DE102019211856A1 - Data structure, control system for reading in such a data structure and method - Google Patents

Abstract

The invention relates to a data structure (10) for transmission from a data source to a data sink by a transmission module, comprising: a file which is subdivided into one or more consecutive data packets (11_1, .., 11_N), one or more different identifier values ( 12_1, ..., 12_N), which is used to identify the respective end of the one or more data packets (11_1, .., 11_N) at the respective end of the one or more data packets (11_1, .., 11_N) in the data structure (10) are stored, a file header (13) which is configured to indicate the position of the one or more different identifier values (12_1, ..., 12_N) in the file, and which is also configured to indicate at least the size and the number of a or several data packets (11_1, .., 11_N) is configured, wherein the file header (13) is further configured in such a way that prior to the transmission of the one or more data packets (11_1, .., 11_N) to a data sink provided for this purpose, hst a reading of a placeholder different from the respective one or more identifier values (12_1, ..., 12_N) into the data sink at the address at which the respective identifier value (12_1, ..., 12_N) of the one or more respective data packet will later be read (11_1, .., 11_N) is read in. Furthermore, the invention relates to a control system for reading in such a data structure and method.

Description

The invention relates to a data structure for transmission from a data source to a data sink by a transmission module. The invention also relates to a control system for reading in such a data structure. The invention also relates to a method for improved loading and further processing of a file which is created on a data source for transmission to a data sink.

Complex control systems, for example SOCs (System-on-a-Chip), nowadays have at least one direct memory access, also known as a DMA controller (Direct Memory Access). Such SOCs are complex, heterogeneous systems that use a hardware platform with z. B. processors, buses, memory and peripherals and software such as. B. device drivers, firmware, operating systems. The SoCs can be used in embedded systems.

Such a DMA controller usually accesses a main memory directly via a bus system. In SOCs, such DMA controllers are used to transport data between peripheral devices and the central, rewritable RAM (Random Access Memory) main memory.

This relieves the central processor; the CPU. DMA controllers require control and parameter information to carry out the data transfer. This information is provided by the CPU. When the DMA controller receives a data transfer request from the peripheral device, it receives bus control rights from the CPU. The DMA controller thus starts writing the data from the peripheral device to the main memory. When the data transfer is complete and the data has been validated / verified, the DMA controller sends a completion signal to the peripheral device and to the CPU. The next data can then be transferred. Simple DMA controllers only indicate the completion of the transferred data.

With simple DMA controllers, however, it is not possible to read out how much data has been read into the main memory; only information about the complete completion of the reading can be provided. This means that further processing of the transferred data can only be started after the loading process has been completed.

The DE 10 2004 006 767 B4 discloses a method for transporting data sections between a memory and at least one peripheral device, comprising the steps to be carried out in the specified order: generating a reference pointer pointing to a peripheral device to complete a DMA data transport from previous configuration data; Transmission of a control signal to a DMA controller (DMA), with which the DMA controller (DMA) is requested to load configuration data via a transport of a data section to be carried out; Loading of configuration data via the data transport to be carried out by the DMA controller (DMA) based on the reference pointer; and performing the data transport by means of the DMA controller (DMA).

The object of the invention is therefore to provide means which enable an improved data transmission between a data sink and a data source.

This object is achieved by a data structure with the features of claim 1 and a control system with the features of claim 6.

The object is also achieved by a method for improved loading and further processing of a file having the features of claim 10.

The subclaims list further advantageous measures that can be combined with one another in any suitable manner in order to achieve further advantages.

• - eine Datei, die in ein oder mehrere aufeinanderfolgende Datenpakete unterteilt ist,
• - einen oder mehrere voneinander verschiedene Kennzeichenwerte, welche zur Kennzeichnung des jeweiligen Endes des eines oder mehreren Datenpaketes am jeweiligen Ende des einen oder mehreren Datenpaketes in der Datenstruktur abgelegt sind,
• - einen Dateikopf, welcher zum Aufzeigen der Position der einen oder mehreren voneinander verschiedenen Kennzeichenwerte in der Datei konfiguriert ist, und welcher ferner zur Angabe zumindest der Größe und der Anzahl der ein oder mehreren Datenpakete konfiguriert ist, wobei der Dateikopf so konfiguriert ist, dass vor dem Übertragen der ein oder mehreren Datenpakete an eine dafür vorgesehene Datensenke zunächst ein Einlesen eines von dem jeweiligen ein oder mehreren Kennzeichenwerten unterschiedlichen Platzhalters in die Datensenke an derjenigen Adresse, an der später der jeweilige Kennzeichenwert des ein oder mehreren jeweiligen Datenpaketes eingelesen ist, bewirkt wird.

The object is achieved by specifying a data structure for transmission from a data source to a data sink by a transmission module, having:

• - a file that is divided into one or more consecutive data packets,
• - one or more different identifier values which are stored in the data structure to identify the respective end of the one or more data packet at the respective end of the one or more data packet,
• a file header which is configured to indicate the position of the one or more mutually different identifier values in the file, and which is further configured to indicate at least the size and the number of the one or more data packets, the file header being configured in such a way that before before transmitting the one or more data packets to a data sink provided for this purpose, initially one of the respective one or more identifier values is read in different placeholders in the data sink at the address at which the respective identifier value of the one or more respective data packet is later read in.

This means that each of the data packets has an identifier value, for example 4 bytes, at its end. These identifier values are added to each data packet when the file is divided into the preferably several individual data packets.

The file header, which is executed before the individual data packets are loaded, stores the position of these identifier values. Initially, the execution of the file header causes placeholders to be read in, which are read in at those addresses in the data sink to which the identifier values of the individual data packets are later written. This means that before the file or the individual data packets are read in, there are other values, namely the placeholders, than the original identifier values at the locations / addresses provided for this in the data sink. If the individual data packets are then read in successively, the placeholders are successively overwritten by the "correct" identifier values.

Depending on the progress of the data transfer, you can now query which of the placeholders has already been overwritten with the correct indicator values. This means that while a transmission module, for example a DMA controller, is working in the background, it can be determined how many data packets have already been written into the data sink. The data packages that have already been fully loaded can be processed further at the same time as the subsequent data packages are being loaded.

Further processing of a data packet can begin as soon as the identifier value associated with the data packet has overwritten the corresponding placeholder in the data sink. If the identifier value associated with the data packet is not yet in the data sink, further processing cannot yet be started, since this data packet has not yet fully arrived in the data sink.

The file header is configured in such a way that it is executed first by the transmission module, for example the DMA controller. The position of the file header can be different in the data to be transmitted, for example as a header at the beginning or at the end or in the middle.

The data structure enables parallel loading and further processing of a file, which is preferably divided into several data packets, independently of the transmission module used, which is used to transmit the data structure from the data source to the data sink.

The file header is preferably configured in such a way that the one or more data packets are read out in a predetermined sequence. The data packets are preferably read out as a data flow, i.e. simply one after the other.

In a preferred embodiment, the file header includes the size of the data structure and the number of data packets. The number can also be one.

Furthermore, the file header is configured in such a way that a space requirement corresponding to the size of the data structure is reserved in the data sink.

The data structure is preferably designed for storage on a data source. The data source is preferably a peripheral storage unit, for example a USB storage device, etc.

• eine Datensenke mit einem Arbeitsspeicher,
• eine direkte Speicherzugriffseinheit zum direkten Einlesen der Datenstruktur von einer Datenquelle in den Arbeitsspeicher der Datensenke, wobei die direkte Speicherzugriffseinheit dazu ausgebildet ist, zunächst den Dateikopf auszuführen, wobei die Speicherzugriffseinheit ferner dazu konfiguriert ist, durch Ausführen des Dateikopfes den für das Einlesen der ein oder mehreren Datenpakete benötigten Arbeitsspeicher mit einer ersten zu beschreibenden Adresse und einer letzten zu beschreibenden Adresse zur Verfügung zu stellen und wobei die direkte Speicherzugriffseinheit dazu konfiguriert ist, in die letzte zu beschreibende Adresse des jeweiligen Datenpaketes einen jeweils von dem Kennzeichenwert unterschiedlichen Platzhalter zu speichern,
• so dass beim Speichern des ein oder mehreren Datenpaketes der jeweilige Platzhalter erst dann mit dem Kennzeichenwert des ein oder mehreren Datenpaketes überschrieben ist, wenn das zu dem jeweiligen Kennzeichenwert gehörende jeweilige Datenpaket vollständig in der Datensenke gespeichert ist.

The object is also achieved by a control system for reading in a data structure as described above, the control system comprising:

• a data sink with a working memory,
• a direct memory access unit for reading the data structure directly from a data source into the main memory of the data sink, the direct memory access unit being designed to initially execute the file header, the memory access unit also being configured to execute the file header for reading in the one or more To provide data packets required main memory with a first address to be written and a last address to be written, and wherein the direct memory access unit is configured to store a placeholder different from the identifier value in the last address of the respective data package,
• so that when the one or more data packets are saved, the respective placeholder is only overwritten with the identifier value of the one or more data packet when the respective data packet belonging to the respective identifier value is completely stored in the data sink.

The control system can be used in all areas where large amounts of data have to be loaded and processed. By the invention Control system, it is possible to query the current transfer status by checking whether one of the placeholders has been overwritten with the intended code value. If this is the case, the data packet belonging to the identifier value has already been fully loaded and can already be processed further while the data packet to be subsequently loaded is being read into the data sink at the same time.

With such a control system, the loading and further processing of data packets can take place in parallel, although the direct memory access unit does not provide any information about how many data or data packets have already been read. Since the entire amount of data is requested from the data source, preferably the peripheral storage unit, the transmission itself takes place without any delay by re-addressing the next data packet. This enables considerable time savings.

The data sink itself is preferably designed as a working memory. This is controlled by the direct memory access unit, preferably via a bus.

The direct memory access unit is also preferably designed as a DMA controller. A DMA controller is to be understood as any hardware that can perform direct memory access either to the main memory itself or to a peripheral memory unit without using the CPU.

The control system is preferably arranged on an SOC (System-on-a-Chip). Such a SOC is preferably used in embedded systems.

• - Laden und Auslesen eines Dateikopfes einer Datenstruktur durch einen DMA-Controller, wobei die Datenstruktur ein oder mehrere Datenpakete mit jeweils einem unterschiedlichen Kennzeichenwert umfasst, wobei der jeweilige Kennzeichenwert am einzulesenden Ende des einen oder mehreren Datenpaketes angeordnet ist,
• - Bereitstellen des benötigten Speicherplatzes für die ein oder mehreren Datenpakete mit den jeweils einem unterschiedlichen Kennzeichenwert in einem Speicher der Datensenke durch den DMA-Controller,
• - Einlesen eines jeweiligen zu den jeweils einem unterschiedlichen Kennzeichenwert korrespondierenden Platzhalters an diejenige Adresse, welche für den jeweils einen unterschiedlichen Kennzeichenwert in der Datensenke vorgesehen ist, wobei sich der jeweilige Platzhalter von dem jeweiligen Kennzeichenwert unterscheidet, so dass beim Speichern des einen oder mehreren Datenpaketes der jeweilige Platzhalter erst dann mit dem jeweiligen Kennzeichenwert des einen oder mehreren Datenpaketes überschrieben ist, wenn das zu dem jeweiligen Kennzeichenwert gehörende Datenpaket vollständig in dem vom DMA-Controller zur Verfügung gestellten Speicher der Datensenke gespeichert ist,
• - Sukzessives Laden und Auslesen der ein oder mehreren Datenpakete mit den unterschiedlichen Kennzeichenwerten.

Furthermore, the object is achieved by a method for improved loading and further processing of a file which is created on a data source for transmission to a data sink, comprising the steps:

• - Loading and reading out a file header of a data structure by a DMA controller, the data structure comprising one or more data packets each with a different identifier value, the respective identifier value being arranged at the end of the one or more data packets to be read,
• - Provision of the required storage space for the one or more data packets, each with a different identifier value in a memory of the data sink by the DMA controller,
• - Reading in a respective placeholder corresponding to a different identifier value to the address which is provided for each different identifier value in the data sink, the respective placeholder differing from the respective identifier value, so that when the one or more data packets are saved the the respective placeholder is only overwritten with the respective identifier value of the one or more data packets when the data packet belonging to the respective identifier value is completely stored in the data sink memory provided by the DMA controller,
• - Successive loading and reading of the one or more data packets with the different identification values.

• - Abfragen, ob der jeweilige Platzhalter für den jeweiligen Kennzeichenwert schon mit dem jeweiligen Kennzeichenwert überschrieben worden ist.

The method preferably comprises the further step:

• - Query whether the respective placeholder for the respective license plate value has already been overwritten with the respective license plate value.

In a further preferred embodiment, when several data packets are provided with an overwritten first placeholder, the second data packet following the first identifier value is loaded with a second identifier value through the corresponding first identifier value of a first data packet, while the first, just loaded data packet is further processed. This is the procedure for all existing data packets. The first and second data packet mean a sequence and are not related to the position of a data packet in the file.

If the data packets are read by the DMA controller, the file header is first read, which contains the identifier values and their position in the file. The DMA controller writes different placeholders for the identifier values to the addresses in the main memory provided for the identifier values. As a result, by reading out the main memory while the DMA controller is working in the background, it can be determined whether a first data packet has arrived in the main memory. If the first identifier value is in the main memory at the specified address, then all the data of the first data packet have been read and the further processing of the first data packet can be started. If the first identifier value is not yet in the main memory, further processing of the first data packet cannot yet be started because it has not yet been completely read.

This continues until all data packets have been read.

This method enables the parallelization of loading and further processing of a file, although the DMA controller does not need to provide any information about how much data has already been read. Since the entire amount of data is requested from the external memory, the transmission itself takes place without any delay by re-addressing the next data packet. This enables considerable time savings.

The method according to the invention and the data structure according to the invention can be used in all areas in which large amounts of data are transmitted to the main memory, for example of an SOC of an embedded system, and then have to be further processed within the embedded system.

This means that the current status of the transmitted data can be queried even with simple DMA controllers, for example. This enables the data to be processed further, even if the data has not yet been fully loaded.

• 1: einen erfindungsgemäßen DMA-Controller,
• 2: ein erfindungsgemäßes System-on-Chip (SOC),
• 3: eine Peripherie-Speichereinheit mit einer erfindungsgemäßen Datenstruktur,
• 4: eine Ausgestaltung eines erfindungsgemäßen Verfahrens.

Further features, properties and advantages of the present invention emerge from the following description with reference to the accompanying figures. They show schematically:

• 1 : a DMA controller according to the invention,
• 2 : a system-on-chip (SOC) according to the invention,
• 3 : a peripheral storage unit with a data structure according to the invention,
• 4th : an embodiment of a method according to the invention.

Although the invention has been illustrated and described in detail by the preferred exemplary embodiments, the invention is not restricted by the examples disclosed. Variations on this can be derived by the person skilled in the art without departing from the scope of protection of the invention as defined by the following patent claims.

1 shows an example of a DMA controller 1 according to the invention.

Such a DMA (direct memory access) controller 1 accesses directly via preferably a bus 6th ( 2 ) to a data sink with a working memory 5 ( 2 ) to read and / or write data. Computer with DMA controller 1 can transfer data to and from a peripheral storage unit, for example 7th transferred with low CPU usage.

The DMA controller 1 is used to transport data between the main memory 5 ( 2 ) and preferably the peripheral storage unit 7th . On the peripheral storage unit 4th is a data structure according to the invention 10 ( 3 ) with the data packets 11_1 , .. 11_N ( 3 ) and the file header 13 ( 3 ) saved.

There are also systems with a central DMA controller 1 , but also systems in which each component has its own DMA controller 1 disposes. As a DMA controller 1 Any block can be understood that has a memory access either to the main memory 5 ( 2 ) itself or on a peripheral storage unit 7th can perform without using an external CPU 8th ( 2 ) to claim something.

The DMA controller 1 reads the data packets 11_1 , .. 11_N ( 3 ) from the peripheral storage unit 7th into a data sink, which is used here as working memory 5 ( 2 ) is designed, a.

A simple DMA controller 1 preferably has at least one DMA control unit 2 and at least one data buffer 3 and at least one register 4th . There are several addressing methods. In a first addressing process, the DMA controller fetches 1 first of all a data byte and saves it in an internal register 4th . Then it addresses the main memory 5 ( 2 ) and transfers the data to him. In the second addressing method, the DMA controller addresses 1 the data to be fetched and puts them immediately in the working memory 5 ( 2 ) by.

Such a DMA controller 1 is for example system-on-chips (SOC) 14th ( 2 ) to find the CPU 8th ( 2 ) to relieve memory access.

2 shows such a system-on-chip (SOC) very schematically 14th with a control system 15th , which is a DMA controller 1 has as a transmission module. The DMA controller 1 is with a memory 5 and a CPU 5 on the system-on-chip (SOC) 14th arranged. Alternatively, the RAM 5 not on the system-on-chip (SOC) 14th arranged, but for data transfer with the CPU 8th or the DMA controller 1 be connected.

The system-on-chip 14 also has an input-output unit 9 on. The individual components as well as the input-output unit 9 can have one or more buses 6th be connected to each other. The DMA controller 1 reads the data packets 11_1 , .. 11_N ( 3 ) for example via the bus 6th directly into the main memory 5 on.

3 shows a data source, which is used here as a peripheral storage unit 7th is trained. On the peripheral storage unit 7th is a data structure according to the invention 10 filed.

The data structure 10 includes a file that is in one or more consecutive data packets 11_1 , .. 11_N , is divided. At the respective end of a respective data packet 11_1 , .., 11_N are characteristic values 12_1 , ..., 12_N which is the end of a data packet 11_1 , .., 11_N mark.

There is also a file header 13 which is the size of the data structure 10 , the number of the respective data packets 11_1 , .., 11_N as well as the address / position of the respective license plate value 12_1 , ..., 12_N in the file includes. The file header 13 can be used as a header or head of the data structure 10 be designed which is read out first.

The file header is also 13 for reading out the order of the one or more data packets 11_1 , .. 11_N configured. The order is preferably one after the other, that is, the data packets 11_1 , .. 11_N are retrieved sequentially as a data stream.

The file header is also 13 for reading in one of the respective license plate value 12_1 , ..., 12_N different placeholders in the data sink, here the main memory 5 ( 1 ), at the address at which the respective license plate value will later 12_1 , ..., 12_N of a respective data package 11_1 , .., 11_N is read, designed.

The file header 13 Instructions for reserving one of the size of the data packets 11_1 , .. 11_N corresponding space requirements in the main memory 5 ( 1 ) or another data sink.

4th shows a method according to the invention for improved loading and further processing of a file from a data source to a data sink.

As the first process step S1 is the loading and reading of the file header 13 ( 3 ) a data structure according to the invention 10 ( 3 ) from a peripheral storage unit 7th ( 3 ) through the DMA controller 1 ( 1 ) accomplished.

The data structure 10 ( 3 ), the size and the number of data packets 11_1 , .. 11_N ( 3 ) as well as the position of the license plate values 12_1 , .., 12_N ( 3 ) to the DMA controller 1 transfer.

In a second process step S2 provides the DMA controller 1 ( 1 ) the required storage space for the data packages 11_1 , .., 11_N ( 3 ) on the RAM 5 ( 1 ) ready and writes to the addresses which are for the license plate values 12_1 , .., 12_N ( 3 ) are provided, each with a placeholder that differs from the identifier values 12_1 , .., 12_N ( 3 ) differs.

In a third process step S3 are according to the data structure 10 ( 3 ) the individual data packages 11_1 , .., 11_N ( 3 ) into memory 5 ( 1 ), starting with the first data packet 11_1 ( 3 ).

In a fourth process step S4 it is checked whether the first placeholder for the first data packet 11_1 ( 3 ) with the first license plate value 12_1 ( 3 ), which is the end of the first data packet 11_1 ( 3 ) has been overwritten.

In a fifth process step S5 are the first placeholders for the first data packet 11_1 ( 3 ) with the first license plate value 12_1 ( 3 ) has not yet been overwritten and the DMA controller 1 ( 1 ) reads in the first data packet 11_1 ( 3 ) into memory 5 ( 1 ) continued.

In an alternative sixth process step S6 are the first placeholders for the first data packet 11_1 ( 3 ) with the first license plate value 12_1 ( 3 ) has already been overwritten. The overwrite can be queried by the memory 5 ( 1 ) while the DMA controller 1 ( 1 ) continues to work in the background, is read out. As a result of the overwrite, it is known that the first data packet 11_1 ( 3 ) already completely in the main memory 5 ( 1 ) has been read in and the first data packet can be processed further 11_1 ( 3 ) can be made while the DMA controller 1 in the background with the reading of the second data packet 11_2 ( 3 ) begins.

In a seventh process step S7 are the second placeholders for the second data packet 11_2 ( 3 ) with the second number plate value 12_2 ( 3 ) has not yet been overwritten and the DMA controller 1 ( 1 ) reads in the second data packet 11_2 ( 3 ) into memory 5 ( 1 ) continued.

In an alternative eighth process step S8 are the second placeholders for the second data packet 11_2 ( 3 ) with the second number plate value 12_2 ( 3 ) has already been overwritten. The overwrite can be queried by the memory 5 ( 1 ), while the DMA controller 1 ( 1 ) continues to work in the background, is read out. As a result of the overwrite, it is known that the second data packet 11_2 ( 3 ) already completely in the main memory 5 ( 1 ) has been read in and the second data packet can be processed further 11_2 ( 3 ) can be made while the DMA controller 1 begins in the background with the reading of the third data packet 11_3.

This is done until all the existing data packets 11_1 , .., 11_N ( 3 ) are read and all placeholders for the respective data packets 11_1 , .., 11_N ( 3 ) with the correct license plate values 12_1 , .., 12_N ( 3 ) have been overwritten.

The method according to the invention enables the loading and further processing of data, for example data packets, to be parallelized, although the hardware, here the DMA controller 1 ( 1 ), does not return automated information or provides how many data packets 11_1 , .., 11_N ( 3 ) or data has already been read.

The advantages of the invention can also be used in particular in control systems in vehicles that make large data available, exchange or retrieve. This is particularly the case in vehicles that are to be used in highly automated driving.

QUOTES INCLUDED IN THE DESCRIPTION

This list of the documents listed by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent literature cited

• DE 102004006767 B4 [0006]

Claims (12)

A data structure (10) for transmission from a data source to a data sink by a transmission module comprising: - a file which is divided into one or more consecutive data packets (11_1, .., 11_N), - one or more different identifier values (12_1, ..., 12_N), which are used to identify the respective end of the one or more data packets (11_1, .., 11_N) at the respective end of the one or more data packets (11_1, .., 11_N) are stored in the data structure (10), - A file header (13) which is configured to indicate the position of the one or more different identifier values (12_1, ..., 12_N) in the file, and which is also configured to indicate at least the size and the number of the one or more Data packets (11_1, .., 11_N) is configured, the file header (13) further configured so that before the one or more data packets (11_1, .., 11_N) are transmitted to a data sink provided for this purpose, one of the respective one or more identifier values (12_1, ..., 12_N) of different placeholders in the data sink at the address at which the respective identifier value (12_1, ..., 12_N) of the one or more respective data packets (11_1, .. , 11_N) is read in, is effected. Data structure (10) Claim 1 , characterized in that the file header (13) is configured in such a way that the one or more data packets (11_1, .., 11_N) are read out in a predetermined order. Data structure (10) Claim 1 or 2 , characterized in that the file header (13) includes the size of the data structure (10) and the number of data packets (11_1, .., 11_N). Data structure (10) Claim 3 , characterized in that the file header (13) is configured in such a way that a space requirement corresponding to the size of the data structure (10) is reserved in the data sink. Data structure (10) according to one of the preceding claims, characterized in that the data structure (10) is designed for storage on a data source. Control system (15) for reading in a data structure (10) according to one of the preceding claims, the control system (15) comprising: a data sink with a main memory (5), a direct memory access unit for reading the data structure (10) directly from a data source into the main memory (5) of the data sink, the direct memory access unit being designed to first execute the file header (13), the memory access unit also being configured to perform the File header (13) to provide the main memory (5) required for reading in the one or more data packets (11_1, .., 11_N) with a first address to be written and a last address to be written, and the direct memory access unit being configured for this purpose to store a placeholder different from the identifier value (12_1, ..., 12_N) in the last address of the respective data packet (11_1, .., 11_N) to be written, so that when the one or more data packets (11_1, .., 11_N) are saved, the respective placeholder is only overwritten with the identifier value (12_1, ..., 12_N) of the one or more data packets (11_1, .., 11_N), when the respective data packet (11_1, .., 11_N) belonging to the respective identifier value (12_1, ..., 12_N) is completely stored in the data sink. Control system (15) Claim 6 , characterized in that the data sink itself is designed as a working memory (5). Control system (15) Claim 6 or 7th , characterized in that the direct memory access unit is designed as a DMA controller (1). Control system (15) according to one of the preceding Claims 6 to 8th , characterized in that the control system (15) is arranged on an SOC (System-on-a-Chip) (14). Method for improved loading and further processing of a file which is created on a data source for transmission to a data sink, characterized by the following steps: loading and reading out a file header (13) of a data structure (10) by a DMA controller (1), wherein the data structure (10) comprises one or more data packets (11_1, .., 11_N) each with a different identifier value (12_1, ..., 12_N), the respective identifier value (12_1, ..., 12_N) at the end of the one or several data packets (11_1, .., 11_N) is arranged, - provision of the required storage space for the one or more data packets (11_1, .., 11_N) each with a different identifier value (12_1, ..., 12_N) in a working memory (5) the data sink by the DMA controller (1), - reading in of a respective placeholder corresponding to each of the different identifier values (12_1, ..., 12_N) to the address which is used for the respective different identifier value (12_1 ,. .., 12_N) is provided in the data sink, whereby the respective placeholder differs from the respective identifier value (12_1, ..., 12_N) so that when the one or more data packets (11_1, .., 11_N) are saved, the respective Placeholder is only overwritten with the respective identifier value (12_1, ..., 12_N) of the one or more data packets (11_1, .., 11_N) when the data packet belonging to the respective identifier value (12_1, ..., 12_N) ( 11_1, .., 11_N) vo is permanently stored in the main memory (5) of the data sink made available by the DMA controller (1), - successive loading and reading of the one or more data packets (11_1, .., 11_N) with the different identifier values (12_1, ... , 12_N). Method for improved loading and further processing of a file Claim 10 , characterized by the further step: - Query whether the respective placeholder for the respective identifier value (12_1, ..., 12_N) has already been overwritten with the respective identifier value (12_1, ..., 12_N). Method for improved loading and further processing of a file Claim 10 or 11 , characterized in that when a plurality of data packets (11_1, .., 11_N) are provided, with an overwritten first placeholder by the corresponding first identifier value (12_1, ..., 12_N) of a first data packet (11_1, .., 11_N), the second data packet (11_2, .., 11_N) following the first identifier value (12_1, ..., 12_N) is loaded with a second identifier value (12_2, ..., 12_N), while the first, just loaded data packet (11_1, .., 11_N) is further processed.

Images