From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from moutng.kundenserver.de (moutng.kundenserver.de [212.227.126.186]) by ozlabs.org (Postfix) with ESMTP id 9E55CDDE0A for ; Fri, 5 Dec 2008 04:34:54 +1100 (EST) From: Arnd Bergmann To: Cliff Wickman , Mijo Safradin , Christoph Lameter , Lee Schermerhorn , linuxppc-dev@ozlabs.org Subject: [PATCH] numactl: fix libnuma on big-endian 64-bit systems Date: Thu, 4 Dec 2008 18:34:45 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Message-Id: <200812041834.45931.arnd@arndb.de> List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , The read-mask function assumes that it is running in 32-bit mode, by addressing the bitmask as a series of int values, instead of longs. This is broken as can easily be reproduced by running numademo on a bit-endian 64-bit system. Changing the addressing to use 'long' values fixes the problem. Reported-by: Mijo Safradin --- Note: the set_nodemask_size() function is broken as well, it seems to always set the nodemask size to "17" with the s2nbits implementation. The fallback path in there looks correct. --- a/libnuma.c 2008-12-04 14:25:30.000000000 +0100 +++ b/libnuma.c 2008-11-20 13:40:29.000000000 +0100 @@ -392,9 +372,9 @@ read_mask(char *s, struct bitmask *bmp) { char *end = s; char *prevend; - unsigned long *start = bmp->maskp; - unsigned long *p = start; - unsigned long *q; + unsigned int *start = (unsigned int *)bmp->maskp; + unsigned int *p = start; + unsigned int *q; unsigned int i; unsigned int n = 0; @@ -431,14 +411,14 @@ } /* Poor mans fls() */ - for(i = sizeof(long) * 8 - 1; i >= 0; i--) + for(i = 31; i >= 0; i--) if (test_bit(i, start + n)) break; /* * Return the last bit set */ - return ((sizeof(unsigned long)*8) * n) + i; + return ((sizeof(unsigned int)*8) * n) + i; } /*