Re: [PATCH] Add MALLOC_CHECK_ and MALLOC_PERTURB_ libc env to the test suite for detecting heap corruption

From: Junio C Hamano <gitster@pobox.com>
Date: Sep 17, 2012, 20:28 UTC
Message-ID: <7vzk4otmwb.fsf@alter.siamese.dyndns.org>
In-Reply-To: <CA+EOSBk0hN+v+PK8jzCCZgG0Ko7SsHOSB=cM7nnaSsrUOaJ-7Q@mail.gmail.com>

Elia Pinto <gitter.spiros@gmail.com> writes:

Show 12 quoted lines

>>  - That "165" thing I mentioned earlier.
>
> Thank you so much for the comments, that's fine. A single
> consideration for  MALLOC_PERTURB.
>
> You can use any value between 1..255 for MALLOC_PERTURB_
> That chooses the byte that glibc will use to memset all freed buffers.
> In general it is defined as
>
>     export MALLOC_PERTURB_=$(($RANDOM % 255 + 1))
>
> (as drepper pointed out http://udrepper.livejournal.com/11429.html)

Drepper never recommends RANDOM there.

> Using a random value is slightly better than using a fixed one
> in case your fixed value is someday just the right/wrong value to mask
> a problem.

Quite the contrary. When you use a fixed pattern, it is easy which other pieces of memory has uninitailized contents. When you use a random value, you sometimes get an error and sometimes the test mysteriously pass, which does not help debugging.

openSUSE folks seem to use a fixed value for this exact reason of repeatability of tests.

http://jaegerandi.blogspot.com/2012/01/finding-subtile-malloc-bugs.html

> So OK per the original expression?

No.

I am not convinced 165 is the perfect value, but I am fairly certain any fixed value is better than using a random to deliberately worsen repeatability of the tests.