Initial reading and analysis
Week 1:
The plan is to create a virtual filesystem that hooks on to the TCP/IP stack and exposes it to user-space as a /net filesystem just like Plan 9 From Bell Labs does. I looked at debugfs as a sample virtual fs implementation. They all use libfs which is a library in Linux provided by the kernel that does most of the redundant boilerplate code required to get the kernel to identify the filesystem.
From initial reading, it seems like there is an API that lets me describe the name of the module, and implement functions like create_file, create_dir, write_file and read_file. These functions get called when we use any shell script or any other user-space program to manipulate the filestructure.
I have managed to get a reference book, which contains all the implementation details of Socket and TCP/IP stack of Linux. I have access to many other Plan 9 documentation from over the years.
More reading to follow in subsequent weeks…
No comments yet.
Leave a Reply
| Next »
-
Recent
- A test script to fetch over HTTP on command-line
- Where to hook onto the tcp stack?
- Basic framework ready!
- “fsnet_create” a good plan?
- Glendix on Routers?
- DNS: A hard to crack nut!
- My first /net code commited!
- Learning to write a Systhetic Filesystem on Linux
- Discussion on 9Fans
- Plan 9 Cheat Sheet
- Network internals of Plan 9
- Initial reading and analysis
-
Links
-
Archives
- April 2009 (3)
- March 2009 (3)
- February 2009 (5)
- January 2009 (1)
-
Categories
-
RSS
Entries RSS
Comments RSS