From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
        id S1755529AbcIJJve (ORCPT <rfc822;w@1wt.eu>);
        Sat, 10 Sep 2016 05:51:34 -0400
Received: from ozlabs.org ([103.22.144.67]:38532 "EHLO ozlabs.org"
        rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
        id S1750861AbcIJJvc (ORCPT <rfc822;linux-kernel@vger.kernel.org>);
        Sat, 10 Sep 2016 05:51:32 -0400
From: Michael Ellerman <mpe@ellerman.id.au>
To: Meelis Roos <mroos@linux.ee>,
        Linux Kernel list <linux-kernel@vger.kernel.org>,
        linuxppc-dev@lists.ozlabs.org
Subject: Re: UBSAN: Undefined behaviour in arch/powerpc/kernel/cputable.c
In-Reply-To: <alpine.LRH.2.20.1609091556420.29746@math.ut.ee>
References: <alpine.LRH.2.20.1609091556420.29746@math.ut.ee>
User-Agent: Notmuch/0.21 (https://notmuchmail.org)
Date: Sat, 10 Sep 2016 19:51:29 +1000
Message-ID: <87r38snkv2.fsf@concordia.ellerman.id.au>
MIME-Version: 1.0
Content-Type: text/plain
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

Meelis Roos <mroos@linux.ee> writes:

> Got this on my PowerPC G4 (533 MHz, PowerPC 7410, PowerMac3,4 platform), 
> with gcc version 5.3.1 20160509 (Debian 5.3.1-19):
>
> [    0.000000] ================================================================================
> [    0.000000] UBSAN: Undefined behaviour in arch/powerpc/kernel/cputable.c:2240:25
> [    0.000000] shift exponent 32 is too large for 32-bit type 'long unsigned int'
> [    0.000000] CPU: 0 PID: 0 Comm: swapper Not tainted 4.8.0-rc5-00075-gd71f058 #106
> [    0.000000] Call Trace:
> [    0.000000] [c0a5ff10] [c03aa588] ubsan_epilogue+0x18/0x4c (unreliable)
> [    0.000000] [c0a5ff20] [c03aab38] __ubsan_handle_shift_out_of_bounds+0xe8/0x13c
> [    0.000000] [c0a5ffa0] [c07c42d0] cpu_feature_keys_init+0xa0/0xb8
> [    0.000000] [c0a5ffd0] [c07ca1fc] setup_feature_keys+0x14/0x24
> [    0.000000] [c0a5ffe0] [c07c6c74] machine_init+0x18/0x58
> [    0.000000] [c0a5fff0] [c000341c] start_here+0x38/0x78
> [    0.000000] ================================================================================

Oops, my bad.

#define NUM_CPU_FTR_KEYS	64

...

	for (i = 0; i < NUM_CPU_FTR_KEYS; i++) {
		unsigned long f = 1ul << i;

Does this fix it?

cheers

diff --git a/arch/powerpc/include/asm/cpu_has_feature.h b/arch/powerpc/include/asm/cpu_has_feature.h
index 2ef55f8968a2..b312b152461b 100644
--- a/arch/powerpc/include/asm/cpu_has_feature.h
+++ b/arch/powerpc/include/asm/cpu_has_feature.h
@@ -15,7 +15,7 @@ static inline bool early_cpu_has_feature(unsigned long feature)
 #ifdef CONFIG_JUMP_LABEL_FEATURE_CHECKS
 #include <linux/jump_label.h>
 
-#define NUM_CPU_FTR_KEYS	64
+#define NUM_CPU_FTR_KEYS	BITS_PER_LONG
 
 extern struct static_key_true cpu_feature_keys[NUM_CPU_FTR_KEYS];