This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: PATCH: remove dot_rdg
On Wed, Dec 16, 2009 at 9:45 PM, Dave Korn
<dave.korn.cygwin@googlemail.com> wrote:
> Richard Guenther wrote:
>> On Wed, Dec 16, 2009 at 12:36 PM, Diego Novillo <dnovillo@google.com> wrote:
>>> On 12/15/09 23:12 , Ben Elliston wrote:
>>>
>>>> 2009-12-16 ?Ben Elliston ?<bje@au.ibm.com>
>>>>
>>>> ? ? ? * tree-data-ref.h (dot_rdg): Remove prototype.
>>>> ? ? ? * tree-data-ref.c (dog_rdg_1, dot_rdg): Remove.
>>> OK
>>
>> Well, you can clearly see that the function is for debugging purposes only.
>> So - do we care? ?The function is probably useful when analyzing testcases.
>
> ?I'm inferring it's a race condition problem with the /tmp/rdg.dot file we're
> talking about here. ?I can just about imagine two users on the same shared
> system and one of them socially engineering the other into debugging a
> carefully-crafted compiler-crashing testcase that might hopefully prompt them
> to call dot_rdg whilst running a symlink-racer to seize control of the file
> when they create it, but I couldn't think of a remote way to exploit it; it's
> a fairly low risk vulnerability, as far as I can see.
>
> ?If we wanted to fix it, libiberty has mkstemp for us, IIRC.
Or we can make the filename a function argument.
Richard.