------- Additional Comments From dlstevens at us dot ibm dot com 2006-03-21 23:56 -------
OK, I did some new testing and I agree there is no bug. I was able to
reproduce some segmentation violations, which I think were coming from
the fact that I was mixing brk()'s with the sbrk()'s, to set the end of
data segment back after growing. I expect doing debugging printf's after
an sbrk() resulted in internal glibc allocations which a later brk() was
then making invalid pointers. :-) I haven't seen any SEGV's when the sbrk()
and brk() don't have anything in between them.
I don't have the original malloc() test, but I was likely doing something
similarly silly there.
Sorry for the confusion, and thanks for all your efforts! I agree this one can