linux-ext4 - Re: [Ext2-devel] rfc: [patch 3/6] more e2fsprogs fixes

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Date:	Mon, 4 Sep 2006 16:59:41 +0200
From:	Alexandre Ratchov <alexandre.ratchov@...l.net>
To:	linux-ext4@...r.kernel.org
Subject: Re: [Ext2-devel] rfc: [patch 3/6] more e2fsprogs fixes

On Fri, Aug 25, 2006 at 05:50:49PM -0600, Andreas Dilger wrote:
> On Aug 25, 2006  15:18 +0200, Alexandre Ratchov wrote:
> > add bg_{block_bitmap,inode_bitmap,inode_table}_hi fields in
> > struct ext2_group_desc and use absolute block numbers. While
> > this is not necessary, IMHO this is simpler than relative
> > addressing, and this change make e2fsprogs "just" work.
> > 
> >  {
> > +	__u32	bg_block_bitmap;	/* Blocks bitmap block */
> > +	__u32	bg_inode_bitmap;	/* Inodes bitmap block */
> >  	__u32	bg_inode_table;		/* Inodes table block */
> >  	__u16	bg_free_blocks_count;	/* Free blocks count */
> >  	__u16	bg_free_inodes_count;	/* Free inodes count */
> >  	__u16	bg_used_dirs_count;	/* Directories count */
> >  	__u16	bg_flags;
> > +	__u32	bg_block_bitmap_hi;	/* Blocks bitmap block */
> > +	__u32	bg_inode_bitmap_hi;	/* Inodes bitmap block */
> > +	__u32	bg_inode_table_hi;	/* Inodes table block */
> > +	/*__u16	bg_reserved[3];*/
> 
> This consumes all of the blockgroup reserved fields, not leaving any
> for checksumming.  In particular, for the UNINIT group changes this
> needs at least 16 bits for the checksum, and it would be good to have
> a 32-bit checksum for the block and inode bitmaps.
> 
> Instead, why not put a 16-bit "_block_bitmap_hi" field here and then
> the inode bitmap and inode table are relative to that?  Alternately,
> if we want a full 32-bit "hi" value, we can maybe live with smaller
> checksums on the bitmaps...
>  

ok, that works for me. However there is still an issue:

in order tu use 48bit adressing, we have to use larger block groups (because
all group descriptors must fit in one block group).  So we'll need 16bit
"_hi" bits for blocks_count, inodes_count and perhaps used_dirs_count. And
this leaves no room for additional fields.

There's already a patch that makes the descriptor structure size variable
(similarly to the inode size), see marc.theaimsgroup.com/?m=114968738307171
Here is the updated patch set:

linux-2.6.18-rc4:

http://www.bullopensource.org/ext4/20060904/ext4-linux-2.6.18-rc4.tar.gz

e2fsprogs-1.39:

http://www.bullopensource.org/ext4/20060904/ext4-e2fsprogs-1.39.tar.gz

cheers,

-- 
Alexandre
-
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html