On Thu, 28 Aug 2025 13:38:33 -0700
Linus Torvalds [off-list ref] wrote:

On Thu, 28 Aug 2025 at 13:17, Steven Rostedt [off-list ref] wrote:

quoted

quoted

That should be simple and straightforward, and hashing two pointers
should be simple and straightforward.  

Would a hash of these pointers have any collisions? That would be bad.  

What? Collisions in 64 bits when you have a handful of cases around?
Not an issue unless you picked your hash to be something ridiculous.

Since I only need a unique identifier, and it appears that the
vma->vm_file->f_inode pointer is unique, would just using that be OK?

I could run it through the same hash algorithm that "%p" goes through so
that it's not a real memory address.

As getting to the path does require some more logic to get to. Not to
mention, this may later need to handle JIT code (and we'll need a way
to map to that too).

-- Steve