[Bug 293382] Dead lock and kernel crash around closefp_impl

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=293382

--- Comment #89 from Kyle Evans <kevans@freebsd.org> ---
(In reply to Konstantin Belousov from comment #86)

Here's where I think the code diverges from the description:

```
This instruction invalidates the TLB entry or entries, regardless of the page
size (4 Kbytes, 2 Mbytes, 4
Mbytes, or 1 Gbyte). 
```

va being aligned to a PDE doesn't mean that describes a 2M page, right?  If
that's not a superpage, we're actually looking at 512 4K pages.  My
interpretation of the above is that if va is a 4K page, it will invalid just
the one 4K page.  Combine that with the wording about incrementing, I think the
way the code is written is *actually* invalidating a single 4K page at the
beginning of a PDE and then jumping by <count> PDEs (skipping all of the 4K
pages after the first), when it was meant to invalidate <count> PDEs and jump
by as many.

-- 
You are receiving this mail because:
You are the assignee for the bug.