TestNtCreateFile, TestPipeCreateNamedPipeOverlapped
- These tests are currently only expected to succeed on _WIN32
- Also reflect the reverse meaning of this fact in the return values
TestSynchWaitableTimer, TestSynchWaitableTimerAPC:
- These tests are currently expected to fail on __APPLE__
- Also reflect the reverse meaning of this fact in the return values
This logic makes sure that we don't forget to fix the tests if the
corresponding WinPR implementations are fixed.
TestLibrary:
- TestLibraryA and TestLibraryB must always get built as shared libraries
Let the compiler know that we're comparing a volatile value.
Otherwise the compiler might nuke the comparison operation
and produce code that will spin endlessly.
The SYNCHRONIZATION_BARRIER_FLAGS_SPIN_ONLY flag caused this test
to run extremely long if the system has very few processors.
Although this is expected (thread starvation) this will cause a
intolerably long execution time for automated tests.
Changed the number of threads to be calculated dyamically based
on the number of processors.
Also do proper cleanup to prevent memory leaks.
The current experimental/incomplete WinPR timer queue implementation
has several race conditions.
This commit fixes a segfault caused by not unklinking freed timers
from the timer queue timers list.
pool:
- the winpr implementation fallback was not used on older windows editions
- drop useless and conflicting TP_CALLBACK_ENVIRON_V3
- fix race conditions by using use proper one-time initialization
- on win32 WinPR tried to load several pool/callback_environment functions
from kernel32.dll but since these are defined as inline functions in the
windows headers, no windows edition has ever exported them in any dll.
- removed callback_environment.c and added corresponding static inline
function to pool.h
- fix segfault in TestPoolWork: CloseThreadpoolWork() must not be called
if there is a cleanup group associated with the work object since calling
CloseThreadpoolCleanupGroupMember() already releases the work object
sync:
- The windows headers incorrectly define InitializeCriticalEx support if
_WIN32_WINNT >= 0x0403 instead of >= 0x0600 (Vista)
- created a compatible define to deal with this issue
Since the current winpr implementation for overlapped operations is
incomplete and buggy, all affected functions will now fail if they are
called with a set FILE_FLAG_OVERLAPPED flag or a non-null pointer to
a OVERLAPPED structure.
winpr/nt:
- use proper one-time initialization on win32
- fix TestNtCreateFile
- fix broken/incomplete _RtlAnsiStringToUnicodeString
- unimplemented functions return appropriate error codes
winpr/pipe:
- improved TestPipeCreateNamedPipe
- rewrite the completely broken TestPipeCreateNamedPipeOverlapped test
rdtk:
- improve test and don't blindly return success
winpr/synch:
- fix race condition in TestSynchTimerQueue
winpr/ssspi:
- fix TestEnumerateSecurityPackages printf output
- fix TestQuerySecurityPackageInfo printf output
winpr/environment:
- fix GetEnvironmentStrings printf output
winpr/comm:
- unimplemented functions return appropriate error codes
winpr/io:
- unimplemented functions return appropriate error codes
winpr/thread:
- implement SwitchToThread() via sched_yield()
TestSynchTimerQueue:
- fixed race condition
TestSynchWaitableTimerAPC:
- Use WaitForSingleObjectEx since the thread must be in an alterable state
TestSynch is now expected to succeed on WIN32
- Mutex is recursive on Windows; as a consequence we have to use
the pthread PTHREAD_MUTEX_RECURSIVE type
- Adapt MutexCloseHandle accordingly
- ReleaseMutex returned TRUE even if pthread_mutex_unlock failed
- Fixed and improved the TestSynchMutex ctest
SetEventFileDescriptor overrides the internal file descriptor of the
event but didn't close it. Now if the descriptor is closed if it isn't
marked as attached.
winsock.h pulls in a lot of defines and dependencies that are not
required and partially unwanted in winpr's core (for parts that are not
related to network). In order to get rid of this dependency and have an
independent defines for extended winpr functions the WINPR_FD_* defines
are used internally (and for exposed functions). Where required, like in
WSAEventSelect, the FD_* is mapped to WINPR_FD_*.
Allows the WinPR HANDLE functions WaitForSingleObject and
WaitForMultipleObjects to signal in case of write events.
This is used by CreateFileDescriptor and SetEventFileDescriptor,
which got an API change accomodating for this new feature.
Unless there is an invalid parameter or internal error, ResetEvent
must return TRUE, even if the object is in nonsignaled state.
Also fixed and improved the CTest TestSynchEvent to test for the
expected/correct SetEvent and ResetEvent return values.
MSDN documentation says it is ensured that all threads in the barrier
have finished using it before allowing the barrier to be released in
DeleteSynchronizationBarrier(). The winpr re-implementation wasn't
keeping to that requirement, which was causing occasional crashes
when shadow client tried to access already freed barrier structure.
The crash was occuring in winpr_Handle_cleanup() after finished
waiting on a barrier's event.
uClibc master branch, as well as uclibc-ng add eventfd_{read,write}
definitions. Instead of testing for __UCLIBC__, have cmake explicitly check
for the existence of eventfd_read and save the result in
WITH_EVENTFD_READ_WRITE.
Fixes build errors like:
.../winpr/libwinpr/synch/event.c:120:12: error: static declaration of 'eventfd_read' follows non-static declaration
static int eventfd_read(int fd, eventfd_t* value)
^
In file included from .../winpr/libwinpr/synch/event.c:39:0:
.../usr/include/sys/eventfd.h:37:12: note: previous declaration of 'eventfd_read' was here
extern int eventfd_read (int __fd, eventfd_t *__value);
^
.../winpr/libwinpr/synch/event.c:125:12: error: static declaration of 'eventfd_write' follows non-static declaration
static int eventfd_write(int fd, eventfd_t value)
^
In file included from .../winpr/libwinpr/synch/event.c:39:0:
.../usr/include/sys/eventfd.h:40:12: note: previous declaration of 'eventfd_write' was here
extern int eventfd_write (int __fd, eventfd_t __value);
1)
Added missing checks for CreateEvent which also required the
following related changes:
- changed freerdp_context_new API to BOOL
- changed freerdp_peer_context_new API to BOOL
- changed pRdpClientNew callback to BOOL
- changed pContextNew callback to BOOL
- changed psPeerAccepted callback to BOOL
- changed psPeerContextNew callback to BOOL
2)
Fixed lots of missing alloc and error checks in the
changed code's neighbourhood.
3)
Check freerdp_client_codecs_prepare result to avoid segfaults
caused by using non-initialized codecs.
4)
Fixed deadlocks in x11 caused by missing xf_unlock_x11() calls
in some error handlers
5)
Some fixes in thread pool:
- DEFAULT_POOL assignment did not match TP_POOL definition
- don't free the pool pointer if it points to the static DEFAULT_POOL
- added error handling and cleanup in InitializeThreadpool
* Borrowed eventfd_read and eventfd_write from bionic for uclibc compatibility (uclibc headers are broken unfortunately). Bionic and
FreeRDP are both under the Apache 2.0 license.
freerdp/winpr had the following issues:
* The non reentrant SSL_library_init() was called concurrently (crash)
* Missing code/api to set the eventually required OpenSSL static and dynamic locking callbacks
* Missing code/api to free the application-global or thread-local OpenSSL data and tables
This commit creates two new winpr functions:
BOOL winpr_InitializeSSL(DWORD flags):
Use the flag WINPR_SSL_INIT_ALREADY_INITIALIZED if you want to tell winpr that
your application has already initialized OpenSSL.
If required use the flag WINPR_SSL_INIT_ENABLE_LOCKING to tell winpr that it
should set the OpenSSL static and dynamic locking callbacks.
Otherwise just call it with the flag WINPR_SSL_INIT_DEFAULT.
The recommended way is that your application calls this function once before
any threads are created. However, in order to support lazy OpenSSL library
initialization winpr_InitializeSSL() can also safely be called multiple times
and concurrently because it uses the new InitOnceExecuteOnce() function to
guarantee that the initialization is only performed successfully once during
the life time of the calling process.
BOOL winpr_CleanupSSL(DWORD flags):
If you create a thread that uses SSL you should call this function before the
thread returns using the flag WINPR_SSL_CLEANUP_THREAD in order to clean up
the thread-local OpenSSL data and tables.
Call the function with the flag WINPR_SSL_CLEANUP_GLOBAL before terminating
your application.
Note: This commit only replaced the current occurences of the
SSL_load_error_strings(); SSL_library_init(); pairs in the freerdp source
with winpr_InitializeSSL(). None of the server or client applications has been
changed according to the recommended usage described above (TBDL).
This resulted in huge memory leaks - 8MB per thread, depending on
the system's default stack size.
The leak happend even if CloseHandle() was correctly used to "detach"
the thread but WaitForSingleObject was never called.
winpr is now always build as single library.
The build option MONOLITHIC_BUILD doesn't influence this behavior anymore.
The only exception is winpr-makecert-tool which is still build as extra
library.
This obsoletes complex_libraries for winpr.
select() has the major drawback that it cannot handle file descriptor
that are bigger than 1024. This patch makes use of poll() instead of
select() when poll() support is available.
# By Marc-André Moreau (27) and others
# Via Marc-André Moreau
* 'master' of git://github.com/awakecoding/FreeRDP: (48 commits)
afreerdp: fix build on Android
libwinpr-wtsapi: include wtsapi32.h on Windows
wfreerdp: fix target exporting
wfreerdp: fix building against OpenSSL with MONOLITHIC_BUILD and shared libraries
wfreerdp: fix test build issues
libwinpr-crypto: implement CryptProtectMemory/CryptUnprotectMemory
libwinpr-memory: add missing definitions
libwinpr-memory: start stubbing
Increased size of stream on GCC conference create request from 512 to 1024. For large numbers of static virtual channels, the stream was being overflowed and was causing crashes.
Fixes to process new command line options (autodetect, heartbeat, multitransport), join the MCS message channel and process auto-detect PDUs during the connection sequence.
wfreerdp: fix build against static openssl library
winpr-hash: fix building against openssl in static mode
libfreerdp-gdi: fix invalidation of region with line drawing
mfreerdp: fix hungarian keyboard '0' and 'i' key inversion
Fixes build issue on Windows
libwinpr-synch: add active/inactive linked list for timer queue
libwinpr-synch: improve timer queue implementation
libwinpr-synch: make use of head's expiration time in timer queue
libwinpr-synch: make use of timespec for timer queue
libwinpr-synch: improve timer queue implementation
...
Conflicts:
client/common/CMakeLists.txt
