[PATCH] um: clean up mm creation

[Johannes Berg]

On Fri, 2023-09-22 at 14:41 +0100, Anton Ivanov wrote:
> On 22/09/2023 12:16, Johannes Berg wrote:
> > From: Johannes Berg <johannes.berg at intel.com>
> > 
> > While enabling PREEMPT on UML, we found that the call to
> > force_flush_all() cannot be done where it is, it sleeps
> > while atomic.
> > 
> > Further investigation shows that all this seems at least
> > a bit roundabout and possibly wrong wrong in the first
> > place - we copy from the 'current' process and then flush
> > when it starts running.
> > 
> > What we really want is to start fresh with an empty mm
> > process, then have it all set up by the kernel (copying
> > the original mm contents as needed), and then sync it
> > in arch_dup_mmap().
> 
> Is there a way we can come up with COW here?

Well, I kind of thought we _did_ if we have arch_dup_mmap()?

Need to understand this better.

> > 
> > We should do the same for the LDT, so need to split that
> > to be able to do this.
> > 
> > Note that this fixes what seems like an issue - we look
> > at current->mm when we copy, but that doesn't seem right
> > in the case of clone() without copying the MM. This is
> > probably saved by the forced flush later right now.
> 
> We will need to work on this.
> 
> It is nearly twice slower than the current approach on a find /usr -type f -exec cat {} > /dev/null \;

Hm, that's annoying. Guess I have to figure out why.

johannes