> Assuming that you are using the older libc and not the newer uClibc, it
> is a known problem with libc. There were some patches to correct for
> this, about a year ago. Otherwise, if this is uClibc that is giving
> this problem, then tell the guys over on the uClibc mailing list, I am
> sure they would want to know!
I'm not sure that this is the case, at least not in what I see,
because the malloc() in the 'old' uClibc simply calls mmap()
and the memory leak in fopen() is really just due to the
memory leak as a side-effect of malloc().
Which patch from 'about a year ago' are you refering to?
I can't see how the 'new' uClibc can fix mmap() (other than
by a workaround like the memory carving that my patches to
Are you sure this problem is gone in the 'new' uClibc under
m68k uClinux 2.0.x? Most people don't seem to run across
this issue, probably because they don't leave their uCsimm
running very long or don't have deamons that run forever
(e.g. inetd exec/exit).
And I'm not so sure the uClibc folks really are interested in
this problem since they seem to be taking uCibc down the x86
direction and few of them seem to be using m68k uClinux. I'm
not even sure anybody over there tests anything on m68k anymore.
This message resent by the email@example.com list server http://www.uClinux.com/
This archive was generated by hypermail 2b30 : Sun Apr 07 2002 - 00:01:43 EST