This website uses cookies and similar technologies to understand visitors' experiences. By continuing to use this website, you accept our use of cookies and similar technologies,Terms of Use, and Privacy Policy.

Nov 30 2009 - 10:13 AM
OS Class - 30th Nov
FileSystems ----------- Ext2 - Groups - How Big? 8192 blocks - Index - 12 direct - 1 Single - 1 Double - 1 Triple - Filenames - Directory entries, variable size, 255 chars Ext3 - Compatible W/Ext2 - Journaling/Logging Journal | Earliest | | Latest 1) Journal - Everything to Journal - Slow 2) Ordered - Log Metadata to Jounal metadata written after data. 3) Writeback - Log metadata to journal but no ordering. What about new devices which are available in the market today called SSD (Solid State devices). - Which are a bunch of chips and we can use electrons to get the data as compared to a mechanical movement of the hard disks. Our assumptions like Random Access or Seeks are expensive does not apply here. Also there are other things like Writes are really really slow. Linux has a bunch of file systems catering to SSDs for e.g. JFFS Virtual File System - VFS (Unix and Linux) File System - fs fs.h - /include/linux (this file has definitions for important file table structures). super operations are performed on Super Block. Super block is a key aspect of the file system. struct inode (619 line) - inode numbers, uid, gid inode_operations - Standard file system, system calls. for each specific file systems we would implement the functions for the new filesystems. file_operations is also linked from inode and also from file too. for e.g. Mmap(fs.h #1225) VFS is an independent structure
|By: Pranav Garg|5733 Reads