File I/O via memory mapping

Both in Appendix A, File I/O Essentials, and in this chapter, we have on several occasions mentioned how the Linux kernel's page cache helps greatly enhance performance by caching the content of files within it (alleviating the need to each time go to the really slow storage device and instead just read or write data chunks within RAM). However, though we gain performance via the page cache, there remains a hidden problem with using both the traditional read(2), write(2) APIs or even the faster SG-I/O (the [p][read|write][v][2](2)) APIs.

..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset