A recursive OPENSSL_init_crypto(OPENSSL_INIT_LOAD_CRYPTO_STRINGS) call
may happen if an out-of-memory error happens at the first callstack,
and the dead-lock happens at the second callstack, because ossl_err_get_state_int
calls OPENSSL_init_crypto(OPENSSL_INIT_LOAD_CRYPTO_STRINGS) although that
call is currently already executing.
At least on posix system this causes the process to freeze at this
point, and must be avoided whatever it takes.
The fix is using err_shelve_state around the critical region, which
makes ossl_err_get_state_int return early and not call the recursive
OPENSSL_init_crypto(OPENSSL_INIT_LOAD_CRYPTO_STRINGS).
This can be reproduced with my error injection patch.
The test vector has been validated on the master branch:
$ ERROR_INJECT=1692279870 ../util/shlib_wrap.sh ./asn1parse-test ./corpora/asn1parse/027f6e82ba01d9db9a9167b83e56cc9f2c602550
ERROR_INJECT=1692279870
#0 0x7f280b42fef8 in __sanitizer_print_stack_trace ../../../../src/libsanitizer/asan/asan_stack.cpp:86
#1 0x5610a3f396b4 in my_malloc fuzz/test-corpus.c:114
#2 0x7f280a2eb94c in CRYPTO_malloc crypto/mem.c:177
#3 0x7f280a2dafdb in OPENSSL_LH_insert crypto/lhash/lhash.c:114
#4 0x7f280a1c87fe in err_load_strings crypto/err/err.c:264
#5 0x7f280a1c87fe in err_load_strings crypto/err/err.c:259
#6 0x7f280a1c87fe in ERR_load_strings_const crypto/err/err.c:301
#7 0x7f280a6f513b in ossl_err_load_PROV_strings providers/common/provider_err.c:233
#8 0x7f280a1cf015 in ossl_err_load_crypto_strings crypto/err/err_all.c:109
#9 0x7f280a2e9b8c in ossl_init_load_crypto_strings crypto/init.c:190
#10 0x7f280a2e9b8c in ossl_init_load_crypto_strings_ossl_ crypto/init.c:181
#11 0x7f2808cfbf67 (/lib/x86_64-linux-gnu/libc.so.6+0x99f67)
#12 0x7f280a32301e in CRYPTO_THREAD_run_once crypto/threads_pthread.c:154
#13 0x7f280a2ea1da in OPENSSL_init_crypto crypto/init.c:553
#14 0x5610a3f38e2f in FuzzerInitialize fuzz/asn1parse.c:29
#15 0x5610a3f38783 in main fuzz/test-corpus.c:194
#16 0x7f2808c8bd8f (/lib/x86_64-linux-gnu/libc.so.6+0x29d8f)
#17 0x7f2808c8be3f in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x29e3f)
#18 0x5610a3f38d34 in _start (/home/runner/work/openssl/openssl/fuzz/asn1parse-test+0x3d34)
AddressSanitizer:DEADLYSIGNAL
=================================================================
==27629==ERROR: AddressSanitizer: ABRT on unknown address 0x03e900006e23 (pc 0x7f2808cfbef8 bp 0x7f280b36afe0 sp 0x7ffd545b2460 T0)
#0 0x7f2808cfbef8 (/lib/x86_64-linux-gnu/libc.so.6+0x99ef8)
#1 0x7f280a32301e in CRYPTO_THREAD_run_once crypto/threads_pthread.c:154
#2 0x7f280a2ea1da in OPENSSL_init_crypto crypto/init.c:553
#3 0x7f280a1c935e in ossl_err_get_state_int crypto/err/err.c:705
#4 0x7f280a1cf1f9 in ERR_new crypto/err/err_blocks.c:20
#5 0x7f280a2eb9ac in CRYPTO_malloc crypto/mem.c:205
#6 0x7f280a2dafdb in OPENSSL_LH_insert crypto/lhash/lhash.c:114
#7 0x7f280a1c87fe in err_load_strings crypto/err/err.c:264
#8 0x7f280a1c87fe in err_load_strings crypto/err/err.c:259
#9 0x7f280a1c87fe in ERR_load_strings_const crypto/err/err.c:301
#10 0x7f280a6f513b in ossl_err_load_PROV_strings providers/common/provider_err.c:233
#11 0x7f280a1cf015 in ossl_err_load_crypto_strings crypto/err/err_all.c:109
#12 0x7f280a2e9b8c in ossl_init_load_crypto_strings crypto/init.c:190
#13 0x7f280a2e9b8c in ossl_init_load_crypto_strings_ossl_ crypto/init.c:181
#14 0x7f2808cfbf67 (/lib/x86_64-linux-gnu/libc.so.6+0x99f67)
#15 0x7f280a32301e in CRYPTO_THREAD_run_once crypto/threads_pthread.c:154
#16 0x7f280a2ea1da in OPENSSL_init_crypto crypto/init.c:553
#17 0x5610a3f38e2f in FuzzerInitialize fuzz/asn1parse.c:29
#18 0x5610a3f38783 in main fuzz/test-corpus.c:194
#19 0x7f2808c8bd8f (/lib/x86_64-linux-gnu/libc.so.6+0x29d8f)
#20 0x7f2808c8be3f in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x29e3f)
#21 0x5610a3f38d34 in _start (/home/runner/work/openssl/openssl/fuzz/asn1parse-test+0x3d34)
AddressSanitizer can not provide additional info.
SUMMARY: AddressSanitizer: ABRT (/lib/x86_64-linux-gnu/libc.so.6+0x99ef8)
==27629==ABORTING
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
(Merged from https://github.com/openssl/openssl/pull/21683)
Since OPENSSL_malloc() and friends report ERR_R_MALLOC_FAILURE, and
at least handle the file name and line number they are called from,
there's no need to report ERR_R_MALLOC_FAILURE where they are called
directly, or when SSLfatal() and RLAYERfatal() is used, the reason
`ERR_R_MALLOC_FAILURE` is changed to `ERR_R_CRYPTO_LIB`.
There were a number of places where `ERR_R_MALLOC_FAILURE` was reported
even though it was a function from a different sub-system that was
called. Those places are changed to report ERR_R_{lib}_LIB, where
{lib} is the name of that sub-system.
Some of them are tricky to get right, as we have a lot of functions
that belong in the ASN1 sub-system, and all the `sk_` calls or from
the CRYPTO sub-system.
Some extra adaptation was necessary where there were custom OPENSSL_malloc()
wrappers, and some bugs are fixed alongside these changes.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19301)
Otherwise we might cause a recursive locking.
Fixes#18535
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/18545)
The #ifdefs weren't quite correct at times.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/18503)
Rather than relying on the locale code working, instead implement these
functions directly.
Fixes#18322
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/18344)
Even though the function is not part of the public api, it is not
entirely removed, in order to minimize the chance of breakage,
because it is exported from libcrypto. Instead, we keep a dummy
implementation.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/17974)
Including e_os.h with a path from a header file doesn't work well on
certain exotic platform. It simply fails to build.
Since we don't seem to be able to stop ourselves, the better move is
to move e_os.h to an include directory that's part of the inclusion
path given to the compiler.
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/17641)
If objects are added in a config file the OPENSSL_INIT_LOAD_CONFIG
will be called recursively which results in hang in RUN_ONCE.
Fixes#16186
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/16210)
The new names are ossl_err_load_xxx_strings.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15446)
There is no need to make the suggested changes in the 3.0 timescale.
These are just suggested improvements for the future.
Fixes#14375
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14556)
Some functions that lock things are void, so we just return early.
Also make ossl_namemap_empty return 0 on error. Updated the docs, and added
some code to ossl_namemap_stored() to handle the failure, and updated the
tests to allow for failure.
Fixes: #14230
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14238)
If everything has already been initialised we can check this with a
single test at the beginning of OPENSSL_init_crypto() and therefore
reduce the amount of time spent in this function. Since this is called
via very many codepaths this should have significant performance benefits.
Partially fixes#13725 and #13578
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/13733)
Most of the time we don't have any explicit settings when loading a
config file. Therefore we optimise things so that we don't need to use
a lock in that instance.
Partially addresses performance issues in #13725
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/13731)
This includes error reporting for libcrypto sub-libraries in surprising
places.
This was done using util/err-to-raise
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/13318)
These methods should ultimately be deprecated. The move is to insulate
non-UNIX platforms from these undefined symbols.
CLA: Permission is granted by the author to the OpenSSL team to use
these modifications.
Fixes#13273
Signed-off-by: Randall S. Becker <rsbecker@nexbridge.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/13276)
Many of the new types introduced by OpenSSL 3.0 have an OSSL_ prefix,
e.g., OSSL_CALLBACK, OSSL_PARAM, OSSL_ALGORITHM, OSSL_SERIALIZER.
The OPENSSL_CTX type stands out a little by using a different prefix.
For consistency reasons, this type is renamed to OSSL_LIB_CTX.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/12621)
Currently, there are two different directories which contain internal
header files of libcrypto which are meant to be shared internally:
While header files in 'include/internal' are intended to be shared
between libcrypto and libssl, the files in 'crypto/include/internal'
are intended to be shared inside libcrypto only.
To make things complicated, the include search path is set up in such
a way that the directive #include "internal/file.h" could refer to
a file in either of these two directoroes. This makes it necessary
in some cases to add a '_int.h' suffix to some files to resolve this
ambiguity:
#include "internal/file.h" # located in 'include/internal'
#include "internal/file_int.h" # located in 'crypto/include/internal'
This commit moves the private crypto headers from
'crypto/include/internal' to 'include/crypto'
As a result, the include directives become unambiguous
#include "internal/file.h" # located in 'include/internal'
#include "crypto/file.h" # located in 'include/crypto'
hence the superfluous '_int.h' suffixes can be stripped.
The files 'store_int.h' and 'store.h' need to be treated specially;
they are joined into a single file.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9333)
Also includes CRMF (RFC 4211) and HTTP transfer (RFC 6712)
CMP and CRMF API is added to libcrypto, and the "cmp" app to the openssl CLI.
Adds extensive man pages and tests. Integration into build scripts.
Incremental pull request based on OpenSSL commit 8869ad4a39 of 2019-04-02
4th chunk: CMP context/parameters and utilities
in crypto/cmp/cmp_ctx.c, crypto/cmp/cmp_util.c, and related files
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9107)
When the new OpenSSL CSPRNG was introduced in version 1.1.1,
it was announced in the release notes that it would be fork-safe,
which the old CSPRNG hadn't been.
The fork-safety was implemented using a fork count, which was
incremented by a pthread_atfork handler. Initially, this handler
was enabled by default. Unfortunately, the default behaviour
had to be changed for other reasons in commit b5319bdbd0, so
the new OpenSSL CSPRNG failed to keep its promise.
This commit restores the fork-safety using a different approach.
It replaces the fork count by a fork id, which coincides with
the process id on UNIX-like operating systems and is zero on other
operating systems. It is used to detect when an automatic reseed
after a fork is necessary.
To prevent a future regression, it also adds a test to verify that
the child reseeds after fork.
CVE-2019-1549
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9832)
Now that all our MACs have moved to the default provider, we let it
take over completely
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/8877)
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)
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)
In later commits this will allow providers to subscribe to thread stop
events. We will need this in the FIPS module. We also make thread stop
handling OPENSSL_CTX aware (different OPENSSL_CTXs may have different
thread local data that needs cleaning up).
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9040)
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)
In preparation for moving the RAND code into the FIPS module we make
drbg_lib.c OPENSSL_CTX aware.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9039)
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)
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)
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)
'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)
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)
