- 第一章 SystemVerilog导论
- 第二章 文本值
- 第三章 数据类型
- 第四章 数组
- 第五章 数据声明
- 第六章 属性
- 第七章 操作符与表达式
- 第八章 过程语句和控制流
- 第九章 进程
- 第十章 任务与函数
- 第十一章 类
- 第十二章 随机约束
- 第十三章 进程间的同步与通信
- 第十四章 调度语义
- 第十五章 时钟控制块
- 第十六章 程序块
- 第十七章 断言
- 第十八章 层次
- 第十九章 接口
- 第二十章 覆盖
- 第二十一章 参数
- 第二十二章 配置库
- 第二十三章 系统任务与系统函数
- 23.1 简介(一般信息)
- 23.2 确立时的typeof函数
- 23.3 typename函数
- 23.4 表达式尺寸系统函数
- 23.5 范围系统函数
- 23.6 Shortreal转换
- 23.7 数组查询系统函数
- 23.8 断言严重性系统任务
- 23.9 断言控制系统任务
- 23.10 断言系统函数
- 23.11 随机数系统函数
- 23.12 程序控制
- 23.13 覆盖系统函数
- 23.14 对Verilog-2001系统任务的增强
- 23.15 $readmemb与$readmemh
- 23.16 $writememb and $writememh
- 23.17 File format considerations for multi-dimensional unpacked arrays
- 23.18 System task arguments for multi-dimensional unpacked arrays
- 第二十四章 VCD数据
- 第二十五章 编译器指令
- 第二十六章 考虑从SystemVerilog中删除的功能
- 第二十七章 直接编程接口(DPI)
- 27.1 概述
- 27.2 Two layers of the DPI
- 27.3 Global name space of imported and exported functions
- 27.4 导入的任务和函数
- 27.5 Calling imported functions
- 27.6 Exported functions
- 27.7 Exported tasks
- 27.8 Disabling DPI tasks and functions
- 第二十八章 SystemVerilog断言API
- 第二十九章 SystemVerilog覆盖API
- 29.1 需求
- 29.2 SystemVerilog real-time coverage access
- 29.3 FSM recognition
- 29.3.1 Specifying the signal that holds the current state
- 29.3.2 Specifying the part-select that holds the current state
- 29.3.3 Specifying the concatenation that holds the current state
- 29.3.4 Specifying the signal that holds the next state
- 29.3.5 Specifying the current and next state signals in the same declaration
- 29.3.6 Specifying the possible states of the FSM
- 29.3.7 Pragmas in one-line comments
- 29.3.8 Example
- 29.4 VPI coverage extensions
- 第三十章 SystemVerilog数据读API
- 30.1 简介(一般信息)
- 30.2 需求
- 30.3 Extensions to VPI enumerations
- 30.4 VPI object type additions
- 30.5 Object model diagrams
- 30.6 Usage extensions to VPI routines
- 30.7 VPI routines added in SystemVerilog
- 30.8 Reading data
- 30.9 Optionally unloading the data
- 30.10 Reading data from multiple databases and/or different read library providers
- 30.11 VPI routines extended in SystemVerilog
- 30.12 VPI routines added in SystemVerilog
- 30.12.1 VPI reader routines
- 第三十一章 SystemVerilog VPI Object Model
- 31.1 简介(一般信息)
- 31.2 Instance
- 31.3 Interface
- 31.4 Program
- 31.5 Module (supersedes IEEE 1364-2001 26.6.1)
- 31.6 Modport
- 31.7 Interface tf decl
- 31.8 Ports (supersedes IEEE 1364-2001 26.6.5)
- 31.9 Ref Obj
- 31.9.1 Examples
- 31.10 Variables (supersedes IEEE 1364-2001 section 26.6.8)
- 31.11 Var Select (supersedes IEEE 1364-2001 26.6.8)
- 31.12 Typespec
- 31.13 Variable Drivers and Loads (supersedes IEEE 1364-2001 26.6.23)
- 31.14 Instance Arrays (supersedes IEEE 1364-2001 26.6.2)
- 31.15 Scope (supersedes IEEE 1364-2001 26.6.3)
- 31.16 IO Declaration (supersedes IEEE 1364-2001 26.6.4)
- 31.17 Clocking Block
- 31.18 Class Object Definition
- 31.19 Constraint, constraint ordering, distribution,
- 31.20 Constraint expression
- 31.21 Class Variables
- 31.22 Structure/Union
- 31.23 Named Events (supersedes IEEE 1364-2001 26.6.11)
- 31.24 Task, Function Declaration (supersedes IEEE 1364-2001 26.6.18)
- 31.25 Alias Statement
- 31.25.1 Examples
- 31.26 Frames (supersedes IEEE 1364-2001 26.6.20)
- 31.27 Threads
- 31.28 tf call (supersedes IEEE 1364-2001 26.6.19)
- 31.29 Module path, path term (supersedes IEEE 1364-2001 26.6.15)
- 31.30 Concurrent assertions
- 31.31 Property Decl
- 31.32 Property Specification
- 31.33 Multiclock Sequence Expression
- 31.34 Sequence Declaration
- 31.35 Sequence Expression
- 31.36 Attribute (supersedes IEEE 1364-2001 26.6.42)
- 31.37 Atomic Statement (supersedes IEEE 1364-2001 26.6.27)
- 31.38 If, if else, return, case, do while (supersedes IEEE 1364-2001 26.6.35, 26.6.36)
- 31.39 waits, disables, expect, foreach (supersedes IEEE 1364 26.6.38)
- 31.40 Simple expressions (supersedes IEEE 1364-2001 26.6.25)
- 31.41 Expressions (supersedes IEEE 1364-2001 26.6.26)
- 31.42 Event control (supersedes IEEE 1364-2001 26.6.30)
- 31.43 Event stmt (supersedes IEEE 1364-2001 26.6.27)
- 31.44 Process (supersedes IEEE 1364-2001 26.6.27)
- 31.45 Assignment (supersedes IEEE 1364-2001 26.6.28)
- 附录A 形式语法
- A.1 源文本
- A.2 声明
- A.3 Primitive instances
- A.4 Module, interface and generated instantiation
- A.5 UDP declaration and instantiation
- A.6 Behavioral statements
- A.6.1 Continuous assignment and net alias statements
- A.6.2 Procedural blocks and assignments
- A.6.3 Parallel and sequential blocks
- A.6.4 Statements
- A.6.5 Timing control statements
- A.6.6 Conditional statements
- A.6.7 Case statements
- A.6.8 Looping statements
- A.6.9 Subroutine call statements
- A.6.10 Assertion statements
- A.6.11 Clocking block
- A.6.12 Randsequence
- A.7 Specify section
- A.8 Expressions
- A.9 General
- A.10 Footnotes (normative)
- 附录B 关键字
- 附录C 标准包
- 附录D 链表
- 附录E DPI C-layer
- E.1 概述
- E.2 Naming conventions
- E.3 Portability
- E.4 Include files
- E.5 Semantic constraints
- E.6 Data types
- E.7 Argument passing modes
- E.8 Context tasks and functions
- E.9 Include files
- E.10 Arrays
- E.11 Open arrays
- E.11.1 Actual ranges
- E.11.2 Array querying functions
- E.11.3 Access functions
- E.11.4 Access to the actual representation
- E.11.5 Access to elements via canonical representation
- E.11.6 Access to scalar elements (bit and logic)
- E.11.7 Access to array elements of other types
- E.11.8 Example 4— two-dimensional open array
- E.11.9 Example 5 — open array
- E.11.10 Example 6 — access to packed arrays
- E.11.11 Example 7 — binary compatible calls of exported functions
- 附录F 包含文件
- 附录G 包含外部语言代码
- 附录H 并发断言的形式语义
- 附录I svvpiuser.h
- 附录J 术语表
- 附录K 参考书目
- 其他
30.8.1 VPI read initialization and load access initialization
Selecting an object is done in 3 steps:
1) The first step is to initialize the read access with a call to vpi_load_extension() to load the reader extension and set:
a) Name of the reader library to be used specified as a character string. This is either a full pathname to this library or the single filename (without path information) of this library, assuming a vendor specific way of defining the location of such a library. The latter method is more portable and therefore recommended. Neither the full pathname, nor the single filename shall include an extension, the name of the library must be unique and the appropriate extension for the actual platform should be provided by the application loading this library More details are in Section 30.10.
b) Name of the database holding the stored data or flush database in case of vpiAccessPostProcess or vpiAccessInteractive respectively; a NULL can be used in case of vpiAccessLimitedInteractive. This is the logical name of a database, not the name of a file in the file system. It is implementation dependent whether there is any relationship to an actual on-disk object and the provided name. See access mode below for more details on the access modes.
c) Access mode: The following VPI properties set the mode of access
— vpiAccessLimitedInteractive: Means that the access shall be done for the data stored in the tool memory (e.g. simulator), the history (or future) that the tool stores is implementation dependent. If the tool does not store the requested info then the querying routines shall return a fail. The database name argument to vpi_load_extension() in this mode shall be ignored (even if not NULL).
— vpiAccessInteractive: Means that the access shall be done interactively. The tool shall then use the database specified as a “flush” area for its data. This mode is very similar to the vpiAccessLimitedInteractive with the additional requirement that all the past history (before current time) shall be stored (for the specified scope/collection, see the access scope/collection description of vpi_load_init().
— vpiAccessPostProcess: Means that the access shall be done through the specified database. All data queries shall return the data stored in the specified database. Data history depends on what is stored in the database, and can be nothing (i.e. no data).
vpi_load_extension() can be called multiple times for different reader interface libraries (coming from different tools), database specification, and/or read access. A call with vpiAccessInteractive means that the user is querying the data stored inside the simulator database and uses the VPI routines supported by the simulator. A call with vpiAccessPostProcess means that the user is accessing the data stored in the database and uses the VPI services provided by the waveform tool. The application, if accessing several databases and/or using multiple read API libraries, can use the routine vpi_get(vpiBelong, <vpiHandle>) to check whether a handle belongs to that database. The call is performed as follows:
reader_extension_ptr->vpi_get(vpiBelong, <vpiHandle>);
where reader_extension_ptr is the reader library pointer returned by the call to vpi_load_extension(). TRUE is returned if the passed handle belongs to that extension, and FALSE otherwise.
If the application uses the built-in library (i.e. the one provided by the tool it is running under), there is no need to use indirection to call the VPI routines; they can be called directly. An initial call must however be made to set the access mode, specify the database, and check for error indicated by a NULL return.
vpi_close() shall be called in case of:
- vpiAcessLimitedInteractive to perform any tool cleanup. The validity of VPI handles after this call is left up to the particular reader implementation.
- vpiAccessPostProcess or vpiAccessInteractive mode to perform any tool cleanup and close the opened database. Handles obtained before the call to vpi_close() are no longer valid after this call.
2) Next step is to specify the elements that shall be accessed. This is accomplished by calling vpi_load_init() and specifying a scope and/or an item collection. At least one of the two (scope or collection) needs to be specified. If both are specified then the union of all the object elements forms the entire set of objects the user can access.
- Access scope: The specified scope handle, and nesting mode govern the scope that access returns. Data queries outside this scope (and its sub-scopes as governed by the nesting mode) shall return a fail in the access routines unless the object belongs to access collection described below. It can be used either in a complementary or in an exclusive fashion to access collection. NULL is to be passed to the collection when access scope is used in an exclusive fashion.
- Access collection: The specified collection stores the traverse object handles to be loaded. It can be used either in a complementary or in an exclusive fashion to access scope. NULL is to be passed to the scope when access collection is used in an exclusive fashion.
如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱,如果其他用户或者作者回复了您的评论,将不能在第一时间通知您!
发布评论