Re: [PATCH v3 1/4] x86/vm86/32: Switch to flush_tlb_mm_range() in mark_screen_rdonly()

[Andy Lutomirski <luto@kernel.org>]

On Wed, Apr 26, 2017 at 5:02 PM, Nadav Amit <namit@vmware.com> wrote:
> And besides, it looks as if the code was meant to flush the entire
> TLB in some cases (e.g., if pgd_none_or_clear_bad() is true).
>
> On 4/26/17, 4:56 PM, "Nadav Amit" <namit@vmware.com> wrote:
>
>     It may be benign, but I don’t think that flushing the TLB without
>     holding the ptl or the mmap_sem (for no apparent reason) is a good
>     practice.
>
>     On 4/22/17, 12:01 AM, "Andy Lutomirski" <luto@kernel.org> wrote:
>
>         mark_screen_rdonly() is the last remaining caller of flush_tlb().
>         flush_tlb_mm_range() is potentially faster and isn't obsolete.
>
>         Compile-tested only because I don't know whether software that uses
>         this mechanism even exists.
>
>         Cc: Rik van Riel <riel@redhat.com>
>         Cc: Dave Hansen <dave.hansen@intel.com>
>         Cc: Nadav Amit <namit@vmware.com>
>         Cc: Michal Hocko <mhocko@suse.com>
>         Cc: Sasha Levin <sasha.levin@oracle.com>
>         Cc: Andrew Morton <akpm@linux-foundation.org>
>         Signed-off-by: Andy Lutomirski <luto@kernel.org>
>         ---
>          arch/x86/kernel/vm86_32.c | 2 +-
>          1 file changed, 1 insertion(+), 1 deletion(-)
>
>         diff --git a/arch/x86/kernel/vm86_32.c b/arch/x86/kernel/vm86_32.c
>         index 23ee89ce59a9..3eda76b3c835 100644
>         --- a/arch/x86/kernel/vm86_32.c
>         +++ b/arch/x86/kernel/vm86_32.c
>         @@ -193,7 +193,7 @@ static void mark_screen_rdonly(struct mm_struct *mm)
>                 pte_unmap_unlock(pte, ptl);
>          out:
>                 up_write(&mm->mmap_sem);
>         -       flush_tlb();
>         +       flush_tlb_mm_range(mm, 0xA0000, 0xA0000 + 32*PAGE_SIZE, 0UL);
>          }
>

Those should probably be pgd_none(), not pgd_none_or_clear_bad().

But this whole function is just garbage.  It mucks with page
protections without even looking up the VMA.  What happens if the
pages are file-backed?  How about chardevs?

I'd like to delete it.  Stas, do you know if there's any code at all
that uses VM86_SCREEN_BITMAP?  Some Googling didn't turn any up at
all.