Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Fri, 2 Nov 2001 18:42:35 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Fri, 2 Nov 2001 18:42:25 -0500 Received: from femail8.sdc1.sfba.home.com ([24.0.95.88]:56718 "EHLO femail8.sdc1.sfba.home.com") by vger.kernel.org with ESMTP id ; Fri, 2 Nov 2001 18:42:13 -0500 Message-ID: <3BE32EC9.842913E6@didntduck.org> Date: Fri, 02 Nov 2001 18:39:53 -0500 From: Brian Gerst X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.4.10 i686) X-Accept-Language: en MIME-Version: 1.0 To: Ken Ashcraft CC: linux-kernel@vger.kernel.org Subject: Re: null pointer questions In-Reply-To: <3BE32E24.E89E83CD@didntduck.org> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Brian Gerst wrote: > > Ken Ashcraft wrote: > > > > > > 2. What happens if I pass a null pointer as the destination parameter > > > > to copy_from_user? Does copy_from_user handle it safely or will the > > > > kernel seg fault? > > > > > > The kernel won't crash, but it might fail (depending on whether 0 is a > > > valid user space address or not). > > > > Why does it matter if 0 is a valid user space or not? If I make the call > > > > copy_from_user(0, user_ptr, 4); > > > > the null pointer is the kernel address, not the user address. Can you > > clarify please? > > copy_from_user uses the string move instruction on the x86, so the > exception code would assume the source faulted not the dest. It would > return -EFAULT instead of causing an oops. Err, would return non-zero instead of -EFAULT. -- Brian Gerst - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/