{"id":102845,"date":"2022-01-03T15:36:38","date_gmt":"2022-01-03T13:36:38","guid":{"rendered":"https:\/\/prohoster.info\/blog\/novosti-interneta\/%d0%be%d0%bf%d1%83%d0%b1%d0%bb%d0%b8%d0%ba%d0%be%d0%b2%d0%b0%d0%bd-%d0%bd%d0%b0%d0%b1%d0%be%d1%80-%d0%bf%d0%b0%d1%82%d1%87%d0%b5%d0%b9-%d1%83%d1%81%d0%ba%d0%be%d1%80%d1%8f%d1%8e%d1%89%d0%b8%d1%85"},"modified":"2022-01-03T15:36:38","modified_gmt":"2022-01-03T13:36:38","slug":"%d0%be%d0%bf%d1%83%d0%b1%d0%bb%d0%b8%d0%ba%d0%be%d0%b2%d0%b0%d0%bd-%d0%bd%d0%b0%d0%b1%d0%be%d1%80-%d0%bf%d0%b0%d1%82%d1%87%d0%b5%d0%b9-%d1%83%d1%81%d0%ba%d0%be%d1%80%d1%8f%d1%8e%d1%89%d0%b8%d1%85","status":"publish","type":"post","link":"https:\/\/prohoster.info\/en\/blog\/news\/%d0%be%d0%bf%d1%83%d0%b1%d0%bb%d0%b8%d0%ba%d0%be%d0%b2%d0%b0%d0%bd-%d0%bd%d0%b0%d0%b1%d0%be%d1%80-%d0%bf%d0%b0%d1%82%d1%87%d0%b5%d0%b9-%d1%83%d1%81%d0%ba%d0%be%d1%80%d1%8f%d1%8e%d1%89%d0%b8%d1%85","title":{"rendered":"A set of patches has been released that speeds up the Linux kernel build by 50-80%.","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"
Ingo Molnar, a renowned Linux kernel developer and author of the Completely Fair Scheduler (CFS), has proposed a series of patches for discussion on the Linux kernel developers' mailing list. These patches impact more than half of all files in the kernel's source code and promise to increase the speed of full kernel recompilation by 50-80%, depending on the configuration. The implemented optimization is notable for introducing the largest set of changes in the history of kernel development \u2014 a total of 2,297 patches modifying over 25,000 files (10,000 header files in the \u2018include\/\u2019 and \u2018arch\/*\/include\/\u2019 directories and 15,000 source code files). <\/p>\n
The performance gain is achieved by changing the method of processing header files. It is noted that after thirty years of kernel development, the state of header files has become deplorable due to a large number of cross-dependencies among files. The work on restructuring the header files took over a year and required significant reworking of the hierarchy and dependencies. During the restructuring, effort was made to separate type and API definitions for different kernel subsystems. <\/p>\n
Among the changes made are: the separation of high-level header files from one another, the exclusion of linking header files for inline functions, the categorization of header files for types and APIs, and ensuring isolated compilation of header files (about 80 files had indirect dependencies that hindered compilation through other headers). Additionally, automatic addition of dependencies to '.h' and '.c' files, stepwise optimization of header files, use of the 'CONFIG_KALLSYMS_FAST=y' mode, and selective consolidation of C files into build blocks to reduce the number of object files were implemented. <\/p>\n
As a result, the work completed has allowed for a reduction in the size of header files processed during the post-processing stage by 1-2 orders of magnitude. For example, prior to optimization, using the 'linux\/gfp.h' header file resulted in the addition of 13,543 lines of code and the inclusion of 303 dependent header files. After optimization, the size was reduced to 181 lines and 26 dependent files. Another example: during the preprocessing of the 'kernel\/pid.c' file without the patch, 94,000 lines of code are included, most of which are not used in pid.c. The separation of header files has allowed for a threefold decrease in the volume of processed code, reducing the number of lines processed to 36,000. <\/p>\n
During a complete kernel rebuild using the command «make -j96 vmlinux» on a test system, applying patches resulted in a reduction of the build time for branch v5.16-rc7 from 231.34 to 129.97 seconds (from 15.5 to 27.7 builds per hour), as well as an increase in the CPU core usage efficiency during the build. In incremental builds, the optimization effect is even more pronounced \u2014 the time taken for a kernel rebuild after making changes to header files has decreased significantly (from 112% to 173% depending on the modified header file). Currently, optimizations are available only for ARM64, MIPS, Sparc, and x86 (both 32-bit and 64-bit) architectures. \n Source: opennet.ru<\/a> <\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"