onn@utgpu.UUCP (09/02/87)
I am having problems with fmalloc() on SCO Xenix 2.1.3. I am compiling
an assembler, and sometimes the machine will dump core, with a segmentation
violation. Adb stack backtrace shows that it is stopping in fmalloc(), called
with an argument of 6. Actually it stops in a function of fmalloc() called
_amalloc().
Since I can't see any problems with the arguments to fmalloc(), I suspect a
problem within fmalloc(), probably something to do with the segmented
architecture of the 80286.
I am compiling the source code with 'cc -Ml -LARGE -O -o asm *.c'. There
are no compilation errors or warnings from the compiler, other than something
to the effect saying that function yyparse() is to large for post optimization.
The code has been linted.
The actual calls in the source are calls to malloc(), but the linker seems
to change these appropriately to fmalloc().
BTW, this program compiles and runs properly on VAXen running both 4.2 and
SVR2, on NCR Towers running SVR2, and on a Sun, running Sun release 3.2.
Does anyone have any ideas?
Thanks,
Brian.
--
--------------
Brian A. Onn
University of Toronto Computing Services
Erindale College
...{utzoo, ihnp4, decvax}!utcs!onn