Re: What I suspect

Olivier Galibert (galibert@pobox.com)
Thu, 9 Dec 1999 02:05:21 -0500


On Wed, Dec 08, 1999 at 07:19:07PM -0500, Brandon S. Allbery KF8NH wrote:
> Say I want to enable malloc checking at runtime. I use LD_PRELOAD for
> that... whoops! Doesn't work. In fact, the best I can hope for is complete
> ineffectiveness: it is likely to cause complete breakage, if some things
> use my LD_PRELOAD'ed malloc() et al. and others use the standard one.

But you know at rld time that you used LD_PRELOAD, so you can avoid
the fast path.

Look better at what Linus wants: a prelinked fast path for the usual
case. What he proposes stays valid if all the cases where the fast
path is inappropriate are either detected (LD_PRELOAD is easily
detectable, since the code which loads the libraries is the one that
handles it) or the lack of detection gives an acceptable result (lack
of override of a weak symbol in a library which has been changed after
the link and for which the symbol did not exist in the previous
version).

OG.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/