Comments

• 

  about 10 years ago by eeich | Reply

  It turned out that implementing this was a lot easier than I thought. The unix socket struct of the kernel has a member which points directly to the socket of the peer. From this the inode of the peer can be obtained easily. I wonder who this information hasn't been made available in /proc/net/unix. To avoid modifying this proc file a new proc file was added: /proc/net/unix_peer listing the sockets which have peers (are not in listening state) and their peers. The iterator funciton is just 16 lines of code plus some overhead to add the new entry. Pipe pairs are easy to find - they share the same inode. To identify the processes which communicate over these pipes and sockets more easily there are two perl scripts processing the information from /proc. To ease processing in user space there are

×
Edit Comment 830
# A First Level Header
## A Second Level Header

Use one asterisk to *emphasize*

Use two asterisks for **strong emphasis**

- Use hyphens
- for unordereed
- lists

This is an [link to example.com](http://example.com/)

This is an image ![an openSUSE geeko icon](https://en.opensuse.org/images/d/d0/Icon-distribution.png)

This is a user link @hans

This is a project link hw#some-cool-title
More Complex Markdown Help
Formatting Help
• Edit
• Preview
It turned out that implementing this was a lot easier than I thought. The unix socket struct of the kernel has a member which points directly to the socket of the peer. From this the inode of the peer can be obtained easily. I wonder who this information hasn't been made available in /proc/net/unix. To avoid modifying this proc file a new proc file was added: /proc/net/unix_peer listing the sockets which have peers (are not in listening state) and their peers. The iterator funciton is just 16 lines of code plus some overhead to add the new entry. Pipe pairs are easy to find - they share the same inode. To identify the processes which communicate over these pipes and sockets more easily there are two perl scripts processing the information from /proc. To ease processing in user space there are
 
×
Reply to eeich
It turned out that implementing this was a lot easier than I thought. The unix socket struct of the kernel has a member which points directly to the socket of the peer. From this the inode of the peer can be obtained easily. I wonder who this information hasn't been made available in /proc/net/unix. To avoid modifying this proc file a new proc file was added: /proc/net/unix_peer listing the sockets which have peers (are not in listening state) and their peers. The iterator funciton is just 16 lines of code plus some overhead to add the new entry. Pipe pairs are easy to find - they share the same inode. To identify the processes which communicate over these pipes and sockets more easily there are two perl scripts processing the information from /proc. To ease processing in user space there are
---
# A First Level Header
## A Second Level Header

Use one asterisk to *emphasize*

Use two asterisks for **strong emphasis**

- Use hyphens
- for unordereed
- lists

This is an [link to example.com](http://example.com/)

This is an image ![an openSUSE geeko icon](https://en.opensuse.org/images/d/d0/Icon-distribution.png)

This is a user link @hans

This is a project link hw#some-cool-title
More Complex Markdown Help
Formatting Help
• Edit
• Preview