From: Andreas Dilger Subject: Re: [PATCH] libext2fs: Always swab MMP block on BE machines Date: Fri, 30 Sep 2011 15:09:53 -0600 Message-ID: <322E8AB3-3078-4854-8D8B-38ABA988B089@dilger.ca> References: <20110930194126.GU12086@tux1.beaverton.ibm.com> Mime-Version: 1.0 (Apple Message framework v1084) Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 8BIT Cc: "Theodore Ts'o" , linux-ext4 To: djwong@us.ibm.com Return-path: Received: from idcmail-mo2no.shaw.ca ([64.59.134.9]:39719 "EHLO idcmail-mo2no.shaw.ca" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751413Ab1I3VJy convert rfc822-to-8bit (ORCPT ); Fri, 30 Sep 2011 17:09:54 -0400 In-Reply-To: <20110930194126.GU12086@tux1.beaverton.ibm.com> Sender: linux-ext4-owner@vger.kernel.org List-ID: On 2011-09-30, at 1:41 PM, Darrick J. Wong wrote: > The MMP code in libext2fs tries to gate MMP block swab'ing with this test: > > #ifdef EXT2FS_ENABLE_SWAPFS > if (fs->super->s_magic == ext2fs_swab16(EXT2_SUPER_MAGIC)) We've never run Lustre on big-endian servers (clients generally work), so it's possible this has been in our patch for years. Thanks for the catch. > However, EXT2FS_ENABLE_SWAPFS never seems to be defined anywhere (all possible > #define's were cleaned out of e2fsprogs ages ago), and even if any still > existed, the field fs->super->s_magic is always in host byteorder, so the test > always fails. So, we can change the #ifdef to WORDS_BIGENDIAN (which is > conditionally defined on BE platforms) and get rid of the broken if test. > > (This fix came up while testing the metadata checksumming patchset) > > Signed-off-by: Darrick J. Wong > --- > > lib/ext2fs/mmp.c | 15 ++++++--------- > 1 files changed, 6 insertions(+), 9 deletions(-) > > diff --git a/lib/ext2fs/mmp.c b/lib/ext2fs/mmp.c > index ed6ee42..91f4fb2 100644 > --- a/lib/ext2fs/mmp.c > +++ b/lib/ext2fs/mmp.c > @@ -91,9 +91,8 @@ errcode_t ext2fs_mmp_read(ext2_filsys fs, blk64_t mmp_blk, void *buf) > } > > mmp_cmp = fs->mmp_cmp; > -#ifdef EXT2FS_ENABLE_SWAPFS > - if (fs->flags & EXT2_FLAG_SWAP_BYTES) > - ext2fs_swap_mmp(mmp_cmp); > +#ifdef WORDS_BIGENDIAN > + ext2fs_swap_mmp(mmp_cmp); > #endif > > if (buf != NULL && buf != fs->mmp_cmp) > @@ -122,18 +121,16 @@ errcode_t ext2fs_mmp_write(ext2_filsys fs, blk64_t mmp_blk, void *buf) > fs->super->s_mmp_block > ext2fs_blocks_count(fs->super)) > return EXT2_ET_MMP_BAD_BLOCK; > > -#ifdef EXT2FS_ENABLE_SWAPFS > - if (fs->super->s_magic == ext2fs_swab16(EXT2_SUPER_MAGIC)) > - ext2fs_swap_mmp(mmp_s); > +#ifdef WORDS_BIGENDIAN > + ext2fs_swap_mmp(mmp_s); > #endif > > /* I was tempted to make this use O_DIRECT and the mmp_fd, but > * this caused no end of grief, while leaving it as-is works. */ > retval = io_channel_write_blk64(fs->io, mmp_blk, -fs->blocksize, buf); > > -#ifdef EXT2FS_ENABLE_SWAPFS > - if (fs->super->s_magic == ext2fs_swab16(EXT2_SUPER_MAGIC)) > - ext2fs_swap_mmp(mmp_s); > +#ifdef WORDS_BIGENDIAN > + ext2fs_swap_mmp(mmp_s); > #endif > > /* Make sure the block gets to disk quickly */ > -- > To unsubscribe from this list: send the line "unsubscribe linux-ext4" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html Cheers, Andreas