| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
| |
Do the work ahead of time and cache the results so that fetching config
descriptors becomes a trivial operation.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
| |
Backend functions dealing with interfaces and alternate settings should
use a type whose range represents that of valid values for interfaces
and alternate settings. Switch to use uint8_t instead of int so that
backends do not have to cast values or do range checks.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
| |
* [-Wdiscarded-qualifiers] assignment discards 'const' qualifier from pointer target type
* [-Wpointer-sign] pointer targets in passing argument N of 'func' differ in signedness
* [-Wsign-compare] comparision between signed and unsigned integer expressions
* [-Wunused-function] 'func' declared 'static' but never defined
* [-Wunused-parameter] unused parameter 'p'
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
| |
* [-Wformat=] format 'S' expects argument of type 'T1', but argument N has type 'T2'
* [-Wmissing-declarations] no previous declaration for 'func'
* [-Wreorder] 'Class::Member' will be initialized after
* [-Wsign-compare] comparison between signed and unsigned integer expressions
* [-Wunused-but-set-variable] variable 'v' set but not used
* [-Wunused-parameter] unused parameter 'p'
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Simplify the library by moving device descriptor initialization to the
backend, while the device is being set up. This removes the duplication
of essentially the same code in every backend.
Add some missing calls to libusb_le16_to_cpu() when reading multi-byte
fields from the "raw" device descriptor. It has worked thus far because
the platforms not using the calls happen to be the same endianness as
the USB bus.
While here, throw in some static assertions to ensure there is no
mismatch between the libusb device descriptor structure and any
device descriptor structure provided by the platform headers.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
All of the API functions should take the typedef'ed versions of the
opaque libusb structures, but some recent additions to the API did not
follow this convention. Fix this by making the changes to the
declarations and definitions of the functions.
Make the placement of the asterisk in pointer variable declarations
consistent (always with the variable name, not the type).
Remove some unnecessary casts and initializations relating to
dynamically allocated memory. While at it, make casts within the core
library consistent in style with no space after the closing parenthesis
of the cast. Most of the core already used this style.
When using the 'sizeof' operator, dereference the pointer instead of
using the type. Most of the core was already doing this, so fix up the
few places that weren't.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
| |
This removes the need for pointer casts when calling backend functions.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
| |
The 'host_endian' argument exists only for a special case in the Linux
backend, that being when the device descriptor is read using usbfs
rather than sysfs. It does not apply to any other descriptor types nor
any other backends, so remove it.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
| |
The public libusb header provides all the definitions for the various
descriptor sizes, so use them instead of defining them again with
different names.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
| |
The Linux backend was the only caller of this function, but with the
packed structures introduced in commit d06cc52851 its use is no longer
necessary. Convert the function to static and remove the return type as
no callers paid attention to it.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
| |
These are going to be used in future commits to cleanup some code. Note
that these are prefixed as 'usbi' rather than 'usb' to avoid conflicts
with definitions provided by some system headers.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
| |
Closes #710
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The sole caller of this function only cares about the device interface
path, so change the calling convention to be like that of the
get_interface_details() function. This also adds more precise error
reporting.
Since this function is specific to the libusb0 filter driver, do not
require the caller to provide the GUID for the libusb0 filter driver.
Remove the use of strtok() on the result of this function. The strtok()
function is not reentrant and is less-than-optimal for locating the
start of the GUID component in the device interface path.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Once upon a time the sanitize_path() function was needed to generate a
consistent path format in order to hash the resulting string for use as
session IDs. Since commit 71a779d078 ("Windows: Rework WinUSB
enumeration process to fix issues on Win8+"), this hashing method is no
longer used for session IDs, thus the sanitize_path() function was no
longer explicitly needed.
User lo1ol also reports in issue #633 that the function actually causes
issues with devices where there is a path component following the device
interface GUID. Rather than patching the function to fix this specific
behavior, just replace it with a simpler function that returns an
uppercased duplicate of the input string.
Closes #633, Closes #662
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Help catch more errors by enabling additional build errors and warnings.
Address some of the warnings seen with these new flags, including moving
the libusb_transfer structure back out of the usbi_transfer structure to
address 'warning: invalid use of structure with flexible array member'.
Apparently a structure ending with a flexible array member is not okay
with the compiler as the last member within another structure.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
C11 compiler support has been available for many years now. It is not
unreasonable to require this now, and doing so allows some cleanup to
the configure script. It is no longer necessary to check for compiler
support of the '-fvibility' flag because any compiler that supports C11
will support this flag as well.
Fix up the way that compiler and linker flags are passed down to the
various makefiles. The compiler flags should be shared by all, but the
linker flags and libraries should be separated between the library and
the examples/tests. The visibility flag is only relevant for the
library and the thread flags are only relevant for sources using thread
constructs, so provide them as needed.
Rearrange configure.ac to group similar functionality and consolidate
where possible.
Based on these changes, update the Travis configuration file to include
newer versions of test platforms to ensure proper C11 compiler support.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add support for real thread IDs on macOS 10.6 and later using the new
pthread_threadid_np() function.
Add support for thread IDs on Haiku, NetBSD and Solaris.
Provide a fallback value other than -1 when direct support is not
available. This should suffice as a unique identifier since pthread_t,
while opaque, is still distinct amongst active threads.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Deferring the cached devices cleanup until the "destructor" function is
called makes it appear as though libusb is leaking memory, especially if
heap allocations are analyzed after calling libusb_exit(). It can also
lead to devices staying on the list longer than they should, as seen by
the following sequence of events:
libusb_init() <-- init_count is 0, async thread is started
devices_scan_devices() <-- enumerates devices
libusb_exit() <-- init_count is 0, async thread is stopped
[one or more devices disconnected]
Because the async thread is stopped when device(s) are disconnected in
the above sequence, the disconnection event(s) will not be processed and
thus darwin_devices_detached() will not be called and the list will have
one or more stale entries until the "destructor" function is finally
called.
Address both of these shortcomings by cleaning up the cached devices
list after stopping the async thread.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Since commit 8cfcc63f4f ("haiku_usb_raw: Add missing wrap_sys_device
field to usbi_os_backend"), compilation of the Haiku backend has been
broken. Since the code is C++, named initializers are not supported. Fix
this by going back to the original style of initializing
usbi_os_backend.
Additionally, commit db99ef3451 ("Various fixes for the Haiku port")
further broken some things. The ClearHalt() function was defined as a
member of the USBDevice class but is declared and needed as a member of
the USBDeviceHandle class. The ioctl command code also contained a typo.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
| |
The majority of backends do not have support for resetting a device, so
simplify them all by making the function optional and having the core
return the appropriate error when the function is not implemented.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
| |
In most cases, usbi_clock_gettime() will map to the standard library's
clock_gettime() function. The semantics of this function are that it
returns -1 upon failure with the error code available in the errno
variable. The backends that need to implement this function should
follow the same semantics, and the return value of usbi_clock_gettime()
should not be directly propagated upwards.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Some functions (e.g. libusb_set_interface_alt_setting()) do not perform
sufficient parameter validation, leaving the burden on the backend to
catch invalid user input. Much of this validation is common across all
backends, yet not every backend implemented it. Fix this by moving
parameter validation to the core library functions.
This is also a good opportunity to remove the redundant
'num_configurations' field from the libusb_device structure. The value
of this field is already contained in the 'device_descriptor' member.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Out of all the backends supported by libusb, only two need to provide an
implementation of the clock_gettime() function. Windows completely lacks
such a function and versions of Mac OS prior to 10.12 do not provide it.
In all other cases the backend simply ends up calling the C library's
clock_gettime() function.
Let's optimize for the common case and check for the availability of
clock_gettime() during configure. If available, we will just call it
directly from any part of the library that needs it. If not available,
the backend is required to provide an implementation of
usbi_clock_gettime() that matches the current requirements.
Closes #685
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
| |
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The transfer timeout is structured around time values provided by the
clock_gettime() function. This function uses a timespec structure, but
the usbi_transfer structure was storing its calculated timeout in a
timeval structure. This mismatch introduces extra work when checking for
transfer timeouts as there must be a conversion between these two
structures. Eliminate this by storing the calculated timeout as a
timespec, thus allowing direct comparison.
Note that a conversion to a timeval is still necessary in the
libusb_get_next_timeout() function because the public API uses a timeval
structure, but this is now the only place where such a conversion is
done.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Commit 763668cc92 ("darwin: fix occasional dead-lock on libusb_exit")
resolved the deadlock situation observed in #112, but unfortunately
there is a very rare race condition that can occur when the asynchronous
event thread exits the run loop before CFRunLoopWakeUp() is called. This
can occur when the shutdown source signal is processed as part of other
events in the run loop, in which case the thread was already "awake".
Prior to this change I was able to consistently trigger a segmentation
fault within 10,000 iterations of a libusb_init()/libusb_exit() loop.
With this change I reached over 4 million iterations without issue.
Closes #701
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The backend private data for the internal library structures has been
accessed through a zero-length os_priv array of type unsigned char.
This approach had two particular disadvantages:
1) A special attribute was needed on the 'os_priv' member to ensure
that the field was properly aligned to a natural pointer alignment.
The support needed for this is not available in every compiler.
2) Each access to the private data areas required an explicit cast
from unsigned char to the type required by the backend.
This change reworks the way the private data is accessed by the
backends. New accessor functions return the private data as a void
pointer type, removing the need for an explicit cast (except for Haiku,
which is C++). The special alignment attribute trickery is also replaced
by simple pointer arithmetic.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
The MAX_PATH_LENGTH in libusb Windows backend is used as size of
dev_id buffer. This buffer used for retreiving Device Instance Id
by SetupDiGetDeviceInstanceIdA function. Acording to Microsoft,
Device Instance Id must be less than MAX_DEVICE_ID_LEN = 200.
So, value of 128 maybe not enough.
Closes #699
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Instead of having just the application name as thread name, provide a more
descriptive one, which can e.g. read by htop.
If setting the name fails, the thread will still work as intended, just
raise a warning in this case.
Closes #689
Signed-off-by: Alexander Stein <alexander.stein@mailbox.org>
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Trying to capture debug logs that reproduce a problem can be tricky.
Turning up the debug level will automatically make the program a bit
slower. This alone cane make timing-sensitive bugs "disappear" when
capturing logs. One of the hot paths for debug messages is fetching the
thread ID, which is immeasurably helpful in understanding thread
interactions within the library. Unfortunately most implementations
require a system call to fetch the executing thread's ID, which isn't
exactly going to help in the way of execution time.
Add a check for thread-local storage support when configuring the
library to build. If the toolchain provides this support, only one
system call will be required per thread. This check is only done for
non-Windows systems because thread-local storage is inefficiently
implemented on MinGW.
Closes #682
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Both the UsbDk and WinUSB backends perform common steps when handling
transfers in order to interact with the poll abstraction, both during
submission and when processing transfer completion. With some
rearranging of shared structures, this can be yanked from the individual
backends and placed in the common area. This allows for several
functions to be removed outright from each backend.
The cancellation logic can also be simplified by attempting CancelIoEx()
at the highest level and delegating to the backend if there are
alternatives to try should CancelIoEx() fail.
After some analysis of how Windows processes asychronous (OVERLAPPED)
requests that the underlying driver completes synchronously, it is now
evident that such requests need not be handled in any special fashion.
Each function that called a driver function that was expected to
complete asynchronously had logic to handle the case of a synchronous
completion, so this has all been killed off. This significantly cleans
up these call sites as now they must only check for an error condition.
Finally, the initialization code for the WinUSB backend has been
reworked to load the WinUSB DLL independent of the libusbK DLL.
Previously when the libusbK DLL was present, all requests to devices
using WinUSB would first be sent through the libusbK DLL where
they would then be forwarded to the WinUSB DLL. This is slightly
inefficient but is also limiting when using Windows 8.1 or later because
support for isochronous transfers through WinUSB will be lost.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
| |
The refactoring in commit df61c0c3a3 ("Windows: improve poll
abstraction") introduced a bug in builds where NDEBUG is defined because
of a statement with side-effects that was put inside an assertion. When
this statement is not evaluated, the file descriptor table gets corrupt.
Fix this by moving the statement outside of the assertion.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
WinUSB control transfers that complete synchronously are incorrectly
having the actual transfer length set to the size of the transfer
buffer. If the control transfer is a read, the device may return less
data than the transfer buffer size. Fix this by reporting the actual
bytes transferred.
Closes #667
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The Linux backend plays lots of games to try and work with older
versions of the kernel that do not have certain features. Lets simplify
the backend by requiring at least 2.6.32 to use libusb. The only thing
remaining that still requires explicit version checking is the maximum
iso frame packet size.
Anything running 2.6.32 or later is sure to have a functional monotonic
clock, so this change also allows the removal of the get_timerfd_clock()
function from the backend as well as the check for a functional
monotonic clock during initialization.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The Visual Studio compiler considers a long to always be 32-bits, so the
official Windows API headers define the DWORD and ULONG types as
unsigned long proper. GCC (and possibly other compilers) vary the width
of a long to match the build target, so this complicates printf format
strings for these two types because the underlying type is inconsistent.
Address this mess by introducing a macro that casts as necessary for the
compiler.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Older versions of the Visual Studio compiler are picky about macros
constructed with the 'do { ... } while (0)' construct. Convert these
internal ones to static inline functions. The result is functionally
equivalent but gets us type checking and a bit more readability.
Also address some compiler warnings due to some header files that are
being included in a different order than before.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The formatting and coding style varied across the whole file. Adopt the
following consistent style:
- Align function arguments to the opening parenthesis
- Do not check for NULL before calling free()
- Reduce indentation where possible in loops by continuing in the
success case
- Remove space between function name and opening parenthesis
- Remove pointless pointer casts from void *
- Replace comparisons with NULL or 0 by a negation operator
- When comparing against rvalues, place the rvalue on the right side
of the expression
- Where possible, have the debug message string on the same line as
the usbi_* call. This makes it easier to grep for specific strings.
Also update the definitions in linux_usbfs.h to exactly match that of
the kernel and remove definitions that are not needed.
A number of functions declared stack buffers of size PATH_MAX. This is
generally 4K, which is very much overkill for a lot of the strings and
is not friendly for embedded environments. Replace many of these buffers
with reasonably-sized ones, in many cases using exactly the size needed.
When reading the descriptors during device enumeration, we were starting
with a 1K buffer and doubling as needed. The vast majority of devices
will not have a large set of descriptors, so change the allocation logic
to grow the buffer in steps of 256 bytes.
Introduce a new parsing function for reading sysfs attributes. Using the
fdopen() function to use fscanf() results in excessive memory
allocation, one for the FILE object and another for the buffer into
which the C library will read the data. The sysfs attributes of interest
are generally just a few characters, so use a small stack buffer and
some rigorous parsing to read these attributes. This also consolidates
error checking (e.g. negative values or larger than expected values).
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Make the formatting consistent across the entire file. In particular:
- Always quote strings whose values are derived
- Use tabs consistently
- Wrap all arguments with square brackets
Replace the use of '-a' with '&&' to be more portable.
Rearrange some of the feature checks to be conditional upon the platform
or backend. For example, there is no need to check for nfds_t on Windows
because poll() doesn't exist there. Similarly we now only check for
timerfd on Linux and Solaris. This translates into slightly faster
configure times.
Explicitly define tokens for both the poll and thread implementations.
This makes the preprocessor conditionals much nicer since it is not
necessary to enumerate all possible OS_* tokens. Also replace
POLL_NFDS_TYPE with a proper typedef that is based on the availability
of the nfds_t type.
Migrate to config definition names that are more consistent with
autoconf. The check for timerfd actually verifies the presence of the
library function instead of just the header definitions, and the token
USBI_TIMERFD_AVAILABLE is now HAVE_TIMERFD. Similarly the check for
syslog results in a definition of HAVE_SYSLOG.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Commit 0bf84e4d51 ("core: fix build warning on newer versions of gcc")
addressed compiler warnings for zero-length printf format strings in the
core library files, but there are some additional remaining in some of
the backends. Address these remaining ones in the same manner.
Also remove the usbi_dbg() call in netbsd_clock_gettime(). This causes
infinite recursion since usbi_dbg() calls the backend's clock_gettime()
function. This was similarly addressed for the OpenBSD backend in commit
6acbd8b405 ("Remove infinite recursion in OpenBSD backend").
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Remove the clear_transfer_priv() function from all backends besides
Linux. This function is only needed if the backend calls
usbi_handle_disconnect(), which only Linux does.
Remove the {attach,detach}_kernel_driver() functions from the Darwin
backend. They return LIBUSB_ERROR_NOT_SUPPORTED, but the same result is
achieved by having those functions be NULL.
Remove the init() and exit() functions from the SunOS backend. They are
optional and as no-ops are pointless.
Remove NULL and 0 initializers from usbi_backend structures.
Use named initializers in the NetBSD backend.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Refactor libusbi.h to include the set of common header files needed by
every main source file in the library and change these source files to
include libusbi.h first, followed by any non-common headers. Including
libusbi.h first ensures that the config definitions are pulled in and
will eliminate redundant includes in the individual sources files.
Also clean up some whitespace errors and remove unnecessary definitions
in the manually generated config.h files.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
| |
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Commit 395e5a8a6f ("windows: remove total fds (256) limitations") and
commit c730a8410c ("windows: workaround WaitForMultipleObjects max 64
events limitation.") lifted some hard-coded limits in the number of
HANDLEs that can be used within the library. This change improves on
these changes to make them more efficient.
A bitmap has been added to provide an efficient lookup mechanism for
located unused file descriptor indices. This avoids the O(n) lookup time
for traversing the entire fd_table. This bitmap is dynamically resized
along with the fd_table.
The incremental size of the fd_table has been reduced from 256 to 64.
The vast majority of applications won't need to use 256 HANDLEs, so we
can optimize memory usage a bit.
Commit fb864b7cde ("fix windows crash when multi-thread do sync
transfer") added usbi_inc_fds_ref() and usbi_dec_fds_ref() functions to
work around a reference count issue. Remove these functions and change
the implementation of usbi_poll() to take a reference to each file
descriptor upon entry and drop the references when returning. If the
application experiences any kind of crash, there is a problem elsewhere.
Finally, make the thread executing usbi_poll() take part in the waiting.
The original implementation had this thread simply waiting on a single
event while separate threads waited on the HANDLEs. Now this thread will
wait on MAXIMUM_WAIT_OBJECTS - 1 HANDLEs, thereby reducing the number of
threads that are created. Additionally there is now only a single event
object that is shared amongst all waiting threads.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
| |
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
According to Microsoft, anything prior to Vista could provide
inconsistent results for the value of QueryPerformanceCounter() across
different cores. Now that XP is no longer supported, we can drop the
significant overhead of using a dedicated thread pinned to a single core
just to read a timestamp.
The C++11 steady_clock implementation directly wraps
QueryPerformanceCounter(), so if it is good enough for that then it is
good enough for our purposes.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
| |
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
| |
XP is nearly 20 years old and there are hoops we jump through to keep
supporting it. Time to say goodbye and simplify some things.
Closes #267
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Commit a9b34d170a ("Adding support for ARM & ARM64 Windows Platform")
introduced a dependency on a particular version of the Windows 10 SDK
for *all* platforms. This is particularly annoying for most users who
will only be building for Windows. Fix this by specifying the SDK
dependency only for the ARM/ARM64 platforms and bump to the latest.
Commit 77037c4dd6 ("Adds /utf-8 to compile options") added this compiler
option to all versions of the Visual Studio project files. This results
in a number of warnings with the older versions that don't recognize
this option. Fix this by keeping this option only for 2015 and newer.
Explicitly specify library dependencies for non-static targets. With a
small change in the UsbDk backend we can completely remove all
depenencies other than kernel32.lib.
Lastly, remove the 'MinimalRebuild' option for 2015 and newer project
files as this option is now deprecated and results in a warning for each
project.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
| |
There appears to be no need for the WinCE backend anymore, and it is
increasingly difficult to keep healthy as the rest of the library
changes.
Require at least Visual Studio 2013 to compile. This simplifies matters
as there is some semblance of C99 support there.
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Prior to this change, the URBs allocated for an individual transfer were
freed when the last URB in the transfer was reaped. Normally this causes
no issues because URBs are reaped in the order they were submitted. If
the device is disconnected while multiple URBs are queued, these URBs
may be reaped in an order that does not match that of submission.
Change the logic to free the URBs when all the URBs of a transfer have
been reaped rather than the last one. While in here, improve some debug
messages.
Closes #607
Signed-off-by: Chris Dickens <christopher.a.dickens@gmail.com>
|