]> sigrok.org Git - libsigrok.git/log
libsigrok.git
8 years agoAdd SR_T_MQLIST.
Bert Vermeulen [Sun, 16 Aug 2015 23:43:22 +0000 (01:43 +0200)]
Add SR_T_MQLIST.

This type consists of an array, with each item a two-member tuple,
representing an MQ/MQflags pair: the first item is the MQ (type
G_VARIANT_TYPE_UINT32), and the second is the MQ flags value
(G_VARIANT_TYPE_UINT64).

A GVariant of type SR_T_MQLIST can thus always represent more than
one MQ/MQflag pair.

8 years agoAdd key info tables for MQ and MQflags.
Bert Vermeulen [Sun, 16 Aug 2015 23:38:10 +0000 (01:38 +0200)]
Add key info tables for MQ and MQflags.

8 years agoReplace sr_config_info with sr_key_info.
Bert Vermeulen [Sun, 16 Aug 2015 23:30:14 +0000 (01:30 +0200)]
Replace sr_config_info with sr_key_info.

The tables defined with this struct can now be used for information
on items other than config keys.

Functions to access these tables have been renamed sr_key_info_[name_]get.
These take an extra argument, keytype, which should be set to SR_KEY_CONFIG
to get the config key tables. Other key types will be added.

8 years agoscpi/usbtmc: Use sr_usb_close().
Bert Vermeulen [Sun, 16 Aug 2015 18:02:21 +0000 (20:02 +0200)]
scpi/usbtmc: Use sr_usb_close().

8 years agoAdd sr_usb_close().
Bert Vermeulen [Sun, 16 Aug 2015 18:00:26 +0000 (20:00 +0200)]
Add sr_usb_close().

8 years agoscpi: Strip leading/trailing spaces from *IDN? response.
Bert Vermeulen [Mon, 17 Aug 2015 00:02:03 +0000 (02:02 +0200)]
scpi: Strip leading/trailing spaces from *IDN? response.

8 years agoscpi: Propagate error codes in API functions.
Bert Vermeulen [Sun, 16 Aug 2015 18:44:37 +0000 (20:44 +0200)]
scpi: Propagate error codes in API functions.

This allows client drivers to detect e.g. timeout conditions.

8 years agoMake sr_next_enabled_channel() from scpi-pps available library-wide.
Bert Vermeulen [Sun, 16 Aug 2015 13:54:02 +0000 (15:54 +0200)]
Make sr_next_enabled_channel() from scpi-pps available library-wide.

8 years agoscpi: Make helper functions from scpi-pps available library-wide.
Bert Vermeulen [Sun, 16 Aug 2015 13:12:34 +0000 (15:12 +0200)]
scpi: Make helper functions from scpi-pps available library-wide.

8 years agoscpi: Move SCPI-related definitions to separate header file.
Bert Vermeulen [Sat, 15 Aug 2015 23:04:04 +0000 (01:04 +0200)]
scpi: Move SCPI-related definitions to separate header file.

8 years agosigrok.m4: License header consistency fixes.
Uwe Hermann [Tue, 25 Aug 2015 15:01:22 +0000 (17:01 +0200)]
sigrok.m4: License header consistency fixes.

Use the same wording and formatting as in the rest of the codebase.

8 years agoBuild: Limit system features to C99 and POSIX
Daniel Elstner [Tue, 25 Aug 2015 10:50:35 +0000 (12:50 +0200)]
Build: Limit system features to C99 and POSIX

Use -std=c99 to try and enforce standard compliance.

8 years agoBuild: Fix SR_CHECK_COMPILE_FLAGS
Daniel Elstner [Mon, 24 Aug 2015 22:17:09 +0000 (00:17 +0200)]
Build: Fix SR_CHECK_COMPILE_FLAGS

Do not put "no" into the flags if the check fails.

8 years agoBuild: Reduce unnecessary config substitutions
Daniel Elstner [Mon, 24 Aug 2015 20:20:30 +0000 (22:20 +0200)]
Build: Reduce unnecessary config substitutions

Variable substitutions with a constant value may as well be
set directly in Makefile.am.

8 years agoBuild: Leave LIBS alone during configure
Daniel Elstner [Mon, 24 Aug 2015 20:06:27 +0000 (22:06 +0200)]
Build: Leave LIBS alone during configure

Put the extra libraries into SR_EXTRA_LIBS instead of LIBS.
Create an SR_CHECK_LIBS macro to make that easy. Substitute
SR_EXTRA_LIBS into libsigrok.pc, too.

8 years agoBuild: Simplify driver dependency check
Daniel Elstner [Mon, 24 Aug 2015 09:18:09 +0000 (11:18 +0200)]
Build: Simplify driver dependency check

8 years agoBuild: Add private copy of C++11 check macro
Daniel Elstner [Mon, 24 Aug 2015 07:25:37 +0000 (09:25 +0200)]
Build: Add private copy of C++11 check macro

Place a copy of ax_cxx_compile_stdcxx_11.m4 from the Autoconf
macro archive into our private m4/ directory. This is cleaner
than trying to parse M4 file versions etc. Plus, the macro is
now always available.

8 years agoBuild: Prefer newer versions of libftdi and SWIG
Daniel Elstner [Sun, 23 Aug 2015 22:08:01 +0000 (00:08 +0200)]
Build: Prefer newer versions of libftdi and SWIG

8 years agoBuild: Use commas to separate missing dependencies
Daniel Elstner [Sun, 23 Aug 2015 21:53:01 +0000 (23:53 +0200)]
Build: Use commas to separate missing dependencies

Beautify the output. Also fix the PyGObject test.

8 years agoBuild: Reduce autogen.sh to a trivial stub
Daniel Elstner [Sun, 23 Aug 2015 19:17:36 +0000 (21:17 +0200)]
Build: Reduce autogen.sh to a trivial stub

Use autoreconf instead of invoking the various Auto tools
separately. Get rid of the Darwin-specific guesswork -- it
does not make sense to handle this at the level of libsigrok.

People should set up their ACLOCAL_PATH themselves as appropriate
for their own system; just as they already need to set up various
other paths.

8 years agoBuild: Check for accepted compiler flags
Daniel Elstner [Sun, 23 Aug 2015 19:08:54 +0000 (21:08 +0200)]
Build: Check for accepted compiler flags

Introduce the SR_CHECK_COMPILE_FLAGS macro and use it to check
for additional compiler flags. Put the accepted flags into the
separate substitution variable SR_EXTRA_CFLAGS.

With this and the preceding changes, bug #578 should now be fixed.

8 years agoBuild: Move package check macros to m4/sigrok.m4
Daniel Elstner [Sun, 23 Aug 2015 18:07:38 +0000 (20:07 +0200)]
Build: Move package check macros to m4/sigrok.m4

Define generic macros for pkg-config dependency checks etc. in
m4/sigrok.m4, so they can be easily reused by downstream modules.

8 years agoBuild: Make compiler warnings configurable
Daniel Elstner [Sun, 23 Aug 2015 13:02:30 +0000 (15:02 +0200)]
Build: Make compiler warnings configurable

Use the SR_ARG_ENABLE_WARNINGS macro to configure and check for
the availability of compiler warning flags. Maintain separate
sets of warning flags for C and C++.

The configure option --enable-warnings=[min|max|fatal|no] can
be used to set the compiler warning level for all languages.
The default level is "max".

8 years agoBuild: Move custom Autoconf macros to separate file
Daniel Elstner [Sun, 23 Aug 2015 12:52:00 +0000 (14:52 +0200)]
Build: Move custom Autoconf macros to separate file

Place custom Autoconf macros which other sigrok modules may
re-use into a separate file m4/sigrok.m4. Also, introduce new
macros for defining the package and library versions, and for
gathering compiler warning flags.

8 years agoClean up .gitignore
Daniel Elstner [Sat, 22 Aug 2015 18:27:53 +0000 (20:27 +0200)]
Clean up .gitignore

8 years agoBuild: Place Autoconf macros into m4/ directory
Daniel Elstner [Sat, 22 Aug 2015 18:58:11 +0000 (20:58 +0200)]
Build: Place Autoconf macros into m4/ directory

This separates the M4 macros from the other cruft in autostuff/
to prepare for the introduction of custom macro files.

8 years agoBuild: Do not omit first argument to m4_warn()
Daniel Elstner [Fri, 21 Aug 2015 22:07:48 +0000 (00:07 +0200)]
Build: Do not omit first argument to m4_warn()

It seems that contrary to what the documentation says, leaving
the category argument to m4_warn() empty is not allowed. Use
the "unsupported" category for lack of a better choice.

8 years agoFreeBSD: Also check for swig3.0.
Uwe Hermann [Wed, 19 Aug 2015 13:44:43 +0000 (15:44 +0200)]
FreeBSD: Also check for swig3.0.

On FreeBSD (for example) the SWIG binary can also be called "swig3.0".

This fixes (together with the recent "swig2.0" addition) bug #557.

8 years agoBuild: Get rid of recursive make invocations
Daniel Elstner [Tue, 18 Aug 2015 19:24:39 +0000 (21:24 +0200)]
Build: Get rid of recursive make invocations

Do not invoke $(MAKE) just to execute some clean rule.
This gets rid of the "jobserver unavailable" warnings.

8 years agoBuild: Tell setup.py where to find SWIG
Daniel Elstner [Tue, 18 Aug 2015 18:09:53 +0000 (20:09 +0200)]
Build: Tell setup.py where to find SWIG

8 years agoFreeBSD: Also check for swig2.0 in addition to swig.
Uwe Hermann [Tue, 18 Aug 2015 16:01:54 +0000 (18:01 +0200)]
FreeBSD: Also check for swig2.0 in addition to swig.

On FreeBSD (for example) the SWIG binary is called "swig2.0".

This fixes bug #557.

8 years agoFreeBSD: Fix compile/link error due to missing libusb_get_version().
Uwe Hermann [Tue, 18 Aug 2015 15:58:28 +0000 (17:58 +0200)]
FreeBSD: Fix compile/link error due to missing libusb_get_version().

The FreeBSD libusb-1.0 API implementation doesn't have libusb_get_version().
Use our CONF_LIBUSB_1_0_VERSION macro instead.

8 years agoBuild: Avoid $< in explicit rules
Daniel Elstner [Tue, 18 Aug 2015 16:06:08 +0000 (18:06 +0200)]
Build: Avoid $< in explicit rules

Although useful, makes other than GNU make do not like this.
Name the prerequisite explicitly instead, and circumvent any
VPATH substitution other makes may do.

8 years agoBuild: Skip C++ CFLAGS and LIBS check when disabled
Daniel Elstner [Tue, 18 Aug 2015 15:46:29 +0000 (17:46 +0200)]
Build: Skip C++ CFLAGS and LIBS check when disabled

It is not a good idea to invoke PKG_CHECK_MODULES with an empty
list of modules to check, so skip the invocation in that case.

8 years agoBuild: Fix typo that broke the driver enable options
Daniel Elstner [Tue, 18 Aug 2015 13:02:01 +0000 (15:02 +0200)]
Build: Fix typo that broke the driver enable options

8 years agoBuild: Make SR_DRIVER() handle dependencies
Daniel Elstner [Mon, 17 Aug 2015 03:08:48 +0000 (05:08 +0200)]
Build: Make SR_DRIVER() handle dependencies

Turn the hardware driver dependency handling on its head, by
making SR_DRIVER() take a list of dependencies to check for.

8 years agoBuild: Make dependency on git change more robust
Daniel Elstner [Mon, 17 Aug 2015 00:08:39 +0000 (02:08 +0200)]
Build: Make dependency on git change more robust

It can sometimes happen that .git/HEAD or .git/refs/head/*, which
are added as config.status dependencies during configure, do not
exist anymore at build time.  For instance, when the current branch
is deleted after switching to a different one.

Wrap the dependencies inside $(wildcard ...) to avoid this problem.
Note that this is a GNU make feature.  However, it should be fine
as it is only used for git builds.  Even if a non-GNU make is used,
the construct will hopefully just expand to nothing.

8 years agoBuild: Show glib/libzip version requirements in configure summary.
Uwe Hermann [Sun, 16 Aug 2015 21:56:43 +0000 (23:56 +0200)]
Build: Show glib/libzip version requirements in configure summary.

8 years agoBuild: Fix a typo causing libftdi1 detection issues.
Uwe Hermann [Sun, 16 Aug 2015 21:53:23 +0000 (23:53 +0200)]
Build: Fix a typo causing libftdi1 detection issues.

8 years agoBuild: Properly namespace the version defines
Daniel Elstner [Sun, 16 Aug 2015 16:53:32 +0000 (18:53 +0200)]
Build: Properly namespace the version defines

Use the CONF_ prefix for the configured host define as well as
the various build-time version macros.

8 years agoBuild: Cleanup up and modernize autotools setup
Daniel Elstner [Sat, 15 Aug 2015 13:59:52 +0000 (15:59 +0200)]
Build: Cleanup up and modernize autotools setup

Note that some --enable options of configure have become
--with or --without options.

8 years agoBuild: Simplify the hardware driver selection
Daniel Elstner [Sat, 15 Aug 2015 00:28:54 +0000 (02:28 +0200)]
Build: Simplify the hardware driver selection

Replace DRIVER() and DRIVER2() by a single SR_DRIVER() macro.
Derive the names of shell variables and preprocessor defines
programatically to cut down on repetition.

8 years agoAdd dist tarball to gitignore
Daniel Elstner [Sat, 15 Aug 2015 17:55:02 +0000 (19:55 +0200)]
Add dist tarball to gitignore

8 years agosetup.py: Do VPATH search for swig/enums.i
Daniel Elstner [Sat, 15 Aug 2015 17:43:29 +0000 (19:43 +0200)]
setup.py: Do VPATH search for swig/enums.i

Also, in swig/classes.i include swig/enums.i rather than plain
enums.i, to guard against future name clashes.

8 years agoBuild: Depend on swig/enums.i where appropriate
Daniel Elstner [Sat, 15 Aug 2015 17:18:36 +0000 (19:18 +0200)]
Build: Depend on swig/enums.i where appropriate

8 years agoenums.py: Place swig/enums.i into build directory
Daniel Elstner [Sat, 15 Aug 2015 17:02:15 +0000 (19:02 +0200)]
enums.py: Place swig/enums.i into build directory

8 years agoBuild: Do not distribute generated headers
Daniel Elstner [Sat, 15 Aug 2015 16:33:05 +0000 (18:33 +0200)]
Build: Do not distribute generated headers

version.h and enums.hpp are generated at build time and should
not be included in the distribution tarball.  This seems to have
triggered a weird error when doing a VPATH build of the dist
tarball.

8 years agoBuild: Pass nostdinc option to automake
Daniel Elstner [Sat, 15 Aug 2015 16:10:50 +0000 (18:10 +0200)]
Build: Pass nostdinc option to automake

It seems automake automatically adds the directory containing
the generated version.h to the include path.  Use nostdinc to
disable default includes altogether.

8 years agoBuild: Use angle brackets for enums.cpp include
Daniel Elstner [Sat, 15 Aug 2015 15:54:14 +0000 (17:54 +0200)]
Build: Use angle brackets for enums.cpp include

8 years agoBuild: Distribute bindings sources
Daniel Elstner [Sat, 15 Aug 2015 14:59:47 +0000 (16:59 +0200)]
Build: Distribute bindings sources

A couple of files referenced during the build were missing from
the distribution, causing a tarball build failure.

8 years agoBuild: Prefix directory to include of enums.hpp
Daniel Elstner [Sat, 15 Aug 2015 14:10:36 +0000 (16:10 +0200)]
Build: Prefix directory to include of enums.hpp

This fixes a build problem due to the reduced include search paths
introduced by my recent changes.  Also fix a couple of other
includes to use angle brackets.

8 years agoBuild: Append git revision hash to version
Daniel Elstner [Fri, 14 Aug 2015 19:01:59 +0000 (21:01 +0200)]
Build: Append git revision hash to version

Append the git revision hash to the libsigrok package version,
unless HEAD exactly matches a release tag.  Note that this does
not affect the version known to autoconf -- e.g. source tarballs
created by make dist will not receive a revision suffix.

Changes to git HEAD automatically trigger a reconfiguration.
Uncommitted changes do not, which is why I left out the -dirty
suffix.

8 years agoConsistently use SR_PACKAGE_VERSION instead of VERSION
Daniel Elstner [Wed, 12 Aug 2015 22:38:47 +0000 (00:38 +0200)]
Consistently use SR_PACKAGE_VERSION instead of VERSION

8 years agoBuild: Delay expansion of $datadir in FIRMWARE_DIR
Daniel Elstner [Fri, 14 Aug 2015 16:03:47 +0000 (18:03 +0200)]
Build: Delay expansion of $datadir in FIRMWARE_DIR

Make it so that $(datadir) is resolved at make time, as per
autotools recommendations.  Note that $datadir is not fully
resolved at configure time to begin with, i.e. part of it
already was evaluated at make time.

8 years agoBuild: Make version.h a configuration header
Daniel Elstner [Fri, 14 Aug 2015 15:53:06 +0000 (17:53 +0200)]
Build: Make version.h a configuration header

Use the proper tool for the job and make libsigrok/version.h
a secondary configuration header, so that autoconf's AC_DEFINE
machinery can be used to generate it.  Note that the header
template is still hand-written, enabling fine control of the
content.

8 years agotests: Include <libsigrok/libsigrok.h>
Daniel Elstner [Thu, 13 Aug 2015 00:06:41 +0000 (02:06 +0200)]
tests: Include <libsigrok/libsigrok.h>

This matches the global style and avoids the "../" path
components.

8 years agoBuild: Set local include directories in Makefile.am
Daniel Elstner [Wed, 12 Aug 2015 23:18:51 +0000 (01:18 +0200)]
Build: Set local include directories in Makefile.am

Move the include flags for files in the source tree from
configure.ac to Makefile.am where they belong.  Also use
AM_CPPFLAGS instead of CFLAGS/CXXFLAGS to make sure the
files in the build/source tree are always picked up first.

Also, remove the include/libsigrok sub-directory from the
search path, thereby making the <libsigrok/> prefix mandatory
when building libsigrok itself.  This matches the convention
already imposed on users of the library.

8 years agoIntroduce OutputFlag
Soeren Apel [Sat, 15 Aug 2015 21:40:29 +0000 (23:40 +0200)]
Introduce OutputFlag

8 years agooutput/srzip: Prevent memory leak in case filename is empty
Soeren Apel [Sat, 1 Aug 2015 19:41:32 +0000 (21:41 +0200)]
output/srzip: Prevent memory leak in case filename is empty

8 years agobackend: Output lib versions, host, and endianness in sr_init().
Uwe Hermann [Sat, 15 Aug 2015 19:06:30 +0000 (21:06 +0200)]
backend: Output lib versions, host, and endianness in sr_init().

This should help to more easily debug issues reported by users.

8 years agobrymen-bm86x: add some error checking
Aurelien Jacobs [Thu, 13 Aug 2015 23:03:57 +0000 (01:03 +0200)]
brymen-bm86x: add some error checking

8 years agoVarious minor cosmetics and consistency fixes.
Uwe Hermann [Sat, 15 Aug 2015 15:16:01 +0000 (17:16 +0200)]
Various minor cosmetics and consistency fixes.

8 years agoDrop comment mentioning non-existing function.
Uwe Hermann [Sat, 15 Aug 2015 15:09:29 +0000 (17:09 +0200)]
Drop comment mentioning non-existing function.

8 years agoDoxygen consistency fixes (@foo instead of \foo).
Uwe Hermann [Wed, 5 Aug 2015 15:23:54 +0000 (17:23 +0200)]
Doxygen consistency fixes (@foo instead of \foo).

8 years agoVarious errno/strerror() related fixes.
Uwe Hermann [Wed, 29 Jul 2015 17:31:43 +0000 (19:31 +0200)]
Various errno/strerror() related fixes.

 - Don't #include <errno.h> in files that don't actually need it.

 - Don't use strerror() on error codes from functions that don't set
   errno. Replace strerror() with sr_strerror() for libsigrok functions.

8 years agolibsigrok: demo: close pipes on dev_acquisition_stop()
Hubert CHAUMETTE [Fri, 31 Jul 2015 07:44:09 +0000 (09:44 +0200)]
libsigrok: demo: close pipes on dev_acquisition_stop()

Fixes a bug where new acquisition failes due to leftover pipes from
previous acquisitions:
sr: demo: dev_acquisition_start: pipe() failed
Indeed, PulseView had 2024 pipes opened. With this fix, it stabilizes at
33 with sampling thread active.

Signed-off-by: Hubert CHAUMETTE <redacted>
8 years agooutput/srzip: Use sr_output->filename instead of option
Soeren Apel [Thu, 30 Jul 2015 05:29:11 +0000 (07:29 +0200)]
output/srzip: Use sr_output->filename instead of option

This fixes parts of bug #570.

8 years agoAdd filename field to sr_output and make it accessible
Soeren Apel [Wed, 29 Jul 2015 19:19:49 +0000 (21:19 +0200)]
Add filename field to sr_output and make it accessible

This fixes parts of bug #570.

8 years agoOnly build BayLibre ACME driver if <sys/timerfd.h> is available.
Uwe Hermann [Sat, 25 Jul 2015 17:50:36 +0000 (19:50 +0200)]
Only build BayLibre ACME driver if <sys/timerfd.h> is available.

This fixes bug #610 (Android build issue).

8 years agools: Don't #include <libserialport.h> directly.
Uwe Hermann [Sat, 25 Jul 2015 14:59:04 +0000 (16:59 +0200)]
ols: Don't #include <libserialport.h> directly.

The driver is not using sp_* calls directly, so no need to #include
the header file.

8 years agoBump version to 0.4.0 (the upcoming next major release).
Uwe Hermann [Sat, 18 Jul 2015 16:03:29 +0000 (18:03 +0200)]
Bump version to 0.4.0 (the upcoming next major release).

8 years agoAppend "-git" to the version string.
Uwe Hermann [Sat, 18 Jul 2015 15:54:41 +0000 (17:54 +0200)]
Append "-git" to the version string.

This avoids confusion with the released tarballs.

8 years agoConvert a few more occurrences of d->priv to d->context.
Uwe Hermann [Thu, 9 Jul 2015 19:05:44 +0000 (21:05 +0200)]
Convert a few more occurrences of d->priv to d->context.

8 years agoFix #442 by renaming sr_dev_driver.priv to .context
Soeren Apel [Mon, 6 Jul 2015 19:55:36 +0000 (21:55 +0200)]
Fix #442 by renaming sr_dev_driver.priv to .context

8 years agofx2lafw: Use the USB_INTERFACE #define.
Uwe Hermann [Thu, 9 Jul 2015 23:03:16 +0000 (01:03 +0200)]
fx2lafw: Use the USB_INTERFACE #define.

8 years agofx2lafw: Fix segfault wrt libusb_detach_kernel_driver().
Uwe Hermann [Thu, 9 Jul 2015 22:59:48 +0000 (00:59 +0200)]
fx2lafw: Fix segfault wrt libusb_detach_kernel_driver().

Move the libusb_detach_kernel_driver() call after the code that
sets the usb->devhdl pointer, otherwise it'll be NULL and result
in a segfault.

  #0  libusb_kernel_driver_active (dev=0x0, interface_number=0) at libusb/core.c:1711
  #1  dev_open (sdi=0x12d99f0) at src/hardware/fx2lafw/api.c:374
  [...]

Tested on a device with the default Cypress VID/PID and one with
the Saleae Logic VID/PID, both works fine.

8 years agofx2lafw: Fix #445 by detaching the kernel driver (if any)
Soeren Apel [Mon, 6 Jul 2015 19:20:26 +0000 (21:20 +0200)]
fx2lafw: Fix #445 by detaching the kernel driver (if any)

This avoids the need to run "rmmod usbtest" on Linux for devices
with the standard Cypress FX2 USB VID/PID (04b4:8613).

8 years agoDemo: Fix #314 by always honoring sample limit changes
Daniel King [Mon, 6 Jul 2015 19:16:29 +0000 (21:16 +0200)]
Demo: Fix #314 by always honoring sample limit changes

8 years agoFix #505 by providing a separate property list for the analog group
Soeren Apel [Sun, 5 Jul 2015 19:06:05 +0000 (21:06 +0200)]
Fix #505 by providing a separate property list for the analog group

8 years agoFix #574 by setting up the transfer first, then starting acquisition
Soeren Apel [Sun, 5 Jul 2015 16:01:45 +0000 (18:01 +0200)]
Fix #574 by setting up the transfer first, then starting acquisition

8 years agobaylibre-acme: Remove unnecessary close().
Bartosz Golaszewski [Wed, 1 Jul 2015 10:30:09 +0000 (12:30 +0200)]
baylibre-acme: Remove unnecessary close().

The timerfd descriptor is closed automatically by
g_io_channel_shutdown(). No need to close it manually.

Signed-off-by: Bartosz Golaszewski <redacted>
8 years agobaylibre-acme: Use SR_HZ_TO_NS() when configuring the timerfd.
Bartosz Golaszewski [Wed, 1 Jul 2015 08:47:38 +0000 (10:47 +0200)]
baylibre-acme: Use SR_HZ_TO_NS() when configuring the timerfd.

Signed-off-by: Bartosz Golaszewski <redacted>
8 years agobaylibre-acme: Add a workaround for slow data acquisition.
Bartosz Golaszewski [Tue, 30 Jun 2015 13:07:50 +0000 (15:07 +0200)]
baylibre-acme: Add a workaround for slow data acquisition.

At high sampling rates and maximum channels we are not able to acquire
samples fast enough, even though frontends still think that samples
arrive on time. This causes visible shifts in frontend plots.

To compensate for the delay introduce the following workaround: check
if we are late (if any clock events have been missed) and resend the
last frame n times (n == number of missed clock events).

Signed-off-by: Bartosz Golaszewski <redacted>
8 years agobaylibre-acme: Use timerfd instead of a fake pipe.
Daniel Lezcano [Thu, 4 Jun 2015 17:01:27 +0000 (19:01 +0200)]
baylibre-acme: Use timerfd instead of a fake pipe.

Currently baylibre-acme uses a fake pipe as the input channel required by
libsigrok API and calls sleep() in the data acquisition callback to create
intervals between measurements.

Switch to a more elegant approach: use Linux' timerfd and set a periodic
timer equal to the sampling rate. Then read the data every time the timer
expires.

Signed-off-by: Daniel Lezcano <redacted>
Signed-off-by: Bartosz Golaszewski <redacted>
8 years agobaylibre-acme: Close sysfs files after stopping the acquisition.
Bartosz Golaszewski [Thu, 28 May 2015 13:45:56 +0000 (15:45 +0200)]
baylibre-acme: Close sysfs files after stopping the acquisition.

Add a missing call to dev_acquisition_close() in dev_acquisition_stop().

Signed-off-by: Bartosz Golaszewski <redacted>
8 years agobaylibre-acme: Optimize reading of values from sysfs.
Daniel Lezcano [Wed, 27 May 2015 18:22:02 +0000 (20:22 +0200)]
baylibre-acme: Optimize reading of values from sysfs.

Opening a file has a cost (security, allocation, syscalls). The
read_sample() function always does an open/read/close sequence.

In order to optimize that, let's open the file at the moment the
acquisition starts, close it when the acquisition stops and make
read_sample() only lseek() to the beginning of the file and read
the value.

Signed-off-by: Daniel Lezcano <redacted>
Signed-off-by: Bartosz Golaszewski <redacted>
8 years agosysclk-lwla: Do not create channels in reverse order
Daniel Elstner [Wed, 24 Jun 2015 17:16:49 +0000 (19:16 +0200)]
sysclk-lwla: Do not create channels in reverse order

This was originally done as an optimization in combination with a list
reversal which has since been removed from the code.  Thus, un-reverse
the channels so that the UI lists them in the correct order again.

8 years agoyokogawa-dlm: Minor cosmetics, consistency fixes, typos.
Uwe Hermann [Mon, 22 Jun 2015 22:39:59 +0000 (00:39 +0200)]
yokogawa-dlm: Minor cosmetics, consistency fixes, typos.

8 years agoyokogawa-dlm: Return correct value for SR_CONF_HDIV
Soeren Apel [Sun, 7 Jun 2015 09:52:30 +0000 (11:52 +0200)]
yokogawa-dlm: Return correct value for SR_CONF_HDIV

8 years agoyokogawa-dlm: Remove dlm_setup_channels()
Soeren Apel [Sun, 7 Jun 2015 09:51:40 +0000 (11:51 +0200)]
yokogawa-dlm: Remove dlm_setup_channels()

8 years agoyokogawa-dlm: Introduce config_channel_set()
Soeren Apel [Sun, 7 Jun 2015 09:50:29 +0000 (11:50 +0200)]
yokogawa-dlm: Introduce config_channel_set()

8 years agoyokogawa-dlm: Config get/set/list handler updates
Soeren Apel [Sun, 7 Jun 2015 09:40:39 +0000 (11:40 +0200)]
yokogawa-dlm: Config get/set/list handler updates

8 years agoyokogawa-dlm: Fix number of digital groups
Soeren Apel [Sun, 7 Jun 2015 09:37:27 +0000 (11:37 +0200)]
yokogawa-dlm: Fix number of digital groups

8 years agoyokogawa-dlm: Prevent duplicate channel index
Soeren Apel [Sun, 7 Jun 2015 09:36:38 +0000 (11:36 +0200)]
yokogawa-dlm: Prevent duplicate channel index

8 years agoyokogawa-dlm: Add support for the DL9000 series
Soeren Apel [Sun, 7 Jun 2015 09:35:21 +0000 (11:35 +0200)]
yokogawa-dlm: Add support for the DL9000 series

8 years agofx2lafw: Add DSLogic Pro and DSCope firmware loading support
Daniel Gröber [Sat, 2 May 2015 22:31:49 +0000 (00:31 +0200)]
fx2lafw: Add DSLogic Pro and DSCope firmware loading support

8 years agodslogic: Fix incorrect samplerate setting.
Carl-Fredrik Sundström [Mon, 22 Jun 2015 14:53:51 +0000 (16:53 +0200)]
dslogic: Fix incorrect samplerate setting.

8 years agobindings/python: Handle import failures without crashing
Alexandru Gagniuc [Wed, 6 May 2015 07:32:30 +0000 (00:32 -0700)]
bindings/python: Handle import failures without crashing

When the import of gi.repository.GLib failed, we would get a NULL
pointer that we passed along without any checking. In this situation,
the entire program would crash with a segmentation fault, and no
message to indicate the problem.

When the import fails, abort the SWIG init and print a message. The
Python interpreter then prints out a backtrace, which can be useful
in tracking down the problem.

8 years agoscpi-pps: Chroma 62000P: Get capabilities from device name
Alexandru Gagniuc [Sun, 17 May 2015 00:30:47 +0000 (17:30 -0700)]
scpi-pps: Chroma 62000P: Get capabilities from device name

The Chroma 62000P series comes in various models with different
current and voltage capabilities. These are encoded in the *IDN
string, so just get them from there, rather than needing a profile
for every model.

8 years agoscpi-pps: Add profile for Chroma 62024P-80-60 DC Source
Alexandru Gagniuc [Sat, 9 May 2015 03:54:08 +0000 (20:54 -0700)]
scpi-pps: Add profile for Chroma 62024P-80-60 DC Source