Re: [patch] cache flush bug in mm/filemap.c (all kernels >=

[Andrew Morton]

"David S. Miller" <[e-mail]> wrote:
>
>    From: Andrew Morton <[e-mail]>
>    Date: Fri, 23 May 2003 02:12:04 -0700
>    
>    install_page() is prefaulting pages into pagetables, so perhaps it should
>    have an update_mmu_cache()?
> 
> I agree.  Someone should take a close look at the do_file_page()
> code paths to make sure that's still kosher after such a change.

What would one be looking for?  I don't know what the sideeffects of
update_mmu_cache() might be.

It looks to be the same as do_no_page() though: the update_mmu_cache() is
the last substantive thing which happens in the fault or the syscall.