[PATCH] scsi/sr.c resource alloc checking

[Arnaldo Carvalho de Melo]

Alan/Jens,

        Please consider applying, a similar patch is already in 2.4. In
sr_init we don't need zeroing the data allocated with scsi_init_malloc, as
scsi_init_malloc already does this for us.

- Arnaldo

--- linux-2.2.18-pre25/drivers/scsi/sr.c        Sat Dec  9 15:08:24 2000
+++ linux-2.2.18-pre25.acme/drivers/scsi/sr.c   Sat Dec  9 17:33:54 2000
@@ -26,6 +26,8 @@
  *     Modified by Jens Axboe <[EMAIL PROTECTED]> - support DVD-RAM
  *     transparently and loose the GHOST hack
  *
+ *     Modified by Arnaldo Carvalho de Melo <[EMAIL PROTECTED]>
+ *     check resource allocation in sr_init and some cleanups - 2000/12/09
  */
 
 #include <linux/module.h>
@@ -70,10 +72,10 @@
        sr_finish, sr_attach, sr_detach
 };
 
-Scsi_CD *scsi_CDs = NULL;
-static int *sr_sizes = NULL;
+Scsi_CD *scsi_CDs;
+static int *sr_sizes;
 
-static int *sr_blocksizes = NULL;
+static int *sr_blocksizes;
 
 static int sr_open(struct cdrom_device_info *, int);
 void get_sectorsize(int);
@@ -1113,16 +1115,21 @@
        }
        if (scsi_CDs)
                return 0;
-       sr_template.dev_max =
-           sr_template.dev_noticed + SR_EXTRA_DEVS;
-       scsi_CDs = (Scsi_CD *) scsi_init_malloc(sr_template.dev_max * sizeof(Scsi_CD), 
GFP_ATOMIC);
-       memset(scsi_CDs, 0, sr_template.dev_max * sizeof(Scsi_CD));
-
-       sr_sizes = (int *) scsi_init_malloc(sr_template.dev_max * sizeof(int), 
GFP_ATOMIC);
-       memset(sr_sizes, 0, sr_template.dev_max * sizeof(int));
-
-       sr_blocksizes = (int *) scsi_init_malloc(sr_template.dev_max *
-                                                sizeof(int), GFP_ATOMIC);
+       sr_template.dev_max = sr_template.dev_noticed + SR_EXTRA_DEVS;
+       scsi_CDs = scsi_init_malloc(sr_template.dev_max * sizeof(Scsi_CD),
+                                       GFP_ATOMIC);
+       if (!scsi_CDs)
+               goto cleanup_register;
+
+       sr_sizes = scsi_init_malloc(sr_template.dev_max * sizeof(int),
+                                       GFP_ATOMIC);
+       if (!sr_sizes)
+               goto cleanup_cds;
+
+       sr_blocksizes = scsi_init_malloc(sr_template.dev_max * sizeof(int),
+                                       GFP_ATOMIC);
+       if (!sr_blocksizes)
+               goto cleanup_sizes;
 
        /*
         * These are good guesses for the time being.
@@ -1131,6 +1138,18 @@
                sr_blocksizes[i] = 2048;
        blksize_size[MAJOR_NR] = sr_blocksizes;
        return 0;
+cleanup_sizes:
+       scsi_init_free((char *) sr_sizes, sr_template.dev_max * sizeof(int));
+       sr_sizes = NULL;
+cleanup_cds:
+       scsi_init_free((char *) scsi_CDs,
+                      (sr_template.dev_noticed + SR_EXTRA_DEVS) *
+                      sizeof(Scsi_CD));
+       scsi_CDs = NULL;
+cleanup_register:
+       unregister_blkdev(MAJOR_NR, "sr");
+       sr_registered--;
+       return 1;
 }
 
 void sr_finish()
@@ -1241,7 +1260,7 @@
                scsi_init_free((char *) scsi_CDs,
                               (sr_template.dev_noticed + SR_EXTRA_DEVS)
                               * sizeof(Scsi_CD));
-
+               scsi_CDs = NULL;
                scsi_init_free((char *) sr_sizes, sr_template.dev_max * sizeof(int));
                sr_sizes = NULL;
 
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
Please read the FAQ at http://www.tux.org/lkml/