Currently I am learning Kernel Module Programming reading "https://www.tldp.org/LDP/lkmpg/2.6/lkmpg.pdf" ,and in that book there is a chapter titled as "/proc File System", I have learnt how modules are used as device drivers, but I have not been able to understand the importance of creating "/proc/file_name" by an kernel module. I know that in /proc directory kernel maintains info about processes. I need to understand the importance of creating "/proc/file_name" by an kernel module.


The beginning of chapter 5.1 says this about the /proc filesystem:

Originally designed to allow easy access to information about processes (hence the name), it is now used by every bit of the kernel which has something interesting to report, [...]

You don't have to create a file in the /proc filesystem, but if you want or need to have an easy way to report statistics or other information about the state of your module, the /proc filesystem is one way to do it. Because of the way the /proc filesystem was developed, it can be a bit disorganized. Instead of creating several things at the root of the /proc filesystem, you might want to create a sub-directory for your module at e.g. /proc/driver, to minimize the chance of conflicts with the main kernel, other third-party modules or future development.

Note that about the time that book was being written, the /sys filesystem was also introduced to be a more systematically arranged place (and so more friendly for integrating user-space tools) for any tunable settings your kernel module might have. Unfortunately, the chapter 6 that was supposed to be about it is just a TODO: Write a chapter about sysfs placeholder in that PDF.

(Historical note: before the /sys filesystem existed, the /proc filesystem was used as a place for run-time tweakable settings, usually identifiable as writeable files. Some of them might still not have been completely migrated to the /sys filesystem, or even not going to migrate at all to maintain backward compatibility. But for any modern development, it would be better to put any tunables in /sys instead of /proc.)

In modern kernels, there is also the debugfs, that can be mounted to /sys/kernel/debug, that could be a good place if your kernel module needs a way to output some debugging information. It might be a good place for outputting more detailed/sensitive debug information, as debugfs is designed for system administrator access only. It is also completely optional: in production systems the debugfs can be simply left unmounted unless there is a need for in-depth troubleshooting.

So, in a nutshell: if your module needs to provide a way to tune some settings at run-time, the /sys filesystem is a good place for them. For debugging information in modern kernels, use the debugfs. For statistics and similar stuff that can be seen by non-root users, /proc can be used.

| improve this answer | |
  • you need to say that if my module wants to show some stats , just like kernel shows in "/proc/modules" which indicates that which of the modules are currently inserted , I can do that by creating file in "/proc" file system and writing the info. – Sunny Khandare Mar 8 '19 at 3:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.