On Sat, Mar 28, 2020 at 10:14:53PM +0100, Alexey Gladkov wrote:

On Sat, Mar 28, 2020 at 01:28:28PM -0700, Kees Cook wrote:

quoted

On Fri, Mar 27, 2020 at 06:23:30PM +0100, Alexey Gladkov wrote:

quoted

[...]
+	if (!kstrtouint(param->string, base, &result.uint_32)) {
+		ctx->hidepid = result.uint_32;

This need to bounds-check the value with a call to valid_hidepid(), yes?

The kstrtouint returns 0 on success and -ERANGE on overflow [1].

No, I mean, hidepid cannot be just any uint32 value. It must be in the
enum. Is that checked somewhere else? It looked like the call to
valid_hidepid() was removed.

-- 
Kees Cook