Re: [RFC PATCH] Remove flush_icache_user_range()

[David Miller <davem@davemloft.net>]

From: Russell King - ARM Linux <linux@arm.linux.org.uk>
Date: Wed, 9 Apr 2014 17:15:55 +0100

> On Wed, Apr 09, 2014 at 12:09:26PM -0400, David Miller wrote:
>> From: Russell King - ARM Linux <linux@arm.linux.org.uk>
>> Date: Wed, 9 Apr 2014 11:01:50 +0100
>> 
>> > What has been proposed for uprobes is a uprobe specific function -
>> > flush_uprobe_xol_access() which is used after uprobes writes via the
>> > kmap_atomic() mapping of the page:
>> > 
>> > +       xol_page_kaddr = kmap_atomic(area->page);
>> > +
>> > +       /* Initialize the slot */
>> > +       memcpy(xol_page_kaddr + (xol_vaddr & ~PAGE_MASK),
>> > +              &uprobe->arch.ixol,
>> > +              sizeof(uprobe->arch.ixol));
>> > +
>> > +       arch_uprobe_flush_xol_access(area->page, xol_vaddr,
>> > +                                    xol_page_kaddr + (xol_vaddr & ~PAGE_MASK),
>> > +                                    sizeof(uprobe->arch.ixol));
>> > +
>> > +       kunmap_atomic(xol_page_kaddr);
>> > 
>> > Comments?
>> 
>> ptrace() accesses (via __access_remote_vm()) already use an existing
>> helper function for these sorts of situations, in the form of
>> copy_{to,from}_user_page().  I would suggest that uprobes uses that
>> as well.
> 
> Yes, I agree with you.
> 
> I made that suggestion too (because it's really the same problem)
> but uprobes people are concerned that it's too much of an overhead:

They should use it at first, then if there is a problem we can identify
exactly what can be optimized or needs to be, and expand the interface
as needed.

Let's not optimize first.