From: "Aneesh Kumar K.V" Subject: Re: ext4 compile error on m68k Date: Thu, 10 Apr 2008 19:28:07 +0530 Message-ID: <20080410135807.GA7526@skywalker> References: <20080405102408.GB30987@cs181133002.pp.htv.fi> <20080410104059.GA23116@skywalker> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Geert Uytterhoeven , Adrian Bunk , sct@redhat.com, akpm@linux-foundation.org, adilger@clusterfs.com, zippel@linux-m68k.org, linux-ext4@vger.kernel.org, linux-m68k@vger.kernel.org To: Andreas Schwab Return-path: Received: from E23SMTP06.au.ibm.com ([202.81.18.175]:37374 "EHLO e23smtp06.au.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755305AbYDJN6b (ORCPT ); Thu, 10 Apr 2008 09:58:31 -0400 Content-Disposition: inline In-Reply-To: Sender: linux-ext4-owner@vger.kernel.org List-ID: On Thu, Apr 10, 2008 at 01:48:55PM +0200, Andreas Schwab wrote: > "Aneesh Kumar K.V" writes: > > > But m68k doesn't want to use the GENERIC_FIND_NEXT_BIT. Can I request > > somebody knowledgeable about m68k to give a try in implementing > > generic_find_next_le_bit equivalent on m68k ? > > What's the exact semantic of generic_find_next_le_bit? Why is it not > documented? > Should be same ext2_find_next_zero_bit except that it find the next set bit rather than the zero bit. The API is used to find the set bit in bitmap. To explain the little endian dependency, what we have is 0 x n [.......1.......] Now we read this as (char *) and try to find the first set bit which in the above example is x. To speed up most of the architectures would like to consider it as an array of long. That implies that the value in the array should be considered as a little endian long, hence the *_le_bit. -aneesh