On 3/6/18 9:05 AM, J Freyensee wrote:

quoted

  +#ifdef CONFIG_PROTECTABLE_MEMORY
+void lkdtm_WRITE_RO_PMALLOC(void)
+{
+    struct gen_pool *pool;
+    int *i;
+
+    pool = pmalloc_create_pool("pool", 0);
+    if (unlikely(!pool)) {
+        pr_info("Failed preparing pool for pmalloc test.");
+        return;
+    }
+
+    i = (int *)pmalloc(pool, sizeof(int), GFP_KERNEL);
+    if (unlikely(!i)) {
+        pr_info("Failed allocating memory for pmalloc test.");
+        pmalloc_destroy_pool(pool);
+        return;
+    }
+
+    *i = INT_MAX;
+    pmalloc_protect_pool(pool);
+
+    pr_info("attempting bad pmalloc write at %p\n", i);
+    *i = 0;

Opps, disregard this, this is the last series of this patch series, not 
the most recent one :-(.

Seems harmless, but I don't get why *i local variable needs to be set 
to 0 at the end of this function.


Otherwise,

Reviewed-by: Jay Freyensee <redacted>