Analysis And Demonstration Of Kernel Function’s Running Time And Path

As an famous open source operating system over the years, Linux has been the research focus of IT. The popular Android system is also based on the free and open source code of Linux. With more and more applications of Android system in mobile terminal, More and more computer fans, research institutions and companies come into the study of the Linux kernel. Over the years many Linux Kernel Developer keep adding and maintaining its functions, Linux kernel source code becomes more complex.In this paper,by analyzing different kernel probing tool, we identify their main features, advantages and disadvantages. Most track results format of the detection range of tools are directed to same basic data and difficulty to read large amounts of data. However Gprof performance analysis tool can automatically generate the results and convert data to SVG diagram and do a simple analysis of the time information. But Gprof performance analysis tool can not track and analyze the Linux kernel. In response to these problems, this paper presents the results of the analysis output of Ftrace, Systemtap and S2 E and other kernel trace detection tools combined with Gprof, so that the output of each tool can be in a good uniform and easy to read.In order to ensure the possibility of combining these tools.In this paper, the relationship between the calling function and the sampling process Gprof time performance analysis tool for a detailed analysis and Ftrace, Systemtap and S2 E probing tools such as tracking the data format conversion, so the data needs to adapt to Gprof.In this paper, we abandon Gprof’s data acquisition section, retains the ability to generate and analyze the data, improve the deficiencies and limitations.Finally completed with other tools of the output.Through the use of improved Gprof, function relations and time informations will be called to generate the results and present to the user with SVG function call graph to help users understand the function relations. The improved Gprof combined with LXR online browsing tools. According to LXR’s workflow,we get users’ input from the front-end, analysis and process users’ input, check user input data from the back-end, call Gprof tool to generate function call graph, display the function call graph in front-end. By binding the Gprof to LXR, allowing users to easily and quickly check with the call to the called relationship of a single function.