It is a bug with gnu-make, but it was worked around by enabling 
glibc-malloc-compat in uClibc. I'd like to disable that malloc option again 
because it is often abused, but I think there are a couple blfs packages 
affected too.

robert

On March 28, 2006 02:07 pm, Osor wrote:
> Hi all, I noticed that a few months back, there was a problem installing a
> modular kernel (2.6.14.3-grsec) with uClibc (0.9.28). I don't know if this
> is still a problem with 2.6.14.6 or whatever. I did, however, figure out
> where the bug was: it wasn't in uClibc, it wasn't the in kernel (directly),
> and it wasn't even a typo in the HLFS book.
>
> This bug was located in...   the `make' program itself! The make program
> (v. 3.80) has a buffer-related bug in expand.c. A very long string in an
> eval command in a Makefile would cause an error, whereby make thinks it
> cannot allocate enough memory. As kernel makefiles are wont to be
> extra-complicated (but usually not buggy), there there was one Makefile in
> this specific kernel (2.6.14.3) configured with this specific C library
> (0.9.28), which controlled modules_install, which also broke make.
> (grammar?)
>
> As it turns out, this bug has already been submitted to the people over at
> make, and there are patches already available. Also the bug-fix has been
> incorporated into the the alpha branch for make.
>
> I don't even know if this applies, since the 2.6.14 tree has progressed
> from .3 to .7 (which, AFAIK, still works with the grsecurity patch).
>
> Take care,
> Osor
-- 
http://linuxfromscratch.org/mailman/listinfo/hlfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page

Reply via email to