[rnnr]

Memory mapped files also known as sections in VMS/NT have just two advantages:

  * Fewer context switches among user space / kernel syscalls

  * No need to copy _modified_ data into the swap. The behavior for read only data doesn't change

That's it, nothing miraculous about it.

[astrange]

No, it also shares them between multiple runs of the same process, and it reuses the pages that were in your file cache anyway.

> * Fewer context switches among user space / kernel syscalls

This is not necessarily true. There's lots of cases where it's actually slower.

[jlokier]

> Fewer context switches among user space / kernel syscalls

Note that each page fault to read an mmaped page is also a context switch from user space into the kernel. The kernel entry/exit paths for syscalls and page faults have much in common.

Mmaped pages can in principle use fault-ahead to reduce the number of page faults when sequential access is detected. An equivalent reduction is available for read syscalls by reading larger blocks at a time.

In practice, mmap files are faster in some scenarios compared with read syscalls and slower in others. Same with O_DIRECT reads, those are faster in some scenarios and slower in others.