* Both libusbk and libusb0.sys rely on libusbk.dll being
installed to provides a WinUSB-like API for driver access.
* In case libusbk.dll is not present, fall back to WinUSB.
* Closes#11
* Closes#12
* Using sizeof() in the UNUSED() macro didn't silence all
unused variable warnings.
* Also fix an "unused ctx" warning when ENABLE_DEBUG_LOGGING is on
* Also improve debug output on libusb_open() failure
* Intel USB 3.0 root hubs are similar to the Nec/Renesas USB 3.0 root
hubs as they aren't listed under "USB" either but under "IUSB3".
* Extend the existing workaround for Nec/Renesas USB 3.0 root hubs to
handle the Intel ones as well.
* Closes#39
* Existing code appears disarms the timerfd always, which cancels
pending timeouts as soon as one packet completes.
* This fix moves the disarming of the timerfd to
arm_timerfd_for_next_timeout(), where it is now done conditionally.
It also avoids calling disarm outside of the above call.
* This patch also ensures that all handling of the timerfd is done
under the flying transfers lock.
* Issue reported by Hans de Goede. For more info, see:
https://sourceforge.net/mailarchive/message.php?msg_id=29442693
* Other platforms (Linux, OS-X) appear to use 1 => follow suit.
* For non root hub devices, we simply increment the USB device address.
* Collisions with device address 0 are not expected, but we add
an assertion for just in case.
* When libusb_get_device_list() is called mutliple times, the HID device
path was unconditionally duplicated in the list of device's interfaces.
* Because array boundaries were not checked, this caused overflow and crash.
* This patch adds an out of bound check and also ensures that duplication
of data, for HID and composite, does not occur
* It also renames the private composite_api_flags to api_flags, as well as
reorganizes the private attributes
* Bug report and part of the fix provided by Toby Gray
* The LOG_LEVEL_ enums, that were moved to the public API
in 933a319469bcccc962031c989e39d9d1f44f2885 may conflict
with applications/headers that also define their own
LOG_LEVEL_ values internally.
* As a matter of fact, as per Trac #31, this produces a
conflict with libusb-compat, as it defines its own levels.
* http://sourceforge.net/mailarchive/message.php?msg_id=29418038
* core.c:700:4: warning: Function call argument is an uninitialized value
* darwin_usb.c:1713:11: warning: Access to field 'cfSource' results in a
dereference of a null pointer (loaded from variable 'hpriv')
* sync.c/dpfp.c/dpfp_threaded.c: warning: Result of 'malloc' is converted
to a pointer of type 'unsigned char', which is incompatible with sizeof
operand type
* The library is now compiled with warning level 4 for VS2010
* Move silencing of 4200, 28125 and 28719 to msvc/config.h
* Add fixes in core to silence unused variables warnings
* Ensure that spinlock is always set interlocked in poll_windows
* Add missing check for calloc return value
* Fix data assignation in conditionals warnings
* Fix an OACR/Prefast error related to the use of strncpy in xusb.c
* Also fixes whitespace inconsistencies in core
* Issues reported by Orin Eman and Xiaofan Chen. See:
https://sourceforge.net/mailarchive/message.php?msg_id=29412656
* windows_usb.c(161): warning #181: argument is incompatible
with corresponding format string conversion
* windows_usb.c(2124): warning #111: statement is unreachable
usbi_dbg("ERROR: broken timer thread");
* multiple: warning #188: enumerated type mixed with another
* Issues reported by Orin Eman
* Result of 'calloc' is converted to a pointer of type 'unsigned char *',
which is incompatible with sizeof operand type 'PUSB_CONFIGURATION_DESCRIPTOR'
* priv->config_descriptor is indeed an array of pointers, with each descriptor
allocated, rather than a sequential list of fixed descriptor.
core.c:
* Result of 'malloc' is converted to a pointer of type 'struct libusb_device *',
which is incompatible with sizeof operand type 'void *'
* Memory is never released; potential leak of memory pointed to by 'devs'
* Assigned value is garbage or undefined (due to potentially empty and
uninitialized device list)
descriptor.c:
* Function call argument is an uninitialized value
io.c:
* Call to 'malloc' has an allocation size of 0 bytes
* Branch condition evaluates to a garbage value (due to get_next_timeout
returning a negative error code instead of zero on error)
* Because we use the WINAPI calling convention, the def file MUST have the @n
aliases. There is no way around this as MinGW's .o use decoration always
for __stdcall, which can't be turned off.
* dlltool must therefore be invoked to create a proper import lib from the .def,
using the --kill-at option.
* To do that, a CREATE_IMPORT_LIB autotools variable is introduced.
* Note: the .def file is currently maintained manually.
* p = realloc(p, new_size) does not free the original buffer in case of
a realloc failure.
* reallocf() can be used to do so, but is not available on all platforms.
* This patch introduces usbi_reallocf() in libusbi.h and use that instead of realloc
* Issue and original patch submitted by Moritz Lipp (trac #27)
* With gcc-4.6 the option -fomit-frame-pointer is turned on per default
even for the win32 target. This results in segfaults for any application
using setjmp/longjmp and MSVCRT.dll on WinXP.
See http://gcc.gnu.org/ml/gcc/2011-10/msg00351.html
* Issue reported by Lars Kanis
* Fix naked "exit", which no exit code, to prevent a build failure
from being noticed automatically if a build dependency is lacking.
* Also printed error message to stderr rather than stdout.
* On some Linux platforms, libusbx compilation breaks with:
error: implicit declaration of function ‘pthread_mutexattr_settype’
* This regression, introduced in 463dda06db5da5de0eab32820c7af60605625afe,
is due to pthread.h needing __USE_UNIX98, which is tied to _XOPEN_SOURCE
or _GNU_SOURCE being correctly defined, and which the inclusion of
<unistd.h> before the _XOPEN_SOURCE override modified
* As _GNU_LINUX ensures the definition of __USE_UNIX98 and we require it
for syscalls, we now only define _GNU_SOURCE for Linux.
* Because of missing [], "LIBUSB_MICROLIBUSB_RC" was being displayed
in lieu of version number
* Regression introduced with c9d41fe5f0324d969f758f42843e1fd05859f438
* Further alignment of log messages according to severity
* Fix "'darwin_error_str' defined but not used" warning when
option --disable-log is used (reported by Xiaofan Chen)
* Adds libusb_get_port_number, libusb_get_parent and libusb_get_port_path
* Linux implementation provided by Alan Stern, OS X by Nathan Hjelm
* Unsupported for *BSD platforms
* Failure to read the config descriptor through regular API calls can
occur for some hubs (eg. 05E3:0608) and is of no major consequence.
* Later failures still reported with error severity as they may hide
something more dramatic
* Issue reported by Jerry from NovaDSP
* As version.h processed by autotools, the automatic updating of the
nano there can result in unneeded reconfs (eg. after issuing a git
pull, regardless of whether the conf is still valid)
* Avoid this by placing the nano in its own file, as it won't be
processed by autoconf. This also simplifies the versioning git hooks
and removes the sed dependency.
