| Panel | ||||||||
|---|---|---|---|---|---|---|---|---|
| ||||||||
This page is a work in progress. Improvements to content and formatting are underway. |
Overview
As shown below, there are 3 main components of an Audio Weaver system that require communication in order to tune an audio system. The Designer GUI will generate text based tuning commands (AWS) based on the layout and any user tuning commands. These commands are sent to AWE Server, which in turn compiles the AWS commands into binary commands (AWB) that can be sent to the target running the AWE Core. The target processes the AWB commands and then generates a reply message that is sent back to the AWE Server.
...
| Expand | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
...
If the server is connected to a target, the file_name argument is not permitted. If the file is not found as specified, and does not have an absolute path, it is searched for in the audio path. See set_path.
audio_stop (BSP)
Syntax
audio_stop
...
When several device are merged for audio input (numbered 0 - N-1) close one of these devices by its index, causing zero samples to be generated in its place. This command is a specialized command for users that need to disconnect from a device that is about to be reconfigured. See reopen_input.
close_output (BSP)
Syntax:
...
When several device are merged for audio output (numbered 0 - N-1) close one of these devices by its index. Samples being written to that device are discarded. This command is a specialized command for users that need to disconnect from a device that is about to be reconfigured. See reopen_output,
cmd
Syntax:
cmd,opcode [,arg1, ... ,argN]
...
destroy_dialog,dialog
where:
dialog must be a dialog created by create_dialog,
...
dialog_state,dialog,state
where:
dialog must be a dialog created by create_dialog,
state must be 0 or 1
This command sets the given dialog to its initial size if zero, otherwise to its alternate size. If the alternate size dimensions given to create_dialog were zero or the same as the initial size, the command has no effect.
...
Terminates logging of binary commands and writes the file. This command works in conjunction with make_binary.
exists_dialog (BSP, GUI only)
...
Checks if a dialog with the specified name already exists. The function returns either:
success,0 (Dialog does not exist)
success,1 (Dialog does exist)
failed, argument count (Incorrect number of arguments to the function)
exit (BSP)
Syntax:
exit
This command destroys the server.
...
Plays the input file through the layout as fast as possible, optionally recording the layout output in a WAV file. The output file will be at whatever rate the layout specifies, and have has as many channels as the output wire.
...
Reads the specified number of float elements in an array of data and returns the result as binary data rather than as text. This command is only supported through the MATLAB AWEClient.dll and is not for general use. The format of the binary reply packet is described in Internal Binary Packets.
fast_read_int (Matlab)
Syntax:
...
Reads the specified number of integer elements in an array of data and returns the result as binary data rather than as text. This command is only supported through the MATLAB AWEClient.dll and is not for general use. The format of the binary reply packet is described in section 6 Internal Binary Packets.
fast_write (Matlab)
Syntax:
...
Writes the float array passed by Matlab to the target layout starting at the address given. This command is only supported through the MATLAB AWEClient.dll and is not for general use. It causes the AWE server to receive a binary packet containing the Matlab array values as documented in section 6 Internal Binary Packets.
fast_write_int (Matlab)
Syntax:
...
Writes the integer array passed by Matlab to the target layout starting at the address given. This command is only supported through the MATLAB AWEClient.dll and is not for general use. It causes the AWE server to receive a binary packet containing the Matlab array values as documented in section 6 Internal Binary Packets.
fast_write_partial (Matlab)
...
Writes the float array passed by Matlab to the target layout starting at the address given. This command is only supported through the MATLAB AWEClient.dll and is not for general use. It causes the AWE server to receive a binary packet containing the Matlab array values as documented in section 6 Internal Binary Packets.
The final set call performed by fast_write is suppressed.
...
Writes the integer array passed by Matlab to the target layout starting at the address given. This command is only supported through the MATLAB AWEClient.dll and is not for general use. It causes the AWE server to receive a binary packet containing the Matlab array values as documented in section 6 Internal Binary Packets.
The final set call performed by fast_write_int is suppressed.
file_exists (BSP)
...
This command either succeeds or fails. The possible failures are:
failed,argument count // command takes only one argument
failed,no target // command requires a connected target
If it succeeds the possible replies:
success,0, // name not found in path
success,1,full_path // found this file in the path
success,2,full_path // found this directory in the path
If found, the full_path will be in the convention of the OS on which the server is running. The only supported OSs are Windows and Linux.
...
YYYY/MM/DD HH:MM:SS.mmm: << message
The form of sent message log items is:
YYYY/MM/DD HH:MM:SS.mmm: >> message
In each case, mmm is the milliseconds past the second.
...
The available proxies are:
0 - Native
1 - RS232
2 - USB
3 - Ethernet
4 - SPI
8 - FTDI
9 - TotalPhaseSPI
Only those proxies enabled in the INI file can be specified. Linux supports only the Ethernet proxy. The default proxy is Native, and is always available.
Replies:
failed,framework %d is current framework
failed,no such framework as %d
failed,can't create framework %d: old connection %d restored
success
getini
Syntax:
getini,section,key
Reads the AWE server INI file.
Replies:
success,section=section_name,key=key_name,value=key_value
A return value of '0,no such key' indicates the INI file does not contain the specified key.
...
Performs a get call on the specified module with the given mask. See module documentation for specifics. If the module has no get function, does nothing.
Replies:
success,module_name=instanceID
get_filesystem_info (BSP)
...
This command queries the target file system properties. On success it returns:
success,type,size,available,overhead,deleted,inuse,free,sizes
where:
type is 1 for Native, or 2 for FLASH.
size is the target device size in words – note that the implementation may only use a portion of the total FLASH storage for the file system.
available is the number of available storage words.
overhead is number of words used for internal data structures
deleted is the number of words used by deleted files
inuse is the number of words used for all purposes
sizes is (block size in words << 16) | max filename length
Note that the file system does not release storage from deleted files – that storage is lost. Repeatedly creating and deleting files will consume all storage. The file system can be restored to its initial empty state with erase_all.
get_first_core (BSP)
Syntax:
...
Returns the target info for the first core in the system which should have ID zero. See get_target_info.
get_next_core (BSP)
Syntax:
...
Returns the target info for the core following prevCoreID in the system. If there is no successor core, it fails. See get_target_info.
get_cores (BSP)
Syntax:
get_cores
Returns the number of cores (or instances) defined by the target.
success,numberOfCores
get_core_list (BSP)
Syntax:
...
Returns the number of cores (or instances) defined by the target, and all their IDs. See also get_instance_table below which is an alias.
success,number of cores, <list of core IDs>
get_instance_table (BSP)
Syntax:
...
This command reads the specified file from the target file system, and writes the file as AWE_directory/filename to your hard drive, in which case it reports success. There are many possible failures including file not found and the target not having a file system.
...
This command is intended for rare cases where a device has to be reconfigured , but can't be while AWE has an open handle on it.
...
Reopens a device previously closed by close_output. The same parameters are used as when the device was originally opened when audio started. It is an error to reopen an already open device, or use an index that does not exist.
...
This command writes the specified local hard disk file to the target file system with the file attribute specified. If a file of that name exists on the target, it is first deleted (see delete_file). There are many possible failures including file not found on your hard disk, not enough space on the target, and the target not having a file system.
The attribute value may be any 7 bit value constructed by orring the following together expressed as decimal:
| Code Block |
|---|
#define LOAD_IMAGE 0x01 |
...
#define STARTUP_FILE 0x02 |
...
#define DATA_FILE 0x04 // Any file of type "Other" |
...
#define COMPILED_SCRIPT 0x08 |
...
#define COMMAND_SCRIPT 0x10 |
...
#define PRESET_SCRIPT 0x20 |
...
#define COMPILED_PRESET_SCRIPT 0x28 |
...
#define LOADER_FILE 0x40 |
Common useful values are 0x18 (= decimal 24) for a compiled AWB file, and 0x1a (= decimal 26) for a bootable compiled AWB file. Other possible attribute combinations are generally not useful.
...
A useful set of commands to compile a script file, and load it into a file system is:
erase_all
compile,1,source_file.aws, destination_file.awb
write_file, destination_file.awb,26
When you next reset the target, the layout should be running. Note that the AWS and AWB extensions are convention only, you can use anything you like.
get_extended_info
Syntax:
get_exterdedextended_info
This command returns user version followed by 12 undefined words
success,user_version,<12 undefined words>
get_first_io
Syntax:
get_first_io
This command returns the first I/O object, as in this example:
success,Input,1,1,48000,1074003968,256,4194304,
The format is
success,instance_name,instanceID,boundID,sampleRate,info1,info2,info3,
Where:
boundID
...
– zero if the pin is not bound, othereise the ID of the bound wire.
info1
...
– a packed bitfield of 10 bit channels | 17 bits blockSize | 4 bit sample size bytes | 1 bit complex
info2
...
– a packed bitfield of 17 bits blockSize | 6 bits data type
info3
...
– a packed bitfield of 10 bits rows | 10 bits cols | 1 bit isIPC | 1 bit isPrivate | 1 bit clock master | 1 bit special
For more information on these fields see Framework.h.
See get_next_io.
get_first_object
Syntax:
get_first_object
...
where each member is formatted as:
member_name=member_type:value
The layout of all classes is given in the schema file, where each member is named and its type given: className will be found in the schema file. The value is displayed appropriately for the type: float values are displayed using %g, all other values are displayed as decimal unsigned integers.
If the member is an array of fixed bounds in the schema, then each element of the array is displayed in the form:
member_name[subscript]=type:value
where the subscript ranges from 0 to N-1.
...
get_heap_count
This command returns:
success, number_of_heaps
Currently this value is always 3.
...
success, free1,free2,free3,size1,size2,size3
where:
freeN
...
– the number words available in heap N.
sizeN –
...
the total size of heap N
All sizes are in 32-bit words.
...
When first created, modules are active. See set_module_state.
get_moduleclass_count
Syntax:
...
get_moduleclass_info,module_class_index
where:
module_class_index must be in the range 0 to one less than the value returned by get_moduleclass_count.
On success, the return value is:
success,className,classID,nParams,DLLName
where:
className
...
– the name of the class as it appears in the schema file,
classID
...
– the numeric value of the class id,
nParams
...
– the number of public and private parameters an instance of the module may take. The values are packed as separate 16 bit numbers into a 32 bit value. The high 16 bits represent the number of private words; the lower 16 bits represent the number of public words.
DLLname is the library the module is found in.
...
Returns the next I/O object in the form described in get_first_io, or:
failed, no more I/O pins
...
Returns the next object in the forms described in get_first_object, or:
failed, no more objects
...
get_object_byname,instanceName
where:
instanceName is some identifier
The command looks up instanceName in the object symbol table. If found, the reply value is as described in get_first_object, otherwise it is:
failed, 'instanceName' is undefined
...
get_schema,className
where:
className is some identifier
The command looks up className in the schema symbol table. If found, the reply value is:
success,Class=className,ClassID=id,member,...
where:
className is the argument to the command,
id is the numeric id of the class,
each member is formatted as:
member_name=type
otherwise, the reply is:
failed, class 'className' is undefined
...
success,sampleRate,profileFreq,packetBufLen,nCores,nThreads,nInputs,nOutputs,baseBlockSize,packedInfo,version,CpuType,targetName,proxyName,CpuFreq,instanceID,isSMP,nInputPins,nOutputPins,featureBits
Where:
sampleRate - target sample rate in Hz
profileFreq - target profiling clock frequency in Hz
packetBufferLen - the size in words of the target's packet buffer
nCores - the number of cores the target currently has. For SMP targets the default 2. nCores can be changed using the set_cores command.
isSMP – flag indicating if core supports SMP
nInputPins – number of input pins
nOutputPins – number of output pins
baseBlockSize - the base block size for audio. Layouts must use an integer multiple of this nThreads - the number of concurrent threads per core. Typically 2 for embedded targets and 4 for SMP targets.
nInputs - the number of input channels
nOutputs - the number of output channels
packedInfo – 7 bits CPU type | 8 bits output channels | 8 bits input channels | 1 bit floating point | 1 bit file system | 4 bits sizeof(int)
version - 32 bit version number usually expressed as w.x.y.z where each field is a byte of the word most significant first
targetName - an up to 8 character name for the target
CpuFreq - the frequency of the target clock in Hz
instanceID - the ID of this core in the range ((0…15) * 16) i.e. (0, 16, 32, …, 240)
isSMP - 1 if the core is SMP (Windows or Linux)
featureBits - usually 0, reserved for DSP Concepts internal use
get_type (BSP)
Syntax:
get_type,expression
Evaluates the type of expression and returns the type as an integer if the expression is legal. The possible reported values are:
0 - integer
1 - unsigned integer
2 - float
3 - fract
4 - object
5 - pointer to integer
6 - pointer to unsigned integer
7 - pointer to float
8 - pointer to fract
9 - pointer to object
get_value
Syntax:
get_value,expression
where expression is formed as follows:
instanceName [. memberName]
InstanceName must be the name of some object. The first memberName must name a member of the class of which instanceName is an instance. Subsequent terms depend on the type of the member as follows:
Member Type | Followed by |
int | nothing, reply is success,address,int,intvalue |
float | nothing, reply is success,address,float,floatvalue |
[N]int | [0 : N-1], reply is success,address,int,intvalue |
[N]float | [0 : N-1], reply is success,address,float,floatvalue |
*className | .member belonging to className (follows pointer) |
**className | [subscript].member belonging to className (follows subscripted pointer) |
className | .member belonging to className (accesses member) |
Note that the final three type name members: if the types of those members are not one of the first 4 scalar forms, then more members must be named to complete the expression. This continues iteratively until the expression reaches one of the first 4 scalar forms.
If the expression is not legal according to these rules, one of the following may be returned:
failed, 'string' is not an identifier
failed, 'name' requires dot expression
failed, no such member of 'class' as 'string'
get_version (BSP)
Syntax:
get_version
...
Begins logging of binary commands sent from the Server to the target. The commands are buffered in internal memory on the PC. When complete, call end_binary to write the commands to the specified file filename.
make_binary is used to create compiled scripts on the target. Only a subset of commands are stored – only those needed to actually instantiate the system and begin processing. The commands logged are:
bind_wire
audio_pump
create_layout
create_module
create_wire
destroy
set_module_state
set_value
write_float_array
write_fract_array
write_int_array
open_web_page (BSP, GUI only)
...
If there are no layouts to pump, the error message is:
Server response: "failed,no layout(s) to pump"
If server is not connected to the target, the error message is:
Server response: "failed,not connected to the target"
otherwise the replay is:
success, cycles, cycles interval
This call is intended to be used with the command write_pump_read for testing. See also fast_write and fast_read.
pump_layout
Syntax:
pump_layout,layout_instance_name [,pump cycles]
...
This command pumps a single layout as though by pump above. It is intended to be used with writing an input wire and reading an output wire for testing. See also fast_write and fast_read.
On success the reply is
success, instance_name=instanceID [, pump cycles]
...
pump_module,module_instance_name
where module_instance_name must be an object created by create_module.
This command pumps a single module as though by pump_layout above. It is intended to be used with writing an input wire and reading an output wire for testing. See also fast_write and fast_read.
On success the reply is
success,Name=instanceID
...
On success, the reply is as described in get_first_io.
query_pump (BSP)
Syntax:
query_pump
This command reports the audio pump status as an integer. The possible values are:
0 - target has no layout, not pumping
1 - target has a layout, not pumping
2 - target has no layout, pumping
3 - target has a layout, pumping
This command is intended for use by an external monitoring process which watches for the pump dying by transitioning from 3 to 1, which usually indicates an I/O error on audio hardware which can only happen on Linux systems.
...
where wireName must be an object created by create_wire.
On success the reply is
success,wireName=objectID,sampleRate,info1,info2
...
read_float_array,expression,count
where:
expression is an expression that evaluates to an array element, so must be subscripted
count is the number of values to read starting at that element
The reply is
success,val[0], ..., val[count-1]
...
read_fract_array,expression,count
where:
expression is an expression that evaluates to an array element, so must be subscripted
count is the number of values to read
The reply is
success,val[0], ..., val[count-1]
where each value is reported as float interpreting each value as fract32, and so constrianed constrained to report a value in the range -1 to 1.
...
read_int_array,expression,count
where:
expression is an expression that evaluates to an array element, so must be subscripted
count is the number of values to read
The reply is
success,val[0], ..., val[count-1]
...
This command executes the commands stored in fileName. Generally, these files have an AWS extension, and are generated by Designer.
...
Assigns to or creates in the INI file an item of the form:
| Code Block |
|---|
[section] |
...
key=value |
set_call
Syntax:
set_call,module_name,mask
...
If the description file has errors, the offending error is reported by failed,parse error, and the detailed error will be in awelog.txt. It is outside the scope of this document to describe the description file format and the possible parse errors which are legion. On error, the existing system is unchanged.
Any attempt to set the description file currently in use reports success,no change. If the command succeeded in loading a new file (or reverting to SMP) the reply is success.
set_cores (BSP,
...
Windows, Linux)
Syntax:
set_cores,N
For SMP targets only, destroys all existing core objects, and creates N new ones. The allowed range is 1-16. The number of cores on server start is 2. The current number of cores can be found using get_cores. Embedded targets with multiple cores have a fixed number of physical cares and do not implement this command.
...
As objects are created, they are assigned IDs starting at 1. If a layout needs to assign a new ID, it must be >= 30000, and not in use by another object. The object_name must be that of an existing object.
...
set_module_state,module_instance_name,state
where:
module_instance_name is the name of a module created with create_module, or is a dot-expression naming a member of an object that is a module
state is a decimal value, and one of
0: active
1: bypass
2: mute
3: inactive
This command sets the state of the module. On success the reply is:
success, module_instance_name=instanceID
See also get_module_state.
set_path (BSP)
Syntax:
set_path,path_item1 [, path_item2 ... ,path_itemN ]
This command takes any number of arguments, each of which is a file system path. It is used to set the search path for audio files specified to audio_pump. The paths are persisted in the server INI file as:
| Code Block |
|---|
{AudioPath] |
...
Path=item1| ... |itemN |
When audio_pump is searching for a file, it works through this list in order, and uses the first match found.
...
Sets the communication time out between Matlab and the server. By default, the value is 4000, or 4 seconds. (This command is useful to prevent issues when you issue a command, such as erase_all, which may take a long time to execute.)
...
set_value,expression,value [, expression,value]*
where:
expression is as described in get_value,
value is a number to be assigned to the location described by expression
There may be any number of [expression,value] pairs given to the command. On completion of the last assignment, the set function of each unique module instance (if any) referenced by any of the expressions are called.
On success the reply is:
...
If the server has any dialogs created by create_dialog, then show,0 causes the server dialog to be hidden. The dialog is un-hidden by destroy, using destroy_dialog to destroy the last child dialog, or by show,1. The show,0 command does nothing if there are no child dialogs.
...
write_float_array,expression,val0,....,valN-1
where:
expression is an expression that evaluates to an array element, and so must be subscripted
This command writes the values to each successive element location starting at expression. The reply is
...
write_float_array_partial,expression,val0,....,valN-1
As write_float_aray except the final set call is suppressed.
...
write_fract_array,expression,val0,....,valN-1
where:
expression is an expression that evaluates to an array element, and so must be subscripted
This command writes the values to each successive element location starting at expression. The reply is
success
Misuse of the command can corrupt storage, because there is no bounds checking for arrays.
...
write_fract_array_partial,expression,val0,....,valN-1
As write_fract_aray except the final set call is suppressed.
...
write_int_array,expression,val0,...,valN-1
where:
expression is an expression that evaluates to an array element, and must be subscripted
This command writes the values to each successive int location starting at expression. The reply is
success
Misuse of the command can corrupt storage, because there is no bounds checking for arrays.
...
write_int_array_partial,expression,val0,...,valN-1
As write_int_aray except the final set call is suppressed.
...
write_pump_read,layout,inputWire,outputWire,val1, ... ,valN
This command writes the vali vali to inputWire, then pumps the layout, then reads outputWire and replies:
success,cycles,val1, ... ,valN
...
Commands can produce error messages from the following table:
Text | Description |
failed, heap type index range | A heap index was not in the range of heaps |
failed, awe_fwMalloc no more storage | The given heap does not have enough storage to satisfy the requested size |
failed, awe_fwMallocScratch no more storage | The scratch heap does not have enough storage to satisfy the requested size |
failed, constructor argument count | A create_xx call has an incorrect number of arguments |
failed, class index out of range | The given class index is not in the range of classes |
failed, class not found | The named class was not found in the symbol table |
failed, module already owned | An attempt was made to give a module to a layout when it is already in another layout |
failed, address outside heap | An attempt was made to assign to a location not in any heap |
failed, not a wire | A wire argument to create_module is not actually a wire |
failed, number of inputs and outputs must match | Some modules require that the number of inputs and outputs are the same |
failed, input pin types must be the same | Some modules require that the types of input and output pins be the same |
failed, module needs at least one input | Many modules require at least one input |
failed, module needs at least one output | Many modules require at least one output |
failed, inputs must match corresponding outputs | Some module require that each |
...
ith input have the same type as each |
...
ith output | |
failed, not a module | An attempt was made to give an object not a module instance to add_module |
failed, I/O count error | The input/output count is not acceptable |
failed, parameter error | A parameter given to create_module is wrong for the specified module class |
failed, no more objects | There are no more objects for get_next_object to display |
failed, not object pointer | An expression expected to be of pointer type is not |
failed, not input pin | The pin must be an input pin. |
failed, I/O pin in use | An attempt was made to bind an I/O pin that was already bound with bind_wire |
failed, pin types not compatible | An attempt was made using bind_wire to bind a wire incompatible with the I/O pin |
failed, pin sizes not compatible | An attempt was made using bind_wire to bind a wire not an integer multiple of the I/O pin size |
failed, not output pin | The pin must be an output pin. |
failed, no more I/O pins | There are no more pin objects for get_next_io to display |
failed, no layouts to pump | 'pump' was called when no layouts exist |
failed, module must have only one output | Many modules require only one output |
failed, output wire must have only one sample | Some modules require that an output have only a single sample |
failed, incompatible block sizes | All contained modules must have the same block size |
failed, wire index out of range | A container wire vector indexed a wire out of range |
failed, unknown error %d | An unknown error occurred |
failed, argument count | A command had an invalid number of arguments |
failed, not implemented on target | The target does not implement the command |
failed, instance name '%s' not identifier | The argument must be an identifier |
failed, instance name '%s' is already used | The instance name has already been defined |
failed, class name '%s' is not defined | An attempt was made to use an undefined class name |
failed, class name '%s' has different classID than created instance | An object was created, but then found to have a different class than it should have |
failed, instance name '%s' is not a pin type | The argument must be a pin type |
failed, name '%s' undefined | A name was seen that has not been defined |
failed, expression error | The expression given to get_value or set_value had an error |
failed, wire name '%s' undefined | A supposed wire name given to create_module is not defined |
failed, wire name '%s' is not a Wire | A supposed wire name given to create_module is not of type Wire |
failed, module name '%s' undefined | A supposed module name given to add_module is undefined |
failed, '%s' is not a module | A supposed module name given to add_module is not a module |
failed, unknown argument | A command that takes a symbolic argument had an unknown string argument |
failed, open sound card for input returned an error | 'audio_pump' could not open the sound card for input |
failed, player create returned 0x%08x | 'audio_pump' could not open the sound card for output |
failed, renderer create returned 0x%08x | 'audio_pump' could not create |
...
output object | |
failed, empty filename | A required filename was empty |
failed, unknown command '%s' | The command keyword is unknown |
failed, empty command | The command was empty |
failed, can't find instance class | An attempt to lookup the class of an instance failed |
failed, can't find instance | An attempt to lookup an instance failed |
failed, '%s' requires subscript | An expression requires a subscript |
failed, '%s' syntax error: missing ']' | Malformed subscript |
failed, '%s' subscript %d out of range | A subscript is outside the array bounds (static arrays only, very rare) |
failed, '%s' requires dot expression | An expression stopped early |
failed, no such member of '%s' as '%s' | The member name given is not a member |
failed, %s(%d): empty class name | Empty class name in schema file |
failed, %s(%d): syntax error in alias of class '%s' | Error while aliasing one class to another in schema file |
failed, %s(%d): unknown base class'%s' in alias of class '%s' | Reference to unknown base class while aliasing one class to another in schema file |
failed, %s(%d): comma expected in class '%s' | Missing comma in schema file |
failed, %s(%d): syntax error in derivation of class '%s' | Syntax error parsing derived class in schema file |
failed, %s(%d): unknown base class '%s' in alias of class '%s' | Attempt to derive a class from an unknown base in schema file |
failed, %s(%d): unexpected '{' in body of class %s | There can only be one level of bracing in schema files |
failed, %s(%d): empty member name in class %s | Member name is empty in schema file |
failed, %s(%d): non-numeric dimension in class %s | Array dimension has non-numeric subscript in schema file |
failed, %s(%d): expected ']' to close dimension in class %s | Missing close bracket in array dimension in schema file |
failed, %s(%d): empty type name in class %s | A type name is empty in schema file |
failed, %s(%d): unknown type name '%s' in class %s | A type name is undefined in schema file. |
failed, %s(%d): unexpected end of file in class %s | Unexpected end of file in schema file |
failed, no such core | A core ID was specified that the target does not have. |
failed, too many bound wires | An attempt was made to bind more than 17 times to an input pin |
Schema Files
Schema files provide a means for describing the layout of DSP storage that is compact and has a simple grammar, and does not need the complexity of the C/C++ type system.
The server has a file Schemas.sch that defines all the classes in the DSP. Each schema corresponds to a structure in the code. Class names and member names must be identifiers in the C/C++ sense. Schema files support C++ comments only.
The form of a schema is:
ClassName
{
member1 type
….
memberN type
}
This is the simplest form, and directly maps to a C struct.
| Code Block |
|---|
ClassName { member1 type …. memberN type } |
The supported types are as follows:
Type | Description |
int | 32 bit integer |
float | 32 bit IEEE float |
[N]int | array of integer with N elements |
[N]float | array of floats with N elements |
*int | pointer to array of integer with unknown number of elements |
*float | pointer to array of float with an unknown number of elements |
*className | pointer to a class instance |
**className | pointer to an array of pointers to class instance with unknown number of elements |
className | a nested structure |
To support mapping to DSP code, class names may have an associated class ID like this:
| Code Block |
|---|
className value |
...
{ |
...
…. |
...
} |
If value is not present, the value zero (unknown ID) is used. The value may be in hex or decimal.
Classes may derive from other classes like this:
| Code Block |
|---|
A |
...
{ |
...
…. |
...
} |
...
B, A |
...
{ |
...
…. |
...
} |
The meaning is the same as public derivation in C++. In the example above, B inherits all the members of A.
The use of a class ID may be combined with inheritance like this:
| Code Block |
|---|
className value, baseClass |
...
{ |
...
…. |
...
} |
As expected, the new class gets the given class ID, and also inherits all the members of the base class. There is no limit to inheritance depth.
All type names must be declared before use. This means that a circular definition such as:
| Code Block |
|---|
A |
...
{ |
...
m *B |
...
} |
...
B |
...
{ |
...
m *A |
...
} |
can't be written, since an attempt is made to refer to B before it is declared.
...
The packet header looks like this:
struct SBinaryPacket
{
| Code Block |
|---|
struct SBinaryPacket { /** Magic packet header value. */ |
...
unsigned int m_magic; |
...
/** Length of data in bytes. */ |
...
unsigned int m_len; |
...
/** Length of data in floats. */ |
...
unsigned int m_nFloats; |
...
/** Command opcode. */ |
...
unsigned int m_opcode; |
...
}; |
m_magic – contains 0x07ff0003
m_len – total packet size in bytes
m_nFloats – payload size in words – note that payload data is not constrained to floats
m_opcode – command opcode, always 30 to server, always 29 from server
It is always required that string data is also sent with a command to the server to specify a destination address as an expression. This data follows the last payload word. Let us assume a payload of 32 words, and string value containing 10 characters including the terminating NULL. Then the length values will be:
...
Any binary message with an opcode other than 30 causes a server assertion failure, since binary messages are intended for internal AWE use only, and would be a serious error with other opcodes. The reply to this message will be success or failed,<reason>, as with other messages. This command is sent to the AWE server by the Maltalb DLL when it processes fast_write.
The text command documented earlier fast_read generates a binary reply with payload of the number of words requested, and with no string part or a string message failed,<reason>. For that message, we have:
...
Binary packets have the form:
word 0: 16 bit length | 8 bit core ID | 8 bit opcode
word 1 - N-2 command payload
word N-1: XOR sum of all preceding words
Reply packets have the form:
word 0: 16 bit length in words | 16 bit zero
word 1 - N-2 reply payload
word N-1: XOR sum of all preceding words
Commands may have no payload. Replies always have at least one payload word, and for almost all of them word 1 is the return value - frequently the error code. Many replies only have this one payload word.
For more about binary packets, including a detailed list of available commands, see the ‘Tuning AWE Core Tuning Command Syntax and Protocol’ section of the AWECore Integration Guide at this location: https://w.dspconcepts.com/docsProtocol.