gpu: nvgpu: Fix coverity defect in page allocator

Fix use-after-free overity defect in page allocator. The alloc struct
was getting used after a call for __gk20a_free_pages() which frees
the alloc struct passed in.

Coverity ID: 468942
Bug 200192125

Change-Id: I4f5d32f245efae967050f93c7806290b4bf3591c
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1272730
GVS: Gerrit_Virtual_Submit
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>

drivers/gpu/nvgpu/gk20a/gk20a_allocator_page.c

@@ -734,19 +734,20 @@ static void gk20a_page_free_fixed(struct gk20a_allocator *__a,
 		alloc = (struct gk20a_page_alloc *) (uintptr_t) base;
 	}
 
+	palloc_dbg(a, "Free  [fixed] 0x%010llx + 0x%llx\n",
+		   alloc->base, alloc->length);
+
+	a->nr_fixed_frees++;
+	a->pages_freed += (alloc->length >> a->page_shift);
+
 	/*
 	 * This works for the time being since the buddy allocator
 	 * uses the same free function for both fixed and regular
 	 * allocs. This would have to be updated if the underlying
 	 * allocator were to change.
 	 */
-	palloc_dbg(a, "Free  [fixed] 0x%010llx + 0x%llx\n",
-		   alloc->base, alloc->length);
 	__gk20a_free_pages(a, alloc, true);
 
-	a->nr_fixed_frees++;
-	a->pages_freed += (alloc->length >> a->page_shift);
-
 done:
 	alloc_unlock(__a);
 }