The check to test that "bus_no" was valid came after we had already used
it as an array offset. This patch moves it forward.
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
continue;
}
bus_no = buf[0]>>16;
+ if (bus_no >= MAX_CHANNEL) { /* Something wrong skip it */
+ printk(KERN_WARNING
+ "%s: Channel number %d out of range\n",
+ pHba->name, bus_no);
+ continue;
+ }
+
scsi_id = buf[1];
scsi_lun = (buf[2]>>8 )&0xff;
pDev = pHba->channel[bus_no].device[scsi_id];
adpt_i2o_report_hba_unit(pHba, d);
adpt_i2o_install_device(pHba, d);
- if(bus_no >= MAX_CHANNEL) { // Something wrong skip it
- printk(KERN_WARNING"%s: Channel number %d out of range \n", pHba->name, bus_no);
- continue;
- }
pDev = pHba->channel[bus_no].device[scsi_id];
if( pDev == NULL){
pDev =