linux-ext4 - Review of ext4-online-defrag-move-victim-files.patch

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 for Android: free password hash cracker in your pocket

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

Date:	Sat, 13 Sep 2008 17:16:11 -0400
From:	"Theodore Ts'o" <tytso@....EDU>
To:	a-fujita@...jp.nec.com
cc:	Takashi Sato <t-sato@...jp.nec.com>, linux-ext4@...r.kernel.org
Subject: Review of ext4-online-defrag-move-victim-files.patch

Hi,

I've been looking over the defrag patches with an eye towards getting
them merged.  This patch concerns me for a number of reasons.  First of
all, it is calling into a number of private functions which had
previously been static functions in fs/ext4/balloc.c.  Specifically,
goal_in_my_reservation(), rsv_window_remove(), rsv_is_empty(),
alloc_new_reservation(), try_to_extend_reservation().  

This is bad for a couple of reasons.  First of all, the functions
weren't renamed, so it results in the namespace leakage.  In general,
non-static functions should be prefixed by ext4_ so that we know they
came from the ext4 filesystem code.  Secondly, these were internal
functions were intended for use in an older set of block allocation
functions that may be removed in the future --- they had previously only
been used by the function ext4_old_new_blocks(), which is used only when
the mount option nomballoc is given.  Given the superiority of the new
multi-block allocator, it's likely that this old code will be going
away.

One of the things which further worries me is that your patch seems to
be making changes to the mballoc() code as well.  Given that the
reservations code in fs/ext4/balloc.c was never intended to used at the
same time as the multi-block allocator code in mballoc(), I suspect
there will be a problem here if the goal of reserving blocks using the
reservation code was to prevent some other inode using allocating those
blocks, since the multi-block allocator does not honor reservations made
by the reservations code, since normally the reservations code is not
active when the mballoc code is active (the two are mutually exclusive).

Best regards,

						- Ted

--
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