What's the difference between `lsof -p <pid> | wc -l` and `ls /proc/<pid>/fd | wc -l`?

Question

Background: I'm playing around with monitoring the ulimit for running processes for a particular user. (I had occasionally seen processes that were getting started with an incorrect limit.) I asked a couple self-professed Linux gurus, and one suggested lsof -p <pid>, while the other suggested ls /proc/<pid>/fd, but neither was positive about which more accurately reflects the actual count towards the max open files limit for a process.

So, which is it?

lsof -p <pid> | wc -l

Or

ls /proc/<pid>/fd | wc -l

Please elaborate on the difference. Thanks!

Answer 1

lsof will also give you memory mapped .so-files - which technically isn't the same as a file handle the application has control over. /proc/<pid>/fd is the measuring point for open file descriptors - however: Mentioned in the proc-man page - if the main thread of a multithreaded program has terminated, this directory will be unavailable.

lsof -p <pid> | grep -v mem | egrep -v '^COMMAND PID' | wc -l will show you the same items as ls /proc/<pid>/fd | wc -l.

The memory maps is available in /proc/<pid>/maps.