Bug 85313

From Bugzilla Helper:
User-Agent: Mozilla/4.7 [en] (X11; I; Linux 2.2.13 i686)

Description of problem:
I've verified that this problem affects at least raidtools 0.90 through 1.00
(Red Hat 7.2, 7.3, 8.0, and Advanced Server 2.1, at least).

In the code that checks to see if the array is already active (check_active(),
raid_io.c), the device matching is not done properly, and substring matches are
mistakenly considered matches. 

The patch (against raidtools-1.00.2-1.3, from Red Hat 7.3) for the fix is:

--- raidtools-1.00.2/raid_io.c  Mon Apr 15 04:09:11 2002
+++ raidtools-1.00.2steeleye/raid_io.c  Thu Feb 27 15:22:07 2003
@@ -515,8 +515,10 @@
 }
 
 
+
 int check_active (md_cfg_entry_t *p)
 {
+       char md_dev_chars[] = "0123456789md";
        char buffer[MAX_LINE_LENGTH], line[MAX_LINE_LENGTH], *ch;
        FILE *fp;
 
@@ -528,11 +530,15 @@
        while (1) {
                if ((fgets(line, MAX_LINE_LENGTH, fp)) == NULL)
                        break;
-               if (strstr(line, buffer) && !strstr(line, "inactive")) {
-                       fprintf(stderr, "%s: array is active -- run raidstop fir
st.\n", p->md_name);
-                       fclose(fp);
-                       return 1;
-               }
+               if (!(ch = strstr(line, buffer)))
+                       continue;
+               if (strspn(ch, md_dev_chars) != strspn(buffer, md_dev_chars))
+                       continue;
+               if (strstr(line, "inactive"))
+                       continue;
+               fprintf(stderr, "%s: array is active -- run raidstop first.\n", 
p->md_name);
+               fclose(fp);
+               return 1;
        }
        fclose(fp);
        return 0;





Version-Release number of selected component (if applicable):


How reproducible:
Always

Steps to Reproduce:
1. mkraid -c <conf_file> /dev/md10

2. mkraid -c <conf_file> /dev/md1     **ALWAYS FAILS**
    

Additional info: