1 commit ab9cd5a7be637f7b793987971a706b1d11c27ded
2 Author: Peter Stuge <peter@stuge.se>
3 Date: Fri Apr 20 08:10:11 2012 +0200
7 commit 696400ad83531bcf944c71164a46f9ccb6a33df3
8 Author: Peter Stuge <peter@stuge.se>
9 Date: Fri Apr 20 08:08:34 2012 +0200
11 AUTHORS: Update list of contributors
13 commit 19f94537ea55586dd85f8d627fcfaa6173069ea1
14 Author: Peter Stuge <peter@stuge.se>
15 Date: Fri Apr 20 06:41:34 2012 +0200
17 AUTHORS: Add Martin Pieuchot as copyright holder
19 commit 1a1c633d7912a2565f6af0c9294e10a759d40e05
20 Author: Peter Stuge <peter@stuge.se>
21 Date: Fri Apr 20 06:41:12 2012 +0200
23 AUTHORS: Add Hans de Goede as copyright holder
25 commit 9477f793341d59cc861acb283e6cb03b5730ac91
26 Author: Peter Stuge <peter@stuge.se>
27 Date: Fri Apr 20 06:40:45 2012 +0200
29 AUTHORS: Add Peter Stuge as copyright holder
31 commit 4b0dfe9770356af3c51103a9b3d63c5f8cc63d8b
32 Author: Pete Batard <pete@akeo.ie>
33 Date: Mon Apr 2 15:47:53 2012 +0100
35 AUTHORS: Update Pete Batard's copyright years and email address
37 libusbx.git commit f98eaca0985364be00fcfde3fd3884075092a4e9
39 [stuge: Remove unrelated changes]
41 commit 9f0ea78357b7b1654cb3e7de6ff57f5f378b68f1
42 Author: Peter Stuge <peter@stuge.se>
43 Date: Fri Apr 20 08:35:04 2012 +0200
45 AUTHORS: Update Nathan Hjelm's copyright years
47 commit 8d38506ebc9fb57a40113e5a7739572a42b826d8
48 Author: Peter Stuge <peter@stuge.se>
49 Date: Fri Apr 20 06:36:08 2012 +0200
51 README: Add the BSDs, punctuation, mailinglist link, and remove Daniel
53 commit 05cbe2de45806db309343cdf298d8569c55863e5
54 Author: James Hanko <james.hanko@tridentmicro.com>
55 Date: Wed Oct 19 17:32:34 2011 -0700
57 Linux: Search for /dev/usbdev<bus>.<device> USB device special files
59 If neither the (now deprecated) usbfs filesystem nor udev is available
60 then libusb searches for device nodes also in /dev where they may be
61 created if the kernel was built with the option CONFIG_USB_DEVICE_CLASS.
62 This helps on embedded systems such as Android, and all mdev users.
66 commit a806d169e6b5f675289cb9307fcd6b908edd9219
67 Author: Peter Stuge <peter@stuge.se>
68 Date: Fri Apr 20 04:32:47 2012 +0200
70 Add a timestamp in log messages
74 commit 03f4d40c748506c85f88427dd53afd5c663a2929
75 Author: Peter Stuge <peter@stuge.se>
76 Date: Fri Apr 20 02:24:44 2012 +0200
78 Include library version in the very first libusb_init() debug message
80 commit 84d5eafd1f24a1c6f77f0c2a2ad40e1fbfdea381
81 Author: Martin Pieuchot <mpi@openbsd.org>
82 Date: Thu Apr 19 13:33:36 2012 +0200
84 OpenBSD: Provide libusb_get_device_speed() data
86 Because the different speed values used by the libusb and OpenBSD's
87 usb_device_info structure are compatible, keep the code simple and
88 just copy the value returned by the USB_GET_DEVICEINFO ioctl() call.
92 commit c145ec09732fac2eef53accc07314040d22b64a5
93 Author: Peter Stuge <peter@stuge.se>
94 Date: Fri Apr 20 00:11:01 2012 +0200
96 doc/Makefile.am: Remove username from SF web rsync upload command
98 Commit c407a2d97b6f8715d3452257a0adac40a20178e6 did it for tarballs.
100 libusbx.git commit ca159f15737e9097a4d031c9950f281c42af040a
102 commit 6506013c6d4face34cfec79dd0584e9f528cbc6a
103 Author: Pete Batard <pete@akeo.ie>
104 Date: Tue Apr 10 13:30:24 2012 +0100
106 Windows: Remove #define compile time options in the backend source code
108 * The DYNAMIC_FDS, AUTO_CLAIM and FORCE_INSTANT_TIMEOUTS options
109 were introduced for development/testing and don't appear to be
110 used by the Windows backend users => remove them.
112 libusbx.git commit 4cccbed825fe1dc138121640a6497d7d05aac159
114 commit e291e369b975692dfa1a0f029244af597fa61220
115 Author: Pete Batard <pete@akeo.ie>
116 Date: Tue Apr 10 13:05:37 2012 +0100
118 Windows: Warn on one libusb_get_device_list() memory allocation failure
120 libusbx.git commit 24d595aa71e7fd0f14d40e933a33f852f7269c8b
122 commit 2f40a03ea0552cd8b5ce3d7667b20dc04da061da
123 Author: Pete Batard <pete@akeo.ie>
124 Date: Tue Apr 10 13:05:37 2012 +0100
126 Windows: Add textual error name to debug message on auto-release errors
128 libusbx.git commit 24d595aa71e7fd0f14d40e933a33f852f7269c8b
130 commit 3c7f85a49b42ac36c6d41294858518c554ac4694
131 Author: Pete Batard <pete@akeo.ie>
132 Date: Tue Apr 10 13:05:37 2012 +0100
134 Windows: Silence warning about unused function guid_to_string()
136 Commit 6b1982b1bd321b55698ce537e64648e8aec85f6f removed the #ifdef
137 around guid_to_string() but this ended up causing a warning when
138 the function wasn't being called. Add it back with correct _MSC_VER.
140 libusbx.git commit 24d595aa71e7fd0f14d40e933a33f852f7269c8b
142 [stuge: Rewrite commit message for libusb.git]
144 commit eb55963a6b33799ea76c34f538f21e86f235ea62
145 Author: Pete Batard <pete@akeo.ie>
146 Date: Tue Apr 10 13:05:37 2012 +0100
148 Windows: Base DUMMY_HANDLE on LONG_PTR instead of LONG
150 libusbx.git commit 24d595aa71e7fd0f14d40e933a33f852f7269c8b
152 commit d930865987de523061eca259a7fed11aea626831
153 Author: Pete Batard <pete@akeo.ie>
154 Date: Tue Apr 10 13:05:37 2012 +0100
156 Windows: Prefer calloc() over malloc() and cast all allocations
158 This silences VS2010 intellisense warnings on memory allocation.
160 libusbx.git commit 24d595aa71e7fd0f14d40e933a33f852f7269c8b
162 commit df35117ce58b74fa530baaaccc30adaf432398ea
163 Author: Peter Stuge <peter@stuge.se>
164 Date: Thu Apr 19 22:55:44 2012 +0200
166 Add libusb_get_version() API function
168 Let's see how this will be used. Fixes #53.
170 libusbx.git commit 37dfd16c8c2f36c81c86de303072def0dc405e32
172 [stuge: Remove unrelated changes]
173 [stuge: Add struct libusb_version member rc for release candidate suffix]
174 [stuge: Add struct libusb_version member describe for git describe output]
176 commit 29d5f63c5765f36ea65b55c4cd5f36e1f10a4dae
177 Author: Peter Stuge <peter@stuge.se>
178 Date: Thu Apr 19 22:25:07 2012 +0200
180 Windows: Rename VS2005 solution and add VS2010 project files
182 libusbx.git commit fcad0941e52eb2fd6d57d3327f420db14dfabf0b
184 [stuge: Remove xusb references]
185 [stuge: Add newlines at end of new files]
186 [stuge: Rename libusb_20??.sln to libusb_vs20??.sln]
188 commit b217ec1f904c4382010a3a834a7c58d595b594cf
189 Author: Pete Batard <pete@akeo.ie>
190 Date: Sat Mar 31 22:32:56 2012 +0100
192 libusb/Makefile.am: Fix missing DLL .def file from make dist
194 libusbx.git commit 10dc1ea9a799b13b7423968130da812aa97289bf
196 [stuge: Use EXTRA_DIST in libusb/Makefile.am instead]
198 commit 54cb51d8417e0128eb9088bf628db138c9daa2a0
199 Author: Pete Batard <pete@akeo.ie>
200 Date: Sat Mar 31 22:14:47 2012 +0100
202 Makefile.am: Fix msvc dir not being included in make dist
204 libusbx.git commit a8a80b60a14e3a8269e42dedf34fda66da8e0e50
206 commit 38d5ba472c6cb728ec818a0c8d964b632c8a7992
207 Author: Hans de Goede <hdegoede@redhat.com>
208 Date: Fri Mar 30 13:40:32 2012 +0100
210 Linux: Don't spam the console with 'device unconfigured' errors
212 libusbx.git commit 3c968be61befc47d33562843e5479ff67303b0c4
214 commit 93f14087b2e79d71ba1e53c0b9c7da655dd1258e
215 Author: Nathan Hjelm <hjelmn@me.com>
216 Date: Thu Apr 19 05:52:58 2012 -0600
218 Darwin: Silence warning: cast from pointer to integer of different size
220 As per http://marc.info/?m=133483651109899
222 Reported-by: Xiaofan Chen <xiaofanc@gmail.com>
224 commit 0efc2086646200d8dee86cbbb1e98136383a05c3
225 Author: Pete Batard <pete@akeo.ie>
226 Date: Thu Mar 29 23:21:25 2012 +0100
228 .gitattributes: Add export-ignore for .gitattributs and .gitignore
230 This keeps the git files out of archives created by gitweb.
232 libusbx.git commit 2250152b251243f75c951533ae7b1139a785cfe6
234 [stuge: Removed unrelated whitespace changes]
236 commit a7ef55c02f0116b94cb75831e5d90ad258c662f8
237 Author: Pete Batard <pete@akeo.ie>
238 Date: Wed Mar 28 01:02:02 2012 +0100
240 Windows: fix Clang warnings about unused assignations
242 libusbx.git commit c0e7b289fdb35687b90263b14426319d2b96b132
244 commit ac8eed67e1595ee5a069623e48362d8a85dcb252
245 Author: Pete Batard <pete@akeo.ie>
246 Date: Mon Mar 26 23:34:25 2012 +0100
248 configure.ac: LT_LANG([Windows Resource]) makes windres check redundant
250 libusbx.git commit 4f9fd6a29a712b0e3312983e9ff153e6d7ed0822
252 commit 89ce6f7d05dc8ff4a441a01b192d2e20857f6c48
253 Author: Pete Batard <pete@akeo.ie>
254 Date: Tue Mar 27 00:03:07 2012 +0100
256 Windows: Bring .vcproj x86_64 output directories in line with x86_32
258 libusbx.git commit abdb4437ea1897226076fa706651e350dba1cfd5
260 commit 77743c6f5b3cfa2d06dcb32204bc65d8bd4fec80
261 Author: Hans de Goede <hdegoede@redhat.com>
262 Date: Fri Mar 9 11:31:11 2012 +0100
264 Linux: Fix a comment in handle_iso_completion()
266 Fix the comment at the end of handle_iso_completion, we don't stop on
267 urbs / isoc packets with less data then requested, which is a good
268 thing since it is a normal condition for isoc transfers.
270 Signed-off-by: Hans de Goede <hdegoede@redhat.com>
272 commit 5b87c2c7475a8d68270e19336afec0c5a08579ca
273 Author: Hans de Goede <hdegoede@redhat.com>
274 Date: Fri Mar 9 11:31:10 2012 +0100
276 Linux: Remove bogus status handling from handle_iso_completion()
278 Remove a useless check+set of status because:
279 1. The check is always true; and
280 2. The new value is never used
282 Signed-off-by: Hans de Goede <hdegoede@redhat.com>
284 commit e23d8f67eab947e5d41216e0139f0a50d1342ef8
285 Author: Hans de Goede <hdegoede@redhat.com>
286 Date: Fri Mar 9 11:31:09 2012 +0100
288 Linux: Translate isoc packet status codes to LIBUSB_TRANSFER_ codes
290 During testing of my usbredir code I hit a scenario where my libusb
291 app was seeing EXDEV as status in the transfer's iso_packet_desc.
293 This happened because we didn't translate linux negative errno errors
294 in each isoc packet's status to LIBUSB_TRANSFER_ status codes! This
295 patch adds these translations.
297 Signed-off-by: Hans de Goede <hdegoede@redhat.com>
299 commit 7199bc2c9c351b553b6175a588237ee793719281
300 Author: Hans de Goede <hdegoede@redhat.com>
301 Date: Fri Mar 9 11:31:08 2012 +0100
303 Linux: Fix URB return code handling
305 During testing of my usbredir code I hit a case where EOVERFLOW was not
306 handled in handle_control_completion(). Instead of just fixing this one
307 case I've audited (and fixed where necessary) all handle_foo_completion()
308 functions to know about all errors documented in
309 linux/Documentation/usb/error-codes.txt.
311 Note that for handle_iso_completion() this patch actually removes the
312 handling of some codes, since these can never occur on an iso urb (they
313 can only occur on the iso packets included in the urb, see the next patch
314 in this series). Also, in case an unknown status is encountered on an iso
315 urb this patch sets the urb's status to ERROR rather then leaving it at
318 Signed-off-by: Hans de Goede <hdegoede@redhat.com>
320 commit 98bc7b8d125c2756deadfc889dcefd4e6f2fcd02
321 Author: Hans de Goede <hdegoede@redhat.com>
322 Date: Fri Mar 9 11:31:07 2012 +0100
324 Don't print an error when libusb_cancel_transfer() fails with NOT_FOUND
326 As stated in the documentation for libusb_cancel_transfer(),
327 LIBUSB_ERROR_NOT_FOUND is an expected return value for
328 libusb_cancel_transfer() under certain circumstances, so
329 printing an error every time this happens is undesirable.
331 Even more so because under Linux IOCTL_USBFS_DISCARDURB sets errno
332 to EINVAL when the kernel can't not find the urb in the kernel's
333 urbs-in-flight list, which means that the urb has already completed
334 at the host controller level but it has not necessarily been reaped
335 yet. IOW under Linux libusb_cancel_transfer() may yield a result of
336 LIBUSB_ERROR_NOT_FOUND *before* the transfer's callback has been
337 called! In conclusion there is no way for applications to avoid
338 calling libusb_cancel_transfer() on already completed transfers,
339 and these errors can and do happen frequently for some USB traffic.
341 Signed-off-by: Hans de Goede <hdegoede@redhat.com>
343 commit 4db8275da59996eb1c22d2e9e9c527fa58b288ad
344 Author: Hans de Goede <hdegoede@redhat.com>
345 Date: Fri Mar 9 11:31:06 2012 +0100
347 Linux: Fix cancel_transfer return value when cancelling a multi-urb transfer
349 If we fail to cancel the last urb of a multi-urb transfer because it
350 has already completed (errno == EINVAL on DISCARD_URB), then the entire
351 transfer has already completed, so returning NOT_FOUND is consistent
352 with what the documentation for libusb_cancel_transfer() says.
354 But if we've successfully cancelled the last urb, and then another urb
355 fails with errno == EINVAL, this means that we've still cancelled the
356 transfer, as it has only *partially* completed.
358 Signed-off-by: Hans de Goede <hdegoede@redhat.com>
360 commit b22253ba9cbd399299b534a8c5e3bbac3f240963
361 Author: Peter Stuge <peter@stuge.se>
362 Date: Thu Apr 19 04:05:29 2012 +0200
364 libusb-1.0.pc: Mention OpenBSD and NetBSD in description
366 commit 6aa0fc60d94e3123de230d3fbc9196814c0e466e
367 Author: Peter Stuge <peter@stuge.se>
368 Date: Mon Apr 16 05:59:44 2012 +0200
370 io.c: Add debug message with the callback address on completed transfer
372 commit de41604560a57b2279ac1d0a10b8192a9224d284
373 Author: Xiaofan Chen <xiaofanc@gmail.com>
374 Date: Sun Apr 15 11:50:27 2012 +0800
376 Experimental support for NetBSD using the OpenBSD backend
378 Signed-off-by: Xiaofan Chen <xiaofanc@gmail.com>
380 commit 404054cb508881c9bd6bc575b66c4f509046b4af
381 Author: Peter Stuge <peter@stuge.se>
382 Date: Thu Apr 19 00:41:45 2012 +0200
384 configure.ac: Separate per-OS backend selection from per-backend settings
386 commit 2eba7e4197cf9afa77451186e1a116ee3ae91adc
387 Author: Xiaofan Chen <xiaofanc@gmail.com>
388 Date: Sun Apr 15 07:17:18 2012 +0800
390 Linux: Silence unused variable warnings when without CLOCK_MONOTONIC
392 Signed-off-by: Xiaofan Chen <xiaofanc@gmail.com>
394 commit bc790a9c414871cc1d33f1c7686048434a700e45
395 Author: Peter Stuge <peter@stuge.se>
396 Date: Sat Apr 7 05:20:36 2012 +0200
398 examples: Use snprintf() instead of sprintf() in dpfp and dpfp_threaded
400 The OpenBSD C compiler generates a warning for every use of sprintf()
401 and for a good reason. Reported in http://marc.info/?m=133376187514495
403 Reported-by: Xiaofan Chen <xiaofanc@gmail.com>
405 commit bffa232dca51bcc41df71c5735fa56370b7509ad
406 Author: Peter Stuge <peter@stuge.se>
407 Date: Sat Apr 7 03:55:36 2012 +0200
409 OpenBSD: Fix warning: 'hpriv' may be used uninitialized
411 As reported in http://marc.info/?m=133376187514495
413 The variable would be used uninitialized when the internal list of
414 open devices is empty as obsd_handle_events() checks which device
415 the event fd belongs to, which can not actually happen, but if it
416 ever does happen then we will now return LIBUSB_ERROR_NO_DEVICE.
418 Reported-by: Xiaofan Chen <xiaofanc@gmail.com>
420 commit 6b1982b1bd321b55698ce537e64648e8aec85f6f
421 Author: Moses McKnight <moses@texband.net>
422 Date: Fri Mar 30 21:53:20 2012 -0500
424 Windows: Fix build on VC++2003 which doesn't support variadic macros
426 Change _MSC_VER tests to check for >= 1400 instead of > 1200 because
427 Visual C++ did not support variadic macros until VC++2005, as per
428 http://msdn.microsoft.com/en-us/library/ms177415%28v=VS.80%29.aspx
430 Since usbi_dbg() is a function and not a macro with the old MS C
431 compilers the functions guid_to_string() and windows_error_str()
432 need to be compiled unconditionally, and the compiler should remove
433 the dead code when debug logging is disabled on newer compilers.
435 commit 58ac0355e547080c82d02685a32480b891383d3d
436 Author: Peter Stuge <peter@stuge.se>
437 Date: Sat Mar 31 03:44:40 2012 +0200
439 Windows: Don't use a log function body macro for old MS compilers
441 The duplication is a lot more straightforward than the previous LOG_BODY
442 macro, which also didn't handle the special debug case correctly. The
443 duplicated code is fortunately not very long.
445 commit 76af4c99bd6122c1dd7b58858b669c919d8c8806
446 Author: Peter Stuge <peter@stuge.se>
447 Date: Thu Mar 22 01:52:59 2012 +0100
449 libusbi.h: Add IS_EPIN() IS_EPOUT() IS_XFERIN() and IS_XFEROUT() macros
451 The macros make backend code a little simpler to read, since the
452 transfer and endpoint direction is checked in several places across
453 the supported platforms.
455 commit e3d0a4cb9e2f9872c9fdbb22d7ded169e111fc8f
456 Author: Peter Stuge <peter@stuge.se>
457 Date: Wed Feb 22 14:21:54 2012 +0100
459 Add LIBUSB_TRANSFER_ADD_ZERO_PACKET flag to indicate need for ZLP
461 Some protocols which use USB require an extra zero length data packet
462 to signal end-of-transfer on bulk endpoints, if the last data packet
463 is exactly wMaxPacketSize bytes long.
465 This flag allows applications to inform libusb about this requirement,
466 so that libusb can handle the issue transparently.
468 At the moment the new flag is only supported on Linux, and submitting
469 a transfer with the flag set returns an error at submit time on other
470 systems. Hopefully implementations will soon follow for other systems.
474 commit 7e0c747cc5ed0a31f38dff5eb5f03d4043b05ada
475 Author: Peter Stuge <peter@stuge.se>
476 Date: Sun Feb 19 06:58:16 2012 +0100
478 libusb.h: Add comma after last value in enums to ease future expansion
480 commit c20ec7a927c05e0b960269751c26ab2b6e0b39ab
481 Author: Peter Stuge <peter@stuge.se>
482 Date: Sun Feb 19 05:39:33 2012 +0100
484 Linux: Rework bulk continuation check to generic kernel version check
486 The old check_flag_bulk_continuation() tested for one specific running
487 kernel version. The new kernel_version_ge() instead allows to test the
488 running kernel version against major, minor and sublevel parameters.
490 commit cf8f5a7be821099b143fcea8e4ba553daa22818b
491 Author: Peter Stuge <peter@stuge.se>
492 Date: Sun Feb 19 05:32:10 2012 +0100
494 OpenBSD: Correctly catch attempts to do isochronous OUT transfers
496 LIBUSB_ENDPOINT_OUT is the value of the direction bit, which is 0 for
497 OUT transfers, so the previous condition could never evaluate to true.
499 commit ae8e86cffb81858af5eeddbf2ea1aec6d55f573d
500 Author: Peter Stuge <peter@stuge.se>
501 Date: Sun Feb 19 00:42:20 2012 +0100
503 Darwin: Do not reject devices with product id 0x0000
507 commit 4630fc22cff8ad3e1afa9b223378c0aabe282b5c
508 Author: Peter Stuge <peter@stuge.se>
509 Date: Mon Feb 13 16:59:51 2012 +0100
511 Do not call timerfd functions when timerfd is not being used
513 When libusb was built with timerfd support but used on a system
514 without timerfd support the library would hang indefinitely on
515 completion of the first transfer, since timerfd functions were
516 being called unconditionally and the error returned when timerfd
517 was not being used caused a confused internal state.
519 Many thanks to Ivo Smits for looking into the issue, proposing
520 an initial solution, and helping with testing! Fixes #73.
522 commit c40f3df1efbb9232b525e5f675746762ca2304b8
523 Author: Michael Plante <michael.plante@gmail.com>
524 Date: Sun Feb 12 13:49:35 2012 -0500
526 Windows: Fix MSVC6 IDE warning on line continuation in libusb-1.0.rc
528 * libusb\libusb-1.0.rc(21): Could not find the file LIBUSB_RC.
529 * only shows up first time after opening workspace.
530 * shows up on either build or clean.
531 * not actually due to rc.exe, but prior to it.
532 * probably an IDE bug.
533 * does not show up when running an exported makefile.
535 Signed-off-by: Michael Plante <michael.plante@gmail.com>
537 commit 5526841023101bafa280755c3bd19fc44b9c32a0
538 Author: Pete Batard <pete@akeo.ie>
539 Date: Sat Jan 7 23:30:23 2012 +0000
541 Windows: Add Renesas SuperSpeed support for libusb_get_device_speed()
543 * tests conducted using a Renesas PCIE USB 3.0 controller and driver and
544 a mass storage USB 3.0 confirm that USB_NODE_CONNECTION_INFORMATION_EX
545 uses (undocumented) value 3 for SuperSpeed
547 commit 3df437680f8f5406795ec5f32d96612d327b2ca4
548 Author: Peter Stuge <peter@stuge.se>
549 Date: Fri Feb 10 18:06:43 2012 +0100
551 autogen.sh: Try to use libtoolize before trying glibtoolize
553 This helps on Mac OS X where an old glibtoolize is included in the
554 system and newer, manually installed, versions provide libtoolize.
556 See also http://marc.info/?m=132490560131894
558 commit 4afbc78fe09bfc3dd7a609e5f82e9272c9e6226d
559 Author: Peter Stuge <peter@stuge.se>
560 Date: Wed Feb 8 15:54:21 2012 +0100
562 Rename the lsusb example program to listdevs
564 This is intended to reduce confusion with the much more significant
565 lsusb utility which is part of the usbutils package.
567 commit 57ab55159298b7057e0dd9b3854f7fd1550588ed
568 Author: Sean McBride <sean@rogue-research.com>
569 Date: Fri Dec 9 13:44:21 2011 -0500
571 core.c: Fix clang warning about useless store to a variable
573 commit 32bbbd133e24c24d02646c0f7d1963404d4955f2
574 Author: Sean McBride <sean@rogue-research.com>
575 Date: Fri Dec 9 13:43:46 2011 -0500
577 Darwin: Clarify comments on event thread creation and termination
579 commit 6cf3285a4e0dfaec0fe53b8eb6a9281e3c5bcb6c
580 Author: Sean McBride <sean@rogue-research.com>
581 Date: Fri Dec 9 13:43:46 2011 -0500
583 Darwin: Fix incorrect mutex protection of libusb_darwin_acfl global
585 commit 662ced422cf1ffce68753296a81dccf87f3a1c57
586 Author: Sean McBride <sean@rogue-research.com>
587 Date: Fri Dec 9 13:43:46 2011 -0500
589 Darwin: Use retained runloop variable instead of CFRunLoopGetCurrent()
591 commit 952a9bb6116afc69c4bae4714c2eb0713bf62a6d
592 Author: Sean McBride <sean@rogue-research.com>
593 Date: Fri Dec 9 13:37:22 2011 -0500
595 Darwin: Fix libusb_init() and libusb_exit() refcount race condition
597 Use OSAtomicIncrement32Barrier() and OSAtomicDecrement32Barrier()
598 in darwin_init() and darwin_exit() to be thread safe.
600 commit 721cc1c19fc8e88c4482ab3d1e8a34fa100c403c
601 Author: Sean McBride <sean@rogue-research.com>
602 Date: Fri Dec 9 13:33:28 2011 -0500
604 threads_posix.c: Fix prototype warning by including threads_posix.h
606 commit f6ef0fca4b36657c1a0885c18cc4c4afcb4f7339
607 Author: Dave Camarillo <dave.camarillo@gmail.com>
608 Date: Fri Sep 23 11:56:58 2011 +0100
610 Windows: Fix crash during enumeration with >= 64 connected devices
612 * indexes were outgrowing the unref_list array before realloc,
613 resulting in out of bound access and crash.
615 commit e280760331e8b39ad3530235653644654166b322
616 Author: Pete Batard <pbatard@gmail.com>
617 Date: Tue Oct 26 16:31:13 2010 +0100
619 Windows MSVC6: Remove SetupAPI, AdvAPI32 and OLE32 dependencies
621 * this ensures that libusb dependent applications only need
622 to link with libusb on Windows
623 * (copied from Pete's msvc08 mods to msvc6 by Michael)
625 commit fa66c93253855c9146011a6854d65df6d390039f
626 Author: Pete Batard <pbatard@gmail.com>
627 Date: Tue Dec 13 21:18:10 2011 -0500
629 Windows: Output an error message on calls to libusb_get_pollfds()
631 Signed-off-by: Michael Plante <michael.plante@gmail.com>
633 commit 637145ed29ee7c7b13ac45d10cf1f798fc06cbd9
634 Author: Pete Batard <pbatard@gmail.com>
635 Date: Thu Jun 3 22:14:20 2010 -0500
637 io.c: Fix comment since pthreads isn't the only threading libusb uses
639 commit bb37939ab44a54a054f73e2f381808b12fb9cb33
640 Author: Michael Plante <michael.plante@gmail.com>
641 Date: Mon Aug 2 19:59:12 2010 -0500
643 libusb.h: Cast in libusb_get_string_descriptor() to avoid MSVC6 warning
645 On MSVC6 bitwise OR promotes to int, causing the warning.
647 commit ca09754471acba9a22fe8fc09170e5660d1d5c5d
648 Author: Michael Plante <michael.plante@gmail.com>
649 Date: Fri Dec 23 16:50:21 2011 -0600
651 Windows: guid_to_string() is only discarded when macro-usbi_dbg used
653 * usbi_dbg encloses all references to guid_to_string
654 * MinGW/cygwin warn about an unused function, so the #if
655 squelches this warning
656 * MSVC6 uses a variadic function form of usbi_dbg instead
657 of a macro, so the compiler still "sees" guid_to_string
658 and it therefore needs to always be defined for MSVC6,
659 even if it's only a stub.
660 * So we define it if usbi_dbg is used OR if MSVC6 is used.
662 Signed-off-by: Michael Plante <michael.plante@gmail.com>
664 commit 8e15adcb0089882d10269ab77d35e193770ab90e
665 Author: Pete Batard <pete@akeo.ie>
666 Date: Tue Nov 15 01:11:12 2011 +0000
668 Windows: Fix undefined datatype 'LONG_PTR' in MSVC6
670 * issue reported by Elmi
672 Signed-off-by: Michael Plante <michael.plante@gmail.com>
674 commit 17ad642cae618cc09c5516612fa3880d4b6b8635
675 Author: Pete Batard <pete@akeo.ie>
676 Date: Sun Nov 13 23:10:15 2011 +0000
678 Windows: Fix for MSVC6 preprocessor not accepting blank parameters
680 * unlike later iterations of Visual Studio, MSVC6 does not accept
681 blank parameters on macro calls [eg. CALL(a, ,b)]
682 * blank params were used with the DLL_DECLARE and DLL_LOAD macros
683 * issue reported by Elmi
685 commit c591a7cbf7dbc6fbb4368e65cf58a8cd02582aeb
686 Author: Michael Plante <michael.plante@gmail.com>
687 Date: Tue Oct 26 11:39:38 2010 +0100
689 Windows: Switch MSVC6 .dsp project files from MBCS to Unicode
691 * MBCS (which is different from UTF-8) only makes sense if
692 supporting Windows 95/98, which we don't
693 * (try to match Pete's vcproj changes in MSVC6)
695 commit 536e424fd7d057fc3f20f9939f9de4572e51c815
696 Author: Michael Plante <michael.plante@gmail.com>
697 Date: Tue Dec 13 21:39:40 2011 -0500
699 Windows: Add missing MSVC6 lsusb.dsp project file
701 Signed-off-by: Michael Plante <michael.plante@gmail.com>
703 commit c56828857eb7ba6fb4f7701693dbb3fe4e75d5f6
704 Author: Martin Pieuchot <mpi@openbsd.org>
705 Date: Mon Jan 30 09:59:30 2012 +0100
709 commit d2285744c7fa4007bb411be354268209d350b0f2
710 Author: Xiaofan Chen <xiaofanc@gmail.com>
711 Date: Fri Jan 13 18:07:34 2012 +0100
713 configure.ac: Darwin: Move -lobjc from LIBS to PC_LIBS_PRIVATE
715 Since commit 40327cd134718475f6cec8935b856d4fdff2099c it is neccessary
716 to explicitly include -lobjc not only when linking libusb itself, but
717 also for programs linking statically against libusb. References #63.
719 See also http://marc.info/?m=132505900202378
721 commit 93b0e09d53ed1d177631af9182378492481a790a
722 Author: Xiaofan Chen <xiaofanc@gmail.com>
723 Date: Fri Jan 13 13:04:00 2012 +0100
725 examples: Link only with ../libusb/libusb-1.0.la and not with -lusb-1.0
727 Previous _LDFLAGS included both the freshly built libusb in ../libusb
728 and -lusb-1.0, where libtool would usually resolve the latter to an
729 already-installed libusb library in the system. The extra reference
730 to a second libusb library resulted in failure to build examples on
731 Mac OS X in some cases, and is plain wrong.
733 See also the thread at http://marc.info/?m=132637593623667
735 commit cdc5df8f1b206a7c19bc221e8a9a97d5d98deb85
736 Author: Peter Stuge <peter@stuge.se>
737 Date: Thu Jan 12 10:34:55 2012 +0100
739 libusb/Makefile.am: Avoid recursive variables in .rc silent-rule
741 Not every make supports recursive variable expansion so some automake
742 versions complain about non-POSIX variable names ever since commit
743 70bec4a9f8ec28d36c731011fa24d37c74ad3523 which added support for
744 silent-rules in our rule to compile the Windows .rc file.
746 This commit removes the recursive variables and instead uses the
747 simple and generic GEN message and associated variable.
749 commit a8e2e7cab278fbfcba91909eaf762cf66da9eda5
750 Author: Ludovic Rousseau <ludovic.rousseau+github@gmail.com>
751 Date: Sat Sep 24 18:27:54 2011 +0200
753 Core: Fix #122 warning: no previous prototype for ‘usbi_log_v’
755 Declare the usbi_log_v() function before using it.
757 commit e88ecb1e65604c222f82da586bb9e9066d66fbba
758 Author: Sean McBride <sean@rogue-research.com>
759 Date: Fri Oct 14 14:40:29 2011 -0400
761 Darwin: Set a name for the background thread on Mac OS X 10.6 and later
763 The call to pthread_setname_np() makes it easy to identify the
764 background thread in the Xcode debugger and in crash reports.
766 commit 60bad550dcb31a315a667a6614c52f712c82f45f
767 Author: Peter Stuge <peter@stuge.se>
768 Date: Thu Sep 22 21:42:35 2011 +0200
770 libusb/Makefile.am: Add libusb-1.0.rc dependency on version.h
772 commit 70bec4a9f8ec28d36c731011fa24d37c74ad3523
773 Author: Peter Stuge <peter@stuge.se>
774 Date: Thu Sep 22 21:38:29 2011 +0200
776 libusb/Makefile.am: Make the libusb-1.0.rc rule support silent-rules
778 commit fa9b682a638e54a4556ef6bd343ef56d6300a520
779 Author: Peter Stuge <peter@stuge.se>
780 Date: Thu Sep 22 21:17:37 2011 +0200
782 configure.ac: Enable libtool support for Windows Resource language
784 commit 58c73ddca310f757fc7ca75197bf3d7ca58ce29a
785 Author: Sean McBride <sean@rogue-research.com>
786 Date: Mon Sep 19 12:52:44 2011 -0400
788 descriptor.c: Use size_t for minor 64 bit compliance improvement
792 commit 27ec326a9f1bff4481702b35f39897c65a5db586
793 Author: Sean McBride <sean@rogue-research.com>
794 Date: Mon Sep 19 12:42:33 2011 -0400
796 Darwin: Fix warnings about mismatched signed/unsigned compare
800 commit 59cc79f2c8e5da41fe2c761bf1e1d9cd18ddac81
801 Author: Sean McBride <sean@rogue-research.com>
802 Date: Mon Sep 19 12:38:20 2011 -0400
804 Fix unused variable warnings when without timerfd and/or when on Darwin
808 commit 1cc5b4a9fb984e83681ae5c797fa6b22bc20f809
809 Author: Ludovic Rousseau <ludovic.rousseau+github@gmail.com>
810 Date: Fri Sep 16 18:07:56 2011 +0200
812 Correctly handle LIBUSB_TRANSFER_OVERFLOW in libusb_control_transfer()
814 sync.c: In function `libusb_control_transfer':
815 sync.c:122: warning: enumeration value `LIBUSB_TRANSFER_OVERFLOW' not
820 commit c55f581f43885f8576d8e2ebf4d46cd7ff3c1030
821 Author: Pete Batard <pete@akeo.ie>
822 Date: Mon Sep 19 16:01:37 2011 +0100
824 Windows: do not set altsetting 0 as part of libusb_release_interface()
826 * pointed out by Travis Robinson and Xiaofan Chen
827 * similar to a change advised by Alan Stern for the Linux kernel:
828 http://marc.info/?m=122790204410765
830 commit efd3fae8f95ef7797ecd4b5731b198f100007d6a
831 Author: Pete Batard <pete@akeo.ie>
832 Date: Mon Sep 19 15:45:35 2011 +0100
834 Windows: fix communication with hubs
836 * issue reported by René Haunstrup in http://marc.info/?m=130503019227814
838 commit a8d1eb7963e5811e16e8fb7e157cf3a00d888347
839 Author: Pete Batard <pete@akeo.ie>
840 Date: Mon Sep 19 15:41:39 2011 +0100
842 Windows: add USB 3.0 controller support
844 * Windows 8, NEC/Renesas, TI, Fresco Logic, Etron, VIA, ASMedia
845 (some of which untested!)
846 * includes workaround for NEC/Renesas USB 3.0 root hubs
848 commit 48a19497bdd14325738ac9197d8cf4ebe1daa2d4
849 Author: Pete Batard <pete@akeo.ie>
850 Date: Mon Sep 19 15:28:03 2011 +0100
852 Windows: add static keywords and drop duplicated or unneeded code
854 commit af407286ba2fd346755297e56e365d84b98090d5
855 Author: Pete Batard <pete@akeo.ie>
856 Date: Mon Sep 19 13:59:46 2011 +0100
858 Windows: fix 'windows_error_str' defined but not used when not logging
860 commit abc3143ba1989a48f62246c2afb0eb4c3c5f7917
861 Author: Peter Stuge <peter@stuge.se>
862 Date: Thu Sep 22 10:19:17 2011 +0200
864 libusbi.h: Silence unused context variable warnings when not logging
866 commit 69b71c896a153a307360cf21be67b3b6e1e09a2d
867 Author: Pete Batard <pete@akeo.ie>
868 Date: Mon Sep 19 13:06:22 2011 +0100
870 Windows: copy DLL file regardless of ddk_build.cmd dll argument case
872 commit fdbac8971212bda3a42feeeaeb92fd8e49c7e904
873 Author: Pete Batard <pete@akeo.ie>
874 Date: Sun Sep 18 20:06:45 2011 +0100
876 Windows: Add uint32_t typedef for old MSC so that DDK can build lsusb
878 commit c471d55e909d84012f065715703dad454d1fa421
879 Author: Vitali Lovich <vlovich@gmail.com>
880 Date: Fri Sep 16 08:53:25 2011 -0600
882 Darwin: fix port leak in darwin_clock_gettime()
884 commit 477f3917f8c123e36993f9c0ba8156ca0f16ccfa
885 Author: Vitali Lovich <vlovich@gmail.com>
886 Date: Fri Sep 16 08:50:40 2011 -0600
888 Darwin: remove unnecessary run loop invalidation during libusb_exit()
890 commit f52094f37ccaff8137f5ce39867d2dba9b0f6bf9
891 Author: Vitali Lovich <vlovich@gmail.com>
892 Date: Fri Sep 16 08:49:54 2011 -0600
894 Darwin: fix port leak during libusb_exit()
896 commit 8a2f7b5c0520e6099f2c624cd0994c60867c1628
897 Author: Vitali Lovich <vlovich@gmail.com>
898 Date: Fri Sep 16 08:48:28 2011 -0600
900 Darwin: fix potential leak on libusb_claim_interface() error
902 commit b1ee2ef8d0b67dcc51c15742cb2decb386cffd9b
903 Author: Nathan Hjelm <hjelmn@me.com>
904 Date: Fri Sep 16 08:41:01 2011 -0600
906 Darwin: Fix #117 transfer size 64/32 bit issue for transfer callbacks
908 The transfer size is now always truncated to 32 bits.
910 commit c407a2d97b6f8715d3452257a0adac40a20178e6
911 Author: Peter Stuge <peter@stuge.se>
912 Date: Thu Sep 15 17:04:43 2011 +0200
914 Remove username from SF FRS rsync upload command
916 commit cbc2d7b2de58cf5fb02d9e5a39be0870527f8632
917 Author: Peter Stuge <peter@stuge.se>
918 Date: Thu Sep 15 16:48:35 2011 +0200
920 Support release candidate versions in configure.ac and libusb-1.0.rc
922 commit 384a528ed7adde935681313dacc5cd79e4be9ea1
923 Author: Peter Stuge <peter@stuge.se>
924 Date: Thu Sep 15 16:32:47 2011 +0200
926 Tidy .gitignore a little
928 There's no xusb, we need a glob pattern for matching .exe files, and
929 fpusb tarballs aren't very common in the libusb source dir.
931 commit 9ad3643391a7fc6ce97aa77ffaf97d244fd9e613
932 Author: Hans de Goede <hdegoede@redhat.com>
933 Date: Wed Sep 14 10:53:48 2011 +0200
935 Docs: Clarify that libusb_handle_events_timeout() tv param can't be NULL
937 The example code and API doc for libusb_handle_events_timeout() could be
938 interpreted as it being OK to pass a NULL pointer for the tv argument (I
939 interpreted it like that when I first started coding for libusb).
941 This patch changes the docs to make it clear that one must always supply
942 a tv struct to libusb_handle_events_timeout.
944 Signed-off-by: Hans de Goede <hdegoede@redhat.com>
946 commit e1c06fb070c1328a3b78c7da8b9e779d466edd97
947 Author: Peter Stuge <peter@stuge.se>
948 Date: Wed Sep 14 08:23:43 2011 +0200
950 Makefile.am: Do not remove ChangeLog on make distclean, fixes #116
952 [pbatard: Remove it on make maintainer-clean instead]
954 commit 0c5bf03eb829e51dcf19562fc4f745937235ea51
955 Author: Vitali Lovich <vlovich@aliph.com>
956 Date: Wed Sep 14 07:02:06 2011 +0200
958 Darwin: Reduce race likelihood between cancellation and device disconnect
960 References #88. The race condition still remains, but this change
961 makes it less likely to trigger.
963 commit 3e6e257f55156c4600f3116a0befba8cee62872a
964 Author: Peter Stuge <peter@stuge.se>
965 Date: Wed Sep 14 04:00:33 2011 +0200
967 Document that libusb_get_device_list() can return any libusb_error code
971 commit 737ba04ea40f19564b445fbb489907529e75edc9
972 Author: Pete Batard <pbatard@gmail.com>
973 Date: Thu Jun 16 10:49:13 2011 +0100
975 Linux: Fix #81 URB double free race condition on device disconnect
977 A submitted transfer that has just been failed by the kernel could be
978 picked up by an event handler to be cleaned up, where freeing of URB
979 memory would race with the submit function doing it's own cleanup and
980 freeing as a result of the submit failing.
982 libusb_submit_transfer() always holds itransfer->lock, so the race can
983 be avoided by taking that lock also in the cleanup path and checking
984 that the URB memory has not already been freed before freeing it there.
986 As http://libusb.org/ticket/81#comment:14 notes there is still another
987 possible, but unlikely, race condition between libusb_submit_transfer()
988 and an event handling thread. That will require more work to solve.
990 [stuge: Add check in cleanup path that URBs have not already been freed]
992 commit 1b0f758a334d78d3832d17d694ac0c931da36050
993 Author: Peter Stuge <peter@stuge.se>
994 Date: Wed Sep 14 03:00:14 2011 +0200
996 Add libusb_error_name() API function
998 Until we have i18ned error messages we can at least offer easy
999 translation from numeric error code to the error code's name.
1001 [hansg: switch() on enum to get a compiler warning if a case is missing]
1002 [hansg: Better return value when called with an unknown error code]
1003 [hansg: Fix DEFAULT_VISIBILITY in definition to compile on all platforms]
1005 commit e16805130a04b2642891e86b8f479e6343ca17da
1006 Author: Peter Stuge <peter@stuge.se>
1007 Date: Wed Sep 14 02:30:18 2011 +0200
1009 Add libusb_has_capability() API function
1011 Since it is currently not planned to change the filename of the libusb
1012 library between libusb-1.0 versions at least on Windows systems it's
1013 important to have a capability API.
1015 Currently there exists only one capability; LIBUSB_CAP_HAS_CAPABILITY,
1016 which tests support for the libusb_has_capability() API.
1018 commit 2dc2fa2a56408e515d25359323b4ea2f1bbec14a
1019 Author: Sebastian Pipping <sebastian@pipping.org>
1020 Date: Tue Sep 13 20:38:26 2011 +0200
1022 Fix typos in introductory documentation
1024 commit 0a828082aeedc6d19b6de4e3b10cc9762352fa7f
1025 Author: Pekka Nikander <pekka.nikander@senseg.com>
1026 Date: Tue Sep 13 18:57:43 2011 +0200
1028 Linux: Add libusb_open() debug and error messages with the device path
1032 commit 6583372125d73543fe0db9fdc59850b9c6aa50d9
1033 Author: Pekka Nikander <pekka.nikander@senseg.com>
1034 Date: Tue Sep 13 18:52:36 2011 +0200
1036 Add debug message with libusb_open() error code
1040 commit f797ecb3b4e653594f8ebc0c9d9e2ca579062779
1041 Author: Hans de Goede <hdegoede@redhat.com>
1042 Date: Mon Sep 12 18:11:20 2011 +0200
1044 Document libusb_handle_events_completed() and _timeout_completed()
1046 Signed-off-by: Hans de Goede <hdegoede@redhat.com>
1047 [stuge: Note that the old racy functions should be avoided by new code]
1049 commit 6696512aade99bb15d6792af90ae329af270eba6
1050 Author: Graeme Gill <graeme2@argyllcms.com>
1051 Date: Sat Sep 10 15:47:05 2011 +0200
1053 Fix #56 race condition causing delayed completion of sync transfers
1055 The sync API had a race where it would check a condition to know if it
1056 needed to call a libusb_handle_events() function. However, the check
1057 was done outside of the lock that is held while the condition is set,
1058 so another thread could completely serve whatever was needed to make
1059 the condition true between it being checked and the event handler being
1060 called. This situation would be detected after a libusb-internal timeout
1061 of 60 seconds, after which the transfer would be completed without
1062 error, but with significant delay.
1064 Original patch at http://marc.info/?l=libusb-devel&m=127252114815709
1066 Changes by Hans de Goede:
1067 - Renamed the "race-proof" functions from libusb_handle_events*_check()
1068 to libusb_handle_events*_completed()
1069 - Drop r = 0 setting in libusb_handle_events_timeout_completed()
1070 (to make both completed checking cases identical flow wise)
1072 Signed-off-by: Hans de Goede <hdegoede@redhat.com>
1073 [stuge: Simplify libusb_handle_events_timeout() change with a goto]
1074 [pbatard: Fix _handle_events_timeout() and _completed() definitions]
1076 commit cfee5446cdb2974c1e90c7c379a365c495975aef
1077 Author: Nathan Hjelm <hjelmn@me.com>
1078 Date: Fri Aug 26 23:06:40 2011 -0600
1080 Darwin: Use a condition to signal async thread ready in libusb_init()
1082 commit d1bd234f7e1b1ff4168b8f605e2038e87a0e361f
1083 Author: Nathan Hjelm <hjelmn@me.com>
1084 Date: Thu Jul 21 22:34:12 2011 -0600
1086 Fix #64 use of reserved identifiers throughout libusb
1088 commit a97762a8a121de1ad3deb9e861088b69bb6a155b
1089 Author: Nathan Hjelm <hjelmn@me.com>
1090 Date: Thu Jul 21 22:31:09 2011 -0600
1092 Darwin: fix libusb_get_device_list() and libusb_claim_interface() leaks
1094 commit 8723df191fcedc6161732fd06d12df392c6dea3f
1095 Author: Nathan Hjelm <hjelmn@me.com>
1096 Date: Thu Jul 21 22:28:47 2011 -0600
1098 Darwin: use kIOMasterPortDefault instead of creating a master port
1100 commit 0eb7e43759003610d1afe6148aa6ff4f5859ccb0
1101 Author: Peter Stuge <peter@stuge.se>
1102 Date: Wed Jul 27 03:15:03 2011 +0200
1104 Allow devices with zero configurations to be discovered
1106 At least in Linux, wireless USB devices have zero configurations until
1107 they have been authorized. This device state can be set for any USB
1110 echo 0 > /sys/bus/usb/devices/?-?/authorized
1112 An unauthorized device would previously make usbi_sanitize_device()
1113 return LIBUSB_ERROR_IO, although there really was no I/O error.
1115 Reported-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
1117 commit 5010c3e80423d13acee6eeb4af974e8a9bcdd006
1118 Author: Peter Stuge <peter@stuge.se>
1119 Date: Wed Jul 27 06:17:48 2011 +0200
1121 Linux: Continue enumeration even if one device returns an error
1123 commit 2bf8848fb5c85ace0f445c0820a0c173cf76bf43
1124 Author: Hans de Goede <hdegoede@redhat.com>
1125 Date: Fri Jun 17 10:19:00 2011 +0200
1127 Linux: Handle single- and two-digit kernel versions like 3.0 and 4
1129 The code has been tested with various version strings.
1131 [stuge: Remove bogus string length check and optimize for newer kernels]
1133 commit 0288bf24f91eca9f433ebad91a7eb6c0e3e29133
1134 Author: Pete Batard <pbatard@gmail.com>
1135 Date: Thu Mar 3 18:35:35 2011 +0000
1137 Windows: Provide libusb_get_device_speed() data
1139 * use (currently superspeed unaware) USB_NODE_CONNECTION_INFORMATION_EX
1143 [stuge: Explicitly convert OS device speeds to libusb device speeds]
1144 [pbatard: Add breaks overlooked when adding explicit conversion]
1146 commit 382b7deaadd7e2f1d08c9e921e460bb3abf130b2
1147 Author: Trygve Laugstøl <trygvis@inamo.no>
1148 Date: Tue Jul 19 23:20:40 2011 +0200
1150 Darwin: Provide libusb_get_device_speed() data
1154 commit 44d6319f57e7e2c77347da228ed032e62716b438
1155 Author: Hans de Goede <hdegoede@redhat.com>
1156 Date: Thu Feb 17 15:06:40 2011 +0100
1158 Linux: Provide libusb_get_device_speed() data from sysfs
1162 Signed-off-by: Hans de Goede <hdegoede@redhat.com>
1164 commit c8c40bcbd22abca922ff31c4ed6a75fd7af78b59
1165 Author: Hans de Goede <hdegoede@redhat.com>
1166 Date: Thu Feb 17 13:17:55 2011 +0100
1168 Add new API function libusb_get_device_speed()
1170 Reworked version of my earlier patch, largely modelled after (if not
1171 copy and pasted from) Trygve Laugstøl's <trygvis@inamo.no> similar
1172 commit 5a6541e0d80fb1f21e2b960bc2337a612f9d74fb in
1173 git://git.libusb.org/libusb-trygvis.git
1174 http://git.libusb.org/libusb-trygvis.git
1175 http://git.libusb.org/?p=libusb-trygvis.git;a=commitdiff;h=5a6541e0d
1177 This patch does not add any OS specific code. The supporting code in
1178 each backend follows in separate patches. References #45.
1180 Signed-off-By: Hans de Goede <hdegoede@redhat.com>
1182 commit 0db8cba7f2c372f0127c7df9724e5a35e93f2069
1183 Author: Nathan Hjelm <hjelmn@me.com>
1184 Date: Sun Mar 20 00:25:44 2011 -0600
1186 Darwin: Improve device enumeration performance
1188 The device location is saved, and later used whenever iterating over
1189 devices. Fixes #111 and #112.
1191 [stuge: Formatting fixes and split out libusb_get_device_speed() change]
1192 [stuge: Combine with patch that uses the saved device location]
1194 commit 511ec4301deb516691e39e3df5b90fe28cb2bb0f
1195 Author: Hans de Goede <hdegoede@redhat.com>
1196 Date: Thu Feb 17 12:49:36 2011 +0100
1198 Linux: Use __read_sysfs_attr() to get busnum and devaddr
1200 Remove code duplication.
1202 Signed-off-by: Hans de Goede <hdegoede@redhat.com>
1204 commit e05bbc59ca7b896e2b825bde59bf883ac8c09f5c
1205 Author: Hans de Goede <hdegoede@redhat.com>
1206 Date: Thu Feb 17 12:45:10 2011 +0100
1208 Linux: Add __read_sysfs_attr() helper function
1210 On Linux we often need to read (postive) integers from sysfs. This patch
1211 adds a helper function for this. This is a preparation patch for adding
1212 a libusb_get_device_speed() function to libusb.
1214 Signed-off-by: Hans de Goede <hdegoede@redhat.com>
1216 commit 20e26dda0f86dafa3436b152e090b5f3410c8ce4
1217 Author: Hans de Goede <hdegoede@redhat.com>
1218 Date: Tue Feb 8 16:37:18 2011 +0100
1220 Linux: Stop kernel from re-attaching in-kernel driver after reset
1222 When an interface is bound to the usbfs driver (iow claimed), the
1223 kernel will unbind it, and then after the reset do a device_attach
1224 on the interface, which will bind the default in kernel driver to
1227 So if an app has detached the in kernel driver, and claimed the
1228 interface and then does a libusb_reset_device. Things end up with
1229 the interface no longer being bound to the usbfs driver (so no longer
1230 claimed) and instead it is bound to the in kernel driver (iow the in
1231 kernel driver is re-attached).
1233 We can stop this from happening by releasing all claimed interfaces
1234 before the reset, as the kernel will not do the device attach after
1235 reset, if no driver was bound to the interface before the reset.
1237 Signed-off-by: Hans de Goede <hdegoede@redhat.com>
1239 commit 4e294dc9848aab86b31d2fff228cf97e7eadf913
1240 Author: Hans de Goede <hdegoede@redhat.com>
1241 Date: Thu Feb 17 12:35:46 2011 +0100
1243 Core: Use USB_MAXINTERFACES rather then sizeof(claimed_interfaces)
1245 claimed_interfaces is a long and thus its size differs on 32
1246 versus 64 bits, this patch uses USB_MAXINTERFACES instead of
1247 sizeof(claimed_interfaces) on interface number range checks
1248 for consistent behavior independent of the arch.
1250 Signed-off-by: Hans de Goede <hdegoede@redhat.com>
1252 commit a68f956dd4c29d6ed799f04f07a3346317843999
1253 Author: Hans de Goede <hdegoede@redhat.com>
1254 Date: Wed Feb 9 12:40:35 2011 +0100
1256 Linux: Fix warnings seen when compiling with gcc-4.6
1258 Compiling with gcc-4.6 results in the following warnings:
1259 os/linux_usbfs.c: In function 'op_get_configuration':
1260 os/linux_usbfs.c:1144:6: warning: variable 'r' set but not used ...
1261 os/linux_usbfs.c: In function 'op_handle_events':
1262 os/linux_usbfs.c:2091:40: warning: 'status' may be used uninitialized ...
1263 os/linux_usbfs.c:2044:6: note: 'status' was declared here
1264 dpfp.c: In function 'save_to_file':
1265 dpfp.c:228:9: warning: variable 'ignore' set but not used ...
1266 dpfp_threaded.c: In function 'save_to_file':
1267 dpfp_threaded.c:257:9: warning: variable 'ignore' set but not used ...
1269 This patch fixes these. (The second comes from handle_control_completion()
1270 which gcc seems to inline into op_handle_events().)
1272 Signed-off-by: Hans de Goede <hdegoede@redhat.com>
1273 [stuge: Add fixes for dpfp examples and update source references]
1275 commit 209b5ccd827039b933b730e3f29abd0d72d3c6c6
1276 Author: Vitali Lovich <vlovich@aliph.com>
1277 Date: Tue Jun 14 08:51:57 2011 +0200
1279 Darwin: Fix #108 crash when a device is unplugged
1281 commit d3543e9140536a9751811b44f8496489be1e6bb7
1282 Author: Peter Stuge <peter@stuge.se>
1283 Date: Mon Jun 13 22:58:23 2011 +0200
1285 libusb.h: Add missing LIBUSB_CLASS_ codes currently defined by usb.org
1287 LIBUSB_CLASS_PHYSICAL
1288 LIBUSB_CLASS_SMART_CARD
1289 LIBUSB_CLASS_CONTENT_SECURITY
1291 LIBUSB_CLASS_PERSONAL_HEALTHCARE
1292 LIBUSB_CLASS_DIAGNOSTIC_DEVICE
1294 Per http://www.usb.org/developers/defined_class
1296 commit 08bc7ffb6a3b65afc6795cd048a890ee8da7d954
1297 Author: Peter Stuge <peter@stuge.se>
1298 Date: Mon Jun 13 22:53:40 2011 +0200
1300 libusb.h: Add LIBUSB_CLASS_IMAGE in addition to LIBUSB_CLASS_PTP
1302 The old LIBUSB_CLASS_PTP name is kept for compatibility. Fixes #106.
1304 commit 3d7938d85a33cd98553c0c898a8dd0490826b077
1305 Author: Peter Stuge <peter@stuge.se>
1306 Date: Mon Jun 13 22:52:48 2011 +0200
1308 libusb.h: Fix LIBUSB_CLASS_PRINTER description typo
1312 commit 5d0526af1a922afab02cdbdf4bb7078702759571
1313 Author: Sean McBride <sean@rogue-research.com>
1314 Date: Fri Feb 11 11:28:04 2011 -0500
1316 Darwin: Fix #100 warning cast to pointer from integer of different size
1318 commit bb8854ff5680136e29b57505300f131c2646f258
1319 Author: Sean McBride <sean@rogue-research.com>
1320 Date: Mon Jun 13 22:26:43 2011 +0200
1322 configure.ac: Fix #97 clang warning about -fgnu89-inline
1324 Remove the test that sets the flag since GNU89 inline semantics are not
1327 [stuge: Also remove reference to the test result]
1329 commit cb3c2e488f298b78027c9d5fc0bec0e5ba27d761
1330 Author: Pete Batard <pbatard@gmail.com>
1331 Date: Wed Jun 15 01:38:58 2011 +0200
1333 Windows: Make cancel_io() in poll_windows.c static
1337 commit 6635dbb4acb5ee556a020e431b15c954edfcbd29
1338 Author: Peter Stuge <peter@stuge.se>
1339 Date: Mon Jun 13 22:21:02 2011 +0200
1341 Windows: Make upperize() in windows_usb.h static
1345 commit 9479ce9dae3d0dbe2ce693bd02ccceb018a75fe9
1346 Author: Vitali Lovich <vlovich@aliph.com>
1347 Date: Wed Mar 16 17:33:57 2011 -0700
1349 descriptor.c: Fix buffer read overflow caught by valgrind
1351 In parse_interface() an unexpected descriptor would be parsed without
1352 validating the descriptor's length. It is possible for size to be 0 at
1353 this point, which means that the parsing would read past the end of the
1354 source buffer. Fix #83 by checking the length of the remaining buffer
1357 commit c775c2f43037cd235b65410583179195e25f9c4a
1358 Author: Vitali Lovich <vlovich@aliph.com>
1359 Date: Wed Mar 16 19:51:40 2011 -0700
1361 Clean up in-flight transfers and device handle when closing a device
1363 Any in-flight transfers should properly invalidate their references
1364 to device handles that are being closed. Additionally, they should be
1365 removed from the transfer-in-flight list. This is done with the events
1366 lock held to protect against another thread processing the same transfer.
1367 The events lock is initialized as a recursive mutex, because the device
1368 close code might itself be called while an event is being handled.
1371 [stuge: Trivial rework to reduce indenting]
1373 commit 74282582cc879f091ad1d847411337bc3fa78a2b
1374 Author: Vitali Lovich <vlovich@aliph.com>
1375 Date: Wed Mar 16 19:43:30 2011 -0700
1377 Add recursive mutexes to threading abstraction
1379 This is necessary for the device close path which needs to attain the
1380 events lock, but which might itself be called while handling an event.
1381 The events lock is necessary to properly clean up transfers which might
1382 still be pointing to the device. References #82.
1384 [stuge: Move usbi_mutex_init_recursive() into threads_posix.c]
1385 [stuge: Must also #define _XOPEN_SOURCE 500 to be able to build]
1386 [pbatard: Un-inline usbi_mutex_init_recursive() to make Cygwin happy]
1388 commit 8f1bc0659136faf312dabe8fe67380a798299d33
1389 Author: Vitali Lovich <vlovich@aliph.com>
1390 Date: Wed Mar 16 19:48:35 2011 -0700
1392 Add USBI_TRANSFER_CANCELLING and _DEVICE_DISAPPEARED status flags
1394 The flags are used to indicate if a cancellation has started, and if
1395 a cancellation has failed because the device is no longer available.
1398 commit 87ba8a9ead4fd9e81ae3c5a894f698779de1f2c2
1399 Author: Vitali Lovich <vlovich@aliph.com>
1400 Date: Wed Mar 16 19:50:04 2011 -0700
1402 Linux: Detect and report that device has gone away while discarding URBs
1404 Return LIBUSB_ERROR_NO_DEVICE if the device is no longer available
1405 (ENODEV) to discard the urbs. References #82.
1407 commit 2c85c363cddd72b9ab425f05e2ee03e0b1df24f1
1408 Author: Peter Stuge <peter@stuge.se>
1409 Date: Mon Jun 6 02:46:46 2011 +0200
1411 libusbi.h: Add comma after last value in enum usbi_transfer_flags
1413 commit 5d7be3e1337f2444097504931ac0f2f0997b2b05
1414 Author: Peter Stuge <peter@stuge.se>
1415 Date: Sun Jun 5 23:28:07 2011 +0200
1417 threads_posix.h: Whitespace fix
1419 commit 18db4813e844fc6929680afcc9b793abcb06876a
1420 Author: Peter Stuge <peter@stuge.se>
1421 Date: Sun Jun 5 23:21:08 2011 +0200
1423 libusb/Makefile.am: Correct threading files in libusb_1_0_la_SOURCES
1425 commit 5f30c81f66e8dd61f8eae16de548697708f9bd18
1426 Author: Alan Ott <alan@signal11.us>
1427 Date: Thu Jul 21 16:37:48 2011 +0200
1429 Linux: Fix #70 race condition in sysfs_get_device_list()
1431 Change the way libusb chooses between using sysfs and usbfs for information
1432 about the attached devies. Using the old method, a race condition could
1433 occur if a device was unplugged just before (or during) the call to
1434 libusb_get_device_list(), corrupting the internal sysfs_can_relate_devices
1435 and sysfs_has_descriptors variables and preventing libusb_get_device_list()
1436 from working in future calls.
1438 The old method was based on the assumption that if certain sysfs files
1439 (eg: busnum) could not be opened, that indicated an inadequacy of sysfs
1440 (ie: the running kernel's sysfs version did not contain those files),
1441 when in reality those files couldn't be opened because the device had
1444 The new method checks the adequacy of sysfs during libusb_init()
1445 (op_init()) and if a sysfs file cannot be opened, it is now assumed that
1446 it is because the device has been unplugged, not because sysfs is
1449 Signed-off-by: Alan Ott <alan@signal11.us>
1450 [stuge: Include closedir() bugfix posted in ticket by Arne Laansoo]
1451 [stuge: Remove dead code in sysfs_scan_device() found by Hans de Goede]
1453 commit 40327cd134718475f6cec8935b856d4fdff2099c
1454 Author: Sean McBride <sean@rogue-research.com>
1455 Date: Thu Feb 10 17:11:29 2011 -0500
1457 Darwin: Fix #63 error when apps use Objective-C garbage collection
1459 commit 3ba2fae24886fec89410e5f2295f65363edcc2df
1460 Author: Sean McBride <sean@rogue-research.com>
1461 Date: Fri Feb 5 11:50:34 2010 -0500
1463 Darwin: Fix #28 clang analyzer warning about unbalanced retain/release
1465 commit b67120f047f7eafa15c88c66fa61cef40805ec1f
1466 Author: Sean McBride <sean@rogue-research.com>
1467 Date: Fri Feb 5 11:50:34 2010 -0500
1469 io.c: Fix clang analyzer warning about unused variable
1473 commit 51af85d03a098d7d99ea544e8e9535b5476bab33
1474 Author: Peter Stuge <peter@stuge.se>
1475 Date: Tue May 10 10:44:18 2011 +0200
1477 .gitignore: Add m4 subdirectory created by libtool
1479 commit c569e5b3b314538f45d0b3b6f14ce3dbdfee1be9
1480 Author: Ludovic Rousseau <rousseau@debian.org>
1481 Date: Mon Nov 15 14:45:50 2010 +0100
1483 Linux: Do not include useless pthread.h
1485 The code does not use threads so no need to #include <pthread.h>
1488 commit 82af9f9d440ef49015652c699157cf62020f097a
1489 Author: Pete Batard <pbatard@gmail.com>
1490 Date: Tue Mar 8 17:41:05 2011 +0000
1492 Windows: minor code improvements
1494 * mostly warnings from VS2010's Intellisense
1496 commit 7e636dbe06001b9ac68e452a2fdc05e33fbb4e3a
1497 Author: Pete Batard <pbatard@gmail.com>
1498 Date: Tue Mar 8 17:40:39 2011 +0000
1500 Windows: prevent set_configuration request from being sent using WinUSB
1502 * multiple WinUSB set configuration requests can stall a device and
1503 WinUSB does not support setting of alt configuration in the first place
1504 * issue reported and investigated by Xiaofan Chen
1506 commit 578c50c99646e5ebbcfbea653524a519b843ebcd
1507 Author: Graeme Gill <graeme@argyllcms.com>
1508 Date: Tue Mar 8 17:39:16 2011 +0000
1510 Windows: fixed default WinUSB timeout and ineffective policy settings
1512 * default WinUSB timeout was 5 seconds, which may be to small.
1513 PIPE_TRANSFER_TIMEOUT policy is now set to infinite for all EPs.
1514 * other WinUSB policy settings were called before the endpoints
1515 had been set. This is now fixed.
1517 commit 25fa63ea1cfe02e784cb62d3b869fd15e62b5b6e
1518 Author: Pete Batard <pbatard@gmail.com>
1519 Date: Tue Mar 8 17:37:40 2011 +0000
1521 Windows: Don't leak pipe fds
1523 use _open() and _close() rather than _open_osfhandle() and CloseHandle()
1525 * use of CloseHandle() prevented the pipe fds from being
1526 relinquished on libusb_exit()
1527 * leaked fds could lead to the OS running out of new fds
1528 and LIBUSB_ERROR_NO_MEM being returned as a result
1529 * using _open() avoids _open_osfhandle() redef for cygwin
1530 * issue reported by Stephano Antonelli
1532 commit 3ccd9bddec316ee4b867959fd7f616eb4d410997
1533 Author: Peter Stuge <peter@stuge.se>
1534 Date: Tue May 10 10:05:29 2011 +0200
1536 Windows: enumeration overhaul
1538 * uses multiple passes in a single call
1539 * uses a hash table, in anticipation for hotplug
1540 * adds a (dummy) HUB interface for harmonization
1541 * adds calloc on device struct creation in core
1542 (to detect and avoid unnecessary double initialization)
1544 commit bb108cbe39802785d10fe025b74ca5ee19734151
1545 Author: Pete Batard <pbatard@gmail.com>
1546 Date: Tue May 10 00:56:18 2011 +0100
1548 Windows: MS project files part 4 - MSVC6 project files
1552 commit 78896ae99b9c2d5672fc93c5bad0f5e8e75c53ea
1553 Author: Pete Batard <pbatard@gmail.com>
1554 Date: Tue May 10 00:53:09 2011 +0100
1556 Windows: MS project files part 3 - VS2005 project files
1560 commit 9e5f4277e379a82f5d32e81db1bd0ffaa6637596
1561 Author: Pete Batard <pbatard@gmail.com>
1562 Date: Tue May 10 00:38:49 2011 +0100
1564 Windows: MS project files part 2 - WDK project files
1568 commit c1e75431418cdfc10ec963d823fce479cdae6840
1569 Author: Pete Batard <pbatard@gmail.com>
1570 Date: Tue May 10 00:38:12 2011 +0100
1572 Windows: MS project files part 1 - support files & preliminaries
1574 * adds the MSVC support files used by MSVC6, VS2005 and WDK
1575 * also adds missing defines required by MS compilers in libusb.h
1576 * also anticipatory sets .gitattributes for CRLF handling
1577 * also adds manually maintained .def file for import lib generation
1581 commit 85a14f43e0ce27cd47f84fdaeacbaa81cebdb886
1582 Author: Peter Stuge <peter@stuge.se>
1583 Date: Mon May 9 08:12:24 2011 +0200
1585 Move library version number from configure.ac to libusb/version.h
1587 This is neccessary to support native MS builds. The Windows resource
1588 file libusb/libusb-1.0.rc must include the release version, which was
1589 previously only available after configure had run and had substituted
1590 the numbers into a generated libusb/libusb-1.0.rc file.
1592 The version atoms are now stored as CPP style #defines in libusb/version.h
1593 so that the .rc no longer needs to be generated but can simply include the
1594 header file and access the version information directly.
1596 The m4 macro LU_DEFINE_VERSION_ATOM() was added to configure.ac to get
1597 version atoms from libusb/version.h for use in AC_INIT(). The macro
1598 handles C and C++ style comments in version.h, but can easily be made
1599 to fail by obscuring the file. Please don't do that.
1601 Tested with MinGW using autoconf, and manual compile of libusb-1.0.rc
1602 using RC.EXE Version 5.2.3690.0 from Visual C++ 2005 Express Edition.
1604 commit 3481f52b014110f6c291e5a15e79ba774dcdfacf
1605 Author: Peter Stuge <peter@stuge.se>
1606 Date: Tue May 10 08:36:32 2011 +0200
1608 Windows: Debug builds create the same .dll file name as non-debug
1610 Thanks to Pete for pointing this out!
1612 commit 9da8cf5c9ed9bd381a77d8170fc0556edbcce4dc
1613 Author: Pete Batard <pbatard@gmail.com>
1614 Date: Tue Mar 8 17:18:48 2011 +0000
1616 Windows: explicit use of ANSI or WideChar calls
1618 * This is in anticipation of adding MSVC project files
1619 that use Unicode rather than MBCS
1621 commit ae675c1f49dc00b695fce7296903f60998cc28ed
1622 Author: Pete Batard <pbatard@gmail.com>
1623 Date: Tue Mar 8 17:16:23 2011 +0000
1625 Windows: Remove SetupAPI, AdvAPI32 and OLE32 link-time dependencies
1627 * this ensures that libusb dependent applications only need
1628 to explicitly link against libusb on Windows
1630 The run-time dependency on Cfgmgr32.dll, OLE32.dll and SetupAPI.dll remains.
1632 commit 356e312511e4892977636da52ec497ad2c716032
1633 Author: Pete Batard <pbatard@gmail.com>
1634 Date: Tue Mar 8 17:12:33 2011 +0000
1636 Add .gitattributes to avoid Windows line ending issues with autocrlf
1638 commit 61c120f025f3dee04eeb88159d7c0403a8bf91fe
1639 Author: Peter Stuge <peter@stuge.se>
1640 Date: Sun Feb 27 09:37:54 2011 +0100
1642 Windows: Touch up FileDescription and ProductName resource strings
1644 commit 99b4ddcd893834a4be8b5884037b650e57ba0ba2
1645 Author: Peter Stuge <peter@stuge.se>
1646 Date: Sun Feb 27 07:20:40 2011 +0100
1648 Windows: Remove emulated direct device access via HID API
1650 Either use a library such as HIDAPI to communicate with HID class
1651 devices, or replace the kernel driver for the device with one that
1652 offers support for the complete libusb API.
1654 commit 4be84ab49c838d534d3a1b8a64ffa89774984ee7
1655 Author: Peter Stuge <peter@stuge.se>
1656 Date: Sun Feb 27 06:59:15 2011 +0100
1658 Revert libusb_strerror() until we have i18n and l10n
1660 commit 0e177f22fcc69d9aab6fb410e85f3413f478c94c
1661 Author: Pete Batard <pbatard@gmail.com>
1662 Date: Wed Jan 19 14:00:12 2011 +0000
1664 dpfp_threaded: return instead of pthread_exit() to avoid Cygwin warning
1666 commit e65878f21b0a0bb96debcb292dc03ce3c4ebb701
1667 Author: Pete Batard <pbatard@gmail.com>
1668 Date: Wed Jan 19 13:52:08 2011 +0000
1670 Windows: Rename various variables named "index" to avoid shadow warnings
1672 commit 019bf73bb2677dc7720bc8280527316adcbaa506
1673 Author: Pete Batard <pbatard@gmail.com>
1674 Date: Tue Oct 5 12:50:53 2010 +0100
1676 Windows: Fix logic in enumeration of driver name and port number
1678 commit 24494ed5db27737fc607fc35dda722521c9fe4d4
1679 Author: Pete Batard <pbatard@gmail.com>
1680 Date: Tue Oct 5 12:50:13 2010 +0100
1682 Windows: Allow claiming any interface in composite device using WinUSB
1684 Issue reported by Benjamin Dobell, see #48. The call to
1685 WinUsb_QueryInterfaceSettings() is removed since it doesn't work
1686 on any other interface but the first (a WinUSB limitation).
1688 This allows the use of WinUSB.sys in lieu of the usbccgp.sys
1689 "USB Generic Parent" default driver for composite devices.
1691 commit 9c8e6545f2fbb171f81a9451eaf2d09e1fbc13b3
1692 Author: Pete Batard <pbatard@gmail.com>
1693 Date: Tue Oct 5 12:47:15 2010 +0100
1695 Windows: Allow arbitrary bConfigurationValue in config descriptors
1697 Unlike bInterfaceNumber, bConfigurationValue does not have to be
1698 either sequential or contiguous. The bConfigurationValue is now
1699 included in debug output. Fixes #48 reported by Benjamin Dobell.
1701 commit dcd002974e511879320233fd5b0600e06bdc5620
1702 Author: Thomas Röfer <Thomas.Roefer@dfki.de>
1703 Date: Fri Nov 19 22:03:11 2010 +0100
1705 Darwin: Schedule isochronous transfers further in the future
1707 This is required to make OpenKinect transfer data at 30 fps. This
1708 change was originally made by Theo Watson, and Thomas improved on it.
1709 The commit came via OpenKinect into libusb thanks to Kyle Machulis and
1710 Hector Martin. Fixes #66.
1712 commit e83fad33dbaf9f37811da6d4cd9d3be554817944
1713 Author: Thomas Röfer <Thomas.Roefer@dfki.de>
1714 Date: Fri Nov 19 22:03:11 2010 +0100
1716 Darwin: Fix #65 memory leak in submit_iso_transfer()
1718 This was also independently fixed by Hoi-Ho Chan, but Thomas caught
1719 an additional problem with his fix.
1721 This commit came via OpenKinect into libusb thanks to Kyle Machulis
1724 commit 1a2d21fe42ad5a157651c787e2f500fc162c325b
1725 Author: Peter Stuge <peter@stuge.se>
1726 Date: Sat Feb 26 04:21:05 2011 +0100
1728 configure.ac: Refactor Windows backend settings into one occurence
1730 The Windows backend is used both with MinGW and Cygwin, and since most
1731 settings are identical they can be set in one place.
1733 commit 717f47621d18de16b87e237f3740299f43e4ca1c
1734 Author: Peter Stuge <peter@stuge.se>
1735 Date: Mon Nov 15 19:58:51 2010 +0100
1737 configure.ac: Rename AM_LDFLAGS to LTLDFLAGS and actually use them
1739 The new variable name tries to clarify that libtool is being used.
1740 Linker flags must thus always be specified with -Wl.
1742 Factor out the libtool flag -no-undefined from host specific cases.
1743 The flag is required to build a Windows DLL, but is correct also for
1744 the other supported systems.
1746 Also, start actually using LTLDFLAGS in libusb/Makefile.am, so that
1747 libtool will see the options set by configure.
1749 commit 3cbe5736469c1c52049752c137ae11387e9b079d
1750 Author: Peter Stuge <peter@stuge.se>
1751 Date: Sun Nov 21 23:38:46 2010 +0100
1753 configure.ac: Clean up PC_LIBS_PRIVATE and AM_LDFLAGS
1755 Move linker options for when linking the library itself from
1756 PC_LIBS_PRIVATE into AM_LDFLAGS. PC_LIBS_PRIVATE should only
1757 contain flags needed to link applications statically against libusb.
1760 commit ef39d1510580b9fcbacf13443fa7a4e553a9a88e
1761 Author: Peter Stuge <peter@stuge.se>
1762 Date: Fri Feb 25 02:33:29 2011 +0100
1764 configure.ac: Call AC_CONFIG_FILES() for each output file
1766 This avoids a problem with Windows line endings.
1768 commit 6e52ea067711e6e7e1b2882757fb61b11b6f8feb
1769 Author: Peter Stuge <peter@stuge.se>
1770 Date: Sun Nov 21 23:37:27 2010 +0100
1772 configure.ac: Whitespace changes and trivial reordering
1774 commit 44b85dabc3f4efe0918a598f981675aa3765cc26
1775 Author: Peter Stuge <peter@stuge.se>
1776 Date: Sun Nov 21 23:10:35 2010 +0100
1778 configure.ac: Quote AC_COMPILE_IFELSE() input
1780 commit 19d5882c398e49a453ff78494db23b7f4d13c37c
1781 Author: Peter Stuge <peter@stuge.se>
1782 Date: Sun Nov 21 23:09:02 2010 +0100
1784 configure.ac: Define booleans to 1 when set, instead of an empty string
1786 This makes the generated config.h look a lot nicer.
1788 commit c24649d796c8b5c97af79a41856422def2e593bc
1789 Author: Peter Stuge <peter@stuge.se>
1790 Date: Sat Oct 30 22:28:59 2010 +0200
1792 configure.ac: Clean up redundancy and fix LIBS on Linux
1794 LIBS, AM_LDFLAGS, OS_ conditionals and THREADS_POSIX had a bit of
1795 redundancy throughout the file, and on Linux LIBS ended up missing
1798 commit 0bd4a4dd89ff64b99ab55bfc0c5782502c10c676
1799 Author: Peter Stuge <peter@stuge.se>
1800 Date: Sat Feb 26 04:08:38 2011 +0100
1802 configure.ac: Touch up Darwin and Cygwin OS messages
1804 commit 16cf11a1213272afd616598fa3aba0941e66ec3c
1805 Author: Peter Stuge <peter@stuge.se>
1806 Date: Sat Oct 30 22:13:05 2010 +0200
1808 configure.ac: Do not use -pthread on Darwin
1810 It's not needed because Darwin has POSIX Threads in libc. Fixes #96.
1812 commit a120747d561929682cd0276ea4c13fb22fad8554
1813 Author: Peter Stuge <peter@stuge.se>
1814 Date: Sat Oct 30 22:07:37 2010 +0200
1816 configure.ac: Check for poll.h, and for nfds_t on Darwin
1818 On Linux, assume nfds_t is always available.
1820 On Darwin, fall back to unsigned int when poll() exists but there
1821 is no nfds_t, such as on Mac OS X before 10.4.
1823 On Windows (both MinGW and Cygwin), always use unsigned int instead
1824 of nfds_t, and don't check for poll.h because we use our own poll()
1827 commit 21d03d12f92c0d9bd7990861ca0ebf5009f82799
1828 Author: Nathan Hjelm <hjelmn@me.com>
1829 Date: Tue Oct 5 20:08:08 2010 -0600
1831 Darwin: Add support for control requests on endpoints other than 0
1833 Signed-off-by: Nathan Hjelm <hjelmn@me.com>
1835 commit 9bfacd760b40f1137a7835198c6028cbdaddeb18
1836 Author: Nathan Hjelm <hjelmn@me.com>
1837 Date: Tue Oct 5 23:11:50 2010 -0600
1839 Darwin: Add more error checking for libusb_open()
1841 Now also check the CreateDeviceAsyncEventSource() return code for errors.
1843 Signed-off-by: Nathan Hjelm <hjelmn@me.com>
1845 commit 478cd4a30728595fd742aa192d8149ee4216d4bc
1846 Author: Peter Stuge <peter@stuge.se>
1847 Date: Wed Nov 3 21:46:53 2010 +0100
1849 Core: libusb_get_next_timeout() must consider all flying transfers
1851 If transfer->flags indicated that a transfer had a timeout, but no
1852 timeout was actually set, then libusb_get_next_timeout() would look
1853 no further for a timeout, ignoring any transfers later in the list
1854 which had a timeout set.
1856 Since libusb has an internal 60 second timeout this bug could not
1857 cause complete lockup, but it could cause a 60 second timeout even
1858 when a transfer was submitted with a shorter timeout.
1860 commit 7802e4a3e38bcefccf96ae510ec2c6b7a959b8b5
1861 Author: Ludovic Rousseau <rousseau@debian.org>
1862 Date: Mon Nov 15 14:55:51 2010 +0100
1864 Linux: Correctly catch read() errors for sysfs config descriptors
1866 read(2) returns ssize_t (signed) and not size_t (unsigned) or int.
1868 Silence "warning: comparison of unsigned expression < 0 is always false"
1871 commit 4c7f3c8c3012bddba058afd2abdda792944fcda1
1872 Author: Peter Stuge <peter@stuge.se>
1873 Date: Sun Jan 23 04:01:25 2011 +0100
1875 Linux: Refactor cancellation into one function for all transfer types
1877 commit 011f1f297bc134188fab9960da1b1b9f0bd541aa
1878 Author: Peter Stuge <peter@stuge.se>
1879 Date: Sun Oct 17 08:19:56 2010 +0200
1881 Linux: Refactor discarding of URBs into a function and return all errors
1883 The Linux backend did not always return errors according to the
1884 libusb_cancel_transfer() documentation.
1886 commit 26246df19cb4d88954a5aa9ed5f6b859e451df85
1887 Author: Peter Stuge <peter@stuge.se>
1888 Date: Sun Jan 23 04:01:01 2011 +0100
1890 Linux: Set private number of URBs also for control transfers
1892 commit 295c9d12e25bc2dbdd8b42bd67a1f7120f0631a1
1893 Author: Alan Stern <stern@rowland.harvard.edu>
1894 Date: Sun Oct 17 06:57:06 2010 +0200
1896 Linux: cancel URBs in reverse order
1898 In a multi-URB transfer, URBs should be cancelled in reverse order of
1899 submission. This prevents races that might otherwise occur (after URB N
1900 is cancelled, data may be transferred for URB N+1 before it too can be
1901 cancelled). Fixes #8.
1903 commit 07df377a5df25e9c1811b554dc00eb8c40e5b995
1904 Author: Alan Stern <stern@rowland.harvard.edu>
1905 Date: Sun Oct 17 04:17:32 2010 +0200
1907 Linux: report correct value for urb->actual_length
1909 Transferred bytes are returned correctly by the kernel for every URB,
1910 even when an error occurs. Hence they should always be included in
1911 the transfer statistics. The same is true for isochronous packet
1914 [stuge: set itransfer->transferred so data is stored at correct offset]
1916 commit b26db49208c7a6ef0a99645142af21ce80c57c22
1917 Author: Alan Stern <stern@rowland.harvard.edu>
1918 Date: Fri Oct 22 11:13:34 2010 +0200
1920 Linux: improve reporting of URB error status codes
1922 Fix the way overflow errors are handled (they can occur in any
1923 transaction, not just the last one in a transfer) and add tests
1924 for other commonly-occuring error statuses.
1926 commit efc29733ad31f81883a7ac51a6cc6cda9ad4feb9
1927 Author: Alan Stern <stern@rowland.harvard.edu>
1928 Date: Tue Oct 5 11:08:40 2010 -0400
1930 Linux: improve detection of BULK_CONTINUATION kernel support
1932 Clarify the comments describing the BULK_CONTINATION flag, and improve
1933 the subroutine that checks whether the flag is supported by making it
1934 "future-proof", i.e., by working properly with kernel version numbers
1937 commit bd267cf92269e0c4cd8eef795b91eefb2c63452e
1938 Author: Peter Stuge <peter@stuge.se>
1939 Date: Mon Nov 22 05:14:00 2010 +0100
1941 README: Update webpage URL and add Peter's email address
1943 commit 845b715675703c5f1bc907f6302ece416c9f112c
1944 Author: Peter Stuge <peter@stuge.se>
1945 Date: Wed Aug 4 07:30:22 2010 +0200
1947 libusb is not just for Linux!
1949 commit 67d9ef7b6877e17f2deec81cd41dc0948f6bed4b
1950 Author: Nathan Hjelm <hjelmn@me.com>
1951 Date: Tue Oct 5 19:48:39 2010 -0600
1953 Remove USBI_OS_HANDLES_TIMEOUT and fix int/isoc timeouts on Darwin
1955 Backends set USBI_TRANSFER_OS_HANDLES_TIMEOUT for transfers instead.
1956 Darwin only handles timeouts for bulk and control transfers, so the
1957 backend now sets that flag accordingly, making libusb core handle
1958 timeouts for interrupt and isochronous transfers. Fixes #31.
1960 Signed-off-by: Nathan Hjelm <hjelmn@me.com>
1962 [stuge: rework libusb_get_next_timeout() and enum usbi_transfer_flags]
1963 [stuge: fix typo; set USBI_TRANSFER_TIMED_OUT flag correctly]
1965 commit 370922dfbe0964c0b8a0c1974bb1d7b85ac10607
1966 Author: Peter Stuge <peter@stuge.se>
1967 Date: Sun Nov 14 03:14:00 2010 +0100
1969 Silence "warning: suggest braces around empty body in an 'if' statement"
1971 commit 0f85015bdf87f3bca0d86ed87c9b976176640922
1972 Author: Konrad Rzepecki <hannibal@astral.lodz.pl>
1973 Date: Sat Nov 13 14:09:22 2010 +0100
1975 Disallow libusb_get_string_descriptor_ascii() with index 0
1977 String descriptor 0 is the list of supported language IDs in the device,
1978 which can't have an ASCII representation. Calling the function with
1979 index 0 is thus not really useful. Fixes #43.
1981 commit 3b41c31d3fa2c1f93649cc5fc0f59acbf3efd5a1
1982 Author: Ludovic Rousseau <rousseau@debian.org>
1983 Date: Fri Mar 19 15:35:14 2010 +0100
1985 Use const for the usbi_parse_descriptor() format string
1987 Constant strings should be "const char *" instead of "char *". Silence
1988 "discards qualifiers from pointer target type" warnings. Fixes #61.
1990 commit 7f2e9f0776386997d2b4c4c47598ab88e3caeb7a
1991 Author: Hector Martin <hector@marcansoft.com>
1992 Date: Wed Aug 25 03:21:57 2010 +0200
1994 Linux: Fix usbfs/sysfs config descriptor handling on big-endian
1996 usbfs endian-swaps, sysfs doesn't, not the other way around. Fixes #51;
1997 descriptor endian parsing is backwards using sysfs on big-endian hosts.
2000 Signed-off-by: Hector Martin <hector@marcansoft.com>
2002 [stuge: checked against Documentation/usb/proc_usb_info.txt]
2004 commit 2aa8984e59cf0f492e7fbebb8924f54f6160cc5e
2005 Author: Ludovic Rousseau <rousseau@debian.org>
2006 Date: Tue Aug 17 20:53:04 2010 +0200
2008 examples: #include <libusb.h> without directory name
2010 The libusb-1.0 directory name belongs in the compile command, e.g.
2011 via pkg-config --cflags. Fixes #58.
2013 commit 7da756e09fd97efad2b35b5cee0e2b2550aac2cb
2014 Author: Ludovic Rousseau <rousseau@debian.org>
2015 Date: Thu Jun 24 10:14:24 2010 +0200
2017 Add missing argument to libusb_wait_for_event() documentation, fix #55
2019 In the libusb_wait_for_event() sample code in the section "Letting other
2020 threads do the work for you" the call to libusb_wait_for_event() was not
2021 updated by commit 1df713d622ab4f0b03aad72d903ac7beb8fb3b90, which added
2022 the libusb_context *ctx parameter to the function, so the sample code
2023 was broken until now.
2025 commit 678c242705b29f9cb6cc421b5625a6485ae58047
2026 Author: Peter Stuge <peter@stuge.se>
2027 Date: Sat Oct 16 14:22:30 2010 +0200
2029 examples: Silence warnings about return value of fwrite()
2031 commit 9cd9059bf9baac69a9d6c909f4c1e000592fa260
2032 Author: Peter Stuge <peter@stuge.se>
2033 Date: Sat Oct 16 11:47:14 2010 +0200
2035 Linux: Don't free() URBs prematurely on hardware error; fix #54
2037 When an URB in a multi-URB transfer had an error, handle_bulk_completion()
2038 would free all URBs and segfault once the next one completed, instead of
2039 cancelling the remaining URBs and cleaning up as usual.
2041 This is basically the patch from the ticket, plus some restructuring
2042 for increased readability of the function.
2044 Many thanks to Brian Shirley and National Instruments for finding and
2047 commit 116d34f608b02f4e14668450b158acf5db726f95
2048 Author: Pete Batard <pbatard@gmail.com>
2049 Date: Tue Oct 5 12:43:41 2010 +0100
2051 Windows: add support for filter drivers
2053 * precedence is driver > upper filter > lower filter, and defined
2054 in struct driver_lookup lookup from set_device_paths()
2056 commit eb833fcd837e3aac2e497438b14c8612254c28ce
2057 Author: Pete Batard <pbatard@gmail.com>
2058 Date: Tue Oct 5 12:36:49 2010 +0100
2060 Windows: fix errcode shadow warning, and string and DLL_DECLARE fixes
2062 * use FormatMessageA, and overall improvement of windows_error_str()
2063 * use safe_strlen() always
2064 * better typing/init default
2065 * more explicit debug message for devices with no driver
2067 commit 1feadb5f78e1d2cdfd2aaf0011e041c2848bae37
2068 Author: Pete Batard <pbatard@gmail.com>
2069 Date: Tue Oct 5 12:29:32 2010 +0100
2071 Windows: use IOCTLs for HID input, output and feature reports
2073 * fixes feature reports not providing actual read size
2074 (reported by Axel Rohde http://marc.info/?m=127033070021994)
2075 * removes the USE_HIDD_FOR_REPORTS macro
2076 * IOCTL usage inspired from HIDAPI by Alan Ott
2078 commit d7c25451f4075d2a12eb5cbc91ff3de8c0292da0
2079 Author: Daniel Drake <dan@reactivated.net>
2080 Date: Mon Oct 4 20:03:58 2010 +0100
2082 Fix libtool version setting
2084 This was being set too early to take effect.
2086 commit a6fb54e2eef4e393bd06829bc3bdc746ea9d41a6
2087 Author: Mike Frysinger <vapier@gentoo.org>
2088 Date: Mon Oct 4 20:01:07 2010 +0100
2090 Populate the pkg-config Libs.private field
2092 commit a9b4065f048e750b7317f6434406fd528b69bf33
2093 Author: Mike Frysinger <vapier@gentoo.org>
2094 Date: Fri Aug 20 00:09:46 2010 -0400
2096 Linux: fix build failure when CLOCK_MONOTONIC does not exist
2098 This is not available on all platforms.
2100 commit d192c5bd32736fde0080d756de6221287af34891
2101 Author: Daniel Drake <dan@reactivated.net>
2102 Date: Mon Oct 4 18:45:18 2010 +0100
2104 Update libtool version info
2106 With input from various people on the mailing list, update the libtool
2107 versioning info and start to update this on every release.
2109 The next libusb release will not need a change here. All following ones
2112 commit e65f69a7ef3de357c867ddaac9598efe407078b6
2113 Author: Pete Batard <pbatard@gmail.com>
2114 Date: Thu Aug 5 16:40:09 2010 +0100
2116 Updated documentation for unsupported Darwin/Windows calls
2118 commit c8bab4e19413d70c515a2eaf747a75b7f66fdc0b
2119 Author: Pete Batard <pbatard@gmail.com>
2120 Date: Sun Sep 19 14:24:25 2010 +0100
2122 Windows: Make libusb_pollfd() return error
2124 Windows does not have numerical file descriptors but does have the
2125 concept of event sources. Exposing these event sources will require
2126 some careful thought and design, to be completed later.
2128 commit bd623f70d99fad8b975f8d4ee62ea4b1abf1272f
2129 Author: Daniel Drake <dan@reactivated.net>
2130 Date: Sun Sep 19 14:21:01 2010 +0100
2132 MinGW: Use --add-stdcall-alias linker option
2134 Naming of symbols inside the library is inconsistent on Windows:
2135 http://wyw.dcweb.cn/stdcall.htm
2137 Use this linker option to add aliases which add compatibility with
2138 the "MSVC DLL" platform when a DEF file is used. It also better matches
2139 the appearance of the Windows API itself.
2141 commit 29f9f9e3af3340df6a955881a93caf9d2a6d08d6
2142 Author: Pete Batard <pbatard@gmail.com>
2143 Date: Fri Aug 13 11:59:49 2010 +0100
2145 Introduced calling convention (for Windows)
2147 Under Windows, a variety of compilers and configurations are available,
2148 meaning that the manner of parameter passing (e.g. registers vs stack)
2151 Match the Windows API calling convention and document this appropriately.
2152 This calling convention will be used regardless of the configuration of
2153 the user's development platform.
2155 The only user-level complication is that all functions used as libusb
2156 callbacks must use the same calling convention as libusb. The
2157 LIBUSB_CALL macro is provided to make this easy.
2159 Signed-off-by: Michael Plante <michael.plante@gmail.com>
2160 Signed-off-by: Pete Batard <pbatard@gmail.com>
2161 [dsd: slight change of strategy, add documentation]
2163 commit be523f1fe7c136c3fca06ae9c96aff44a22482ba
2164 Author: Pete Batard <pbatard@gmail.com>
2165 Date: Fri Aug 6 21:56:04 2010 -0600
2167 Windows: fix string copy in sanitize_path
2169 The NULL byte was not being copied.
2171 commit 7fd36af9f749ce8bf96f89cefd519a9f46825573
2172 Author: Pete Batard <pbatard@gmail.com>
2173 Date: Fri Jul 30 00:30:31 2010 +0100
2175 Windows: DuplicateHandle failure is not a warning
2177 This is because WinUSB handles are not duplicable, and generate a
2178 lot of these warnings.
2180 commit 33cd2f68b4d500e852e354e5acae3d9a96a8ec26
2181 Author: Pete Batard <pbatard@gmail.com>
2182 Date: Fri Jul 30 00:23:05 2010 +0100
2184 removed trailing whitespaces
2186 commit e47849afe9381c864683220eac9ce7bcf43f7883
2187 Author: Pete Batard <pbatard@gmail.com>
2188 Date: Wed Aug 4 18:18:32 2010 -0600
2190 Windows: Mark interfaces as unused when releasing
2192 Without this important cleanup, the autoclaim code could attempt to
2193 reuse an unclaimed interface as if it were already claimed.
2195 commit b30ca141a0bdce26e7cfdd91d346e49051532ee7
2196 Author: Pete Batard <pbatard@gmail.com>
2197 Date: Wed Aug 4 18:15:59 2010 -0600
2199 Windows: fix device path string duplication
2201 commit e8d7a89503d5655c4b04da718c6971e25ff48162
2202 Author: Nathan Hjelm <hjelmn@me.com>
2203 Date: Mon Aug 2 21:27:38 2010 -0600
2205 Darwin: Cache device configuration value
2207 Cache device configuration value to bring the Darwin backend more in
2208 line with the libusb spec. To handle buggy devices GetConfiguration is
2209 not called unless the device has more than one configuration.
2211 commit fe5d1d2050f715ffa8cbd35a378ac969a1a01584
2212 Author: Xiaofan Chen <xiaofanc@gmail.com>
2213 Date: Fri Jul 30 15:56:43 2010 +0100
2215 Fix inconsistencies between prototypes and definitions
2217 This removes some warnings on cygwin regarding a name clash against
2218 the standard library "index" function.
2220 commit 3b538c43e11b805a37d93b6283d270d5d5f0d1f5
2221 Author: Hoi-Ho Chan <hoiho.chan@gmail.com>
2222 Date: Tue Jul 27 21:04:17 2010 -0600
2224 Darwin: handle kIOReturnUnderrun
2226 This return code indicates that the device returned a data packet
2227 less than the max packet size. In libusb backend terms, this is
2228 a successful transfer.
2230 commit 9a4249f8a104b98a15a7e3ba7ecae9a385ed9027
2231 Author: Pete Batard <pbatard@gmail.com>
2232 Date: Sat Jul 10 17:51:13 2010 -0600
2236 Via Cygwin/MinGW, libusb now has windows support.
2237 Thanks to contributors: Michael Plante, Orin Eman, Peter Stuge,
2238 Stephan Meyer, Xiaofan Chen.
2240 commit d38dd5e3d2a872f7064eea084ddea8c33811dd7d
2241 Author: Hoi-Ho Chan <hoiho.chan@gmail.com>
2242 Date: Sat Jul 24 13:48:00 2010 -0600
2244 Darwin: write iso transfer details into correct packets
2246 commit fbff5b4b665ff2ebc948162617cf7f4f5d2fabeb
2247 Author: Nathan Hjelm <hjelmn@me.com>
2248 Date: Sat Jul 24 13:42:57 2010 -0600
2250 Darwin: fix interface object leak
2252 commit dffc09835d6860ae9b6a62c39af5be346188cb05
2253 Author: Nathan Hjelm <hjelmn@me.com>
2254 Date: Sat Jul 24 13:41:03 2010 -0600
2256 Darwin: clean up some debug messages
2258 commit 946d1abbca397b8e88a1b0b4df176dcc5e3b7348
2259 Author: Nathan Hjelm <hjelmn@me.com>
2260 Date: Sat Jul 24 13:39:16 2010 -0600
2262 Darwin: Don't open devices during scan unless we have to unsuspend
2264 commit 3b6d9ac82e2599cad7817d21e909a42275ddc4c4
2265 Author: Peter Stuge <peter@stuge.se>
2266 Date: Fri Jun 25 08:08:13 2010 +0200
2268 Linux: Handle early complete of multi-URB transfer
2270 commit a516fcb99376686d2d0028e5ac69950c57a55b9e
2271 Author: Daniel Drake <dan@reactivated.net>
2272 Date: Wed Jul 21 17:49:21 2010 -0600
2274 Only compile dpfp examples when sigaction is available
2276 It's not available on MinGW.
2278 commit a1d831ef82e1700f4ebb9df6ca5cef74be757c08
2279 Author: Aurelien Jarno <aurelien@aurel32.net>
2280 Date: Tue May 25 09:32:19 2010 +0200
2282 configure.ac: fix bashisms
2284 commit 7211aba6e290f9805e911959b2a43a4b3ec56d4f
2285 Author: Daniel Drake <dan@reactivated.net>
2286 Date: Fri Jun 25 12:01:58 2010 -0500
2288 Only include sys/time.h on appropriate platforms
2290 This header doesn't exist on windows.
2291 For libusb, determine at configure-time if the header is available.
2292 For libusb.h, use gcc predefined macros to only include the header
2293 on platforms that need it.
2295 commit b33c3cb9651459de1f3d549677cbac67a017a295
2296 Author: Daniel Drake <dan@reactivated.net>
2297 Date: Tue Jun 22 18:45:38 2010 -0500
2299 Linux: Fix log message commit
2301 Forgot to test this.
2303 commit 139648d5723190734a51c1c66ddf41c3e87adf10
2304 Author: Ludovic Rousseau <rousseau@debian.org>
2305 Date: Mon Jun 21 10:19:29 2010 +0200
2307 Only initialize the default context when it is requested
2309 commit 22d61cd0891d8304dfc1a70579cf154fd8e6644a
2310 Author: Daniel Drake <dan@reactivated.net>
2311 Date: Tue Jun 22 18:20:23 2010 -0500
2313 Linux: fix log message that was being unconditionally sent to stderr (#44)
2315 This message was put in place to aid libusb-compat-0.1 users, who
2316 may run into the change that libusb-0.1 allowed you to usb_open() a
2317 device that you don't have write access to, but libusb-1.0 does not.
2319 As explained on the ticket this change is dangerous so it should go through
2320 the usual logging mechanism. I also added a similar regular log message
2321 to libusb-compat-0.1.
2323 commit 82065a2c5e9d5ad0b2bb28792d12462f2fa150f0
2324 Author: Ludovic Rousseau <rousseau@debian.org>
2325 Date: Thu Jun 10 18:41:27 2010 +0200
2327 Fix 2 documentation typos
2329 commit 351c062aacdde0b0fa7972ee9b7c6b355dcee110
2330 Author: Ludovic Rousseau <rousseau@debian.org>
2331 Date: Wed Apr 21 16:49:33 2010 +0200
2333 add AM_MAINTAINER_MODE
2335 commit a7c056c14867e1cc3ef83682f2706cdbdca7c4ce
2336 Author: Pete Batard <pbatard@gmail.com>
2337 Date: Thu Jun 10 17:00:25 2010 -0500
2339 Factorize event handler interruption code
2341 This will be additionally used by the windows backend which modifies
2342 the fd set frequently.
2344 commit fc0af8e3f76925ad09e0fbc6d233d500753ee375
2345 Author: Pete Batard <pbatard@gmail.com>
2346 Date: Sat Jun 5 12:14:19 2010 -0500
2348 Abstract low-level event handler operations
2350 The Windows backend uses something other than UNIX file descriptors
2351 for event handling. Abstract out the operations to allow for this.
2353 commit 64383fb5b55a74b706967836e81cef61e57df86a
2354 Author: Pete Batard <pbatard@gmail.com>
2355 Date: Tue Jun 1 11:38:14 2010 +0100
2357 fixes the possibility of using a broken timercmp on Windows
2359 Microsoft's implementation of timercmp does not work for >= or <=.
2361 commit f6d60a9a8fcf06629812da886e3a5de592624b61
2362 Author: Pete Batard <pbatard@gmail.com>
2363 Date: Mon Feb 1 19:26:13 2010 +0000
2365 Switched to standard int pointer type for windows 64-bit compat
2367 On 64-bit Windows a pointer is 64 bit, but long is still 32 bit for
2368 backwards compatibility.
2369 uintptr_t nicely hides this difference and also works on Linux.
2371 [dsd: add appropriate include]
2373 commit a2fa855ce3795e31905b255267874af0603168bc
2374 Author: Michael Plante <michael.plante@gmail.com>
2375 Date: Tue Mar 9 16:39:46 2010 -0600
2377 added doxygen, dpfp, msvc, and xusb stuff to gitignore
2379 commit 845d96a9c30b6f053208ed6b5b5d4ad1655e2026
2380 Author: Daniel Drake <dan@reactivated.net>
2381 Date: Mon May 31 19:56:34 2010 -0500
2383 Add reference counting to default context
2385 Michael Plante pointed out that if 2 users call libusb_init(NULL) within
2386 a process, we end up creating 2 default contexts, one of which is lost.
2388 Add reference counting so that the default context is reused and
2389 destroyed only after the last user.
2391 commit 0e77f017fd3d1bdec4aced0429a380d3e434081b
2392 Author: Daniel Drake <dan@reactivated.net>
2393 Date: Thu May 27 19:24:49 2010 -0400
2395 Fix libusb_init() error handling
2397 9996ccaed7 introduced a problem in that mutexes could be destroyed
2398 in the error handling codepath before they had been created.
2399 Pointed out by Michael Plante.
2401 commit b03de9d07bf9a6b0c4e6cba9ec06b5271a22e621
2402 Author: Michael Plante <michael.plante@gmail.com>
2403 Date: Mon May 24 11:57:10 2010 -0300
2405 Modify log functions to compile with MSVC6
2407 MSVC6 does not support variadics.
2409 commit aa77b02da65afb0332b9e6853dae9e4dcda00bfa
2410 Author: Pete Batard <pbatard@gmail.com>
2411 Date: Sun Feb 14 19:46:05 2010 -0600
2413 Add type parameter to the list_for_each_entry() and _safe() macros
2415 typeof() is a GCC extension, not supported by target compilers such
2418 commit b92b16f437079208c8dfb1dc2380893846a07ee9
2419 Author: Daniel Drake <dan@reactivated.net>
2420 Date: Wed May 19 17:37:29 2010 -0300
2422 Fix a usbi_pollfd leak in error path
2424 I accidently missed this when applying Pete Batard's earlier patch.
2425 Pointed out by Michael Plante.
2427 commit 8aceb5c99056aaed6e6f2db1b303817e37e9fe8c
2428 Author: Pete Batard <pbatard@gmail.com>
2429 Date: Sun Feb 14 19:47:30 2010 -0600
2431 Rename all interface parameters to usb_interface or interface_number
2433 Remove the use of the ambiguous "interface" where possible without
2435 'interface' is a preprocessor #define already associated with COM on Windows,
2436 that can cause a conflict as soon as windows.h is included.
2438 There are a few more places where interface is still used, which should
2439 be changed for libusb-2.0
2441 commit d7031ee8e1c449af23b2674fe18d9fc652c9bd80
2442 Author: Pete Batard <pbatard@gmail.com>
2443 Date: Mon May 17 19:30:27 2010 -0300
2445 Stricter types and casts
2447 commit 9996ccaed740f8fcec4287a478ccaaa573f9865a
2448 Author: Pete Batard <pbatard@gmail.com>
2449 Date: Mon May 17 19:16:54 2010 -0300
2451 Fixes to error handling and exit codepaths
2453 Various locks and things were being leaked/left open when handling
2454 errors and during deinitialization.
2458 commit 371ca4e440894bd1c845971418aa24c9d3919e63
2459 Author: Nathan Hjelm <hjelmn@me.com>
2460 Date: Thu May 13 19:38:05 2010 -0300
2462 Darwin: Define IO_OBJECT_NULL if not already defined
2464 Needed on Mac OSX 10.3
2466 commit 7ba92cff94bbba19284749c614c26141d3023f37
2467 Author: Hans Ulrich Niedermann <hun@n-dimensional.de>
2468 Date: Fri Jan 29 12:08:52 2010 +0100
2470 libusb_cpu_to_le16: macro->static inline function
2472 The libusb_cpu_to_le16 macro was a ({ ... }) expression, which
2473 generates the following compiler warning every time it is used,
2474 (and that is several times for every #include <libusb.h>):
2476 libusb.h:880: warning: ISO C forbids braced-groups within expressions
2478 With this patch, #include <libusb.h> stops generating compiler
2479 warnings on gcc 4.4.
2481 As libusb.h heavily relies on the use of static inline functions
2482 those can be relied on to work properly, and there should not be
2483 any significant difference in the code the compiler generates.
2485 commit 3473ac6c6fab32202d02d87679ebdb24e7d2df98
2486 Author: Francesco <francesco.montorsi@gmail.com>
2487 Date: Tue Mar 9 16:35:35 2010 -0600
2489 Add libusb_strerror() to get short error message in English from enum
2493 commit 5b69f3b7b0610d5f518954792ddc93332a41d292
2494 Author: Michael Plante <michael.plante@gmail.com>
2495 Date: Thu Feb 25 19:39:22 2010 -0600
2497 Fix libusb_device::lock memory leak
2499 commit a636df4ca1b87ed3094a87828c022092a80ba613
2500 Author: Pete Batard <pbatard@gmail.com>
2501 Date: Wed May 12 21:46:31 2010 -0300
2505 commit 23b5db8b3e353176dfa0635bfb6f0e2658f6e3de
2506 Author: Peter Stuge <peter@stuge.se>
2507 Date: Tue May 11 00:51:43 2010 +0100
2509 Add internal abstraction for POSIX Threads
2511 This prepares for a Windows backend without dependency on pthreads-w32.
2513 pthread_* is renamed to usbi_* and PTHREAD_* to USBI_*.
2515 A usbi_mutex_static_t and usbi_mutex_static_lock() and _unlock() are
2516 introduced for statically initialized mutexes, since they may be
2517 implemented using other types when pthreads mutexes aren't used.
2519 Move -pthread from libusb/Makefile.am to host-specific THREAD_CFLAGS in
2520 configure.ac. This will enable optional use of -lpthread for cygwin.
2522 [dsd: minor tweaks, and roll in a change based on patches from Pete
2523 Batard to only build dpfp_threaded example when we're using pthreads]
2525 commit a04cbb0095a78aeed3f65aaf06c46069d0ac184a
2526 Author: Daniel Drake <dan@reactivated.net>
2527 Date: Tue May 4 19:01:59 2010 -0300
2531 commit 53b47299531974fa8901b74a163f3c8ebae4eec9
2532 Author: Daniel Drake <dan@reactivated.net>
2533 Date: Thu Apr 22 18:12:51 2010 -0300
2535 Linux: don't set SHORT_NOT_OK on bulk out URBs (#20
2537 Setting this flag is illegal, and the behaviour we're looking for
2538 is already in place for host-to-device transfers without this flag.
2540 commit 1519828b7ee1fce46d4c51fc097d52e01b8e0bb4
2541 Author: Daniel Drake <dan@reactivated.net>
2542 Date: Tue Apr 20 20:14:09 2010 -0300
2544 Linux: Handle failure to read active configuration during enumeration
2546 commit 45168627cc15aee3875192f34286110dbbd27095
2547 Author: Nathan Hjelm <hjelmn@me.com>
2548 Date: Tue Apr 20 19:43:06 2010 -0300
2550 Darwin: don't reuse cached descriptors during enumeration
2552 Descriptor reuse was causing scans to return invalid information when
2553 the device at a location has changed.
2555 commit 3af329db6dcbfb96d24867c757f9e125e5b7b0e5
2556 Author: Daniel Drake <dan@reactivated.net>
2557 Date: Mon Apr 19 19:36:55 2010 -0300
2561 commit ceb8cacd6d2a4189de0db2ee46d45217511c69be
2562 Author: Daniel Drake <dan@reactivated.net>
2563 Date: Mon Apr 19 19:31:12 2010 -0300
2565 Document that fd set contents are an internal implementation detail (#36)
2567 commit 9bea500b5747bdeba7c8251d45608558e71a1db5
2568 Author: Martin Koegler <mkoegler@auto.tuwien.ac.at>
2569 Date: Mon Apr 19 19:22:46 2010 -0300
2571 Linux: correct config descriptor endianness fix
2573 The seek_to_next_config codepath deals with both sysfs and usbfs;
2574 make sure we only convert values in the usbfs path.
2576 commit 02df59a309e813c50b8230de99e69fb4e1814279
2577 Author: Martin Koegler <mkoegler@auto.tuwien.ac.at>
2578 Date: Mon Apr 5 17:11:32 2010 +0200
2580 Linux: Fix endianness handling of usbfs config descriptors (#27)
2582 driver/usb/core/devio.c function usbdev_read translate the follwing files
2585 le16_to_cpus(&temp_desc.bcdUSB);
2586 le16_to_cpus(&temp_desc.idVendor);
2587 le16_to_cpus(&temp_desc.idProduct);
2588 le16_to_cpus(&temp_desc.bcdDevice);
2590 All other data is passed without any change.
2592 libusb/os/linux_usbfs.c calls usbi_parse_descriptor with host_endian=1
2593 for config descriptors. According to the kernel code, they must be
2594 processed with host_endian=0, as they are not translated by the kernel.
2596 Signed-off-by: Martin Koegler <mkoegler@auto.tuwien.ac.at>
2598 commit ec303b01a5d4e51c000a283853af65059fa62285
2599 Author: Nathan Hjelm <hjelmn@mac.com>
2600 Date: Tue Mar 23 16:22:08 2010 -0600
2602 Darwin: fix enumeration of devices with non-consecutive addresses (#23)
2604 commit cd809e2f7cee3874b7ae16b2c482a8b63a90e4a5
2605 Author: Bastien Nocera <hadess@hadess.net>
2606 Date: Mon Mar 8 10:50:51 2010 +0000
2608 Add more interface classes
2612 commit 1ce4aa67d849f5cad8a21072dc1c7b42158ce817
2613 Author: Nathan Hjelm <hjelmn@mac.com>
2614 Date: Thu Mar 18 11:32:32 2010 -0600
2616 Darwin: Cleanup async callback code, catch request timeouts
2618 commit 68af9f8d731f700267335941a8214d34ab518cc8
2619 Author: Nathan Hjelm <hjelmn@mac.com>
2620 Date: Thu Mar 18 11:31:46 2010 -0600
2622 Darwin: fix endianness of control setup packet
2624 IOUSBLib expects the control request to be in host byte order.
2625 Swap the request into host byte order.
2627 commit cfce4d127184f4e4f334976151a0f80594bb5e22
2628 Author: Nathan Hjelm <hjelmn@mac.com>
2629 Date: Mon Feb 15 14:10:08 2010 -0600
2631 Darwin: fix memory leak in process_device
2633 Credit to Mike Colagrosso for finding this bug.
2635 commit 161893cfbefefe315f657677705abe090fc526f2
2636 Author: Nathan Hjelm <hjelmn@mac.com>
2637 Date: Mon Feb 15 14:09:19 2010 -0600
2639 Darwin: use logging functions
2641 Use usbi_warn, usbi_err, and usbi_info instead of _usbi_log.
2643 commit 2a72f38548208044dc3aa62681419d006c35732d
2644 Author: Nathan Hjelm <hjelmn@mac.com>
2645 Date: Mon Feb 15 14:07:44 2010 -0600
2647 Darwin: support multiple calls to libusb_init
2649 Credit to Orin Eman for finding this bug.
2651 commit a4186794d87124503db2f5f51f51ce90bb95daa7
2652 Author: Daniel Drake <dan@reactivated.net>
2653 Date: Sun Nov 22 17:20:53 2009 +0000
2657 commit 8392ff22136fccaf1e15d186157609b8dd127bc5
2658 Author: Ludovic Rousseau <rousseau@debian.org>
2659 Date: Sun Oct 25 10:06:41 2009 +0100
2661 lsusb example: make print_devs() static
2663 lsusb.c:26: warning: no previous prototype for ‘print_devs’
2665 commit 9cc6bfaa15239bb6db1c1570b9beb6df2f848951
2666 Author: Ludovic Rousseau <rousseau@debian.org>
2667 Date: Sun Oct 25 10:05:10 2009 +0100
2669 Darwin: fix warning in darwin_error_str()
2671 os/darwin_usb.c:63: warning: return discards qualifiers from pointer
2674 commit 4c706d2fb6b2c43b10d72ac5dff51cac4d939f1a
2675 Author: Nathan Hjelm <hjelmn@mac.com>
2676 Date: Sat Nov 21 17:06:43 2009 +0000
2678 Darwin: allow devices to be opened multiple times
2680 Allows libusb applications to access multiple interfaces of the same
2681 device in the same application.
2683 Also fixes a set alt interface bug.
2685 commit 0232fc559cdacb9561f982dd6d28feb4435b3e4e
2686 Author: Daniel Drake <dan@reactivated.net>
2687 Date: Sat Nov 21 17:01:32 2009 +0000
2689 Increase libusb_handle_events() timeout to 60 seconds
2691 The internal timing seems to be working, this will be a better test of
2692 it before we make this timeout unlimited.
2694 commit 0bd7ef5d8697973a026c36c15d6276177b4ec4ea
2695 Author: Daniel Drake <dan@reactivated.net>
2696 Date: Sat Nov 21 16:57:25 2009 +0000
2698 Refine timerfd header check (#18)
2700 Require glibc-2.9 for the working timerfd support.
2702 commit 90d8fcab9018b8e6887a7e0592d1e5f692117234
2703 Author: Daniel Drake <dsd@gentoo.org>
2704 Date: Sun Nov 15 12:17:13 2009 +0000
2708 commit 11d591058e3f105b0e90c23bbf58b18de691e690
2709 Author: Daniel Drake <dsd@gentoo.org>
2710 Date: Sat Nov 7 10:43:59 2009 +0000
2712 Update documentation about early completion caveats
2714 commit f796c9528a71aa55326b6f0c9c7c5ec073d2bf92
2715 Author: Ludovic Rousseau <rousseau@debian.org>
2716 Date: Sun Oct 25 09:59:51 2009 +0100
2718 Add libusb_get_max_iso_packet_size() to libusb.h
2720 core.c:777: warning: no previous prototype for 'libusb_get_max_iso_packet_size'
2722 commit 4783008b7e711de9cb31631e60dda995f44068de
2723 Author: Daniel Drake <dsd@gentoo.org>
2724 Date: Wed Oct 28 20:33:49 2009 +0545
2726 Use timerfd for timeout handling
2728 Use a new file descriptor from the timerfd system calls to handle
2729 timeouts. On supported systems, this means that there is less hassle
2730 figuring out when the poll() timeout should be, since
2731 libusb_get_next_timeout() will always return 0 and the timeout events will
2732 be triggered as regular activity on the file descriptor set.
2734 Add API function libusb_pollfds_handle_timeouts() to detect whether
2735 you're on a platform with the timing headache, and flesh out the
2736 surrounding documentation.
2738 commit 9b120c2b3735566533c179aa8ca758fe45899a38
2739 Author: Daniel Drake <dsd@gentoo.org>
2740 Date: Sat Nov 7 10:03:07 2009 +0000
2742 Use AM_SILENT_RULES for building
2744 commit ef6ea6c3ae38e4524f10e16e8cb88177d39c4826
2745 Author: Daniel Drake <dsd@gentoo.org>
2746 Date: Fri Nov 6 21:54:08 2009 +0000
2750 commit 217f57617e0cff0d1bd6d726b243f04c6b179773
2751 Author: David Moore <dcm@acm.org>
2752 Date: Fri Nov 6 21:37:25 2009 +0000
2754 Linux: Add support for the new URB_BULK_CONTINUATION flag
2756 Add support for the new USBDEVFS_URB_BULK_CONTINUATION flag to libusb.
2758 This flag, which is expected to be available in usbfs starting with
2759 kernel 2.6.32, allows the kernel to cancel multiple URBs upon receipt
2760 of a short packet. This capability allows libusb to preserve data
2761 integrity of large bulk transfers that are split into multiple URBs.
2762 Without this support, these URBs must be canceled in userspace upon
2763 receipt of a short packet, a race condition against future transfers
2764 which might partially fill these canceled URBs.
2766 This patch automatically detects whether a supported kernel is present
2767 and enables the use of the flag when possible.
2769 [dsd: tweaks to supported kernel detection, and some inline
2770 documentation of this mechanism]
2772 commit 69830057547396f893f0d7b3125a05d016313b10
2773 Author: Daniel Drake <dsd@gentoo.org>
2774 Date: Wed Oct 28 15:13:22 2009 +0545
2778 At least on Linux, there were some possible races that could occur if
2779 a transfer is cancelled from one thread while another thread is handling
2780 an event for that transfer, or for if a transfer completes while it is
2781 still being submitted from another thread, etc.
2783 On the global level, transfers could be submitted and cancelled at the
2786 Fix those issues with transfer-level locks.
2788 commit 98f1b30d24359cb3185051b8df9ebb663cc10369
2789 Author: Daniel Drake <dsd@gentoo.org>
2790 Date: Mon Sep 14 08:01:24 2009 +0100
2792 Clarify that timeout 0 means unlimited timeout
2794 commit 858684f0dd25921e09565034a88709dbf6f6c61b
2795 Author: Daniel Drake <dsd@gentoo.org>
2796 Date: Fri Sep 11 22:09:12 2009 +0100
2798 Linux: more flexibility with monotonic clock
2800 Some users have reported that CLOCK_MONOTONIC does not work on their
2801 systems - I suspect it is available on x86 but perhaps not some
2802 of the more uncommon architectures. We should fall back on
2803 CLOCK_REALTIME in these cases.
2805 Also, CLOCK_MONOTONIC_RAW seems even more monotonic, so we should use
2806 that if it is available.
2808 We now test different clock IDs during initialization to find the
2809 best one that works.
2811 commit fe0d8dce1ed704915d501e7da700440c78144211
2812 Author: Nathan Hjelm <hjelmn@mac.com>
2813 Date: Fri Sep 11 18:00:29 2009 +0100
2815 Darwin: handle overflows
2817 commit f46716f42040986203fa6e873bfdabe1be2900ec
2818 Author: Daniel Drake <dsd@gentoo.org>
2819 Date: Thu Aug 27 21:14:54 2009 +0545
2823 commit 5b489b8b2a5aba7b8b804e5af6d2628735548238
2824 Author: Toby Peterson <toby@macports.org>
2825 Date: Sun Aug 23 10:04:59 2009 +0545
2827 Darwin: 64-bit type fixes
2829 commit ad8ae04d0b52009af0b1180e005f7554d2bbb26c
2830 Author: Nathan Hjelm <hjelmn@mac.com>
2831 Date: Sun Aug 23 10:02:55 2009 +0545
2833 Darwin: fix crash when reading descriptors after close
2835 Fix a crash which occurs if the user does the following sequence on a
2836 device: open, close, get_configuration_descriptor.
2838 commit 45ae2aecf5512dcff059b2a416534e81c6a00c88
2839 Author: Daniel Drake <dsd@gentoo.org>
2840 Date: Sat Aug 1 13:55:15 2009 +0545
2842 move bug info to bug tracker
2844 Protection needed: http://www.libusb.org/ticket/4
2845 Losing data: fixed in previous commit
2847 commit 126129e174062c2a27423817a459e5113f777789
2848 Author: Daniel Drake <dsd@gentoo.org>
2849 Date: Thu Jul 9 22:09:04 2009 +0100
2851 Linux: try harder not to lose any data
2853 We would previously lose any data that was present on a cancelled URB.
2854 Work harder to make sure this doesn't happen.
2856 commit 0334ee642b47dfe1ca9db64b22e7702ea14b3f09
2857 Author: Daniel Drake <dsd@gentoo.org>
2858 Date: Sun Jun 28 19:49:10 2009 +0100
2860 Add libusb_get_max_iso_packet_size()
2862 As pointed out by Dennis Muhlestein, libusb_get_max_packet_size()
2863 doesn't really do what the documentation might suggest because it
2864 does not consider the number of transaction opportunities per
2867 Add a new function to do what is useful for isochronous I/O.
2869 commit 615f18e64e96ae4ecc8e43d0de00933059a5209a
2870 Author: Daniel Drake <dsd@gentoo.org>
2871 Date: Sat Jun 20 22:33:21 2009 +0100
2873 Linux: fix sending of zero length bulk packets
2875 Note that there are is a kernel bug preventing this from working
2876 properly at the moment, even after this fix.
2878 commit 86f79fbf61c2018bdf009c7ebf92b38f3a16fd0c
2879 Author: Nathan Hjelm <hjelmn@mac.com>
2880 Date: Fri Jun 19 22:18:44 2009 +0100
2882 Darwin: Don't cancel transfers on timeout
2884 ...because the OS does this for us.
2886 commit d4bd9ed4de19a9c766f7a23eea8c852cdd61c18f
2887 Author: Daniel Drake <dsd@gentoo.org>
2888 Date: Sun Jun 14 19:23:21 2009 +0100
2890 Don't terminate enums with commas
2892 g++ -pedantic doesn't like this
2893 Reported by Eberhard Mattes
2895 commit 8be256082eb9bd8e243d89529f742926bb29a21b
2896 Author: Daniel Drake <dsd@gentoo.org>
2897 Date: Sat Jun 13 17:19:07 2009 +0100
2901 commit 060e006e663fd59c281be29b71eb197e02b210e8
2902 Author: Daniel Drake <dsd@gentoo.org>
2903 Date: Wed Jun 10 21:42:05 2009 +0100
2905 Linux: fix config descriptor parsing on big-endian systems
2907 Multi-byte fields in the configuration descriptors that come back from
2908 usbfs are always in bus endian format.
2910 Thanks to Joe Jezak for help investigating and fixing this.
2912 commit c4a905022f684da9a4a853eb9232a81a53df2652
2913 Author: Nathan Hjelm <hjelmn@mac.com>
2914 Date: Sun Jun 7 22:29:35 2009 +0100
2916 Darwin: improve handling of disconnected devices
2918 commit 6b69f54451762ef590b9c938ab000c07cf9099a3
2919 Author: Nathan Hjelm <hjelmn@mac.com>
2920 Date: Sun Jun 7 22:26:37 2009 +0100
2922 Darwin: fix parsing of config descriptors
2924 This was a confusion between configuration numbers and zero-based
2925 configuration indexes.
2927 commit 2b3a9ffa776b383cb2dbc3c55e490e32e4c3c22b
2928 Author: Daniel Drake <dsd@gentoo.org>
2929 Date: Sun Jun 7 22:19:53 2009 +0100
2931 Eliminate -Wsign-compare compiler warnings
2933 This was due to an API inconsistency which can be safely worked around.
2934 Hopefully we'll remember to fix the API next time we come to break
2937 commit cad5cb55c37137e94e35c74fdabfe42a5cbd229b
2938 Author: David Moore <dcm@acm.org>
2939 Date: Wed May 27 23:15:54 2009 -0700
2941 Make synchronous transfer APIs robust against signal interruption
2943 libusb_control_transfer and libusb_bulk_transfer are designed to be
2944 synchronous such that control is not returned until the transfer
2945 definitively succeeds or fails. That assumption is violated if a signal
2946 interrupts these functions because there is no way for the application
2947 to continue waiting for the transfer without resubmitting it. This
2948 patch changes these synchronous APIs so they do not abort in the case of
2949 a signal interruption.
2951 Signed-off-by: David Moore <dcm@acm.org>
2953 commit b501795985a23109f176d296e7b544b4c6354528
2954 Author: Alex Vatchenko <alex@fabulatech.com>
2955 Date: Thu May 28 15:58:50 2009 -0400
2957 pre-gcc-3.4 compatibility
2959 The -fvisibility and -Wno-pointer-sign options are not available on
2962 commit 00bb2805e994887f0a754a825c3ce03d22393386
2963 Author: Daniel Drake <dsd@gentoo.org>
2964 Date: Tue May 26 15:53:26 2009 -0400
2966 Fix memory leak in config descriptor parsing
2968 Pointed out by Martin Koegler.
2970 commit 068ff5b8a83fec0a9a91c80535a25b89a9ae64e8
2971 Author: Daniel Drake <dsd@gentoo.org>
2972 Date: Tue May 12 19:20:51 2009 -0400
2976 commit aa24c04e4043e39674b59ff0d302b2365cd0078f
2977 Author: Daniel Drake <dsd@gentoo.org>
2978 Date: Tue May 12 19:03:37 2009 -0400
2982 commit 44767677447fae4267131f99c591b14117486cac
2983 Author: Daniel Drake <dsd@gentoo.org>
2984 Date: Tue May 12 19:01:44 2009 -0400
2988 Probably missed a couple of outstanding issues
2990 commit e0365a6f22e0e7f330ac931df031cf6f17f453d1
2991 Author: Nathan Hjelm <hjelmn@mac.com>
2992 Date: Sun Mar 22 21:13:29 2009 -0400
2994 Darwin: get_config_descriptor bugfixes
2996 commit 5fd0e8478240fece646a58a3c6114001a73be99f
2997 Author: Daniel Drake <dsd@gentoo.org>
2998 Date: Wed Feb 25 12:33:33 2009 -0300
3000 Fix compilation of Darwin backend
3002 My fault. Reported by ihryamzik@gmail.com
3004 commit b49f6bf5c910d0fd694ecf165d7927673707bff9
3005 Author: Nathan Hjelm <hjelmn@mac.com>
3006 Date: Mon Feb 16 21:39:29 2009 -0300
3010 commit d859158581e9a3250f36cdeeb8ea67cda04053bd
3011 Author: Nathan Hjelm <hjelmn@mac.com>
3012 Date: Mon Feb 16 21:30:49 2009 -0300
3014 Make endianness macros endian-independent
3016 Implementation suggested by David Moore. Needed for proper universal
3017 code support on Darwin.
3019 commit 9196f58bdc8b9e967261df39865215faa5d39cfa
3020 Author: Nathan Hjelm <hjelmn@mac.com>
3021 Date: Mon Feb 16 21:25:18 2009 -0300
3023 Abstract clock reading into OS layer
3025 This will differ on Linux and Darwin, at least.
3027 [dsd: minor style tweaks]
3029 commit e91207860cac09b3afaafc4c14221b78d585c59a
3030 Author: Daniel Drake <dsd@gentoo.org>
3031 Date: Mon Feb 16 21:20:28 2009 -0300
3033 Fix endianness in device descriptors
3035 Pointed out by Nathan Hjelm.
3037 commit ccf6d3d2f6acf9b9a1cb5e1f1b03f382ae509b17
3038 Author: Nathan Hjelm <hjelmn@mac.com>
3039 Date: Mon Feb 16 21:13:52 2009 -0300
3041 Only link with librt on Linux
3043 [dsd: tweak configure.ac change]
3045 commit 894539931e4b4bd85708fe98b956049243cd6fb0
3046 Author: Hans Ulrich Niedermann <hun@n-dimensional.de>
3047 Date: Mon Jan 26 00:31:48 2009 +0100
3049 API docs: describe libusb_transfer_cb_fn type
3051 Add some text describing the libusb_transfer_cb_fn function type
3052 with the semantics I have gathered from reading other parts of
3053 the API docs, referring to the proper section for more details.
3055 [dsd: tweaked the description slightly]
3057 commit abe34a2656f8f9f21e53603796c536585e6233ef
3058 Author: Hans Ulrich Niedermann <hun@n-dimensional.de>
3059 Date: Sun Jan 25 18:21:59 2009 +0100
3061 Make empty array in struct compatible with C99
3063 If the compiler is known to be running in C99 mode,
3064 use "flexible array members" ("foo[]").
3066 If the compiler is running in any other mode, continue
3067 using the non-standard but widely common "foo[0]" syntax.
3069 commit 620075c7400764d9bb539b5c02065c45c2e8251e
3070 Author: Hans Ulrich Niedermann <hun@n-dimensional.de>
3071 Date: Sun Jan 25 18:20:45 2009 +0100
3073 Avoid signedness errors in API docs example code
3075 commit c754ae294cfe96ec4738d6641137c9e6c56330c7
3076 Author: Daniel Drake <dsd@gentoo.org>
3077 Date: Sat Jan 17 00:25:40 2009 +0000
3079 Decrement poll() result when internal pipe has been handled
3081 When we receive data on the internal control pipe, we need to correctly
3082 decrement the number of ready file descriptors before passing on the
3083 remaining work to the OS implementation.
3085 commit b9ca960f2ba271d2b1a58e22b7c70464d69f6c8a
3086 Author: David Moore <dcm@acm.org>
3087 Date: Sun Jan 11 21:46:17 2009 -0800
3089 Prevent transfer from being submitted twice or improperly canceled
3091 This ensures that tpriv->urbs and tpriv->iso_urbs are always set to NULL
3092 whenever a transfer is not submitted. In this way, submit_*_transfer()
3093 and cancel_*_transfer() can error check to ensure that the transfer is
3094 in the correct state to be either submitted or canceled, preventing
3095 potential memory leaks or double frees.
3097 Signed-off-by: David Moore <dcm@acm.org>
3099 commit d2a8ec2da8abcd8f4648ed118da16191011982dd
3100 Author: David Moore <dcm@acm.org>
3101 Date: Wed Jan 7 22:31:09 2009 -0800
3103 Linux: Fix race condition in cancel_bulk_transfer()
3105 This fixes a race condition in cancel_bulk_transfer(). In the old
3106 version, awaiting_reap and awaiting_discard are incremented in
3107 cancel_bulk_transfer() and decremented in handle_bulk_completion().
3108 However, since these events may take place in two different threads,
3109 these variables may reach zero before all URBs have been canceled,
3110 triggered spurious callbacks and duplicate frees.
3112 This changes the logic to use a single variable "num_retired" to replace
3113 both awaiting_reap and awaiting_discard. num_retired is incremented
3114 only in handle_bulk_completion() and thus there is no race. The handler
3115 will know that all URBs have been canceled when num_retired becomes
3118 This change also simplifies a great deal of the logic in both functions
3119 and is a net reduction in the amount of code.
3121 Note that some variables such as "reap_action" probably need to still be
3122 protected by a mutex, and this patch does not address that issue.
3124 Signed-off-by: David Moore <dcm@acm.org>
3126 commit 34b9eebe35d8167d43cffb6ad6175f6b2251b572
3127 Author: Daniel Drake <dsd@gentoo.org>
3128 Date: Sat Dec 13 20:06:49 2008 +0000
3134 commit b1d636dde418dc8fe6a8f037dea0d800e57c6b94
3135 Author: Mikhail Gusarov <dottedmag@dottedmag.ent>
3136 Date: Wed Dec 10 15:24:05 2008 +0600
3138 EOL-whitespace fixes
3140 Signed-off-by: Mikhail Gusarov <dottedmag@dottedmag.ent>
3142 commit 4a9ac382ba5149ae5b3d0962d6a840b5e0dc13aa
3143 Author: Daniel Drake <dsd@gentoo.org>
3144 Date: Tue Dec 9 21:43:37 2008 +0000
3146 Linux: fix reading of active configuration
3148 Shannon Chuang pointed out that we only read 1 byte into an
3149 uninitialized integer, and then return the whole thing.
3151 commit 81a7310d92461ba7b1d98aeff7c4c007fab07120
3152 Author: Daniel Drake <dsd@gentoo.org>
3153 Date: Tue Dec 9 21:39:11 2008 +0000
3155 Fix double free in descriptor parsing error path
3157 Pointed out by Shannon Chuang.
3159 commit 9b4b53453db56ba9c1d707a645bbe6c7a02a3c81
3160 Author: Daniel Drake <dsd@gentoo.org>
3161 Date: Fri Nov 21 13:28:32 2008 +0000
3163 Improvements to multi-threaded I/O system
3165 Documentation brushed up, and I realised that another function is
3166 needed for tight event handling loops -- they must be able to check
3167 if an open/close operation is trying to interrupt it.
3169 commit 0e5b0fcb77a90b8bd95ad23669da472af31ef069
3170 Author: Daniel Drake <dsd@gentoo.org>
3171 Date: Fri Nov 21 11:00:47 2008 +0000
3173 SourceForge website upload system changed
3175 commit a133875e6e2f52eff53ac03f505d59cbdd33e178
3176 Author: Daniel Drake <dsd@gentoo.org>
3177 Date: Fri Nov 21 10:54:57 2008 +0000
3181 commit c32aa662769b676ff3247778664fccc71fc427ec
3182 Author: Daniel Drake <dsd@gentoo.org>
3183 Date: Thu Nov 20 15:31:10 2008 +0000
3185 Pause event handling while opening and closing devices
3187 Ludovic Rousseau found that crashes often occur if you close a device
3188 while another thread is doing event handling.
3190 Fix this by adding an internal control pipe, which the close routines
3191 use to interrupt the event handler and obtain the event handling lock,
3192 ensuring that no other thread is handling events while the device is
3193 closed. After the close completes, it signals all the event handlers
3194 to start up again using the usual mechanism.
3196 Also modified libusb_open() to do a similar thing, so that event
3197 handlers are interrupted in order to realise that a new poll fd has
3200 commit 1d7cf3d0fa8698eae25097cbda1870be90ff6f5e
3201 Author: Bastien Nocera <hadess@hadess.net>
3202 Date: Thu Oct 30 14:35:23 2008 +0000
3204 Add libusb_attach_kernel_driver()
3206 Add support for re-attaching a driver to a device under Linux.
3208 [dsd: fixed handling of return value, and added LIBUSB_ERROR_BUSY case]
3210 commit 914a4e70657c86b5094770aa2d898c978b1cdf41
3211 Author: Daniel Drake <dsd@gentoo.org>
3212 Date: Sun Nov 2 21:45:54 2008 +0000
3214 Linux: handle low-level transfer errors
3216 Handle more URB error status codes, thanks to Lou and Alan Stern.
3218 commit d25b566b3b8febafdda4211de724b4727dd4b7e0
3219 Author: Aurelien Jarno <aurelien@aurel32.net>
3220 Date: Thu Sep 4 13:50:49 2008 +0200
3222 Support for out-of-tree building
3224 Contrary to libusb 0.1, libusb 1.0 does not support out-of-tree building
3225 for the documentation part. This patch fixes that by using a
3226 doxygen.cfg.in file, which contains @top_srcdir@ to refer to the
3227 location of the libusb source code.
3229 commit 8674c67ef78e1cf89db1fa584a4304f7c5ddcc5f
3230 Author: Daniel Drake <dsd@gentoo.org>
3231 Date: Sun Nov 2 15:00:39 2008 +0000
3233 Document the logging style
3237 commit 94936cbcfe3f02eb65c8b91e29896604316259d8
3238 Author: Daniel Drake <dsd@gentoo.org>
3239 Date: Wed Aug 27 22:44:24 2008 -0500
3241 Async I/O documentation touchups
3243 commit 2f8f1b1a900f5b9828e5e2ff93c2b26a44fd9de2
3244 Author: Daniel Drake <dsd@gentoo.org>
3245 Date: Wed Aug 27 21:23:10 2008 -0500
3247 Doc: fix typo in events lock section
3249 commit 046e681e4ac6b697e40458ee0f5baf3cd8763bb0
3250 Author: Daniel Drake <dsd@gentoo.org>
3251 Date: Sat Aug 23 01:00:21 2008 -0500
3255 commit fec6eaa4eecdf7e0b8299157a5dabef94417f193
3256 Author: Daniel Drake <dsd@gentoo.org>
3257 Date: Sat Aug 23 00:57:44 2008 -0500
3259 Reset internal transfer flags on submit
3261 This fixes a problem pointed out by Lou, where resubmitting a transfer
3262 that previously timed out never timed out again, as if the timeout had
3265 commit 81627bd38900ec9701ab69c141aa51a9abea1f60
3266 Author: Daniel Drake <dsd@gentoo.org>
3267 Date: Sun Aug 10 17:09:48 2008 -0500
3269 Add libusb_get_configuration prototype
3273 commit 1631eaefcc7cdf7047c6b9225618a9aa1b3d87d1
3274 Author: Daniel Drake <dsd@gentoo.org>
3275 Date: Sat Jul 19 17:51:28 2008 -0500
3279 commit 0d9c40d7f3eee15ada624b5caab2cfb912144fdc
3280 Author: Daniel Drake <dsd@gentoo.org>
3281 Date: Wed Jul 16 09:22:12 2008 -0500
3283 Linux: fix sysfs directory handling
3285 The length of the directory name varies, e.g. when you have nested hubs.
3286 Use dynamic allocation to be able to deal with any length of name.
3288 commit 66c9847ad54b126ccafd4877292454ab57d0c115
3289 Author: David Engraf <dogge2k@gmx.de>
3290 Date: Wed Jul 16 09:16:38 2008 -0500
3292 Linux: correct usage of MAX_PATH
3294 Thanks to clarification from Artem Egorkine, MAX_PATH already includes
3295 space for the trailing NULL.
3297 commit 6494f07a6d1b73ce786ae581c4ff691e39e61d19
3298 Author: Daniel Drake <dsd@gentoo.org>
3299 Date: Tue Jul 8 21:15:54 2008 -0500
3301 Linux: fix bulk/iso transfer double-cancellation
3303 Reset counters to 0 so that it's possible to cancel a transfer twice
3304 without breaking things. Not sure that I want to support this properly,
3305 but this makes it work at least.
3307 commit 9140f5b414e37a3714cde9c6d6c87c963338a83a
3308 Author: Daniel Drake <dsd@gentoo.org>
3309 Date: Sat Jun 28 21:28:44 2008 -0500
3313 commit 83a029062c1d2bfb584d7a6ee94915583d37464c
3314 Author: Daniel Drake <dsd@gentoo.org>
3315 Date: Sat Jun 28 21:24:49 2008 -0500
3317 Add missing GET_CONTEXT()
3319 commit fbad9a5426f0369394c88987355a8fb06a741ca1
3320 Author: Daniel Drake <dsd@gentoo.org>
3321 Date: Sat Jun 28 21:18:41 2008 -0500
3323 Linux: Compatibility with new sysfs descriptors file
3325 As of 2.6.26, the descriptors file now includes all descriptors, not
3326 just the active one.
3328 commit 819e65f880ca43526036e56c65c415042c91f58f
3329 Author: Daniel Drake <dsd@gentoo.org>
3330 Date: Thu Jun 26 22:47:05 2008 -0500
3332 Allow user data pointer to be passed through pollfd notification API
3334 commit 1df713d622ab4f0b03aad72d903ac7beb8fb3b90
3335 Author: Daniel Drake <dsd@gentoo.org>
3336 Date: Tue Jun 24 23:01:51 2008 -0500
3338 Introduce contexts to the API
3340 Suggested by David Zeuthen. This allows multiple libraries in the same
3341 process to independently use libusb without interfering.
3343 commit 9818151c60a85aea6af24cb0996a92c3726c9864
3344 Author: Daniel Drake <dsd@gentoo.org>
3345 Date: Tue Jun 24 21:31:13 2008 -0500
3347 Revert "Temporary workaround for event handling serialization issue"
3349 This reverts commit 2d3a1111caff40ebb87983c861ff548cdc9e5946.
3350 This was based on the assumption that dying threads would automatically
3351 release mutexes, which is not the case.
3353 commit d5f82893fab3f1c13b1af4ba17aac72479bad7d5
3354 Author: Daniel Drake <dsd@gentoo.org>
3355 Date: Fri Jun 20 23:04:53 2008 -0500
3359 commit 546dee211eefbdd280fd1fc9dee84a9b52105078
3360 Author: Artem Egorkine <arteme@gmail.com>
3361 Date: Tue Jun 17 18:27:38 2008 -0500
3363 a fix for SIGSEGV in handle_bulk_completion()
3365 We cannot dereference tpriv after calling
3366 usbi_handle_transfer_cancellation() because that function may invoke
3367 the user-supplied callback which may free the transfer.
3369 commit 947ba8056456a5215724fb502e3e09d50016f699
3370 Author: Daniel Drake <dsd@gentoo.org>
3371 Date: Mon Jun 16 22:50:50 2008 -0500
3373 Refine configuration selection again
3375 At Alan Stern's suggestion, just offer the bare "set configuration" and
3376 "get configuration" functionality, and let applications worry about the
3377 specific race conditions and unusual situations.
3379 commit e7a7a49d0331ee0e14145f6e7ec39763b36314ac
3380 Author: Daniel Drake <dsd@gentoo.org>
3381 Date: Sun Jun 15 15:29:38 2008 -0500
3385 Add a THANKS file, make copyright notices easily accessible, update TODO
3387 commit 217534e4b396081e038ee5f94c813d1668963673
3388 Author: Daniel Drake <dsd@gentoo.org>
3389 Date: Sun Jun 15 15:23:29 2008 -0500
3391 round up poll timeout
3393 Pointed out by Richard Röjfors, otherwise we end up busy-looping with
3394 a poll() timeout of 0.
3396 commit 5ad79b324bc5e11a75a585398a1e81c26f05e758
3397 Author: Felipe Balbi <me@felipebalbi.com>
3398 Date: Sun Jun 8 12:34:48 2008 +0300
3400 fix doc about libusb_open()
3402 libusb_open() returns int instead of libusb_device_handle.
3404 Signed-off-by: Felipe Balbi <me@felipebalbi.com>
3405 [dsd: small correction]
3407 commit 2d3a1111caff40ebb87983c861ff548cdc9e5946
3408 Author: Daniel Drake <dsd@gentoo.org>
3409 Date: Thu May 29 17:26:26 2008 +0100
3411 Temporary workaround for event handling serialization issue
3413 Ludovic Rousseau pointed out that libusb_unlock_events() is not called
3414 when a thread gets terminated with a signal, meaning that event waiters
3415 will not be woken up in this case.
3417 Add a temporary hack to libusb_event_handler_active() so that at least
3418 the other threads will realise on the next iteration of their event
3421 The real fix will likely involve reworking most of this.
3423 commit 514bb8790cfe8b93ccfff82bc17081b1030acce0
3424 Author: Daniel Drake <dsd@gentoo.org>
3425 Date: Thu May 29 12:35:01 2008 +0100
3427 Refine libusb_set_configuration() semantics
3429 Applications will generally want to set a configuration before claiming
3430 interfaces. The problem is that the interface may already be set, and
3431 someone else may have claimed an interface (meaning that all calls to
3432 set_configuration will fail, even if it's for the same configuration).
3434 There are now 2 options:
3435 1. Use the new libusb_get_configuration() to determine active
3436 configuration before calling libusb_set_configuration()
3437 2. Or just call libusb_set_configuration() as usual, which will do
3438 nothing if that configuration is already active.
3440 commit d1292f8e7300051239a7ed2769d221dc7a6f9fca
3441 Author: Daniel Drake <dsd@gentoo.org>
3442 Date: Sun May 25 23:42:42 2008 +0100
3446 First libusb-1.0 beta release
3448 commit 4cd249388e4ed2ca03f1263984ab67df5030bb65
3449 Author: Daniel Drake <dsd@gentoo.org>
3450 Date: Sun May 25 22:59:09 2008 +0100
3452 Take lock before raising event waiters condition
3454 This avoids a race between the user checking for active event handler
3455 and then blocking on the condition variable
3457 commit 7c525480ab3c4db9205bfe30bb0d2bef9096c444
3458 Author: Daniel Drake <dsd@gentoo.org>
3459 Date: Fri May 23 15:57:07 2008 +0100
3461 Implement serialization of event handlers
3463 Now offers a mechanism to wait for events while another thread is
3464 doing the event handling. Complicates things for MT async apps, but
3465 then again it's a bit of a tricky combination to start with.
3467 commit ff0660a415ecfd0879600eaad1c5899b6d93a30a
3468 Author: Daniel Drake <dsd@gentoo.org>
3469 Date: Sat May 24 21:28:31 2008 +0100
3471 allow LIBUSB_DEBUG env var to control message verbosity at runtime
3473 Based on ideas from Ludovic Rousseau
3475 commit 4d788967e3f8d75eaf3a1ac1ee8e2e8bed0601c1
3476 Author: Daniel Drake <dsd@gentoo.org>
3477 Date: Fri May 23 15:58:11 2008 +0100
3479 don't print messages by default
3481 Add libusb_set_debug() API to set message verbosity.
3483 Ludovic Rousseau pointed out that applications may close stdout/stderr
3484 descriptors, which might then be reused.
3486 commit bef33bb9eba0da04ee7488d9cd5e6ab12bc61c0c
3487 Author: Daniel Drake <dsd@gentoo.org>
3488 Date: Mon May 19 15:43:27 2008 +0100
3490 Fix memory leak in libusb_control_transfer
3492 Pointed out by David Engraf: we weren't freeing the transfer buffer
3494 commit a345bacb6f07ebb2122402041a9e8092d4a20a4d
3495 Author: David Engraf <david.engraf@netcom.eu>
3496 Date: Mon May 19 15:16:32 2008 +0100
3498 Linux: fix handling of ioctl failure
3500 The return value of some ioctl commands in linux_usbfs.c are not
3501 handeled correct. The ioctl function returns != 0 and errno is set with
3504 commit e44396a458ecea9e5edd9a7577e617571c76860d
3505 Author: David Engraf <david.engraf@netcom.eu>
3506 Date: Mon May 19 15:13:38 2008 +0100
3508 critical memory leak in handle_events
3510 This patch closes a critical memory leak in handle_events. The fds
3511 variable is malloced but never freed. When I'm calling
3512 handle_events with a timeout of 0, my system runs out of memory after a
3515 commit eb25630f52bc9848b444e439632c899977d887b0
3516 Author: Daniel Drake <dsd@gentoo.org>
3517 Date: Fri May 16 23:50:20 2008 +0100
3519 More informative libusb_open() return code
3521 Hopefully one of the last API tweaks...
3523 commit 2b2e9c40b195261b09ac52ebdb93eef25c79de90
3524 Author: Daniel Drake <dsd@gentoo.org>
3525 Date: Fri May 16 22:37:56 2008 +0100
3527 Fix endianness with descriptor handling
3529 Alan Stern pointed out that usbfs gives host-endian data, but sysfs gives
3532 commit 1298c51f516a7bf04ca9add1b7db14417cdc66f3
3533 Author: Daniel Drake <dsd@gentoo.org>
3534 Date: Mon May 12 18:46:37 2008 +0100
3536 Backend documentation for porting efforts
3538 Hopefully comprehensive enough for people to get started.
3540 commit ade26afc42c34ceb1c45afcadd2ea5e8240eaca4
3541 Author: Daniel Drake <dsd@gentoo.org>
3542 Date: Mon May 12 15:43:30 2008 +0100
3544 Linux: fix caching of guessed configuration
3546 Reported and tested by Xiaofan Chen
3548 commit e25d590a9198995b4f0b6afeb41ecae318715e7e
3549 Author: Rob Walker <rob@tenfoot.org.uk>
3550 Date: Sun May 11 21:14:18 2008 +0100
3552 Linux: fix bulk transfer early completion
3554 We were forgetting about the remaining urbs when a non-final urb
3557 [dsd: some touchups and a warning message for a corner case that we don't handle]
3559 commit fec7c84163e25b8f811632828334d75da82bcb16
3560 Author: Daniel Drake <dsd@gentoo.org>
3561 Date: Sun May 11 20:31:58 2008 +0100
3563 Handle hot-unplugging
3565 This involved moving from select() to poll() because there is no way to
3566 distinguish usbfs's POLLERR condition with select().
3568 commit aeb905fa5d78cdbba80a680aa7a2bb7338f27f65
3569 Author: Daniel Drake <dsd@gentoo.org>
3570 Date: Sun May 11 16:04:35 2008 +0100
3574 commit ba5d9a45c06311204f51faef41d1ee215bb5b823
3575 Author: Daniel Drake <dsd@gentoo.org>
3576 Date: Sun May 11 15:36:24 2008 +0100
3578 Endianness of control setup packets
3580 Document behaviour where host-endian and bus-endian (little) should be
3583 Also remove packed attribute from libusb_control_setup as all fields
3584 are naturally aligned.
3586 commit 7bedc3b2683f35ee1dd39a2ebe0ec05be0019f38
3587 Author: Daniel Drake <dsd@gentoo.org>
3588 Date: Sun May 11 15:11:59 2008 +0100
3590 Document that releasing interface causes SET_INTERFACE control request
3592 commit f2ede9876cd4f5cfa7751b975670fa449187fe3d
3593 Author: Daniel Drake <dsd@gentoo.org>
3594 Date: Sat May 10 21:45:42 2008 +0100
3596 Support unconfigured devices
3598 commit 8ebb4ccdfaf5f095a1c38787d909d280ea64405c
3599 Author: Daniel Drake <dsd@gentoo.org>
3600 Date: Sat May 10 20:56:44 2008 +0100
3602 Linux: fix caching of active configuration for non-sysfs
3604 commit c3844f7aeb2176636ce6e6ef697659fdb0b30048
3605 Author: Daniel Drake <dsd@gentoo.org>
3606 Date: Sat May 10 14:42:43 2008 +0100
3608 Fetch configurations by index (not value)
3610 Otherwise there is no way to know which values to look for.
3612 commit d77052c0d630e33737c38d601fd633155f6b2229
3613 Author: Daniel Drake <dsd@gentoo.org>
3614 Date: Sat May 10 00:12:53 2008 +0100
3616 Linux: comprehensive sysfs vs usbfs access
3618 Be more flexible when certain parts of sysfs are not available.
3620 commit 74bc842bac1a32a26323da6c3e8af2f66e1b5cfa
3621 Author: Daniel Drake <dsd@gentoo.org>
3622 Date: Fri May 9 19:38:13 2008 +0100
3624 Linux: no need to store usbfs node path
3626 This can be computed from bus number and device address
3628 commit fe4adcc99e30115204ab832ad3e0170c9aca7629
3629 Author: Daniel Drake <dsd@gentoo.org>
3630 Date: Fri May 9 14:34:31 2008 +0100
3632 Rework configuration handling
3634 libusb no longer caches descriptors in libusb_device but backends are
3635 intended to be able to provide copies from memory. In the common linux
3636 case we can use sysfs.
3638 commit 5741bfe01a2481b8c3830c80edc3637bf62a7e16
3639 Author: Daniel Drake <dsd@gentoo.org>
3640 Date: Fri May 9 14:36:14 2008 +0100
3642 libusb_get_device_list() can return negative error code
3644 commit 45413101b78298e9332b22a34bc6bc159000ad8a
3645 Author: Daniel Drake <dsd@gentoo.org>
3646 Date: Thu May 8 23:09:21 2008 +0100
3648 Linux: fix clear_halt implementation
3650 The ioctl expects an integer.
3651 Bug found with the assistance of Soumen Mondal
3653 commit 10d4e427cc171dfd6ad7f43a33ce3cfebcd7aa04
3654 Author: Daniel Drake <dsd@gentoo.org>
3655 Date: Thu May 8 23:04:52 2008 +0100
3659 Partially based on some libusb-devel discussions
3661 commit ded0a249322571a075e3ed3528021864247dfa55
3662 Author: Daniel Drake <dsd@gentoo.org>
3663 Date: Mon May 5 22:57:48 2008 +0100
3665 Linux: enumerate devices and descriptors from sysfs
3667 Suggested by Alan Stern. This avoids waking up any suspended USB devices.
3668 sysfs is not available on all systems, so the usbfs mechanism is still
3669 in place as a fallback.
3671 commit 885c2a5de69d6b7d8902bb55d6d83680a5a1a6e5
3672 Author: Daniel Drake <dsd@gentoo.org>
3673 Date: Mon May 5 21:34:31 2008 +0100
3675 documentation touchups
3677 commit 59c205d542b43d79fe28622dbe8f03a3a3300b6f
3678 Author: Daniel Drake <dsd@gentoo.org>
3679 Date: Mon May 5 21:17:03 2008 +0100
3681 more error code sanitization
3683 commit 88055d4b5913102a90ff666f75fd922c74860dc5
3684 Author: Daniel Drake <dsd@gentoo.org>
3685 Date: Mon May 5 20:57:43 2008 +0100
3687 Isochronous transfer helper functions
3689 commit a95c943ed301dcd20e92b0b3b255568899a4c42e
3690 Author: Daniel Drake <dsd@gentoo.org>
3691 Date: Mon May 5 17:54:33 2008 +0100
3693 free open devices on exit
3695 commit b27fff633843824744df7d334cb89ece329cafa6
3696 Author: Daniel Drake <dsd@gentoo.org>
3697 Date: Mon May 5 17:47:49 2008 +0100
3699 move descriptor parsing into main library
3701 OS modules now provide functionality for fetching device/config
3704 commit a304eca71f22c6df7d70a901483b30f1b8e93378
3705 Author: Daniel Drake <dsd@gentoo.org>
3706 Date: Mon May 5 16:22:33 2008 +0100
3708 Detect endpoint halts and unsupported control requests
3710 commit 7da521954ba661d3f537440c31a84b66e974d56b
3711 Author: Daniel Drake <dsd@gentoo.org>
3712 Date: Mon May 5 01:00:31 2008 +0100
3714 Linux: fix logical URB allocation
3716 We were allocating the wrong number of URBs for transfers of size
3719 commit 470b1bc42bf53373ce678fc76bab9160a54d6881
3720 Author: Daniel Drake <dsd@gentoo.org>
3721 Date: Sun May 4 16:51:23 2008 +0100
3723 add functionality for querying and detaching kernel driver
3725 commit 17ecfb0ecc833596c43755c80d461cddb9b3b0d7
3726 Author: Daniel Drake <dsd@gentoo.org>
3727 Date: Sun May 4 16:20:46 2008 +0100
3729 Descriptor reading functionality
3731 commit bdce367d1bd8691465844b2411c85215498f517d
3732 Author: Daniel Drake <dsd@gentoo.org>
3733 Date: Sun May 4 14:22:16 2008 +0100
3735 sanitized error returns from synchronous I/O functions
3737 commit bfe74e9cd9c17a40fff042ea0647326f51cfecae
3738 Author: Daniel Drake <dsd@gentoo.org>
3739 Date: Sun May 4 14:10:19 2008 +0100
3741 implement libusb_reset_device()
3743 commit c01f744ce07bbeccbe353e956479f0cc5a811a6f
3744 Author: Daniel Drake <dsd@gentoo.org>
3745 Date: Sun May 4 13:46:41 2008 +0100
3747 Add functionality to clear endpoint halt conditions
3749 commit 5878daa85e3364bb3885190425d4f9deaa2d8c36
3750 Author: Daniel Drake <dsd@gentoo.org>
3751 Date: Sun May 4 00:51:59 2008 +0100
3753 Fix libusb_get_device_list return value
3755 It was documented to return the list length, but was returning 0.
3757 commit a74106a9b48b88bae0e099be34a18d653a8f1238
3758 Author: Daniel Drake <dsd@gentoo.org>
3759 Date: Sat May 3 16:50:58 2008 +0100
3761 Well-defined error codes for some operations
3763 Also renamed set_interface_altsetting to set_interface_alt_setting for
3766 commit 0f463fe671455670efcf4a93e526b2a9082b0afe
3767 Author: Daniel Drake <dsd@gentoo.org>
3768 Date: Sat May 3 16:24:11 2008 +0100
3770 Add libusb_set_configuration()
3772 commit e65608994f795f97e4d6adac20329ca08709db2d
3773 Author: Daniel Drake <dsd@gentoo.org>
3774 Date: Sat May 3 15:40:23 2008 +0100
3776 rename LIBUSB_ERROR_NOMEM to LIBUSB_ERROR_NO_MEM
3777 improves consistency
3779 commit 5384bbf300779e7936eac9ebd33ee38f5b533f8e
3780 Author: Daniel Drake <dsd@gentoo.org>
3781 Date: Sat May 3 14:34:54 2008 +0100
3783 Linux: better EACCES handling
3785 commit 1936eeeba3f933a078df1005cc7c1533a8ec2bac
3786 Author: Daniel Drake <dsd@gentoo.org>
3787 Date: Sat May 3 14:26:56 2008 +0100
3789 Rename reference counting API
3791 This improves consistency with the naming style of other functions
3793 commit e2f822b4f17342f614ab2455179a80975fca6cb0
3794 Author: Daniel Drake <dsd@gentoo.org>
3795 Date: Sat May 3 14:24:40 2008 +0100
3797 Linux: use read-only access to enumerate devices and descriptors
3799 Write access is not needed until later.
3800 Also fix handling of open() error code.
3802 commit 81204eed3d0a7744b89b68bd90d567b68d191a91
3803 Author: Daniel Drake <dsd@gentoo.org>
3804 Date: Fri May 2 17:45:23 2008 +0100
3806 document pollfd notifiers
3808 commit e20f8281fb7da32587f81b5eea5818af1eab0fe0
3809 Author: Daniel Drake <dsd@gentoo.org>
3810 Date: Fri May 2 17:33:52 2008 +0100
3812 Allow multiple interfaces to be claimed
3814 commit 66685ff25d81a1f1a2a3dbda882d407638d00672
3815 Author: Daniel Drake <dsd@gentoo.org>
3816 Date: Fri May 2 16:55:36 2008 +0100
3818 Documentation improvements
3820 commit 55bce39cbe0c9746a428b66751d1afe551d86381
3821 Author: Daniel Drake <dsd@gentoo.org>
3822 Date: Fri May 2 12:55:34 2008 +0100
3824 detect short descriptor reads
3826 commit f3f17846dcced6b2dc40f899ed6a2aba151c264b
3827 Author: Daniel Drake <dsd@gentoo.org>
3828 Date: Fri May 2 12:32:46 2008 +0100
3830 remove "references" and "referenced by" listings in documentation
3832 commit fbb017dd9a26ce449578482b95157b3c850fa446
3833 Author: Daniel Drake <dsd@gentoo.org>
3834 Date: Fri May 2 12:32:08 2008 +0100
3836 Use typedef ptr types in parameters and return types
3838 This is passed through to the docs
3840 commit 3675e978fb7a1042f8601931255658bcd14a2298
3841 Author: Daniel Drake <dsd@gentoo.org>
3842 Date: Fri May 2 12:28:31 2008 +0100
3844 Expose bus number and device address
3846 commit 2012898ffc492317cb428d97077bf5ee0e98e1da
3847 Author: Daniel Drake <dsd@gentoo.org>
3848 Date: Tue Apr 29 15:26:46 2008 +0100
3850 standardised error codes
3852 commit 0269c75ae3de7acb2dc555e45928750e35fdb320
3853 Author: Daniel Drake <dsd@gentoo.org>
3854 Date: Tue Apr 29 13:34:26 2008 +0100
3856 Fix descriptor handling memory leaks
3858 commit d3ab4e3bd4068cba41c1e12c4b175018dc3cb343
3859 Author: Daniel Drake <dsd@gentoo.org>
3860 Date: Sun Apr 27 23:50:01 2008 +0100
3862 Rename endpoint_type to transfer_type
3864 Matches the USB specs more closely
3866 commit 379e90e0fdcef6f4db1849fa0a7fd524407cf564
3867 Author: Daniel Drake <dsd@gentoo.org>
3868 Date: Sun Apr 27 23:41:24 2008 +0100
3870 Remove a FIXME comment
3874 commit 2ddf81420557d016bbcbf28dda0d8d1c64d16264
3875 Author: Daniel Drake <dsd@gentoo.org>
3876 Date: Sun Apr 27 23:30:34 2008 +0100
3878 Fix unlocked flying_transfers list access
3880 commit 7ac0a405b4c57db42e88cbcba5f135697f03b646
3881 Author: Daniel Drake <dsd@gentoo.org>
3882 Date: Sun Apr 27 23:27:04 2008 +0100
3884 Remove libusb_cancel_transfer_sync
3886 This wasn't a particularly nice API. Cancellation should be handled by
3887 the completion handler.
3889 commit a5006fd7b9dae6f1db7969c8744086ba3a7c027b
3890 Author: Daniel Drake <dsd@gentoo.org>
3891 Date: Sun Apr 27 19:56:02 2008 +0100
3893 remove extra timeout check
3895 If select() didn't timeout then don't check for more timeouts
3897 commit 3bdafaa359d9c65e1af360b5338e1f25528a8066
3898 Author: Daniel Drake <dsd@gentoo.org>
3899 Date: Sun Apr 27 19:53:51 2008 +0100
3901 Documentation touchups
3903 Detail that interrupt polling interval is fixed to endpoint bInterval,
3904 and that data may be transferred even on a timeout condition.
3906 commit e12618c79d520955b1abf80a277e881d91907162
3907 Author: Daniel Drake <dsd@gentoo.org>
3908 Date: Sat Apr 19 16:17:53 2008 +0100
3910 add threaded example program
3912 commit 1fcdb0678b759569db7cd530457dbc0a5f86fb1d
3913 Author: Daniel Drake <dsd@gentoo.org>
3914 Date: Fri Apr 18 23:58:54 2008 +0100
3916 libusb_device mutex protection
3918 commit 0efd2efa65d5513e5754d717d522b2c5c45332e2
3919 Author: Daniel Drake <dsd@gentoo.org>
3920 Date: Fri Apr 18 23:42:40 2008 +0100
3922 mutex protection for device and handle lists
3924 commit d2ff2b14a15e5099ff0214502944de30f141a39c
3925 Author: Daniel Drake <dsd@gentoo.org>
3926 Date: Fri Apr 18 20:00:31 2008 +0100
3928 lock flying_transfers during transfer removal
3930 commit 2569980762cb66804da31e54bef243702da8dfe9
3931 Author: Daniel Drake <dsd@gentoo.org>
3932 Date: Fri Apr 18 19:59:56 2008 +0100
3934 pollfds mutex protection
3936 commit f5a33e08ec0da51ccb5e95a5adec50c5211a9bf5
3937 Author: Daniel Drake <dsd@gentoo.org>
3938 Date: Thu Apr 17 18:24:54 2008 +0100
3940 flying_transfers mutex protection
3942 This marks the beginnings of thread safety
3944 commit 85e1994123545125ce0653415d122f9790f8d041
3945 Author: Daniel Drake <dsd@gentoo.org>
3946 Date: Thu Apr 17 17:44:28 2008 +0100
3948 Linux: fix control transfer cancellation
3950 commit cf9b831f7bfd97ba88721c1d934aeda522f3692c
3951 Author: Daniel Drake <dsd@gentoo.org>
3952 Date: Mon Apr 14 20:03:27 2008 +0100
3954 set correct endpoint type for isochronous transfers
3956 Pointed out by Felix Domke
3958 commit ad6e2b712c5b54af44424e58a2776686314e26b7
3959 Author: Daniel Drake <dsd@gentoo.org>
3960 Date: Fri Apr 11 21:40:07 2008 +0100
3962 Linux: fire multiple URBs at once for split transfers
3964 This results in a significant performance increase for bulk transfers
3965 larger than 16kb, and fixes a bug where data would be spliced and mixed
3966 between two simultaneously submitted transfers to the same endpoint.
3968 It also allows isochronous transfers larger than 16kb to be submitted.
3970 This commit also improves cancellation - the library now understands
3973 commit d9b2ce2894ed43d34de6850f2dac50ccabf1db55
3974 Author: Daniel Drake <dsd@gentoo.org>
3975 Date: Tue Apr 1 23:28:32 2008 +0100
3977 Constify some return data
3979 Make it clear that apps should not mess with these structures
3981 commit 8d809854e2b19c2b7c27ab05e5d76b34e2a5cead
3982 Author: Daniel Drake <dsd@gentoo.org>
3983 Date: Sun Mar 30 22:19:00 2008 +0100
3985 Rename libusb_poll() to libusb_handle_events()
3987 commit 211f80c9f2a4a58cd2bbf5b7751f45089c8961e7
3988 Author: Daniel Drake <dsd@gentoo.org>
3989 Date: Tue Mar 25 16:24:30 2008 +0000
3991 Isochronous endpoint I/O
3993 Due to variable-sized structures, this involved changing allocation
3994 mechanism. All transfers must now be allocated and freed through
3997 A synchronous function is missing, and I could do with writing a few
3998 more helper functions to simplify things.
4000 commit b1ade6fca668d8aa156d5b5bf3a933f116144dc2
4001 Author: Daniel Drake <dsd@gentoo.org>
4002 Date: Sun Mar 23 21:17:32 2008 +0000
4004 distcheck fix for docs
4006 commit 62077b7fa5b222e1bb9dea3467585aed69d596e2
4007 Author: Daniel Drake <dsd@gentoo.org>
4008 Date: Sun Mar 23 21:13:14 2008 +0000
4010 Support for changing altsetting
4012 Will probably be suject to later consideration w.r.t. claiming of endpoints
4014 commit 0499e9f418607b5786ac0c3e97bc46737a609ca3
4015 Author: Daniel Drake <dsd@gentoo.org>
4016 Date: Thu Mar 20 21:10:01 2008 +0000
4018 it's called bmRequestType
4020 commit 88884ae4bbe721dac117d5190303cb67852fa5d3
4021 Author: Daniel Drake <dsd@gentoo.org>
4022 Date: Thu Mar 20 21:07:00 2008 +0000
4024 Save transfer flags before invoking callback
4026 The callback may free the transfer, in which case we can't read the flags
4027 after it has returned.
4029 commit ead09cde6895df0034a685516987ce253575e9a7
4030 Author: Daniel Drake <dsd@gentoo.org>
4031 Date: Sat Mar 15 16:35:12 2008 +0000
4035 Hopefully mostly complete. Some constants were renamed and move into
4038 commit 285724cc14ea5f993e2c4c92fe0aaf6c335bc139
4039 Author: Vasily Khoruzhick <anarsoul@gmail.com>
4040 Date: Sat Mar 15 13:50:53 2008 +0200
4042 Fix get_device_list for realloc case
4044 get_device_list can modify pointer passed to it with realloc, but this
4045 case wasn't handled and caused crash on my machine. Fixed
4047 Signed-off-by: Vasily Khoruzhick <anarsoul@gmail.com>
4049 commit c0c9432d38b22784070dce3a7874c62c31786a27
4050 Author: Daniel Drake <dsd@gentoo.org>
4051 Date: Thu Mar 13 12:36:56 2008 +0000
4053 Beginnings of cross-platform abstraction
4055 This also includes a libusb_get_pollfds API change
4057 commit de4c5341d168697baa4c0901c406deb47e78aae7
4058 Author: Daniel Drake <dsd@gentoo.org>
4059 Date: Mon Mar 10 18:38:15 2008 +0000
4063 commit aae05f688dc26a013166ab9436fd25018b49d76a
4064 Author: Daniel Drake <dsd@gentoo.org>
4065 Date: Mon Mar 10 11:32:15 2008 +0000
4067 Simplify libusb_fill_control_transfer()
4069 Transfer length can be inferred from the setup packet.
4071 commit 7c5ea95297c2f3f6afc42f60c1bf1bef777bb1fb
4072 Author: Daniel Drake <dsd@gentoo.org>
4073 Date: Mon Mar 10 11:29:24 2008 +0000
4077 Rename API away from "devh" and fix implementation to match header
4079 commit 9e89d1aa82cbe7e21672ee3e2e545238fffc5318
4080 Author: Daniel Drake <dsd@gentoo.org>
4081 Date: Sun Mar 9 16:27:06 2008 +0000
4083 Rename descriptor-getting functions
4085 Based on feedback from Tim Roberts
4087 commit 283ae96c3ccbcee1c3950fce18ae47093ae8200e
4088 Author: Daniel Drake <dsd@gentoo.org>
4089 Date: Sun Mar 9 16:15:02 2008 +0000
4091 Simplify cancellation API
4093 No need to provide the device here.
4095 commit ebad1c79688ba603ad017ed0fe2f3c0dc8edd1ad
4096 Author: Daniel Drake <dsd@gentoo.org>
4097 Date: Sun Mar 9 16:12:08 2008 +0000
4099 Rename libusb_dev_handle to libusb_device_handle
4101 commit 1ac0a7d88f282b6f293c456fac8edb143cbaca3d
4102 Author: Daniel Drake <dsd@gentoo.org>
4103 Date: Sun Mar 9 01:01:57 2008 +0000
4105 Move synchronous I/O implementation to its own file
4107 commit 66348c90ea4570bf999ac301089e006d0cce1926
4108 Author: Daniel Drake <dsd@gentoo.org>
4109 Date: Sun Mar 9 00:58:09 2008 +0000
4111 Separate transfer allocation and submission
4113 Leads to some hefty API changes. Now we're much more similar to the
4116 Problems with dealing with asynchronous control transfers are passed
4117 on to the user, basically you must allocate a buffer, start with the
4118 setup, and put the data after. This won't make much sense until
4119 documented (soon...)
4121 commit d21ebe47ce578c93cd8969be1c933d503e32e5d4
4122 Author: Daniel Drake <dsd@gentoo.org>
4123 Date: Sat Mar 8 12:48:35 2008 +0000
4125 Rework URB API naming
4127 Now refer to everything as "transfers" as consistent with the USB spec
4128 libusb_transfer is now a kind of transfer handle. To reduce confusion
4129 with libusb_bulk_transfer and libusb_control_transfer, those have been
4130 renamed to libusb_{control,bulk}_transfer_request.
4132 commit fd6fb3cc093ad47271353a0e468d16547f991382
4133 Author: Daniel Drake <dsd@gentoo.org>
4134 Date: Sat Mar 8 11:53:33 2008 +0000
4136 Add fallback on /proc/bus/usb
4138 /dev/bus/usb is a relatively new thing probably not present on every
4141 commit f3fdf447916289cd92b7190377681894e8ab611a
4142 Author: Daniel Drake <dsd@gentoo.org>
4143 Date: Sat Mar 8 00:10:16 2008 +0000
4145 Persistent libusb_device storage
4147 Devices are now assigned a session ID (currently busnum:devaddr) which
4148 is used to distinguish unique devices.
4150 Now multiple callers of libusb_get_device_list will get the same
4151 libusb_device structure instances.
4153 commit 23f8fb8bafeeda93ce234cb40eb7219d2e36c19c
4154 Author: Daniel Drake <dsd@gentoo.org>
4155 Date: Thu Mar 6 23:43:57 2008 +0000
4157 Add convenience function to find and open a device by USB VID+PID
4159 Lots of libusb apps I write are simple test apps not intended to be real
4160 apps. Having a function available to quickly locate my device will be
4161 handy in such situations.
4163 commit 9cfdb494fccac53a4277da7c8b6d15f1a72a4959
4164 Author: Daniel Drake <dsd@gentoo.org>
4165 Date: Thu Mar 6 23:25:20 2008 +0000
4167 Rework device discovery API
4169 libusb_find_devices and libusb_get_devices are no more
4171 libusb_get_device_list obtains a list of libusb_device structures for all
4172 known devices in the system.
4174 Each libusb_device now has a reference count, defaulting to 1 on
4175 instantiation. The reference count of 1 refers to the fact that it is
4176 present in the list in this scenario.
4178 Opening a device adds a pointer to the libusb_device structure in the
4179 handle, so that also adds a reference. Closing the device removes that
4182 The function to free the device list can optionally unref all the devices
4185 In future we will make the libusb_device instances all "global" so that if
4186 the app calls get_device_list twice it actually gets the same libusb_device
4187 structure references back. This way we can start to track disconnects, and
4188 we can investigate adding a unique "session ID" to each libusb_device, an
4189 identifier guaranteed to be unique to that device until reboot.
4191 commit 77cea822788e024c848c7f554915f771a2dc1e0f
4192 Author: Daniel Drake <dsd@gentoo.org>
4193 Date: Thu Mar 6 17:05:04 2008 +0000
4195 Remove ctrl debug code
4197 This was unintentionally committed
4199 commit b873521a6f575628d535f89278bb67888004d0ca
4200 Author: Daniel Drake <dsd@gentoo.org>
4201 Date: Tue Feb 26 22:11:34 2008 +0000
4203 Notifications for changes to the fd set
4205 Applications can now be notified when they should start and stop polling
4206 new file descriptors.
4208 commit ca66e16a370c828d08a5bc3f9f203db145c81a4c
4209 Author: Daniel Drake <dsd@gentoo.org>
4210 Date: Fri Feb 1 12:20:20 2008 +0000
4212 Fix header installation path
4214 commit e583100a4086c411093eef5c450880b965df6b1b
4215 Author: Daniel Drake <dsd@gentoo.org>
4216 Date: Wed Jan 30 13:24:54 2008 +0000
4220 commit 637a8d7ff8a11a23588925d9d3003a609bda8075
4221 Author: Daniel Drake <dsd@gentoo.org>
4222 Date: Wed Jan 30 13:19:10 2008 +0000
4224 Remove timers and signalfd
4226 Instead of timers, add a mechanism for informing the parent app when the
4227 next timeout is due to happen, so that it can call us at that time.
4229 As we no longer use signals, signalfd has also been removed.
4231 commit 8c56abe2e13ed4dd8c15737c21be7b6b1e5cf2ac
4232 Author: Daniel Drake <dsd@gentoo.org>
4233 Date: Sat Jan 26 18:16:56 2008 +0000
4235 Poll usbfs descriptors directly
4237 We don't need to use signalfd for this functionality because we can poll
4238 the file descriptors for write events.
4240 commit e3ce855acec33d37dd833bca5dad5e8c3773003a
4241 Author: Daniel Drake <dsd@gentoo.org>
4242 Date: Sat Jan 26 17:26:40 2008 +0000
4244 Fix dpfp example compilation
4246 commit 1d41a5cda2fefe826877ee4445199af30aa188b7
4247 Author: Daniel Drake <dsd@gentoo.org>
4248 Date: Sat Jan 26 17:22:53 2008 +0000
4254 commit 37b7f9a112eb1441d352467f661205b8872d0553
4255 Author: Daniel Drake <dsd@gentoo.org>
4256 Date: Sat Jan 26 14:03:43 2008 +0000
4258 Only use -fgnu89-inline when supported
4260 J F pointed out that this option only works on gcc 4.3 and newer.
4262 commit e3a09ca0b9cb6f46d54a0130f678c6097240a2bd
4263 Author: Daniel Drake <dsd@gentoo.org>
4264 Date: Sat Jan 5 21:59:40 2008 +0000
4266 Remove more fpusb remnants
4268 commit dcd5e0075fa84601d997d29bb60bc8370b97a375
4269 Author: Daniel Drake <dsd@gentoo.org>
4270 Date: Fri Jan 4 01:19:07 2008 +0000
4274 Start listing the things I want to discuss/reconsider before 1.0 API
4277 commit de53d972ed9c1f21f86ffd585560520e986f212c
4278 Author: Daniel Drake <dsd@gentoo.org>
4279 Date: Fri Jan 4 01:17:06 2008 +0000
4281 Rename some _devh_ functions
4283 Decided against naming functions in this style based on advice from
4284 Tim Roberts. Still not really sure about devh naming in general.
4286 commit 132d84bf8ae34bd0892b3d8207a53d87cc0a9cba
4287 Author: Daniel Drake <dsd@gentoo.org>
4288 Date: Fri Jan 4 01:13:25 2008 +0000
4290 Use "transfer" instead of "msg"
4292 Based on a discussion with Tim Roberts.
4294 commit ea6af58a8643137eb01eeb7d2f65404e390feb1b
4295 Author: Daniel Drake <dsd@gentoo.org>
4296 Date: Fri Jan 4 00:57:00 2008 +0000
4298 Plural of bus is busses
4302 commit dbb3fd871e3972b4e670f3161e7cd2f58f357600
4303 Author: Daniel Drake <dsd@gentoo.org>
4304 Date: Fri Jan 4 00:54:57 2008 +0000
4306 API renaming: remove fpi and usb_ stuff
4308 fpi changed to usbi.
4310 We should not expose structures with prefix "usb_" in the public
4311 namespace as it is quite likely there will be some conflict somewhere.
4312 Instead, using "libusb_" should be safer.
4314 commit e9364d72151ae1de9cce4175f330fe1529f02511
4315 Author: Daniel Drake <dsd@gentoo.org>
4316 Date: Fri Jan 4 00:40:49 2008 +0000
4318 Rename to libusb-1.0
4320 I've taken over the libusb project, and what was previously known as fpusb
4321 will eventually be released as libusb-1.0.
4323 commit f4ed0b6a3e0a20555fdfc9e9b8be7e8a2c8ae192
4324 Author: Daniel Drake <dsd@gentoo.org>
4325 Date: Sun Dec 30 22:09:50 2007 +0000
4327 Compiler flags update
4329 Also fix some warnings
4331 commit d3d671efe27491baf1d2ceadbfeeb4e7428d6f9a
4332 Author: Daniel Drake <dsd@gentoo.org>
4333 Date: Sun Dec 30 22:05:17 2007 +0000
4335 Fix signalfd structure size on x86
4337 ptr is a 64-bit value
4339 commit dfac6756faad76071a92945e3e3a16b8a841a7ea
4340 Author: Daniel Drake <dsd@gentoo.org>
4341 Date: Wed Dec 12 23:32:54 2007 +0000
4343 Don't pass ctrl_msg/bulk_msg structures through async paths
4345 These are easiest to construct on the stack, but in the async case it is
4346 likely that the submitting function returns (hence stack is destroyed)
4347 before URB completion.
4349 commit a8d2881eb7c273892acf2ff6e1f3f552631d1d11
4350 Author: Daniel Drake <dsd@gentoo.org>
4351 Date: Mon Dec 3 23:29:22 2007 +0000
4353 Add functionality to access device descriptors
4355 commit b5a7a411087b37a6bf56b4688a9e3290daaed153
4356 Author: Daniel Drake <dsd@gentoo.org>
4357 Date: Mon Dec 3 22:27:21 2007 +0000
4359 Rename fpusb_device_handle functions to devh
4361 Prevents confusion with functions that work on fpusb_dev structures.
4363 commit eacf4505a473df0f521d1b69d5b45448f0709e2d
4364 Author: Daniel Drake <dsd@gentoo.org>
4365 Date: Sun Dec 2 23:16:57 2007 +0000
4367 Add API to access poll fd
4371 commit 852bba4754ec57679c823f33e8feba6e4a564cbe
4372 Author: Daniel Drake <dsd@gentoo.org>
4373 Date: Wed Nov 28 13:48:45 2007 +0000
4377 Basic library structure which supports enumerating detected USB devices