Re: [PATCH 2/5] __FD_ELT: Implement correct buffer overflow check

["Carlos O'Donell" <carlos@redhat.com>]

On 05/01/2013 02:28 AM, KOSAKI Motohiro wrote:
>>> +     ? __fdelt_buffer_warn(__d, __bos0 (s))                          \
>>
>> Space between function and bracket e.g. foo () not foo().
> 
> ah, ok.
> 
>>
>>> +     : __fdelt_buffer_chk(__d, __bos0 (s))                           \
>>> +      : __d / __NFDBITS;                                             \
>>
>> I'm not happy that this isn't very conservative.
>>
>> If __bos0 fails should we fall back to static FD_SETSIZE checking
>> e.g. "__fdelt_buffer_warn (__d, FD_SETSIZE)"?
>>
>> It seems that that would be better than no checking.
> 
> Hmm.. This doesn't cross my mind. All other buffer boundary checks
> of _FORTIFY_SOURCE fall back no checking.  compiler may fails to
> determine a right buffer size in various reasons. at that time, I don't
> want to kill innocent applications.
> 
>> I know why you want to fall back to no check, because that
>> way you don't require any kind of new flag to disable the
>> check in the event it triggers when you don't want it to
>> (when __bos0 fails).
> 
> If you like flag, I'm not putting objection. but if making flag, a lot
> of libraries need
> to turn on "no check" mode because when a buffer is allocated from applications,
> library code can't know a buffer size at least at compile time.
> 
> 
>> Does compiling ruby (or similar code) with this header
>> result in calls to __fdelt_buffer_warn or __fdelt_buffer_chk?
> 
> Unfortunately, No. __builtin_object_size() require compiler know the
> buffer size.
> In the other words, it doesn't work if an allocate function and
> FD_{SET,CLR} functions
> doesn't exist in the same place. This is the same limitation with
> other string buffer
> overflow checks.

Then we need a flag, and ruby needs to use the flag to disable the
check on Linux.

The fundamental truth is that glibc implements POSIX, not "Linux."
And in POSIX there is a limit of FD_SETSIZE.

The default checking should be for POSIX.

We should provide a way to disable _FORTIFY_SOURCE checks that
are POSIX-only.

I still think your current macro is *better* because if __bos0
works then you have a dynamic check that is better than a static
check.

Thus the final solution is a combination of your new __bos0
changes and a flag to disable the check in the event that __bos0
fails.

What do you think?

Cheers,
Carlos.