From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754128AbdBGMfF (ORCPT ); Tue, 7 Feb 2017 07:35:05 -0500 Received: from foss.arm.com ([217.140.101.70]:51164 "EHLO foss.arm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753807AbdBGMfE (ORCPT ); Tue, 7 Feb 2017 07:35:04 -0500 From: Mark Rutland To: linux-arm-kernel@lists.infradead.org Cc: linux-kernel@vger.kernel.org, Mark Rutland , Catalin Marinas , Kees Cook , Will Deacon Subject: [PATCH] arm64: uaccess: consistently check object sizes Date: Tue, 7 Feb 2017 12:33:55 +0000 Message-Id: <1486470835-25956-1-git-send-email-mark.rutland@arm.com> X-Mailer: git-send-email 1.9.1 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Currently in arm64's copy_{to,from}_user, we only check the source/destination object size if access_ok() tells us the user access is permissible. However, in copy_from_user() we'll subsequently zero any remainder on the destination object. If we failed the access_ok() check, that applies to the whole object size, which we didn't check. To ensure that we catch that case, this patch hoists check_object_size() to the start of copy_from_user(), matching __copy_from_user() and __copy_to_user(). To make all of our uaccess copy primitives consistent, the same is done to copy_to_user(). Signed-off-by: Mark Rutland Cc: Catalin Marinas Cc: Kees Cook Cc: Will Deacon --- arch/arm64/include/asm/uaccess.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) Kees, Was there any rationale for not handling the !access_ok() case? I note that other architectures follow the same pattern, and may need a similar fixup. Thanks, Mark. diff --git a/arch/arm64/include/asm/uaccess.h b/arch/arm64/include/asm/uaccess.h index 46da3ea..5308d69 100644 --- a/arch/arm64/include/asm/uaccess.h +++ b/arch/arm64/include/asm/uaccess.h @@ -379,9 +379,9 @@ static inline unsigned long __must_check copy_from_user(void *to, const void __u { unsigned long res = n; kasan_check_write(to, n); + check_object_size(to, n, false); if (access_ok(VERIFY_READ, from, n)) { - check_object_size(to, n, false); res = __arch_copy_from_user(to, from, n); } if (unlikely(res)) @@ -392,9 +392,9 @@ static inline unsigned long __must_check copy_from_user(void *to, const void __u static inline unsigned long __must_check copy_to_user(void __user *to, const void *from, unsigned long n) { kasan_check_read(from, n); + check_object_size(from, n, true); if (access_ok(VERIFY_WRITE, to, n)) { - check_object_size(from, n, true); n = __arch_copy_to_user(to, from, n); } return n; -- 1.9.1