C language: Segmentation Fault: 11. Unclear error

Question

I am trying to use this c program to do the numerical modeling. It has been successfully compiled. However, when I am trying to input:

./forward 0.0 0.0 1.0 3.00 100.0.

It keeps reminding me: segmentation fault: 11.

I do not know what goes wrong. My code is as following:

#include<stdio.h>
#include<stdlib.h>
#include<string.h>
#include<math.h>
#include"util_err.h"
#define pi 3.1415926
#define MAX_Array 16000

typedef struct source {
    float vp;
    float ev_radius;
} source;

source model_search(float evdp, float a_radius, char *file){
    int i;
    float *radius=NULL,*alpha_s=NULL,*beta_s=NULL;
    float *z_s=NULL;
    source sour;
    FILE *f;
    /*radius = vector(0,MAX_Array-1);
    alpha_s = vector(0,MAX_Array-1);
    beta_s = vector(0,MAX_Array-1);
    z_s = vector(0,MAX_Array-1);*/
    f = fopen(file,"r");
    if (f==NULL)
    {
        nrerror("unable to open the file");
    }
    sour.vp = 0;
    sour.ev_radius = 0;
    for(i=0;!feof(f);i++){
        fscanf(f,"%f %f %f",&radius[i],&alpha_s[i],&beta_s[i]);
        z_s[i] = a_radius - radius[i];
        if(i==0) continue;
        if(evdp>=z_s[i-1]&&evdp<=z_s[i]){
            sour.vp = alpha_s[i-1]+(evdp-z_s[i-1])/(z_s[i]-z_s[i-1])
                *(alpha_s[i]-alpha_s[i-1]);
            sour.ev_radius = a_radius-evdp;
            printf("vp = %f\n",sour.vp);
            printf("ev_radius = %f\n",sour.ev_radius);
            break;
        }
    } 
    fclose(f);  
    return sour;
}

float* linespace(float fl, float fr, int n){
    int i;
    static float *u=NULL,interval;
    interval = (fr-fl)/(n-1);
    for(i=0;i<n-1;i++){
        u[i] = fl + i * interval;
    }
    u[n-1] = fr;
    return u;
}
/* Input: x, y, z, t0, evdp*/
int main(int argc,char **argv){
    int i,j;
    int sto,sto1,sa,sa1;
    float x,y,z,origin_time;
    float travel_time,evdp,a_radius=6371.1;
    float *takeoff=NULL,*takeoff1=NULL,*azi=NULL,*azi1=NULL;
    float t0,az;
    char filename[160];
    source sour;
    FILE *fp=NULL,*fp1=NULL;
    fp=fopen("data.dat","w");

    if(fp==NULL){
        nrerror("unable to allocate");
    }
    fp1=fopen("data1.dat","w");
    if(fp1==NULL){
        nrerror("unable to allocate");
    }
    if(argc<6){
        nrerror("Too few arguments!");
    }
    x = atof(argv[1]);
    y = atof(argv[2]);
    z = atof(argv[3]);
    origin_time = atof(argv[4]);
    evdp = atof(argv[5]);
    strcpy(filename,"/Users/a123/Desktop/Ear_relo/plot_seis/iasp91.dat");
    sour = model_search(evdp,a_radius,filename);
    printf("vp = %f, ev_radius = %f",sour.vp,sour.ev_radius);
    sto = 91;
    sto1 = 10;
    sa = 361;
    sa1 = 37;
    takeoff = linespace(0.0,90.0,sto);
    takeoff1 = linespace(0.0,90.0,sto1);
    azi = linespace(0.0,360.0,sa);
    azi1 = linespace(0.0,360.0,sa1);
/*  write data.dat for the background values */
    for(i=0;i<sto;i++){
        t0 = takeoff[i];
        for(j=0;j<sa;j++){
            az = azi[j];
            travel_time = -x*cos(az*pi/180)*sin(t0*pi/180)
                -y*sin(az*pi/180)*sin(t0*pi/180)+z*cos(t0*pi/180);
            travel_time = travel_time/sour.vp + origin_time;
            fprintf(fp, "%5.1f %5.1f %13.8f\n", az,t0,travel_time); 
        }
    }
/* write data1.dat */
    for(i=0;i<sto1;i++){
        t0 = takeoff1[i];
        for(j=0;j<sa1;j++){
            az = azi1[j];
            travel_time = -x*cos(az*pi/180)*sin(t0*pi/180)
                -y*sin(az*pi/180)*sin(t0*pi/180)+z*cos(t0*pi/180);
            travel_time = travel_time/sour.vp + origin_time;
            fprintf(fp1, "%5.1f %5.1f %13.8f\n", az,t0,travel_time);
        }
    }
    free(takeoff);
    free(takeoff1);
    free(azi);
    free(azi1);
    fclose(fp);
    fclose(fp1);
    return 0;

}

This is my iasp91.dat file:

6371 5.8 3.36
6370 5.8 3.36
6369 5.8 3.36
6368 5.8 3.36
6367 5.8 3.36
6366 5.8 3.36
6365 5.8 3.36
6364 5.8 3.36
6363 5.8 3.36
6362 5.8 3.36
6361 5.8 3.36
6360 5.8 3.36
6359 5.8 3.36
6358 5.8 3.36
6357 5.8 3.36
6356 5.8 3.36
6355 5.8 3.36
6354 5.8 3.36
6353 5.8 3.36
6352 5.8 3.36
6351 5.8 3.36
6351 6.5 3.75
6350 6.5 3.75
6349 6.5 3.75
6348 6.5 3.75
6347 6.5 3.75
6346 6.5 3.75
6345 6.5 3.75
6344 6.5 3.75
6343 6.5 3.75
6342 6.5 3.75
6341 6.5 3.75
6340 6.5 3.75
6339 6.5 3.75
6338 6.5 3.75
6337 6.5 3.75
6336 6.5 3.75
6336 8.04 4.47
6331 8.0406 4.4718
6326 8.0412 4.4735
6321 8.0418 4.4753
6311 8.0429 4.4788
6301 8.0441 4.4824
6291 8.0453 4.4859
6281 8.0465 4.4894
6271 8.0476 4.4929
6261 8.0488 4.4965
6251 8.05 4.5
6251 8.05 4.5
6241 8.0778 4.502
6231 8.1056 4.504
6221 8.1333 4.506

And the "util_err.h" declares the function --- nrerror.

I do not know what the exact problem is but it seems to me that the problem is in the block linespace.

Answer 1

source model_search(float evdp, float a_radius, char *file) {
    float *radius=NULL, ...
        fscanf(f,"%f %f %f", &radius[i] ...

&radius[i] is here equal to (uintptr_t)NULL + sizeof(float) * i which is an invalid pointer value. fscanf tries to dereference thater point and assign float value to it. Dereferencing invalid pointer value is undefined behaviour and the assigment most probably results in seg fault on unix systems.

1. MCVE is the minimal example needed to reproduce the error. One could debug your program with simple printf statements until seeing which statement produces seg fault.

2. It sounds like you may need to learn how to use a debugger to step through your code. With a good debugger, you can execute your program line by line and see where it is deviating from what you expect. This is an essential tool if you are going to do any programming. Further reading: How to debug small programs]

3. In code you have for(...; feof(...); ...) which is always wrong. See Why is “while ( !feof (file) )” always wrong?. In your case on the last line, the fscanf finds EOF, so the last line is incorrectly handled in your code. You should check the return value of fscanf.
4. Check errors. Try to check all errors.
5. In main() the float *takeoff=NULL,*takeoff1=NULL,*azi=NULL,*azi1=NULL; are never assigned a valid pointer, only NULLs. Inside linspace you check if (... && u == 0) return 0;, so takeoff = linspace(..., takeoff) still results in takeoff = NULL. Accessing NULL at t0 = takeoff[i]; inside main() is undefined behavior and most probably will seg fault too.
6. It's really a personal taste, but for simple programmer errors, it's nice to use assertions. linspace(...., float u[]) { assert(u != NULL); ...}.