摘要:
A method of avoiding unnecessary context switching in a multithreaded environment. A thread of execution of a process waiting on a lock protecting access to a shared resource may wait for the lock to be released by executing in a loop, or “spin”. The waiting thread may continuously check, in a user mode of an operating system, an indicator of whether the lock has been released. After a certain time period, the thread may stop spinning and enter a kernel mode of the operating system. Subsequently, before going to sleep which entails costly context switching, the thread may perform an additional check of the indicator to determine whether the lock has been released. If this is the case, the thread returns to user mode and the unnecessary context switching is avoided.
摘要:
A method of avoiding unnecessary context switching in a multithreaded environment. A thread of execution of a process waiting on a lock protecting access to a shared resource may wait for the lock to be released by executing in a loop, or “spin”. The waiting thread may continuously check, in a user mode of an operating system, an indicator of whether the lock has been released. After a certain time period, the thread may stop spinning and enter a kernel mode of the operating system. Subsequently, before going to sleep which entails costly context switching, the thread may perform an additional check of the indicator to determine whether the lock has been released. If this is the case, the thread returns to user mode and the unnecessary context switching is avoided.
摘要:
A parallel processing method and apparatus for initializing libraries is disclosed. Libraries for an application are identified, an initialization order for the libraries is determined, and the libraries are initialized in asynchronous stages. The initialization order is determined by forming a library tree of the libraries' references and determining a load order for the references according to the levels of the references in the library tree. The asynchronous stages comprise a loading stage that includes a load queue, a snapping stage that includes a snap queue, and an initializing stage that includes an initialize queue.
摘要:
Deterministically resolving cycles in a library tree is disclosed. Resolving cycles supports certain processes such as safe library initialization. Cycles in the library tree are identified; at least one soft link in each identified cycle is identified; and the at least one soft link in each identified cycle is broken. If a cycle has no soft links, notification is provided indicating that the cycle cannot be broken. Identifying at least one soft link in each identified cycle comprises, for each link in the cycle, determining the dependent and supporting libraries; and determining if one or more functions in the supporting library are required for initializing the dependent library.
摘要:
Aspects of the present invention are directed at providing safe and efficient ways for a program to perform a one-time initialization of a data item in a multi-threaded environment. In accordance with one embodiment, a method is provided that allows a program to perform a synchronized initialization of a data item that may be accessed by multiple threads. More specifically, the method includes receiving a request to initialize the data item from a current thread. In response to receiving the request, the method determines whether the current thread is the first thread to attempt to initialize the data item. If the current thread is the first thread to attempt to initialize the data item, the method enforces mutual exclusion and blocks other attempts to initialize the data item made by concurrent threads. Then, the current thread is allowed to execute program code provided by the program to initialize the data item.
摘要:
Deterministically resolving cycles in a library tree is disclosed. Resolving cycles supports certain processes such as safe library initialization. Cycles in the library tree are identified; at least one soft link in each identified cycle is identified; and the at least one soft link in each identified cycle is broken. If a cycle has no soft links, notification is provided indicating that the cycle cannot be broken. Identifying at least one soft link in each identified cycle comprises, for each link in the cycle, determining the dependent and supporting libraries; and determining if one or more functions in the supporting library are required for initializing the dependent library.
摘要:
Sharing access to resources using an inter-process communication (“IPC”) provides a connection in which references to resources are passed from a sender to a receiver in a trusted third party environment. A sender in possession of a reference to a resource, such as a handle to an object, may initiate the connection with the receiver. In turn, the receiver may accept or refuse the connection, and may further specify the types of resources in which the receiver is interested when accepting through the connection. Sharing access to resources in this manner advantageously insures that only a process that already has access to a resource is able to share that access with another process, and further that only processes that wish to do so will accept such access.
摘要:
A barrier with local spinning. The barrier is described as a barrier object having a bit vector embedded as a pointer. If the vector bit is zero, the object functions as a counter; if the vector bit is one, the object operates as a pointer to a stack. The object includes the total number of threads required to rendezvous at the barrier to trigger release of the threads. The object points to a stack block list that describes each thread that has arrived at the barrier. Arriving at the barrier involves reading the top stack block, pushing onto the list a stack block for the thread that just arrived, decrementing the thread count, and spinning on corresponding local memory locations or timing out and blocking. When the last thread arrives at the barrier, the barrier is reset and all threads at the barrier are awakened for the start of the next process.
摘要:
A parallel processing method and apparatus for initializing libraries is disclosed. Libraries for an application are identified, an initialization order for the libraries is determined, and the libraries are initialized in asynchronous stages. The initialization order is determined by forming a library tree of the libraries' references and determining a load order for the references according to the levels of the references in the library tree. The asynchronous stages comprise a loading stage that includes a load queue, a snapping stage that includes a snap queue, and an initializing stage that includes an initialize queue.
摘要:
In a first thread of a process a determination is made that a current value at a target address is not a desired value. In response to this determination, a first application programming interface (API) is invoked to indicate that the first thread is to sleep and be woken up when a second thread modifies the value at the target address. When a second thread modifies the value at the target address, the second thread invokes a second API to indicate that the value at the target address has been modified. In response to the second API being invoked, the first thread is woken up.