For pointer subtraction, the result is of type "ptrdiff_t". According to
the C11 standard (Committee Draft - April 12, 2011):
"When two pointers are subtracted, both shall point to elements of the
same array object, or one past the last element of the array object; the
result is the difference of the subscripts of the two array elements. The
size of the result is implementation-defined, and its type (a signed
integer type) is ptrdiff_t defined in the <stddef.h> header. If the result
is not representable in an object of that type, the behavior is
undefined."
In our codes, there are cases that the pointer subtraction is not
performed by pointers to elements of the same array object. This might
lead to potential issues, since the behavior is undefined according to C11
standard.
Also, since the size of type "ptrdiff_t" is implementation-defined. Some
static code checkers may warn that the pointer subtraction might underflow
first and then being cast to a bigger size. For example:
UINT8 *Ptr1, *Ptr2;
UINTN PtrDiff;
...
PtrDiff = (UINTN) (Ptr1 - Ptr2);
The commit will refine the pointer subtraction expressions by casting each
pointer to UINTN first and then perform the subtraction:
PtrDiff = (UINTN) Ptr1 - (UINTN) Ptr2;
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Hao Wu <hao.a.wu@intel.com>
Acked-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Qin Long <qin.long@intel.com>
v2:
Re-generate the patch after the new OpensslLibCrypto instance.
OpenSSL 1.0.2k was released with several severity fixes at
26-Jan-2017 (https://www.openssl.org/news/secadv/20170126.txt).
This patch is to upgrade the supported OpenSSL version in
CryptoPkg/OpensslLib to catch the latest release 1.0.2k.
Cc: Ye Ting <ting.ye@intel.com>
Cc: Wu Jiaxin <jiaxin.wu@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Qin Long <qin.long@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Tested-by: Laszlo Ersek <lersek@redhat.com>
Commit 32387e0081 ("CryptoPkg: Enable ssl build in OpensslLib directly",
2016-12-14) pulls OpenSSL's libssl files into the "OpensslLib.inf" library
instance unconditionally.
If a platform doesn't include the TLS modules, such as
- CryptoPkg/Library/TlsLib/TlsLib.inf
- NetworkPkg/TlsAuthConfigDxe/TlsAuthConfigDxe.inf
- NetworkPkg/TlsDxe/TlsDxe.inf
then the platform never actually uses the libssl functionality that gets
built into "OpensslLib.inf".
Tomas Hoger from Red Hat Product Security tells me that security
evaluation is less demanding if we can actually *exclude* the libssl files
from such OVMF builds that don't specify -D TLS_ENABLE (rather than just
trust modules not to call libssl functions if we don't specify -D
TLS_ENABLE).
This patch introduces a parallel OpensslLib instance called
"OpensslLibCrypto" that is appropriate for platform builds without TLS
enablement. It does not build C source files in vain, and it eases
security review -- all libssl vulnerabilities can be excluded at once.
"OpensslLibCrypto.inf" is created as a copy of "OpensslLib.inf", modifying
the BASE_NAME, MODULE_UNI_FILE and FILE_GUID defines.
"process_files.sh" is extended to auto-generate the list of OpenSSL files
for both library instances accordingly. This list is updated in
"OpensslLibCrypto.inf" at once.
"OpensslLibCrypto.uni" is introduced as a copy of "OpensslLib.uni",
highlighting the difference.
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Gary Lin <glin@suse.com>
Cc: Jiaxin Wu <jiaxin.wu@intel.com>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Cc: Qin Long <qin.long@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Ting Ye <ting.ye@intel.com>
Cc: Tomas Hoger <thoger@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Commit 32387e0081 ("CryptoPkg: Enable ssl build in OpensslLib directly",
2016-12-14) removed the "no-queue" configuration option in
"process_files.sh", plus it enabled "process_files.sh" to place all libssl
source files into "OpensslLib.inf".
However, the patch apparently failed to capture two changes originating
from the above actions:
- the definitions of the OPENSSL_NO_PQUEUE and NO_PQUEUE macros were not
removed from "opensslconf.h",
- "ssl/ssl_conf.c" was not added to "OpensslLib.inf".
Refresh these files, completing commit 32387e0081.
I built OVMF with -D SECURE_BOOT_ENABLE -D TLS_ENABLE, and ArmVirtQemu
with -D SECURE_BOOT_ENABLE, after this fix, and experienced no regression.
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Gary Lin <glin@suse.com>
Cc: Jiaxin Wu <jiaxin.wu@intel.com>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Cc: Qin Long <qin.long@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Ting Ye <ting.ye@intel.com>
Cc: Tomas Hoger <thoger@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Qin Long <qin.long@intel.com>
v2:
* Code refine and Typo fix:
TlsHandeAlert -> TlsHandleAlert
This patch is used to add new TlsLib library, which is wrapped
over OpenSSL. The implementation provides TLS library functions
for EFI TLS protocol and EFI TLS Configuration Protocol.
Cc: Ye Ting <ting.ye@intel.com>
Cc: Long Qin <qin.long@intel.com>
Cc: Fu Siyuan <siyuan.fu@intel.com>
Cc: Zhang Lubo <lubo.zhang@intel.com>
Cc: Thomas Palmer <thomas.palmer@hpe.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Wu Jiaxin <jiaxin.wu@intel.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Reviewed-by: Ye Ting <ting.ye@intel.com>
Correct the unaligned parameter names in comments (BaseCryptLib.h and
HMAC-SHA256 wrapper implementation)
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Dandan Bi <dandan.bi@intel.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Add one new API (Pkcs5HashPassword) to provide PKCS#5 v2.0 PBKDF2
support (Password based encryption key derivation function, specified
in RFC 2898).
Also update the Cryptest utility to include the new API testing (with
the test vector from RFC6070).
Cc: Ting Ye <ting.ye@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Qin Long <qin.long@intel.com>
Reviewed-by: Ting Ye <ting.ye@intel.com>
Add new HMAC-SHA256 cipher support in CryptoPkg to meet more security
and industry requirements,
and update Cryptest utility to include new HMAC-SHA256 test case.
Cc: Ting Ye <ting.ye@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Qin Long <qin.long@intel.com>
Reviewed-by: Ting Ye <ting.ye@intel.com>
Add new xxxxHashAll APIs to facilitate the digest computation of blob
data. New APIs include: Md4HashAll(), Md5HashAll(), Sha1HashAll(),
Sha256HashAll(), Sha384HashAll(), and Sha512HashAll().
The corresponding test cases were added in Cryptest utility.
Cc: Ting Ye <ting.ye@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Qin Long <qin.long@intel.com>
Reviewed-by: Ting Ye <ting.ye@intel.com>
Two official releases (OpenSSL 1.0.2i and 1.0.2j) were available
with several severity fixes at 22-Sep-2016 and 26-Sep-2016.
Refer to
https://www.openssl.org/news/secadv/20160922.txt and
https://www.openssl.org/news/secadv/20160926.txt.
This patch is to upgrade the supported OpenSSL version in
CryptoPkg/OpensslLib to catch the latest release 1.0.2j.
Cc: Ting Ye <ting.ye@intel.com>
Cc: David Woodhouse <David.Woodhouse@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Qin Long <qin.long@intel.com>
Reviewed-by: Ting Ye <ting.ye@intel.com>
Tested-by: Laszlo Ersek <lersek@redhat.com>
Remove "TSCounterSignature" from TSVerify.c, which is not being
used by anyone.
Cc: Ting Ye <ting.ye@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Qin Long <qin.long@intel.com>
Reviewed-by: Ting Ye <ting.ye@intel.com>
It's "OpensslLib", not "OpenSslLib" - not a big issue, but the
typo is annoying when trying to copy-n-paste the path name to
use it on the command line on Linux.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Reviewed-By: Wu Jiaxin <jiaxin.wu@intel.com>
Add the following definition in the [BuildOptions] section in package DSC
files to disable APIs that are deprecated:
[BuildOptions]
*_*_*_CC_FLAGS = -D DISABLE_NEW_DEPRECATED_INTERFACES
Cc: Qin Long <qin.long@intel.com>
Cc: Ting Ye <ting.ye@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Hao Wu <hao.a.wu@intel.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Reviewed-by: Ye Ting <ting.ye@intel.com>
This is never set anymore, so unsetting it or testing whether it is unset
no longer makes any sense.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Tested-by: Laszlo Ersek <lersek@redhat.com>
Tested-By: Liming Gao <liming.gao@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
Set the #define NO_MSABI_VA_FUNCS that will be introduced in a subsequent
patch to avoid the use of the MS ABI in variadic functions. In EDK2, such
functions normally require the EFIAPI modifier to be used, but for external
libraries such as OpenSSL, which lack these annotations, it is easier to
simply revert to the default SysV style VA_LIST ABI.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Tested-by: Laszlo Ersek <lersek@redhat.com>
Tested-By: Liming Gao <liming.gao@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
OpenSSL 1.0.2h was released with several severity fixes at
03-May-2016 (https://www.openssl.org/news/secadv/20160503.txt).
Upgrade the supported OpenSSL version in CryptoPkg/OpensslLib to
catch the latest release 1.0.2h.
Cc: Ting Ye <ting.ye@intel.com>
Cc: David Woodhouse <David.Woodhouse@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Qin Long <qin.long@intel.com>
Reviewed-by: Ye Ting <ting.ye@intel.com>
Tested-by: Laszlo Ersek <lersek@redhat.com>
Some fields in structure 'CertCtx' might be used uninitialized in function
Pkcs7GetCertificatesList().
This commit makes sure that 'CertCtx' gets initialized before being used.
Cc: Long Qin <qin.long@intel.com>
Cc: Ye Ting <ting.ye@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Hao Wu <hao.a.wu@intel.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Reviewed-by: Ye Ting <ting.ye@intel.com>
This commit modifies the code logic to avoid passing NULL pointer to
function BN_bn2bin().
Cc: Long Qin <qin.long@intel.com>
Cc: Ye Ting <ting.ye@intel.com>
Cc: Fu Siyuan <siyuan.fu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Hao Wu <hao.a.wu@intel.com>
Reviewed-by: Qin Long <qin.long@intel.com>
- availabe to available
Cc: Qin Long <qin.long@intel.com>
Cc: Ting Ye <ting.ye@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Giri P Mudusuru <giri.p.mudusuru@intel.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Getting openssl 1.0.2g building with ARM RVCT requires a change to
ignore an unset variable used before set was necessary.
(NOTE: This was fixed in OpenSSL 1.1 HEAD with commit
d9b8b89bec4480de3a10bdaf9425db371c19145b, and can be dropped then.)
corrects x509_vfy.c(875): error C3017: ok may be used before being set
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Eugene Cohen <eugene@hp.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Enable AES cipher support for SmmCryptLib instance.
Cc: Ting Ye <ting.ye@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Qin Long <qin.long@intel.com>
Reviewed-by: Ting Ye <ting.ye@intel.com>
This patch is used to fix the potential system hang
caused by the NULL 'time' parameter usage.
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: Long Qin <qin.long@intel.com>
Cc: Ye Ting <ting.ye@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jiaxin Wu <jiaxin.wu@intel.com>
Reviewed-by: David Woodhouse <David.Woodhouse@intel.com>
OpenSSL 1.0.2g was released with several severity fixes at
01-Mar-2016(https://www.openssl.org/news/secadv/20160301.txt).
Upgrade the supported OpenSSL version in CryptoPkg/OpensslLib to
catch the latest release 1.0.2g.
(NOTE: RT4175 from David Woodhouse was included in 1.0.2g. The
new-generated patch will remove this part. And the line
endings were still kept as before in this version for
consistency)
CC: Ting Ye <ting.ye@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Qin Long <qin.long@intel.com>
Reviewed-by: David Woodhouse <David.Woodhouse@intel.com>
Until we fix the git repository to store line endings properly and then
just check them out in the appropriate form for the platform, let's make
process_files.sh convert the opensslconf.h to DOS line endings when it
creates it.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Reviewed-by: Qin Long <qin.long@intel.com>
This got broken in committing, due to a catalogue of broken practices.
Firstly, we should *pull* git submissions, never recommit them. You
preserve the correct history then, and don't risk rebasing to result in
a history which *never* worked in the form that gets preserved.
That would have kept the authorship attrbution correct too.
Secondly, we shouldn't be storing CRLF line endings in the objects that
git stores in its database. It is designed to store simple LF line endings,
and then check that out as appropriate for the system (resulting in CRLF
in the working tree for Windows users, as they expect). That would avoid
this problem, and all the other problems we have with patches being
exchanged.
Make it executable too, which also got lost in the commit mess.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
OpenSSL 1.1 (as well as our backport to 1.0.2) now allows us to run its
standard Configure script and import the result into the EDK II source
repository for others to build natively. The opensslconf.h file and the
list of files in OpensslLib.inf don't need to be managed manually.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Tested-by: Qin Long <qin.long@intel.com>
This is pull request #755 for OpenSSL 1.1, along with a little extra fix
in the RSA_NET code which has been removed from 1.1 so we can't fix it
there.
https://github.com/openssl/openssl/pull/755
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Tested-by: Qin Long <qin.long@intel.com>
Support for the UEFI target has been added to OpenSSL in commit 4d60c7e10.
Drop our partial implementation and use a backported version of what's
upstream. This includes a couple of fixes which will be needed when we
automatically generate the file list and opensslconf.h instead of
manually maintaining those.
This includes the subsequent fix in commit fb4844bbc.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Tested-by: Qin Long <qin.long@intel.com>
Instead of commenting out the Signed Certificate Timestamps purely based
on the OPENSSL_SYS_UEFI flag, OpenSSL 1.1 supports a no-sct configuration
option, added in commit 05d7bf6c5. Drop our own hack and use that.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Tested-by: Qin Long <qin.long@intel.com>
A more complete implementation of the X509_V_FLAG_NO_CHECK_TIME flag was
added to OpenSSL 1.1 as commit d35ff2c0a. Drop our own version and use a
backport of what was committed upstream.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Tested-by: Qin Long <qin.long@intel.com>
A more complete fix for the no-cms configuration has been added to
OpenSSL 1.1 as commit e968561d5. Drop our own version and use a
backport of what was committed upstream.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Tested-by: Qin Long <qin.long@intel.com>
A different fix for the excessive stack usage has been merged into
OpenSSL 1.1 as commit 8e704858f. Drop our own version and use a backport
of what was committed upstream.
Note: This requires the free() function to work correctly when passed
a NULL argument (qv).
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Tested-by: Qin Long <qin.long@intel.com>
A complete implementation of the no-filenames configuration option was
added to OpenSSL 1.1 in commit 02f7114a7. Drop our own version and use
a backport of what was committed upstream.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Tested-by: Qin Long <qin.long@intel.com>
Extensive fixes for the no-stdio configuration have been merged into
OpenSSL 1.1, primarily in commit 984d6c605.
The backport to 1.0.2 is slightly different because we still have a
mixture of no-fp-api and no-stdio in 1.0.2, although they are hopelessly
intertwined. Nevertheless, drop our own original version and switch to
a backported version of what went into 1.1.
This includes subsequent fixes in commit c0cf5b84d for the TS code.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Tested-by: Qin Long <qin.long@intel.com>
A different fix for the PKCS7_verify() regression on Authenticode
signatures has landed in the OpenSSL 1.0.2 branch as commit c436c990f
and will be present in the 1.0.2g release.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Tested-by: Qin Long <qin.long@intel.com>
All the OpenSSL changes we carry in our EDKII_openssl patch for 1.0.2
are now merged into upstream OpenSSL and will be in the upcoming 1.1
release.
As a first step towards switching out our original hacks for backported
versions of the commits which were actually accepted into OpenSSL 1.1,
just regenerate the *existing* patch against the 1.0.2f release using
'git diff'.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Tested-by: Qin Long <qin.long@intel.com>
This can be an auto-generated file, and it *isn't* in the OpenSSL git tree;
it's only in the generated tarballs. So rather than including it in our
OpenSSL patch, just have the user copy it into place.
This makes it easier to manage changes, and is a step towards better
integration.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Tested-by: Qin Long <qin.long@intel.com>
The standard OpenSSL 1.0.2 configuration and build process will already
symlink or copy the necessary header files to the include/openssl/
directory within the OpenSSL source tree.
When we transition to OpenSSL 1.1 it won't even be necessary to link
or copy the files there; they have just been moved outright.
So let's use them from there. Change the include directory specified
in CryptoPkg/CryptoPkg.dec, and modify the Install.cmd and Install.sh
scripts to copy the files to the normal directory within the OpenSSL
source tree, instead of CryptoPkg/Include/openssl/.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Qin Long <qin.long@intel.com>
Tested-by: Qin Long <qin.long@intel.com>
The ISO C standard says about realloc(),
If ptr is a null pointer, the realloc function behaves like the malloc
function for the specified size.
The realloc() implementation doesn't conform to this currently, so add a
check and call malloc() if appropriate.
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: Qin Long <qin.long@intel.com>
Cc: Ting Ye <ting.ye@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Qin Long <qin.long@intel.com>
The ISO C standard says about free(),
If ptr is a null pointer, no action occurs.
This is not true of the RuntimeFreeMem() internal function. Therefore we
must not forward the argument of free() to RuntimeFreeMem() without
checking.
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: Qin Long <qin.long@intel.com>
Cc: Ting Ye <ting.ye@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Qin Long <qin.long@intel.com>
The ISO C standard says about free(),
If ptr is a null pointer, no action occurs.
This is not true of the FreePool() interface of the MemoryAllocationLib
class:
Buffer must have been allocated on a previous call to the pool
allocation services of the Memory Allocation Library. [...] If Buffer
was not allocated with a pool allocation function in the Memory
Allocation Library, then ASSERT().
Therefore we must not forward the argument of free() to FreePool() without
checking.
This bug can be triggered by upstream OpenSSL commit 8e704858f219
("RT3955: Reduce some stack usage"), for example.
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: Qin Long <qin.long@intel.com>
Cc: Ting Ye <ting.ye@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Qin Long <qin.long@intel.com>
OpenSSL has released version 1.0.2f with two security fixes
(http://www.openssl.org/news/secadv/20160128.txt) at 28-Jan-2016.
Upgrade the supported OpenSSL version in CryptoPkg/OpensslLib
to catch the latest release 1.0.2f.
(NOTE: The patch file was just re-generated, and no new source
changes was introduced for 1.0.2f enabling)
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Qin Long <qin.long@intel.com>
Reviewed-by: Ting Ye <ting.ye@intel.com>
Although the function qsort receives as an argument a "compare" function
which returns an "int", QuickSortWorker (the function used internally by
qsort to do its job) receives as an argument a "CompareFunction" which
returns an "INTN". In a 32-bit machine, "INTN" is defined as "INT32",
which is defined as "int" and everything works well. However, when qsort
is compiled for a 64-bit machine, "INTN" is defined as "INT64" and the
return values of the compare functions become incompatible ("int" for
qsort and "INT64" for QuickSortWorker), causing malfunction.
For example, let's assume qsort is being compiled for a 64-bit machine.
As stated before, the "compare" function will be returning an "int",
and "CompareFunction" will be returning an "INT64". When, for example,
the "compare" function (which was passed as an argument to qsort and,
then, re-passed as an argument to QuickSortWorker) returns -1 (or
0xffffffff, in a 32-bit integer, its original return type) from inside
a call to QuickSortWorker, its return value is interpreted as being an
"INT64" value - which turns out to be 4294967295 (or 0x00000000ffffffff,
in a 64-bit integer) -, making the function QuickSortWorker to behave
unexpectedly.
Note that this unexpected (or incorrect) conversion does not happen when
casting an "INT32" to an "INT64" directly, but does happen when casting
function types.
The issue is fixed by changing the return type of SORT_COMPARE (the type
of "CompareFunction", used by QuickSortWorker) from "INTN" to "int".
This way, both qsort and QuickSortWorker use compatible definitions for
their compare functions.
Contributed-under: TianoCore Contribution Agreement 1.0
Acked-by: Paulo Alcantara Cavalcanti <paulo.alc.cavalcanti@hp.com>
Signed-off-by: Karyne Mayer <kmayer@hp.com>
Signed-off-by: Rodrigo Dias Correa <rodrigo.dia.correa@hp.com>
Signed-off-by: Arthur Crippa Burigo <acb@hp.com>
Reviewed-by: Qin Long <qin.long@intel.com>
git-svn-id: https://svn.code.sf.net/p/edk2/code/trunk/edk2@19748 6f19259b-4bc3-4df7-8a09-765794883524
Jiewen Yao
Hao Wu
Qin Long
Laszlo Ersek
Jiaxin Wu
Dandan Bi
Gary Lin
Thomas Huth
Liming Gao
Ard Biesheuvel
Giri P Mudusuru
Eugene Cohen
David Woodhouse
Karyne Mayer