Want to give the interview a wave of official explanations, Then explain it in vernacular. What is pessimism and optimism, We should not be optimistic when we interview. This problem extends to a lot of questions, For example, thread safety ,CAS principle, Advantages and disadvantages. ĭo you understand optimistic lock and pessimistic lock ？ Įpoll Suitable for a large number of less active connections. Select Suitable for a small number of active connections, Usually thousands of. call epoll_wait The ready event in the linked list will be returned directly, Efficient.
Treesize professional full version driver#
Double linked list stores ready events, All added to epoll The events in will be related to the device ( network card ) The driver establishes a callback relationship, in other words, This callback method is called when the corresponding event occurs, This callback method is called in the kernel ep_poll_callback, It will add what happened to rdlist In the double linked list. Is maintained in the kernel Red and black trees And a Double linked list, Red and black trees are stored through epoll_ctl Method direction epoll The event added to the object, So you don't need to call epoll_wait Copy all event structures in full. The number of file descriptors that a single process can monitor There is a maximum limit, Usually 1024, Of course, you can change the quantityĮpoll A. Kernel / User space memory copy problem ,select The handle data structure set in the kernel is changed every time, So every time select All handle data structures need to be copied from user space to kernel space, It's a huge expense
Select Is triggered by Level trigger, Application if not completed on a ready file descriptor IO operation, So every time after that select The call still notifies the process of these file descriptors Select Returns an array containing the entire handle, The application needs to traverse the entire array to discover which handles have an event The same routine here, Let's talk about the purpose of both, And then the pros and cons of both. Tell me about select and epoll The difference between ？.So what's the difference between signal drive and signal drive ? Program the kernel to start an operation, And let the kernel operate throughout （ This includes copying data from the kernel to the application's buffer ） Notify the application when finished. When the data is ready, The process will receive a SIGIO The signal, It can be called in the signal processing function I/O Operating functions process data.
Treesize professional full version install#
In signal driven mode I/O In the model, The application program uses socket interface to drive signal I/O, And install a signal processing function, The process continues to run without blocking. IO Multiplexing means that the standard input can be 、 Socket and so on IO All the way to, Any way IO Something happened, Will inform the corresponding application to process the corresponding IO event, In our view, it's repeated meanwhile Can handle multiple things. When using fgets Waiting for standard input, If there is data in the socket but it can't be read out.
When using non blocking functions, And blocking IO analogy, The kernel will immediately return, Get enough when you return CPU Time to go on doing other things. When used Blocking IO When, Applications will be merciless Hang up, Wait for the kernel to complete the operation, Because the kernel at this point might be CPU Time switches to other processes that need to be done, In the eyes of our app, we feel stuck ( Blocking ) 了. We know that when calling a function, there are only two situations, or Go back to, Then according to the return value, the following business processing is carried out.
0 kommentar(er)
