Issue with function pointer

Question

I'm having a problem with a function pointer that is using the wrong arguments.

The portion of the code which is giving me trouble is this

654     for (k = 0; k < o->nk; k++) {
655       val[0] = o->f(o->xyz+3*k, o->constant, o->f_cent, o->f_sig, val+1);
          ...
659     }

The goal is to call a series of functions (whose pointers are stored in o->f), each of which is called o->nk times in this loop.

The type of o->f is

typedef double (*FLDfunc_t)(double x[], double A, double *ct, double *sig, double grd[]);

All functions run correctly, except for the last one (multsinfunc). Output from gdb is:

#0  0x000000000048712f in multsinfunc (xyz=0x3923c55, A_in=6.28318530717958, cent_in=0x3568ad7, sig_in=0x3568b6e, grad=0x3243f6a879aff)
at MODELS/fields/functions.c:176
#1  0x00000000004863c2 in do_field_operation (o=0x7fffffffdde0, mdl=0x7ffff7e11010, l=CELL_C, kvar=4, kfrom=0, kto=8)
at MODELS/fields/fields.c:655

However, when I go into "do_field_operation", the values for the parameters are different:

gdb$ p o->xyz+3*k
$13 = (double *) 0x918150

gdb$ p o->constant
$14 = 1

gdb$ p o->f_cent
$15 = (double *) 0x880130

gdb$ p o->f_sig
$16 = (double *) 0x880148

The result is that I get a segmentation fault when I try to access "xyz" in multsinfunc. Maybe this is obvious, but I have no idea why this happens. Everything seems to be declared properly. What could be causing this?

Maybe xyz+3*(o->nk-1) is out of bound. But it's only a guess — janisz
– janisz, Commented Jun 14, 2013 at 19:12

Casey · Accepted Answer · 2013-06-14 19:24:55Z

1

You are passing 5 arguments through your call of o->f, but - looking at the gdb output - do_field_operation has 6 parameters. Calling a function through a function pointer of incorrect type invokes undefined behavior.

answered Jun 14, 2013 at 19:24

Casey

42.9k7 gold badges101 silver badges127 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

DanielF Over a year ago

But do_field_operation is the one who calls o->f. I think I figured it out. -> and () have the same precedence. Adding some parenthesis around (o->f) seems to fix the problem, though my K&R book says they associate "left to right", so o->f should evaluate first anyway. That also doesn't explain why there were no errors with the other function pointers.

Collectives™ on Stack Overflow

Issue with function pointer

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related