} srd_Decoder;
/* This is only used for nicer srd_dbg() output. */
-static const char *output_type_name(unsigned int idx)
+SRD_PRIV const char *output_type_name(unsigned int idx)
{
- static const char names[][16] = {
+ static const char *names[] = {
"OUTPUT_ANN",
"OUTPUT_PYTHON",
"OUTPUT_BINARY",
+ "OUTPUT_LOGIC",
"OUTPUT_META",
"(invalid)"
};
return;
if (pda->ann_text)
g_strfreev(pda->ann_text);
- g_free(pda);
}
static int convert_annotation(struct srd_decoder_inst *di, PyObject *obj,
goto err;
}
- pda = g_malloc(sizeof(struct srd_proto_data_annotation));
+ pda = pdata->data;
pda->ann_class = ann_class;
pda->ann_text = ann_text;
- pdata->data = pda;
PyGILState_Release(gstate);
return SRD_ERR_PYTHON;
}
+static void release_logic(struct srd_proto_data_logic *pdl)
+{
+ if (!pdl)
+ return;
+ g_free((void *)pdl->data);
+}
+
+static int convert_logic(struct srd_decoder_inst *di, PyObject *obj,
+ struct srd_proto_data *pdata)
+{
+ struct srd_proto_data_logic *pdl;
+ PyObject *py_tmp;
+ Py_ssize_t size;
+ int logic_group;
+ char *group_name, *buf;
+ PyGILState_STATE gstate;
+
+ gstate = PyGILState_Ensure();
+
+ /* Should be a list of [logic group, bytes]. */
+ if (!PyList_Check(obj)) {
+ srd_err("Protocol decoder %s submitted non-list for SRD_OUTPUT_LOGIC.",
+ di->decoder->name);
+ goto err;
+ }
+
+ /* Should have 2 elements. */
+ if (PyList_Size(obj) != 2) {
+ srd_err("Protocol decoder %s submitted SRD_OUTPUT_LOGIC list "
+ "with %zd elements instead of 2", di->decoder->name,
+ PyList_Size(obj));
+ goto err;
+ }
+
+ /* The first element should be an integer. */
+ py_tmp = PyList_GetItem(obj, 0);
+ if (!PyLong_Check(py_tmp)) {
+ srd_err("Protocol decoder %s submitted SRD_OUTPUT_LOGIC list, "
+ "but first element was not an integer.", di->decoder->name);
+ goto err;
+ }
+ logic_group = PyLong_AsLong(py_tmp);
+ if (!(group_name = g_slist_nth_data(di->decoder->logic_output_channels, logic_group))) {
+ srd_err("Protocol decoder %s submitted SRD_OUTPUT_LOGIC with "
+ "unregistered logic group %d.", di->decoder->name, logic_group);
+ goto err;
+ }
+
+ /* Second element should be bytes. */
+ py_tmp = PyList_GetItem(obj, 1);
+ if (!PyBytes_Check(py_tmp)) {
+ srd_err("Protocol decoder %s submitted SRD_OUTPUT_LOGIC list, "
+ "but second element was not bytes.", di->decoder->name);
+ goto err;
+ }
+
+ /* Consider an empty set of bytes a bug. */
+ if (PyBytes_Size(py_tmp) == 0) {
+ srd_err("Protocol decoder %s submitted SRD_OUTPUT_LOGIC "
+ "with empty data set.", di->decoder->name);
+ goto err;
+ }
+
+ if (PyBytes_AsStringAndSize(py_tmp, &buf, &size) == -1)
+ goto err;
+
+ PyGILState_Release(gstate);
+
+ pdl = pdata->data;
+ pdl->logic_group = logic_group;
+ /* pdl->repeat_count is set by the caller as it depends on the sample range */
+ if (!(pdl->data = g_try_malloc(size)))
+ return SRD_ERR_MALLOC;
+ memcpy((void *)pdl->data, (const void *)buf, size);
+
+ return SRD_OK;
+
+err:
+ PyGILState_Release(gstate);
+
+ return SRD_ERR_PYTHON;
+}
+
static void release_binary(struct srd_proto_data_binary *pdb)
{
if (!pdb)
return;
g_free((void *)pdb->data);
- g_free(pdb);
}
static int convert_binary(struct srd_decoder_inst *di, PyObject *obj,
PyGILState_Release(gstate);
- pdb = g_malloc(sizeof(struct srd_proto_data_binary));
+ pdb = pdata->data;
pdb->bin_class = bin_class;
pdb->size = size;
- if (!(pdb->data = g_try_malloc(pdb->size))) {
- g_free(pdb);
+ if (!(pdb->data = g_try_malloc(pdb->size)))
return SRD_ERR_MALLOC;
- }
memcpy((void *)pdb->data, (const void *)buf, pdb->size);
- pdata->data = pdb;
return SRD_OK;
return SRD_ERR_PYTHON;
}
-static struct srd_decoder_inst *srd_sess_inst_find_by_obj(
+static inline struct srd_decoder_inst *srd_sess_inst_find_by_obj(
struct srd_session *sess, const GSList *stack, const PyObject *obj)
{
const GSList *l;
*
* @since 0.1.0
*/
-static struct srd_decoder_inst *srd_inst_find_by_obj(
+static inline struct srd_decoder_inst *srd_inst_find_by_obj(
const GSList *stack, const PyObject *obj)
{
struct srd_decoder_inst *di;
struct srd_session *sess;
GSList *l;
+ /* Performance shortcut: Handle the most common case first. */
+ sess = sessions->data;
+ di = sess->di_list->data;
+ if (di->py_inst == obj)
+ return di;
+
di = NULL;
for (l = sessions; di == NULL && l != NULL; l = l->next) {
sess = l->data;
g_variant_unref(gvar);
}
+PyDoc_STRVAR(Decoder_put_doc,
+ "Put an annotation for the specified span of samples.\n"
+ "\n"
+ "Arguments: start and end sample number, stream id, annotation data.\n"
+ "Annotation data's layout depends on the output stream type."
+);
+
static PyObject *Decoder_put(PyObject *self, PyObject *args)
{
GSList *l;
struct srd_decoder_inst *di, *next_di;
struct srd_pd_output *pdo;
struct srd_proto_data pdata;
+ struct srd_proto_data_annotation pda;
+ struct srd_proto_data_binary pdb;
+ struct srd_proto_data_logic pdl;
uint64_t start_sample, end_sample;
int output_id;
struct srd_pd_callback *cb;
}
pdo = l->data;
- srd_spew("Instance %s put %" PRIu64 "-%" PRIu64 " %s on oid %d.",
- di->inst_id, start_sample, end_sample,
- output_type_name(pdo->output_type), output_id);
+ /* Upon SRD_OUTPUT_PYTHON for stacked PDs, we have a nicer log message later. */
+ if (pdo->output_type != SRD_OUTPUT_PYTHON && di->next_di != NULL) {
+ srd_spew("Instance %s put %" PRIu64 "-%" PRIu64 " %s on "
+ "oid %d (%s).", di->inst_id, start_sample, end_sample,
+ output_type_name(pdo->output_type), output_id,
+ pdo->proto_id);
+ }
pdata.start_sample = start_sample;
pdata.end_sample = end_sample;
case SRD_OUTPUT_ANN:
/* Annotations are only fed to callbacks. */
if ((cb = srd_pd_output_callback_find(di->sess, pdo->output_type))) {
+ pdata.data = &pda;
/* Convert from PyDict to srd_proto_data_annotation. */
if (convert_annotation(di, py_data, &pdata) != SRD_OK) {
/* An error was already logged. */
case SRD_OUTPUT_PYTHON:
for (l = di->next_di; l; l = l->next) {
next_di = l->data;
- srd_spew("Sending %" PRIu64 "-%" PRIu64 " to instance %s",
- start_sample, end_sample, next_di->inst_id);
+ srd_spew("Instance %s put %" PRIu64 "-%" PRIu64 " %s "
+ "on oid %d (%s) to instance %s.", di->inst_id,
+ start_sample,
+ end_sample, output_type_name(pdo->output_type),
+ output_id, pdo->proto_id, next_di->inst_id);
if (!(py_res = PyObject_CallMethod(
next_di->py_inst, "decode", "KKO", start_sample,
end_sample, py_data))) {
break;
case SRD_OUTPUT_BINARY:
if ((cb = srd_pd_output_callback_find(di->sess, pdo->output_type))) {
+ pdata.data = &pdb;
/* Convert from PyDict to srd_proto_data_binary. */
if (convert_binary(di, py_data, &pdata) != SRD_OK) {
/* An error was already logged. */
release_binary(pdata.data);
}
break;
+ case SRD_OUTPUT_LOGIC:
+ if ((cb = srd_pd_output_callback_find(di->sess, pdo->output_type))) {
+ pdata.data = &pdl;
+ /* Convert from PyDict to srd_proto_data_logic. */
+ if (convert_logic(di, py_data, &pdata) != SRD_OK) {
+ /* An error was already logged. */
+ break;
+ }
+ if (end_sample <= start_sample) {
+ srd_err("Ignored SRD_OUTPUT_LOGIC with invalid sample range.");
+ break;
+ }
+ pdl.repeat_count = (end_sample - start_sample) - 1;
+ Py_BEGIN_ALLOW_THREADS
+ cb->cb(&pdata, cb->cb_data);
+ Py_END_ALLOW_THREADS
+ release_logic(pdata.data);
+ }
+ break;
case SRD_OUTPUT_META:
if ((cb = srd_pd_output_callback_find(di->sess, pdo->output_type))) {
/* Annotations need converting from PyObject. */
return NULL;
}
-static PyObject *Decoder_register(PyObject *self, PyObject *args,
- PyObject *kwargs)
+PyDoc_STRVAR(Decoder_register_doc,
+ "Register a new output stream."
+);
+
+static PyObject *Decoder_register(PyObject *self,
+ PyObject *args, PyObject *kwargs)
{
struct srd_decoder_inst *di;
struct srd_pd_output *pdo;
return py_new_output_id;
}
- srd_dbg("Instance %s creating new output type %d for %s.",
- di->inst_id, output_type, proto_id);
-
pdo = g_malloc(sizeof(struct srd_pd_output));
/* pdo_id is just a simple index, nothing is deleted from this list anyway. */
PyGILState_Release(gstate);
+ srd_dbg("Instance %s creating new output type %s as oid %d (%s).",
+ di->inst_id, output_type_name(output_type), pdo->pdo_id,
+ proto_id);
+
return py_new_output_id;
err:
/* A channelmap value of -1 means "unused optional channel". */
if (di->dec_channelmap[i] == -1) {
/* Value of unused channel is 0xff, instead of 0 or 1. */
- PyTuple_SetItem(py_pinvalues, i, PyLong_FromLong(0xff));
+ PyTuple_SetItem(py_pinvalues, i, PyLong_FromUnsignedLong(0xff));
} else {
sample_pos = di->inbuf + ((di->abs_cur_samplenum - di->abs_start_samplenum) * di->data_unitsize);
byte_offset = di->dec_channelmap[i] / 8;
bit_offset = di->dec_channelmap[i] % 8;
sample = *(sample_pos + byte_offset) & (1 << bit_offset) ? 1 : 0;
- PyTuple_SetItem(py_pinvalues, i, PyLong_FromLong(sample));
+ PyTuple_SetItem(py_pinvalues, i, PyLong_FromUnsignedLong(sample));
}
}
*
* If there are no terms in the condition, 'term_list' will be NULL.
*
+ * @param di The decoder instance to use. Must not be NULL.
* @param py_dict A Python dict containing terms. Must not be NULL.
* @param term_list Pointer to a GSList which will be set to the newly
* created list of terms. Must not be NULL.
*
* @return SRD_OK upon success, a negative error code otherwise.
*/
-static int create_term_list(PyObject *py_dict, GSList **term_list)
+static int create_term_list(struct srd_decoder_inst *di,
+ PyObject *py_dict, GSList **term_list)
{
Py_ssize_t pos = 0;
PyObject *py_key, *py_value;
struct srd_term *term;
- uint64_t num_samples_to_skip;
+ int64_t num_samples_to_skip;
char *term_str;
PyGILState_STATE gstate;
/* Check whether the current key is a string or a number. */
if (PyLong_Check(py_key)) {
/* The key is a number. */
- /* TODO: Check if the number is a valid channel. */
/* Get the value string. */
if ((py_pydictitem_as_str(py_dict, py_key, &term_str)) != SRD_OK) {
srd_err("Failed to get the value.");
term = g_malloc(sizeof(struct srd_term));
term->type = get_term_type(term_str);
term->channel = PyLong_AsLong(py_key);
+ if (term->channel < 0 || term->channel >= di->dec_num_channels)
+ term->type = SRD_TERM_ALWAYS_FALSE;
g_free(term_str);
} else if (PyUnicode_Check(py_key)) {
/* The key is a string. */
- /* TODO: Check if it's "skip". */
+ /* TODO: Check if the key is "skip". */
if ((py_pydictitem_as_long(py_dict, py_key, &num_samples_to_skip)) != SRD_OK) {
srd_err("Failed to get number of samples to skip.");
goto err;
term->type = SRD_TERM_SKIP;
term->num_samples_to_skip = num_samples_to_skip;
term->num_samples_already_skipped = 0;
+ if (num_samples_to_skip < 0)
+ term->type = SRD_TERM_ALWAYS_FALSE;
} else {
srd_err("Term key is neither a string nor a number.");
goto err;
num_conditions = PyList_Size(py_conditionlist);
if (num_conditions == 0)
goto ret_9999; /* The PD invoked self.wait([]). */
- Py_IncRef(py_conditionlist);
+ Py_INCREF(py_conditionlist);
} else if (PyDict_Check(py_conds)) {
/* 'py_conds' is a dict. */
if (PyDict_Size(py_conds) == 0)
goto ret_9999; /* The PD invoked self.wait({}). */
/* Make a list and put the dict in there for convenience. */
py_conditionlist = PyList_New(1);
- Py_IncRef(py_conds);
+ Py_INCREF(py_conds);
PyList_SetItem(py_conditionlist, 0, py_conds);
num_conditions = 1;
} else {
}
/* Create the list of terms in this condition. */
- if ((ret = create_term_list(py_dict, &term_list)) < 0)
+ if ((ret = create_term_list(di, py_dict, &term_list)) < 0)
break;
/* Add the new condition to the PD instance's condition list. */
return SRD_OK;
}
+PyDoc_STRVAR(Decoder_wait_doc,
+ "Wait for one or more conditions to occur.\n"
+ "\n"
+ "Returns the sample data at the next position where the condition\n"
+ "is seen. When the optional condition is missing or empty, the next\n"
+ "sample number is used. The condition can be a dictionary with one\n"
+ "condition's details, or a list of dictionaries specifying multiple\n"
+ "conditions of which at least one condition must be true. Dicts can\n"
+ "contain one or more key/value pairs, all of which must be true for\n"
+ "the dict's condition to be considered true. The key either is a\n"
+ "channel index or a keyword, the value is the operation's parameter.\n"
+ "\n"
+ "Supported parameters for channel number keys: 'h', 'l', 'r', 'f',\n"
+ "or 'e' for level or edge conditions. Other supported keywords:\n"
+ "'skip' to advance over the given number of samples.\n"
+);
+
static PyObject *Decoder_wait(PyObject *self, PyObject *args)
{
int ret;
unsigned int i;
gboolean found_match;
struct srd_decoder_inst *di;
- PyObject *py_pinvalues, *py_matched;
+ PyObject *py_pinvalues, *py_matched, *py_samplenum;
PyGILState_STATE gstate;
if (!self || !args)
/* If there's a match, set self.samplenum etc. and return. */
if (found_match) {
/* Set self.samplenum to the (absolute) sample number that matched. */
- PyObject_SetAttrString(di->py_inst, "samplenum",
- PyLong_FromLong(di->abs_cur_samplenum));
+ py_samplenum = PyLong_FromUnsignedLongLong(di->abs_cur_samplenum);
+ PyObject_SetAttrString(di->py_inst, "samplenum", py_samplenum);
+ Py_DECREF(py_samplenum);
if (di->match_array && di->match_array->len > 0) {
py_matched = PyTuple_New(di->match_array->len);
for (i = 0; i < di->match_array->len; i++)
PyTuple_SetItem(py_matched, i, PyBool_FromLong(di->match_array->data[i]));
PyObject_SetAttrString(di->py_inst, "matched", py_matched);
+ Py_DECREF(py_matched);
match_array_free(di);
} else {
PyObject_SetAttrString(di->py_inst, "matched", Py_None);
/* Signal the main thread that we handled all samples. */
g_cond_signal(&di->handled_all_samples_cond);
+ /*
+ * When EOF was provided externally, communicate the
+ * Python EOFError exception to .decode() and return
+ * from the .wait() method call. This is motivated by
+ * the use of Python context managers, so that .decode()
+ * methods can "close" incompletely accumulated data
+ * when the sample data is exhausted.
+ */
+ if (di->communicate_eof) {
+ srd_dbg("%s: %s: Raising EOF from wait().",
+ di->inst_id, __func__);
+ g_mutex_unlock(&di->data_mutex);
+ PyErr_SetString(PyExc_EOFError, "samples exhausted");
+ goto err;
+ }
+
/*
* When termination of wait() and decode() was requested,
* then exit the loop after releasing the mutex.
return NULL;
}
+PyDoc_STRVAR(Decoder_has_channel_doc,
+ "Check whether input data is supplied for a given channel.\n"
+ "\n"
+ "Argument: A channel index.\n"
+ "Returns: A boolean, True if the channel is connected,\n"
+ "False if the channel is open (won't see any input data).\n"
+);
+
/**
* Return whether the specified channel was supplied to the decoder.
*
int idx, count;
struct srd_decoder_inst *di;
PyGILState_STATE gstate;
+ PyObject *bool_ret;
if (!self || !args)
return NULL;
PyGILState_Release(gstate);
- return (di->dec_channelmap[idx] == -1) ? Py_False : Py_True;
+ bool_ret = (di->dec_channelmap[idx] == -1) ? Py_False : Py_True;
+ Py_INCREF(bool_ret);
+ return bool_ret;
err:
PyGILState_Release(gstate);
return NULL;
}
+PyDoc_STRVAR(Decoder_doc, "sigrok Decoder base class");
+
static PyMethodDef Decoder_methods[] = {
- { "put", Decoder_put, METH_VARARGS,
- "Accepts a dictionary with the following keys: startsample, endsample, data" },
- { "register", (PyCFunction)Decoder_register, METH_VARARGS|METH_KEYWORDS,
- "Register a new output stream" },
- { "wait", Decoder_wait, METH_VARARGS,
- "Wait for one or more conditions to occur" },
- { "has_channel", Decoder_has_channel, METH_VARARGS,
- "Report whether a channel was supplied" },
- {NULL, NULL, 0, NULL}
+ { "put",
+ Decoder_put, METH_VARARGS,
+ Decoder_put_doc,
+ },
+ { "register",
+ (PyCFunction)(void(*)(void))Decoder_register, METH_VARARGS | METH_KEYWORDS,
+ Decoder_register_doc,
+ },
+ { "wait",
+ Decoder_wait, METH_VARARGS,
+ Decoder_wait_doc,
+ },
+ { "has_channel",
+ Decoder_has_channel, METH_VARARGS,
+ Decoder_has_channel_doc,
+ },
+ ALL_ZERO,
};
/**
{
PyType_Spec spec;
PyType_Slot slots[] = {
- { Py_tp_doc, "sigrok Decoder base class" },
+ { Py_tp_doc, Decoder_doc },
{ Py_tp_methods, Decoder_methods },
{ Py_tp_new, (void *)&PyType_GenericNew },
- { 0, NULL }
+ ALL_ZERO,
};
PyObject *py_obj;
PyGILState_STATE gstate;