Henry,

Henry Ptasinski wrote:
On Mon, Sep 27, 2010 at 01:37:18PM -0700, Jason Cooper wrote:
@@ -1847,7 +1858,12 @@ dhd_add_if(dhd_info_t *dhd, int ifidx, void *handle, 
char *name,
        ASSERT(dhd && (ifidx < DHD_MAX_IFS));
ifp = dhd->iflist[ifidx];
-       if (!ifp && !(ifp = MALLOC(dhd->pub.osh, sizeof(dhd_if_t)))) {
+       if (!ifp) {
+               DHD_ERROR(("%s: dhd->iflist[ifidx] null\n", __func__));
+               return -EINVAL;
+       }
+       ifp = MALLOC(dhd->pub.osh, sizeof(dhd_if_t));
+       if (!ifp) {
                DHD_ERROR(("%s: OOM - dhd_if_t\n", __func__));
                return -ENOMEM;
        }

I think you changed the logic here from AND to OR.  I believe this would
be more correct:

        ifp = MALLOC(dhd->pub.osh, sizeof(dhd_if_t));
        if (!(dhd->iflist[ifidx]) && (!ifp)) {
                DHD_ERROR(("%s: OOM - dhd_if_t\n", __func__));
                return -ENOMEM;
        }


I was attempting to remove the checkpatch.pl error with as little interpretation 
as possible.  The current code executes the MALLOC() if and only if ifp != NULL.  
eg, if and only if dhd->iflist[ifidx] != NULL.

Do you really want to bail only when _both_ are NULL?  What if the MALLOC() 
failed?  Or, what if the dhd->iflist[ifidx] was NULL?


thx,

Jason.
_______________________________________________
devel mailing list
[email protected]
http://driverdev.linuxdriverproject.org/mailman/listinfo/devel

Reply via email to