A method for cross-domain memory sharing in a multi-core heterogeneous system-on-a-chip.
By combining a layered file system and a communication module, writable access to read-only memory partitions in multi-core heterogeneous system-on-a-chip is achieved, solving the problem of difficult cross-domain data access and improving the timeliness and consistency of data synchronization.
Patent Information
- Application Number
- CN202611611131.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-16
- Publication Date
- 2026-11-17
AI Technical Summary
In multi-core heterogeneous system-on-a-chip (SoC), cross-domain data access is difficult. Existing technologies cannot write or modify read-only partitions, and cannot meet the needs of cross-domain debugging and configuration updates.
By using the layered file system mounting module, the read-only physical storage partition is mounted as the underlying directory and the virtual storage device is mounted as the upper-level directory, generating a merged directory. The communication module transmits data change notifications between domains, enabling cross-domain storage sharing and data synchronization.
It enables timely and proactive reflection of data changes in read-write partitions without modifying the original data in the read-only physical storage partition, thus improving the timeliness and consistency of cross-domain data synchronization.
Smart Images

Figure CN123115050A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of chip technology, and in particular to a multi-core heterogeneous system-on-a-chip and a cross-domain memory sharing method. Background Technology
[0002] Multi-core heterogeneous system-on-chips (SoCs) are increasingly used in fields such as automotive smart cockpits and autonomous driving. These chips typically integrate multiple processing cores with different architectures and are internally divided into multiple independent application domains. Each application domain has its own independent storage partition for storing its own operating system image, applications, and data. Some of these storage partitions are usually configured as read-only to ensure system stability and security. However, this partitioning isolation mechanism also brings difficulties to cross-domain data sharing; the read-only physical storage partition of application domain 1 is only accessible within its own domain, and application domain 2 cannot directly read or modify the data in that read-only partition.
[0003] To address cross-domain data access issues, existing technologies typically employ Network File System (NFS), exposing read-only partitions in application domain 1 to application domain 2 via network links such as Ethernet, in the form of a network shared directory. However, read-only partitions shared via NFS remain read-only in application domain 2, preventing write or modification operations. This fails to meet the needs of application scenarios requiring writes to read-only partitions, such as cross-domain debugging and configuration updates. Summary of the Invention
[0004] This application is made in view of at least one of the above-mentioned technical problems existing in the prior art, and this application enables cross-domain storage sharing.
[0005] In a first aspect, embodiments of this application provide a multi-core heterogeneous system-on-a-chip, including: a first application domain and a second application domain; The first application domain is configured with a read-only physical storage partition, a virtual storage device, a first communication module, and a stacked file system mounting module. The second application domain is configured with a read-write physical storage partition and a second communication module. The physical storage area of the virtual storage device corresponds to the read-write physical storage partition of the second application domain. The layered file system mounting module is used to mount the read-only physical storage partition as the bottom directory and the virtual storage device as the top directory for the layered file system mounting, and generate a merged directory. In response to the first communication module receiving a data change notification, the merged directory is updated. The second communication module is used to send a data change notification to the first communication module via inter-domain communication in response to a data change operation of the second application domain on the readable and writable physical storage partition.
[0006] Secondly, embodiments of this application provide a cross-domain storage sharing method, applied to any of the aforementioned multi-core heterogeneous system-on-a-chip, comprising: The overlay file system mounting module uses the read-only physical storage partition of the first application domain as the underlying directory and the virtual storage device as the upper directory, and generates a merged directory through the overlay file system mounting. In response to the data change operation of the second application domain on the readable and writable physical storage partition, the second communication module sends a data change notification to the first communication module through inter-domain communication; The stacked file system mounting module updates the merged directory in response to the data change notification received by the first communication module.
[0007] This application provides a multi-core heterogeneous system-on-a-chip and a cross-domain storage sharing method. By mapping the physical storage area of a virtual storage device in a first application domain to a read-write physical storage partition in a second application domain, the first application domain can use the storage space of the second application domain through the virtual storage device, achieving cross-domain storage sharing and capacity expansion. A layered file system mounting module is used to mount the read-only physical storage partition as the bottom-level directory and the virtual storage device as the top-level directory, generating a merged directory. The first application domain can obtain the directory content after the bottom-level read-only data and the top-level writable data are superimposed through the merged directory as a unified entry point, thus achieving writable access to the read-only data without modifying the original data of the read-only physical storage partition. A second communication module responds to the second application domain's data change operations on the read-write physical storage partition by sending a data change notification to the first communication module via inter-domain communication. The layered file system mounting module then updates the merged directory in response to this notification, ensuring that data changes in the second application domain are proactively and promptly reflected in the merged directory of the first application domain, improving the timeliness and consistency of cross-domain data synchronization. Attached Figure Description
[0008] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0009] Figure 1 This is a schematic diagram of a multi-core heterogeneous system-on-a-chip provided in one embodiment of this application; Figure 2 This is a flowchart illustrating a cross-domain storage sharing method provided in one embodiment of this application; Figure 3This is a timing diagram of a cross-domain storage sharing method provided in one embodiment of this application. Detailed Implementation
[0010] To enable those skilled in the art to better understand the technical solutions of the embodiments of this application, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0011] like Figure 1 As shown in the figure, this application embodiment provides a multi-core heterogeneous system-on-a-chip, including: a first application domain and a second application domain.
[0012] A multi-core heterogeneous system-on-a-chip (SoC) refers to a system-on-a-chip that integrates at least two processing cores or core clusters with different architectures, instruction sets, or performance and power consumption characteristics within the same chip. For example, a multi-core heterogeneous SoC may include at least two of the following: application processor cores, real-time processor cores, graphics processors, neural network processors, and digital signal processors. The first application domain and the second application domain can run on different processing cores, thus forming independent or isolated operating environments.
[0013] The first and second application domains are independent or isolated operating environments within a multi-core heterogeneous system-on-a-chip (SoC). Each application domain can run an independent operating system instance; for example, the first application domain runs a first operating system, and the second application domain runs a second operating system. The first and second application domains can be isolated from each other through hardware isolation mechanisms, such as memory management units or bus firewalls. These isolation mechanisms ensure that each application domain has its own independent address space, interrupt resources, and memory resources.
[0014] The first application domain is configured with a read-only physical storage partition, a virtual storage device, a first communication module, and a stacked file system mounting module. The second application domain is configured with a read-write physical storage partition and a second communication module. The physical storage area of the virtual storage device corresponds to the read-write physical storage partition of the second application domain.
[0015] A read-only physical storage partition (ROM) is a physical storage area configured within the first application domain that cannot be written to or modified during normal operation. ROMs can use non-volatile storage media, such as embedded multimedia card partitions or general-purpose flash memory partitions. ROMs can store the operating system image and factory configuration data of the first application domain.
[0016] A read-write physical storage partition refers to a physical storage area configured in a second application domain that allows write and modification operations. Read-write physical storage partitions can also use non-volatile storage media, such as embedded multimedia card partitions, general-purpose flash memory partitions, and non-flash memory partitions. Read-write physical storage partitions are used to provide the physical storage backend for subsequent virtual storage devices.
[0017] A virtual storage device is a software-simulated storage device located within a first application domain. To the operating system of the first application domain, the virtual storage device appears as a read-write block or character device, but it does not necessarily possess its own independent physical storage medium. The physical storage area of the virtual storage device corresponds to the read-write physical storage partitions of the second application domain.
[0018] The read-write physical storage partition can be dedicated to cross-domain storage sharing in this embodiment, or it can share the same physical partition with other data in the second application domain.
[0019] The layered file system mounting module is used to mount a read-only physical storage partition as the bottom layer directory and a virtual storage device as the top layer directory to form a layered file system, and to generate a merged directory. In response to the data change notification received by the first communication module, the merged directory is updated.
[0020] The second communication module is used to send a data change notification to the first communication module through inter-domain communication in response to a data change operation of the second application domain on the readable and writable physical storage partition.
[0021] The first communication module is located in the first application domain, and the second communication module is located in the second application domain. The first and second communication modules are used for information exchange through inter-domain communication. Inter-domain communication can be implemented based on the on-chip hardware resources of a multi-core heterogeneous system-on-a-chip.
[0022] The layered file system mounting module is set in the first application domain. It is used to mount the layered file system with the read-only physical storage partition as the bottom directory and the virtual storage device as the top directory, and generate a merged directory.
[0023] A layered file system is a union file system that can stack multiple directories into a unified directory view. The bottom layer is read-only, and the top layer is writable. The merged directory is the unified view presented to users of the first application domain after the bottom and top directories are stacked. The mounting module for a layered file system can be implemented using a layered file system or other union file systems.
[0024] The layered file system mounting module updates the merged directory in response to a data change notification received by the first communication module. Updates may include refreshing the directory entry cache, refreshing the inode cache, and re-performing the directory merge. When the second application domain writes to or modifies a read-write physical storage partition, new data appears in the upper-level directory. After the layered file system mounting module updates the merged directory, users in the first application domain can see the updated file content through the merged directory. When the second application domain deletes data from a read-write physical storage partition, a mask file appears in the upper-level directory. The layered file system mounting module uses this mask file to hide the corresponding original file in the lower-level directory, thus presenting the deleted file's appearance in the merged directory.
[0025] The physical storage area of the virtual storage device corresponds to the read / write physical storage partition of the second application domain, giving the upper-level directory of the first application domain real physical disk space. The overlay file system mounting module mounts the read-only physical storage partition as the bottom-level directory and the virtual storage device as the upper-level directory, enabling the first application domain to simultaneously present read-only raw data and cross-domain write data in the form of a merged directory. The second communication module responds to data change operations in the second application domain by sending a data change notification to the first communication module via inter-domain communication. The overlay file system mounting module responds to the first communication module receiving the data change notification by updating the merged directory. Thus, data changes made by the second application domain to the read / write physical storage partition can be promptly detected by the first application domain and reflected in the merged directory.
[0026] This application embodiment maps the physical storage area of the virtual storage device in the first application domain to the read / write physical storage partition of the second application domain. The first application domain can then utilize the storage space of the second application domain through the virtual storage device, achieving cross-domain storage sharing and capacity expansion. A layered file system mounting module mounts the read-only physical storage partition as the bottom-level directory and the virtual storage device as the top-level directory, generating a merged directory. The first application domain can access the directory content—combining the bottom-level read-only data and the top-level writable data—through the merged directory as a unified entry point, thus enabling writable access to the read-only data without modifying the original data in the read-only physical storage partition. A second communication module responds to data change operations by the second application domain on the read / write physical storage partition by sending a data change notification to the first communication module via inter-domain communication. The layered file system mounting module then updates the merged directory in response to this notification, ensuring that data changes in the second application domain are proactively and promptly reflected in the merged directory of the first application domain, improving the timeliness and consistency of cross-domain data synchronization.
[0027] In one embodiment of this application, both the read-only physical storage partition and the read-write physical storage partition are embedded multimedia card partitions, and the virtual storage device is a virtual embedded multimedia card partition.
[0028] A read-only physical storage partition for an embedded multimedia card partition means dividing the storage space of the embedded multimedia card corresponding to the first application domain into at least one logical partition and configuring the logical partition as read-only.
[0029] The read-write physical storage partition is also an embedded multimedia card partition, that is, at least one logical partition in the embedded multimedia card storage space corresponding to the second application domain, and this partition allows write and modification operations. The read-write physical storage partition serves as the local storage area of the second application domain and can be used to store the operating system data, application data, etc. of the second application domain. In this embodiment, it also serves as the physical storage backend of the virtual storage device of the first application domain.
[0030] Read-write physical storage partitions and read-only physical storage partitions can reside on the same embedded multimedia card or on different embedded multimedia cards. When they are on the same embedded multimedia card, different logical partitions can be configured as read-only and read-write respectively through a partition table; when they are on different embedded multimedia cards, they can be accessed independently through their respective storage controllers.
[0031] In other implementations, the readable and writable physical storage partition can also be replaced by non-volatile storage media such as general-purpose flash storage partitions and non-flash partitions.
[0032] A virtual storage device is a virtual embedded multimedia card partition, which refers to an embedded multimedia card device or partition that is simulated by software and set up in the first application domain.
[0033] Embedded multimedia cards have a mature automotive-grade application foundation, standardized interface protocols, and extensive driver support in the field of in-vehicle intelligent cockpits. By using embedded multimedia card partitions as physical storage media and virtual device types, existing embedded multimedia card driver stacks, block device layers, and overlay file systems in the operating system kernel can be directly reused without developing dedicated storage drivers or file system interfaces, significantly reducing the implementation complexity and development cost of cross-domain storage sharing solutions.
[0034] In one embodiment of this application, inter-domain communication transmits data change notifications via a hardware mailbox and transfers change data via shared memory.
[0035] Hardware mailboxes and shared memory can also employ other collaborative methods. For example, a hardware mailbox can carry the address and length information of a shared memory buffer along with the transmission of notifications, and the first communication module can read the changed data from the shared memory based on this address and length information.
[0036] Hardware mailboxes are specifically designed for transmitting data change notifications. With small data volumes and low transmission latency, they ensure that the first application domain can promptly detect data changes made by the second application domain, improving the real-time nature and determinism of cross-domain notifications. Shared memory is dedicated to transmitting changed data, offering large storage capacity and high read / write bandwidth. This meets the capacity requirements for transmitting changed data, avoiding the efficiency bottlenecks associated with transmitting large amounts of data via hardware mailboxes.
[0037] In one embodiment of this application, in response to a write operation or modification operation of a second application domain on a readable and writable physical storage partition, the second communication module is used to send a data change notification to the first communication module through inter-domain communication. The overlay file system mounting module is used to refresh the view of the upper-level directory to read the changed data in response to data change notifications, and update the merged directory based on the changed data.
[0038] A write operation refers to the second application domain writing new data to a read-write physical storage partition. Write operations can include at least one of the following: creating a new file, writing file content, appending file content, writing directory entries, writing file metadata, and allocating storage blocks. A modification operation refers to the second application domain altering data that already exists in the read-write physical storage partition. Modification operations can include at least one of the following: modifying file content, overwriting existing data, updating file attributes, renaming files or directories, and adjusting directory structure.
[0039] Write or modification operations can be initiated by applications, system services, upgrade programs, etc. in the second application domain.
[0040] Data change notifications may include at least one of the following: change type, indicating whether the operation is a write or modify operation; change location, indicating the physical address, logical block address, partition identifier, file path, or directory entry identifier where the change occurred; change length, indicating the amount of data changed; change timestamp, indicating the time the change occurred; file identifier, indicating the file that was changed; operator identifier, indicating the process or user that initiated the change; and verification information to ensure the integrity of the notification. Data change notifications typically involve small amounts of data, making them suitable for transmission via low-latency inter-domain communication mechanisms.
[0041] Refreshing the view of the upper-level directory refers to the overlay file system mounting module updating or invalidating at least one of the following: the upper-level directory's cache, directory entries, inodes, page cache, and block mapping relationships. This allows the overlay file system mounting module to reread the latest data in the upper-level directory. Since the physical storage area of the virtual storage device corresponds to the read / write physical storage partition of the second application domain, the write or modification operations of the second application domain on the read / write physical storage partition have directly affected the shared physical backend. Therefore, by refreshing the view of the upper-level directory, the overlay file system mounting module can read the changed data that the second application domain has just written or modified.
[0042] Updating the merge directory can include: reflecting the changed data in the directory entries of the merge directory; updating the contents of the corresponding files in the merge directory; updating the attributes of the corresponding files in the merge directory, etc.
[0043] The stacked file system mounting module responds to data change notifications by refreshing the view of the upper-level directory to read the changed data. This ensures that the first application domain reads the latest data from the shared physical backend, avoiding data staleness issues caused by inconsistent caching and improving cross-domain data consistency.
[0044] In one embodiment of this application, in response to a deletion operation of a read-write physical storage partition by a second application domain and the generation of a whitelist file, a second communication module is used to send a whitelist file notification to a first communication module via inter-domain communication. The layered file system mounting module is used to refresh the view of the upper-level directory in response to white-screen file notifications to read the white-screen files, and to mask the files in the corresponding lower-level directories in the merge directory based on the white-screen files.
[0045] A deletion operation refers to the act of a second application domain deleting existing files, directories, directory entries, or other data objects from a read-write physical storage partition. Deletion operations can be initiated by applications, system services, etc., within the second application domain.
[0046] White screen files are special files used in layered file systems to represent deletion semantics. In the mechanism of a layered file system, when a file in a lower-level directory needs to be deleted, the original file in the lower-level directory is not directly modified. Instead, a white screen file corresponding to the file to be deleted is created in the upper-level directory. White screen files are used to mask the corresponding original file in the lower-level directory when merging lower-level and upper-level directories in a layered file system, making the original file invisible in the merged directory, thus achieving the same effect as deletion.
[0047] Whitespace file notifications may include at least one of the following information: change type, indicating that the operation is a deletion operation; original file identifier, indicating the name, path, inode number, or block address of the deleted file; whitespace file identifier, indicating the name, path, or storage location of the generated whitespace file; change location, indicating the physical address, logical block address, or partition identifier of the whitespace file; change timestamp, indicating the time when the deletion operation occurred; operator identifier, indicating the process or user that initiated the deletion operation; and verification information to ensure the integrity of the notification. Whitespace file notifications typically have a small data size, making them suitable for transmission via low-latency inter-domain communication mechanisms.
[0048] Since the masking file only exists in the upper-level directory, while the lower-level directory is still a read-only physical storage partition of the first application domain, the masking operation will not modify the original data in the read-only physical storage partition.
[0049] The overlay file system mounting module responds to whitelist file notifications by refreshing the view of the upper-level directory to read the whitelist file. This ensures that the first application domain reads the latest whitelist file from the shared physical backend, avoiding deletion operation failures due to cache inconsistencies and improving cross-domain data consistency. By using the whitelist file to mask files in the corresponding underlying directories within the merged directory, the first application domain can obtain the latest content after deletion operations through a unified entry point via the merged directory. This achieves synchronous presentation of cross-domain deletions without modifying the original data in the read-only physical storage partition, protecting the security and integrity of the underlying read-only data.
[0050] In one embodiment of this application, the overlay file system mounting module is configured to, in response to a first application domain's read operation on the merged directory, search for target data in the upper-level directory, and, in response to the failure to read the target data from the upper-level directory, read the target data from the lower-level directory.
[0051] A merged directory is a unified directory generated by the overlay file system mount module by combining the underlying and upper-level directories. Applications, system services, or users in the first application domain can initiate read operations on the merged directory through the file system interface. Read operations can include at least one of the following: opening a file, reading file content, reading a directory list, querying file attributes, querying directory entries, and performing a file search.
[0052] The upper-level directory is a virtual storage device. Since the physical storage area of the virtual storage device corresponds to the read / write physical storage partition of the second application domain, the upper-level directory may contain the latest data written or modified across domains by the second application domain. Therefore, after receiving a read operation on the merged directory, the overlay file system mount module first searches for the target data in the upper-level directory.
[0053] Searching for target data in the parent directory can include: searching for the filename or directory name corresponding to the target data in the directory entries of the parent directory; searching for the inode corresponding to the target data in the inode of the parent directory; searching for the data block corresponding to the target data in the page cache or block cache of the parent directory; and searching for the attribute information corresponding to the target data in the extended attributes of the parent directory.
[0054] When the overlay file system mount module does not find the target data in the upper-level directory, it indicates that the target data has not been written or modified across domains by the second application domain, or that the target data does not have a corresponding blank file in the upper-level directory. In this case, the overlay file system mount module reads the target data from the lower-level directory.
[0055] The underlying directory is a read-only physical storage partition of the first application domain. This underlying directory stores the original data of the first application domain, including the operating system image and factory configuration data. Because the underlying directory is read-only, the overlay file system mount module does not modify the original data in the underlying directory when reading target data from it.
[0056] By prioritizing the search for target data in the upper-level directory, the first application domain can obtain the latest data written or modified across domains by the second application domain, ensuring that the data presented in the merged directory is consistent with the latest state in the shared physical backend, thus improving the timeliness and accuracy of cross-domain data reading.
[0057] In one embodiment of this application, a first application domain runs multiple operating systems, and a read-only physical storage partition includes multiple read-only physical sub-partitions corresponding to the multiple operating systems.
[0058] Running multiple operating systems in the first application domain refers to running two or more operating system instances simultaneously or in a time-sharing manner within the same application domain. These multiple operating systems can be of the same type or different types.
[0059] A read-only physical storage partition comprises multiple read-only physical sub-partitions, each corresponding to a different operating system. This refers to physically or logically dividing the read-only physical storage partition configured in the first application domain into multiple sub-partitions, each corresponding to a different operating system. Each read-only physical sub-partition stores data that does not need to be modified during runtime, such as the system image, root file system, and factory configuration data of its corresponding operating system.
[0060] The correspondence between multiple read-only physical subpartitions and multiple operating systems can be one-to-one, meaning one operating system corresponds to one read-only physical subpartition. One operating system can also correspond to multiple read-only physical subpartitions; for example, one operating system might correspond to one system image subpartition and one static resource subpartition. Alternatively, multiple operating systems can share different directories within the same read-only physical subpartition.
[0061] This application embodiment includes multiple read-only physical sub-partitions corresponding to multiple operating systems in the read-only physical storage partition, so that each operating system has an independent read-only storage space, avoiding mutual interference between multiple operating systems due to sharing the same read-only partition, and improving the independence and security of multi-system operation.
[0062] like Figure 2 As shown, this application provides a cross-domain storage sharing method, applied to any of the above-mentioned multi-core heterogeneous system-on-a-chip, the method comprising: Step 201: The overlay file system mounting module uses the read-only physical storage partition of the first application domain as the underlying directory and the virtual storage device as the upper-level directory, and generates a merged directory by mounting the overlay file system.
[0063] The layered file system mounting module mounts underlying and upper-level directories together using a layered file system. A layered file system is a union file system that can stack multiple directories into a unified directory view. The underlying directory is read-only, and the upper-level directory is writable. The merged directory provides a unified view presented to users of the first application domain after the underlying and upper-level directories are stacked.
[0064] Step 202: In response to the data change operation of the second application domain on the readable and writable physical storage partition, the second communication module sends a data change notification to the first communication module through inter-domain communication.
[0065] Data modification operations refer to data modification actions performed by the second application domain on a read-write physical storage partition. Data modification operations can include at least one of write, modify, or delete operations. Write operations can include creating new files, writing file content, appending file content, writing directory entries, and writing file metadata. Modification operations can include modifying file content, overwriting existing data, updating file attributes, and renaming files or directories. Deletion operations can include deleting files, deleting directories, and deleting directory entries.
[0066] Step 203: The stacked file system mounting module updates the merged directory in response to the data change notification received by the first communication module.
[0067] Updating the merged directory can include: refreshing the view of the parent directory; reading the changed data in the parent directory; re-performing the merge of the lower and upper directories based on the changed data; and updating the directory entries, file content, file attributes, or directory structure of the merged directory.
[0068] This application embodiment uses a layered file system mounting module to mount the read-only physical storage partition of the first application domain as the bottom-level directory and the virtual storage device as the top-level directory, generating a merged directory. This allows the first application domain to access the directory content, which combines the bottom-level read-only data and the top-level writable data, through the merged directory as a unified entry point. This enables writable access to the read-only data without modifying the original data in the read-only physical storage partition. In response to data change operations on the read-write physical storage partition by the second application domain, the second communication module proactively sends a data change notification to the first communication module via inter-domain communication. This allows the first application domain to promptly learn of cross-domain data changes without polling, improving the real-time nature and proactivity of cross-domain storage sharing.
[0069] In one embodiment of this application, the data change operation is a write operation or a modification operation; The overlay file system mounting module responds to the data change notification received by the first communication module by updating the merged directory, including: In response to data change notifications, the overlay file system mount module refreshes the view of the parent directory to read the changed data and updates the merged directory based on the changed data.
[0070] Data change operations, also known as write or modify operations, refer to the data change behavior performed by the second application domain on a readable and writable physical storage partition. Specifically, this means writing new data to the partition or modifying existing data in the partition.
[0071] Because the stacked file system merges directories in the order of upper-level directories first and lower-level directories second when reading the merged directory, the updated merged directory can automatically reflect the changes when the data in the upper-level directory changes.
[0072] In this embodiment, the overlay file system mounting module responds to data change notifications by refreshing the view of the upper-level directory to read the changed data. This ensures that the first application domain reads the latest data in the shared physical backend, avoiding data staleness issues caused by inconsistent caching and improving cross-domain data consistency.
[0073] In one embodiment of this application, in response to a data change operation by a second application domain on a readable and writable physical storage partition, the second communication module sends a data change notification to the first communication module via inter-domain communication, including: In response to the second application domain performing a deletion operation on the readable and writable physical storage partition and generating a whitelist file, the second communication module sends a whitelist file notification to the first communication module through inter-domain communication. The overlay file system mounting module responds to the data change notification received by the first communication module by updating the merged directory, including: The overlay file system mount module responds to the white screen file notification by refreshing the view of the upper-level directory to read the white screen file, and then masks the files in the corresponding lower-level directory in the merge directory based on the white screen file.
[0074] A deletion operation refers to the act of a second application domain deleting existing files, directories, directory entries, or other data objects in a read-write physical storage partition.
[0075] The second communication module sends a whitelist file notification, not the whitelist file itself, to the first communication module via inter-domain communication. Since the whitelist file has already been generated in a readable and writable physical storage partition of the second application domain, and this partition corresponds to the physical storage area of the virtual storage device, the whitelist file is physically located in the shared storage backend corresponding to the virtual storage device. Upon receiving the whitelist file notification, the first communication module only needs to trigger the overlay file system mount module to refresh the view of the upper-level directory to read the whitelist file.
[0076] The overlay file system mount module responds to whitelist file notifications by refreshing the view of the upper-level directory to read the whitelist file. This ensures that the first application domain reads the latest whitelist file in the shared physical backend, avoiding deletion operation failures caused by cache inconsistencies and improving cross-domain data consistency.
[0077] like Figure 3 The diagram shown is a timing diagram of a cross-domain storage sharing method provided in an embodiment of this application. The multi-core heterogeneous system-on-a-chip adopts a multi-core heterogeneous architecture. Application domain A runs an instrument operating system, and application domain B runs an entertainment and debugging operating system. Application domain A is configured with a read-only physical storage partition, a virtual storage device, a first communication module, and an overlay file system mounting module. The read-only physical storage partition is an embedded multimedia card partition, including two read-only sub-partitions, EMMC_A_1 and EMMC_A_2, which store the system image and static resource files of the instrument operating system, respectively. The virtual storage device is a virtual embedded multimedia card partition, denoted as v_emmc_a_overlay; the first communication module is used for inter-domain communication with application domain B; and the overlay file system mounting module is used to perform overlay file system mounting.
[0078] Application domain B is configured with a read / write physical storage partition and a second communication module. The read / write physical storage partition, denoted as emmc_b_overlay, is an embedded multimedia card partition used to store data for entertainment and debugging the operating system, and serves as the physical storage backend for the virtual storage device of application domain A. The second communication module is used for inter-domain communication with application domain A. The physical storage area of the virtual storage device v_emmc_a_overlay corresponds to that of application domain B's emmc_b_overlay, and both share the same physical storage backend.
[0079] Inter-domain communication between application domain A and application domain B is implemented using a hardware mailbox and shared memory on a multi-core heterogeneous system-on-a-chip. The hardware mailbox is used to transmit notification messages, and the shared memory is used to store changed data or change metadata.
[0080] During system startup, the overlay file system mounting module of application domain A performs the overlay file system mounting operation. Specifically, the overlay file system mounting module uses the read-only physical storage partitions of application domain A, namely EMMC_A_1 and EMMC_A_2, as the bottom-level directory and the virtual storage device v_emmc_a_overlay as the top-level directory, generating a merged directory through the overlay file system mounting. The merged directory is a unified directory view after the bottom-level and top-level directories are superimposed. Applications in application domain A access files through this merged directory. `lowdir` is an abbreviation for lower directory. `A_Low` is the abbreviation for the read-only physical storage partition of application domain A in the diagram. `updir` is an abbreviation for upper directory. `A_Disk` is the abbreviation for the virtual storage device v_emmc_a_overlay of application domain A in the diagram. The `.wh` file is the whitelist file.
[0081] Since the physical storage area of the virtual storage device v_emmc_a_overlay corresponds to emmc_b_overlay in application domain B, the physical disk location of the upper-level directory is actually located in emmc_b_overlay in application domain B. When application domain A reads files through the merged directory, the overlay file system mount module first searches for the target data in the upper-level directory; if the target data is not found in the upper-level directory, it reads the target data from the lower-level directory.
[0082] When debugging or configuration updates of the read-only partition of application domain A are required through application domain B, application domain B performs write or modification operations on the read-write physical storage partition emmc_b_overlay. For example, the debugging tools of application domain B create new files or modify the contents of existing files in emmc_b_overlay.
[0083] In response to the write or modify operation, the second communication module of application domain B generates a data change notification and sends it to the first communication module of application domain A via inter-domain communication. Specifically, the second communication module writes the data change notification to a hardware mailbox and triggers the hardware mailbox to send an interrupt signal to application domain A; the first communication module responds to the interrupt signal and reads the data change notification from the hardware mailbox. The data change notification includes information such as change type, change location, and change length.
[0084] After receiving a data change notification, the first communication module of application domain A triggers the overlay file system mounting module to perform an update operation. In response to this data change notification, the overlay file system mounting module refreshes the view of the upper-level directory to read the changed data that application domain B has just written or modified. Since the virtual storage devices v_emmc_a_overlay and emmc_b_overlay share the same physical storage backend, the write or modification operations of application domain B have directly affected this shared physical backend. Therefore, the overlay file system mounting module can read the changed data simply by refreshing the view of the upper-level directory.
[0085] Subsequently, the overlay file system mount module updates the merged directory based on the read change data. In the updated merged directory, users in application domain A can see the file contents written or modified by application domain B. The original data in the read-only physical storage partitions EMMC_A_1 and EMMC_A_2 of application domain A remains unchanged.
[0086] When a file needs to be deleted from the read-only partition of application domain A via application domain B, application domain B performs the deletion operation on the read-write physical storage partition emmc_b_overlay. In response to this deletion operation, application domain B generates a white screen file in emmc_b_overlay. The white screen file is used to mask the corresponding original file in the underlying directory during the merging of overlay file systems.
[0087] The second communication module of application domain B generates a whitelist file notification and sends it to the first communication module of application domain A via inter-domain communication. Specifically, the second communication module writes the whitelist file notification to a hardware mailbox and triggers the hardware mailbox to send an interrupt signal to application domain A; the first communication module responds to the interrupt signal and reads the whitelist file notification from the hardware mailbox. The whitelist file notification includes information such as the original file identifier, the whitelist file identifier, and the changed location.
[0088] Upon receiving the whiteout file notification, the first communication module of application domain A triggers the overlay file system mounting module to perform an update operation. In response to the whiteout file notification, the overlay file system mounting module refreshes the view of the upper-level directory to read the whiteout file just generated by application domain B. Since the virtual storage devices v_emmc_a_overlay and emmc_b_overlay share the same physical storage backend, the whiteout file already exists in the shared physical backend, and the overlay file system mounting module can read the whiteout file by refreshing the view of the upper-level directory.
[0089] Subsequently, the overlay file system mounting module uses this masking file to block files in the corresponding underlying directories within the merged directory. Specifically, when the overlay file system merges underlying and upper-level directories, if a masking file exists in the upper-level directory corresponding to a file in the underlying directory, the overlay file system will not display the original file in that underlying directory in the merged directory, or will mark the original file as deleted. Thus, users in application domain A will see that the file has been deleted through the merged directory, but the original file in the underlying directories EMMC_A_1 and EMMC_A_2 still physically exists and has not been modified. If the masking file is subsequently deleted, the original file in the underlying directory can be re-displayed in the merged directory.
[0090] When an application in application domain A reads target data through the merged directory, the overlay file system mount module responds to the read operation by first searching for the target data in the upper-level directory. If the target data is found in the upper-level directory, it is returned; otherwise, it is read from the lower-level directory and returned. Through this read mechanism, application domain A can use the merged directory as a unified entry point to simultaneously access both read-only data at the lower level and writable data at the upper level.
[0091] This embodiment of the cross-domain storage sharing method utilizes the overlay file system mounting module of application domain A to mount a read-only physical storage partition as the underlying directory and a virtual storage device as the upper-level directory, generating a merged directory. Combined with data modification operations on the read-write physical storage partition by application domain B and inter-domain communication between the second and first communication modules, this enables cross-domain write, modification, and deletion of the read-only partition of application domain A by application domain B. Throughout the process, the original data in the read-only physical storage partition of application domain A remains unmodified, and data changes by application domain B are promptly reflected in the merged directory of application domain A through a shared physical backend and inter-domain notification mechanism. This method does not rely on external network protocol stacks and physical network connections, offering advantages such as low latency, high security, and ease of deployment. It is particularly suitable for scenarios involving debugging, configuration updates, and storage expansion across multiple domains in in-vehicle intelligent cockpits.
[0092] This application's multi-core heterogeneous system-on-a-chip and its cross-domain storage sharing method can also be applied to the field of embodied intelligence. Embodied intelligence systems typically need to integrate multiple functions such as perception, decision-making, planning, control, and artificial intelligence reasoning on the same system-on-a-chip, and tend to adopt a multi-core heterogeneous architecture to divide different functions into mutually isolated application domains. Each application domain runs an independent operating system and is configured with independent physical storage partitions. Using the technical solution of this application, a correspondence between virtual storage devices and physical storage partitions can be established between multiple application domains of embodied intelligence devices. By using a layered file system, read-only physical storage partitions are mounted as the bottom-level directory and virtual storage devices are mounted as the upper-level directory to generate a merged directory. Data change notifications are transmitted through on-chip inter-domain communication mechanisms, enabling one application domain to share read-only data of another application domain in a writable manner without relying on external networks or modifying the original read-only partitions. Therefore, this application can meet the needs of low-latency, high-security, and isolated cross-domain data sharing, model or configuration updates, debugging, and storage expansion between multiple domains in embodied intelligence devices, and is particularly suitable for domain controllers or main control chips of embodied intelligence devices such as humanoid robots, mobile robots, robotic arms, and drones. Those skilled in the art can adaptively configure the number of application domains, storage partitioning, mounting method of stacked file systems, and inter-domain communication mechanism according to specific intelligent application scenarios.
[0093] In this application, "vehicle" can refer to "automobile," "vehicle," or "complete vehicle," or other similar terms, including general motor vehicles such as sedans, SUVs, MPVs, buses, trucks, and other freight or passenger vehicles; water transport vehicles including various boats and vessels; and aircraft, including hybrid vehicles, electric vehicles, gasoline vehicles, plug-in hybrid vehicles, fuel cell vehicles, and other alternative fuel vehicles. Hybrid vehicles refer to vehicles with two or more power sources, and electric vehicles include pure electric vehicles and range-extended electric vehicles; this application does not specifically limit their use.
[0094] It should be understood that in the embodiments of this application, the processor may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.
[0095] It should also be understood that the memory mentioned in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. Specifically, non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DR RAM).
[0096] It should be noted that when the processor is a general-purpose processor, DSP, ASIC, FPGA, or other programmable logic device, discrete gate or transistor logic device, or discrete hardware component, the memory (storage module) is integrated into the processor.
[0097] It should be noted that the memories described herein are intended to include, but are not limited to, these and any other suitable types of memories.
[0098] In addition to the data bus, this bus may also include a power bus, a control bus, and a status signal bus. However, for clarity, all buses are labeled "bus" in the diagram.
[0099] It should also be understood that the first, second, third, fourth and various numerical designations used herein are merely for descriptive convenience and are not intended to limit the scope of this application.
[0100] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.
[0101] In implementation, each step of the above method can be completed by integrated logic circuits in the processor's hardware or by instructions in software. The steps of the method disclosed in the embodiments of this application can be directly implemented by a hardware processor, or by a combination of hardware and software modules in the processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method. To avoid repetition, detailed descriptions are omitted here.
[0102] In the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.
[0103] Those skilled in the art will recognize that the various illustrative logical blocks (ILBs) and steps described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.
[0104] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.
[0105] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0106] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.
[0107] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive), etc.
[0108] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
Claims
1. A multi-core heterogeneous system-on-a-chip, characterized in that, include: First application domain and second application domain; The first application domain is configured with a read-only physical storage partition, a virtual storage device, a first communication module, and a stacked file system mounting module. The second application domain is configured with a read-write physical storage partition and a second communication module. The physical storage area of the virtual storage device corresponds to the read-write physical storage partition of the second application domain. The layered file system mounting module is used to mount the read-only physical storage partition as the bottom directory and the virtual storage device as the top directory for the layered file system mounting, and generate a merged directory. In response to the first communication module receiving a data change notification, the merged directory is updated. The second communication module is used to send a data change notification to the first communication module via inter-domain communication in response to a data change operation of the second application domain on the readable and writable physical storage partition.
2. The multi-core heterogeneous system-on-a-chip as described in claim 1, characterized in that, in, Both the read-only physical storage partition and the read-write physical storage partition are embedded multimedia card partitions, and the virtual storage device is a virtual embedded multimedia card partition.
3. The multi-core heterogeneous system-on-a-chip as described in claim 1, characterized in that, in, The inter-domain communication transmits the data change notification via a hardware mailbox and transfers the changed data via shared memory.
4. The multi-core heterogeneous system-on-a-chip as described in claim 1, characterized in that, In response to a write or modification operation of the second application domain on a readable and writable physical storage partition, the second communication module is configured to send a data change notification to the first communication module through the inter-domain communication. The layered file system mounting module is used to refresh the view of the upper-level directory to read the changed data in response to the data change notification, and update the merged directory based on the changed data.
5. The multi-core heterogeneous system-on-a-chip as described in claim 1, characterized in that, In response to the second application domain's deletion operation on the readable and writable physical storage partition and the generation of a whitelist file, the second communication module is used to send a whitelist file notification to the first communication module through the inter-domain communication. The layered file system mounting module is used to refresh the view of the upper-level directory to read the white-screen file in response to the white-screen file notification, and to block the files in the merged directory corresponding to the lower-level directory based on the white-screen file.
6. The multi-core heterogeneous system-on-a-chip according to claim 1, characterized in that, The stacked file system mounting module is configured to, in response to the first application domain's read operation on the merged directory, search for target data in the upper-level directory, and, in response to the failure to read the target data in the upper-level directory, read the target data from the lower-level directory.
7. The multi-core heterogeneous system-on-a-chip according to claim 1, characterized in that, in, The first application domain runs multiple operating systems, and the read-only physical storage partition includes multiple read-only physical sub-partitions corresponding to the multiple operating systems.
8. A cross-domain storage sharing method, characterized in that, Applied to the multi-core heterogeneous system-on-a-chip according to any one of claims 1-7, comprising: The overlay file system mounting module uses the read-only physical storage partition of the first application domain as the underlying directory and the virtual storage device as the upper directory, and generates a merged directory through the overlay file system mounting. In response to the data change operation of the second application domain on the readable and writable physical storage partition, the second communication module sends a data change notification to the first communication module through inter-domain communication; The stacked file system mounting module updates the merged directory in response to the data change notification received by the first communication module.
9. The method as described in claim 8, characterized in that, The data change operation is either a write operation or a modification operation; The overlay file system mounting module updates the merged directory in response to the data change notification received by the first communication module, including: In response to the data change notification, the layered file system mounting module refreshes the view of the upper-level directory to read the changed data and updates the merged directory based on the changed data.
10. The method as described in claim 8, characterized in that, In response to a data change operation by the second application domain on the readable and writable physical storage partition, the second communication module sends a data change notification to the first communication module via inter-domain communication, including: In response to the second application domain performing a deletion operation on the readable and writable physical storage partition and generating a whitelist file, the second communication module sends a whitelist file notification to the first communication module through inter-domain communication. The overlay file system mounting module updates the merged directory in response to the data change notification received by the first communication module, including: In response to the white screen file notification, the layered file system mounting module refreshes the view of the upper-level directory to read the white screen file, and blocks the files in the merged directory corresponding to the lower-level directory based on the white screen file.
