Closed GoogleCodeExporter closed 8 years ago
The crash occurs in the file calls.c in the gcc directory.
if (stack_usage_map_buf) {
free (stack_usage_map_buf);
}
at the end of function : emit_library_call_value_1
The variable is set in the routine with :
#ifdef ARGS_GROW_DOWNWARD
highest_outgoing_arg_in_use = MAX (initial_highest_arg_in_use,
needed + 1);
#else
highest_outgoing_arg_in_use = MAX (initial_highest_arg_in_use,
needed);
#endif
stack_usage_map_buf = XNEWVEC (char, highest_outgoing_arg_in_use);
The call comes from a higher level routine in builtins.c :
expand_builtin_powi
Original comment by fred.tre...@googlemail.com
on 29 Oct 2008 at 2:42
With the current version of sx-gcc, the test case above compiles and executes.
The
same holds for the "pr19402-2.c" test case.
Fred, could you check if this issue also disappears for your sx-gcc
configuration?
Original comment by nou...@gmail.com
on 27 Nov 2008 at 10:19
I still see this problem. For example in the last test results, running on
the testmachine (clean svn).
http://source.tree.se/sx/sx-gcc/testrun/142-180-180-sx-2.19/gcc.log
/home/fred/tmp/sx/src/sx-gcc/gcc/testsuite/gcc.dg/pr19402-2.c: In function
'powicl':
/home/fred/tmp/sx/src/sx-gcc/gcc/testsuite/gcc.dg/pr19402-2.c:51: internal
compiler
error: Aborted
...
** glibc detected *** /home/fred/tmp/sx/sx-build/gcc/cc1: free(): invalid next
size
(fast): 0x0000000001f33430 ***
In what svn revision and testsetup did the failure disapear ?
Compiling on x86-64, running on sx6i.
Original comment by fred.tre...@googlemail.com
on 2 Dec 2008 at 10:33
Funny thing, on my computer (i686), with the current release (r150) of sx-gcc,
the
"pr19402-2.c" test case compiles just fine. The test also executes fine if I
run it
on SX8.
So it must be the x86_64 issue...
Original comment by nou...@gmail.com
on 2 Dec 2008 at 11:07
what's the status of this, folks? imho it should have been fixed with current
trunk
that includes all the fp magic we did since reported ... someone in the know,
pls
check and resolve.
Original comment by jmoc...@gmail.com
on 15 Jan 2009 at 12:31
caused by buffer overflow in emit_library_call_value_1. this has been fixed in
our
trunk since r225. patch also sent upstream.
Original comment by jmoc...@gmail.com
on 6 Mar 2009 at 12:22
Original issue reported on code.google.com by
fred.tre...@googlemail.com
on 29 Oct 2008 at 1:02