With the sigrok session running in a worker thread, if sr_session_stop is called
from another thread, it shuts down the pollfds used by the hardware drivers,
without ensuring that the sigrok event loop is no longer using those pollfds.
On the demo driver, this involves shutting down the GIOChannels, causing a
segfault when the sigrok event loop tries to use them. This is evident when
using the Stop button in PulseView, while the session is running.
This isn't a problem with just the demo driver; any driver's resources may be
freed by sr_session_stop concurrently with the sigrok session running.
To solve this problem, we don't touch the session itself in sr_session_stop().
Instead, we mark it for decommissioning and return. The session polls this flag,
and shuts itself down when requested.
This fixes bug 4.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Number of timebases is needed to properly interpret a frame's worth
of samples. Number of vertical divisions isn't, but may nevertheless
be interesting for a frontend that wants to reproduce the scope's
native display.
sr_config_get() of SR_CONF_SAMPLERATE now returns a GVARIANT_TYPE_DICTIONARY.
This dictionary contains a single key-value pair. Possible values for the key:
- "samplerates": the value is an array of GVARIANT_TYPE_UINT64 representing
all valid samplerates.
- "samplerate-steps": the value is an array of GVARIANT_TYPE_UINT64 with
exactly three members, which represent the lowest samplerate, highest
samplerate, and the minimum step, respectively.
sr_config_get() provides a GVariant owned by the caller, so it must be
released with g_variant_unref() when done.
sr_config_set() takes a GVariant from the caller which may be floating;
it will be properly sunk and release after use by this function. Thus
the output of g_variant_new_*() may be used as an argument.
sr_config_list() also provides a GVariant owned by the caller, to be
unreferenced when done.
sr_config_make() can take a floating reference.
Expresses the time between samples, in milliseconds. This can be used
for devices with a samplerate > 1 second, such as dataloggers, which
cannot be expressed with SR_CONF_SAMPLERATE.
This is essential if a format contains e.g. the number of probes; the
init() function needs to initialize the sr_dev_inst struct, but needs
access to the file to properly add the probes to it.
This takes an sr_config key and returns a list of possible values for
that key to be submitted with config_set(). The format of the list and
its contents is dependent on the key.
This will replace the SR_DI_* keys that returned such a list before,
such as SR_DI_SAMPLERATES.
This patch marks packet structures and their payloads as const.
This indicates to packet receivers that modifications to these are
not allowed. In general all pointers should be marked const unless
modification of the referenced data is explicitly allowed.
Having concentration as a unit is vague, as it can be expressed in
many ways. In the context of sigrok, concentration means a normalized
number from 0 to 1.
Document its meaning.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Not all drivers use subdrivers. The only reason the subdriver field was
introduced was to accomodate the model of serial-dmm.
The sr_dev_driver struct is available to the frontend. Exposing the subdriver
field creates the problem of exposing knowledge of libsigrok's internal driver
layout, even though the drivers are designed to be a flat list to the frontend.
Store the subdriver in the dev_context struct of serial-dmm.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Store/use the receive_data() function and a pointer to the driver struct
in the dmms[] array. Use a ".subdriver" entry in the driver struct.
Use a macro to simplify hw_init() wrappers.
In the case of USB drivers, a driver's dev_acquisition_stop() cannot
simply remove its fd sources from the session and close its devices:
a USB transfer might still be underway, and it needs to be finished
(and its memory freed) properly.
An sr_dev_inst->status value is added: SR_ST_STOPPING, which should
be set when the driver's dev_acquisition_stop() is called, and acts
as a marker for the USB event handler to wind up its operations.
In order for dev_acquisition_stop() to be able to set the sdi status,
however, it needs to be unconstified.
Whenever we added a field to the enums in libsigrok.h, we were running
the risk of breaking ABI compatibility. Any reordering of existing values
or the addition of entries other than at the bottom of the list would
cause an ABI breakage.
By hardcoding the values we ensure that any added field will have an
unused value, and will not take a value previously used by a different
flag. By doing this, we avoid confusing frontends compiled with an
earlier API.
We use 10000 as the first entry of each enum, and each "category" gets
an explicit number (10000, 20000, 30000, and so on).
Also avoid making the first value of an enum zero. A value of zero is
used as terminator item in some lists.
Remove explicit "DUMMY" (terminator) enum entries, they're not needed.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
- Turn SR_OK/SR_ERR_* and SR_LOG_* into proper enums.
- Use /**< Foo. */ for Doxygen comments that are on the same line as
the (e.g.) enum entry they document. If the comment is not on the
same line (but rather directly above the enum entry) a simple
/** Foo. */ comment is sufficient.
- Use /** */ instead of /* */ in some places, so that Doxygen actually uses
the respective comments.
- Various smaller cosmetic fixes or cleanups.
The Radioshack 22-812 can measure parameters such as transistor gain,
dbm, etc. Add those to libsigrok.h.
The SR_UNIT_UNITLESS is for quantities that do not have units. Any
ratio or gain are just factors, whic do not have units. Specifically,
a transistor's gain, or hFE, is a unitless quantity.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
The new output module callbacks will be init, recv and cleanup. The
existing data and event callbacks still work, but will be phased out
as existing modules get converted.
The recv() callback gets a copy of every packet on the session bus,
and thus has visibility of all metadata, allowing it to properly
output any acquired data.
These are used to list the device instances currently known to the driver,
and clear that list.
Drivers that don't necessarily clear their list of instances on every scan,
such as genericdmm, need to provide these to the frontend to keep instance
management sane.
Since probes now live in a struct sr_dev_inst owned by the driver, it
already knows about them. Instead of a frontend telling the driver to
configure probes, all driver now do this just before starting acquisition.
It's up to the driver to keep its GSList of struct sr_dev_inst * in there.
Anything else the driver wants to keep driver-global should also go in
there, such as libusb_context.
It's obsolete: no frontend ever used it, and neither did libsigrok.
The sdi->status field is only used internally by some drivers, and
should probably be moved to the driver-specific context structs.
This will replace dev_info_get(), and will be used to fetch both
driver and device instance-specific information. The sr_dev_inst
argument is NULL in case of a driver info fetch. In line with the
libsigrok wrapper, this function returns an error code, using the
supplied void ** to return the requested data.
There is no point in libsigrok copying probe lists around. The driver now
builds a list of probes according to the model device it found, and will
make that available to a frontend. The frontend thus has a reference of
what the driver has, including default names, and doesn't need libsigrok
to provide an unnecessary level of abstraction.
The sr_probe_new() library-private function is a helper for drivers.
init() now only does whatever administrative stuff it needs (typically not
much), and returns an error code.
scan() can be called multiple times during the life of an application, and
returns a GSList of struct sr_dev_inst * of devices found during that scan.
The instances are a copy of the ones stored in the driver's own instance
list, to be freed by the caller with g_slist_free() only.
The scan() call can be passed a GSList of struct sr_hwopt *, to direct the
scanning.
SR_HWOPT_* entries are driver options, not device instance parameters, so
they will never be mixed together.
Also, driver options are always passed in a GSList, where the data field
is a struct sr_hwopt.
This changes the semantics of the init() call as well. That now only
initializes the driver -- an administrative affair, no hardware gets
touched during this call. It returns a standard SR_OK or SR_ERR* code.
The scan() call does a discovery run for devices it knows, and returns
the number found. It can be called at any time.
It was actually used in one way: the session file loaded abused it for
passing in the filename -- something it definitely wasn't intended for.
This now uses the proper way to pass arguments to a driver: the new
SR_HWCAP_SESSIONFILE.
The OLS driver could also use it as an indication of the serial port to
use instead of actively probing all serial ports on the system, but there
wasn't any frontend code that passed in such a parameter, making it
entirely useless. That will soon be handled differently with the new
scan() API call, regardless.
All frontends will have to include <libsigrok/libsigrok.h> from now on.
This header includes proto.h and version.h, both installed from the
distribution into $INCLUDE/libsigrok/ as well.
The only dynamically changed header is now version.h, which has both
libsigrok and libtool compile-time versions in it.
Bert Vermeulen
Uwe Hermann
Martin Ling
Alexandru Gagniuc
Joel Holdsworth
Peter Stuge
Lars-Peter Clausen