{"id":155127,"date":"2026-01-25T17:11:59","date_gmt":"2026-01-25T15:11:59","guid":{"rendered":"https:\/\/prohoster.info\/blog\/novosti-interneta\/predstavlena-fs-daxfs-pohozhaya-na-ramfs-i-tmpfs-no-rabotayushhaya-v-obhod-stranichnogo-kesha"},"modified":"2026-01-25T17:11:59","modified_gmt":"2026-01-25T15:11:59","slug":"predstavlena-fs-daxfs-pohozhaya-na-ramfs-i-tmpfs-no-rabotayushhaya-v-obhod-stranichnogo-kesha","status":"publish","type":"post","link":"https:\/\/prohoster.info\/et\/blog\/news\/predstavlena-fs-daxfs-pohozhaya-na-ramfs-i-tmpfs-no-rabotayushhaya-v-obhod-stranichnogo-kesha","title":{"rendered":"Esitletud on failis\u00fcsteem DAXFS, mis sarnaneb ramfs ja tmpfs-iga, kuid t\u00f6\u00f6tab lehek\u00fclje puhvrist m\u00f6\u00f6da.","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"
Cong Wang, who is responsible for the traffic control subsystem in the Linux kernel and is developing the Multikernel project, has introduced a new filesystem, DAXFS, to the list of Linux kernel developers. DAXFS utilizes RAM for data storage, operates in read-only mode, and resembles the functionality of trimmed-down filesystems ramfs and tmpfs. The kernel module implementing DAXFS and the mkdaxfs utility for creating the filesystem have been published under the GPLv2 license. Following a review, the possibility of integrating DAXFS into the main kernel composition will be considered. <\/p>\n
A key feature of the new filesystem is the use of the DAX (Direct Access) subsystem to bypass the page cache and directly access shared physical memory. In ramfs and tmpfs, the contents are stored in the page cache, which, when using a filesystem in multiple containers or kernels launched using the Multikernel technology, leads to multiple copies of data being placed in physical memory. Filling ramfs and tmpfs also requires performing unnecessary copy operations without the possibility of mapping an existing memory area. <\/p>\n
Direct memory access in DAXFS allows for the elimination of unnecessary copy operations and facilitates operation outside the traditional input\/output stack based on block devices and buffering mechanisms, as well as avoids memory fragmentation when sharing one filesystem across different containers and multikernel kernels. DAXFS operates in Zero-Copy mode and directly delivers the content of files from memory without unnecessary data duplication in the page cache, with the possibility of mapping continuous areas of physical memory into each instance of the container or kernel. Additionally, the use of the dma-buf API for using memory exported by GPU, FPGA\/SmartNIC, or CXL devices is supported. <\/p>\n
The filesystem operates in read-only mode and is initialized by loading a pre-prepared image of the filesystem into memory, which simplifies implementation, eliminates the need for memory allocation operations during operation, and avoids complicated device management mechanisms. <\/p>\n
DAXFS-i peamine rakendusala on kasutamine koos Multikernel tehnoloogiaga, mis v\u00f5imaldab \u00fchel f\u00fc\u00fcsilisel arvutil k\u00e4itada mitut s\u00f5ltumatut Linuxi tuuma, millel on otsene juurdep\u00e4\u00e4s riistvarainele ja mis v\u00f5ivad olla kasutusel mitmete isoleeritud s\u00fcsteemi keskkondade k\u00e4itamiseks. <\/p>\n
DAXFS-i saab kasutada ka mitme konteineri v\u00f5i tuuma \u00fchise juurdep\u00e4\u00e4su korraldamiseks p\u00f5hisisu pildile, mis on paigaldatud jagatud m\u00e4llu; riistvara kiirendite m\u00e4lule juurdep\u00e4\u00e4su kulude v\u00e4hendamiseks; m\u00e4lule juurdep\u00e4\u00e4su organiseerimiseks mitmest hostist ilma v\u00f5rgu sisend-v\u00e4ljundita v\u00f5i p\u00fcsivale m\u00e4llu juurdep\u00e4\u00e4suks CXL (Compute Express Link) baasil seadmetele. \n Allikas: opennet.ru<\/a> <\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"