linux-kernel - Re: [RFC/PATCH 1/2] stackfs: generic functions for obtaining hidden object

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

Message-Id: <200610131543.k9DFh05m016578@agora.fsl.cs.sunysb.edu>
Date:	Fri, 13 Oct 2006 11:43:00 -0400
From:	Erez Zadok <ezk@...sunysb.edu>
To:	Pekka J Enberg <penberg@...helsinki.fi>
Cc:	linux-kernel@...r.kernel.org, linux-fsdevel@...r.kernel.org,
	jsipek@...sunysb.edu, ezk@...sunysb.edu, mhalcrow@...ibm.com,
	phillip@...lewell.homeip.net
Subject: Re: [RFC/PATCH 1/2] stackfs: generic functions for obtaining hidden object 

In message <Pine.LNX.4.64.0610131615370.563@...-30.cs.Helsinki.FI>, Pekka J Enberg writes:
> From: Pekka Enberg <penberg@...helsinki.fi>
> 
> Add generic functions for obtaining the hidden object (superblock, inode,
> dentry, and dentry mount-point) in a stacked filesystem.  As fan-out 
> stacked filesystems have multiple hidden objects, we store them in a 
> statically allocated array of pointers.  The current hard-coded limit 
> STACKFS_MAX_BRANCHES is not enough for unionfs (for which users have more 
> than 100 branches).  That, however, can be fixed later for unionfs.

I think we should do it right the first time (i.e., now :-)

> +#define STACKFS_MAX_BRANCHES (8)

> +struct stackfs_sb_info {
> +	struct super_block *hidden_sbs[STACKFS_MAX_BRANCHES];
> +};

Why not make it something more dynamic, such as a mount-time option per sb?
Even at 8, you waste most of that space for non-fan-out stackable file
systems such as ecryptfs; and those unionfs users who want more, will have
to _recompile_ the code.

And given that this code is shared, if just one f/s needs 100 branches, why
should ecryptfs waste 99*sizeof(pointer) bytes for every *_info structure?

> +static inline struct super_block *
> +__stackfs_hidden_sb(struct super_block *sb, unsigned long branch_idx)
> +{
> +	struct stackfs_sb_info *info = sb->s_fs_info;
> +	return info->hidden_sbs[branch_idx];
> +}

Also, the functions don't check array bounds.  Where, if at all, this gets
checked against the STACKFS_MAX_BRANCHES value?  Shouldn't we at least put
some ASSERT's in there to catch bugs?

Of course, I realize that the above code is rather simple now and changing
it as I propose will require kmalloc/kfree (or containers) carefully used
throughout.

Thanks,
Erez.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/