lkml.org 
[lkml]   [2013]   [Dec]   [10]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [PATCH 0/1] uprobes: Kill __replace_page(), change uprobe_write_opcode() to rely on gup(WRITE)
From
On Tue, Dec 10, 2013 at 11:18 AM, Oleg Nesterov <oleg@redhat.com> wrote:
>
> Well, if we use get_user_pages() we have vma for free.

No we don't. get_user_pages() is expensive as hell.

We'd be *much* better off using get_user_pages_fast() if possible -
and I bet _is_ possible in 99% of all cases.

So saying that we get vma "for free" is complete BS. We get it "for
very expensive."

> But I am not sure about "and verify it for now" above. Do you mean
> that uprobe_write_opcode() should do another valid_vma() ?

No. I think we should just do something like

if (vma->vm_flags & (VM_SHARED | VM_HUGETLB))
return -EINVAL;

and actually make the logic obvious. We should *also* make sure we do
the right thing for THP, because the VM_HUGETLB check isn't obviously
sufficient. Does the code really do the right thing for a writable
hugepage? I'd like that to be *obvious*.

Quite frankly, I think that adding a few of these kinds of *obvious*
checks and then just looking up the page table _once_ - instead of
three times - would make the code safer and faster.

As it is, it is neither safe nor fast. It *may* be safe if callers do
all the right black magic, but it's by no means obvious.

Linus


\
 
 \ /
  Last update: 2013-12-10 21:01    [W:0.081 / U:0.008 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site