From: Coly Li Subject: Re: [PATCH 0/2] Add inode checksum support to ext4 Date: Sat, 30 Jul 2011 15:25:32 +0800 Message-ID: <4E33B1EC.9030004__29028.6702839241$1312010505$gmane$org@gmail.com> References: <20110406224410.GB24354@tux1.beaverton.ibm.com> <1302290868.4461.7.camel@mingming-laptop> <20110727082730.GG20655@tux1.beaverton.ibm.com> <20110728165615.GI20655@tux1.beaverton.ibm.com> <20110728220735.GA27253@noexit.corp.google.com> <0E795C1D-AD1E-4CC4-9426-2B58D98B14DC@dilger.ca> <20110729131937.GB5910@noexit.corp.google.com> Reply-To: colyli@gmail.com Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE To: Andreas Dilger , "Darrick J. Wong" , Andreas Dilger , Mingming Cao , Theodore Ts'o , l Return-path: Received: from oproxy3-pub.bluehost.com ([69.89.21.8]:40928 "HELO oproxy3-pub.bluehost.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1753112Ab1G3HVY (ORCPT ); Sat, 30 Jul 2011 03:21:24 -0400 In-Reply-To: <20110729131937.GB5910@noexit.corp.google.com> Sender: linux-ext4-owner@vger.kernel.org List-ID: On 2011=E5=B9=B407=E6=9C=8829=E6=97=A5 21:19, Joel Becker Wrote: > On Fri, Jul 29, 2011 at 03:48:45AM -0600, Andreas Dilger wrote: >> On 2011-07-28, at 4:07 PM, Joel Becker wrote: >>> We use ethernet crc32 in ocfs2. btrfs uses crc32c. Frankly, I >>> could have used crc32c if I'd really thought about the hardware >>> acceleration benefits. I think it's a good idea for ext4. >> >> The problem with crc32[c] is that if you don't have hardware acceler= ation >> it is terribly slow. >=20 > We find ethernet crc32 just fine in ocfs2. I use the kernel's > implementation, which survives everyone's network traffic, and of cou= rse > we added the triggers to jbd2 so we only have to do the calculations = on > read and write. >=20 Ext4 supports non-journal mode, and there are a few users (Google, Taob= ao, etc.). A trigger of jbd2 may not work well for non-journal Ext4 ... And in non-journal mode, there is not copy of any meta data block in jb= d2, we need to be more careful in check summing, e.g. inode/block bitmap blocks... >> Yes, it makes sense to just put a "fake" dirent at the end of the le= af block, >> or similar. I don't think it is necessary to modify existing direct= ories or >> extent blocks to add these structures in, if there is no room, but f= or new >> blocks, or blocks with space it is enough. >=20 > We have tunefs.ocfs2 code to adjust existing directories to add > the trailer. It's not too bad, really. Agree, enable/disable the trailer isn't that difficult. We need more ey= es to take care the user space tools. Coly -- To unsubscribe from this list: send the line "unsubscribe linux-ext4" i= n the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html