLOCK)
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
&& !RUN_ONCE(&engine_padlock, ossl_init_engine_padlock))
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
return 0;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
# endif
# if defined(OPENSSL_SYS_WIN32) && !defined(OPENSSL_NO_CAPIENG)
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
if ((opts & OPENSSL_INIT_ENGINE_CAPI)
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
&& !RUN_ONCE(&engine_capi, ossl_init_engine_capi))
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
return 0;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
# endif
Rework based on feedback: 1. Cleaned up eventfd handling 2. Reworked socket setup code to allow other algorithms to be added in future 3. Fixed compile errors for static build 4. Added error to error stack in all cases of ALG_PERR/ALG_ERR 5. Called afalg_aes_128_cbc() from bind() to avoid race conditions 6. Used MAX_INFLIGHT define in io_getevents system call 7. Coding style fixes Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org>
2016-02-23 08:01:01 +00:00
# if !defined(OPENSSL_NO_AFALGENG)
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
if ((opts & OPENSSL_INIT_ENGINE_AFALG)
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
&& !RUN_ONCE(&engine_afalg, ossl_init_engine_afalg))
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
return 0;
Rework based on feedback: 1. Cleaned up eventfd handling 2. Reworked socket setup code to allow other algorithms to be added in future 3. Fixed compile errors for static build 4. Added error to error stack in all cases of ALG_PERR/ALG_ERR 5. Called afalg_aes_128_cbc() from bind() to avoid race conditions 6. Used MAX_INFLIGHT define in io_getevents system call 7. Coding style fixes Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org>
2016-02-23 08:01:01 +00:00
# endif
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
# endif
if (opts & (OPENSSL_INIT_ENGINE_ALL_BUILTIN
Don't try to init dasync internally Since dasync isn't installed, and is only ever used as a dynamic engine, there's no reason to consider it for initialization when building static engines. Reviewed-by: Ben Laurie <ben@openssl.org>
2016-08-17 13:06:23 +00:00
| OPENSSL_INIT_ENGINE_OPENSSL
Rework based on feedback: 1. Cleaned up eventfd handling 2. Reworked socket setup code to allow other algorithms to be added in future 3. Fixed compile errors for static build 4. Added error to error stack in all cases of ALG_PERR/ALG_ERR 5. Called afalg_aes_128_cbc() from bind() to avoid race conditions 6. Used MAX_INFLIGHT define in io_getevents system call 7. Coding style fixes Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org>
2016-02-23 08:01:01 +00:00
| OPENSSL_INIT_ENGINE_AFALG)) {
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
ENGINE_register_all_complete();
}
#endif
Fix no-comp build Reviewed-by: Matt Caswell <matt@openssl.org>
2016-03-11 20:13:19 +00:00
#ifndef OPENSSL_NO_COMP
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
if ((opts & OPENSSL_INIT_ZLIB)
Change all our uses of CRYPTO_THREAD_run_once to use RUN_ONCE instead That way, we have a way to check if the init function was successful or not. Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2016-07-19 17:42:11 +00:00
&& !RUN_ONCE(&zlib, ossl_init_zlib))
Swap the init code to use CRYPTO_ONCE The init code was using its own "once" implementation. Now that we have the new thread API we should use that instead. Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-03-02 14:51:00 +00:00
return 0;
Fix no-comp build Reviewed-by: Matt Caswell <matt@openssl.org>
2016-03-11 20:13:19 +00:00
#endif
The new init functions can now fail so shouldn't be void The new init functions can fail if the library has already been stopped. We should be able to indicate failure with a 0 return value. Reviewed-by: Rich Salz <rsalz@openssl.org>
2016-02-10 13:59:15 +00:00
return 1;
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
}
Rename INIT funtions, deprecate old ones. Man, there were a lot of renamings :) Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 16:52:40 +00:00
int OPENSSL_atexit(void (*handler)(void))
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
{
OPENSSL_INIT_STOP *newhand;
Rework DSO API conditions and configuration option 'no-dso' is meaningless, as it doesn't get any macro defined. Therefore, we remove all checks of OPENSSL_NO_DSO. However, there may be some odd platforms with no DSO scheme. For those, we generate the internal macro DSO_NONE aand use it. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/#8622)
2019-03-31 14:15:02 +00:00
#if !defined(OPENSSL_USE_NODELETE)\
Introduce a no-pinshared option This option prevents OpenSSL from pinning itself in memory. Fixes #7598 [extended tests] Reviewed-by: Tim Hudson <tjh@openssl.org> (Merged from https://github.com/openssl/openssl/pull/7647)
2018-11-16 14:05:14 +00:00
&& !defined(OPENSSL_NO_PINSHARED)
Ensure that libcrypto and libssl do not unload until the process exits Because we use atexit() to cleanup after ourselves, this will cause a problem if we have been dynamically loaded and then unloaded again: the atexit() handler may no longer be there. Most modern atexit() implementations can handle this, however there are still difficulties if libssl gets unloaded before libcrypto, because of the atexit() callback that libcrypto makes to libssl. The most robust solution seems to be to ensure that libcrypto and libssl never unload. This is done by simply deliberately leaking a dlopen() reference to them. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-18 13:13:25 +00:00
{
union {
void *sym;
void (*func)(void);
} handlersym;
handlersym.func = handler;
For the lack of GetModuleHandleEx(), we use DSO route for WinCE. Revert win32_pathbyaddr() which is used in DSO_dsobyaddr(). Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from https://github.com/openssl/openssl/pull/8596)
2019-03-27 08:30:47 +00:00
# if defined(DSO_WIN32) && !defined(_WIN32_WCE)
Implement GET_MODULE_HANDLE_EX_FLAG_PIN for windows Rather than leaking a reference, just call GetModuleHandleEx and pin the module on Windows. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-28 10:03:22 +00:00
{
HMODULE handle = NULL;
BOOL ret;
Ensure that libcrypto and libssl do not unload until the process exits Because we use atexit() to cleanup after ourselves, this will cause a problem if we have been dynamically loaded and then unloaded again: the atexit() handler may no longer be there. Most modern atexit() implementations can handle this, however there are still difficulties if libssl gets unloaded before libcrypto, because of the atexit() callback that libcrypto makes to libssl. The most robust solution seems to be to ensure that libcrypto and libssl never unload. This is done by simply deliberately leaking a dlopen() reference to them. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-18 13:13:25 +00:00
Implement GET_MODULE_HANDLE_EX_FLAG_PIN for windows Rather than leaking a reference, just call GetModuleHandleEx and pin the module on Windows. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-28 10:03:22 +00:00
/*
* We don't use the DSO route for WIN32 because there is a better
* way
*/
ret = GetModuleHandleEx(GET_MODULE_HANDLE_EX_FLAG_FROM_ADDRESS
| GET_MODULE_HANDLE_EX_FLAG_PIN,
handlersym.sym, &handle);
if (!ret)
return 0;
}
Rework DSO API conditions and configuration option 'no-dso' is meaningless, as it doesn't get any macro defined. Therefore, we remove all checks of OPENSSL_NO_DSO. However, there may be some odd platforms with no DSO scheme. For those, we generate the internal macro DSO_NONE aand use it. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/#8622)
2019-03-31 14:15:02 +00:00
# elif !defined(DSO_NONE)
Implement GET_MODULE_HANDLE_EX_FLAG_PIN for windows Rather than leaking a reference, just call GetModuleHandleEx and pin the module on Windows. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-28 10:03:22 +00:00
/*
* Deliberately leak a reference to the handler. This will force the
* library/code containing the handler to remain loaded until we run the
* atexit handler. If -znodelete has been used then this is
Fix typo in files in crypto folder Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Andy Polyakov <appro@openssl.org> GH: #4093
2017-08-04 05:10:41 +00:00
* unnecessary.
Implement GET_MODULE_HANDLE_EX_FLAG_PIN for windows Rather than leaking a reference, just call GetModuleHandleEx and pin the module on Windows. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-28 10:03:22 +00:00
*/
{
DSO *dso = NULL;
Don't leave stale errors on queue if DSO_dsobyaddr() fails The init code uses DSO_dsobyaddr() to leak a reference to ourselves to ensure we remain loaded until atexit() time. In some circumstances that can fail and leave stale errors on the error queue. Fixes #3372 Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3383)
2017-05-04 11:51:18 +00:00
ERR_set_mark();
Implement GET_MODULE_HANDLE_EX_FLAG_PIN for windows Rather than leaking a reference, just call GetModuleHandleEx and pin the module on Windows. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-28 10:03:22 +00:00
dso = DSO_dsobyaddr(handlersym.sym, DSO_FLAG_NO_UNLOAD_ON_FREE);
Add dladdr() for AIX Although it deviates from the actual prototype of DSO_dsobyaddr(), this is now ISO C compliant and gcc -Wpedantic accepts the code. Added DATA segment checking to catch ptrgl virtual addresses. Avoid memleaks with every AIX/dladdr() call. Removed debug-fprintf()s. Added test case for DSO_dsobyaddr(), which will eventually call dladdr(). Removed unecessary AIX ifdefs again. The implementation can only lookup function symbols, no data symbols. Added PIC-flag to aix*-cc build targets. As AIX is missing a dladdr() implementation it is currently uncertain our exit()-handlers can still be called when the application exits. After dlclose() the whole library might have been unloaded already. Signed-off-by: Matthias Kraft <makr@gmx.eu> Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Rich Salz <rsalz@openssl.org> (Merged from https://github.com/openssl/openssl/pull/5668)
2018-03-19 17:37:46 +00:00
/* See same code above in ossl_init_base() for an explanation. */
Adapt OPENSSL_INIT_DEBUG to the new generic trace API Co-authored-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/8198)
2018-12-12 23:32:57 +00:00
OSSL_TRACE1(INIT,
"atexit: obtained DSO reference? %s\n",
(dso == NULL ? "No!" : "Yes."));
Implement GET_MODULE_HANDLE_EX_FLAG_PIN for windows Rather than leaking a reference, just call GetModuleHandleEx and pin the module on Windows. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-28 10:03:22 +00:00
DSO_free(dso);
Don't leave stale errors on queue if DSO_dsobyaddr() fails The init code uses DSO_dsobyaddr() to leak a reference to ourselves to ensure we remain loaded until atexit() time. In some circumstances that can fail and leave stale errors on the error queue. Fixes #3372 Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3383)
2017-05-04 11:51:18 +00:00
ERR_pop_to_mark();
Implement GET_MODULE_HANDLE_EX_FLAG_PIN for windows Rather than leaking a reference, just call GetModuleHandleEx and pin the module on Windows. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-28 10:03:22 +00:00
}
# endif
Ensure that libcrypto and libssl do not unload until the process exits Because we use atexit() to cleanup after ourselves, this will cause a problem if we have been dynamically loaded and then unloaded again: the atexit() handler may no longer be there. Most modern atexit() implementations can handle this, however there are still difficulties if libssl gets unloaded before libcrypto, because of the atexit() callback that libcrypto makes to libssl. The most robust solution seems to be to ensure that libcrypto and libssl never unload. This is done by simply deliberately leaking a dlopen() reference to them. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-18 13:13:25 +00:00
}
Link using -znodelete Instead of deliberately leaking a reference to ourselves, use nodelete which does this more neatly. Only for Linux at the moment. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-18 14:11:57 +00:00
#endif
Ensure that libcrypto and libssl do not unload until the process exits Because we use atexit() to cleanup after ourselves, this will cause a problem if we have been dynamically loaded and then unloaded again: the atexit() handler may no longer be there. Most modern atexit() implementations can handle this, however there are still difficulties if libssl gets unloaded before libcrypto, because of the atexit() callback that libcrypto makes to libssl. The most robust solution seems to be to ensure that libcrypto and libssl never unload. This is done by simply deliberately leaking a dlopen() reference to them. Reviewed-by: Tim Hudson <tjh@openssl.org>
2016-10-18 13:13:25 +00:00
Set error code on alloc failures Almost all *alloc failures now set an error code. Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> (Merged from https://github.com/openssl/openssl/pull/5842)
2018-04-03 15:31:16 +00:00
if ((newhand = OPENSSL_malloc(sizeof(*newhand))) == NULL) {
CRYPTOerr(CRYPTO_F_OPENSSL_ATEXIT, ERR_R_MALLOC_FAILURE);
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
return 0;
Set error code on alloc failures Almost all *alloc failures now set an error code. Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> (Merged from https://github.com/openssl/openssl/pull/5842)
2018-04-03 15:31:16 +00:00
}
Provide framework for auto initialise/deinitialise of the library This commit provides the basis and core code for an auto initialisation and deinitialisation framework for libcrypto and libssl. The intention is to remove the need (in many circumstances) to call explicit initialise and deinitialise functions. Explicit initialisation will still be an option, and if non-default initialisation is needed then it will be required. Similarly for de-initialisation (although this will be a lot easier since it will bring all de-initialisation into a single function). Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-02-09 11:26:14 +00:00
newhand->handler = handler;
newhand->next = stop_handlers;
stop_handlers = newhand;
return 1;
}
Add fork handlers, based on pthread_atfork Only for Unix platforms Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3754)
2017-06-22 18:00:06 +00:00
Revert the crypto "global lock" implementation Conceptually, this is a squashed version of: Revert "Address feedback" This reverts commit 75551e07bd2339dfea06ef1d31d69929e13a4495. and Revert "Add CRYPTO_thread_glock_new" This reverts commit ed6b2c7938ec6f07b15745d4183afc276e74c6dd. But there were some intervening commits that made neither revert apply cleanly, so instead do it all as one shot. The crypto global locks were an attempt to cope with the awkward POSIX semantics for pthread_atfork(); its documentation (the "RATIONALE" section) indicates that the expected usage is to have the prefork handler lock all "global" locks, and the parent and child handlers release those locks, to ensure that forking happens with a consistent (lock) state. However, the set of functions available in the child process is limited to async-signal-safe functions, and pthread_mutex_unlock() is not on the list of async-signal-safe functions! The only synchronization primitives that are async-signal-safe are the semaphore primitives, which are not really appropriate for general-purpose usage. However, the state consistency problem that the global locks were attempting to solve is not actually a serious problem, particularly for OpenSSL. That is, we can consider four cases of forking application that might use OpenSSL: (1) Single-threaded, does not call into OpenSSL in the child (e.g., the child calls exec() immediately) For this class of process, no locking is needed at all, since there is only ever a single thread of execution and the only reentrancy is due to signal handlers (which are themselves limited to async-signal-safe operation and should not be doing much work at all). (2) Single-threaded, calls into OpenSSL after fork() The application must ensure that it does not fork() with an unexpected lock held (that is, one that would get unlocked in the parent but accidentally remain locked in the child and cause deadlock). Since OpenSSL does not expose any of its internal locks to the application and the application is single-threaded, the OpenSSL internal locks will be unlocked for the fork(), and the state will be consistent. (OpenSSL will need to reseed its PRNG in the child, but that is an orthogonal issue.) If the application makes use of locks from libcrypto, proper handling for those locks is the responsibility of the application, as for any other locking primitive that is available for application programming. (3) Multi-threaded, does not call into OpenSSL after fork() As for (1), the OpenSSL state is only relevant in the parent, so no particular fork()-related handling is needed. The internal locks are relevant, but there is no interaction with the child to consider. (4) Multi-threaded, calls into OpenSSL after fork() This is the case where the pthread_atfork() hooks to ensure that all global locks are in a known state across fork() would come into play, per the above discussion. However, these "calls into OpenSSL after fork()" are still subject to the restriction to async-signal-safe functions. Since OpenSSL uses all sorts of locking and libc functions that are not on the list of safe functions (e.g., malloc()), this case is not currently usable and is unlikely to ever be usable, independently of the locking situation. So, there is no need to go through contortions to attempt to support this case in the one small area of locking interaction with fork(). In light of the above analysis (thanks @davidben and @achernya), go back to the simpler implementation that does not need to distinguish "library-global" locks or to have complicated atfork handling for locks. Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> (Merged from https://github.com/openssl/openssl/pull/5089)
2018-01-16 15:49:54 +00:00
#ifdef OPENSSL_SYS_UNIX
Add fork handlers, based on pthread_atfork Only for Unix platforms Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3754)
2017-06-22 18:00:06 +00:00
/*
* The following three functions are for OpenSSL developers. This is
* where we set/reset state across fork (called via pthread_atfork when
* it exists, or manually by the application when it doesn't).
*
* WARNING! If you put code in either OPENSSL_fork_parent or
* OPENSSL_fork_child, you MUST MAKE SURE that they are async-signal-
* safe. See this link, for example:
* http://man7.org/linux/man-pages/man7/signal-safety.7.html
*/
void OPENSSL_fork_prepare(void)
{
}
void OPENSSL_fork_parent(void)
{
}
void OPENSSL_fork_child(void)
{
Make RAND_DRBG fork-safe Use atfork to count child forks, and reseed DRBG when the counts don't match. Reviewed-by: Paul Dale <paul.dale@oracle.com> (Merged from https://github.com/openssl/openssl/pull/4101)
2017-08-06 22:12:28 +00:00
rand_fork();
Move the rand_nonce_lock code into drbg_lib.c It was previously rand_lib but it makes more sense in drbg_lib.c since all the functions that use this lock are only ever called from drbg_lib.c We add some FIPS_MODE defines in preparation for later moving this code into the FIPS module. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/9039)
2019-05-24 15:36:44 +00:00
/* TODO(3.0): Inform all providers about a fork event */
Add fork handlers, based on pthread_atfork Only for Unix platforms Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3754)
2017-06-22 18:00:06 +00:00
}
#endif
Reference in a new issue Copy permalink