lkml.org 
[lkml]   [2016]   [Jan]   [19]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
Date
SubjectRe: [PATCH v3] kernel: add kcov code coverage
On Mon, Jan 18, 2016 at 10:52 PM, Kirill A. Shutemov
<kirill@shutemov.name> wrote:
> On Mon, Jan 18, 2016 at 08:28:04PM +0100, Dmitry Vyukov wrote:
>> On Mon, Jan 18, 2016 at 1:42 PM, Kirill A. Shutemov
>> <kirill@shutemov.name> wrote:
>> > On Thu, Jan 14, 2016 at 03:22:21PM +0100, Dmitry Vyukov wrote:
>> >> + area = t->kcov_area;
>> >> + /* The first u32 is number of subsequent PCs. */
>> >> + pos = READ_ONCE(area[0]) + 1;
>> >> + if (likely(pos < t->kcov_size)) {
>> >> + area[pos] = (u32)_RET_IP_;
>> >
>> > If area[0] is UINT32_MAX the condition will be true and you'll make
>> > area[0] temporary set to (u32)_RET_IP_. That's may be fun. :)
>>
>> Well, if user wants to corrupt this region, he is free to do so. But
>> it wont't corrupt kernel, right?
>
> Right. But it's easy enough to avoid overflow too:
> "if (likely(pos && pos < t->kcov_size)"


It is easy but I don't understand what we are trying to fix.
Such check won't fix operation of kcov if user makes random memory
writes. User can also write random values which are not UINT32_MAX to
area[0]. In such case coverage will be intermixed and some pieces will
be missing.
Also this function is quite hot. A single register check and branch is
not too expensive, but it does not worth it if it does not bring any
value.


>> >> +
>> >> + page = vmalloc_to_page(kcov->area + off);
>> >> + get_page(page);
>> >> + vmf->page = page;
>> >> + return 0;
>> >> +}
>> >
>> > BTW, since all pages pre-allocated, you don't really need fault handler --
>> > just setup all pages table enties during mmap. This would shorten warm up
>> > period for userspace. See vm_insert_page().
>>
>> Done in v4
>> I don't need to get_page() before vm_insert_page(), right?
>
> Right. vm_insert_page() will take the referenece for you.


ack

\
 
 \ /
  Last update: 2016-01-19 14:01    [W:2.533 / U:0.008 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site